diff options
Diffstat (limited to 'lib')
176 files changed, 52551 insertions, 0 deletions
diff --git a/lib/_Noreturn.h b/lib/_Noreturn.h new file mode 100644 index 0000000..94fdfaf --- /dev/null +++ b/lib/_Noreturn.h @@ -0,0 +1,14 @@ +#ifndef _Noreturn +# if 201103 <= (defined __cplusplus ? __cplusplus : 0) +# define _Noreturn [[noreturn]] +# elif (201112 <= (defined __STDC_VERSION__ ? __STDC_VERSION__ : 0) \ + || 4 < __GNUC__ + (7 <= __GNUC_MINOR__)) + /* _Noreturn works as-is. */ +# elif 2 < __GNUC__ + (8 <= __GNUC_MINOR__) || 0x5110 <= __SUNPRO_C +# define _Noreturn __attribute__ ((__noreturn__)) +# elif 1200 <= (defined _MSC_VER ? _MSC_VER : 0) +# define _Noreturn __declspec (noreturn) +# else +# define _Noreturn +# endif +#endif diff --git a/lib/acl-errno-valid.c b/lib/acl-errno-valid.c new file mode 100644 index 0000000..96577eb --- /dev/null +++ b/lib/acl-errno-valid.c @@ -0,0 +1,52 @@ +/* Test whether ACLs are well supported on this system. + + Copyright 2013-2018 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. + + Written by Paul Eggert. */ + +#include <config.h> + +#include <acl.h> + +#include <errno.h> + +/* Return true if errno value ERRNUM indicates that ACLs are well + supported on this system. ERRNUM should be an errno value obtained + after an ACL-related system call fails. */ +bool +acl_errno_valid (int errnum) +{ + /* Recognize some common errors such as from an NFS mount that does + not support ACLs, even when local drives do. */ + switch (errnum) + { + case EBUSY: return false; + case EINVAL: return false; +#if defined __APPLE__ && defined __MACH__ + case ENOENT: return false; +#endif + case ENOSYS: return false; + +#if defined ENOTSUP && ENOTSUP != EOPNOTSUPP +# if ENOTSUP != ENOSYS /* Needed for the MS-Windows port of GNU Emacs. */ + case ENOTSUP: return false; +# endif +#endif + + case EOPNOTSUPP: return false; + default: return true; + } +} diff --git a/lib/acl-internal.c b/lib/acl-internal.c new file mode 100644 index 0000000..92e7b9b --- /dev/null +++ b/lib/acl-internal.c @@ -0,0 +1,507 @@ +/* Test whether a file has a nontrivial ACL. -*- coding: utf-8 -*- + + Copyright (C) 2002-2003, 2005-2018 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. + + Written by Paul Eggert, Andreas Grünbacher, and Bruno Haible. */ + +#include <config.h> + +#include "acl.h" + +#include "acl-internal.h" + +#if USE_ACL && HAVE_ACL_GET_FILE /* Linux, FreeBSD, Mac OS X, IRIX, Tru64, Cygwin >= 2.5 */ + +# if HAVE_ACL_TYPE_EXTENDED /* Mac OS X */ + +/* ACL is an ACL, from a file, stored as type ACL_TYPE_EXTENDED. + Return 1 if the given ACL is non-trivial. + Return 0 if it is trivial. */ +int +acl_extended_nontrivial (acl_t acl) +{ + /* acl is non-trivial if it is non-empty. */ + return (acl_entries (acl) > 0); +} + +# else /* Linux, FreeBSD, IRIX, Tru64, Cygwin >= 2.5 */ + +/* ACL is an ACL, from a file, stored as type ACL_TYPE_ACCESS. + Return 1 if the given ACL is non-trivial. + Return 0 if it is trivial, i.e. equivalent to a simple stat() mode. + Return -1 and set errno upon failure to determine it. */ +int +acl_access_nontrivial (acl_t acl) +{ + /* acl is non-trivial if it has some entries other than for "user::", + "group::", and "other::". Normally these three should be present + at least, allowing us to write + return (3 < acl_entries (acl)); + but the following code is more robust. */ +# if HAVE_ACL_FIRST_ENTRY /* Linux, FreeBSD, Cygwin >= 2.5 */ + + acl_entry_t ace; + int got_one; + + for (got_one = acl_get_entry (acl, ACL_FIRST_ENTRY, &ace); + got_one > 0; + got_one = acl_get_entry (acl, ACL_NEXT_ENTRY, &ace)) + { + acl_tag_t tag; + if (acl_get_tag_type (ace, &tag) < 0) + return -1; + if (!(tag == ACL_USER_OBJ || tag == ACL_GROUP_OBJ || tag == ACL_OTHER)) + return 1; + } + return got_one; + +# elif HAVE_ACL_TO_SHORT_TEXT /* IRIX */ + /* Don't use acl_get_entry: it is undocumented. */ + + int count = acl->acl_cnt; + int i; + + for (i = 0; i < count; i++) + { + acl_entry_t ace = &acl->acl_entry[i]; + acl_tag_t tag = ace->ae_tag; + + if (!(tag == ACL_USER_OBJ || tag == ACL_GROUP_OBJ + || tag == ACL_OTHER_OBJ)) + return 1; + } + return 0; + +# elif HAVE_ACL_FREE_TEXT /* Tru64 */ + /* Don't use acl_get_entry: it takes only one argument and does not work. */ + + int count = acl->acl_num; + acl_entry_t ace; + + for (ace = acl->acl_first; count > 0; ace = ace->next, count--) + { + acl_tag_t tag; + acl_perm_t perm; + + tag = ace->entry->acl_type; + if (!(tag == ACL_USER_OBJ || tag == ACL_GROUP_OBJ || tag == ACL_OTHER)) + return 1; + + perm = ace->entry->acl_perm; + /* On Tru64, perm can also contain non-standard bits such as + PERM_INSERT, PERM_DELETE, PERM_MODIFY, PERM_LOOKUP, ... */ + if ((perm & ~(ACL_READ | ACL_WRITE | ACL_EXECUTE)) != 0) + return 1; + } + return 0; + +# else + + errno = ENOSYS; + return -1; +# endif +} + +int +acl_default_nontrivial (acl_t acl) +{ + /* acl is non-trivial if it is non-empty. */ + return (acl_entries (acl) > 0); +} + +# endif + +#elif USE_ACL && HAVE_FACL && defined GETACL /* Solaris, Cygwin < 2.5, not HP-UX */ + +/* Test an ACL retrieved with GETACL. + Return 1 if the given ACL, consisting of COUNT entries, is non-trivial. + Return 0 if it is trivial, i.e. equivalent to a simple stat() mode. */ +int +acl_nontrivial (int count, aclent_t *entries) +{ + int i; + + for (i = 0; i < count; i++) + { + aclent_t *ace = &entries[i]; + + /* Note: If ace->a_type = USER_OBJ, ace->a_id is the st_uid from stat(). + If ace->a_type = GROUP_OBJ, ace->a_id is the st_gid from stat(). + We don't need to check ace->a_id in these cases. */ + if (!(ace->a_type == USER_OBJ + || ace->a_type == GROUP_OBJ + || ace->a_type == OTHER_OBJ + /* Note: Cygwin does not return a CLASS_OBJ ("mask:") entry + sometimes. */ + || ace->a_type == CLASS_OBJ)) + return 1; + } + return 0; +} + +# ifdef ACE_GETACL + +/* A shortcut for a bitmask. */ +# define NEW_ACE_WRITEA_DATA (NEW_ACE_WRITE_DATA | NEW_ACE_APPEND_DATA) + +/* Test an ACL retrieved with ACE_GETACL. + Return 1 if the given ACL, consisting of COUNT entries, is non-trivial. + Return 0 if it is trivial, i.e. equivalent to a simple stat() mode. */ +int +acl_ace_nontrivial (int count, ace_t *entries) +{ + int i; + + /* The flags in the ace_t structure changed in a binary incompatible way + when ACL_NO_TRIVIAL etc. were introduced in <sys/acl.h> version 1.15. + How to distinguish the two conventions at runtime? + In the old convention, usually three ACEs have a_flags = ACE_OWNER / + ACE_GROUP / ACE_OTHER, in the range 0x0100..0x0400. In the new + convention, these values are not used. */ + int old_convention = 0; + + for (i = 0; i < count; i++) + if (entries[i].a_flags & (OLD_ACE_OWNER | OLD_ACE_GROUP | OLD_ACE_OTHER)) + { + old_convention = 1; + break; + } + + if (old_convention) + /* Running on Solaris 10. */ + for (i = 0; i < count; i++) + { + ace_t *ace = &entries[i]; + + /* Note: + If ace->a_flags = ACE_OWNER, ace->a_who is the st_uid from stat(). + If ace->a_flags = ACE_GROUP, ace->a_who is the st_gid from stat(). + We don't need to check ace->a_who in these cases. */ + if (!(ace->a_type == OLD_ALLOW + && (ace->a_flags == OLD_ACE_OWNER + || ace->a_flags == OLD_ACE_GROUP + || ace->a_flags == OLD_ACE_OTHER))) + return 1; + } + else + { + /* Running on Solaris 10 (newer version) or Solaris 11. */ + unsigned int access_masks[6] = + { + 0, /* owner@ deny */ + 0, /* owner@ allow */ + 0, /* group@ deny */ + 0, /* group@ allow */ + 0, /* everyone@ deny */ + 0 /* everyone@ allow */ + }; + + for (i = 0; i < count; i++) + { + ace_t *ace = &entries[i]; + unsigned int index1; + unsigned int index2; + + if (ace->a_type == NEW_ACE_ACCESS_ALLOWED_ACE_TYPE) + index1 = 1; + else if (ace->a_type == NEW_ACE_ACCESS_DENIED_ACE_TYPE) + index1 = 0; + else + return 1; + + if (ace->a_flags == NEW_ACE_OWNER) + index2 = 0; + else if (ace->a_flags == (NEW_ACE_GROUP | NEW_ACE_IDENTIFIER_GROUP)) + index2 = 2; + else if (ace->a_flags == NEW_ACE_EVERYONE) + index2 = 4; + else + return 1; + + access_masks[index1 + index2] |= ace->a_access_mask; + } + + /* The same bit shouldn't be both allowed and denied. */ + if (access_masks[0] & access_masks[1]) + return 1; + if (access_masks[2] & access_masks[3]) + return 1; + if (access_masks[4] & access_masks[5]) + return 1; + + /* Check minimum masks. */ + if ((NEW_ACE_WRITE_NAMED_ATTRS + | NEW_ACE_WRITE_ATTRIBUTES + | NEW_ACE_WRITE_ACL + | NEW_ACE_WRITE_OWNER) + & ~ access_masks[1]) + return 1; + access_masks[1] &= ~(NEW_ACE_WRITE_NAMED_ATTRS + | NEW_ACE_WRITE_ATTRIBUTES + | NEW_ACE_WRITE_ACL + | NEW_ACE_WRITE_OWNER); + if ((NEW_ACE_READ_NAMED_ATTRS + | NEW_ACE_READ_ATTRIBUTES + | NEW_ACE_READ_ACL + | NEW_ACE_SYNCHRONIZE) + & ~ access_masks[5]) + return 1; + access_masks[5] &= ~(NEW_ACE_READ_NAMED_ATTRS + | NEW_ACE_READ_ATTRIBUTES + | NEW_ACE_READ_ACL + | NEW_ACE_SYNCHRONIZE); + + /* Check the allowed or denied bits. */ + switch ((access_masks[0] | access_masks[1]) + & ~(NEW_ACE_READ_NAMED_ATTRS + | NEW_ACE_READ_ATTRIBUTES + | NEW_ACE_READ_ACL + | NEW_ACE_SYNCHRONIZE)) + { + case 0: + case NEW_ACE_READ_DATA: + case NEW_ACE_WRITEA_DATA: + case NEW_ACE_READ_DATA | NEW_ACE_WRITEA_DATA: + case NEW_ACE_EXECUTE: + case NEW_ACE_READ_DATA | NEW_ACE_EXECUTE: + case NEW_ACE_WRITEA_DATA | NEW_ACE_EXECUTE: + case NEW_ACE_READ_DATA | NEW_ACE_WRITEA_DATA | NEW_ACE_EXECUTE: + break; + default: + return 1; + } + switch ((access_masks[2] | access_masks[3]) + & ~(NEW_ACE_READ_NAMED_ATTRS + | NEW_ACE_READ_ATTRIBUTES + | NEW_ACE_READ_ACL + | NEW_ACE_SYNCHRONIZE)) + { + case 0: + case NEW_ACE_READ_DATA: + case NEW_ACE_WRITEA_DATA: + case NEW_ACE_READ_DATA | NEW_ACE_WRITEA_DATA: + case NEW_ACE_EXECUTE: + case NEW_ACE_READ_DATA | NEW_ACE_EXECUTE: + case NEW_ACE_WRITEA_DATA | NEW_ACE_EXECUTE: + case NEW_ACE_READ_DATA | NEW_ACE_WRITEA_DATA | NEW_ACE_EXECUTE: + break; + default: + return 1; + } + switch ((access_masks[4] | access_masks[5]) + & ~(NEW_ACE_WRITE_NAMED_ATTRS + | NEW_ACE_WRITE_ATTRIBUTES + | NEW_ACE_WRITE_ACL + | NEW_ACE_WRITE_OWNER)) + { + case 0: + case NEW_ACE_READ_DATA: + case NEW_ACE_WRITEA_DATA: + case NEW_ACE_READ_DATA | NEW_ACE_WRITEA_DATA: + case NEW_ACE_EXECUTE: + case NEW_ACE_READ_DATA | NEW_ACE_EXECUTE: + case NEW_ACE_WRITEA_DATA | NEW_ACE_EXECUTE: + case NEW_ACE_READ_DATA | NEW_ACE_WRITEA_DATA | NEW_ACE_EXECUTE: + break; + default: + return 1; + } + + /* Check that the NEW_ACE_WRITE_DATA and NEW_ACE_APPEND_DATA bits are + either both allowed or both denied. */ + if (((access_masks[0] & NEW_ACE_WRITE_DATA) != 0) + != ((access_masks[0] & NEW_ACE_APPEND_DATA) != 0)) + return 1; + if (((access_masks[2] & NEW_ACE_WRITE_DATA) != 0) + != ((access_masks[2] & NEW_ACE_APPEND_DATA) != 0)) + return 1; + if (((access_masks[4] & NEW_ACE_WRITE_DATA) != 0) + != ((access_masks[4] & NEW_ACE_APPEND_DATA) != 0)) + return 1; + } + + return 0; +} + +# endif + +#elif USE_ACL && HAVE_GETACL /* HP-UX */ + +/* Return 1 if the given ACL is non-trivial. + Return 0 if it is trivial, i.e. equivalent to a simple stat() mode. */ +int +acl_nontrivial (int count, struct acl_entry *entries) +{ + int i; + + if (count > 3) + return 1; + + for (i = 0; i < count; i++) + { + struct acl_entry *ace = &entries[i]; + + if (ace->uid != ACL_NSUSER && ace->gid != ACL_NSGROUP) + return 1; + } + return 0; +} + +# if HAVE_ACLV_H /* HP-UX >= 11.11 */ + +/* Return 1 if the given ACL is non-trivial. + Return 0 if it is trivial, i.e. equivalent to a simple stat() mode. */ +int +aclv_nontrivial (int count, struct acl *entries) +{ + int i; + + for (i = 0; i < count; i++) + { + struct acl *ace = &entries[i]; + + /* Note: If ace->a_type = USER_OBJ, ace->a_id is the st_uid from stat(). + If ace->a_type = GROUP_OBJ, ace->a_id is the st_gid from stat(). + We don't need to check ace->a_id in these cases. */ + if (!(ace->a_type == USER_OBJ /* no need to check ace->a_id here */ + || ace->a_type == GROUP_OBJ /* no need to check ace->a_id here */ + || ace->a_type == CLASS_OBJ + || ace->a_type == OTHER_OBJ)) + return 1; + } + return 0; +} + +# endif + +#elif USE_ACL && (HAVE_ACLX_GET || HAVE_STATACL) /* AIX */ + +/* Return 1 if the given ACL is non-trivial. + Return 0 if it is trivial, i.e. equivalent to a simple stat() mode. */ +int +acl_nontrivial (struct acl *a) +{ + /* The normal way to iterate through an ACL is like this: + struct acl_entry *ace; + for (ace = a->acl_ext; ace != acl_last (a); ace = acl_nxt (ace)) + { + struct ace_id *aei; + switch (ace->ace_type) + { + case ACC_PERMIT: + case ACC_DENY: + case ACC_SPECIFY: + ...; + } + for (aei = ace->ace_id; aei != id_last (ace); aei = id_nxt (aei)) + ... + } + */ + return (acl_last (a) != a->acl_ext ? 1 : 0); +} + +# if HAVE_ACLX_GET && defined ACL_AIX_WIP /* newer AIX */ + +/* Return 1 if the given ACL is non-trivial. + Return 0 if it is trivial, i.e. equivalent to a simple stat() mode. */ +int +acl_nfs4_nontrivial (nfs4_acl_int_t *a) +{ +# if 1 /* let's try this first */ + return (a->aclEntryN > 0 ? 1 : 0); +# else + int count = a->aclEntryN; + int i; + + for (i = 0; i < count; i++) + { + nfs4_ace_int_t *ace = &a->aclEntry[i]; + + if (!((ace->flags & ACE4_ID_SPECIAL) != 0 + && (ace->aceWho.special_whoid == ACE4_WHO_OWNER + || ace->aceWho.special_whoid == ACE4_WHO_GROUP + || ace->aceWho.special_whoid == ACE4_WHO_EVERYONE) + && ace->aceType == ACE4_ACCESS_ALLOWED_ACE_TYPE + && ace->aceFlags == 0 + && (ace->aceMask & ~(ACE4_READ_DATA | ACE4_LIST_DIRECTORY + | ACE4_WRITE_DATA | ACE4_ADD_FILE + | ACE4_EXECUTE)) == 0)) + return 1; + } + return 0; +# endif +} + +# endif + +#elif USE_ACL && HAVE_ACLSORT /* NonStop Kernel */ + +/* Test an ACL retrieved with ACL_GET. + Return 1 if the given ACL, consisting of COUNT entries, is non-trivial. + Return 0 if it is trivial, i.e. equivalent to a simple stat() mode. */ +int +acl_nontrivial (int count, struct acl *entries) +{ + int i; + + for (i = 0; i < count; i++) + { + struct acl *ace = &entries[i]; + + /* Note: If ace->a_type = USER_OBJ, ace->a_id is the st_uid from stat(). + If ace->a_type = GROUP_OBJ, ace->a_id is the st_gid from stat(). + We don't need to check ace->a_id in these cases. */ + if (!(ace->a_type == USER_OBJ /* no need to check ace->a_id here */ + || ace->a_type == GROUP_OBJ /* no need to check ace->a_id here */ + || ace->a_type == CLASS_OBJ + || ace->a_type == OTHER_OBJ)) + return 1; + } + return 0; +} + +#endif + +void +free_permission_context (struct permission_context *ctx) +{ +#if USE_ACL +# if HAVE_ACL_GET_FILE /* Linux, FreeBSD, Mac OS X, IRIX, Tru64, Cygwin >= 2.5 */ + if (ctx->acl) + acl_free (ctx->acl); +# if !HAVE_ACL_TYPE_EXTENDED + if (ctx->default_acl) + acl_free (ctx->default_acl); +# endif + +# elif defined GETACL /* Solaris, Cygwin < 2.5 */ + free (ctx->entries); +# ifdef ACE_GETACL + free (ctx->ace_entries); +# endif + +# elif HAVE_GETACL /* HP-UX */ + +# if HAVE_ACLV_H +# endif + +# elif HAVE_STATACL /* older AIX */ + +# elif HAVE_ACLSORT /* NonStop Kernel */ + +# endif +#endif +} diff --git a/lib/acl-internal.h b/lib/acl-internal.h new file mode 100644 index 0000000..2da7c5a --- /dev/null +++ b/lib/acl-internal.h @@ -0,0 +1,302 @@ +/* Internal implementation of access control lists. -*- coding: utf-8 -*- + + Copyright (C) 2002-2003, 2005-2018 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. + + Written by Paul Eggert, Andreas Grünbacher, and Bruno Haible. */ + +#include "acl.h" + +#include <stdbool.h> +#include <stdlib.h> + +/* All systems define the ACL related API in <sys/acl.h>. */ +#if HAVE_SYS_ACL_H +# include <sys/acl.h> +#endif +#if defined HAVE_FACL && ! defined GETACLCNT && defined ACL_CNT +# define GETACLCNT ACL_CNT +#endif + +/* On Linux and Cygwin >= 2.5, additional ACL related API is available in + <acl/libacl.h>. */ +#ifdef HAVE_ACL_LIBACL_H +# include <acl/libacl.h> +#endif + +/* On HP-UX >= 11.11, additional ACL API is available in <aclv.h>. */ +#if HAVE_ACLV_H +# include <sys/types.h> +# include <aclv.h> +/* HP-UX 11.11 lacks these declarations. */ +extern int acl (char *, int, int, struct acl *); +extern int aclsort (int, int, struct acl *); +#endif + +#include <errno.h> + +#include <limits.h> +#ifndef MIN +# define MIN(a,b) ((a) < (b) ? (a) : (b)) +#endif + +#ifndef SIZE_MAX +# define SIZE_MAX ((size_t) -1) +#endif + +#ifndef HAVE_FCHMOD +# define HAVE_FCHMOD false +# define fchmod(fd, mode) (-1) +#endif + +#ifndef _GL_INLINE_HEADER_BEGIN + #error "Please include config.h first." +#endif +_GL_INLINE_HEADER_BEGIN +#ifndef ACL_INTERNAL_INLINE +# define ACL_INTERNAL_INLINE _GL_INLINE +#endif + +#if USE_ACL + +# if HAVE_ACL_GET_FILE +/* POSIX 1003.1e (draft 17 -- abandoned) specific version. */ +/* Linux, FreeBSD, Mac OS X, IRIX, Tru64, Cygwin >= 2.5 */ + +# ifndef MIN_ACL_ENTRIES +# define MIN_ACL_ENTRIES 4 +# endif + +/* POSIX 1003.1e (draft 17) */ +# ifdef HAVE_ACL_GET_FD +/* Most platforms have a 1-argument acl_get_fd, only OSF/1 has a 2-argument + macro(!). */ +# if HAVE_ACL_FREE_TEXT /* OSF/1 */ +ACL_INTERNAL_INLINE acl_t +rpl_acl_get_fd (int fd) +{ + return acl_get_fd (fd, ACL_TYPE_ACCESS); +} +# undef acl_get_fd +# define acl_get_fd rpl_acl_get_fd +# endif +# else +# define HAVE_ACL_GET_FD false +# undef acl_get_fd +# define acl_get_fd(fd) (NULL) +# endif + +/* POSIX 1003.1e (draft 17) */ +# ifdef HAVE_ACL_SET_FD +/* Most platforms have a 2-argument acl_set_fd, only OSF/1 has a 3-argument + macro(!). */ +# if HAVE_ACL_FREE_TEXT /* OSF/1 */ +ACL_INTERNAL_INLINE int +rpl_acl_set_fd (int fd, acl_t acl) +{ + return acl_set_fd (fd, ACL_TYPE_ACCESS, acl); +} +# undef acl_set_fd +# define acl_set_fd rpl_acl_set_fd +# endif +# else +# define HAVE_ACL_SET_FD false +# undef acl_set_fd +# define acl_set_fd(fd, acl) (-1) +# endif + +/* POSIX 1003.1e (draft 13) */ +# if ! HAVE_ACL_FREE_TEXT +# define acl_free_text(buf) acl_free (buf) +# endif + +/* Linux-specific */ +/* Cygwin >= 2.5 implements this function, but it returns 1 for all + directories, thus is unusable. */ +# if !defined HAVE_ACL_EXTENDED_FILE || defined __CYGWIN__ +# undef HAVE_ACL_EXTENDED_FILE +# define HAVE_ACL_EXTENDED_FILE false +# define acl_extended_file(name) (-1) +# endif + +# if ! defined HAVE_ACL_FROM_MODE && ! defined HAVE_ACL_FROM_TEXT +# define acl_from_mode (NULL) +# endif + +/* Set to 0 if a file's mode is stored independently from the ACL. */ +# if (HAVE_ACL_COPY_EXT_NATIVE && HAVE_ACL_CREATE_ENTRY_NP) || defined __sgi /* Mac OS X, IRIX */ +# define MODE_INSIDE_ACL 0 +# endif + +/* Return the number of entries in ACL. + Return -1 and set errno upon failure to determine it. */ +/* Define a replacement for acl_entries if needed. (Only Linux has it.) */ +# if !HAVE_ACL_ENTRIES +# define acl_entries rpl_acl_entries +extern int acl_entries (acl_t); +# endif + +# if HAVE_ACL_TYPE_EXTENDED /* Mac OS X */ +/* ACL is an ACL, from a file, stored as type ACL_TYPE_EXTENDED. + Return 1 if the given ACL is non-trivial. + Return 0 if it is trivial. */ +extern int acl_extended_nontrivial (acl_t); +# else +/* ACL is an ACL, from a file, stored as type ACL_TYPE_ACCESS. + Return 1 if the given ACL is non-trivial. + Return 0 if it is trivial, i.e. equivalent to a simple stat() mode. + Return -1 and set errno upon failure to determine it. */ +extern int acl_access_nontrivial (acl_t); + +/* ACL is an ACL, from a file, stored as type ACL_TYPE_DEFAULT. + Return 1 if the given ACL is non-trivial. + Return 0 if it is trivial, i.e. equivalent to a simple stat() mode. + Return -1 and set errno upon failure to determine it. */ +extern int acl_default_nontrivial (acl_t); +# endif + +# elif HAVE_FACL && defined GETACL /* Solaris, Cygwin < 2.5, not HP-UX */ + +/* Set to 0 if a file's mode is stored independently from the ACL. */ +# if defined __CYGWIN__ /* Cygwin */ +# define MODE_INSIDE_ACL 0 +# endif + +/* Return 1 if the given ACL is non-trivial. + Return 0 if it is trivial, i.e. equivalent to a simple stat() mode. */ +extern int acl_nontrivial (int count, aclent_t *entries) _GL_ATTRIBUTE_PURE; + +# ifdef ACE_GETACL /* Solaris 10 */ + +/* Test an ACL retrieved with ACE_GETACL. + Return 1 if the given ACL, consisting of COUNT entries, is non-trivial. + Return 0 if it is trivial, i.e. equivalent to a simple stat() mode. */ +extern int acl_ace_nontrivial (int count, ace_t *entries) _GL_ATTRIBUTE_PURE; + +/* Definitions for when the built executable is executed on Solaris 10 + (newer version) or Solaris 11. */ +/* For a_type. */ +# define OLD_ALLOW 0 +# define OLD_DENY 1 +# define NEW_ACE_ACCESS_ALLOWED_ACE_TYPE 0 /* replaces ALLOW */ +# define NEW_ACE_ACCESS_DENIED_ACE_TYPE 1 /* replaces DENY */ +/* For a_flags. */ +# define OLD_ACE_OWNER 0x0100 +# define OLD_ACE_GROUP 0x0200 +# define OLD_ACE_OTHER 0x0400 +# define NEW_ACE_OWNER 0x1000 +# define NEW_ACE_GROUP 0x2000 +# define NEW_ACE_IDENTIFIER_GROUP 0x0040 +# define NEW_ACE_EVERYONE 0x4000 +/* For a_access_mask. */ +# define NEW_ACE_READ_DATA 0x001 /* corresponds to 'r' */ +# define NEW_ACE_WRITE_DATA 0x002 /* corresponds to 'w' */ +# define NEW_ACE_APPEND_DATA 0x004 +# define NEW_ACE_READ_NAMED_ATTRS 0x008 +# define NEW_ACE_WRITE_NAMED_ATTRS 0x010 +# define NEW_ACE_EXECUTE 0x020 +# define NEW_ACE_DELETE_CHILD 0x040 +# define NEW_ACE_READ_ATTRIBUTES 0x080 +# define NEW_ACE_WRITE_ATTRIBUTES 0x100 +# define NEW_ACE_DELETE 0x10000 +# define NEW_ACE_READ_ACL 0x20000 +# define NEW_ACE_WRITE_ACL 0x40000 +# define NEW_ACE_WRITE_OWNER 0x80000 +# define NEW_ACE_SYNCHRONIZE 0x100000 + +# endif + +# elif HAVE_GETACL /* HP-UX */ + +/* Return 1 if the given ACL is non-trivial. + Return 0 if it is trivial, i.e. equivalent to a simple stat() mode. */ +extern int acl_nontrivial (int count, struct acl_entry *entries); + +# if HAVE_ACLV_H /* HP-UX >= 11.11 */ + +/* Return 1 if the given ACL is non-trivial. + Return 0 if it is trivial, i.e. equivalent to a simple stat() mode. */ +extern int aclv_nontrivial (int count, struct acl *entries); + +# endif + +# elif HAVE_ACLX_GET && 0 /* AIX */ + +/* TODO */ + +# elif HAVE_STATACL /* older AIX */ + +/* Return 1 if the given ACL is non-trivial. + Return 0 if it is trivial, i.e. equivalent to a simple stat() mode. */ +extern int acl_nontrivial (struct acl *a); + +# elif HAVE_ACLSORT /* NonStop Kernel */ + +/* Return 1 if the given ACL is non-trivial. + Return 0 if it is trivial, i.e. equivalent to a simple stat() mode. */ +extern int acl_nontrivial (int count, struct acl *entries); + +# endif + +/* Set to 1 if a file's mode is implicit by the ACL. */ +# ifndef MODE_INSIDE_ACL +# define MODE_INSIDE_ACL 1 +# endif + +#endif + +struct permission_context { + mode_t mode; +#if USE_ACL +# if HAVE_ACL_GET_FILE /* Linux, FreeBSD, Mac OS X, IRIX, Tru64, Cygwin >= 2.5 */ + acl_t acl; +# if !HAVE_ACL_TYPE_EXTENDED + acl_t default_acl; +# endif + bool acls_not_supported; + +# elif defined GETACL /* Solaris, Cygwin < 2.5 */ + int count; + aclent_t *entries; +# ifdef ACE_GETACL + int ace_count; + ace_t *ace_entries; +# endif + +# elif HAVE_GETACL /* HP-UX */ + struct acl_entry entries[NACLENTRIES]; + int count; +# if HAVE_ACLV_H + struct acl aclv_entries[NACLVENTRIES]; + int aclv_count; +# endif + +# elif HAVE_STATACL /* older AIX */ + union { struct acl a; char room[4096]; } u; + bool have_u; + +# elif HAVE_ACLSORT /* NonStop Kernel */ + struct acl entries[NACLENTRIES]; + int count; + +# endif +#endif +}; + +int get_permissions (const char *, int, mode_t, struct permission_context *); +int set_permissions (struct permission_context *, const char *, int); +void free_permission_context (struct permission_context *); + +_GL_INLINE_HEADER_END diff --git a/lib/acl.h b/lib/acl.h new file mode 100644 index 0000000..d0c122c --- /dev/null +++ b/lib/acl.h @@ -0,0 +1,35 @@ +/* acl.c - access control lists + + Copyright (C) 2002, 2008-2018 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. + + Written by Paul Eggert. */ + +#ifndef _GL_ACL_H +#define _GL_ACL_H 1 + +#include <stdbool.h> +#include <sys/types.h> +#include <sys/stat.h> + +bool acl_errno_valid (int) _GL_ATTRIBUTE_CONST; +int file_has_acl (char const *, struct stat const *); +int qset_acl (char const *, int, mode_t); +int set_acl (char const *, int, mode_t); +int qcopy_acl (char const *, int, char const *, int, mode_t); +int copy_acl (char const *, int, char const *, int, mode_t); +int chmod_or_fchmod (char const *, int, mode_t); + +#endif diff --git a/lib/acl_entries.c b/lib/acl_entries.c new file mode 100644 index 0000000..ce730d4 --- /dev/null +++ b/lib/acl_entries.c @@ -0,0 +1,75 @@ +/* Return the number of entries in an ACL. + + Copyright (C) 2002-2003, 2005-2018 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. + + Written by Paul Eggert and Andreas Gruenbacher. */ + +#include <config.h> + +#include "acl-internal.h" + +/* This file assumes POSIX-draft like ACLs + (Linux, FreeBSD, Mac OS X, IRIX, Tru64, Cygwin >= 2.5). */ + +/* Return the number of entries in ACL. + Return -1 and set errno upon failure to determine it. */ + +int +acl_entries (acl_t acl) +{ + int count = 0; + + if (acl != NULL) + { +#if HAVE_ACL_FIRST_ENTRY /* Linux, FreeBSD, Mac OS X, Cygwin >= 2.5 */ +# if HAVE_ACL_TYPE_EXTENDED /* Mac OS X */ + /* acl_get_entry returns 0 when it successfully fetches an entry, + and -1/EINVAL at the end. */ + acl_entry_t ace; + int got_one; + + for (got_one = acl_get_entry (acl, ACL_FIRST_ENTRY, &ace); + got_one >= 0; + got_one = acl_get_entry (acl, ACL_NEXT_ENTRY, &ace)) + count++; +# else /* Linux, FreeBSD, Cygwin >= 2.5 */ + /* acl_get_entry returns 1 when it successfully fetches an entry, + and 0 at the end. */ + acl_entry_t ace; + int got_one; + + for (got_one = acl_get_entry (acl, ACL_FIRST_ENTRY, &ace); + got_one > 0; + got_one = acl_get_entry (acl, ACL_NEXT_ENTRY, &ace)) + count++; + if (got_one < 0) + return -1; +# endif +#else /* IRIX, Tru64 */ +# if HAVE_ACL_TO_SHORT_TEXT /* IRIX */ + /* Don't use acl_get_entry: it is undocumented. */ + count = acl->acl_cnt; +# endif +# if HAVE_ACL_FREE_TEXT /* Tru64 */ + /* Don't use acl_get_entry: it takes only one argument and does not + work. */ + count = acl->acl_num; +# endif +#endif + } + + return count; +} diff --git a/lib/alignof.h b/lib/alignof.h new file mode 100644 index 0000000..26d71dd --- /dev/null +++ b/lib/alignof.h @@ -0,0 +1,50 @@ +/* Determine alignment of types. + Copyright (C) 2003-2004, 2006, 2009-2018 Free Software Foundation, Inc. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3, or (at your option) + any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, see <https://www.gnu.org/licenses/>. */ + +#ifndef _ALIGNOF_H +#define _ALIGNOF_H + +#include <stddef.h> + +/* alignof_slot (TYPE) + Determine the alignment of a structure slot (field) of a given type, + at compile time. Note that the result depends on the ABI. + This is the same as alignof (TYPE) and _Alignof (TYPE), defined in + <stdalign.h> if __alignof_is_defined is 1. + Note: The result cannot be used as a value for an 'enum' constant, + due to bugs in HP-UX 10.20 cc and AIX 3.2.5 xlc. */ +#if defined __cplusplus + template <class type> struct alignof_helper { char __slot1; type __slot2; }; +# define alignof_slot(type) offsetof (alignof_helper<type>, __slot2) +#else +# define alignof_slot(type) offsetof (struct { char __slot1; type __slot2; }, __slot2) +#endif + +/* alignof_type (TYPE) + Determine the good alignment of an object of the given type at compile time. + Note that this is not necessarily the same as alignof_slot(type). + For example, with GNU C on x86 platforms: alignof_type(double) = 8, but + - when -malign-double is not specified: alignof_slot(double) = 4, + - when -malign-double is specified: alignof_slot(double) = 8. + Note: The result cannot be used as a value for an 'enum' constant, + due to bugs in HP-UX 10.20 cc and AIX 3.2.5 xlc. */ +#if defined __GNUC__ || defined __IBM__ALIGNOF__ +# define alignof_type __alignof__ +#else +# define alignof_type alignof_slot +#endif + +#endif /* _ALIGNOF_H */ diff --git a/lib/alloca.c b/lib/alloca.c new file mode 100644 index 0000000..ee0f018 --- /dev/null +++ b/lib/alloca.c @@ -0,0 +1,478 @@ +/* alloca.c -- allocate automatically reclaimed memory + (Mostly) portable public-domain implementation -- D A Gwyn + + This implementation of the PWB library alloca function, + which is used to allocate space off the run-time stack so + that it is automatically reclaimed upon procedure exit, + was inspired by discussions with J. Q. Johnson of Cornell. + J.Otto Tennant <jot@cray.com> contributed the Cray support. + + There are some preprocessor constants that can + be defined when compiling for your specific system, for + improved efficiency; however, the defaults should be okay. + + The general concept of this implementation is to keep + track of all alloca-allocated blocks, and reclaim any + that are found to be deeper in the stack than the current + invocation. This heuristic does not reclaim storage as + soon as it becomes invalid, but it will do so eventually. + + As a special case, alloca(0) reclaims storage without + allocating any. It is a good idea to use alloca(0) in + your main control loop, etc. to force garbage collection. */ + +#include <config.h> + +#include <alloca.h> + +#include <string.h> +#include <stdlib.h> + +#ifdef emacs +# include "lisp.h" +# include "blockinput.h" +# ifdef EMACS_FREE +# undef free +# define free EMACS_FREE +# endif +#else +# define memory_full() abort () +#endif + +/* If compiling with GCC 2, this file's not needed. */ +#if !defined (__GNUC__) || __GNUC__ < 2 + +/* If someone has defined alloca as a macro, + there must be some other way alloca is supposed to work. */ +# ifndef alloca + +# ifdef emacs +# ifdef static +/* actually, only want this if static is defined as "" + -- this is for usg, in which emacs must undefine static + in order to make unexec workable + */ +# ifndef STACK_DIRECTION +you +lose +-- must know STACK_DIRECTION at compile-time +/* Using #error here is not wise since this file should work for + old and obscure compilers. */ +# endif /* STACK_DIRECTION undefined */ +# endif /* static */ +# endif /* emacs */ + +/* If your stack is a linked list of frames, you have to + provide an "address metric" ADDRESS_FUNCTION macro. */ + +# if defined (CRAY) && defined (CRAY_STACKSEG_END) +long i00afunc (); +# define ADDRESS_FUNCTION(arg) (char *) i00afunc (&(arg)) +# else +# define ADDRESS_FUNCTION(arg) &(arg) +# endif + +/* Define STACK_DIRECTION if you know the direction of stack + growth for your system; otherwise it will be automatically + deduced at run-time. + + STACK_DIRECTION > 0 => grows toward higher addresses + STACK_DIRECTION < 0 => grows toward lower addresses + STACK_DIRECTION = 0 => direction of growth unknown */ + +# ifndef STACK_DIRECTION +# define STACK_DIRECTION 0 /* Direction unknown. */ +# endif + +# if STACK_DIRECTION != 0 + +# define STACK_DIR STACK_DIRECTION /* Known at compile-time. */ + +# else /* STACK_DIRECTION == 0; need run-time code. */ + +static int stack_dir; /* 1 or -1 once known. */ +# define STACK_DIR stack_dir + +static int +find_stack_direction (int *addr, int depth) +{ + int dir, dummy = 0; + if (! addr) + addr = &dummy; + *addr = addr < &dummy ? 1 : addr == &dummy ? 0 : -1; + dir = depth ? find_stack_direction (addr, depth - 1) : 0; + return dir + dummy; +} + +# endif /* STACK_DIRECTION == 0 */ + +/* An "alloca header" is used to: + (a) chain together all alloca'ed blocks; + (b) keep track of stack depth. + + It is very important that sizeof(header) agree with malloc + alignment chunk size. The following default should work okay. */ + +# ifndef ALIGN_SIZE +# define ALIGN_SIZE sizeof(double) +# endif + +typedef union hdr +{ + char align[ALIGN_SIZE]; /* To force sizeof(header). */ + struct + { + union hdr *next; /* For chaining headers. */ + char *deep; /* For stack depth measure. */ + } h; +} header; + +static header *last_alloca_header = NULL; /* -> last alloca header. */ + +/* Return a pointer to at least SIZE bytes of storage, + which will be automatically reclaimed upon exit from + the procedure that called alloca. Originally, this space + was supposed to be taken from the current stack frame of the + caller, but that method cannot be made to work for some + implementations of C, for example under Gould's UTX/32. */ + +void * +alloca (size_t size) +{ + auto char probe; /* Probes stack depth: */ + register char *depth = ADDRESS_FUNCTION (probe); + +# if STACK_DIRECTION == 0 + if (STACK_DIR == 0) /* Unknown growth direction. */ + STACK_DIR = find_stack_direction (NULL, (size & 1) + 20); +# endif + + /* Reclaim garbage, defined as all alloca'd storage that + was allocated from deeper in the stack than currently. */ + + { + register header *hp; /* Traverses linked list. */ + +# ifdef emacs + BLOCK_INPUT; +# endif + + for (hp = last_alloca_header; hp != NULL;) + if ((STACK_DIR > 0 && hp->h.deep > depth) + || (STACK_DIR < 0 && hp->h.deep < depth)) + { + register header *np = hp->h.next; + + free (hp); /* Collect garbage. */ + + hp = np; /* -> next header. */ + } + else + break; /* Rest are not deeper. */ + + last_alloca_header = hp; /* -> last valid storage. */ + +# ifdef emacs + UNBLOCK_INPUT; +# endif + } + + if (size == 0) + return NULL; /* No allocation required. */ + + /* Allocate combined header + user data storage. */ + + { + /* Address of header. */ + register header *new; + + size_t combined_size = sizeof (header) + size; + if (combined_size < sizeof (header)) + memory_full (); + + new = malloc (combined_size); + + if (! new) + memory_full (); + + new->h.next = last_alloca_header; + new->h.deep = depth; + + last_alloca_header = new; + + /* User storage begins just after header. */ + + return (void *) (new + 1); + } +} + +# if defined (CRAY) && defined (CRAY_STACKSEG_END) + +# ifdef DEBUG_I00AFUNC +# include <stdio.h> +# endif + +# ifndef CRAY_STACK +# define CRAY_STACK +# ifndef CRAY2 +/* Stack structures for CRAY-1, CRAY X-MP, and CRAY Y-MP */ +struct stack_control_header + { + long shgrow:32; /* Number of times stack has grown. */ + long shaseg:32; /* Size of increments to stack. */ + long shhwm:32; /* High water mark of stack. */ + long shsize:32; /* Current size of stack (all segments). */ + }; + +/* The stack segment linkage control information occurs at + the high-address end of a stack segment. (The stack + grows from low addresses to high addresses.) The initial + part of the stack segment linkage control information is + 0200 (octal) words. This provides for register storage + for the routine which overflows the stack. */ + +struct stack_segment_linkage + { + long ss[0200]; /* 0200 overflow words. */ + long sssize:32; /* Number of words in this segment. */ + long ssbase:32; /* Offset to stack base. */ + long:32; + long sspseg:32; /* Offset to linkage control of previous + segment of stack. */ + long:32; + long sstcpt:32; /* Pointer to task common address block. */ + long sscsnm; /* Private control structure number for + microtasking. */ + long ssusr1; /* Reserved for user. */ + long ssusr2; /* Reserved for user. */ + long sstpid; /* Process ID for pid based multi-tasking. */ + long ssgvup; /* Pointer to multitasking thread giveup. */ + long sscray[7]; /* Reserved for Cray Research. */ + long ssa0; + long ssa1; + long ssa2; + long ssa3; + long ssa4; + long ssa5; + long ssa6; + long ssa7; + long sss0; + long sss1; + long sss2; + long sss3; + long sss4; + long sss5; + long sss6; + long sss7; + }; + +# else /* CRAY2 */ +/* The following structure defines the vector of words + returned by the STKSTAT library routine. */ +struct stk_stat + { + long now; /* Current total stack size. */ + long maxc; /* Amount of contiguous space which would + be required to satisfy the maximum + stack demand to date. */ + long high_water; /* Stack high-water mark. */ + long overflows; /* Number of stack overflow ($STKOFEN) calls. */ + long hits; /* Number of internal buffer hits. */ + long extends; /* Number of block extensions. */ + long stko_mallocs; /* Block allocations by $STKOFEN. */ + long underflows; /* Number of stack underflow calls ($STKRETN). */ + long stko_free; /* Number of deallocations by $STKRETN. */ + long stkm_free; /* Number of deallocations by $STKMRET. */ + long segments; /* Current number of stack segments. */ + long maxs; /* Maximum number of stack segments so far. */ + long pad_size; /* Stack pad size. */ + long current_address; /* Current stack segment address. */ + long current_size; /* Current stack segment size. This + number is actually corrupted by STKSTAT to + include the fifteen word trailer area. */ + long initial_address; /* Address of initial segment. */ + long initial_size; /* Size of initial segment. */ + }; + +/* The following structure describes the data structure which trails + any stack segment. I think that the description in 'asdef' is + out of date. I only describe the parts that I am sure about. */ + +struct stk_trailer + { + long this_address; /* Address of this block. */ + long this_size; /* Size of this block (does not include + this trailer). */ + long unknown2; + long unknown3; + long link; /* Address of trailer block of previous + segment. */ + long unknown5; + long unknown6; + long unknown7; + long unknown8; + long unknown9; + long unknown10; + long unknown11; + long unknown12; + long unknown13; + long unknown14; + }; + +# endif /* CRAY2 */ +# endif /* not CRAY_STACK */ + +# ifdef CRAY2 +/* Determine a "stack measure" for an arbitrary ADDRESS. + I doubt that "lint" will like this much. */ + +static long +i00afunc (long *address) +{ + struct stk_stat status; + struct stk_trailer *trailer; + long *block, size; + long result = 0; + + /* We want to iterate through all of the segments. The first + step is to get the stack status structure. We could do this + more quickly and more directly, perhaps, by referencing the + $LM00 common block, but I know that this works. */ + + STKSTAT (&status); + + /* Set up the iteration. */ + + trailer = (struct stk_trailer *) (status.current_address + + status.current_size + - 15); + + /* There must be at least one stack segment. Therefore it is + a fatal error if "trailer" is null. */ + + if (trailer == 0) + abort (); + + /* Discard segments that do not contain our argument address. */ + + while (trailer != 0) + { + block = (long *) trailer->this_address; + size = trailer->this_size; + if (block == 0 || size == 0) + abort (); + trailer = (struct stk_trailer *) trailer->link; + if ((block <= address) && (address < (block + size))) + break; + } + + /* Set the result to the offset in this segment and add the sizes + of all predecessor segments. */ + + result = address - block; + + if (trailer == 0) + { + return result; + } + + do + { + if (trailer->this_size <= 0) + abort (); + result += trailer->this_size; + trailer = (struct stk_trailer *) trailer->link; + } + while (trailer != 0); + + /* We are done. Note that if you present a bogus address (one + not in any segment), you will get a different number back, formed + from subtracting the address of the first block. This is probably + not what you want. */ + + return (result); +} + +# else /* not CRAY2 */ +/* Stack address function for a CRAY-1, CRAY X-MP, or CRAY Y-MP. + Determine the number of the cell within the stack, + given the address of the cell. The purpose of this + routine is to linearize, in some sense, stack addresses + for alloca. */ + +static long +i00afunc (long address) +{ + long stkl = 0; + + long size, pseg, this_segment, stack; + long result = 0; + + struct stack_segment_linkage *ssptr; + + /* Register B67 contains the address of the end of the + current stack segment. If you (as a subprogram) store + your registers on the stack and find that you are past + the contents of B67, you have overflowed the segment. + + B67 also points to the stack segment linkage control + area, which is what we are really interested in. */ + + stkl = CRAY_STACKSEG_END (); + ssptr = (struct stack_segment_linkage *) stkl; + + /* If one subtracts 'size' from the end of the segment, + one has the address of the first word of the segment. + + If this is not the first segment, 'pseg' will be + nonzero. */ + + pseg = ssptr->sspseg; + size = ssptr->sssize; + + this_segment = stkl - size; + + /* It is possible that calling this routine itself caused + a stack overflow. Discard stack segments which do not + contain the target address. */ + + while (!(this_segment <= address && address <= stkl)) + { +# ifdef DEBUG_I00AFUNC + fprintf (stderr, "%011o %011o %011o\n", this_segment, address, stkl); +# endif + if (pseg == 0) + break; + stkl = stkl - pseg; + ssptr = (struct stack_segment_linkage *) stkl; + size = ssptr->sssize; + pseg = ssptr->sspseg; + this_segment = stkl - size; + } + + result = address - this_segment; + + /* If you subtract pseg from the current end of the stack, + you get the address of the previous stack segment's end. + This seems a little convoluted to me, but I'll bet you save + a cycle somewhere. */ + + while (pseg != 0) + { +# ifdef DEBUG_I00AFUNC + fprintf (stderr, "%011o %011o\n", pseg, size); +# endif + stkl = stkl - pseg; + ssptr = (struct stack_segment_linkage *) stkl; + size = ssptr->sssize; + pseg = ssptr->sspseg; + result += size; + } + return (result); +} + +# endif /* not CRAY2 */ +# endif /* CRAY */ + +# endif /* no alloca */ +#endif /* not GCC 2 */ diff --git a/lib/alloca.in.h b/lib/alloca.in.h new file mode 100644 index 0000000..5ad8af8 --- /dev/null +++ b/lib/alloca.in.h @@ -0,0 +1,65 @@ +/* Memory allocation on the stack. + + Copyright (C) 1995, 1999, 2001-2004, 2006-2018 Free Software Foundation, + Inc. + + This program is free software; you can redistribute it and/or modify it + under the terms of the GNU General Public License as published + by the Free Software Foundation; either version 3, or (at your option) + any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + General Public License for more details. + + You should have received a copy of the GNU General Public + License along with this program; if not, see + <https://www.gnu.org/licenses/>. + */ + +/* Avoid using the symbol _ALLOCA_H here, as Bison assumes _ALLOCA_H + means there is a real alloca function. */ +#ifndef _GL_ALLOCA_H +#define _GL_ALLOCA_H + +/* alloca (N) returns a pointer to N bytes of memory + allocated on the stack, which will last until the function returns. + Use of alloca should be avoided: + - inside arguments of function calls - undefined behaviour, + - in inline functions - the allocation may actually last until the + calling function returns, + - for huge N (say, N >= 65536) - you never know how large (or small) + the stack is, and when the stack cannot fulfill the memory allocation + request, the program just crashes. + */ + +#ifndef alloca +# ifdef __GNUC__ +# define alloca __builtin_alloca +# elif defined _AIX +# define alloca __alloca +# elif defined _MSC_VER +# include <malloc.h> +# define alloca _alloca +# elif defined __DECC && defined __VMS +# define alloca __ALLOCA +# elif defined __TANDEM && defined _TNS_E_TARGET +# ifdef __cplusplus +extern "C" +# endif +void *_alloca (unsigned short); +# pragma intrinsic (_alloca) +# define alloca _alloca +# elif defined __MVS__ +# include <stdlib.h> +# else +# include <stddef.h> +# ifdef __cplusplus +extern "C" +# endif +void *alloca (size_t); +# endif +#endif + +#endif /* _GL_ALLOCA_H */ diff --git a/lib/arg-nonnull.h b/lib/arg-nonnull.h new file mode 100644 index 0000000..5f03408 --- /dev/null +++ b/lib/arg-nonnull.h @@ -0,0 +1,26 @@ +/* A C macro for declaring that specific arguments must not be NULL. + Copyright (C) 2009-2018 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify it + under the terms of the GNU General Public License as published + by the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +/* _GL_ARG_NONNULL((n,...,m)) tells the compiler and static analyzer tools + that the values passed as arguments n, ..., m must be non-NULL pointers. + n = 1 stands for the first argument, n = 2 for the second argument etc. */ +#ifndef _GL_ARG_NONNULL +# if (__GNUC__ == 3 && __GNUC_MINOR__ >= 3) || __GNUC__ > 3 +# define _GL_ARG_NONNULL(params) __attribute__ ((__nonnull__ params)) +# else +# define _GL_ARG_NONNULL(params) +# endif +#endif diff --git a/lib/basename-lgpl.c b/lib/basename-lgpl.c new file mode 100644 index 0000000..33f9994 --- /dev/null +++ b/lib/basename-lgpl.c @@ -0,0 +1,75 @@ +/* basename.c -- return the last element in a file name + + Copyright (C) 1990, 1998-2001, 2003-2006, 2009-2018 Free Software + Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +#include <config.h> + +#include "dirname.h" + +#include <string.h> + +/* Return the address of the last file name component of NAME. If + NAME has no relative file name components because it is a file + system root, return the empty string. */ + +char * +last_component (char const *name) +{ + char const *base = name + FILE_SYSTEM_PREFIX_LEN (name); + char const *p; + bool saw_slash = false; + + while (ISSLASH (*base)) + base++; + + for (p = base; *p; p++) + { + if (ISSLASH (*p)) + saw_slash = true; + else if (saw_slash) + { + base = p; + saw_slash = false; + } + } + + return (char *) base; +} + +/* Return the length of the basename NAME. Typically NAME is the + value returned by base_name or last_component. Act like strlen + (NAME), except omit all trailing slashes. */ + +size_t +base_len (char const *name) +{ + size_t len; + size_t prefix_len = FILE_SYSTEM_PREFIX_LEN (name); + + for (len = strlen (name); 1 < len && ISSLASH (name[len - 1]); len--) + continue; + + if (DOUBLE_SLASH_IS_DISTINCT_ROOT && len == 1 + && ISSLASH (name[0]) && ISSLASH (name[1]) && ! name[2]) + return 2; + + if (FILE_SYSTEM_DRIVE_PREFIX_CAN_BE_RELATIVE && prefix_len + && len == prefix_len && ISSLASH (name[prefix_len])) + return prefix_len + 1; + + return len; +} diff --git a/lib/binary-io.c b/lib/binary-io.c new file mode 100644 index 0000000..f9cc4dd --- /dev/null +++ b/lib/binary-io.c @@ -0,0 +1,37 @@ +/* Binary mode I/O. + Copyright 2017-2018 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +#include <config.h> + +#define BINARY_IO_INLINE _GL_EXTERN_INLINE +#include "binary-io.h" + +#if defined __DJGPP__ || defined __EMX__ +# include <errno.h> +# include <unistd.h> + +int +__gl_setmode_check (int fd) +{ + if (isatty (fd)) + { + errno = EINVAL; + return -1; + } + else + return 0; +} +#endif diff --git a/lib/binary-io.h b/lib/binary-io.h new file mode 100644 index 0000000..1f21fc0 --- /dev/null +++ b/lib/binary-io.h @@ -0,0 +1,81 @@ +/* Binary mode I/O. + Copyright (C) 2001, 2003, 2005, 2008-2018 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +#ifndef _BINARY_H +#define _BINARY_H + +/* For systems that distinguish between text and binary I/O. + O_BINARY is guaranteed by the gnulib <fcntl.h>. */ +#include <fcntl.h> + +/* The MSVC7 <stdio.h> doesn't like to be included after '#define fileno ...', + so we include it here first. */ +#include <stdio.h> + +#ifndef _GL_INLINE_HEADER_BEGIN + #error "Please include config.h first." +#endif +_GL_INLINE_HEADER_BEGIN +#ifndef BINARY_IO_INLINE +# define BINARY_IO_INLINE _GL_INLINE +#endif + +#if O_BINARY +# if defined __EMX__ || defined __DJGPP__ || defined __CYGWIN__ +# include <io.h> /* declares setmode() */ +# define __gl_setmode setmode +# else +# define __gl_setmode _setmode +# undef fileno +# define fileno _fileno +# endif +#else + /* On reasonable systems, binary I/O is the only choice. */ + /* Use a function rather than a macro, to avoid gcc warnings + "warning: statement with no effect". */ +BINARY_IO_INLINE int +__gl_setmode (int fd _GL_UNUSED, int mode _GL_UNUSED) +{ + return O_BINARY; +} +#endif + +#if defined __DJGPP__ || defined __EMX__ +extern int __gl_setmode_check (int); +#else +BINARY_IO_INLINE int +__gl_setmode_check (int fd _GL_UNUSED) { return 0; } +#endif + +/* Set FD's mode to MODE, which should be either O_TEXT or O_BINARY. + Return the old mode if successful, -1 (setting errno) on failure. + Ordinarily this function would be called 'setmode', since that is + its name on MS-Windows, but it is called 'set_binary_mode' here + to avoid colliding with a BSD function of another name. */ + +BINARY_IO_INLINE int +set_binary_mode (int fd, int mode) +{ + int r = __gl_setmode_check (fd); + return r != 0 ? r : __gl_setmode (fd, mode); +} + +/* This macro is obsolescent. */ +#define SET_BINARY(fd) ((void) set_binary_mode (fd, O_BINARY)) + +_GL_INLINE_HEADER_END + +#endif /* _BINARY_H */ diff --git a/lib/btowc.c b/lib/btowc.c new file mode 100644 index 0000000..54124b2 --- /dev/null +++ b/lib/btowc.c @@ -0,0 +1,39 @@ +/* Convert unibyte character to wide character. + Copyright (C) 2008, 2010-2018 Free Software Foundation, Inc. + Written by Bruno Haible <bruno@clisp.org>, 2008. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +#include <config.h> + +/* Specification. */ +#include <wchar.h> + +#include <stdio.h> +#include <stdlib.h> + +wint_t +btowc (int c) +{ + if (c != EOF) + { + char buf[1]; + wchar_t wc; + + buf[0] = c; + if (mbtowc (&wc, buf, 1) >= 0) + return wc; + } + return WEOF; +} diff --git a/lib/c++defs.h b/lib/c++defs.h new file mode 100644 index 0000000..72ff1ff --- /dev/null +++ b/lib/c++defs.h @@ -0,0 +1,316 @@ +/* C++ compatible function declaration macros. + Copyright (C) 2010-2018 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify it + under the terms of the GNU General Public License as published + by the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +#ifndef _GL_CXXDEFS_H +#define _GL_CXXDEFS_H + +/* Begin/end the GNULIB_NAMESPACE namespace. */ +#if defined __cplusplus && defined GNULIB_NAMESPACE +# define _GL_BEGIN_NAMESPACE namespace GNULIB_NAMESPACE { +# define _GL_END_NAMESPACE } +#else +# define _GL_BEGIN_NAMESPACE +# define _GL_END_NAMESPACE +#endif + +/* The three most frequent use cases of these macros are: + + * For providing a substitute for a function that is missing on some + platforms, but is declared and works fine on the platforms on which + it exists: + + #if @GNULIB_FOO@ + # if !@HAVE_FOO@ + _GL_FUNCDECL_SYS (foo, ...); + # endif + _GL_CXXALIAS_SYS (foo, ...); + _GL_CXXALIASWARN (foo); + #elif defined GNULIB_POSIXCHECK + ... + #endif + + * For providing a replacement for a function that exists on all platforms, + but is broken/insufficient and needs to be replaced on some platforms: + + #if @GNULIB_FOO@ + # if @REPLACE_FOO@ + # if !(defined __cplusplus && defined GNULIB_NAMESPACE) + # undef foo + # define foo rpl_foo + # endif + _GL_FUNCDECL_RPL (foo, ...); + _GL_CXXALIAS_RPL (foo, ...); + # else + _GL_CXXALIAS_SYS (foo, ...); + # endif + _GL_CXXALIASWARN (foo); + #elif defined GNULIB_POSIXCHECK + ... + #endif + + * For providing a replacement for a function that exists on some platforms + but is broken/insufficient and needs to be replaced on some of them and + is additionally either missing or undeclared on some other platforms: + + #if @GNULIB_FOO@ + # if @REPLACE_FOO@ + # if !(defined __cplusplus && defined GNULIB_NAMESPACE) + # undef foo + # define foo rpl_foo + # endif + _GL_FUNCDECL_RPL (foo, ...); + _GL_CXXALIAS_RPL (foo, ...); + # else + # if !@HAVE_FOO@ or if !@HAVE_DECL_FOO@ + _GL_FUNCDECL_SYS (foo, ...); + # endif + _GL_CXXALIAS_SYS (foo, ...); + # endif + _GL_CXXALIASWARN (foo); + #elif defined GNULIB_POSIXCHECK + ... + #endif +*/ + +/* _GL_EXTERN_C declaration; + performs the declaration with C linkage. */ +#if defined __cplusplus +# define _GL_EXTERN_C extern "C" +#else +# define _GL_EXTERN_C extern +#endif + +/* _GL_FUNCDECL_RPL (func, rettype, parameters_and_attributes); + declares a replacement function, named rpl_func, with the given prototype, + consisting of return type, parameters, and attributes. + Example: + _GL_FUNCDECL_RPL (open, int, (const char *filename, int flags, ...) + _GL_ARG_NONNULL ((1))); + */ +#define _GL_FUNCDECL_RPL(func,rettype,parameters_and_attributes) \ + _GL_FUNCDECL_RPL_1 (rpl_##func, rettype, parameters_and_attributes) +#define _GL_FUNCDECL_RPL_1(rpl_func,rettype,parameters_and_attributes) \ + _GL_EXTERN_C rettype rpl_func parameters_and_attributes + +/* _GL_FUNCDECL_SYS (func, rettype, parameters_and_attributes); + declares the system function, named func, with the given prototype, + consisting of return type, parameters, and attributes. + Example: + _GL_FUNCDECL_SYS (open, int, (const char *filename, int flags, ...) + _GL_ARG_NONNULL ((1))); + */ +#define _GL_FUNCDECL_SYS(func,rettype,parameters_and_attributes) \ + _GL_EXTERN_C rettype func parameters_and_attributes + +/* _GL_CXXALIAS_RPL (func, rettype, parameters); + declares a C++ alias called GNULIB_NAMESPACE::func + that redirects to rpl_func, if GNULIB_NAMESPACE is defined. + Example: + _GL_CXXALIAS_RPL (open, int, (const char *filename, int flags, ...)); + + Wrapping rpl_func in an object with an inline conversion operator + avoids a reference to rpl_func unless GNULIB_NAMESPACE::func is + actually used in the program. */ +#define _GL_CXXALIAS_RPL(func,rettype,parameters) \ + _GL_CXXALIAS_RPL_1 (func, rpl_##func, rettype, parameters) +#if defined __cplusplus && defined GNULIB_NAMESPACE +# define _GL_CXXALIAS_RPL_1(func,rpl_func,rettype,parameters) \ + namespace GNULIB_NAMESPACE \ + { \ + static const struct _gl_ ## func ## _wrapper \ + { \ + typedef rettype (*type) parameters; \ + \ + inline operator type () const \ + { \ + return ::rpl_func; \ + } \ + } func = {}; \ + } \ + _GL_EXTERN_C int _gl_cxxalias_dummy +#else +# define _GL_CXXALIAS_RPL_1(func,rpl_func,rettype,parameters) \ + _GL_EXTERN_C int _gl_cxxalias_dummy +#endif + +/* _GL_CXXALIAS_RPL_CAST_1 (func, rpl_func, rettype, parameters); + is like _GL_CXXALIAS_RPL_1 (func, rpl_func, rettype, parameters); + except that the C function rpl_func may have a slightly different + declaration. A cast is used to silence the "invalid conversion" error + that would otherwise occur. */ +#if defined __cplusplus && defined GNULIB_NAMESPACE +# define _GL_CXXALIAS_RPL_CAST_1(func,rpl_func,rettype,parameters) \ + namespace GNULIB_NAMESPACE \ + { \ + static const struct _gl_ ## func ## _wrapper \ + { \ + typedef rettype (*type) parameters; \ + \ + inline operator type () const \ + { \ + return reinterpret_cast<type>(::rpl_func); \ + } \ + } func = {}; \ + } \ + _GL_EXTERN_C int _gl_cxxalias_dummy +#else +# define _GL_CXXALIAS_RPL_CAST_1(func,rpl_func,rettype,parameters) \ + _GL_EXTERN_C int _gl_cxxalias_dummy +#endif + +/* _GL_CXXALIAS_SYS (func, rettype, parameters); + declares a C++ alias called GNULIB_NAMESPACE::func + that redirects to the system provided function func, if GNULIB_NAMESPACE + is defined. + Example: + _GL_CXXALIAS_SYS (open, int, (const char *filename, int flags, ...)); + + Wrapping func in an object with an inline conversion operator + avoids a reference to func unless GNULIB_NAMESPACE::func is + actually used in the program. */ +#if defined __cplusplus && defined GNULIB_NAMESPACE +# define _GL_CXXALIAS_SYS(func,rettype,parameters) \ + namespace GNULIB_NAMESPACE \ + { \ + static const struct _gl_ ## func ## _wrapper \ + { \ + typedef rettype (*type) parameters; \ + \ + inline operator type () const \ + { \ + return ::func; \ + } \ + } func = {}; \ + } \ + _GL_EXTERN_C int _gl_cxxalias_dummy +#else +# define _GL_CXXALIAS_SYS(func,rettype,parameters) \ + _GL_EXTERN_C int _gl_cxxalias_dummy +#endif + +/* _GL_CXXALIAS_SYS_CAST (func, rettype, parameters); + is like _GL_CXXALIAS_SYS (func, rettype, parameters); + except that the C function func may have a slightly different declaration. + A cast is used to silence the "invalid conversion" error that would + otherwise occur. */ +#if defined __cplusplus && defined GNULIB_NAMESPACE +# define _GL_CXXALIAS_SYS_CAST(func,rettype,parameters) \ + namespace GNULIB_NAMESPACE \ + { \ + static const struct _gl_ ## func ## _wrapper \ + { \ + typedef rettype (*type) parameters; \ + \ + inline operator type () const \ + { \ + return reinterpret_cast<type>(::func); \ + } \ + } func = {}; \ + } \ + _GL_EXTERN_C int _gl_cxxalias_dummy +#else +# define _GL_CXXALIAS_SYS_CAST(func,rettype,parameters) \ + _GL_EXTERN_C int _gl_cxxalias_dummy +#endif + +/* _GL_CXXALIAS_SYS_CAST2 (func, rettype, parameters, rettype2, parameters2); + is like _GL_CXXALIAS_SYS (func, rettype, parameters); + except that the C function is picked among a set of overloaded functions, + namely the one with rettype2 and parameters2. Two consecutive casts + are used to silence the "cannot find a match" and "invalid conversion" + errors that would otherwise occur. */ +#if defined __cplusplus && defined GNULIB_NAMESPACE + /* The outer cast must be a reinterpret_cast. + The inner cast: When the function is defined as a set of overloaded + functions, it works as a static_cast<>, choosing the designated variant. + When the function is defined as a single variant, it works as a + reinterpret_cast<>. The parenthesized cast syntax works both ways. */ +# define _GL_CXXALIAS_SYS_CAST2(func,rettype,parameters,rettype2,parameters2) \ + namespace GNULIB_NAMESPACE \ + { \ + static const struct _gl_ ## func ## _wrapper \ + { \ + typedef rettype (*type) parameters; \ + \ + inline operator type () const \ + { \ + return reinterpret_cast<type>((rettype2 (*) parameters2)(::func)); \ + } \ + } func = {}; \ + } \ + _GL_EXTERN_C int _gl_cxxalias_dummy +#else +# define _GL_CXXALIAS_SYS_CAST2(func,rettype,parameters,rettype2,parameters2) \ + _GL_EXTERN_C int _gl_cxxalias_dummy +#endif + +/* _GL_CXXALIASWARN (func); + causes a warning to be emitted when ::func is used but not when + GNULIB_NAMESPACE::func is used. func must be defined without overloaded + variants. */ +#if defined __cplusplus && defined GNULIB_NAMESPACE +# define _GL_CXXALIASWARN(func) \ + _GL_CXXALIASWARN_1 (func, GNULIB_NAMESPACE) +# define _GL_CXXALIASWARN_1(func,namespace) \ + _GL_CXXALIASWARN_2 (func, namespace) +/* To work around GCC bug <https://gcc.gnu.org/bugzilla/show_bug.cgi?id=43881>, + we enable the warning only when not optimizing. */ +# if !__OPTIMIZE__ +# define _GL_CXXALIASWARN_2(func,namespace) \ + _GL_WARN_ON_USE (func, \ + "The symbol ::" #func " refers to the system function. " \ + "Use " #namespace "::" #func " instead.") +# elif __GNUC__ >= 3 && GNULIB_STRICT_CHECKING +# define _GL_CXXALIASWARN_2(func,namespace) \ + extern __typeof__ (func) func +# else +# define _GL_CXXALIASWARN_2(func,namespace) \ + _GL_EXTERN_C int _gl_cxxalias_dummy +# endif +#else +# define _GL_CXXALIASWARN(func) \ + _GL_EXTERN_C int _gl_cxxalias_dummy +#endif + +/* _GL_CXXALIASWARN1 (func, rettype, parameters_and_attributes); + causes a warning to be emitted when the given overloaded variant of ::func + is used but not when GNULIB_NAMESPACE::func is used. */ +#if defined __cplusplus && defined GNULIB_NAMESPACE +# define _GL_CXXALIASWARN1(func,rettype,parameters_and_attributes) \ + _GL_CXXALIASWARN1_1 (func, rettype, parameters_and_attributes, \ + GNULIB_NAMESPACE) +# define _GL_CXXALIASWARN1_1(func,rettype,parameters_and_attributes,namespace) \ + _GL_CXXALIASWARN1_2 (func, rettype, parameters_and_attributes, namespace) +/* To work around GCC bug <https://gcc.gnu.org/bugzilla/show_bug.cgi?id=43881>, + we enable the warning only when not optimizing. */ +# if !__OPTIMIZE__ +# define _GL_CXXALIASWARN1_2(func,rettype,parameters_and_attributes,namespace) \ + _GL_WARN_ON_USE_CXX (func, rettype, parameters_and_attributes, \ + "The symbol ::" #func " refers to the system function. " \ + "Use " #namespace "::" #func " instead.") +# elif __GNUC__ >= 3 && GNULIB_STRICT_CHECKING +# define _GL_CXXALIASWARN1_2(func,rettype,parameters_and_attributes,namespace) \ + extern __typeof__ (func) func +# else +# define _GL_CXXALIASWARN1_2(func,rettype,parameters_and_attributes,namespace) \ + _GL_EXTERN_C int _gl_cxxalias_dummy +# endif +#else +# define _GL_CXXALIASWARN1(func,rettype,parameters_and_attributes) \ + _GL_EXTERN_C int _gl_cxxalias_dummy +#endif + +#endif /* _GL_CXXDEFS_H */ diff --git a/lib/c-ctype.c b/lib/c-ctype.c new file mode 100644 index 0000000..5d9d4d8 --- /dev/null +++ b/lib/c-ctype.c @@ -0,0 +1,3 @@ +#include <config.h> +#define C_CTYPE_INLINE _GL_EXTERN_INLINE +#include "c-ctype.h" diff --git a/lib/c-ctype.h b/lib/c-ctype.h new file mode 100644 index 0000000..d55d4f6 --- /dev/null +++ b/lib/c-ctype.h @@ -0,0 +1,366 @@ +/* Character handling in C locale. + + These functions work like the corresponding functions in <ctype.h>, + except that they have the C (POSIX) locale hardwired, whereas the + <ctype.h> functions' behaviour depends on the current locale set via + setlocale. + + Copyright (C) 2000-2003, 2006, 2008-2018 Free Software Foundation, Inc. + +This program is free software; you can redistribute it and/or modify +it under the terms of the GNU General Public License as published by +the Free Software Foundation; either version 3 of the License, or +(at your option) any later version. + +This program is distributed in the hope that it will be useful, +but WITHOUT ANY WARRANTY; without even the implied warranty of +MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +GNU General Public License for more details. + +You should have received a copy of the GNU General Public License +along with this program; if not, see <https://www.gnu.org/licenses/>. */ + +#ifndef C_CTYPE_H +#define C_CTYPE_H + +#include <stdbool.h> + +#ifndef _GL_INLINE_HEADER_BEGIN + #error "Please include config.h first." +#endif +_GL_INLINE_HEADER_BEGIN +#ifndef C_CTYPE_INLINE +# define C_CTYPE_INLINE _GL_INLINE +#endif + +#ifdef __cplusplus +extern "C" { +#endif + + +/* The functions defined in this file assume the "C" locale and a character + set without diacritics (ASCII-US or EBCDIC-US or something like that). + Even if the "C" locale on a particular system is an extension of the ASCII + character set (like on BeOS, where it is UTF-8, or on AmigaOS, where it + is ISO-8859-1), the functions in this file recognize only the ASCII + characters. */ + + +#if (' ' == 32) && ('!' == 33) && ('"' == 34) && ('#' == 35) \ + && ('%' == 37) && ('&' == 38) && ('\'' == 39) && ('(' == 40) \ + && (')' == 41) && ('*' == 42) && ('+' == 43) && (',' == 44) \ + && ('-' == 45) && ('.' == 46) && ('/' == 47) && ('0' == 48) \ + && ('1' == 49) && ('2' == 50) && ('3' == 51) && ('4' == 52) \ + && ('5' == 53) && ('6' == 54) && ('7' == 55) && ('8' == 56) \ + && ('9' == 57) && (':' == 58) && (';' == 59) && ('<' == 60) \ + && ('=' == 61) && ('>' == 62) && ('?' == 63) && ('A' == 65) \ + && ('B' == 66) && ('C' == 67) && ('D' == 68) && ('E' == 69) \ + && ('F' == 70) && ('G' == 71) && ('H' == 72) && ('I' == 73) \ + && ('J' == 74) && ('K' == 75) && ('L' == 76) && ('M' == 77) \ + && ('N' == 78) && ('O' == 79) && ('P' == 80) && ('Q' == 81) \ + && ('R' == 82) && ('S' == 83) && ('T' == 84) && ('U' == 85) \ + && ('V' == 86) && ('W' == 87) && ('X' == 88) && ('Y' == 89) \ + && ('Z' == 90) && ('[' == 91) && ('\\' == 92) && (']' == 93) \ + && ('^' == 94) && ('_' == 95) && ('a' == 97) && ('b' == 98) \ + && ('c' == 99) && ('d' == 100) && ('e' == 101) && ('f' == 102) \ + && ('g' == 103) && ('h' == 104) && ('i' == 105) && ('j' == 106) \ + && ('k' == 107) && ('l' == 108) && ('m' == 109) && ('n' == 110) \ + && ('o' == 111) && ('p' == 112) && ('q' == 113) && ('r' == 114) \ + && ('s' == 115) && ('t' == 116) && ('u' == 117) && ('v' == 118) \ + && ('w' == 119) && ('x' == 120) && ('y' == 121) && ('z' == 122) \ + && ('{' == 123) && ('|' == 124) && ('}' == 125) && ('~' == 126) +/* The character set is ASCII or one of its variants or extensions, not EBCDIC. + Testing the value of '\n' and '\r' is not relevant. */ +# define C_CTYPE_ASCII 1 +#elif ! (' ' == '\x40' && '0' == '\xf0' \ + && 'A' == '\xc1' && 'J' == '\xd1' && 'S' == '\xe2' \ + && 'a' == '\x81' && 'j' == '\x91' && 's' == '\xa2') +# error "Only ASCII and EBCDIC are supported" +#endif + +#if 'A' < 0 +# error "EBCDIC and char is signed -- not supported" +#endif + +/* Cases for control characters. */ + +#define _C_CTYPE_CNTRL \ + case '\a': case '\b': case '\f': case '\n': \ + case '\r': case '\t': case '\v': \ + _C_CTYPE_OTHER_CNTRL + +/* ASCII control characters other than those with \-letter escapes. */ + +#if C_CTYPE_ASCII +# define _C_CTYPE_OTHER_CNTRL \ + case '\x00': case '\x01': case '\x02': case '\x03': \ + case '\x04': case '\x05': case '\x06': case '\x0e': \ + case '\x0f': case '\x10': case '\x11': case '\x12': \ + case '\x13': case '\x14': case '\x15': case '\x16': \ + case '\x17': case '\x18': case '\x19': case '\x1a': \ + case '\x1b': case '\x1c': case '\x1d': case '\x1e': \ + case '\x1f': case '\x7f' +#else + /* Use EBCDIC code page 1047's assignments for ASCII control chars; + assume all EBCDIC code pages agree about these assignments. */ +# define _C_CTYPE_OTHER_CNTRL \ + case '\x00': case '\x01': case '\x02': case '\x03': \ + case '\x07': case '\x0e': case '\x0f': case '\x10': \ + case '\x11': case '\x12': case '\x13': case '\x18': \ + case '\x19': case '\x1c': case '\x1d': case '\x1e': \ + case '\x1f': case '\x26': case '\x27': case '\x2d': \ + case '\x2e': case '\x32': case '\x37': case '\x3c': \ + case '\x3d': case '\x3f' +#endif + +/* Cases for lowercase hex letters, and lowercase letters, all offset by N. */ + +#define _C_CTYPE_LOWER_A_THRU_F_N(N) \ + case 'a' + (N): case 'b' + (N): case 'c' + (N): case 'd' + (N): \ + case 'e' + (N): case 'f' + (N) +#define _C_CTYPE_LOWER_N(N) \ + _C_CTYPE_LOWER_A_THRU_F_N(N): \ + case 'g' + (N): case 'h' + (N): case 'i' + (N): case 'j' + (N): \ + case 'k' + (N): case 'l' + (N): case 'm' + (N): case 'n' + (N): \ + case 'o' + (N): case 'p' + (N): case 'q' + (N): case 'r' + (N): \ + case 's' + (N): case 't' + (N): case 'u' + (N): case 'v' + (N): \ + case 'w' + (N): case 'x' + (N): case 'y' + (N): case 'z' + (N) + +/* Cases for hex letters, digits, lower, punct, and upper. */ + +#define _C_CTYPE_A_THRU_F \ + _C_CTYPE_LOWER_A_THRU_F_N (0): \ + _C_CTYPE_LOWER_A_THRU_F_N ('A' - 'a') +#define _C_CTYPE_DIGIT \ + case '0': case '1': case '2': case '3': \ + case '4': case '5': case '6': case '7': \ + case '8': case '9' +#define _C_CTYPE_LOWER _C_CTYPE_LOWER_N (0) +#define _C_CTYPE_PUNCT \ + case '!': case '"': case '#': case '$': \ + case '%': case '&': case '\'': case '(': \ + case ')': case '*': case '+': case ',': \ + case '-': case '.': case '/': case ':': \ + case ';': case '<': case '=': case '>': \ + case '?': case '@': case '[': case '\\': \ + case ']': case '^': case '_': case '`': \ + case '{': case '|': case '}': case '~' +#define _C_CTYPE_UPPER _C_CTYPE_LOWER_N ('A' - 'a') + + +/* Function definitions. */ + +/* Unlike the functions in <ctype.h>, which require an argument in the range + of the 'unsigned char' type, the functions here operate on values that are + in the 'unsigned char' range or in the 'char' range. In other words, + when you have a 'char' value, you need to cast it before using it as + argument to a <ctype.h> function: + + const char *s = ...; + if (isalpha ((unsigned char) *s)) ... + + but you don't need to cast it for the functions defined in this file: + + const char *s = ...; + if (c_isalpha (*s)) ... + */ + +C_CTYPE_INLINE bool +c_isalnum (int c) +{ + switch (c) + { + _C_CTYPE_DIGIT: + _C_CTYPE_LOWER: + _C_CTYPE_UPPER: + return true; + default: + return false; + } +} + +C_CTYPE_INLINE bool +c_isalpha (int c) +{ + switch (c) + { + _C_CTYPE_LOWER: + _C_CTYPE_UPPER: + return true; + default: + return false; + } +} + +/* The function isascii is not locale dependent. + Its use in EBCDIC is questionable. */ +C_CTYPE_INLINE bool +c_isascii (int c) +{ + switch (c) + { + case ' ': + _C_CTYPE_CNTRL: + _C_CTYPE_DIGIT: + _C_CTYPE_LOWER: + _C_CTYPE_PUNCT: + _C_CTYPE_UPPER: + return true; + default: + return false; + } +} + +C_CTYPE_INLINE bool +c_isblank (int c) +{ + return c == ' ' || c == '\t'; +} + +C_CTYPE_INLINE bool +c_iscntrl (int c) +{ + switch (c) + { + _C_CTYPE_CNTRL: + return true; + default: + return false; + } +} + +C_CTYPE_INLINE bool +c_isdigit (int c) +{ + switch (c) + { + _C_CTYPE_DIGIT: + return true; + default: + return false; + } +} + +C_CTYPE_INLINE bool +c_isgraph (int c) +{ + switch (c) + { + _C_CTYPE_DIGIT: + _C_CTYPE_LOWER: + _C_CTYPE_PUNCT: + _C_CTYPE_UPPER: + return true; + default: + return false; + } +} + +C_CTYPE_INLINE bool +c_islower (int c) +{ + switch (c) + { + _C_CTYPE_LOWER: + return true; + default: + return false; + } +} + +C_CTYPE_INLINE bool +c_isprint (int c) +{ + switch (c) + { + case ' ': + _C_CTYPE_DIGIT: + _C_CTYPE_LOWER: + _C_CTYPE_PUNCT: + _C_CTYPE_UPPER: + return true; + default: + return false; + } +} + +C_CTYPE_INLINE bool +c_ispunct (int c) +{ + switch (c) + { + _C_CTYPE_PUNCT: + return true; + default: + return false; + } +} + +C_CTYPE_INLINE bool +c_isspace (int c) +{ + switch (c) + { + case ' ': case '\t': case '\n': case '\v': case '\f': case '\r': + return true; + default: + return false; + } +} + +C_CTYPE_INLINE bool +c_isupper (int c) +{ + switch (c) + { + _C_CTYPE_UPPER: + return true; + default: + return false; + } +} + +C_CTYPE_INLINE bool +c_isxdigit (int c) +{ + switch (c) + { + _C_CTYPE_DIGIT: + _C_CTYPE_A_THRU_F: + return true; + default: + return false; + } +} + +C_CTYPE_INLINE int +c_tolower (int c) +{ + switch (c) + { + _C_CTYPE_UPPER: + return c - 'A' + 'a'; + default: + return c; + } +} + +C_CTYPE_INLINE int +c_toupper (int c) +{ + switch (c) + { + _C_CTYPE_LOWER: + return c - 'a' + 'A'; + default: + return c; + } +} + +#ifdef __cplusplus +} +#endif + +_GL_INLINE_HEADER_END + +#endif /* C_CTYPE_H */ diff --git a/lib/c-strcase.h b/lib/c-strcase.h new file mode 100644 index 0000000..41f439e --- /dev/null +++ b/lib/c-strcase.h @@ -0,0 +1,56 @@ +/* Case-insensitive string comparison functions in C locale. + Copyright (C) 1995-1996, 2001, 2003, 2005, 2009-2018 Free Software + Foundation, Inc. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3, or (at your option) + any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, see <https://www.gnu.org/licenses/>. */ + +#ifndef C_STRCASE_H +#define C_STRCASE_H + +#include <stddef.h> + + +/* The functions defined in this file assume the "C" locale and a character + set without diacritics (ASCII-US or EBCDIC-US or something like that). + Even if the "C" locale on a particular system is an extension of the ASCII + character set (like on BeOS, where it is UTF-8, or on AmigaOS, where it + is ISO-8859-1), the functions in this file recognize only the ASCII + characters. More precisely, one of the string arguments must be an ASCII + string; the other one can also contain non-ASCII characters (but then + the comparison result will be nonzero). */ + + +#ifdef __cplusplus +extern "C" { +#endif + + +/* Compare strings S1 and S2, ignoring case, returning less than, equal to or + greater than zero if S1 is lexicographically less than, equal to or greater + than S2. */ +extern int c_strcasecmp (const char *s1, const char *s2) _GL_ATTRIBUTE_PURE; + +/* Compare no more than N characters of strings S1 and S2, ignoring case, + returning less than, equal to or greater than zero if S1 is + lexicographically less than, equal to or greater than S2. */ +extern int c_strncasecmp (const char *s1, const char *s2, size_t n) + _GL_ATTRIBUTE_PURE; + + +#ifdef __cplusplus +} +#endif + + +#endif /* C_STRCASE_H */ diff --git a/lib/c-strcasecmp.c b/lib/c-strcasecmp.c new file mode 100644 index 0000000..2b1ac99 --- /dev/null +++ b/lib/c-strcasecmp.c @@ -0,0 +1,56 @@ +/* c-strcasecmp.c -- case insensitive string comparator in C locale + Copyright (C) 1998-1999, 2005-2006, 2009-2018 Free Software Foundation, Inc. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3, or (at your option) + any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, see <https://www.gnu.org/licenses/>. */ + +#include <config.h> + +/* Specification. */ +#include "c-strcase.h" + +#include <limits.h> + +#include "c-ctype.h" + +int +c_strcasecmp (const char *s1, const char *s2) +{ + register const unsigned char *p1 = (const unsigned char *) s1; + register const unsigned char *p2 = (const unsigned char *) s2; + unsigned char c1, c2; + + if (p1 == p2) + return 0; + + do + { + c1 = c_tolower (*p1); + c2 = c_tolower (*p2); + + if (c1 == '\0') + break; + + ++p1; + ++p2; + } + while (c1 == c2); + + if (UCHAR_MAX <= INT_MAX) + return c1 - c2; + else + /* On machines where 'char' and 'int' are types of the same size, the + difference of two 'unsigned char' values - including the sign bit - + doesn't fit in an 'int'. */ + return (c1 > c2 ? 1 : c1 < c2 ? -1 : 0); +} diff --git a/lib/c-strcaseeq.h b/lib/c-strcaseeq.h new file mode 100644 index 0000000..9bde9de --- /dev/null +++ b/lib/c-strcaseeq.h @@ -0,0 +1,181 @@ +/* Optimized case-insensitive string comparison in C locale. + Copyright (C) 2001-2002, 2007, 2009-2018 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify it + under the terms of the GNU General Public License as published + by the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +/* Written by Bruno Haible <bruno@clisp.org>. */ + +#include "c-strcase.h" +#include "c-ctype.h" + +/* STRCASEEQ allows to optimize string comparison with a small literal string. + STRCASEEQ (s, "UTF-8", 'U','T','F','-','8',0,0,0,0) + is semantically equivalent to + c_strcasecmp (s, "UTF-8") == 0 + just faster. */ + +/* Help GCC to generate good code for string comparisons with + immediate strings. */ +#if defined (__GNUC__) && defined (__OPTIMIZE__) + +/* Case insensitive comparison of ASCII characters. */ +# if C_CTYPE_ASCII +# define CASEEQ(other,upper) \ + (c_isupper (upper) ? ((other) & ~0x20) == (upper) : (other) == (upper)) +# else +# define CASEEQ(other,upper) \ + (c_toupper (other) == (upper)) +# endif + +static inline int +strcaseeq9 (const char *s1, const char *s2) +{ + return c_strcasecmp (s1 + 9, s2 + 9) == 0; +} + +static inline int +strcaseeq8 (const char *s1, const char *s2, char s28) +{ + if (CASEEQ (s1[8], s28)) + { + if (s28 == 0) + return 1; + else + return strcaseeq9 (s1, s2); + } + else + return 0; +} + +static inline int +strcaseeq7 (const char *s1, const char *s2, char s27, char s28) +{ + if (CASEEQ (s1[7], s27)) + { + if (s27 == 0) + return 1; + else + return strcaseeq8 (s1, s2, s28); + } + else + return 0; +} + +static inline int +strcaseeq6 (const char *s1, const char *s2, char s26, char s27, char s28) +{ + if (CASEEQ (s1[6], s26)) + { + if (s26 == 0) + return 1; + else + return strcaseeq7 (s1, s2, s27, s28); + } + else + return 0; +} + +static inline int +strcaseeq5 (const char *s1, const char *s2, char s25, char s26, char s27, char s28) +{ + if (CASEEQ (s1[5], s25)) + { + if (s25 == 0) + return 1; + else + return strcaseeq6 (s1, s2, s26, s27, s28); + } + else + return 0; +} + +static inline int +strcaseeq4 (const char *s1, const char *s2, char s24, char s25, char s26, char s27, char s28) +{ + if (CASEEQ (s1[4], s24)) + { + if (s24 == 0) + return 1; + else + return strcaseeq5 (s1, s2, s25, s26, s27, s28); + } + else + return 0; +} + +static inline int +strcaseeq3 (const char *s1, const char *s2, char s23, char s24, char s25, char s26, char s27, char s28) +{ + if (CASEEQ (s1[3], s23)) + { + if (s23 == 0) + return 1; + else + return strcaseeq4 (s1, s2, s24, s25, s26, s27, s28); + } + else + return 0; +} + +static inline int +strcaseeq2 (const char *s1, const char *s2, char s22, char s23, char s24, char s25, char s26, char s27, char s28) +{ + if (CASEEQ (s1[2], s22)) + { + if (s22 == 0) + return 1; + else + return strcaseeq3 (s1, s2, s23, s24, s25, s26, s27, s28); + } + else + return 0; +} + +static inline int +strcaseeq1 (const char *s1, const char *s2, char s21, char s22, char s23, char s24, char s25, char s26, char s27, char s28) +{ + if (CASEEQ (s1[1], s21)) + { + if (s21 == 0) + return 1; + else + return strcaseeq2 (s1, s2, s22, s23, s24, s25, s26, s27, s28); + } + else + return 0; +} + +static inline int +strcaseeq0 (const char *s1, const char *s2, char s20, char s21, char s22, char s23, char s24, char s25, char s26, char s27, char s28) +{ + if (CASEEQ (s1[0], s20)) + { + if (s20 == 0) + return 1; + else + return strcaseeq1 (s1, s2, s21, s22, s23, s24, s25, s26, s27, s28); + } + else + return 0; +} + +#define STRCASEEQ(s1,s2,s20,s21,s22,s23,s24,s25,s26,s27,s28) \ + strcaseeq0 (s1, s2, s20, s21, s22, s23, s24, s25, s26, s27, s28) + +#else + +#define STRCASEEQ(s1,s2,s20,s21,s22,s23,s24,s25,s26,s27,s28) \ + (c_strcasecmp (s1, s2) == 0) + +#endif diff --git a/lib/c-strncasecmp.c b/lib/c-strncasecmp.c new file mode 100644 index 0000000..8151c1a --- /dev/null +++ b/lib/c-strncasecmp.c @@ -0,0 +1,56 @@ +/* c-strncasecmp.c -- case insensitive string comparator in C locale + Copyright (C) 1998-1999, 2005-2006, 2009-2018 Free Software Foundation, Inc. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3, or (at your option) + any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, see <https://www.gnu.org/licenses/>. */ + +#include <config.h> + +/* Specification. */ +#include "c-strcase.h" + +#include <limits.h> + +#include "c-ctype.h" + +int +c_strncasecmp (const char *s1, const char *s2, size_t n) +{ + register const unsigned char *p1 = (const unsigned char *) s1; + register const unsigned char *p2 = (const unsigned char *) s2; + unsigned char c1, c2; + + if (p1 == p2 || n == 0) + return 0; + + do + { + c1 = c_tolower (*p1); + c2 = c_tolower (*p2); + + if (--n == 0 || c1 == '\0') + break; + + ++p1; + ++p2; + } + while (c1 == c2); + + if (UCHAR_MAX <= INT_MAX) + return c1 - c2; + else + /* On machines where 'char' and 'int' are types of the same size, the + difference of two 'unsigned char' values - including the sign bit - + doesn't fit in an 'int'. */ + return (c1 > c2 ? 1 : c1 < c2 ? -1 : 0); +} diff --git a/lib/canonicalize-lgpl.c b/lib/canonicalize-lgpl.c new file mode 100644 index 0000000..7831b3f --- /dev/null +++ b/lib/canonicalize-lgpl.c @@ -0,0 +1,428 @@ +/* Return the canonical absolute name of a given file. + Copyright (C) 1996-2018 Free Software Foundation, Inc. + This file is part of the GNU C Library. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +#ifndef _LIBC +/* Don't use __attribute__ __nonnull__ in this compilation unit. Otherwise gcc + optimizes away the name == NULL test below. */ +# define _GL_ARG_NONNULL(params) + +# define _GL_USE_STDLIB_ALLOC 1 +# include <config.h> +#endif + +#if !HAVE_CANONICALIZE_FILE_NAME || !FUNC_REALPATH_WORKS || defined _LIBC + +/* Specification. */ +#include <stdlib.h> + +#include <alloca.h> +#include <string.h> +#include <unistd.h> +#include <limits.h> +#if HAVE_SYS_PARAM_H || defined _LIBC +# include <sys/param.h> +#endif +#include <sys/stat.h> +#include <errno.h> +#include <stddef.h> + +#ifdef _LIBC +# include <shlib-compat.h> +#else +# define SHLIB_COMPAT(lib, introduced, obsoleted) 0 +# define versioned_symbol(lib, local, symbol, version) extern int dummy +# define compat_symbol(lib, local, symbol, version) +# define weak_alias(local, symbol) +# define __canonicalize_file_name canonicalize_file_name +# define __realpath realpath +# include "pathmax.h" +# include "malloca.h" +# include "dosname.h" +# if HAVE_GETCWD +# if IN_RELOCWRAPPER + /* When building the relocatable program wrapper, use the system's getcwd + function, not the gnulib override, otherwise we would get a link error. + */ +# undef getcwd +# endif +# if defined VMS && !defined getcwd + /* We want the directory in Unix syntax, not in VMS syntax. + The gnulib override of 'getcwd' takes 2 arguments; the original VMS + 'getcwd' takes 3 arguments. */ +# define __getcwd(buf, max) getcwd (buf, max, 0) +# else +# define __getcwd getcwd +# endif +# else +# define __getcwd(buf, max) getwd (buf) +# endif +# define __readlink readlink +# define __set_errno(e) errno = (e) +# ifndef MAXSYMLINKS +# ifdef SYMLOOP_MAX +# define MAXSYMLINKS SYMLOOP_MAX +# else +# define MAXSYMLINKS 20 +# endif +# endif +#endif + +#ifndef DOUBLE_SLASH_IS_DISTINCT_ROOT +# define DOUBLE_SLASH_IS_DISTINCT_ROOT 0 +#endif + +/* Define this independently so that stdint.h is not a prerequisite. */ +#ifndef SIZE_MAX +# define SIZE_MAX ((size_t) -1) +#endif + +#if !FUNC_REALPATH_WORKS || defined _LIBC + +static void +alloc_failed (void) +{ +#if defined _WIN32 && ! defined __CYGWIN__ + /* Avoid errno problem without using the malloc or realloc modules; see: + https://lists.gnu.org/r/bug-gnulib/2016-08/msg00025.html */ + errno = ENOMEM; +#endif +} + +/* Return the canonical absolute name of file NAME. A canonical name + does not contain any ".", ".." components nor any repeated path + separators ('/') or symlinks. All path components must exist. If + RESOLVED is null, the result is malloc'd; otherwise, if the + canonical name is PATH_MAX chars or more, returns null with 'errno' + set to ENAMETOOLONG; if the name fits in fewer than PATH_MAX chars, + returns the name in RESOLVED. If the name cannot be resolved and + RESOLVED is non-NULL, it contains the path of the first component + that cannot be resolved. If the path can be resolved, RESOLVED + holds the same value as the value returned. */ + +char * +__realpath (const char *name, char *resolved) +{ + char *rpath, *dest, *extra_buf = NULL; + const char *start, *end, *rpath_limit; + long int path_max; + int num_links = 0; + size_t prefix_len; + + if (name == NULL) + { + /* As per Single Unix Specification V2 we must return an error if + either parameter is a null pointer. We extend this to allow + the RESOLVED parameter to be NULL in case the we are expected to + allocate the room for the return value. */ + __set_errno (EINVAL); + return NULL; + } + + if (name[0] == '\0') + { + /* As per Single Unix Specification V2 we must return an error if + the name argument points to an empty string. */ + __set_errno (ENOENT); + return NULL; + } + +#ifdef PATH_MAX + path_max = PATH_MAX; +#else + path_max = pathconf (name, _PC_PATH_MAX); + if (path_max <= 0) + path_max = 8192; +#endif + + if (resolved == NULL) + { + rpath = malloc (path_max); + if (rpath == NULL) + { + alloc_failed (); + return NULL; + } + } + else + rpath = resolved; + rpath_limit = rpath + path_max; + + /* This is always zero for Posix hosts, but can be 2 for MS-Windows + and MS-DOS X:/foo/bar file names. */ + prefix_len = FILE_SYSTEM_PREFIX_LEN (name); + + if (!IS_ABSOLUTE_FILE_NAME (name)) + { + if (!__getcwd (rpath, path_max)) + { + rpath[0] = '\0'; + goto error; + } + dest = strchr (rpath, '\0'); + start = name; + prefix_len = FILE_SYSTEM_PREFIX_LEN (rpath); + } + else + { + dest = rpath; + if (prefix_len) + { + memcpy (rpath, name, prefix_len); + dest += prefix_len; + } + *dest++ = '/'; + if (DOUBLE_SLASH_IS_DISTINCT_ROOT) + { + if (ISSLASH (name[1]) && !ISSLASH (name[2]) && !prefix_len) + *dest++ = '/'; + *dest = '\0'; + } + start = name + prefix_len; + } + + for (end = start; *start; start = end) + { +#ifdef _LIBC + struct stat64 st; +#else + struct stat st; +#endif + + /* Skip sequence of multiple path-separators. */ + while (ISSLASH (*start)) + ++start; + + /* Find end of path component. */ + for (end = start; *end && !ISSLASH (*end); ++end) + /* Nothing. */; + + if (end - start == 0) + break; + else if (end - start == 1 && start[0] == '.') + /* nothing */; + else if (end - start == 2 && start[0] == '.' && start[1] == '.') + { + /* Back up to previous component, ignore if at root already. */ + if (dest > rpath + prefix_len + 1) + for (--dest; dest > rpath && !ISSLASH (dest[-1]); --dest) + continue; + if (DOUBLE_SLASH_IS_DISTINCT_ROOT + && dest == rpath + 1 && !prefix_len + && ISSLASH (*dest) && !ISSLASH (dest[1])) + dest++; + } + else + { + size_t new_size; + + if (!ISSLASH (dest[-1])) + *dest++ = '/'; + + if (dest + (end - start) >= rpath_limit) + { + ptrdiff_t dest_offset = dest - rpath; + char *new_rpath; + + if (resolved) + { + __set_errno (ENAMETOOLONG); + if (dest > rpath + prefix_len + 1) + dest--; + *dest = '\0'; + goto error; + } + new_size = rpath_limit - rpath; + if (end - start + 1 > path_max) + new_size += end - start + 1; + else + new_size += path_max; + new_rpath = (char *) realloc (rpath, new_size); + if (new_rpath == NULL) + { + alloc_failed (); + goto error; + } + rpath = new_rpath; + rpath_limit = rpath + new_size; + + dest = rpath + dest_offset; + } + +#ifdef _LIBC + dest = __mempcpy (dest, start, end - start); +#else + memcpy (dest, start, end - start); + dest += end - start; +#endif + *dest = '\0'; + + /* FIXME: if lstat fails with errno == EOVERFLOW, + the entry exists. */ +#ifdef _LIBC + if (__lxstat64 (_STAT_VER, rpath, &st) < 0) +#else + if (lstat (rpath, &st) < 0) +#endif + goto error; + + if (S_ISLNK (st.st_mode)) + { + char *buf; + size_t len; + ssize_t n; + + if (++num_links > MAXSYMLINKS) + { + __set_errno (ELOOP); + goto error; + } + + buf = malloca (path_max); + if (!buf) + { + __set_errno (ENOMEM); + goto error; + } + + n = __readlink (rpath, buf, path_max - 1); + if (n < 0) + { + int saved_errno = errno; + freea (buf); + __set_errno (saved_errno); + goto error; + } + buf[n] = '\0'; + + if (!extra_buf) + { + extra_buf = malloca (path_max); + if (!extra_buf) + { + freea (buf); + __set_errno (ENOMEM); + goto error; + } + } + + len = strlen (end); + /* Check that n + len + 1 doesn't overflow and is <= path_max. */ + if (n >= SIZE_MAX - len || n + len >= path_max) + { + freea (buf); + __set_errno (ENAMETOOLONG); + goto error; + } + + /* Careful here, end may be a pointer into extra_buf... */ + memmove (&extra_buf[n], end, len + 1); + name = end = memcpy (extra_buf, buf, n); + + if (IS_ABSOLUTE_FILE_NAME (buf)) + { + size_t pfxlen = FILE_SYSTEM_PREFIX_LEN (buf); + + if (pfxlen) + memcpy (rpath, buf, pfxlen); + dest = rpath + pfxlen; + *dest++ = '/'; /* It's an absolute symlink */ + if (DOUBLE_SLASH_IS_DISTINCT_ROOT) + { + if (ISSLASH (buf[1]) && !ISSLASH (buf[2]) && !pfxlen) + *dest++ = '/'; + *dest = '\0'; + } + /* Install the new prefix to be in effect hereafter. */ + prefix_len = pfxlen; + } + else + { + /* Back up to previous component, ignore if at root + already: */ + if (dest > rpath + prefix_len + 1) + for (--dest; dest > rpath && !ISSLASH (dest[-1]); --dest) + continue; + if (DOUBLE_SLASH_IS_DISTINCT_ROOT && dest == rpath + 1 + && ISSLASH (*dest) && !ISSLASH (dest[1]) && !prefix_len) + dest++; + } + } + else if (!S_ISDIR (st.st_mode) && *end != '\0') + { + __set_errno (ENOTDIR); + goto error; + } + } + } + if (dest > rpath + prefix_len + 1 && ISSLASH (dest[-1])) + --dest; + if (DOUBLE_SLASH_IS_DISTINCT_ROOT && dest == rpath + 1 && !prefix_len + && ISSLASH (*dest) && !ISSLASH (dest[1])) + dest++; + *dest = '\0'; + + if (extra_buf) + freea (extra_buf); + + return rpath; + +error: + { + int saved_errno = errno; + if (extra_buf) + freea (extra_buf); + if (resolved == NULL) + free (rpath); + __set_errno (saved_errno); + } + return NULL; +} +versioned_symbol (libc, __realpath, realpath, GLIBC_2_3); +#endif /* !FUNC_REALPATH_WORKS || defined _LIBC */ + + +#if SHLIB_COMPAT(libc, GLIBC_2_0, GLIBC_2_3) +char * +attribute_compat_text_section +__old_realpath (const char *name, char *resolved) +{ + if (resolved == NULL) + { + __set_errno (EINVAL); + return NULL; + } + + return __realpath (name, resolved); +} +compat_symbol (libc, __old_realpath, realpath, GLIBC_2_0); +#endif + + +char * +__canonicalize_file_name (const char *name) +{ + return __realpath (name, NULL); +} +weak_alias (__canonicalize_file_name, canonicalize_file_name) + +#else + +/* This declaration is solely to ensure that after preprocessing + this file is never empty. */ +typedef int dummy; + +#endif diff --git a/lib/cdefs.h b/lib/cdefs.h new file mode 100644 index 0000000..70fb638 --- /dev/null +++ b/lib/cdefs.h @@ -0,0 +1,514 @@ +/* Copyright (C) 1992-2018 Free Software Foundation, Inc. + This file is part of the GNU C Library. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU General Public + License as published by the Free Software Foundation; either + version 3 of the License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + General Public License for more details. + + You should have received a copy of the GNU General Public + License along with the GNU C Library; if not, see + <https://www.gnu.org/licenses/>. */ + +#ifndef _SYS_CDEFS_H +#define _SYS_CDEFS_H 1 + +/* We are almost always included from features.h. */ +#ifndef _FEATURES_H +# include <features.h> +#endif + +/* The GNU libc does not support any K&R compilers or the traditional mode + of ISO C compilers anymore. Check for some of the combinations not + anymore supported. */ +#if defined __GNUC__ && !defined __STDC__ +# error "You need a ISO C conforming compiler to use the glibc headers" +#endif + +/* Some user header file might have defined this before. */ +#undef __P +#undef __PMT + +#ifdef __GNUC__ + +/* All functions, except those with callbacks or those that + synchronize memory, are leaf functions. */ +# if __GNUC_PREREQ (4, 6) && !defined _LIBC +# define __LEAF , __leaf__ +# define __LEAF_ATTR __attribute__ ((__leaf__)) +# else +# define __LEAF +# define __LEAF_ATTR +# endif + +/* GCC can always grok prototypes. For C++ programs we add throw() + to help it optimize the function calls. But this works only with + gcc 2.8.x and egcs. For gcc 3.2 and up we even mark C functions + as non-throwing using a function attribute since programs can use + the -fexceptions options for C code as well. */ +# if !defined __cplusplus && __GNUC_PREREQ (3, 3) +# define __THROW __attribute__ ((__nothrow__ __LEAF)) +# define __THROWNL __attribute__ ((__nothrow__)) +# define __NTH(fct) __attribute__ ((__nothrow__ __LEAF)) fct +# define __NTHNL(fct) __attribute__ ((__nothrow__)) fct +# else +# if defined __cplusplus && __GNUC_PREREQ (2,8) +# define __THROW throw () +# define __THROWNL throw () +# define __NTH(fct) __LEAF_ATTR fct throw () +# define __NTHNL(fct) fct throw () +# else +# define __THROW +# define __THROWNL +# define __NTH(fct) fct +# define __NTHNL(fct) fct +# endif +# endif + +#else /* Not GCC. */ + +# if (defined __cplusplus \ + || (defined __STDC_VERSION__ && __STDC_VERSION__ >= 199901L)) +# define __inline inline +# else +# define __inline /* No inline functions. */ +# endif + +# define __THROW +# define __THROWNL +# define __NTH(fct) fct + +#endif /* GCC. */ + +/* Compilers that are not clang may object to + #if defined __clang__ && __has_extension(...) + even though they do not need to evaluate the right-hand side of the &&. */ +#if defined __clang__ && defined __has_extension +# define __glibc_clang_has_extension(ext) __has_extension (ext) +#else +# define __glibc_clang_has_extension(ext) 0 +#endif + +/* These two macros are not used in glibc anymore. They are kept here + only because some other projects expect the macros to be defined. */ +#define __P(args) args +#define __PMT(args) args + +/* For these things, GCC behaves the ANSI way normally, + and the non-ANSI way under -traditional. */ + +#define __CONCAT(x,y) x ## y +#define __STRING(x) #x + +/* This is not a typedef so `const __ptr_t' does the right thing. */ +#define __ptr_t void * + + +/* C++ needs to know that types and declarations are C, not C++. */ +#ifdef __cplusplus +# define __BEGIN_DECLS extern "C" { +# define __END_DECLS } +#else +# define __BEGIN_DECLS +# define __END_DECLS +#endif + + +/* Fortify support. */ +#define __bos(ptr) __builtin_object_size (ptr, __USE_FORTIFY_LEVEL > 1) +#define __bos0(ptr) __builtin_object_size (ptr, 0) + +#if __GNUC_PREREQ (4,3) +# define __warndecl(name, msg) \ + extern void name (void) __attribute__((__warning__ (msg))) +# define __warnattr(msg) __attribute__((__warning__ (msg))) +# define __errordecl(name, msg) \ + extern void name (void) __attribute__((__error__ (msg))) +#else +# define __warndecl(name, msg) extern void name (void) +# define __warnattr(msg) +# define __errordecl(name, msg) extern void name (void) +#endif + +/* Support for flexible arrays. + Headers that should use flexible arrays only if they're "real" + (e.g. only if they won't affect sizeof()) should test + #if __glibc_c99_flexarr_available. */ +#if defined __STDC_VERSION__ && __STDC_VERSION__ >= 199901L && !defined __HP_cc +# define __flexarr [] +# define __glibc_c99_flexarr_available 1 +#elif __GNUC_PREREQ (2,97) +/* GCC 2.97 supports C99 flexible array members as an extension, + even when in C89 mode or compiling C++ (any version). */ +# define __flexarr [] +# define __glibc_c99_flexarr_available 1 +#elif defined __GNUC__ +/* Pre-2.97 GCC did not support C99 flexible arrays but did have + an equivalent extension with slightly different notation. */ +# define __flexarr [0] +# define __glibc_c99_flexarr_available 1 +#else +/* Some other non-C99 compiler. Approximate with [1]. */ +# define __flexarr [1] +# define __glibc_c99_flexarr_available 0 +#endif + + +/* __asm__ ("xyz") is used throughout the headers to rename functions + at the assembly language level. This is wrapped by the __REDIRECT + macro, in order to support compilers that can do this some other + way. When compilers don't support asm-names at all, we have to do + preprocessor tricks instead (which don't have exactly the right + semantics, but it's the best we can do). + + Example: + int __REDIRECT(setpgrp, (__pid_t pid, __pid_t pgrp), setpgid); */ + +#if defined __GNUC__ && __GNUC__ >= 2 + +# define __REDIRECT(name, proto, alias) name proto __asm__ (__ASMNAME (#alias)) +# ifdef __cplusplus +# define __REDIRECT_NTH(name, proto, alias) \ + name proto __THROW __asm__ (__ASMNAME (#alias)) +# define __REDIRECT_NTHNL(name, proto, alias) \ + name proto __THROWNL __asm__ (__ASMNAME (#alias)) +# else +# define __REDIRECT_NTH(name, proto, alias) \ + name proto __asm__ (__ASMNAME (#alias)) __THROW +# define __REDIRECT_NTHNL(name, proto, alias) \ + name proto __asm__ (__ASMNAME (#alias)) __THROWNL +# endif +# define __ASMNAME(cname) __ASMNAME2 (__USER_LABEL_PREFIX__, cname) +# define __ASMNAME2(prefix, cname) __STRING (prefix) cname + +/* +#elif __SOME_OTHER_COMPILER__ + +# define __REDIRECT(name, proto, alias) name proto; \ + _Pragma("let " #name " = " #alias) +*/ +#endif + +/* GCC has various useful declarations that can be made with the + `__attribute__' syntax. All of the ways we use this do fine if + they are omitted for compilers that don't understand it. */ +#if !defined __GNUC__ || __GNUC__ < 2 +# define __attribute__(xyz) /* Ignore */ +#endif + +/* At some point during the gcc 2.96 development the `malloc' attribute + for functions was introduced. We don't want to use it unconditionally + (although this would be possible) since it generates warnings. */ +#if __GNUC_PREREQ (2,96) +# define __attribute_malloc__ __attribute__ ((__malloc__)) +#else +# define __attribute_malloc__ /* Ignore */ +#endif + +/* Tell the compiler which arguments to an allocation function + indicate the size of the allocation. */ +#if __GNUC_PREREQ (4, 3) +# define __attribute_alloc_size__(params) \ + __attribute__ ((__alloc_size__ params)) +#else +# define __attribute_alloc_size__(params) /* Ignore. */ +#endif + +/* At some point during the gcc 2.96 development the `pure' attribute + for functions was introduced. We don't want to use it unconditionally + (although this would be possible) since it generates warnings. */ +#if __GNUC_PREREQ (2,96) +# define __attribute_pure__ __attribute__ ((__pure__)) +#else +# define __attribute_pure__ /* Ignore */ +#endif + +/* This declaration tells the compiler that the value is constant. */ +#if __GNUC_PREREQ (2,5) +# define __attribute_const__ __attribute__ ((__const__)) +#else +# define __attribute_const__ /* Ignore */ +#endif + +/* At some point during the gcc 3.1 development the `used' attribute + for functions was introduced. We don't want to use it unconditionally + (although this would be possible) since it generates warnings. */ +#if __GNUC_PREREQ (3,1) +# define __attribute_used__ __attribute__ ((__used__)) +# define __attribute_noinline__ __attribute__ ((__noinline__)) +#else +# define __attribute_used__ __attribute__ ((__unused__)) +# define __attribute_noinline__ /* Ignore */ +#endif + +/* Since version 3.2, gcc allows marking deprecated functions. */ +#if __GNUC_PREREQ (3,2) +# define __attribute_deprecated__ __attribute__ ((__deprecated__)) +#else +# define __attribute_deprecated__ /* Ignore */ +#endif + +/* Since version 4.5, gcc also allows one to specify the message printed + when a deprecated function is used. clang claims to be gcc 4.2, but + may also support this feature. */ +#if __GNUC_PREREQ (4,5) || \ + __glibc_clang_has_extension (__attribute_deprecated_with_message__) +# define __attribute_deprecated_msg__(msg) \ + __attribute__ ((__deprecated__ (msg))) +#else +# define __attribute_deprecated_msg__(msg) __attribute_deprecated__ +#endif + +/* At some point during the gcc 2.8 development the `format_arg' attribute + for functions was introduced. We don't want to use it unconditionally + (although this would be possible) since it generates warnings. + If several `format_arg' attributes are given for the same function, in + gcc-3.0 and older, all but the last one are ignored. In newer gccs, + all designated arguments are considered. */ +#if __GNUC_PREREQ (2,8) +# define __attribute_format_arg__(x) __attribute__ ((__format_arg__ (x))) +#else +# define __attribute_format_arg__(x) /* Ignore */ +#endif + +/* At some point during the gcc 2.97 development the `strfmon' format + attribute for functions was introduced. We don't want to use it + unconditionally (although this would be possible) since it + generates warnings. */ +#if __GNUC_PREREQ (2,97) +# define __attribute_format_strfmon__(a,b) \ + __attribute__ ((__format__ (__strfmon__, a, b))) +#else +# define __attribute_format_strfmon__(a,b) /* Ignore */ +#endif + +/* The nonnull function attribute marks pointer parameters that + must not be NULL. Do not define __nonnull if it is already defined, + for portability when this file is used in Gnulib. */ +#ifndef __nonnull +# if __GNUC_PREREQ (3,3) +# define __nonnull(params) __attribute__ ((__nonnull__ params)) +# else +# define __nonnull(params) +# endif +#endif + +/* If fortification mode, we warn about unused results of certain + function calls which can lead to problems. */ +#if __GNUC_PREREQ (3,4) +# define __attribute_warn_unused_result__ \ + __attribute__ ((__warn_unused_result__)) +# if defined __USE_FORTIFY_LEVEL && __USE_FORTIFY_LEVEL > 0 +# define __wur __attribute_warn_unused_result__ +# endif +#else +# define __attribute_warn_unused_result__ /* empty */ +#endif +#ifndef __wur +# define __wur /* Ignore */ +#endif + +/* Forces a function to be always inlined. */ +#if __GNUC_PREREQ (3,2) +/* The Linux kernel defines __always_inline in stddef.h (283d7573), and + it conflicts with this definition. Therefore undefine it first to + allow either header to be included first. */ +# undef __always_inline +# define __always_inline __inline __attribute__ ((__always_inline__)) +#else +# undef __always_inline +# define __always_inline __inline +#endif + +/* Associate error messages with the source location of the call site rather + than with the source location inside the function. */ +#if __GNUC_PREREQ (4,3) +# define __attribute_artificial__ __attribute__ ((__artificial__)) +#else +# define __attribute_artificial__ /* Ignore */ +#endif + +/* GCC 4.3 and above with -std=c99 or -std=gnu99 implements ISO C99 + inline semantics, unless -fgnu89-inline is used. Using __GNUC_STDC_INLINE__ + or __GNUC_GNU_INLINE is not a good enough check for gcc because gcc versions + older than 4.3 may define these macros and still not guarantee GNU inlining + semantics. + + clang++ identifies itself as gcc-4.2, but has support for GNU inlining + semantics, that can be checked fot by using the __GNUC_STDC_INLINE_ and + __GNUC_GNU_INLINE__ macro definitions. */ +#if (!defined __cplusplus || __GNUC_PREREQ (4,3) \ + || (defined __clang__ && (defined __GNUC_STDC_INLINE__ \ + || defined __GNUC_GNU_INLINE__))) +# if defined __GNUC_STDC_INLINE__ || defined __cplusplus +# define __extern_inline extern __inline __attribute__ ((__gnu_inline__)) +# define __extern_always_inline \ + extern __always_inline __attribute__ ((__gnu_inline__)) +# else +# define __extern_inline extern __inline +# define __extern_always_inline extern __always_inline +# endif +#endif + +#ifdef __extern_always_inline +# define __fortify_function __extern_always_inline __attribute_artificial__ +#endif + +/* GCC 4.3 and above allow passing all anonymous arguments of an + __extern_always_inline function to some other vararg function. */ +#if __GNUC_PREREQ (4,3) +# define __va_arg_pack() __builtin_va_arg_pack () +# define __va_arg_pack_len() __builtin_va_arg_pack_len () +#endif + +/* It is possible to compile containing GCC extensions even if GCC is + run in pedantic mode if the uses are carefully marked using the + `__extension__' keyword. But this is not generally available before + version 2.8. */ +#if !__GNUC_PREREQ (2,8) +# define __extension__ /* Ignore */ +#endif + +/* __restrict is known in EGCS 1.2 and above. */ +#if !__GNUC_PREREQ (2,92) +# if defined __STDC_VERSION__ && __STDC_VERSION__ >= 199901L +# define __restrict restrict +# else +# define __restrict /* Ignore */ +# endif +#endif + +/* ISO C99 also allows to declare arrays as non-overlapping. The syntax is + array_name[restrict] + GCC 3.1 supports this. */ +#if __GNUC_PREREQ (3,1) && !defined __GNUG__ +# define __restrict_arr __restrict +#else +# ifdef __GNUC__ +# define __restrict_arr /* Not supported in old GCC. */ +# else +# if defined __STDC_VERSION__ && __STDC_VERSION__ >= 199901L +# define __restrict_arr restrict +# else +/* Some other non-C99 compiler. */ +# define __restrict_arr /* Not supported. */ +# endif +# endif +#endif + +#if __GNUC__ >= 3 +# define __glibc_unlikely(cond) __builtin_expect ((cond), 0) +# define __glibc_likely(cond) __builtin_expect ((cond), 1) +#else +# define __glibc_unlikely(cond) (cond) +# define __glibc_likely(cond) (cond) +#endif + +#ifdef __has_attribute +# define __glibc_has_attribute(attr) __has_attribute (attr) +#else +# define __glibc_has_attribute(attr) 0 +#endif + +#if (!defined _Noreturn \ + && (defined __STDC_VERSION__ ? __STDC_VERSION__ : 0) < 201112 \ + && !__GNUC_PREREQ (4,7)) +# if __GNUC_PREREQ (2,8) +# define _Noreturn __attribute__ ((__noreturn__)) +# else +# define _Noreturn +# endif +#endif + +#if __GNUC_PREREQ (8, 0) +/* Describes a char array whose address can safely be passed as the first + argument to strncpy and strncat, as the char array is not necessarily + a NUL-terminated string. */ +# define __attribute_nonstring__ __attribute__ ((__nonstring__)) +#else +# define __attribute_nonstring__ +#endif + +#if (!defined _Static_assert && !defined __cplusplus \ + && (defined __STDC_VERSION__ ? __STDC_VERSION__ : 0) < 201112 \ + && (!__GNUC_PREREQ (4, 6) || defined __STRICT_ANSI__)) +# define _Static_assert(expr, diagnostic) \ + extern int (*__Static_assert_function (void)) \ + [!!sizeof (struct { int __error_if_negative: (expr) ? 2 : -1; })] +#endif + +/* The #ifndef lets Gnulib avoid including these on non-glibc + platforms, where the includes typically do not exist. */ +#ifndef __WORDSIZE +# include <bits/wordsize.h> +# include <bits/long-double.h> +#endif + +#if defined __LONG_DOUBLE_MATH_OPTIONAL && defined __NO_LONG_DOUBLE_MATH +# define __LDBL_COMPAT 1 +# ifdef __REDIRECT +# define __LDBL_REDIR1(name, proto, alias) __REDIRECT (name, proto, alias) +# define __LDBL_REDIR(name, proto) \ + __LDBL_REDIR1 (name, proto, __nldbl_##name) +# define __LDBL_REDIR1_NTH(name, proto, alias) __REDIRECT_NTH (name, proto, alias) +# define __LDBL_REDIR_NTH(name, proto) \ + __LDBL_REDIR1_NTH (name, proto, __nldbl_##name) +# define __LDBL_REDIR1_DECL(name, alias) \ + extern __typeof (name) name __asm (__ASMNAME (#alias)); +# define __LDBL_REDIR_DECL(name) \ + extern __typeof (name) name __asm (__ASMNAME ("__nldbl_" #name)); +# define __REDIRECT_LDBL(name, proto, alias) \ + __LDBL_REDIR1 (name, proto, __nldbl_##alias) +# define __REDIRECT_NTH_LDBL(name, proto, alias) \ + __LDBL_REDIR1_NTH (name, proto, __nldbl_##alias) +# endif +#endif +#if !defined __LDBL_COMPAT || !defined __REDIRECT +# define __LDBL_REDIR1(name, proto, alias) name proto +# define __LDBL_REDIR(name, proto) name proto +# define __LDBL_REDIR1_NTH(name, proto, alias) name proto __THROW +# define __LDBL_REDIR_NTH(name, proto) name proto __THROW +# define __LDBL_REDIR_DECL(name) +# ifdef __REDIRECT +# define __REDIRECT_LDBL(name, proto, alias) __REDIRECT (name, proto, alias) +# define __REDIRECT_NTH_LDBL(name, proto, alias) \ + __REDIRECT_NTH (name, proto, alias) +# endif +#endif + +/* __glibc_macro_warning (MESSAGE) issues warning MESSAGE. This is + intended for use in preprocessor macros. + + Note: MESSAGE must be a _single_ string; concatenation of string + literals is not supported. */ +#if __GNUC_PREREQ (4,8) || __glibc_clang_prereq (3,5) +# define __glibc_macro_warning1(message) _Pragma (#message) +# define __glibc_macro_warning(message) \ + __glibc_macro_warning1 (GCC warning message) +#else +# define __glibc_macro_warning(msg) +#endif + +/* Generic selection (ISO C11) is a C-only feature, available in GCC + since version 4.9. Previous versions do not provide generic + selection, even though they might set __STDC_VERSION__ to 201112L, + when in -std=c11 mode. Thus, we must check for !defined __GNUC__ + when testing __STDC_VERSION__ for generic selection support. + On the other hand, Clang also defines __GNUC__, so a clang-specific + check is required to enable the use of generic selection. */ +#if !defined __cplusplus \ + && (__GNUC_PREREQ (4, 9) \ + || __glibc_clang_has_extension (c_generic_selections) \ + || (!defined __GNUC__ && defined __STDC_VERSION__ \ + && __STDC_VERSION__ >= 201112L)) +# define __HAVE_GENERIC_SELECTION 1 +#else +# define __HAVE_GENERIC_SELECTION 0 +#endif + +#endif /* sys/cdefs.h */ diff --git a/lib/close-stream.c b/lib/close-stream.c new file mode 100644 index 0000000..3d5a52e --- /dev/null +++ b/lib/close-stream.c @@ -0,0 +1,78 @@ +/* Close a stream, with nicer error checking than fclose's. + + Copyright (C) 1998-2002, 2004, 2006-2018 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +#include <config.h> + +#include "close-stream.h" + +#include <errno.h> +#include <stdbool.h> + +#include "fpending.h" + +#if USE_UNLOCKED_IO +# include "unlocked-io.h" +#endif + +/* Close STREAM. Return 0 if successful, EOF (setting errno) + otherwise. A failure might set errno to 0 if the error number + cannot be determined. + + A failure with errno set to EPIPE may or may not indicate an error + situation worth signaling to the user. See the documentation of the + close_stdout_set_ignore_EPIPE function for details. + + If a program writes *anything* to STREAM, that program should close + STREAM and make sure that it succeeds before exiting. Otherwise, + suppose that you go to the extreme of checking the return status + of every function that does an explicit write to STREAM. The last + printf can succeed in writing to the internal stream buffer, and yet + the fclose(STREAM) could still fail (due e.g., to a disk full error) + when it tries to write out that buffered data. Thus, you would be + left with an incomplete output file and the offending program would + exit successfully. Even calling fflush is not always sufficient, + since some file systems (NFS and CODA) buffer written/flushed data + until an actual close call. + + Besides, it's wasteful to check the return value from every call + that writes to STREAM -- just let the internal stream state record + the failure. That's what the ferror test is checking below. */ + +int +close_stream (FILE *stream) +{ + const bool some_pending = (__fpending (stream) != 0); + const bool prev_fail = (ferror (stream) != 0); + const bool fclose_fail = (fclose (stream) != 0); + + /* Return an error indication if there was a previous failure or if + fclose failed, with one exception: ignore an fclose failure if + there was no previous error, no data remains to be flushed, and + fclose failed with EBADF. That can happen when a program like cp + is invoked like this 'cp a b >&-' (i.e., with standard output + closed) and doesn't generate any output (hence no previous error + and nothing to be flushed). */ + + if (prev_fail || (fclose_fail && (some_pending || errno != EBADF))) + { + if (! fclose_fail) + errno = 0; + return EOF; + } + + return 0; +} diff --git a/lib/close-stream.h b/lib/close-stream.h new file mode 100644 index 0000000..be3d419 --- /dev/null +++ b/lib/close-stream.h @@ -0,0 +1,2 @@ +#include <stdio.h> +int close_stream (FILE *stream); diff --git a/lib/closeout.c b/lib/closeout.c new file mode 100644 index 0000000..0672732 --- /dev/null +++ b/lib/closeout.c @@ -0,0 +1,136 @@ +/* Close standard output and standard error, exiting with a diagnostic on error. + + Copyright (C) 1998-2002, 2004, 2006, 2008-2018 Free Software Foundation, + Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +#include <config.h> + +#include "closeout.h" + +#include <errno.h> +#include <stdbool.h> +#include <stdio.h> +#include <unistd.h> + +#include "gettext.h" +#define _(msgid) gettext (msgid) + +#include "close-stream.h" +#include "error.h" +#include "exitfail.h" +#include "quotearg.h" + +#ifndef __has_feature +# define __has_feature(a) false +#endif + +#if defined __SANITIZE_ADDRESS__ || __has_feature (address_sanitizer) +enum { SANITIZE_ADDRESS = true }; +#else +enum { SANITIZE_ADDRESS = false }; +#endif + +static const char *file_name; + +/* Set the file name to be reported in the event an error is detected + by close_stdout. */ +void +close_stdout_set_file_name (const char *file) +{ + file_name = file; +} + +static bool ignore_EPIPE /* = false */; + +/* Specify the reaction to an EPIPE error during the closing of stdout: + - If ignore = true, it shall be ignored. + - If ignore = false, it shall evoke a diagnostic, along with a nonzero + exit status. + The default is ignore = false. + + This setting matters only if the SIGPIPE signal is ignored (i.e. its + handler set to SIG_IGN) or blocked. Only particular programs need to + temporarily ignore SIGPIPE. If SIGPIPE is ignored or blocked because + it was ignored or blocked in the parent process when it created the + child process, it usually is a bug in the parent process: It is bad + practice to have SIGPIPE ignored or blocked while creating a child + process. + + EPIPE occurs when writing to a pipe or socket that has no readers now, + when SIGPIPE is ignored or blocked. + + The ignore = false setting is suitable for a scenario where it is normally + guaranteed that the pipe writer terminates before the pipe reader. In + this case, an EPIPE is an indication of a premature termination of the + pipe reader and should lead to a diagnostic and a nonzero exit status. + + The ignore = true setting is suitable for a scenario where you don't know + ahead of time whether the pipe writer or the pipe reader will terminate + first. In this case, an EPIPE is an indication that the pipe writer can + stop doing useless write() calls; this is what close_stdout does anyway. + EPIPE is part of the normal pipe/socket shutdown protocol in this case, + and should not lead to a diagnostic message. */ + +void +close_stdout_set_ignore_EPIPE (bool ignore) +{ + ignore_EPIPE = ignore; +} + +/* Close standard output. On error, issue a diagnostic and _exit + with status 'exit_failure'. + + Also close standard error. On error, _exit with status 'exit_failure'. + + Since close_stdout is commonly registered via 'atexit', POSIX + and the C standard both say that it should not call 'exit', + because the behavior is undefined if 'exit' is called more than + once. So it calls '_exit' instead of 'exit'. If close_stdout + is registered via atexit before other functions are registered, + the other functions can act before this _exit is invoked. + + Applications that use close_stdout should flush any streams + other than stdout and stderr before exiting, since the call to + _exit will bypass other buffer flushing. Applications should + be flushing and closing other streams anyway, to check for I/O + errors. Also, applications should not use tmpfile, since _exit + can bypass the removal of these files. + + It's important to detect such failures and exit nonzero because many + tools (most notably 'make' and other build-management systems) depend + on being able to detect failure in other tools via their exit status. */ + +void +close_stdout (void) +{ + if (close_stream (stdout) != 0 + && !(ignore_EPIPE && errno == EPIPE)) + { + char const *write_error = _("write error"); + if (file_name) + error (0, errno, "%s: %s", quotearg_colon (file_name), + write_error); + else + error (0, errno, "%s", write_error); + + _exit (exit_failure); + } + + /* Close stderr only if not sanitizing, as sanitizers may report to + stderr after this function returns. */ + if (!SANITIZE_ADDRESS && close_stream (stderr) != 0) + _exit (exit_failure); +} diff --git a/lib/closeout.h b/lib/closeout.h new file mode 100644 index 0000000..6629d49 --- /dev/null +++ b/lib/closeout.h @@ -0,0 +1,36 @@ +/* Close standard output and standard error. + + Copyright (C) 1998, 2000, 2003-2004, 2006, 2008-2018 Free Software + Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +#ifndef CLOSEOUT_H +# define CLOSEOUT_H 1 + +# include <stdbool.h> + +# ifdef __cplusplus +extern "C" { +# endif + +void close_stdout_set_file_name (const char *file); +void close_stdout_set_ignore_EPIPE (bool ignore); +void close_stdout (void); + +# ifdef __cplusplus +} +# endif + +#endif diff --git a/lib/copy-acl.c b/lib/copy-acl.c new file mode 100644 index 0000000..f5699ef --- /dev/null +++ b/lib/copy-acl.c @@ -0,0 +1,61 @@ +/* Copy access control list from one file to file. -*- coding: utf-8 -*- + + Copyright (C) 2002-2003, 2005-2018 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. + + Written by Paul Eggert, Andreas Grünbacher, and Bruno Haible. */ + +#include <config.h> + +#include "acl.h" + +#include <errno.h> + +#include "quote.h" +#include "error.h" +#include "gettext.h" +#define _(msgid) gettext (msgid) + + +/* Copy access control lists from one file to another. If SOURCE_DESC is + a valid file descriptor, use file descriptor operations, else use + filename based operations on SRC_NAME. Likewise for DEST_DESC and + DST_NAME. + If access control lists are not available, fchmod the target file to + MODE. Also sets the non-permission bits of the destination file + (S_ISUID, S_ISGID, S_ISVTX) to those from MODE if any are set. + Return 0 if successful, otherwise output a diagnostic and return a + negative error code. */ + +int +copy_acl (const char *src_name, int source_desc, const char *dst_name, + int dest_desc, mode_t mode) +{ + int ret = qcopy_acl (src_name, source_desc, dst_name, dest_desc, mode); + switch (ret) + { + case -2: + error (0, errno, "%s", quote (src_name)); + break; + + case -1: + error (0, errno, _("preserving permissions for %s"), quote (dst_name)); + break; + + default: + break; + } + return ret; +} diff --git a/lib/ctype.in.h b/lib/ctype.in.h new file mode 100644 index 0000000..cfcb118 --- /dev/null +++ b/lib/ctype.in.h @@ -0,0 +1,57 @@ +/* A substitute for ISO C99 <ctype.h>, for platforms on which it is incomplete. + + Copyright (C) 2009-2018 Free Software Foundation, Inc. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3, or (at your option) + any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, see <https://www.gnu.org/licenses/>. */ + +/* Written by Bruno Haible. */ + +/* + * ISO C 99 <ctype.h> for platforms on which it is incomplete. + * <http://www.opengroup.org/onlinepubs/9699919799/basedefs/ctype.h.html> + */ + +#ifndef _@GUARD_PREFIX@_CTYPE_H + +#if __GNUC__ >= 3 +@PRAGMA_SYSTEM_HEADER@ +#endif +@PRAGMA_COLUMNS@ + +/* Include the original <ctype.h>. */ +/* The include_next requires a split double-inclusion guard. */ +#@INCLUDE_NEXT@ @NEXT_CTYPE_H@ + +#ifndef _@GUARD_PREFIX@_CTYPE_H +#define _@GUARD_PREFIX@_CTYPE_H + +/* The definitions of _GL_FUNCDECL_RPL etc. are copied here. */ + +/* The definition of _GL_WARN_ON_USE is copied here. */ + +/* Return non-zero if c is a blank, i.e. a space or tab character. */ +#if @GNULIB_ISBLANK@ +# if !@HAVE_ISBLANK@ +_GL_EXTERN_C int isblank (int c); +# endif +#elif defined GNULIB_POSIXCHECK +# undef isblank +# if HAVE_RAW_DECL_ISBLANK +_GL_WARN_ON_USE (isblank, "isblank is unportable - " + "use gnulib module isblank for portability"); +# endif +#endif + +#endif /* _@GUARD_PREFIX@_CTYPE_H */ +#endif /* _@GUARD_PREFIX@_CTYPE_H */ diff --git a/lib/dfa.c b/lib/dfa.c new file mode 100644 index 0000000..c9be5e8 --- /dev/null +++ b/lib/dfa.c @@ -0,0 +1,4344 @@ +/* dfa.c - deterministic extended regexp routines for GNU + Copyright (C) 1988, 1998, 2000, 2002, 2004-2005, 2007-2018 Free Software + Foundation, Inc. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3, or (at your option) + any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, write to the Free Software + Foundation, Inc., + 51 Franklin Street - Fifth Floor, Boston, MA 02110-1301, USA */ + +/* Written June, 1988 by Mike Haertel + Modified July, 1988 by Arthur David Olson to assist BMG speedups */ + +#include <config.h> + +#include "dfa.h" + +#include <assert.h> +#include <ctype.h> +#include <stdint.h> +#include <stdio.h> +#include <stdlib.h> +#include <limits.h> +#include <string.h> +#include <locale.h> + +static bool +streq (char const *a, char const *b) +{ + return strcmp (a, b) == 0; +} + +static bool +isasciidigit (char c) +{ + return '0' <= c && c <= '9'; +} + +#include "gettext.h" +#define _(str) gettext (str) + +#include <wchar.h> + +#include "intprops.h" +#include "xalloc.h" +#include "localeinfo.h" + +#ifndef FALLTHROUGH +# if __GNUC__ < 7 +# define FALLTHROUGH ((void) 0) +# else +# define FALLTHROUGH __attribute__ ((__fallthrough__)) +# endif +#endif + +#ifndef MIN +# define MIN(a,b) ((a) < (b) ? (a) : (b)) +#endif + +/* HPUX defines these as macros in sys/param.h. */ +#ifdef setbit +# undef setbit +#endif +#ifdef clrbit +# undef clrbit +#endif + +/* First integer value that is greater than any character code. */ +enum { NOTCHAR = 1 << CHAR_BIT }; + +/* This represents part of a character class. It must be unsigned and + at least CHARCLASS_WORD_BITS wide. Any excess bits are zero. */ +typedef unsigned long int charclass_word; + +/* CHARCLASS_WORD_BITS is the number of bits used in a charclass word. + CHARCLASS_PAIR (LO, HI) is part of a charclass initializer, and + represents 64 bits' worth of a charclass, where LO and HI are the + low and high-order 32 bits of the 64-bit quantity. */ +#if ULONG_MAX >> 31 >> 31 < 3 +enum { CHARCLASS_WORD_BITS = 32 }; +# define CHARCLASS_PAIR(lo, hi) lo, hi +#else +enum { CHARCLASS_WORD_BITS = 64 }; +# define CHARCLASS_PAIR(lo, hi) (((charclass_word) (hi) << 32) + (lo)) +#endif + +/* An initializer for a charclass whose 32-bit words are A through H. */ +#define CHARCLASS_INIT(a, b, c, d, e, f, g, h) \ + {{ \ + CHARCLASS_PAIR (a, b), CHARCLASS_PAIR (c, d), \ + CHARCLASS_PAIR (e, f), CHARCLASS_PAIR (g, h) \ + }} + +/* The maximum useful value of a charclass_word; all used bits are 1. */ +static charclass_word const CHARCLASS_WORD_MASK + = ((charclass_word) 1 << (CHARCLASS_WORD_BITS - 1) << 1) - 1; + +/* Number of words required to hold a bit for every character. */ +enum +{ + CHARCLASS_WORDS = (NOTCHAR + CHARCLASS_WORD_BITS - 1) / CHARCLASS_WORD_BITS +}; + +/* Sets of unsigned characters are stored as bit vectors in arrays of ints. */ +typedef struct { charclass_word w[CHARCLASS_WORDS]; } charclass; + +/* Convert a possibly-signed character to an unsigned character. This is + a bit safer than casting to unsigned char, since it catches some type + errors that the cast doesn't. */ +static unsigned char +to_uchar (char ch) +{ + return ch; +} + +/* Contexts tell us whether a character is a newline or a word constituent. + Word-constituent characters are those that satisfy iswalnum, plus '_'. + Each character has a single CTX_* value; bitmasks of CTX_* values denote + a particular character class. + + A state also stores a context value, which is a bitmask of CTX_* values. + A state's context represents a set of characters that the state's + predecessors must match. For example, a state whose context does not + include CTX_LETTER will never have transitions where the previous + character is a word constituent. A state whose context is CTX_ANY + might have transitions from any character. */ + +enum + { + CTX_NONE = 1, + CTX_LETTER = 2, + CTX_NEWLINE = 4, + CTX_ANY = 7 + }; + +/* Sometimes characters can only be matched depending on the surrounding + context. Such context decisions depend on what the previous character + was, and the value of the current (lookahead) character. Context + dependent constraints are encoded as 9-bit integers. Each bit that + is set indicates that the constraint succeeds in the corresponding + context. + + bit 6-8 - valid contexts when next character is CTX_NEWLINE + bit 3-5 - valid contexts when next character is CTX_LETTER + bit 0-2 - valid contexts when next character is CTX_NONE + + succeeds_in_context determines whether a given constraint + succeeds in a particular context. Prev is a bitmask of possible + context values for the previous character, curr is the (single-bit) + context value for the lookahead character. */ +static int +newline_constraint (int constraint) +{ + return (constraint >> 6) & 7; +} +static int +letter_constraint (int constraint) +{ + return (constraint >> 3) & 7; +} +static int +other_constraint (int constraint) +{ + return constraint & 7; +} + +static bool +succeeds_in_context (int constraint, int prev, int curr) +{ + return !! (((curr & CTX_NONE ? other_constraint (constraint) : 0) \ + | (curr & CTX_LETTER ? letter_constraint (constraint) : 0) \ + | (curr & CTX_NEWLINE ? newline_constraint (constraint) : 0)) \ + & prev); +} + +/* The following describe what a constraint depends on. */ +static bool +prev_newline_dependent (int constraint) +{ + return ((constraint ^ constraint >> 2) & 0111) != 0; +} +static bool +prev_letter_dependent (int constraint) +{ + return ((constraint ^ constraint >> 1) & 0111) != 0; +} + +/* Tokens that match the empty string subject to some constraint actually + work by applying that constraint to determine what may follow them, + taking into account what has gone before. The following values are + the constraints corresponding to the special tokens previously defined. */ +enum + { + NO_CONSTRAINT = 0777, + BEGLINE_CONSTRAINT = 0444, + ENDLINE_CONSTRAINT = 0700, + BEGWORD_CONSTRAINT = 0050, + ENDWORD_CONSTRAINT = 0202, + LIMWORD_CONSTRAINT = 0252, + NOTLIMWORD_CONSTRAINT = 0525 + }; + +/* The regexp is parsed into an array of tokens in postfix form. Some tokens + are operators and others are terminal symbols. Most (but not all) of these + codes are returned by the lexical analyzer. */ + +typedef ptrdiff_t token; +static ptrdiff_t const TOKEN_MAX = PTRDIFF_MAX; + +/* States are indexed by state_num values. These are normally + nonnegative but -1 is used as a special value. */ +typedef ptrdiff_t state_num; + +/* Predefined token values. */ +enum +{ + END = -1, /* END is a terminal symbol that matches the + end of input; any value of END or less in + the parse tree is such a symbol. Accepting + states of the DFA are those that would have + a transition on END. This is -1, not some + more-negative value, to tweak the speed of + comparisons to END. */ + + /* Ordinary character values are terminal symbols that match themselves. */ + + /* CSET must come last in the following list of special tokens. Otherwise, + the list order matters only for performance. Related special tokens + should have nearby values so that code like (t == ANYCHAR || t == MBCSET + || CSET <= t) can be done with a single machine-level comparison. */ + + EMPTY = NOTCHAR, /* EMPTY is a terminal symbol that matches + the empty string. */ + + QMARK, /* QMARK is an operator of one argument that + matches zero or one occurrences of its + argument. */ + + STAR, /* STAR is an operator of one argument that + matches the Kleene closure (zero or more + occurrences) of its argument. */ + + PLUS, /* PLUS is an operator of one argument that + matches the positive closure (one or more + occurrences) of its argument. */ + + REPMN, /* REPMN is a lexical token corresponding + to the {m,n} construct. REPMN never + appears in the compiled token vector. */ + + CAT, /* CAT is an operator of two arguments that + matches the concatenation of its + arguments. CAT is never returned by the + lexical analyzer. */ + + OR, /* OR is an operator of two arguments that + matches either of its arguments. */ + + LPAREN, /* LPAREN never appears in the parse tree, + it is only a lexeme. */ + + RPAREN, /* RPAREN never appears in the parse tree. */ + + WCHAR, /* Only returned by lex. wctok contains + the wide character representation. */ + + ANYCHAR, /* ANYCHAR is a terminal symbol that matches + a valid multibyte (or single byte) character. + It is used only if MB_CUR_MAX > 1. */ + + BEG, /* BEG is an initial symbol that matches the + beginning of input. */ + + BEGLINE, /* BEGLINE is a terminal symbol that matches + the empty string at the beginning of a + line. */ + + ENDLINE, /* ENDLINE is a terminal symbol that matches + the empty string at the end of a line. */ + + BEGWORD, /* BEGWORD is a terminal symbol that matches + the empty string at the beginning of a + word. */ + + ENDWORD, /* ENDWORD is a terminal symbol that matches + the empty string at the end of a word. */ + + LIMWORD, /* LIMWORD is a terminal symbol that matches + the empty string at the beginning or the + end of a word. */ + + NOTLIMWORD, /* NOTLIMWORD is a terminal symbol that + matches the empty string not at + the beginning or end of a word. */ + + BACKREF, /* BACKREF is generated by \<digit> + or by any other construct that + is not completely handled. If the scanner + detects a transition on backref, it returns + a kind of "semi-success" indicating that + the match will have to be verified with + a backtracking matcher. */ + + MBCSET, /* MBCSET is similar to CSET, but for + multibyte characters. */ + + CSET /* CSET and (and any value greater) is a + terminal symbol that matches any of a + class of characters. */ +}; + + +/* States of the recognizer correspond to sets of positions in the parse + tree, together with the constraints under which they may be matched. + So a position is encoded as an index into the parse tree together with + a constraint. */ +typedef struct +{ + size_t index; /* Index into the parse array. */ + unsigned int constraint; /* Constraint for matching this position. */ +} position; + +/* Sets of positions are stored as arrays. */ +typedef struct +{ + position *elems; /* Elements of this position set. */ + ptrdiff_t nelem; /* Number of elements in this set. */ + ptrdiff_t alloc; /* Number of elements allocated in ELEMS. */ +} position_set; + +/* A state of the dfa consists of a set of positions, some flags, + and the token value of the lowest-numbered position of the state that + contains an END token. */ +typedef struct +{ + size_t hash; /* Hash of the positions of this state. */ + position_set elems; /* Positions this state could match. */ + unsigned char context; /* Context from previous state. */ + unsigned short constraint; /* Constraint for this state to accept. */ + token first_end; /* Token value of the first END in elems. */ + position_set mbps; /* Positions which can match multibyte + characters or the follows, e.g., period. + Used only if MB_CUR_MAX > 1. */ + state_num mb_trindex; /* Index of this state in MB_TRANS, or + negative if the state does not have + ANYCHAR. */ +} dfa_state; + +/* Maximum for any transition table count. This should be at least 3, + for the initial state setup. */ +enum { MAX_TRCOUNT = 1024 }; + +/* A bracket operator. + e.g., [a-c], [[:alpha:]], etc. */ +struct mb_char_classes +{ + ptrdiff_t cset; + bool invert; + wchar_t *chars; /* Normal characters. */ + ptrdiff_t nchars; + ptrdiff_t nchars_alloc; +}; + +struct regex_syntax +{ + /* Syntax bits controlling the behavior of the lexical analyzer. */ + reg_syntax_t syntax_bits; + bool syntax_bits_set; + + /* Flag for case-folding letters into sets. */ + bool case_fold; + + /* True if ^ and $ match only the start and end of data, and do not match + end-of-line within data. */ + bool anchor; + + /* End-of-line byte in data. */ + unsigned char eolbyte; + + /* Cache of char-context values. */ + char sbit[NOTCHAR]; + + /* If never_trail[B], the byte B cannot be a non-initial byte in a + multibyte character. */ + bool never_trail[NOTCHAR]; + + /* Set of characters considered letters. */ + charclass letters; + + /* Set of characters that are newline. */ + charclass newline; +}; + +/* Lexical analyzer. All the dross that deals with the obnoxious + GNU Regex syntax bits is located here. The poor, suffering + reader is referred to the GNU Regex documentation for the + meaning of the @#%!@#%^!@ syntax bits. */ +struct lexer_state +{ + char const *ptr; /* Pointer to next input character. */ + size_t left; /* Number of characters remaining. */ + token lasttok; /* Previous token returned; initially END. */ + size_t parens; /* Count of outstanding left parens. */ + int minrep, maxrep; /* Repeat counts for {m,n}. */ + + /* Wide character representation of the current multibyte character, + or WEOF if there was an encoding error. Used only if + MB_CUR_MAX > 1. */ + wint_t wctok; + + /* Length of the multibyte representation of wctok. */ + int cur_mb_len; + + /* The most recently analyzed multibyte bracket expression. */ + struct mb_char_classes brack; + + /* We're separated from beginning or (, | only by zero-width characters. */ + bool laststart; +}; + +/* Recursive descent parser for regular expressions. */ + +struct parser_state +{ + token tok; /* Lookahead token. */ + size_t depth; /* Current depth of a hypothetical stack + holding deferred productions. This is + used to determine the depth that will be + required of the real stack later on in + dfaanalyze. */ +}; + +/* A compiled regular expression. */ +struct dfa +{ + /* Syntax configuration */ + struct regex_syntax syntax; + + /* Fields filled by the scanner. */ + charclass *charclasses; /* Array of character sets for CSET tokens. */ + ptrdiff_t cindex; /* Index for adding new charclasses. */ + ptrdiff_t calloc; /* Number of charclasses allocated. */ + size_t canychar; /* Index of anychar class, or (size_t) -1. */ + + /* Scanner state */ + struct lexer_state lex; + + /* Parser state */ + struct parser_state parse; + + /* Fields filled by the parser. */ + token *tokens; /* Postfix parse array. */ + size_t tindex; /* Index for adding new tokens. */ + size_t talloc; /* Number of tokens currently allocated. */ + size_t depth; /* Depth required of an evaluation stack + used for depth-first traversal of the + parse tree. */ + size_t nleaves; /* Number of leaves on the parse tree. */ + size_t nregexps; /* Count of parallel regexps being built + with dfaparse. */ + bool fast; /* The DFA is fast. */ + token utf8_anychar_classes[5]; /* To lower ANYCHAR in UTF-8 locales. */ + mbstate_t mbs; /* Multibyte conversion state. */ + + /* The following are valid only if MB_CUR_MAX > 1. */ + + /* The value of multibyte_prop[i] is defined by following rule. + if tokens[i] < NOTCHAR + bit 0 : tokens[i] is the first byte of a character, including + single-byte characters. + bit 1 : tokens[i] is the last byte of a character, including + single-byte characters. + + e.g. + tokens + = 'single_byte_a', 'multi_byte_A', single_byte_b' + = 'sb_a', 'mb_A(1st byte)', 'mb_A(2nd byte)', 'mb_A(3rd byte)', 'sb_b' + multibyte_prop + = 3 , 1 , 0 , 2 , 3 + */ + char *multibyte_prop; + + /* Fields filled by the superset. */ + struct dfa *superset; /* Hint of the dfa. */ + + /* Fields filled by the state builder. */ + dfa_state *states; /* States of the dfa. */ + state_num sindex; /* Index for adding new states. */ + ptrdiff_t salloc; /* Number of states currently allocated. */ + + /* Fields filled by the parse tree->NFA conversion. */ + position_set *follows; /* Array of follow sets, indexed by position + index. The follow of a position is the set + of positions containing characters that + could conceivably follow a character + matching the given position in a string + matching the regexp. Allocated to the + maximum possible position index. */ + bool searchflag; /* We are supposed to build a searching + as opposed to an exact matcher. A searching + matcher finds the first and shortest string + matching a regexp anywhere in the buffer, + whereas an exact matcher finds the longest + string matching, but anchored to the + beginning of the buffer. */ + + /* Fields filled by dfaanalyze. */ + int *constraints; /* Array of union of accepting constraints + in the follow of a position. */ + int *separates; /* Array of contexts on follow of a + position. */ + + /* Fields filled by dfaexec. */ + state_num tralloc; /* Number of transition tables that have + slots so far, not counting trans[-1] and + trans[-2]. */ + int trcount; /* Number of transition tables that have + been built, other than for initial + states. */ + int min_trcount; /* Number of initial states. Equivalently, + the minimum state number for which trcount + counts transitions. */ + state_num **trans; /* Transition tables for states that can + never accept. If the transitions for a + state have not yet been computed, or the + state could possibly accept, its entry in + this table is NULL. This points to two + past the start of the allocated array, + and trans[-1] and trans[-2] are always + NULL. */ + state_num **fails; /* Transition tables after failing to accept + on a state that potentially could do so. + If trans[i] is non-null, fails[i] must + be null. */ + char *success; /* Table of acceptance conditions used in + dfaexec and computed in build_state. */ + state_num *newlines; /* Transitions on newlines. The entry for a + newline in any transition table is always + -1 so we can count lines without wasting + too many cycles. The transition for a + newline is stored separately and handled + as a special case. Newline is also used + as a sentinel at the end of the buffer. */ + state_num initstate_notbol; /* Initial state for CTX_LETTER and CTX_NONE + context in multibyte locales, in which we + do not distinguish between their contexts, + as not supported word. */ + position_set mb_follows; /* Follow set added by ANYCHAR on demand. */ + state_num **mb_trans; /* Transition tables for states with + ANYCHAR. */ + state_num mb_trcount; /* Number of transition tables for states with + ANYCHAR that have actually been built. */ + + /* Information derived from the locale. This is at the end so that + a quick memset need not clear it specially. */ + + /* dfaexec implementation. */ + char *(*dfaexec) (struct dfa *, char const *, char *, + bool, size_t *, bool *); + + /* The locale is simple, like the C locale. These locales can be + processed more efficiently, as they are single-byte, their native + character set is in collating-sequence order, and they do not + have multi-character collating elements. */ + bool simple_locale; + + /* Other cached information derived from the locale. */ + struct localeinfo localeinfo; +}; + +/* User access to dfa internals. */ + +/* S could possibly be an accepting state of R. */ +static bool +accepting (state_num s, struct dfa const *r) +{ + return r->states[s].constraint != 0; +} + +/* STATE accepts in the specified context. */ +static bool +accepts_in_context (int prev, int curr, state_num state, struct dfa const *dfa) +{ + return succeeds_in_context (dfa->states[state].constraint, prev, curr); +} + +static void regexp (struct dfa *dfa); + +/* Store into *PWC the result of converting the leading bytes of the + multibyte buffer S of length N bytes, using D->localeinfo.sbctowc + and updating the conversion state in *D. On conversion error, + convert just a single byte, to WEOF. Return the number of bytes + converted. + + This differs from mbrtowc (PWC, S, N, &D->mbs) as follows: + + * PWC points to wint_t, not to wchar_t. + * The last arg is a dfa *D instead of merely a multibyte conversion + state D->mbs. + * N must be at least 1. + * S[N - 1] must be a sentinel byte. + * Shift encodings are not supported. + * The return value is always in the range 1..N. + * D->mbs is always valid afterwards. + * *PWC is always set to something. */ +static size_t +mbs_to_wchar (wint_t *pwc, char const *s, size_t n, struct dfa *d) +{ + unsigned char uc = s[0]; + wint_t wc = d->localeinfo.sbctowc[uc]; + + if (wc == WEOF) + { + wchar_t wch; + size_t nbytes = mbrtowc (&wch, s, n, &d->mbs); + if (0 < nbytes && nbytes < (size_t) -2) + { + *pwc = wch; + return nbytes; + } + memset (&d->mbs, 0, sizeof d->mbs); + } + + *pwc = wc; + return 1; +} + +#ifdef DEBUG + +static void +prtok (token t) +{ + if (t <= END) + fprintf (stderr, "END"); + else if (0 <= t && t < NOTCHAR) + { + unsigned int ch = t; + fprintf (stderr, "0x%02x", ch); + } + else + { + char const *s; + switch (t) + { + case BEG: + s = "BEG"; + break; + case EMPTY: + s = "EMPTY"; + break; + case BACKREF: + s = "BACKREF"; + break; + case BEGLINE: + s = "BEGLINE"; + break; + case ENDLINE: + s = "ENDLINE"; + break; + case BEGWORD: + s = "BEGWORD"; + break; + case ENDWORD: + s = "ENDWORD"; + break; + case LIMWORD: + s = "LIMWORD"; + break; + case NOTLIMWORD: + s = "NOTLIMWORD"; + break; + case QMARK: + s = "QMARK"; + break; + case STAR: + s = "STAR"; + break; + case PLUS: + s = "PLUS"; + break; + case CAT: + s = "CAT"; + break; + case OR: + s = "OR"; + break; + case LPAREN: + s = "LPAREN"; + break; + case RPAREN: + s = "RPAREN"; + break; + case ANYCHAR: + s = "ANYCHAR"; + break; + case MBCSET: + s = "MBCSET"; + break; + default: + s = "CSET"; + break; + } + fprintf (stderr, "%s", s); + } +} +#endif /* DEBUG */ + +/* Stuff pertaining to charclasses. */ + +static bool +tstbit (unsigned int b, charclass const *c) +{ + return c->w[b / CHARCLASS_WORD_BITS] >> b % CHARCLASS_WORD_BITS & 1; +} + +static void +setbit (unsigned int b, charclass *c) +{ + charclass_word one = 1; + c->w[b / CHARCLASS_WORD_BITS] |= one << b % CHARCLASS_WORD_BITS; +} + +static void +clrbit (unsigned int b, charclass *c) +{ + charclass_word one = 1; + c->w[b / CHARCLASS_WORD_BITS] &= ~(one << b % CHARCLASS_WORD_BITS); +} + +static void +zeroset (charclass *s) +{ + memset (s, 0, sizeof *s); +} + +static void +fillset (charclass *s) +{ + for (int i = 0; i < CHARCLASS_WORDS; i++) + s->w[i] = CHARCLASS_WORD_MASK; +} + +static void +notset (charclass *s) +{ + for (int i = 0; i < CHARCLASS_WORDS; ++i) + s->w[i] = CHARCLASS_WORD_MASK & ~s->w[i]; +} + +static bool +equal (charclass const *s1, charclass const *s2) +{ + charclass_word w = 0; + for (int i = 0; i < CHARCLASS_WORDS; i++) + w |= s1->w[i] ^ s2->w[i]; + return w == 0; +} + +static bool +emptyset (charclass const *s) +{ + charclass_word w = 0; + for (int i = 0; i < CHARCLASS_WORDS; i++) + w |= s->w[i]; + return w == 0; +} + +/* Grow PA, which points to an array of *NITEMS items, and return the + location of the reallocated array, updating *NITEMS to reflect its + new size. The new array will contain at least NITEMS_INCR_MIN more + items, but will not contain more than NITEMS_MAX items total. + ITEM_SIZE is the size of each item, in bytes. + + ITEM_SIZE and NITEMS_INCR_MIN must be positive. *NITEMS must be + nonnegative. If NITEMS_MAX is -1, it is treated as if it were + infinity. + + If PA is null, then allocate a new array instead of reallocating + the old one. + + Thus, to grow an array A without saving its old contents, do + { free (A); A = xpalloc (NULL, &AITEMS, ...); }. */ + +static void * +xpalloc (void *pa, ptrdiff_t *nitems, ptrdiff_t nitems_incr_min, + ptrdiff_t nitems_max, ptrdiff_t item_size) +{ + ptrdiff_t n0 = *nitems; + + /* The approximate size to use for initial small allocation + requests. This is the largest "small" request for the GNU C + library malloc. */ + enum { DEFAULT_MXFAST = 64 * sizeof (size_t) / 4 }; + + /* If the array is tiny, grow it to about (but no greater than) + DEFAULT_MXFAST bytes. Otherwise, grow it by about 50%. + Adjust the growth according to three constraints: NITEMS_INCR_MIN, + NITEMS_MAX, and what the C language can represent safely. */ + + ptrdiff_t n, nbytes; + if (INT_ADD_WRAPV (n0, n0 >> 1, &n)) + n = PTRDIFF_MAX; + if (0 <= nitems_max && nitems_max < n) + n = nitems_max; + + ptrdiff_t adjusted_nbytes + = ((INT_MULTIPLY_WRAPV (n, item_size, &nbytes) || SIZE_MAX < nbytes) + ? MIN (PTRDIFF_MAX, SIZE_MAX) + : nbytes < DEFAULT_MXFAST ? DEFAULT_MXFAST : 0); + if (adjusted_nbytes) + { + n = adjusted_nbytes / item_size; + nbytes = adjusted_nbytes - adjusted_nbytes % item_size; + } + + if (! pa) + *nitems = 0; + if (n - n0 < nitems_incr_min + && (INT_ADD_WRAPV (n0, nitems_incr_min, &n) + || (0 <= nitems_max && nitems_max < n) + || INT_MULTIPLY_WRAPV (n, item_size, &nbytes))) + xalloc_die (); + pa = xrealloc (pa, nbytes); + *nitems = n; + return pa; +} + +/* Ensure that the array addressed by PA holds at least I + 1 items. + Either return PA, or reallocate the array and return its new address. + Although PA may be null, the returned value is never null. + + The array holds *NITEMS items, where 0 <= I <= *NITEMS; *NITEMS + is updated on reallocation. If PA is null, *NITEMS must be zero. + Do not allocate more than NITEMS_MAX items total; -1 means no limit. + ITEM_SIZE is the size of one item; it must be positive. + Avoid O(N**2) behavior on arrays growing linearly. */ +static void * +maybe_realloc (void *pa, ptrdiff_t i, ptrdiff_t *nitems, + ptrdiff_t nitems_max, ptrdiff_t item_size) +{ + if (i < *nitems) + return pa; + return xpalloc (pa, nitems, 1, nitems_max, item_size); +} + +/* In DFA D, find the index of charclass S, or allocate a new one. */ +static ptrdiff_t +charclass_index (struct dfa *d, charclass *s) +{ + ptrdiff_t i; + + for (i = 0; i < d->cindex; ++i) + if (equal (s, &d->charclasses[i])) + return i; + d->charclasses = maybe_realloc (d->charclasses, d->cindex, &d->calloc, + TOKEN_MAX - CSET, sizeof *d->charclasses); + ++d->cindex; + d->charclasses[i] = *s; + return i; +} + +static bool +unibyte_word_constituent (struct dfa const *dfa, unsigned char c) +{ + return dfa->localeinfo.sbctowc[c] != WEOF && (isalnum (c) || (c) == '_'); +} + +static int +char_context (struct dfa const *dfa, unsigned char c) +{ + if (c == dfa->syntax.eolbyte && !dfa->syntax.anchor) + return CTX_NEWLINE; + if (unibyte_word_constituent (dfa, c)) + return CTX_LETTER; + return CTX_NONE; +} + +/* Set a bit in the charclass for the given wchar_t. Do nothing if WC + is represented by a multi-byte sequence. Even for MB_CUR_MAX == 1, + this may happen when folding case in weird Turkish locales where + dotless i/dotted I are not included in the chosen character set. + Return whether a bit was set in the charclass. */ +static bool +setbit_wc (wint_t wc, charclass *c) +{ + int b = wctob (wc); + if (b < 0) + return false; + + setbit (b, c); + return true; +} + +/* Set a bit for B and its case variants in the charclass C. + MB_CUR_MAX must be 1. */ +static void +setbit_case_fold_c (int b, charclass *c) +{ + int ub = toupper (b); + for (int i = 0; i < NOTCHAR; i++) + if (toupper (i) == ub) + setbit (i, c); +} + +/* Return true if the locale compatible with the C locale. */ + +static bool +using_simple_locale (bool multibyte) +{ + /* The native character set is known to be compatible with + the C locale. The following test isn't perfect, but it's good + enough in practice, as only ASCII and EBCDIC are in common use + and this test correctly accepts ASCII and rejects EBCDIC. */ + enum { native_c_charset = + ('\b' == 8 && '\t' == 9 && '\n' == 10 && '\v' == 11 && '\f' == 12 + && '\r' == 13 && ' ' == 32 && '!' == 33 && '"' == 34 && '#' == 35 + && '%' == 37 && '&' == 38 && '\'' == 39 && '(' == 40 && ')' == 41 + && '*' == 42 && '+' == 43 && ',' == 44 && '-' == 45 && '.' == 46 + && '/' == 47 && '0' == 48 && '9' == 57 && ':' == 58 && ';' == 59 + && '<' == 60 && '=' == 61 && '>' == 62 && '?' == 63 && 'A' == 65 + && 'Z' == 90 && '[' == 91 && '\\' == 92 && ']' == 93 && '^' == 94 + && '_' == 95 && 'a' == 97 && 'z' == 122 && '{' == 123 && '|' == 124 + && '}' == 125 && '~' == 126) + }; + + if (!native_c_charset || multibyte) + return false; + else + { + /* Treat C and POSIX locales as being compatible. Also, treat + errors as compatible, as these are invariably from stubs. */ + char const *loc = setlocale (LC_ALL, NULL); + return !loc || streq (loc, "C") || streq (loc, "POSIX"); + } +} + +/* Fetch the next lexical input character from the pattern. There + must at least one byte of pattern input. Set DFA->lex.wctok to the + value of the character or to WEOF depending on whether the input is + a valid multibyte character (possibly of length 1). Then return + the next input byte value, except return EOF if the input is a + multibyte character of length greater than 1. */ +static int +fetch_wc (struct dfa *dfa) +{ + size_t nbytes = mbs_to_wchar (&dfa->lex.wctok, dfa->lex.ptr, dfa->lex.left, + dfa); + dfa->lex.cur_mb_len = nbytes; + int c = nbytes == 1 ? to_uchar (dfa->lex.ptr[0]) : EOF; + dfa->lex.ptr += nbytes; + dfa->lex.left -= nbytes; + return c; +} + +/* If there is no more input, report an error about unbalanced brackets. + Otherwise, behave as with fetch_wc (DFA). */ +static int +bracket_fetch_wc (struct dfa *dfa) +{ + if (! dfa->lex.left) + dfaerror (_("unbalanced [")); + return fetch_wc (dfa); +} + +typedef int predicate (int); + +/* The following list maps the names of the Posix named character classes + to predicate functions that determine whether a given character is in + the class. The leading [ has already been eaten by the lexical + analyzer. */ +struct dfa_ctype +{ + const char *name; + predicate *func; + bool single_byte_only; +}; + +static const struct dfa_ctype prednames[] = { + {"alpha", isalpha, false}, + {"upper", isupper, false}, + {"lower", islower, false}, + {"digit", isdigit, true}, + {"xdigit", isxdigit, false}, + {"space", isspace, false}, + {"punct", ispunct, false}, + {"alnum", isalnum, false}, + {"print", isprint, false}, + {"graph", isgraph, false}, + {"cntrl", iscntrl, false}, + {"blank", isblank, false}, + {NULL, NULL, false} +}; + +static const struct dfa_ctype *_GL_ATTRIBUTE_PURE +find_pred (const char *str) +{ + for (unsigned int i = 0; prednames[i].name; ++i) + if (streq (str, prednames[i].name)) + return &prednames[i]; + return NULL; +} + +/* Parse a bracket expression, which possibly includes multibyte + characters. */ +static token +parse_bracket_exp (struct dfa *dfa) +{ + /* This is a bracket expression that dfaexec is known to + process correctly. */ + bool known_bracket_exp = true; + + /* Used to warn about [:space:]. + Bit 0 = first character is a colon. + Bit 1 = last character is a colon. + Bit 2 = includes any other character but a colon. + Bit 3 = includes ranges, char/equiv classes or collation elements. */ + int colon_warning_state; + + dfa->lex.brack.nchars = 0; + charclass ccl; + zeroset (&ccl); + int c = bracket_fetch_wc (dfa); + bool invert = c == '^'; + if (invert) + { + c = bracket_fetch_wc (dfa); + known_bracket_exp = dfa->simple_locale; + } + wint_t wc = dfa->lex.wctok; + int c1; + wint_t wc1; + colon_warning_state = (c == ':'); + do + { + c1 = NOTCHAR; /* Mark c1 as not initialized. */ + colon_warning_state &= ~2; + + /* Note that if we're looking at some other [:...:] construct, + we just treat it as a bunch of ordinary characters. We can do + this because we assume regex has checked for syntax errors before + dfa is ever called. */ + if (c == '[') + { + c1 = bracket_fetch_wc (dfa); + wc1 = dfa->lex.wctok; + + if ((c1 == ':' && (dfa->syntax.syntax_bits & RE_CHAR_CLASSES)) + || c1 == '.' || c1 == '=') + { + enum { MAX_BRACKET_STRING_LEN = 32 }; + char str[MAX_BRACKET_STRING_LEN + 1]; + size_t len = 0; + for (;;) + { + c = bracket_fetch_wc (dfa); + if (dfa->lex.left == 0 + || (c == c1 && dfa->lex.ptr[0] == ']')) + break; + if (len < MAX_BRACKET_STRING_LEN) + str[len++] = c; + else + /* This is in any case an invalid class name. */ + str[0] = '\0'; + } + str[len] = '\0'; + + /* Fetch bracket. */ + c = bracket_fetch_wc (dfa); + wc = dfa->lex.wctok; + if (c1 == ':') + /* Build character class. POSIX allows character + classes to match multicharacter collating elements, + but the regex code does not support that, so do not + worry about that possibility. */ + { + char const *class + = (dfa->syntax.case_fold && (streq (str, "upper") + || streq (str, "lower")) + ? "alpha" : str); + const struct dfa_ctype *pred = find_pred (class); + if (!pred) + dfaerror (_("invalid character class")); + + if (dfa->localeinfo.multibyte && !pred->single_byte_only) + known_bracket_exp = false; + else + for (int c2 = 0; c2 < NOTCHAR; ++c2) + if (pred->func (c2)) + setbit (c2, &ccl); + } + else + known_bracket_exp = false; + + colon_warning_state |= 8; + + /* Fetch new lookahead character. */ + c1 = bracket_fetch_wc (dfa); + wc1 = dfa->lex.wctok; + continue; + } + + /* We treat '[' as a normal character here. c/c1/wc/wc1 + are already set up. */ + } + + if (c == '\\' + && (dfa->syntax.syntax_bits & RE_BACKSLASH_ESCAPE_IN_LISTS)) + { + c = bracket_fetch_wc (dfa); + wc = dfa->lex.wctok; + } + + if (c1 == NOTCHAR) + { + c1 = bracket_fetch_wc (dfa); + wc1 = dfa->lex.wctok; + } + + if (c1 == '-') + /* build range characters. */ + { + int c2 = bracket_fetch_wc (dfa); + wint_t wc2 = dfa->lex.wctok; + + /* A bracket expression like [a-[.aa.]] matches an unknown set. + Treat it like [-a[.aa.]] while parsing it, and + remember that the set is unknown. */ + if (c2 == '[' && dfa->lex.ptr[0] == '.') + { + known_bracket_exp = false; + c2 = ']'; + } + + if (c2 == ']') + { + /* In the case [x-], the - is an ordinary hyphen, + which is left in c1, the lookahead character. */ + dfa->lex.ptr -= dfa->lex.cur_mb_len; + dfa->lex.left += dfa->lex.cur_mb_len; + } + else + { + if (c2 == '\\' && (dfa->syntax.syntax_bits + & RE_BACKSLASH_ESCAPE_IN_LISTS)) + { + c2 = bracket_fetch_wc (dfa); + wc2 = dfa->lex.wctok; + } + + colon_warning_state |= 8; + c1 = bracket_fetch_wc (dfa); + wc1 = dfa->lex.wctok; + + /* Treat [x-y] as a range if x != y. */ + if (wc != wc2 || wc == WEOF) + { + if (dfa->simple_locale + || (isasciidigit (c) & isasciidigit (c2))) + { + for (int ci = c; ci <= c2; ci++) + if (dfa->syntax.case_fold && isalpha (ci)) + setbit_case_fold_c (ci, &ccl); + else + setbit (ci, &ccl); + } + else + known_bracket_exp = false; + + continue; + } + } + } + + colon_warning_state |= (c == ':') ? 2 : 4; + + if (!dfa->localeinfo.multibyte) + { + if (dfa->syntax.case_fold && isalpha (c)) + setbit_case_fold_c (c, &ccl); + else + setbit (c, &ccl); + continue; + } + + if (wc == WEOF) + known_bracket_exp = false; + else + { + wchar_t folded[CASE_FOLDED_BUFSIZE + 1]; + unsigned int n = (dfa->syntax.case_fold + ? case_folded_counterparts (wc, folded + 1) + 1 + : 1); + folded[0] = wc; + for (unsigned int i = 0; i < n; i++) + if (!setbit_wc (folded[i], &ccl)) + { + dfa->lex.brack.chars + = maybe_realloc (dfa->lex.brack.chars, dfa->lex.brack.nchars, + &dfa->lex.brack.nchars_alloc, -1, + sizeof *dfa->lex.brack.chars); + dfa->lex.brack.chars[dfa->lex.brack.nchars++] = folded[i]; + } + } + } + while ((wc = wc1, (c = c1) != ']')); + + if (colon_warning_state == 7) + dfawarn (_("character class syntax is [[:space:]], not [:space:]")); + + if (! known_bracket_exp) + return BACKREF; + + if (dfa->localeinfo.multibyte && (invert || dfa->lex.brack.nchars != 0)) + { + dfa->lex.brack.invert = invert; + dfa->lex.brack.cset = emptyset (&ccl) ? -1 : charclass_index (dfa, &ccl); + return MBCSET; + } + + if (invert) + { + notset (&ccl); + if (dfa->syntax.syntax_bits & RE_HAT_LISTS_NOT_NEWLINE) + clrbit ('\n', &ccl); + } + + return CSET + charclass_index (dfa, &ccl); +} + +struct lexptr +{ + char const *ptr; + size_t left; +}; + +static void +push_lex_state (struct dfa *dfa, struct lexptr *ls, char const *s) +{ + ls->ptr = dfa->lex.ptr; + ls->left = dfa->lex.left; + dfa->lex.ptr = s; + dfa->lex.left = strlen (s); +} + +static void +pop_lex_state (struct dfa *dfa, struct lexptr const *ls) +{ + dfa->lex.ptr = ls->ptr; + dfa->lex.left = ls->left; +} + +static token +lex (struct dfa *dfa) +{ + bool backslash = false; + + /* Basic plan: We fetch a character. If it's a backslash, + we set the backslash flag and go through the loop again. + On the plus side, this avoids having a duplicate of the + main switch inside the backslash case. On the minus side, + it means that just about every case begins with + "if (backslash) ...". */ + for (int i = 0; i < 2; ++i) + { + if (! dfa->lex.left) + return dfa->lex.lasttok = END; + int c = fetch_wc (dfa); + + switch (c) + { + case '\\': + if (backslash) + goto normal_char; + if (dfa->lex.left == 0) + dfaerror (_("unfinished \\ escape")); + backslash = true; + break; + + case '^': + if (backslash) + goto normal_char; + if (dfa->syntax.syntax_bits & RE_CONTEXT_INDEP_ANCHORS + || dfa->lex.lasttok == END || dfa->lex.lasttok == LPAREN + || dfa->lex.lasttok == OR) + return dfa->lex.lasttok = BEGLINE; + goto normal_char; + + case '$': + if (backslash) + goto normal_char; + if (dfa->syntax.syntax_bits & RE_CONTEXT_INDEP_ANCHORS + || dfa->lex.left == 0 + || ((dfa->lex.left + > !(dfa->syntax.syntax_bits & RE_NO_BK_PARENS)) + && (dfa->lex.ptr[!(dfa->syntax.syntax_bits & RE_NO_BK_PARENS) + & (dfa->lex.ptr[0] == '\\')] + == ')')) + || ((dfa->lex.left + > !(dfa->syntax.syntax_bits & RE_NO_BK_VBAR)) + && (dfa->lex.ptr[!(dfa->syntax.syntax_bits & RE_NO_BK_VBAR) + & (dfa->lex.ptr[0] == '\\')] + == '|')) + || ((dfa->syntax.syntax_bits & RE_NEWLINE_ALT) + && dfa->lex.left > 0 && dfa->lex.ptr[0] == '\n')) + return dfa->lex.lasttok = ENDLINE; + goto normal_char; + + case '1': + case '2': + case '3': + case '4': + case '5': + case '6': + case '7': + case '8': + case '9': + if (backslash && !(dfa->syntax.syntax_bits & RE_NO_BK_REFS)) + { + dfa->lex.laststart = false; + return dfa->lex.lasttok = BACKREF; + } + goto normal_char; + + case '`': + if (backslash && !(dfa->syntax.syntax_bits & RE_NO_GNU_OPS)) + { + /* FIXME: should be beginning of string */ + return dfa->lex.lasttok = BEGLINE; + } + goto normal_char; + + case '\'': + if (backslash && !(dfa->syntax.syntax_bits & RE_NO_GNU_OPS)) + { + /* FIXME: should be end of string */ + return dfa->lex.lasttok = ENDLINE; + } + goto normal_char; + + case '<': + if (backslash && !(dfa->syntax.syntax_bits & RE_NO_GNU_OPS)) + return dfa->lex.lasttok = BEGWORD; + goto normal_char; + + case '>': + if (backslash && !(dfa->syntax.syntax_bits & RE_NO_GNU_OPS)) + return dfa->lex.lasttok = ENDWORD; + goto normal_char; + + case 'b': + if (backslash && !(dfa->syntax.syntax_bits & RE_NO_GNU_OPS)) + return dfa->lex.lasttok = LIMWORD; + goto normal_char; + + case 'B': + if (backslash && !(dfa->syntax.syntax_bits & RE_NO_GNU_OPS)) + return dfa->lex.lasttok = NOTLIMWORD; + goto normal_char; + + case '?': + if (dfa->syntax.syntax_bits & RE_LIMITED_OPS) + goto normal_char; + if (backslash != ((dfa->syntax.syntax_bits & RE_BK_PLUS_QM) != 0)) + goto normal_char; + if (!(dfa->syntax.syntax_bits & RE_CONTEXT_INDEP_OPS) + && dfa->lex.laststart) + goto normal_char; + return dfa->lex.lasttok = QMARK; + + case '*': + if (backslash) + goto normal_char; + if (!(dfa->syntax.syntax_bits & RE_CONTEXT_INDEP_OPS) + && dfa->lex.laststart) + goto normal_char; + return dfa->lex.lasttok = STAR; + + case '+': + if (dfa->syntax.syntax_bits & RE_LIMITED_OPS) + goto normal_char; + if (backslash != ((dfa->syntax.syntax_bits & RE_BK_PLUS_QM) != 0)) + goto normal_char; + if (!(dfa->syntax.syntax_bits & RE_CONTEXT_INDEP_OPS) + && dfa->lex.laststart) + goto normal_char; + return dfa->lex.lasttok = PLUS; + + case '{': + if (!(dfa->syntax.syntax_bits & RE_INTERVALS)) + goto normal_char; + if (backslash != ((dfa->syntax.syntax_bits & RE_NO_BK_BRACES) == 0)) + goto normal_char; + if (!(dfa->syntax.syntax_bits & RE_CONTEXT_INDEP_OPS) + && dfa->lex.laststart) + goto normal_char; + + /* Cases: + {M} - exact count + {M,} - minimum count, maximum is infinity + {,N} - 0 through N + {,} - 0 to infinity (same as '*') + {M,N} - M through N */ + { + char const *p = dfa->lex.ptr; + char const *lim = p + dfa->lex.left; + dfa->lex.minrep = dfa->lex.maxrep = -1; + for (; p != lim && isasciidigit (*p); p++) + dfa->lex.minrep = (dfa->lex.minrep < 0 + ? *p - '0' + : MIN (RE_DUP_MAX + 1, + dfa->lex.minrep * 10 + *p - '0')); + if (p != lim) + { + if (*p != ',') + dfa->lex.maxrep = dfa->lex.minrep; + else + { + if (dfa->lex.minrep < 0) + dfa->lex.minrep = 0; + while (++p != lim && isasciidigit (*p)) + dfa->lex.maxrep + = (dfa->lex.maxrep < 0 + ? *p - '0' + : MIN (RE_DUP_MAX + 1, + dfa->lex.maxrep * 10 + *p - '0')); + } + } + if (! ((! backslash || (p != lim && *p++ == '\\')) + && p != lim && *p++ == '}' + && 0 <= dfa->lex.minrep + && (dfa->lex.maxrep < 0 + || dfa->lex.minrep <= dfa->lex.maxrep))) + { + if (dfa->syntax.syntax_bits & RE_INVALID_INTERVAL_ORD) + goto normal_char; + dfaerror (_("invalid content of \\{\\}")); + } + if (RE_DUP_MAX < dfa->lex.maxrep) + dfaerror (_("regular expression too big")); + dfa->lex.ptr = p; + dfa->lex.left = lim - p; + } + dfa->lex.laststart = false; + return dfa->lex.lasttok = REPMN; + + case '|': + if (dfa->syntax.syntax_bits & RE_LIMITED_OPS) + goto normal_char; + if (backslash != ((dfa->syntax.syntax_bits & RE_NO_BK_VBAR) == 0)) + goto normal_char; + dfa->lex.laststart = true; + return dfa->lex.lasttok = OR; + + case '\n': + if (dfa->syntax.syntax_bits & RE_LIMITED_OPS + || backslash || !(dfa->syntax.syntax_bits & RE_NEWLINE_ALT)) + goto normal_char; + dfa->lex.laststart = true; + return dfa->lex.lasttok = OR; + + case '(': + if (backslash != ((dfa->syntax.syntax_bits & RE_NO_BK_PARENS) == 0)) + goto normal_char; + dfa->lex.parens++; + dfa->lex.laststart = true; + return dfa->lex.lasttok = LPAREN; + + case ')': + if (backslash != ((dfa->syntax.syntax_bits & RE_NO_BK_PARENS) == 0)) + goto normal_char; + if (dfa->lex.parens == 0 + && dfa->syntax.syntax_bits & RE_UNMATCHED_RIGHT_PAREN_ORD) + goto normal_char; + dfa->lex.parens--; + dfa->lex.laststart = false; + return dfa->lex.lasttok = RPAREN; + + case '.': + if (backslash) + goto normal_char; + if (dfa->canychar == (size_t) -1) + { + charclass ccl; + fillset (&ccl); + if (!(dfa->syntax.syntax_bits & RE_DOT_NEWLINE)) + clrbit ('\n', &ccl); + if (dfa->syntax.syntax_bits & RE_DOT_NOT_NULL) + clrbit ('\0', &ccl); + if (dfa->localeinfo.multibyte) + for (int c2 = 0; c2 < NOTCHAR; c2++) + if (dfa->localeinfo.sbctowc[c2] == WEOF) + clrbit (c2, &ccl); + dfa->canychar = charclass_index (dfa, &ccl); + } + dfa->lex.laststart = false; + return dfa->lex.lasttok = (dfa->localeinfo.multibyte + ? ANYCHAR + : CSET + dfa->canychar); + + case 's': + case 'S': + if (!backslash || (dfa->syntax.syntax_bits & RE_NO_GNU_OPS)) + goto normal_char; + if (!dfa->localeinfo.multibyte) + { + charclass ccl; + zeroset (&ccl); + for (int c2 = 0; c2 < NOTCHAR; ++c2) + if (isspace (c2)) + setbit (c2, &ccl); + if (c == 'S') + notset (&ccl); + dfa->lex.laststart = false; + return dfa->lex.lasttok = CSET + charclass_index (dfa, &ccl); + } + + /* FIXME: see if optimizing this, as is done with ANYCHAR and + add_utf8_anychar, makes sense. */ + + /* \s and \S are documented to be equivalent to [[:space:]] and + [^[:space:]] respectively, so tell the lexer to process those + strings, each minus its "already processed" '['. */ + { + struct lexptr ls; + push_lex_state (dfa, &ls, &"^[:space:]]"[c == 's']); + dfa->lex.lasttok = parse_bracket_exp (dfa); + pop_lex_state (dfa, &ls); + } + + dfa->lex.laststart = false; + return dfa->lex.lasttok; + + case 'w': + case 'W': + if (!backslash || (dfa->syntax.syntax_bits & RE_NO_GNU_OPS)) + goto normal_char; + + if (!dfa->localeinfo.multibyte) + { + charclass ccl; + zeroset (&ccl); + for (int c2 = 0; c2 < NOTCHAR; ++c2) + if (dfa->syntax.sbit[c2] == CTX_LETTER) + setbit (c2, &ccl); + if (c == 'W') + notset (&ccl); + dfa->lex.laststart = false; + return dfa->lex.lasttok = CSET + charclass_index (dfa, &ccl); + } + + /* FIXME: see if optimizing this, as is done with ANYCHAR and + add_utf8_anychar, makes sense. */ + + /* \w and \W are documented to be equivalent to [_[:alnum:]] and + [^_[:alnum:]] respectively, so tell the lexer to process those + strings, each minus its "already processed" '['. */ + { + struct lexptr ls; + push_lex_state (dfa, &ls, &"^_[:alnum:]]"[c == 'w']); + dfa->lex.lasttok = parse_bracket_exp (dfa); + pop_lex_state (dfa, &ls); + } + + dfa->lex.laststart = false; + return dfa->lex.lasttok; + + case '[': + if (backslash) + goto normal_char; + dfa->lex.laststart = false; + return dfa->lex.lasttok = parse_bracket_exp (dfa); + + default: + normal_char: + dfa->lex.laststart = false; + /* For multibyte character sets, folding is done in atom. Always + return WCHAR. */ + if (dfa->localeinfo.multibyte) + return dfa->lex.lasttok = WCHAR; + + if (dfa->syntax.case_fold && isalpha (c)) + { + charclass ccl; + zeroset (&ccl); + setbit_case_fold_c (c, &ccl); + return dfa->lex.lasttok = CSET + charclass_index (dfa, &ccl); + } + + return dfa->lex.lasttok = c; + } + } + + /* The above loop should consume at most a backslash + and some other character. */ + abort (); + return END; /* keeps pedantic compilers happy. */ +} + +static void +addtok_mb (struct dfa *dfa, token t, char mbprop) +{ + if (dfa->talloc == dfa->tindex) + { + dfa->tokens = x2nrealloc (dfa->tokens, &dfa->talloc, + sizeof *dfa->tokens); + if (dfa->localeinfo.multibyte) + dfa->multibyte_prop = xnrealloc (dfa->multibyte_prop, dfa->talloc, + sizeof *dfa->multibyte_prop); + } + if (dfa->localeinfo.multibyte) + dfa->multibyte_prop[dfa->tindex] = mbprop; + dfa->tokens[dfa->tindex++] = t; + + switch (t) + { + case QMARK: + case STAR: + case PLUS: + break; + + case CAT: + case OR: + dfa->parse.depth--; + break; + + case BACKREF: + dfa->fast = false; + FALLTHROUGH; + default: + dfa->nleaves++; + FALLTHROUGH; + case EMPTY: + dfa->parse.depth++; + break; + } + if (dfa->parse.depth > dfa->depth) + dfa->depth = dfa->parse.depth; +} + +static void addtok_wc (struct dfa *dfa, wint_t wc); + +/* Add the given token to the parse tree, maintaining the depth count and + updating the maximum depth if necessary. */ +static void +addtok (struct dfa *dfa, token t) +{ + if (dfa->localeinfo.multibyte && t == MBCSET) + { + bool need_or = false; + + /* Extract wide characters into alternations for better performance. + This does not require UTF-8. */ + for (ptrdiff_t i = 0; i < dfa->lex.brack.nchars; i++) + { + addtok_wc (dfa, dfa->lex.brack.chars[i]); + if (need_or) + addtok (dfa, OR); + need_or = true; + } + dfa->lex.brack.nchars = 0; + + /* Wide characters have been handled above, so it is possible + that the set is empty now. Do nothing in that case. */ + if (dfa->lex.brack.cset != -1) + { + addtok (dfa, CSET + dfa->lex.brack.cset); + if (need_or) + addtok (dfa, OR); + } + } + else + { + addtok_mb (dfa, t, 3); + } +} + +/* We treat a multibyte character as a single atom, so that DFA + can treat a multibyte character as a single expression. + + e.g., we construct the following tree from "<mb1><mb2>". + <mb1(1st-byte)><mb1(2nd-byte)><CAT><mb1(3rd-byte)><CAT> + <mb2(1st-byte)><mb2(2nd-byte)><CAT><mb2(3rd-byte)><CAT><CAT> */ +static void +addtok_wc (struct dfa *dfa, wint_t wc) +{ + unsigned char buf[MB_LEN_MAX]; + mbstate_t s = { 0 }; + size_t stored_bytes = wcrtomb ((char *) buf, wc, &s); + + if (stored_bytes != (size_t) -1) + dfa->lex.cur_mb_len = stored_bytes; + else + { + /* This is merely stop-gap. buf[0] is undefined, yet skipping + the addtok_mb call altogether can corrupt the heap. */ + dfa->lex.cur_mb_len = 1; + buf[0] = 0; + } + + addtok_mb (dfa, buf[0], dfa->lex.cur_mb_len == 1 ? 3 : 1); + for (int i = 1; i < dfa->lex.cur_mb_len; i++) + { + addtok_mb (dfa, buf[i], i == dfa->lex.cur_mb_len - 1 ? 2 : 0); + addtok (dfa, CAT); + } +} + +static void +add_utf8_anychar (struct dfa *dfa) +{ + static charclass const utf8_classes[5] = { + /* 80-bf: non-leading bytes. */ + CHARCLASS_INIT (0, 0, 0, 0, 0xffffffff, 0xffffffff, 0, 0), + + /* 00-7f: 1-byte sequence. */ + CHARCLASS_INIT (0xffffffff, 0xffffffff, 0xffffffff, 0xffffffff, 0, 0, 0, 0), + + /* c2-df: 2-byte sequence. */ + CHARCLASS_INIT (0, 0, 0, 0, 0, 0, 0xfffffffc, 0), + + /* e0-ef: 3-byte sequence. */ + CHARCLASS_INIT (0, 0, 0, 0, 0, 0, 0, 0xffff), + + /* f0-f7: 4-byte sequence. */ + CHARCLASS_INIT (0, 0, 0, 0, 0, 0, 0, 0xff0000) + }; + const unsigned int n = sizeof (utf8_classes) / sizeof (utf8_classes[0]); + + /* Define the five character classes that are needed below. */ + if (dfa->utf8_anychar_classes[0] == 0) + for (unsigned int i = 0; i < n; i++) + { + charclass c = utf8_classes[i]; + if (i == 1) + { + if (!(dfa->syntax.syntax_bits & RE_DOT_NEWLINE)) + clrbit ('\n', &c); + if (dfa->syntax.syntax_bits & RE_DOT_NOT_NULL) + clrbit ('\0', &c); + } + dfa->utf8_anychar_classes[i] = CSET + charclass_index (dfa, &c); + } + + /* A valid UTF-8 character is + + ([0x00-0x7f] + |[0xc2-0xdf][0x80-0xbf] + |[0xe0-0xef[0x80-0xbf][0x80-0xbf] + |[0xf0-f7][0x80-0xbf][0x80-0xbf][0x80-0xbf]) + + which I'll write more concisely "B|CA|DAA|EAAA". Factor the [0x00-0x7f] + and you get "B|(C|(D|EA)A)A". And since the token buffer is in reverse + Polish notation, you get "B C D E A CAT OR A CAT OR A CAT OR". */ + unsigned int i; + for (i = 1; i < n; i++) + addtok (dfa, dfa->utf8_anychar_classes[i]); + while (--i > 1) + { + addtok (dfa, dfa->utf8_anychar_classes[0]); + addtok (dfa, CAT); + addtok (dfa, OR); + } +} + +/* The grammar understood by the parser is as follows. + + regexp: + regexp OR branch + branch + + branch: + branch closure + closure + + closure: + closure QMARK + closure STAR + closure PLUS + closure REPMN + atom + + atom: + <normal character> + <multibyte character> + ANYCHAR + MBCSET + CSET + BACKREF + BEGLINE + ENDLINE + BEGWORD + ENDWORD + LIMWORD + NOTLIMWORD + LPAREN regexp RPAREN + <empty> + + The parser builds a parse tree in postfix form in an array of tokens. */ + +static void +atom (struct dfa *dfa) +{ + if ((0 <= dfa->parse.tok && dfa->parse.tok < NOTCHAR) + || dfa->parse.tok >= CSET + || dfa->parse.tok == BEG || dfa->parse.tok == BACKREF + || dfa->parse.tok == BEGLINE || dfa->parse.tok == ENDLINE + || dfa->parse.tok == BEGWORD || dfa->parse.tok == ENDWORD + || dfa->parse.tok == LIMWORD || dfa->parse.tok == NOTLIMWORD + || dfa->parse.tok == ANYCHAR || dfa->parse.tok == MBCSET) + { + if (dfa->parse.tok == ANYCHAR && dfa->localeinfo.using_utf8) + { + /* For UTF-8 expand the period to a series of CSETs that define a + valid UTF-8 character. This avoids using the slow multibyte + path. I'm pretty sure it would be both profitable and correct to + do it for any encoding; however, the optimization must be done + manually as it is done above in add_utf8_anychar. So, let's + start with UTF-8: it is the most used, and the structure of the + encoding makes the correctness more obvious. */ + add_utf8_anychar (dfa); + } + else + addtok (dfa, dfa->parse.tok); + dfa->parse.tok = lex (dfa); + } + else if (dfa->parse.tok == WCHAR) + { + if (dfa->lex.wctok == WEOF) + addtok (dfa, BACKREF); + else + { + addtok_wc (dfa, dfa->lex.wctok); + + if (dfa->syntax.case_fold) + { + wchar_t folded[CASE_FOLDED_BUFSIZE]; + unsigned int n = case_folded_counterparts (dfa->lex.wctok, + folded); + for (unsigned int i = 0; i < n; i++) + { + addtok_wc (dfa, folded[i]); + addtok (dfa, OR); + } + } + } + + dfa->parse.tok = lex (dfa); + } + else if (dfa->parse.tok == LPAREN) + { + dfa->parse.tok = lex (dfa); + regexp (dfa); + if (dfa->parse.tok != RPAREN) + dfaerror (_("unbalanced (")); + dfa->parse.tok = lex (dfa); + } + else + addtok (dfa, EMPTY); +} + +/* Return the number of tokens in the given subexpression. */ +static size_t _GL_ATTRIBUTE_PURE +nsubtoks (struct dfa const *dfa, size_t tindex) +{ + switch (dfa->tokens[tindex - 1]) + { + default: + return 1; + case QMARK: + case STAR: + case PLUS: + return 1 + nsubtoks (dfa, tindex - 1); + case CAT: + case OR: + { + size_t ntoks1 = nsubtoks (dfa, tindex - 1); + return 1 + ntoks1 + nsubtoks (dfa, tindex - 1 - ntoks1); + } + } +} + +/* Copy the given subexpression to the top of the tree. */ +static void +copytoks (struct dfa *dfa, size_t tindex, size_t ntokens) +{ + if (dfa->localeinfo.multibyte) + for (size_t i = 0; i < ntokens; ++i) + addtok_mb (dfa, dfa->tokens[tindex + i], + dfa->multibyte_prop[tindex + i]); + else + for (size_t i = 0; i < ntokens; ++i) + addtok_mb (dfa, dfa->tokens[tindex + i], 3); +} + +static void +closure (struct dfa *dfa) +{ + atom (dfa); + while (dfa->parse.tok == QMARK || dfa->parse.tok == STAR + || dfa->parse.tok == PLUS || dfa->parse.tok == REPMN) + if (dfa->parse.tok == REPMN && (dfa->lex.minrep || dfa->lex.maxrep)) + { + size_t ntokens = nsubtoks (dfa, dfa->tindex); + size_t tindex = dfa->tindex - ntokens; + if (dfa->lex.maxrep < 0) + addtok (dfa, PLUS); + if (dfa->lex.minrep == 0) + addtok (dfa, QMARK); + int i; + for (i = 1; i < dfa->lex.minrep; i++) + { + copytoks (dfa, tindex, ntokens); + addtok (dfa, CAT); + } + for (; i < dfa->lex.maxrep; i++) + { + copytoks (dfa, tindex, ntokens); + addtok (dfa, QMARK); + addtok (dfa, CAT); + } + dfa->parse.tok = lex (dfa); + } + else if (dfa->parse.tok == REPMN) + { + dfa->tindex -= nsubtoks (dfa, dfa->tindex); + dfa->parse.tok = lex (dfa); + closure (dfa); + } + else + { + addtok (dfa, dfa->parse.tok); + dfa->parse.tok = lex (dfa); + } +} + +static void +branch (struct dfa* dfa) +{ + closure (dfa); + while (dfa->parse.tok != RPAREN && dfa->parse.tok != OR + && dfa->parse.tok >= 0) + { + closure (dfa); + addtok (dfa, CAT); + } +} + +static void +regexp (struct dfa *dfa) +{ + branch (dfa); + while (dfa->parse.tok == OR) + { + dfa->parse.tok = lex (dfa); + branch (dfa); + addtok (dfa, OR); + } +} + +/* Main entry point for the parser. S is a string to be parsed, len is the + length of the string, so s can include NUL characters. D is a pointer to + the struct dfa to parse into. */ +static void +dfaparse (char const *s, size_t len, struct dfa *d) +{ + d->lex.ptr = s; + d->lex.left = len; + d->lex.lasttok = END; + d->lex.laststart = true; + + if (!d->syntax.syntax_bits_set) + dfaerror (_("no syntax specified")); + + if (!d->nregexps) + addtok (d, BEG); + + d->parse.tok = lex (d); + d->parse.depth = d->depth; + + regexp (d); + + if (d->parse.tok != END) + dfaerror (_("unbalanced )")); + + addtok (d, END - d->nregexps); + addtok (d, CAT); + + if (d->nregexps) + addtok (d, OR); + + ++d->nregexps; +} + +/* Some primitives for operating on sets of positions. */ + +/* Copy one set to another. */ +static void +copy (position_set const *src, position_set *dst) +{ + if (dst->alloc < src->nelem) + { + free (dst->elems); + dst->elems = xpalloc (NULL, &dst->alloc, src->nelem - dst->alloc, -1, + sizeof *dst->elems); + } + dst->nelem = src->nelem; + if (src->nelem != 0) + memcpy (dst->elems, src->elems, src->nelem * sizeof *dst->elems); +} + +static void +alloc_position_set (position_set *s, size_t size) +{ + s->elems = xnmalloc (size, sizeof *s->elems); + s->alloc = size; + s->nelem = 0; +} + +/* Insert position P in set S. S is maintained in sorted order on + decreasing index. If there is already an entry in S with P.index + then merge (logically-OR) P's constraints into the one in S. + S->elems must point to an array large enough to hold the resulting set. */ +static void +insert (position p, position_set *s) +{ + ptrdiff_t count = s->nelem; + ptrdiff_t lo = 0, hi = count; + while (lo < hi) + { + ptrdiff_t mid = (lo + hi) >> 1; + if (s->elems[mid].index < p.index) + lo = mid + 1; + else if (s->elems[mid].index == p.index) + { + s->elems[mid].constraint |= p.constraint; + return; + } + else + hi = mid; + } + + s->elems = maybe_realloc (s->elems, count, &s->alloc, -1, sizeof *s->elems); + for (ptrdiff_t i = count; i > lo; i--) + s->elems[i] = s->elems[i - 1]; + s->elems[lo] = p; + ++s->nelem; +} + +static void +append (position p, position_set *s) +{ + ptrdiff_t count = s->nelem; + s->elems = maybe_realloc (s->elems, count, &s->alloc, -1, sizeof *s->elems); + s->elems[s->nelem++] = p; +} + +/* Merge S1 and S2 (with the additional constraint C2) into M. The + result is as if the positions of S1, and of S2 with the additional + constraint C2, were inserted into an initially empty set. */ +static void +merge_constrained (position_set const *s1, position_set const *s2, + unsigned int c2, position_set *m) +{ + ptrdiff_t i = 0, j = 0; + + if (m->alloc - s1->nelem < s2->nelem) + { + free (m->elems); + m->alloc = s1->nelem; + m->elems = xpalloc (NULL, &m->alloc, s2->nelem, -1, sizeof *m->elems); + } + m->nelem = 0; + while (i < s1->nelem || j < s2->nelem) + if (! (j < s2->nelem) + || (i < s1->nelem && s1->elems[i].index <= s2->elems[j].index)) + { + unsigned int c = ((i < s1->nelem && j < s2->nelem + && s1->elems[i].index == s2->elems[j].index) + ? s2->elems[j++].constraint & c2 + : 0); + m->elems[m->nelem].index = s1->elems[i].index; + m->elems[m->nelem++].constraint = s1->elems[i++].constraint | c; + } + else + { + if (s2->elems[j].constraint & c2) + { + m->elems[m->nelem].index = s2->elems[j].index; + m->elems[m->nelem++].constraint = s2->elems[j].constraint & c2; + } + j++; + } +} + +/* Merge two sets of positions into a third. The result is exactly as if + the positions of both sets were inserted into an initially empty set. */ +static void +merge (position_set const *s1, position_set const *s2, position_set *m) +{ + merge_constrained (s1, s2, -1, m); +} + +static void +merge2 (position_set *dst, position_set const *src, position_set *m) +{ + if (src->nelem < 4) + { + for (ptrdiff_t i = 0; i < src->nelem; ++i) + insert (src->elems[i], dst); + } + else + { + merge (src, dst, m); + copy (m, dst); + } +} + +/* Delete a position from a set. Return the nonzero constraint of the + deleted position, or zero if there was no such position. */ +static unsigned int +delete (size_t del, position_set *s) +{ + size_t count = s->nelem; + size_t lo = 0, hi = count; + while (lo < hi) + { + size_t mid = (lo + hi) >> 1; + if (s->elems[mid].index < del) + lo = mid + 1; + else if (s->elems[mid].index == del) + { + unsigned int c = s->elems[mid].constraint; + size_t i; + for (i = mid; i + 1 < count; i++) + s->elems[i] = s->elems[i + 1]; + s->nelem = i; + return c; + } + else + hi = mid; + } + return 0; +} + +/* Replace a position with the followed set. */ +static void +replace (position_set *dst, size_t del, position_set *add, + unsigned int constraint, position_set *tmp) +{ + unsigned int c = delete (del, dst) & constraint; + + if (c) + { + copy (dst, tmp); + merge_constrained (tmp, add, c, dst); + } +} + +/* Find the index of the state corresponding to the given position set with + the given preceding context, or create a new state if there is no such + state. Context tells whether we got here on a newline or letter. */ +static state_num +state_index (struct dfa *d, position_set const *s, int context) +{ + size_t hash = 0; + int constraint = 0; + state_num i; + token first_end = 0; + + for (i = 0; i < s->nelem; ++i) + hash ^= s->elems[i].index + s->elems[i].constraint; + + /* Try to find a state that exactly matches the proposed one. */ + for (i = 0; i < d->sindex; ++i) + { + if (hash != d->states[i].hash || s->nelem != d->states[i].elems.nelem + || context != d->states[i].context) + continue; + state_num j; + for (j = 0; j < s->nelem; ++j) + if (s->elems[j].constraint != d->states[i].elems.elems[j].constraint + || s->elems[j].index != d->states[i].elems.elems[j].index) + break; + if (j == s->nelem) + return i; + } + +#ifdef DEBUG + fprintf (stderr, "new state %zd\n nextpos:", i); + for (state_num j = 0; j < s->nelem; j++) + { + fprintf (stderr, " %zu:", s->elems[j].index); + prtok (d->tokens[s->elems[j].index]); + } + fprintf (stderr, "\n context:"); + if (context ^ CTX_ANY) + { + if (context & CTX_NONE) + fprintf (stderr, " CTX_NONE"); + if (context & CTX_LETTER) + fprintf (stderr, " CTX_LETTER"); + if (context & CTX_NEWLINE) + fprintf (stderr, " CTX_NEWLINE"); + } + else + fprintf (stderr, " CTX_ANY"); + fprintf (stderr, "\n"); +#endif + + for (state_num j = 0; j < s->nelem; j++) + { + int c = d->constraints[s->elems[j].index]; + + if (c != 0) + { + if (succeeds_in_context (c, context, CTX_ANY)) + constraint |= c; + if (!first_end) + first_end = d->tokens[s->elems[j].index]; + } + else if (d->tokens[s->elems[j].index] == BACKREF) + constraint = NO_CONSTRAINT; + } + + + /* Create a new state. */ + d->states = maybe_realloc (d->states, d->sindex, &d->salloc, -1, + sizeof *d->states); + d->states[i].hash = hash; + alloc_position_set (&d->states[i].elems, s->nelem); + copy (s, &d->states[i].elems); + d->states[i].context = context; + d->states[i].constraint = constraint; + d->states[i].first_end = first_end; + d->states[i].mbps.nelem = 0; + d->states[i].mbps.elems = NULL; + d->states[i].mb_trindex = -1; + + ++d->sindex; + + return i; +} + +/* Find the epsilon closure of a set of positions. If any position of the set + contains a symbol that matches the empty string in some context, replace + that position with the elements of its follow labeled with an appropriate + constraint. Repeat exhaustively until no funny positions are left. + S->elems must be large enough to hold the result. */ +static void +epsclosure (struct dfa const *d) +{ + position_set tmp; + alloc_position_set (&tmp, d->nleaves); + for (size_t i = 0; i < d->tindex; ++i) + if (d->follows[i].nelem > 0 && d->tokens[i] >= NOTCHAR + && d->tokens[i] != BACKREF && d->tokens[i] != ANYCHAR + && d->tokens[i] != MBCSET && d->tokens[i] < CSET) + { + unsigned int constraint; + switch (d->tokens[i]) + { + case BEGLINE: + constraint = BEGLINE_CONSTRAINT; + break; + case ENDLINE: + constraint = ENDLINE_CONSTRAINT; + break; + case BEGWORD: + constraint = BEGWORD_CONSTRAINT; + break; + case ENDWORD: + constraint = ENDWORD_CONSTRAINT; + break; + case LIMWORD: + constraint = LIMWORD_CONSTRAINT; + break; + case NOTLIMWORD: + constraint = NOTLIMWORD_CONSTRAINT; + break; + default: + constraint = NO_CONSTRAINT; + break; + } + + delete (i, &d->follows[i]); + + for (size_t j = 0; j < d->tindex; j++) + if (i != j && d->follows[j].nelem > 0) + replace (&d->follows[j], i, &d->follows[i], constraint, &tmp); + } + free (tmp.elems); +} + +/* Returns the set of contexts for which there is at least one + character included in C. */ + +static int +charclass_context (struct dfa const *dfa, charclass const *c) +{ + int context = 0; + + for (unsigned int j = 0; j < CHARCLASS_WORDS; ++j) + { + if (c->w[j] & dfa->syntax.newline.w[j]) + context |= CTX_NEWLINE; + if (c->w[j] & dfa->syntax.letters.w[j]) + context |= CTX_LETTER; + if (c->w[j] & ~(dfa->syntax.letters.w[j] | dfa->syntax.newline.w[j])) + context |= CTX_NONE; + } + + return context; +} + +/* Returns the contexts on which the position set S depends. Each context + in the set of returned contexts (let's call it SC) may have a different + follow set than other contexts in SC, and also different from the + follow set of the complement set (sc ^ CTX_ANY). However, all contexts + in the complement set will have the same follow set. */ + +static int _GL_ATTRIBUTE_PURE +state_separate_contexts (struct dfa *d, position_set const *s) +{ + int separate_contexts = 0; + + for (size_t j = 0; j < s->nelem; j++) + separate_contexts |= d->separates[s->elems[j].index]; + + return separate_contexts; +} + +enum +{ + /* Single token is repeated. It is distinguished from non-repeated. */ + OPT_REPEAT = (1 << 0), + + /* Multiple tokens are repeated. This flag is on at head of tokens. The + node is not merged. */ + OPT_LPAREN = (1 << 1), + + /* Multiple branches are joined. The node is not merged. */ + OPT_RPAREN = (1 << 2), + + /* The node is walked. If the node is found in walking again, OPT_RPAREN + flag is turned on. */ + OPT_WALKED = (1 << 3), + + /* The node is queued. The node is not queued again. */ + OPT_QUEUED = (1 << 4) +}; + +static void +merge_nfa_state (struct dfa *d, size_t tindex, char *flags, + position_set *merged) +{ + position_set *follows = d->follows; + ptrdiff_t nelem = 0; + + d->constraints[tindex] = 0; + + for (ptrdiff_t i = 0; i < follows[tindex].nelem; i++) + { + size_t sindex = follows[tindex].elems[i].index; + + /* Skip the node as pruned in future. */ + unsigned int iconstraint = follows[tindex].elems[i].constraint; + if (iconstraint == 0) + continue; + + if (d->tokens[follows[tindex].elems[i].index] <= END) + { + d->constraints[tindex] |= follows[tindex].elems[i].constraint; + continue; + } + + if (!(flags[sindex] & (OPT_LPAREN | OPT_RPAREN))) + { + ptrdiff_t j; + + for (j = 0; j < nelem; j++) + { + size_t dindex = follows[tindex].elems[j].index; + + if (follows[tindex].elems[j].constraint != iconstraint) + continue; + + if (flags[dindex] & (OPT_LPAREN | OPT_RPAREN)) + continue; + + if (d->tokens[sindex] != d->tokens[dindex]) + continue; + + if ((flags[sindex] ^ flags[dindex]) & OPT_REPEAT) + continue; + + if (flags[sindex] & OPT_REPEAT) + delete (sindex, &follows[sindex]); + + merge2 (&follows[dindex], &follows[sindex], merged); + + break; + } + + if (j < nelem) + continue; + } + + follows[tindex].elems[nelem++] = follows[tindex].elems[i]; + flags[sindex] |= OPT_QUEUED; + } + + follows[tindex].nelem = nelem; +} + +static int +compare (const void *a, const void *b) +{ + int aindex; + int bindex; + + aindex = (int) ((position *) a)->index; + bindex = (int) ((position *) b)->index; + + return aindex - bindex; +} + +static void +reorder_tokens (struct dfa *d) +{ + ptrdiff_t nleaves; + ptrdiff_t *map; + token *tokens; + position_set *follows; + int *constraints; + char *multibyte_prop; + + nleaves = 0; + + map = xnmalloc (d->tindex, sizeof *map); + + map[0] = nleaves++; + + for (ptrdiff_t i = 1; i < d->tindex; i++) + map[i] = -1; + + tokens = xnmalloc (d->nleaves, sizeof *tokens); + follows = xnmalloc (d->nleaves, sizeof *follows); + constraints = xnmalloc (d->nleaves, sizeof *constraints); + + if (d->localeinfo.multibyte) + multibyte_prop = xnmalloc (d->nleaves, sizeof *multibyte_prop); + else + multibyte_prop = NULL; + + for (ptrdiff_t i = 0; i < d->tindex; i++) + { + if (map[i] == -1) + { + free (d->follows[i].elems); + d->follows[i].elems = NULL; + d->follows[i].nelem = 0; + continue; + } + + tokens[map[i]] = d->tokens[i]; + follows[map[i]] = d->follows[i]; + constraints[map[i]] = d->constraints[i]; + + if (multibyte_prop != NULL) + multibyte_prop[map[i]] = d->multibyte_prop[i]; + + for (ptrdiff_t j = 0; j < d->follows[i].nelem; j++) + { + if (map[d->follows[i].elems[j].index] == -1) + map[d->follows[i].elems[j].index] = nleaves++; + + d->follows[i].elems[j].index = map[d->follows[i].elems[j].index]; + } + + qsort (d->follows[i].elems, d->follows[i].nelem, + sizeof *d->follows[i].elems, compare); + } + + for (ptrdiff_t i = 0; i < nleaves; i++) + { + d->tokens[i] = tokens[i]; + d->follows[i] = follows[i]; + d->constraints[i] = constraints[i]; + + if (multibyte_prop != NULL) + d->multibyte_prop[i] = multibyte_prop[i]; + } + + d->tindex = d->nleaves = nleaves; + + free (tokens); + free (follows); + free (constraints); + free (multibyte_prop); + free (map); +} + +static void +dfaoptimize (struct dfa *d) +{ + char *flags; + position_set merged0; + position_set *merged; + + flags = xmalloc (d->tindex * sizeof *flags); + memset (flags, 0, d->tindex * sizeof *flags); + + for (size_t i = 0; i < d->tindex; i++) + { + for (ptrdiff_t j = 0; j < d->follows[i].nelem; j++) + { + if (d->follows[i].elems[j].index == i) + flags[d->follows[i].elems[j].index] |= OPT_REPEAT; + else if (d->follows[i].elems[j].index < i) + flags[d->follows[i].elems[j].index] |= OPT_LPAREN; + else if (flags[d->follows[i].elems[j].index] &= OPT_WALKED) + flags[d->follows[i].elems[j].index] |= OPT_RPAREN; + else + flags[d->follows[i].elems[j].index] |= OPT_WALKED; + } + } + + flags[0] |= OPT_QUEUED; + + merged = &merged0; + alloc_position_set (merged, d->nleaves); + + d->constraints = xnmalloc (d->tindex, sizeof *d->constraints); + + for (ptrdiff_t i = 0; i < d->tindex; i++) + if (flags[i] & OPT_QUEUED) + merge_nfa_state (d, i, flags, merged); + + reorder_tokens (d); + + free (merged->elems); + free (flags); +} + +/* Perform bottom-up analysis on the parse tree, computing various functions. + Note that at this point, we're pretending constructs like \< are real + characters rather than constraints on what can follow them. + + Nullable: A node is nullable if it is at the root of a regexp that can + match the empty string. + * EMPTY leaves are nullable. + * No other leaf is nullable. + * A QMARK or STAR node is nullable. + * A PLUS node is nullable if its argument is nullable. + * A CAT node is nullable if both its arguments are nullable. + * An OR node is nullable if either argument is nullable. + + Firstpos: The firstpos of a node is the set of positions (nonempty leaves) + that could correspond to the first character of a string matching the + regexp rooted at the given node. + * EMPTY leaves have empty firstpos. + * The firstpos of a nonempty leaf is that leaf itself. + * The firstpos of a QMARK, STAR, or PLUS node is the firstpos of its + argument. + * The firstpos of a CAT node is the firstpos of the left argument, union + the firstpos of the right if the left argument is nullable. + * The firstpos of an OR node is the union of firstpos of each argument. + + Lastpos: The lastpos of a node is the set of positions that could + correspond to the last character of a string matching the regexp at + the given node. + * EMPTY leaves have empty lastpos. + * The lastpos of a nonempty leaf is that leaf itself. + * The lastpos of a QMARK, STAR, or PLUS node is the lastpos of its + argument. + * The lastpos of a CAT node is the lastpos of its right argument, union + the lastpos of the left if the right argument is nullable. + * The lastpos of an OR node is the union of the lastpos of each argument. + + Follow: The follow of a position is the set of positions that could + correspond to the character following a character matching the node in + a string matching the regexp. At this point we consider special symbols + that match the empty string in some context to be just normal characters. + Later, if we find that a special symbol is in a follow set, we will + replace it with the elements of its follow, labeled with an appropriate + constraint. + * Every node in the firstpos of the argument of a STAR or PLUS node is in + the follow of every node in the lastpos. + * Every node in the firstpos of the second argument of a CAT node is in + the follow of every node in the lastpos of the first argument. + + Because of the postfix representation of the parse tree, the depth-first + analysis is conveniently done by a linear scan with the aid of a stack. + Sets are stored as arrays of the elements, obeying a stack-like allocation + scheme; the number of elements in each set deeper in the stack can be + used to determine the address of a particular set's array. */ +static void +dfaanalyze (struct dfa *d, bool searchflag) +{ + /* Array allocated to hold position sets. */ + position *posalloc = xnmalloc (d->nleaves, 2 * sizeof *posalloc); + /* Firstpos and lastpos elements. */ + position *firstpos = posalloc; + position *lastpos = firstpos + d->nleaves; + position pos; + position_set tmp; + + /* Stack for element counts and nullable flags. */ + struct + { + /* Whether the entry is nullable. */ + bool nullable; + + /* Counts of firstpos and lastpos sets. */ + size_t nfirstpos; + size_t nlastpos; + } *stkalloc = xnmalloc (d->depth, sizeof *stkalloc), *stk = stkalloc; + + position_set merged; /* Result of merging sets. */ + + addtok (d, CAT); + +#ifdef DEBUG + fprintf (stderr, "dfaanalyze:\n"); + for (size_t i = 0; i < d->tindex; ++i) + { + fprintf (stderr, " %zu:", i); + prtok (d->tokens[i]); + } + putc ('\n', stderr); +#endif + + d->searchflag = searchflag; + alloc_position_set (&merged, d->nleaves); + d->follows = xcalloc (d->tindex, sizeof *d->follows); + + for (size_t i = 0; i < d->tindex; ++i) + { + switch (d->tokens[i]) + { + case EMPTY: + /* The empty set is nullable. */ + stk->nullable = true; + + /* The firstpos and lastpos of the empty leaf are both empty. */ + stk->nfirstpos = stk->nlastpos = 0; + stk++; + break; + + case STAR: + case PLUS: + /* Every element in the firstpos of the argument is in the follow + of every element in the lastpos. */ + { + tmp.elems = firstpos - stk[-1].nfirstpos; + tmp.nelem = stk[-1].nfirstpos; + position *p = lastpos - stk[-1].nlastpos; + for (size_t j = 0; j < stk[-1].nlastpos; j++) + { + merge (&tmp, &d->follows[p[j].index], &merged); + copy (&merged, &d->follows[p[j].index]); + } + } + FALLTHROUGH; + case QMARK: + /* A QMARK or STAR node is automatically nullable. */ + if (d->tokens[i] != PLUS) + stk[-1].nullable = true; + break; + + case CAT: + /* Every element in the firstpos of the second argument is in the + follow of every element in the lastpos of the first argument. */ + { + tmp.nelem = stk[-1].nfirstpos; + tmp.elems = firstpos - stk[-1].nfirstpos; + position *p = lastpos - stk[-1].nlastpos - stk[-2].nlastpos; + for (size_t j = 0; j < stk[-2].nlastpos; j++) + { + merge (&tmp, &d->follows[p[j].index], &merged); + copy (&merged, &d->follows[p[j].index]); + } + } + + /* The firstpos of a CAT node is the firstpos of the first argument, + union that of the second argument if the first is nullable. */ + if (stk[-2].nullable) + stk[-2].nfirstpos += stk[-1].nfirstpos; + else + firstpos -= stk[-1].nfirstpos; + + /* The lastpos of a CAT node is the lastpos of the second argument, + union that of the first argument if the second is nullable. */ + if (stk[-1].nullable) + stk[-2].nlastpos += stk[-1].nlastpos; + else + { + position *p = lastpos - stk[-1].nlastpos - stk[-2].nlastpos; + for (size_t j = 0; j < stk[-1].nlastpos; j++) + p[j] = p[j + stk[-2].nlastpos]; + lastpos -= stk[-2].nlastpos; + stk[-2].nlastpos = stk[-1].nlastpos; + } + + /* A CAT node is nullable if both arguments are nullable. */ + stk[-2].nullable &= stk[-1].nullable; + stk--; + break; + + case OR: + /* The firstpos is the union of the firstpos of each argument. */ + stk[-2].nfirstpos += stk[-1].nfirstpos; + + /* The lastpos is the union of the lastpos of each argument. */ + stk[-2].nlastpos += stk[-1].nlastpos; + + /* An OR node is nullable if either argument is nullable. */ + stk[-2].nullable |= stk[-1].nullable; + stk--; + break; + + default: + /* Anything else is a nonempty position. (Note that special + constructs like \< are treated as nonempty strings here; + an "epsilon closure" effectively makes them nullable later. + Backreferences have to get a real position so we can detect + transitions on them later. But they are nullable. */ + stk->nullable = d->tokens[i] == BACKREF; + + /* This position is in its own firstpos and lastpos. */ + stk->nfirstpos = stk->nlastpos = 1; + stk++; + + firstpos->index = lastpos->index = i; + firstpos->constraint = lastpos->constraint = NO_CONSTRAINT; + firstpos++, lastpos++; + + break; + } +#ifdef DEBUG + /* ... balance the above nonsyntactic #ifdef goo... */ + fprintf (stderr, "node %zu:", i); + prtok (d->tokens[i]); + putc ('\n', stderr); + fprintf (stderr, + stk[-1].nullable ? " nullable: yes\n" : " nullable: no\n"); + fprintf (stderr, " firstpos:"); + for (size_t j = 0; j < stk[-1].nfirstpos; j++) + { + fprintf (stderr, " %zu:", firstpos[j - stk[-1].nfirstpos].index); + prtok (d->tokens[firstpos[j - stk[-1].nfirstpos].index]); + } + fprintf (stderr, "\n lastpos:"); + for (size_t j = 0; j < stk[-1].nlastpos; j++) + { + fprintf (stderr, " %zu:", lastpos[j - stk[-1].nlastpos].index); + prtok (d->tokens[lastpos[j - stk[-1].nlastpos].index]); + } + putc ('\n', stderr); +#endif + } + + /* For each follow set that is the follow set of a real position, replace + it with its epsilon closure. */ + epsclosure (d); + + dfaoptimize (d); + +#ifdef DEBUG + for (size_t i = 0; i < d->tindex; ++i) + if (d->tokens[i] == BEG || d->tokens[i] < NOTCHAR + || d->tokens[i] == BACKREF || d->tokens[i] == ANYCHAR + || d->tokens[i] == MBCSET || d->tokens[i] >= CSET) + { + fprintf (stderr, "follows(%zu:", i); + prtok (d->tokens[i]); + fprintf (stderr, "):"); + for (size_t j = 0; j < d->follows[i].nelem; j++) + { + fprintf (stderr, " %zu:", d->follows[i].elems[j].index); + prtok (d->tokens[d->follows[i].elems[j].index]); + } + putc ('\n', stderr); + } +#endif + + pos.index = 0; + pos.constraint = NO_CONSTRAINT; + + alloc_position_set (&tmp, 1); + + append (pos, &tmp); + + d->separates = xnmalloc (d->tindex, sizeof *d->separates); + + for (ptrdiff_t i = 0; i < d->tindex; i++) + { + d->separates[i] = 0; + + if (prev_newline_dependent (d->constraints[i])) + d->separates[i] |= CTX_NEWLINE; + if (prev_letter_dependent (d->constraints[i])) + d->separates[i] |= CTX_LETTER; + + for (ptrdiff_t j = 0; j < d->follows[i].nelem; j++) + { + if (prev_newline_dependent (d->follows[i].elems[j].constraint)) + d->separates[i] |= CTX_NEWLINE; + if (prev_letter_dependent (d->follows[i].elems[j].constraint)) + d->separates[i] |= CTX_LETTER; + } + } + + /* Context wanted by some position. */ + int separate_contexts = state_separate_contexts (d, &tmp); + + /* Build the initial state. */ + if (separate_contexts & CTX_NEWLINE) + state_index (d, &tmp, CTX_NEWLINE); + d->initstate_notbol = d->min_trcount + = state_index (d, &tmp, separate_contexts ^ CTX_ANY); + if (separate_contexts & CTX_LETTER) + d->min_trcount = state_index (d, &tmp, CTX_LETTER); + d->min_trcount++; + d->trcount = 0; + + free (posalloc); + free (stkalloc); + free (merged.elems); + free (tmp.elems); +} + +/* Make sure D's state arrays are large enough to hold NEW_STATE. */ +static void +realloc_trans_if_necessary (struct dfa *d) +{ + state_num oldalloc = d->tralloc; + if (oldalloc < d->sindex) + { + state_num **realtrans = d->trans ? d->trans - 2 : NULL; + ptrdiff_t newalloc1 = realtrans ? d->tralloc + 2 : 0; + realtrans = xpalloc (realtrans, &newalloc1, d->sindex - oldalloc, + -1, sizeof *realtrans); + realtrans[0] = realtrans[1] = NULL; + d->trans = realtrans + 2; + ptrdiff_t newalloc = d->tralloc = newalloc1 - 2; + d->fails = xnrealloc (d->fails, newalloc, sizeof *d->fails); + d->success = xnrealloc (d->success, newalloc, sizeof *d->success); + d->newlines = xnrealloc (d->newlines, newalloc, sizeof *d->newlines); + if (d->localeinfo.multibyte) + { + realtrans = d->mb_trans ? d->mb_trans - 2 : NULL; + realtrans = xnrealloc (realtrans, newalloc1, sizeof *realtrans); + if (oldalloc == 0) + realtrans[0] = realtrans[1] = NULL; + d->mb_trans = realtrans + 2; + } + for (; oldalloc < newalloc; oldalloc++) + { + d->trans[oldalloc] = NULL; + d->fails[oldalloc] = NULL; + if (d->localeinfo.multibyte) + d->mb_trans[oldalloc] = NULL; + } + } +} + +/* + Calculate the transition table for a new state derived from state s + for a compiled dfa d after input character uc, and return the new + state number. + + Do not worry about all possible input characters; calculate just the group + of positions that match uc. Label it with the set of characters that + every position in the group matches (taking into account, if necessary, + preceding context information of s). Then find the union + of these positions' follows, i.e., the set of positions of the + new state. For each character in the group's label, set the transition + on this character to be to a state corresponding to the set's positions, + and its associated backward context information, if necessary. + + When building a searching matcher, include the positions of state + 0 in every state. + + The group is constructed by building an equivalence-class + partition of the positions of s. + + For each position, find the set of characters C that it matches. Eliminate + any characters from C that fail on grounds of backward context. + + Check whether the group's label L has nonempty + intersection with C. If L - C is nonempty, create a new group labeled + L - C and having the same positions as the current group, and set L to + the intersection of L and C. Insert the position in the group, set + C = C - L, and resume scanning. + + If after comparing with every group there are characters remaining in C, + create a new group labeled with the characters of C and insert this + position in that group. */ + +static state_num +build_state (state_num s, struct dfa *d, unsigned char uc) +{ + position_set follows; /* Union of the follows for each + position of the current state. */ + position_set group; /* Positions that match the input char. */ + position_set tmp; /* Temporary space for merging sets. */ + state_num state; /* New state. */ + state_num state_newline; /* New state on a newline transition. */ + state_num state_letter; /* New state on a letter transition. */ + +#ifdef DEBUG + fprintf (stderr, "build state %td\n", s); +#endif + + /* A pointer to the new transition table, and the table itself. */ + state_num **ptrans = (accepting (s, d) ? d->fails : d->trans) + s; + state_num *trans = *ptrans; + + if (!trans) + { + /* MAX_TRCOUNT is an arbitrary upper limit on the number of + transition tables that can exist at once, other than for + initial states. Often-used transition tables are quickly + rebuilt, whereas rarely-used ones are cleared away. */ + if (MAX_TRCOUNT <= d->trcount) + { + for (state_num i = d->min_trcount; i < d->tralloc; i++) + { + free (d->trans[i]); + free (d->fails[i]); + d->trans[i] = d->fails[i] = NULL; + } + d->trcount = 0; + } + + d->trcount++; + *ptrans = trans = xmalloc (NOTCHAR * sizeof *trans); + + /* Fill transition table with a default value which means that the + transited state has not been calculated yet. */ + for (int i = 0; i < NOTCHAR; i++) + trans[i] = -2; + } + + /* Set up the success bits for this state. */ + d->success[s] = 0; + if (accepts_in_context (d->states[s].context, CTX_NEWLINE, s, d)) + d->success[s] |= CTX_NEWLINE; + if (accepts_in_context (d->states[s].context, CTX_LETTER, s, d)) + d->success[s] |= CTX_LETTER; + if (accepts_in_context (d->states[s].context, CTX_NONE, s, d)) + d->success[s] |= CTX_NONE; + + alloc_position_set (&follows, d->nleaves); + + /* Find the union of the follows of the positions of the group. + This is a hideously inefficient loop. Fix it someday. */ + for (size_t j = 0; j < d->states[s].elems.nelem; ++j) + for (size_t k = 0; + k < d->follows[d->states[s].elems.elems[j].index].nelem; ++k) + insert (d->follows[d->states[s].elems.elems[j].index].elems[k], + &follows); + + /* Positions that match the input char. */ + alloc_position_set (&group, d->nleaves); + + /* The group's label. */ + charclass label; + fillset (&label); + + for (size_t i = 0; i < follows.nelem; ++i) + { + charclass matches; /* Set of matching characters. */ + position pos = follows.elems[i]; + bool matched = false; + if (d->tokens[pos.index] >= 0 && d->tokens[pos.index] < NOTCHAR) + { + zeroset (&matches); + setbit (d->tokens[pos.index], &matches); + if (d->tokens[pos.index] == uc) + matched = true; + } + else if (d->tokens[pos.index] >= CSET) + { + matches = d->charclasses[d->tokens[pos.index] - CSET]; + if (tstbit (uc, &matches)) + matched = true; + } + else if (d->tokens[pos.index] == ANYCHAR) + { + matches = d->charclasses[d->canychar]; + if (tstbit (uc, &matches)) + matched = true; + + /* ANYCHAR must match with a single character, so we must put + it to D->states[s].mbps which contains the positions which + can match with a single character not a byte. If all + positions which has ANYCHAR does not depend on context of + next character, we put the follows instead of it to + D->states[s].mbps to optimize. */ + if (succeeds_in_context (pos.constraint, d->states[s].context, + CTX_NONE)) + { + if (d->states[s].mbps.nelem == 0) + alloc_position_set (&d->states[s].mbps, 1); + insert (pos, &d->states[s].mbps); + } + } + else + continue; + + /* Some characters may need to be eliminated from matches because + they fail in the current context. */ + if (pos.constraint != NO_CONSTRAINT) + { + if (!succeeds_in_context (pos.constraint, + d->states[s].context, CTX_NEWLINE)) + for (size_t j = 0; j < CHARCLASS_WORDS; ++j) + matches.w[j] &= ~d->syntax.newline.w[j]; + if (!succeeds_in_context (pos.constraint, + d->states[s].context, CTX_LETTER)) + for (size_t j = 0; j < CHARCLASS_WORDS; ++j) + matches.w[j] &= ~d->syntax.letters.w[j]; + if (!succeeds_in_context (pos.constraint, + d->states[s].context, CTX_NONE)) + for (size_t j = 0; j < CHARCLASS_WORDS; ++j) + matches.w[j] &= d->syntax.letters.w[j] | d->syntax.newline.w[j]; + + /* If there are no characters left, there's no point in going on. */ + if (emptyset (&matches)) + continue; + + /* If we have reset the bit that made us declare "matched", reset + that indicator, too. This is required to avoid an infinite loop + with this command: echo cx | LC_ALL=C grep -E 'c\b[x ]' */ + if (!tstbit (uc, &matches)) + matched = false; + } + +#ifdef DEBUG + fprintf (stderr, " nextpos %zu:", pos.index); + prtok (d->tokens[pos.index]); + fprintf (stderr, " of"); + for (size_t j = 0; j < NOTCHAR; j++) + if (tstbit (j, &matches)) + fprintf (stderr, " 0x%02zx", j); + fprintf (stderr, "\n"); +#endif + + if (matched) + { + for (size_t k = 0; k < CHARCLASS_WORDS; ++k) + label.w[k] &= matches.w[k]; + append (pos, &group); + } + else + { + for (size_t k = 0; k < CHARCLASS_WORDS; ++k) + label.w[k] &= ~matches.w[k]; + } + } + + alloc_position_set (&tmp, d->nleaves); + + if (group.nelem > 0) + { + /* If we are building a searching matcher, throw in the positions + of state 0 as well, if possible. */ + if (d->searchflag) + { + /* If a token in follows.elems is not 1st byte of a multibyte + character, or the states of follows must accept the bytes + which are not 1st byte of the multibyte character. + Then, if a state of follows encounters a byte, it must not be + a 1st byte of a multibyte character nor a single byte character. + In this case, do not add state[0].follows to next state, because + state[0] must accept 1st-byte. + + For example, suppose <sb a> is a certain single byte character, + <mb A> is a certain multibyte character, and the codepoint of + <sb a> equals the 2nd byte of the codepoint of <mb A>. When + state[0] accepts <sb a>, state[i] transits to state[i+1] by + accepting the 1st byte of <mb A>, and state[i+1] accepts the + 2nd byte of <mb A>, if state[i+1] encounters the codepoint of + <sb a>, it must not be <sb a> but the 2nd byte of <mb A>, so do + not add state[0]. */ + + bool mergeit = !d->localeinfo.multibyte; + if (!mergeit) + { + mergeit = true; + for (size_t j = 0; mergeit && j < group.nelem; j++) + mergeit &= d->multibyte_prop[group.elems[j].index]; + } + if (mergeit) + { + merge (&d->states[0].elems, &group, &tmp); + copy (&tmp, &group); + } + } + + /* Find out if the new state will want any context information, + by calculating possible contexts that the group can match, + and separate contexts that the new state wants to know. */ + int possible_contexts = charclass_context (d, &label); + int separate_contexts = state_separate_contexts (d, &group); + + /* Find the state(s) corresponding to the union of the follows. */ + if (possible_contexts & ~separate_contexts) + state = state_index (d, &group, separate_contexts ^ CTX_ANY); + else + state = -1; + if (separate_contexts & possible_contexts & CTX_NEWLINE) + state_newline = state_index (d, &group, CTX_NEWLINE); + else + state_newline = state; + if (separate_contexts & possible_contexts & CTX_LETTER) + state_letter = state_index (d, &group, CTX_LETTER); + else + state_letter = state; + + /* Reallocate now, to reallocate any newline transition properly. */ + realloc_trans_if_necessary (d); + } + + /* If we are a searching matcher, the default transition is to a state + containing the positions of state 0, otherwise the default transition + is to fail miserably. */ + else if (d->searchflag) + { + state_newline = 0; + state_letter = d->min_trcount - 1; + state = d->initstate_notbol; + } + else + { + state_newline = -1; + state_letter = -1; + state = -1; + } + + /* Set the transitions for each character in the label. */ + for (size_t i = 0; i < NOTCHAR; i++) + if (tstbit (i, &label)) + switch (d->syntax.sbit[i]) + { + case CTX_NEWLINE: + trans[i] = state_newline; + break; + case CTX_LETTER: + trans[i] = state_letter; + break; + default: + trans[i] = state; + break; + } + +#ifdef DEBUG + fprintf (stderr, "trans table %td", s); + for (size_t i = 0; i < NOTCHAR; ++i) + { + if (!(i & 0xf)) + fprintf (stderr, "\n"); + fprintf (stderr, " %2td", trans[i]); + } + fprintf (stderr, "\n"); +#endif + + free (group.elems); + free (follows.elems); + free (tmp.elems); + + /* Keep the newline transition in a special place so we can use it as + a sentinel. */ + if (tstbit (d->syntax.eolbyte, &label)) + { + d->newlines[s] = trans[d->syntax.eolbyte]; + trans[d->syntax.eolbyte] = -1; + } + + return trans[uc]; +} + +/* Multibyte character handling sub-routines for dfaexec. */ + +/* Consume a single byte and transit state from 's' to '*next_state'. + This function is almost same as the state transition routin in dfaexec. + But state transition is done just once, otherwise matching succeed or + reach the end of the buffer. */ +static state_num +transit_state_singlebyte (struct dfa *d, state_num s, unsigned char const **pp) +{ + state_num *t; + + if (d->trans[s]) + t = d->trans[s]; + else if (d->fails[s]) + t = d->fails[s]; + else + { + build_state (s, d, **pp); + if (d->trans[s]) + t = d->trans[s]; + else + { + t = d->fails[s]; + assert (t); + } + } + + if (t[**pp] == -2) + build_state (s, d, **pp); + + return t[*(*pp)++]; +} + +/* Transit state from s, then return new state and update the pointer of + the buffer. This function is for a period operator which can match a + multi-byte character. */ +static state_num +transit_state (struct dfa *d, state_num s, unsigned char const **pp, + unsigned char const *end) +{ + wint_t wc; + + int mbclen = mbs_to_wchar (&wc, (char const *) *pp, end - *pp, d); + + /* This state has some operators which can match a multibyte character. */ + d->mb_follows.nelem = 0; + + /* Calculate the state which can be reached from the state 's' by + consuming 'mbclen' single bytes from the buffer. */ + state_num s1 = s; + int mbci; + for (mbci = 0; mbci < mbclen && (mbci == 0 || d->min_trcount <= s); mbci++) + s = transit_state_singlebyte (d, s, pp); + *pp += mbclen - mbci; + + if (wc == WEOF) + { + /* It is an invalid character, so ANYCHAR is not accepted. */ + return s; + } + + /* If all positions which have ANYCHAR do not depend on the context + of the next character, calculate the next state with + pre-calculated follows and cache the result. */ + if (d->states[s1].mb_trindex < 0) + { + if (MAX_TRCOUNT <= d->mb_trcount) + { + state_num s3; + for (s3 = -1; s3 < d->tralloc; s3++) + { + free (d->mb_trans[s3]); + d->mb_trans[s3] = NULL; + } + + for (state_num i = 0; i < d->sindex; i++) + d->states[i].mb_trindex = -1; + d->mb_trcount = 0; + } + d->states[s1].mb_trindex = d->mb_trcount++; + } + + if (! d->mb_trans[s]) + { + enum { TRANSPTR_SIZE = sizeof *d->mb_trans[s] }; + enum { TRANSALLOC_SIZE = MAX_TRCOUNT * TRANSPTR_SIZE }; + d->mb_trans[s] = xmalloc (TRANSALLOC_SIZE); + for (int i = 0; i < MAX_TRCOUNT; i++) + d->mb_trans[s][i] = -1; + } + else if (d->mb_trans[s][d->states[s1].mb_trindex] >= 0) + return d->mb_trans[s][d->states[s1].mb_trindex]; + + if (s == -1) + copy (&d->states[s1].mbps, &d->mb_follows); + else + merge (&d->states[s1].mbps, &d->states[s].elems, &d->mb_follows); + + int separate_contexts = state_separate_contexts (d, &d->mb_follows); + state_num s2 = state_index (d, &d->mb_follows, separate_contexts ^ CTX_ANY); + realloc_trans_if_necessary (d); + + d->mb_trans[s][d->states[s1].mb_trindex] = s2; + + return s2; +} + +/* The initial state may encounter a byte which is not a single byte character + nor the first byte of a multibyte character. But it is incorrect for the + initial state to accept such a byte. For example, in Shift JIS the regular + expression "\\" accepts the codepoint 0x5c, but should not accept the second + byte of the codepoint 0x815c. Then the initial state must skip the bytes + that are not a single byte character nor the first byte of a multibyte + character. + + Given DFA state d, use mbs_to_wchar to advance MBP until it reaches + or exceeds P, and return the advanced MBP. If WCP is non-NULL and + the result is greater than P, set *WCP to the final wide character + processed, or to WEOF if no wide character is processed. Otherwise, + if WCP is non-NULL, *WCP may or may not be updated. + + Both P and MBP must be no larger than END. */ +static unsigned char const * +skip_remains_mb (struct dfa *d, unsigned char const *p, + unsigned char const *mbp, char const *end) +{ + if (d->syntax.never_trail[*p]) + return p; + while (mbp < p) + { + wint_t wc; + mbp += mbs_to_wchar (&wc, (char const *) mbp, + end - (char const *) mbp, d); + } + return mbp; +} + +/* Search through a buffer looking for a match to the struct dfa *D. + Find the first occurrence of a string matching the regexp in the + buffer, and the shortest possible version thereof. Return a pointer to + the first character after the match, or NULL if none is found. BEGIN + points to the beginning of the buffer, and END points to the first byte + after its end. Note however that we store a sentinel byte (usually + newline) in *END, so the actual buffer must be one byte longer. + When ALLOW_NL, newlines may appear in the matching string. + If COUNT is non-NULL, increment *COUNT once for each newline processed. + If MULTIBYTE, the input consists of multibyte characters and/or + encoding-error bytes. Otherwise, it consists of single-byte characters. + Here is the list of features that make this DFA matcher punt: + - [M-N] range in non-simple locale: regex is up to 25% faster on [a-z] + - [^...] in non-simple locale + - [[=foo=]] or [[.foo.]] + - [[:alpha:]] etc. in multibyte locale (except [[:digit:]] works OK) + - back-reference: (.)\1 + - word-delimiter in multibyte locale: \<, \>, \b, \B + See using_simple_locale for the definition of "simple locale". */ + +static inline char * +dfaexec_main (struct dfa *d, char const *begin, char *end, bool allow_nl, + size_t *count, bool multibyte) +{ + if (MAX_TRCOUNT <= d->sindex) + { + for (state_num s = d->min_trcount; s < d->sindex; s++) + { + free (d->states[s].elems.elems); + free (d->states[s].mbps.elems); + } + d->sindex = d->min_trcount; + + if (d->trans) + { + for (state_num s = 0; s < d->tralloc; s++) + { + free (d->trans[s]); + free (d->fails[s]); + d->trans[s] = d->fails[s] = NULL; + } + d->trcount = 0; + } + + if (d->localeinfo.multibyte && d->mb_trans) + { + for (state_num s = -1; s < d->tralloc; s++) + { + free (d->mb_trans[s]); + d->mb_trans[s] = NULL; + } + for (state_num s = 0; s < d->min_trcount; s++) + d->states[s].mb_trindex = -1; + d->mb_trcount = 0; + } + } + + if (!d->tralloc) + realloc_trans_if_necessary (d); + + /* Current state. */ + state_num s = 0, s1 = 0; + + /* Current input character. */ + unsigned char const *p = (unsigned char const *) begin; + unsigned char const *mbp = p; + + /* Copy of d->trans so it can be optimized into a register. */ + state_num **trans = d->trans; + unsigned char eol = d->syntax.eolbyte; /* Likewise for eolbyte. */ + unsigned char saved_end = *(unsigned char *) end; + *end = eol; + + if (multibyte) + { + memset (&d->mbs, 0, sizeof d->mbs); + if (d->mb_follows.alloc == 0) + alloc_position_set (&d->mb_follows, d->nleaves); + } + + size_t nlcount = 0; + for (;;) + { + state_num *t; + while ((t = trans[s]) != NULL) + { + if (s < d->min_trcount) + { + if (!multibyte || d->states[s].mbps.nelem == 0) + { + while (t[*p] == s) + p++; + } + if (multibyte) + p = mbp = skip_remains_mb (d, p, mbp, end); + } + + if (multibyte) + { + s1 = s; + + if (d->states[s].mbps.nelem == 0 + || d->localeinfo.sbctowc[*p] != WEOF || (char *) p >= end) + { + /* If an input character does not match ANYCHAR, do it + like a single-byte character. */ + s = t[*p++]; + } + else + { + s = transit_state (d, s, &p, (unsigned char *) end); + mbp = p; + trans = d->trans; + } + } + else + { + s1 = t[*p++]; + t = trans[s1]; + if (! t) + { + state_num tmp = s; + s = s1; + s1 = tmp; /* swap */ + break; + } + if (s < d->min_trcount) + { + while (t[*p] == s1) + p++; + } + s = t[*p++]; + } + } + + if (s < 0) + { + if (s == -2) + { + s = build_state (s1, d, p[-1]); + trans = d->trans; + } + else if ((char *) p <= end && p[-1] == eol && 0 <= d->newlines[s1]) + { + /* The previous character was a newline. Count it, and skip + checking of multibyte character boundary until here. */ + nlcount++; + mbp = p; + + s = (allow_nl ? d->newlines[s1] + : d->syntax.sbit[eol] == CTX_NEWLINE ? 0 + : d->syntax.sbit[eol] == CTX_LETTER ? d->min_trcount - 1 + : d->initstate_notbol); + } + else + { + p = NULL; + goto done; + } + } + else if (d->fails[s]) + { + if ((d->success[s] & d->syntax.sbit[*p]) + || ((char *) p == end + && accepts_in_context (d->states[s].context, CTX_NEWLINE, s, + d))) + goto done; + + if (multibyte && s < d->min_trcount) + p = mbp = skip_remains_mb (d, p, mbp, end); + + s1 = s; + if (!multibyte || d->states[s].mbps.nelem == 0 + || d->localeinfo.sbctowc[*p] != WEOF || (char *) p >= end) + { + /* If a input character does not match ANYCHAR, do it + like a single-byte character. */ + s = d->fails[s][*p++]; + } + else + { + s = transit_state (d, s, &p, (unsigned char *) end); + mbp = p; + trans = d->trans; + } + } + else + { + build_state (s, d, p[0]); + trans = d->trans; + } + } + + done: + if (count) + *count += nlcount; + *end = saved_end; + return (char *) p; +} + +/* Specialized versions of dfaexec for multibyte and single-byte cases. + This is for performance, as dfaexec_main is an inline function. */ + +static char * +dfaexec_mb (struct dfa *d, char const *begin, char *end, + bool allow_nl, size_t *count, bool *backref) +{ + return dfaexec_main (d, begin, end, allow_nl, count, true); +} + +static char * +dfaexec_sb (struct dfa *d, char const *begin, char *end, + bool allow_nl, size_t *count, bool *backref) +{ + return dfaexec_main (d, begin, end, allow_nl, count, false); +} + +/* Always set *BACKREF and return BEGIN. Use this wrapper for + any regexp that uses a construct not supported by this code. */ +static char * +dfaexec_noop (struct dfa *d, char const *begin, char *end, + bool allow_nl, size_t *count, bool *backref) +{ + *backref = true; + return (char *) begin; +} + +/* Like dfaexec_main (D, BEGIN, END, ALLOW_NL, COUNT, D->localeinfo.multibyte), + but faster and set *BACKREF if the DFA code does not support this + regexp usage. */ + +char * +dfaexec (struct dfa *d, char const *begin, char *end, + bool allow_nl, size_t *count, bool *backref) +{ + return d->dfaexec (d, begin, end, allow_nl, count, backref); +} + +struct dfa * +dfasuperset (struct dfa const *d) +{ + return d->superset; +} + +bool +dfaisfast (struct dfa const *d) +{ + return d->fast; +} + +static void +free_mbdata (struct dfa *d) +{ + free (d->multibyte_prop); + free (d->lex.brack.chars); + free (d->mb_follows.elems); + + if (d->mb_trans) + { + state_num s; + for (s = -1; s < d->tralloc; s++) + free (d->mb_trans[s]); + free (d->mb_trans - 2); + } +} + +/* Return true if every construct in D is supported by this DFA matcher. */ +static bool _GL_ATTRIBUTE_PURE +dfa_supported (struct dfa const *d) +{ + for (size_t i = 0; i < d->tindex; i++) + { + switch (d->tokens[i]) + { + case BEGWORD: + case ENDWORD: + case LIMWORD: + case NOTLIMWORD: + if (!d->localeinfo.multibyte) + continue; + FALLTHROUGH; + case BACKREF: + case MBCSET: + return false; + } + } + return true; +} + +/* Disable use of the superset DFA if it is not likely to help + performance. */ +static void +maybe_disable_superset_dfa (struct dfa *d) +{ + if (!d->localeinfo.using_utf8) + return; + + bool have_backref = false; + for (size_t i = 0; i < d->tindex; ++i) + { + switch (d->tokens[i]) + { + case ANYCHAR: + /* Lowered. */ + abort (); + case BACKREF: + have_backref = true; + break; + case MBCSET: + /* Requires multi-byte algorithm. */ + return; + default: + break; + } + } + + if (!have_backref && d->superset) + { + /* The superset DFA is not likely to be much faster, so remove it. */ + dfafree (d->superset); + free (d->superset); + d->superset = NULL; + } + + free_mbdata (d); + d->localeinfo.multibyte = false; + d->dfaexec = dfaexec_sb; + d->fast = true; +} + +static void +dfassbuild (struct dfa *d) +{ + struct dfa *sup = dfaalloc (); + + *sup = *d; + sup->localeinfo.multibyte = false; + sup->dfaexec = dfaexec_sb; + sup->multibyte_prop = NULL; + sup->superset = NULL; + sup->states = NULL; + sup->sindex = 0; + sup->constraints = NULL; + sup->separates = NULL; + sup->follows = NULL; + sup->tralloc = 0; + sup->trans = NULL; + sup->fails = NULL; + sup->success = NULL; + sup->newlines = NULL; + + sup->charclasses = xnmalloc (sup->calloc, sizeof *sup->charclasses); + if (d->cindex) + { + memcpy (sup->charclasses, d->charclasses, + d->cindex * sizeof *sup->charclasses); + } + + sup->tokens = xnmalloc (d->tindex, 2 * sizeof *sup->tokens); + sup->talloc = d->tindex * 2; + + bool have_achar = false; + bool have_nchar = false; + size_t j; + for (size_t i = j = 0; i < d->tindex; i++) + { + switch (d->tokens[i]) + { + case ANYCHAR: + case MBCSET: + case BACKREF: + { + charclass ccl; + fillset (&ccl); + sup->tokens[j++] = CSET + charclass_index (sup, &ccl); + sup->tokens[j++] = STAR; + if (d->tokens[i + 1] == QMARK || d->tokens[i + 1] == STAR + || d->tokens[i + 1] == PLUS) + i++; + have_achar = true; + } + break; + case BEGWORD: + case ENDWORD: + case LIMWORD: + case NOTLIMWORD: + if (d->localeinfo.multibyte) + { + /* These constraints aren't supported in a multibyte locale. + Ignore them in the superset DFA. */ + sup->tokens[j++] = EMPTY; + break; + } + FALLTHROUGH; + default: + sup->tokens[j++] = d->tokens[i]; + if ((0 <= d->tokens[i] && d->tokens[i] < NOTCHAR) + || d->tokens[i] >= CSET) + have_nchar = true; + break; + } + } + sup->tindex = j; + + if (have_nchar && (have_achar || d->localeinfo.multibyte)) + d->superset = sup; + else + { + dfafree (sup); + free (sup); + } +} + +/* Parse and analyze a single string of the given length. */ +void +dfacomp (char const *s, size_t len, struct dfa *d, bool searchflag) +{ + dfaparse (s, len, d); + dfassbuild (d); + + if (dfa_supported (d)) + { + maybe_disable_superset_dfa (d); + dfaanalyze (d, searchflag); + } + else + { + d->dfaexec = dfaexec_noop; + } + + if (d->superset) + { + d->fast = true; + dfaanalyze (d->superset, searchflag); + } +} + +/* Free the storage held by the components of a dfa. */ +void +dfafree (struct dfa *d) +{ + free (d->charclasses); + free (d->tokens); + + if (d->localeinfo.multibyte) + free_mbdata (d); + + free (d->constraints); + free (d->separates); + + for (size_t i = 0; i < d->sindex; ++i) + { + free (d->states[i].elems.elems); + free (d->states[i].mbps.elems); + } + free (d->states); + + if (d->follows) + { + for (size_t i = 0; i < d->tindex; ++i) + free (d->follows[i].elems); + free (d->follows); + } + + if (d->trans) + { + for (size_t i = 0; i < d->tralloc; ++i) + { + free (d->trans[i]); + free (d->fails[i]); + } + + free (d->trans - 2); + free (d->fails); + free (d->newlines); + free (d->success); + } + + if (d->superset) + { + dfafree (d->superset); + free (d->superset); + } +} + +/* Having found the postfix representation of the regular expression, + try to find a long sequence of characters that must appear in any line + containing the r.e. + Finding a "longest" sequence is beyond the scope here; + we take an easy way out and hope for the best. + (Take "(ab|a)b"--please.) + + We do a bottom-up calculation of sequences of characters that must appear + in matches of r.e.'s represented by trees rooted at the nodes of the postfix + representation: + sequences that must appear at the left of the match ("left") + sequences that must appear at the right of the match ("right") + lists of sequences that must appear somewhere in the match ("in") + sequences that must constitute the match ("is") + + When we get to the root of the tree, we use one of the longest of its + calculated "in" sequences as our answer. + + The sequences calculated for the various types of node (in pseudo ANSI c) + are shown below. "p" is the operand of unary operators (and the left-hand + operand of binary operators); "q" is the right-hand operand of binary + operators. + + "ZERO" means "a zero-length sequence" below. + + Type left right is in + ---- ---- ----- -- -- + char c # c # c # c # c + + ANYCHAR ZERO ZERO ZERO ZERO + + MBCSET ZERO ZERO ZERO ZERO + + CSET ZERO ZERO ZERO ZERO + + STAR ZERO ZERO ZERO ZERO + + QMARK ZERO ZERO ZERO ZERO + + PLUS p->left p->right ZERO p->in + + CAT (p->is==ZERO)? (q->is==ZERO)? (p->is!=ZERO && p->in plus + p->left : q->right : q->is!=ZERO) ? q->in plus + p->is##q->left p->right##q->is p->is##q->is : p->right##q->left + ZERO + + OR longest common longest common (do p->is and substrings common + leading trailing to q->is have same p->in and + (sub)sequence (sub)sequence q->in length and content) ? + of p->left of p->right + and q->left and q->right p->is : NULL + + If there's anything else we recognize in the tree, all four sequences get set + to zero-length sequences. If there's something we don't recognize in the + tree, we just return a zero-length sequence. + + Break ties in favor of infrequent letters (choosing 'zzz' in preference to + 'aaa')? + + And ... is it here or someplace that we might ponder "optimizations" such as + egrep 'psi|epsilon' -> egrep 'psi' + egrep 'pepsi|epsilon' -> egrep 'epsi' + (Yes, we now find "epsi" as a "string + that must occur", but we might also + simplify the *entire* r.e. being sought) + grep '[c]' -> grep 'c' + grep '(ab|a)b' -> grep 'ab' + grep 'ab*' -> grep 'a' + grep 'a*b' -> grep 'b' + + There are several issues: + + Is optimization easy (enough)? + + Does optimization actually accomplish anything, + or is the automaton you get from "psi|epsilon" (for example) + the same as the one you get from "psi" (for example)? + + Are optimizable r.e.'s likely to be used in real-life situations + (something like 'ab*' is probably unlikely; something like is + 'psi|epsilon' is likelier)? */ + +static char * +icatalloc (char *old, char const *new) +{ + size_t newsize = strlen (new); + if (newsize == 0) + return old; + size_t oldsize = strlen (old); + char *result = xrealloc (old, oldsize + newsize + 1); + memcpy (result + oldsize, new, newsize + 1); + return result; +} + +static void +freelist (char **cpp) +{ + while (*cpp) + free (*cpp++); +} + +static char ** +enlist (char **cpp, char *new, size_t len) +{ + new = memcpy (xmalloc (len + 1), new, len); + new[len] = '\0'; + /* Is there already something in the list that's new (or longer)? */ + size_t i; + for (i = 0; cpp[i] != NULL; ++i) + if (strstr (cpp[i], new) != NULL) + { + free (new); + return cpp; + } + /* Eliminate any obsoleted strings. */ + for (size_t j = 0; cpp[j] != NULL; ) + if (strstr (new, cpp[j]) == NULL) + ++j; + else + { + free (cpp[j]); + if (--i == j) + break; + cpp[j] = cpp[i]; + cpp[i] = NULL; + } + /* Add the new string. */ + cpp = xnrealloc (cpp, i + 2, sizeof *cpp); + cpp[i] = new; + cpp[i + 1] = NULL; + return cpp; +} + +/* Given pointers to two strings, return a pointer to an allocated + list of their distinct common substrings. */ +static char ** +comsubs (char *left, char const *right) +{ + char **cpp = xzalloc (sizeof *cpp); + + for (char *lcp = left; *lcp != '\0'; lcp++) + { + size_t len = 0; + char *rcp = strchr (right, *lcp); + while (rcp != NULL) + { + size_t i; + for (i = 1; lcp[i] != '\0' && lcp[i] == rcp[i]; ++i) + continue; + if (i > len) + len = i; + rcp = strchr (rcp + 1, *lcp); + } + if (len != 0) + cpp = enlist (cpp, lcp, len); + } + return cpp; +} + +static char ** +addlists (char **old, char **new) +{ + for (; *new; new++) + old = enlist (old, *new, strlen (*new)); + return old; +} + +/* Given two lists of substrings, return a new list giving substrings + common to both. */ +static char ** +inboth (char **left, char **right) +{ + char **both = xzalloc (sizeof *both); + + for (size_t lnum = 0; left[lnum] != NULL; ++lnum) + { + for (size_t rnum = 0; right[rnum] != NULL; ++rnum) + { + char **temp = comsubs (left[lnum], right[rnum]); + both = addlists (both, temp); + freelist (temp); + free (temp); + } + } + return both; +} + +typedef struct must must; + +struct must +{ + char **in; + char *left; + char *right; + char *is; + bool begline; + bool endline; + must *prev; +}; + +static must * +allocmust (must *mp, size_t size) +{ + must *new_mp = xmalloc (sizeof *new_mp); + new_mp->in = xzalloc (sizeof *new_mp->in); + new_mp->left = xzalloc (size); + new_mp->right = xzalloc (size); + new_mp->is = xzalloc (size); + new_mp->begline = false; + new_mp->endline = false; + new_mp->prev = mp; + return new_mp; +} + +static void +resetmust (must *mp) +{ + freelist (mp->in); + mp->in[0] = NULL; + mp->left[0] = mp->right[0] = mp->is[0] = '\0'; + mp->begline = false; + mp->endline = false; +} + +static void +freemust (must *mp) +{ + freelist (mp->in); + free (mp->in); + free (mp->left); + free (mp->right); + free (mp->is); + free (mp); +} + +struct dfamust * +dfamust (struct dfa const *d) +{ + must *mp = NULL; + char const *result = ""; + bool exact = false; + bool begline = false; + bool endline = false; + bool need_begline = false; + bool need_endline = false; + bool case_fold_unibyte = d->syntax.case_fold && MB_CUR_MAX == 1; + + for (size_t ri = 1; ri + 1 < d->tindex; ri++) + { + token t = d->tokens[ri]; + switch (t) + { + case BEGLINE: + mp = allocmust (mp, 2); + mp->begline = true; + need_begline = true; + break; + case ENDLINE: + mp = allocmust (mp, 2); + mp->endline = true; + need_endline = true; + break; + case LPAREN: + case RPAREN: + assert (!"neither LPAREN nor RPAREN may appear here"); + + case EMPTY: + case BEGWORD: + case ENDWORD: + case LIMWORD: + case NOTLIMWORD: + case BACKREF: + case ANYCHAR: + case MBCSET: + mp = allocmust (mp, 2); + break; + + case STAR: + case QMARK: + resetmust (mp); + break; + + case OR: + { + char **new; + must *rmp = mp; + must *lmp = mp = mp->prev; + size_t j, ln, rn, n; + + /* Guaranteed to be. Unlikely, but ... */ + if (streq (lmp->is, rmp->is)) + { + lmp->begline &= rmp->begline; + lmp->endline &= rmp->endline; + } + else + { + lmp->is[0] = '\0'; + lmp->begline = false; + lmp->endline = false; + } + /* Left side--easy */ + size_t i = 0; + while (lmp->left[i] != '\0' && lmp->left[i] == rmp->left[i]) + ++i; + lmp->left[i] = '\0'; + /* Right side */ + ln = strlen (lmp->right); + rn = strlen (rmp->right); + n = ln; + if (n > rn) + n = rn; + for (i = 0; i < n; ++i) + if (lmp->right[ln - i - 1] != rmp->right[rn - i - 1]) + break; + for (j = 0; j < i; ++j) + lmp->right[j] = lmp->right[(ln - i) + j]; + lmp->right[j] = '\0'; + new = inboth (lmp->in, rmp->in); + freelist (lmp->in); + free (lmp->in); + lmp->in = new; + freemust (rmp); + } + break; + + case PLUS: + mp->is[0] = '\0'; + break; + + case END: + assert (!mp->prev); + for (size_t i = 0; mp->in[i] != NULL; ++i) + if (strlen (mp->in[i]) > strlen (result)) + result = mp->in[i]; + if (streq (result, mp->is)) + { + if ((!need_begline || mp->begline) && (!need_endline + || mp->endline)) + exact = true; + begline = mp->begline; + endline = mp->endline; + } + goto done; + + case CAT: + { + must *rmp = mp; + must *lmp = mp = mp->prev; + + /* In. Everything in left, plus everything in + right, plus concatenation of + left's right and right's left. */ + lmp->in = addlists (lmp->in, rmp->in); + if (lmp->right[0] != '\0' && rmp->left[0] != '\0') + { + size_t lrlen = strlen (lmp->right); + size_t rllen = strlen (rmp->left); + char *tp = xmalloc (lrlen + rllen); + memcpy (tp, lmp->right, lrlen); + memcpy (tp + lrlen, rmp->left, rllen); + lmp->in = enlist (lmp->in, tp, lrlen + rllen); + free (tp); + } + /* Left-hand */ + if (lmp->is[0] != '\0') + lmp->left = icatalloc (lmp->left, rmp->left); + /* Right-hand */ + if (rmp->is[0] == '\0') + lmp->right[0] = '\0'; + lmp->right = icatalloc (lmp->right, rmp->right); + /* Guaranteed to be */ + if ((lmp->is[0] != '\0' || lmp->begline) + && (rmp->is[0] != '\0' || rmp->endline)) + { + lmp->is = icatalloc (lmp->is, rmp->is); + lmp->endline = rmp->endline; + } + else + { + lmp->is[0] = '\0'; + lmp->begline = false; + lmp->endline = false; + } + freemust (rmp); + } + break; + + case '\0': + /* Not on *my* shift. */ + goto done; + + default: + if (CSET <= t) + { + /* If T is a singleton, or if case-folding in a unibyte + locale and T's members all case-fold to the same char, + convert T to one of its members. Otherwise, do + nothing further with T. */ + charclass *ccl = &d->charclasses[t - CSET]; + int j; + for (j = 0; j < NOTCHAR; j++) + if (tstbit (j, ccl)) + break; + if (! (j < NOTCHAR)) + { + mp = allocmust (mp, 2); + break; + } + t = j; + while (++j < NOTCHAR) + if (tstbit (j, ccl) + && ! (case_fold_unibyte + && toupper (j) == toupper (t))) + break; + if (j < NOTCHAR) + { + mp = allocmust (mp, 2); + break; + } + } + + size_t rj = ri + 2; + if (d->tokens[ri + 1] == CAT) + { + for (; rj < d->tindex - 1; rj += 2) + { + if ((rj != ri && (d->tokens[rj] <= 0 + || NOTCHAR <= d->tokens[rj])) + || d->tokens[rj + 1] != CAT) + break; + } + } + mp = allocmust (mp, ((rj - ri) >> 1) + 1); + mp->is[0] = mp->left[0] = mp->right[0] + = case_fold_unibyte ? toupper (t) : t; + + size_t i; + for (i = 1; ri + 2 < rj; i++) + { + ri += 2; + t = d->tokens[ri]; + mp->is[i] = mp->left[i] = mp->right[i] + = case_fold_unibyte ? toupper (t) : t; + } + mp->is[i] = mp->left[i] = mp->right[i] = '\0'; + mp->in = enlist (mp->in, mp->is, i); + break; + } + } + done:; + + struct dfamust *dm = NULL; + if (*result) + { + dm = xmalloc (sizeof *dm); + dm->exact = exact; + dm->begline = begline; + dm->endline = endline; + dm->must = xstrdup (result); + } + + while (mp) + { + must *prev = mp->prev; + freemust (mp); + mp = prev; + } + + return dm; +} + +void +dfamustfree (struct dfamust *dm) +{ + free (dm->must); + free (dm); +} + +struct dfa * +dfaalloc (void) +{ + return xmalloc (sizeof (struct dfa)); +} + +/* Initialize DFA. */ +void +dfasyntax (struct dfa *dfa, struct localeinfo const *linfo, + reg_syntax_t bits, int dfaopts) +{ + memset (dfa, 0, offsetof (struct dfa, dfaexec)); + dfa->dfaexec = linfo->multibyte ? dfaexec_mb : dfaexec_sb; + dfa->simple_locale = using_simple_locale (linfo->multibyte); + dfa->localeinfo = *linfo; + + dfa->fast = !dfa->localeinfo.multibyte; + + dfa->canychar = -1; + dfa->lex.cur_mb_len = 1; + dfa->syntax.syntax_bits_set = true; + dfa->syntax.case_fold = (bits & RE_ICASE) != 0; + dfa->syntax.anchor = (dfaopts & DFA_ANCHOR) != 0; + dfa->syntax.eolbyte = dfaopts & DFA_EOL_NUL ? '\0' : '\n'; + dfa->syntax.syntax_bits = bits; + + for (int i = CHAR_MIN; i <= CHAR_MAX; ++i) + { + unsigned char uc = i; + + dfa->syntax.sbit[uc] = char_context (dfa, uc); + switch (dfa->syntax.sbit[uc]) + { + case CTX_LETTER: + setbit (uc, &dfa->syntax.letters); + break; + case CTX_NEWLINE: + setbit (uc, &dfa->syntax.newline); + break; + } + + /* POSIX requires that the five bytes in "\n\r./" (including the + terminating NUL) cannot occur inside a multibyte character. */ + dfa->syntax.never_trail[uc] = (dfa->localeinfo.using_utf8 + ? (uc & 0xc0) != 0x80 + : strchr ("\n\r./", uc) != NULL); + } +} + +/* vim:set shiftwidth=2: */ diff --git a/lib/dfa.h b/lib/dfa.h new file mode 100644 index 0000000..fea815d --- /dev/null +++ b/lib/dfa.h @@ -0,0 +1,117 @@ +/* dfa.h - declarations for GNU deterministic regexp compiler + Copyright (C) 1988, 1998, 2007, 2009-2018 Free Software Foundation, Inc. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3, or (at your option) + any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, write to the Free Software + Foundation, Inc., + 51 Franklin Street - Fifth Floor, Boston, MA 02110-1301, USA */ + +/* Written June, 1988 by Mike Haertel */ + +#include <regex.h> +#include <stdbool.h> +#include <stddef.h> + +struct localeinfo; /* See localeinfo.h. */ + +/* Element of a list of strings, at least one of which is known to + appear in any R.E. matching the DFA. */ +struct dfamust +{ + bool exact; + bool begline; + bool endline; + char *must; +}; + +/* The dfa structure. It is completely opaque. */ +struct dfa; + +/* Entry points. */ + +/* Allocate a struct dfa. The struct dfa is completely opaque. + The returned pointer should be passed directly to free() after + calling dfafree() on it. */ +extern struct dfa *dfaalloc (void) _GL_ATTRIBUTE_MALLOC; + +/* DFA options that can be ORed together, for dfasyntax's 4th arg. */ +enum + { + /* ^ and $ match only the start and end of data, and do not match + end-of-line within data. This is always false for grep, but + possibly true for other apps. */ + DFA_ANCHOR = 1 << 0, + + /* '\0' in data is end-of-line, instead of the traditional '\n'. */ + DFA_EOL_NUL = 1 << 1 + }; + +/* Initialize or reinitialize a DFA. This must be called before + any of the routines below. The arguments are: + 1. The DFA to operate on. + 2. Information about the current locale. + 3. Syntax bits described in regex.h. + 4. Additional DFA options described above. */ +extern void dfasyntax (struct dfa *, struct localeinfo const *, + reg_syntax_t, int); + +/* Build and return the struct dfamust from the given struct dfa. */ +extern struct dfamust *dfamust (struct dfa const *); + +/* Free the storage held by the components of a struct dfamust. */ +extern void dfamustfree (struct dfamust *); + +/* Compile the given string of the given length into the given struct dfa. + Final argument is a flag specifying whether to build a searching or an + exact matcher. */ +extern void dfacomp (char const *, size_t, struct dfa *, bool); + +/* Search through a buffer looking for a match to the given struct dfa. + Find the first occurrence of a string matching the regexp in the + buffer, and the shortest possible version thereof. Return a pointer to + the first character after the match, or NULL if none is found. BEGIN + points to the beginning of the buffer, and END points to the first byte + after its end. Note however that we store a sentinel byte (usually + newline) in *END, so the actual buffer must be one byte longer. + When ALLOW_NL is true, newlines may appear in the matching string. + If COUNT is non-NULL, increment *COUNT once for each newline processed. + Finally, if BACKREF is non-NULL set *BACKREF to indicate whether we + encountered a back-reference. The caller can use this to decide + whether to fall back on a backtracking matcher. */ +extern char *dfaexec (struct dfa *d, char const *begin, char *end, + bool allow_nl, size_t *count, bool *backref); + +/* Return a superset for D. The superset matches everything that D + matches, along with some other strings (though the latter should be + rare, for efficiency reasons). Return a null pointer if no useful + superset is available. */ +extern struct dfa *dfasuperset (struct dfa const *d) _GL_ATTRIBUTE_PURE; + +/* The DFA is likely to be fast. */ +extern bool dfaisfast (struct dfa const *) _GL_ATTRIBUTE_PURE; + +/* Free the storage held by the components of a struct dfa. */ +extern void dfafree (struct dfa *); + +/* Error handling. */ + +/* dfawarn() is called by the regexp routines whenever a regex is compiled + that likely doesn't do what the user wanted. It takes a single + argument, a NUL-terminated string describing the situation. The user + must supply a dfawarn. */ +extern void dfawarn (const char *); + +/* dfaerror() is called by the regexp routines whenever an error occurs. It + takes a single argument, a NUL-terminated string describing the error. + The user must supply a dfaerror. */ +extern _Noreturn void dfaerror (const char *); diff --git a/lib/dirname-lgpl.c b/lib/dirname-lgpl.c new file mode 100644 index 0000000..a40f6a9 --- /dev/null +++ b/lib/dirname-lgpl.c @@ -0,0 +1,86 @@ +/* dirname.c -- return all but the last element in a file name + + Copyright (C) 1990, 1998, 2000-2001, 2003-2006, 2009-2018 Free Software + Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +#include <config.h> + +#include "dirname.h" + +#include <stdlib.h> +#include <string.h> + +/* Return the length of the prefix of FILE that will be used by + dir_name. If FILE is in the working directory, this returns zero + even though 'dir_name (FILE)' will return ".". Works properly even + if there are trailing slashes (by effectively ignoring them). */ + +size_t +dir_len (char const *file) +{ + size_t prefix_length = FILE_SYSTEM_PREFIX_LEN (file); + size_t length; + + /* Advance prefix_length beyond important leading slashes. */ + prefix_length += (prefix_length != 0 + ? (FILE_SYSTEM_DRIVE_PREFIX_CAN_BE_RELATIVE + && ISSLASH (file[prefix_length])) + : (ISSLASH (file[0]) + ? ((DOUBLE_SLASH_IS_DISTINCT_ROOT + && ISSLASH (file[1]) && ! ISSLASH (file[2]) + ? 2 : 1)) + : 0)); + + /* Strip the basename and any redundant slashes before it. */ + for (length = last_component (file) - file; + prefix_length < length; length--) + if (! ISSLASH (file[length - 1])) + break; + return length; +} + + +/* In general, we can't use the builtin 'dirname' function if available, + since it has different meanings in different environments. + In some environments the builtin 'dirname' modifies its argument. + + Return the leading directories part of FILE, allocated with malloc. + Works properly even if there are trailing slashes (by effectively + ignoring them). Return NULL on failure. + + If lstat (FILE) would succeed, then { chdir (dir_name (FILE)); + lstat (base_name (FILE)); } will access the same file. Likewise, + if the sequence { chdir (dir_name (FILE)); + rename (base_name (FILE), "foo"); } succeeds, you have renamed FILE + to "foo" in the same directory FILE was in. */ + +char * +mdir_name (char const *file) +{ + size_t length = dir_len (file); + bool append_dot = (length == 0 + || (FILE_SYSTEM_DRIVE_PREFIX_CAN_BE_RELATIVE + && length == FILE_SYSTEM_PREFIX_LEN (file) + && file[2] != '\0' && ! ISSLASH (file[2]))); + char *dir = malloc (length + append_dot + 1); + if (!dir) + return NULL; + memcpy (dir, file, length); + if (append_dot) + dir[length++] = '.'; + dir[length] = '\0'; + return dir; +} diff --git a/lib/dirname.h b/lib/dirname.h new file mode 100644 index 0000000..8b51111 --- /dev/null +++ b/lib/dirname.h @@ -0,0 +1,54 @@ +/* Take file names apart into directory and base names. + + Copyright (C) 1998, 2001, 2003-2006, 2009-2018 Free Software Foundation, + Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +#ifndef DIRNAME_H_ +# define DIRNAME_H_ 1 + +# include <stdbool.h> +# include <stddef.h> +# include "dosname.h" + +# ifndef DIRECTORY_SEPARATOR +# define DIRECTORY_SEPARATOR '/' +# endif + +# ifndef DOUBLE_SLASH_IS_DISTINCT_ROOT +# define DOUBLE_SLASH_IS_DISTINCT_ROOT 0 +# endif + +#ifdef __cplusplus +extern "C" { +#endif + +# if GNULIB_DIRNAME +char *base_name (char const *file) _GL_ATTRIBUTE_MALLOC; +char *dir_name (char const *file); +# endif + +char *mdir_name (char const *file); +size_t base_len (char const *file) _GL_ATTRIBUTE_PURE; +size_t dir_len (char const *file) _GL_ATTRIBUTE_PURE; +char *last_component (char const *file) _GL_ATTRIBUTE_PURE; + +bool strip_trailing_slashes (char *file); + +#ifdef __cplusplus +} /* extern "C" */ +#endif + +#endif /* not DIRNAME_H_ */ diff --git a/lib/dosname.h b/lib/dosname.h new file mode 100644 index 0000000..fef3b6d --- /dev/null +++ b/lib/dosname.h @@ -0,0 +1,52 @@ +/* File names on MS-DOS/Windows systems. + + Copyright (C) 2000-2001, 2004-2006, 2009-2018 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. + + From Paul Eggert and Jim Meyering. */ + +#ifndef _DOSNAME_H +#define _DOSNAME_H + +#if (defined _WIN32 || defined __CYGWIN__ \ + || defined __EMX__ || defined __MSDOS__ || defined __DJGPP__) + /* This internal macro assumes ASCII, but all hosts that support drive + letters use ASCII. */ +# define _IS_DRIVE_LETTER(C) (((unsigned int) (C) | ('a' - 'A')) - 'a' \ + <= 'z' - 'a') +# define FILE_SYSTEM_PREFIX_LEN(Filename) \ + (_IS_DRIVE_LETTER ((Filename)[0]) && (Filename)[1] == ':' ? 2 : 0) +# ifndef __CYGWIN__ +# define FILE_SYSTEM_DRIVE_PREFIX_CAN_BE_RELATIVE 1 +# endif +# define ISSLASH(C) ((C) == '/' || (C) == '\\') +#else +# define FILE_SYSTEM_PREFIX_LEN(Filename) 0 +# define ISSLASH(C) ((C) == '/') +#endif + +#ifndef FILE_SYSTEM_DRIVE_PREFIX_CAN_BE_RELATIVE +# define FILE_SYSTEM_DRIVE_PREFIX_CAN_BE_RELATIVE 0 +#endif + +#if FILE_SYSTEM_DRIVE_PREFIX_CAN_BE_RELATIVE +# define IS_ABSOLUTE_FILE_NAME(F) ISSLASH ((F)[FILE_SYSTEM_PREFIX_LEN (F)]) +# else +# define IS_ABSOLUTE_FILE_NAME(F) \ + (ISSLASH ((F)[0]) || FILE_SYSTEM_PREFIX_LEN (F) != 0) +#endif +#define IS_RELATIVE_FILE_NAME(F) (! IS_ABSOLUTE_FILE_NAME (F)) + +#endif /* DOSNAME_H_ */ diff --git a/lib/errno.in.h b/lib/errno.in.h new file mode 100644 index 0000000..b95c4e9 --- /dev/null +++ b/lib/errno.in.h @@ -0,0 +1,279 @@ +/* A POSIX-like <errno.h>. + + Copyright (C) 2008-2018 Free Software Foundation, Inc. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3, or (at your option) + any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, see <https://www.gnu.org/licenses/>. */ + +#ifndef _@GUARD_PREFIX@_ERRNO_H + +#if __GNUC__ >= 3 +@PRAGMA_SYSTEM_HEADER@ +#endif +@PRAGMA_COLUMNS@ + +/* The include_next requires a split double-inclusion guard. */ +#@INCLUDE_NEXT@ @NEXT_ERRNO_H@ + +#ifndef _@GUARD_PREFIX@_ERRNO_H +#define _@GUARD_PREFIX@_ERRNO_H + + +/* On native Windows platforms, many macros are not defined. */ +# if defined _WIN32 && ! defined __CYGWIN__ + +/* These are the same values as defined by MSVC 10, for interoperability. */ + +# ifndef ENOMSG +# define ENOMSG 122 +# define GNULIB_defined_ENOMSG 1 +# endif + +# ifndef EIDRM +# define EIDRM 111 +# define GNULIB_defined_EIDRM 1 +# endif + +# ifndef ENOLINK +# define ENOLINK 121 +# define GNULIB_defined_ENOLINK 1 +# endif + +# ifndef EPROTO +# define EPROTO 134 +# define GNULIB_defined_EPROTO 1 +# endif + +# ifndef EBADMSG +# define EBADMSG 104 +# define GNULIB_defined_EBADMSG 1 +# endif + +# ifndef EOVERFLOW +# define EOVERFLOW 132 +# define GNULIB_defined_EOVERFLOW 1 +# endif + +# ifndef ENOTSUP +# define ENOTSUP 129 +# define GNULIB_defined_ENOTSUP 1 +# endif + +# ifndef ENETRESET +# define ENETRESET 117 +# define GNULIB_defined_ENETRESET 1 +# endif + +# ifndef ECONNABORTED +# define ECONNABORTED 106 +# define GNULIB_defined_ECONNABORTED 1 +# endif + +# ifndef ECANCELED +# define ECANCELED 105 +# define GNULIB_defined_ECANCELED 1 +# endif + +# ifndef EOWNERDEAD +# define EOWNERDEAD 133 +# define GNULIB_defined_EOWNERDEAD 1 +# endif + +# ifndef ENOTRECOVERABLE +# define ENOTRECOVERABLE 127 +# define GNULIB_defined_ENOTRECOVERABLE 1 +# endif + +# ifndef EINPROGRESS +# define EINPROGRESS 112 +# define EALREADY 103 +# define ENOTSOCK 128 +# define EDESTADDRREQ 109 +# define EMSGSIZE 115 +# define EPROTOTYPE 136 +# define ENOPROTOOPT 123 +# define EPROTONOSUPPORT 135 +# define EOPNOTSUPP 130 +# define EAFNOSUPPORT 102 +# define EADDRINUSE 100 +# define EADDRNOTAVAIL 101 +# define ENETDOWN 116 +# define ENETUNREACH 118 +# define ECONNRESET 108 +# define ENOBUFS 119 +# define EISCONN 113 +# define ENOTCONN 126 +# define ETIMEDOUT 138 +# define ECONNREFUSED 107 +# define ELOOP 114 +# define EHOSTUNREACH 110 +# define EWOULDBLOCK 140 +# define GNULIB_defined_ESOCK 1 +# endif + +# ifndef ETXTBSY +# define ETXTBSY 139 +# define ENODATA 120 /* not required by POSIX */ +# define ENOSR 124 /* not required by POSIX */ +# define ENOSTR 125 /* not required by POSIX */ +# define ETIME 137 /* not required by POSIX */ +# define EOTHER 131 /* not required by POSIX */ +# define GNULIB_defined_ESTREAMS 1 +# endif + +/* These are intentionally the same values as the WSA* error numbers, defined + in <winsock2.h>. */ +# define ESOCKTNOSUPPORT 10044 /* not required by POSIX */ +# define EPFNOSUPPORT 10046 /* not required by POSIX */ +# define ESHUTDOWN 10058 /* not required by POSIX */ +# define ETOOMANYREFS 10059 /* not required by POSIX */ +# define EHOSTDOWN 10064 /* not required by POSIX */ +# define EPROCLIM 10067 /* not required by POSIX */ +# define EUSERS 10068 /* not required by POSIX */ +# define EDQUOT 10069 +# define ESTALE 10070 +# define EREMOTE 10071 /* not required by POSIX */ +# define GNULIB_defined_EWINSOCK 1 + +# endif + + +/* On OSF/1 5.1, when _XOPEN_SOURCE_EXTENDED is not defined, the macros + EMULTIHOP, ENOLINK, EOVERFLOW are not defined. */ +# if @EMULTIHOP_HIDDEN@ +# define EMULTIHOP @EMULTIHOP_VALUE@ +# define GNULIB_defined_EMULTIHOP 1 +# endif +# if @ENOLINK_HIDDEN@ +# define ENOLINK @ENOLINK_VALUE@ +# define GNULIB_defined_ENOLINK 1 +# endif +# if @EOVERFLOW_HIDDEN@ +# define EOVERFLOW @EOVERFLOW_VALUE@ +# define GNULIB_defined_EOVERFLOW 1 +# endif + + +/* On OpenBSD 4.0 and on native Windows, the macros ENOMSG, EIDRM, ENOLINK, + EPROTO, EMULTIHOP, EBADMSG, EOVERFLOW, ENOTSUP, ECANCELED are not defined. + Likewise, on NonStop Kernel, EDQUOT is not defined. + Define them here. Values >= 2000 seem safe to use: Solaris ESTALE = 151, + HP-UX EWOULDBLOCK = 246, IRIX EDQUOT = 1133. + + Note: When one of these systems defines some of these macros some day, + binaries will have to be recompiled so that they recognizes the new + errno values from the system. */ + +# ifndef ENOMSG +# define ENOMSG 2000 +# define GNULIB_defined_ENOMSG 1 +# endif + +# ifndef EIDRM +# define EIDRM 2001 +# define GNULIB_defined_EIDRM 1 +# endif + +# ifndef ENOLINK +# define ENOLINK 2002 +# define GNULIB_defined_ENOLINK 1 +# endif + +# ifndef EPROTO +# define EPROTO 2003 +# define GNULIB_defined_EPROTO 1 +# endif + +# ifndef EMULTIHOP +# define EMULTIHOP 2004 +# define GNULIB_defined_EMULTIHOP 1 +# endif + +# ifndef EBADMSG +# define EBADMSG 2005 +# define GNULIB_defined_EBADMSG 1 +# endif + +# ifndef EOVERFLOW +# define EOVERFLOW 2006 +# define GNULIB_defined_EOVERFLOW 1 +# endif + +# ifndef ENOTSUP +# define ENOTSUP 2007 +# define GNULIB_defined_ENOTSUP 1 +# endif + +# ifndef ENETRESET +# define ENETRESET 2011 +# define GNULIB_defined_ENETRESET 1 +# endif + +# ifndef ECONNABORTED +# define ECONNABORTED 2012 +# define GNULIB_defined_ECONNABORTED 1 +# endif + +# ifndef ESTALE +# define ESTALE 2009 +# define GNULIB_defined_ESTALE 1 +# endif + +# ifndef EDQUOT +# define EDQUOT 2010 +# define GNULIB_defined_EDQUOT 1 +# endif + +# ifndef ECANCELED +# define ECANCELED 2008 +# define GNULIB_defined_ECANCELED 1 +# endif + +/* On many platforms, the macros EOWNERDEAD and ENOTRECOVERABLE are not + defined. */ + +# ifndef EOWNERDEAD +# if defined __sun + /* Use the same values as defined for Solaris >= 8, for + interoperability. */ +# define EOWNERDEAD 58 +# define ENOTRECOVERABLE 59 +# elif defined _WIN32 && ! defined __CYGWIN__ + /* We have a conflict here: pthreads-win32 defines these values + differently than MSVC 10. It's hairy to decide which one to use. */ +# if defined __MINGW32__ && !defined USE_WINDOWS_THREADS + /* Use the same values as defined by pthreads-win32, for + interoperability. */ +# define EOWNERDEAD 43 +# define ENOTRECOVERABLE 44 +# else + /* Use the same values as defined by MSVC 10, for + interoperability. */ +# define EOWNERDEAD 133 +# define ENOTRECOVERABLE 127 +# endif +# else +# define EOWNERDEAD 2013 +# define ENOTRECOVERABLE 2014 +# endif +# define GNULIB_defined_EOWNERDEAD 1 +# define GNULIB_defined_ENOTRECOVERABLE 1 +# endif + +# ifndef EILSEQ +# define EILSEQ 2015 +# define GNULIB_defined_EILSEQ 1 +# endif + +#endif /* _@GUARD_PREFIX@_ERRNO_H */ +#endif /* _@GUARD_PREFIX@_ERRNO_H */ diff --git a/lib/error.c b/lib/error.c new file mode 100644 index 0000000..adc8733 --- /dev/null +++ b/lib/error.c @@ -0,0 +1,411 @@ +/* Error handler for noninteractive utilities + Copyright (C) 1990-1998, 2000-2007, 2009-2018 Free Software Foundation, Inc. + This file is part of the GNU C Library. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +/* Written by David MacKenzie <djm@gnu.ai.mit.edu>. */ + +#if !_LIBC +# include <config.h> +#endif + +#include "error.h" + +#include <stdarg.h> +#include <stdio.h> +#include <stdlib.h> +#include <string.h> + +#if !_LIBC && ENABLE_NLS +# include "gettext.h" +# define _(msgid) gettext (msgid) +#endif + +#ifdef _LIBC +# include <libintl.h> +# include <stdbool.h> +# include <stdint.h> +# include <wchar.h> +# define mbsrtowcs __mbsrtowcs +# define USE_UNLOCKED_IO 0 +# define _GL_ATTRIBUTE_FORMAT_PRINTF(a, b) +# define _GL_ARG_NONNULL(a) +#else +# include "getprogname.h" +#endif + +#if USE_UNLOCKED_IO +# include "unlocked-io.h" +#endif + +#ifndef _ +# define _(String) String +#endif + +/* If NULL, error will flush stdout, then print on stderr the program + name, a colon and a space. Otherwise, error will call this + function without parameters instead. */ +void (*error_print_progname) (void); + +/* This variable is incremented each time 'error' is called. */ +unsigned int error_message_count; + +#ifdef _LIBC +/* In the GNU C library, there is a predefined variable for this. */ + +# define program_name program_invocation_name +# include <errno.h> +# include <limits.h> +# include <libio/libioP.h> + +/* In GNU libc we want do not want to use the common name 'error' directly. + Instead make it a weak alias. */ +extern void __error (int status, int errnum, const char *message, ...) + __attribute__ ((__format__ (__printf__, 3, 4))); +extern void __error_at_line (int status, int errnum, const char *file_name, + unsigned int line_number, const char *message, + ...) + __attribute__ ((__format__ (__printf__, 5, 6))); +# define error __error +# define error_at_line __error_at_line + +# include <libio/iolibio.h> +# define fflush(s) _IO_fflush (s) +# undef putc +# define putc(c, fp) _IO_putc (c, fp) + +# include <bits/libc-lock.h> + +#else /* not _LIBC */ + +# include <fcntl.h> +# include <unistd.h> + +# if defined _WIN32 && ! defined __CYGWIN__ +/* Get declarations of the native Windows API functions. */ +# define WIN32_LEAN_AND_MEAN +# include <windows.h> +/* Get _get_osfhandle. */ +# if GNULIB_MSVC_NOTHROW +# include "msvc-nothrow.h" +# else +# include <io.h> +# endif +# endif + +/* The gnulib override of fcntl is not needed in this file. */ +# undef fcntl + +# if !(GNULIB_STRERROR_R_POSIX || HAVE_DECL_STRERROR_R) +# ifndef HAVE_DECL_STRERROR_R +"this configure-time declaration test was not run" +# endif +# if STRERROR_R_CHAR_P +char *strerror_r (int errnum, char *buf, size_t buflen); +# else +int strerror_r (int errnum, char *buf, size_t buflen); +# endif +# endif + +#define program_name getprogname () + +# if GNULIB_STRERROR_R_POSIX || HAVE_STRERROR_R || defined strerror_r +# define __strerror_r strerror_r +# endif /* GNULIB_STRERROR_R_POSIX || HAVE_STRERROR_R || defined strerror_r */ +#endif /* not _LIBC */ + +#if !_LIBC +/* Return non-zero if FD is open. */ +static int +is_open (int fd) +{ +# if defined _WIN32 && ! defined __CYGWIN__ + /* On native Windows: The initial state of unassigned standard file + descriptors is that they are open but point to an INVALID_HANDLE_VALUE. + There is no fcntl, and the gnulib replacement fcntl does not support + F_GETFL. */ + return (HANDLE) _get_osfhandle (fd) != INVALID_HANDLE_VALUE; +# else +# ifndef F_GETFL +# error Please port fcntl to your platform +# endif + return 0 <= fcntl (fd, F_GETFL); +# endif +} +#endif + +static void +flush_stdout (void) +{ +#if !_LIBC + int stdout_fd; + +# if GNULIB_FREOPEN_SAFER + /* Use of gnulib's freopen-safer module normally ensures that + fileno (stdout) == 1 + whenever stdout is open. */ + stdout_fd = STDOUT_FILENO; +# else + /* POSIX states that fileno (stdout) after fclose is unspecified. But in + practice it is not a problem, because stdout is statically allocated and + the fd of a FILE stream is stored as a field in its allocated memory. */ + stdout_fd = fileno (stdout); +# endif + /* POSIX states that fflush (stdout) after fclose is unspecified; it + is safe in glibc, but not on all other platforms. fflush (NULL) + is always defined, but too draconian. */ + if (0 <= stdout_fd && is_open (stdout_fd)) +#endif + fflush (stdout); +} + +static void +print_errno_message (int errnum) +{ + char const *s; + +#if _LIBC || GNULIB_STRERROR_R_POSIX || defined HAVE_STRERROR_R + char errbuf[1024]; +# if _LIBC || (!GNULIB_STRERROR_R_POSIX && STRERROR_R_CHAR_P) + s = __strerror_r (errnum, errbuf, sizeof errbuf); +# else + if (__strerror_r (errnum, errbuf, sizeof errbuf) == 0) + s = errbuf; + else + s = 0; +# endif +#else + s = strerror (errnum); +#endif + +#if !_LIBC + if (! s) + s = _("Unknown system error"); +#endif + +#if _LIBC + __fxprintf (NULL, ": %s", s); +#else + fprintf (stderr, ": %s", s); +#endif +} + +static void _GL_ATTRIBUTE_FORMAT_PRINTF (3, 0) _GL_ARG_NONNULL ((3)) +error_tail (int status, int errnum, const char *message, va_list args) +{ +#if _LIBC + if (_IO_fwide (stderr, 0) > 0) + { + size_t len = strlen (message) + 1; + wchar_t *wmessage = NULL; + mbstate_t st; + size_t res; + const char *tmp; + bool use_malloc = false; + + while (1) + { + if (__libc_use_alloca (len * sizeof (wchar_t))) + wmessage = (wchar_t *) alloca (len * sizeof (wchar_t)); + else + { + if (!use_malloc) + wmessage = NULL; + + wchar_t *p = (wchar_t *) realloc (wmessage, + len * sizeof (wchar_t)); + if (p == NULL) + { + free (wmessage); + fputws_unlocked (L"out of memory\n", stderr); + return; + } + wmessage = p; + use_malloc = true; + } + + memset (&st, '\0', sizeof (st)); + tmp = message; + + res = mbsrtowcs (wmessage, &tmp, len, &st); + if (res != len) + break; + + if (__builtin_expect (len >= SIZE_MAX / sizeof (wchar_t) / 2, 0)) + { + /* This really should not happen if everything is fine. */ + res = (size_t) -1; + break; + } + + len *= 2; + } + + if (res == (size_t) -1) + { + /* The string cannot be converted. */ + if (use_malloc) + { + free (wmessage); + use_malloc = false; + } + wmessage = (wchar_t *) L"???"; + } + + __vfwprintf (stderr, wmessage, args); + + if (use_malloc) + free (wmessage); + } + else +#endif + vfprintf (stderr, message, args); + + ++error_message_count; + if (errnum) + print_errno_message (errnum); +#if _LIBC + __fxprintf (NULL, "\n"); +#else + putc ('\n', stderr); +#endif + fflush (stderr); + if (status) + exit (status); +} + + +/* Print the program name and error message MESSAGE, which is a printf-style + format string with optional args. + If ERRNUM is nonzero, print its corresponding system error message. + Exit with status STATUS if it is nonzero. */ +void +error (int status, int errnum, const char *message, ...) +{ + va_list args; + +#if defined _LIBC && defined __libc_ptf_call + /* We do not want this call to be cut short by a thread + cancellation. Therefore disable cancellation for now. */ + int state = PTHREAD_CANCEL_ENABLE; + __libc_ptf_call (pthread_setcancelstate, (PTHREAD_CANCEL_DISABLE, &state), + 0); +#endif + + flush_stdout (); +#ifdef _LIBC + _IO_flockfile (stderr); +#endif + if (error_print_progname) + (*error_print_progname) (); + else + { +#if _LIBC + __fxprintf (NULL, "%s: ", program_name); +#else + fprintf (stderr, "%s: ", program_name); +#endif + } + + va_start (args, message); + error_tail (status, errnum, message, args); + va_end (args); + +#ifdef _LIBC + _IO_funlockfile (stderr); +# ifdef __libc_ptf_call + __libc_ptf_call (pthread_setcancelstate, (state, NULL), 0); +# endif +#endif +} + +/* Sometimes we want to have at most one error per line. This + variable controls whether this mode is selected or not. */ +int error_one_per_line; + +void +error_at_line (int status, int errnum, const char *file_name, + unsigned int line_number, const char *message, ...) +{ + va_list args; + + if (error_one_per_line) + { + static const char *old_file_name; + static unsigned int old_line_number; + + if (old_line_number == line_number + && (file_name == old_file_name + || (old_file_name != NULL + && file_name != NULL + && strcmp (old_file_name, file_name) == 0))) + + /* Simply return and print nothing. */ + return; + + old_file_name = file_name; + old_line_number = line_number; + } + +#if defined _LIBC && defined __libc_ptf_call + /* We do not want this call to be cut short by a thread + cancellation. Therefore disable cancellation for now. */ + int state = PTHREAD_CANCEL_ENABLE; + __libc_ptf_call (pthread_setcancelstate, (PTHREAD_CANCEL_DISABLE, &state), + 0); +#endif + + flush_stdout (); +#ifdef _LIBC + _IO_flockfile (stderr); +#endif + if (error_print_progname) + (*error_print_progname) (); + else + { +#if _LIBC + __fxprintf (NULL, "%s:", program_name); +#else + fprintf (stderr, "%s:", program_name); +#endif + } + +#if _LIBC + __fxprintf (NULL, file_name != NULL ? "%s:%u: " : " ", + file_name, line_number); +#else + fprintf (stderr, file_name != NULL ? "%s:%u: " : " ", + file_name, line_number); +#endif + + va_start (args, message); + error_tail (status, errnum, message, args); + va_end (args); + +#ifdef _LIBC + _IO_funlockfile (stderr); +# ifdef __libc_ptf_call + __libc_ptf_call (pthread_setcancelstate, (state, NULL), 0); +# endif +#endif +} + +#ifdef _LIBC +/* Make the weak alias. */ +# undef error +# undef error_at_line +weak_alias (__error, error) +weak_alias (__error_at_line, error_at_line) +#endif diff --git a/lib/error.h b/lib/error.h new file mode 100644 index 0000000..36401e1 --- /dev/null +++ b/lib/error.h @@ -0,0 +1,75 @@ +/* Declaration for error-reporting function + Copyright (C) 1995-1997, 2003, 2006, 2008-2018 Free Software Foundation, + Inc. + This file is part of the GNU C Library. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +#ifndef _ERROR_H +#define _ERROR_H 1 + +/* The __attribute__ feature is available in gcc versions 2.5 and later. + The __-protected variants of the attributes 'format' and 'printf' are + accepted by gcc versions 2.6.4 (effectively 2.7) and later. + We enable _GL_ATTRIBUTE_FORMAT only if these are supported too, because + gnulib and libintl do '#define printf __printf__' when they override + the 'printf' function. */ +#if __GNUC__ > 2 || (__GNUC__ == 2 && __GNUC_MINOR__ >= 7) +# define _GL_ATTRIBUTE_FORMAT(spec) __attribute__ ((__format__ spec)) +#else +# define _GL_ATTRIBUTE_FORMAT(spec) /* empty */ +#endif + +/* On mingw, the flavor of printf depends on whether the extensions module + * is in use; the check for <stdio.h> determines the witness macro. */ +#ifndef _GL_ATTRIBUTE_SPEC_PRINTF +# if GNULIB_PRINTF_ATTRIBUTE_FLAVOR_GNU +# define _GL_ATTRIBUTE_SPEC_PRINTF __gnu_printf__ +# else +# define _GL_ATTRIBUTE_SPEC_PRINTF __printf__ +# endif +#endif + +#ifdef __cplusplus +extern "C" { +#endif + +/* Print a message with 'fprintf (stderr, FORMAT, ...)'; + if ERRNUM is nonzero, follow it with ": " and strerror (ERRNUM). + If STATUS is nonzero, terminate the program with 'exit (STATUS)'. */ + +extern void error (int __status, int __errnum, const char *__format, ...) + _GL_ATTRIBUTE_FORMAT ((_GL_ATTRIBUTE_SPEC_PRINTF, 3, 4)); + +extern void error_at_line (int __status, int __errnum, const char *__fname, + unsigned int __lineno, const char *__format, ...) + _GL_ATTRIBUTE_FORMAT ((_GL_ATTRIBUTE_SPEC_PRINTF, 5, 6)); + +/* If NULL, error will flush stdout, then print on stderr the program + name, a colon and a space. Otherwise, error will call this + function without parameters instead. */ +extern void (*error_print_progname) (void); + +/* This variable is incremented each time 'error' is called. */ +extern unsigned int error_message_count; + +/* Sometimes we want to have at most one error per line. This + variable controls whether this mode is selected or not. */ +extern int error_one_per_line; + +#ifdef __cplusplus +} +#endif + +#endif /* error.h */ diff --git a/lib/exitfail.c b/lib/exitfail.c new file mode 100644 index 0000000..19c5cba --- /dev/null +++ b/lib/exitfail.c @@ -0,0 +1,24 @@ +/* Failure exit status + + Copyright (C) 2002-2003, 2005-2007, 2009-2018 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +#include <config.h> + +#include "exitfail.h" + +#include <stdlib.h> + +int volatile exit_failure = EXIT_FAILURE; diff --git a/lib/exitfail.h b/lib/exitfail.h new file mode 100644 index 0000000..a42057b --- /dev/null +++ b/lib/exitfail.h @@ -0,0 +1,18 @@ +/* Failure exit status + + Copyright (C) 2002, 2009-2018 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +extern int volatile exit_failure; diff --git a/lib/fcntl.in.h b/lib/fcntl.in.h new file mode 100644 index 0000000..a1e7d35 --- /dev/null +++ b/lib/fcntl.in.h @@ -0,0 +1,366 @@ +/* Like <fcntl.h>, but with non-working flags defined to 0. + + Copyright (C) 2006-2018 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +/* written by Paul Eggert */ + +#if __GNUC__ >= 3 +@PRAGMA_SYSTEM_HEADER@ +#endif +@PRAGMA_COLUMNS@ + +#if defined __need_system_fcntl_h +/* Special invocation convention. */ + +/* Needed before <sys/stat.h>. + May also define off_t to a 64-bit type on native Windows. */ +#include <sys/types.h> +/* On some systems other than glibc, <sys/stat.h> is a prerequisite of + <fcntl.h>. On glibc systems, we would like to avoid namespace pollution. + But on glibc systems, <fcntl.h> includes <sys/stat.h> inside an + extern "C" { ... } block, which leads to errors in C++ mode with the + overridden <sys/stat.h> from gnulib. These errors are known to be gone + with g++ version >= 4.3. */ +#if !(defined __GLIBC__ || defined __UCLIBC__) || (defined __cplusplus && defined GNULIB_NAMESPACE && (defined __ICC || !(__GNUC__ > 4 || (__GNUC__ == 4 && __GNUC_MINOR__ >= 3)))) +# include <sys/stat.h> +#endif +#@INCLUDE_NEXT@ @NEXT_FCNTL_H@ + +#else +/* Normal invocation convention. */ + +#ifndef _@GUARD_PREFIX@_FCNTL_H + +/* Needed before <sys/stat.h>. + May also define off_t to a 64-bit type on native Windows. */ +#include <sys/types.h> +/* On some systems other than glibc, <sys/stat.h> is a prerequisite of + <fcntl.h>. On glibc systems, we would like to avoid namespace pollution. + But on glibc systems, <fcntl.h> includes <sys/stat.h> inside an + extern "C" { ... } block, which leads to errors in C++ mode with the + overridden <sys/stat.h> from gnulib. These errors are known to be gone + with g++ version >= 4.3. */ +#if !(defined __GLIBC__ || defined __UCLIBC__) || (defined __cplusplus && defined GNULIB_NAMESPACE && (defined __ICC || !(__GNUC__ > 4 || (__GNUC__ == 4 && __GNUC_MINOR__ >= 3)))) +# include <sys/stat.h> +#endif +/* The include_next requires a split double-inclusion guard. */ +#@INCLUDE_NEXT@ @NEXT_FCNTL_H@ + +#ifndef _@GUARD_PREFIX@_FCNTL_H +#define _@GUARD_PREFIX@_FCNTL_H + +#ifndef __GLIBC__ /* Avoid namespace pollution on glibc systems. */ +# include <unistd.h> +#endif + +/* Native Windows platforms declare open(), creat() in <io.h>. */ +#if (@GNULIB_OPEN@ || defined GNULIB_POSIXCHECK) \ + && (defined _WIN32 && ! defined __CYGWIN__) +# include <io.h> +#endif + + +/* The definitions of _GL_FUNCDECL_RPL etc. are copied here. */ + +/* The definition of _GL_ARG_NONNULL is copied here. */ + +/* The definition of _GL_WARN_ON_USE is copied here. */ + + +/* Declare overridden functions. */ + +#if @GNULIB_FCNTL@ +# if @REPLACE_FCNTL@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef fcntl +# define fcntl rpl_fcntl +# endif +_GL_FUNCDECL_RPL (fcntl, int, (int fd, int action, ...)); +_GL_CXXALIAS_RPL (fcntl, int, (int fd, int action, ...)); +# else +# if !@HAVE_FCNTL@ +_GL_FUNCDECL_SYS (fcntl, int, (int fd, int action, ...)); +# endif +_GL_CXXALIAS_SYS (fcntl, int, (int fd, int action, ...)); +# endif +_GL_CXXALIASWARN (fcntl); +#elif defined GNULIB_POSIXCHECK +# undef fcntl +# if HAVE_RAW_DECL_FCNTL +_GL_WARN_ON_USE (fcntl, "fcntl is not always POSIX compliant - " + "use gnulib module fcntl for portability"); +# endif +#endif + +#if @GNULIB_OPEN@ +# if @REPLACE_OPEN@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef open +# define open rpl_open +# endif +_GL_FUNCDECL_RPL (open, int, (const char *filename, int flags, ...) + _GL_ARG_NONNULL ((1))); +_GL_CXXALIAS_RPL (open, int, (const char *filename, int flags, ...)); +# else +_GL_CXXALIAS_SYS (open, int, (const char *filename, int flags, ...)); +# endif +/* On HP-UX 11, in C++ mode, open() is defined as an inline function with a + default argument. _GL_CXXALIASWARN does not work in this case. */ +# if !defined __hpux +_GL_CXXALIASWARN (open); +# endif +#elif defined GNULIB_POSIXCHECK +# undef open +/* Assume open is always declared. */ +_GL_WARN_ON_USE (open, "open is not always POSIX compliant - " + "use gnulib module open for portability"); +#endif + +#if @GNULIB_OPENAT@ +# if @REPLACE_OPENAT@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef openat +# define openat rpl_openat +# endif +_GL_FUNCDECL_RPL (openat, int, + (int fd, char const *file, int flags, /* mode_t mode */ ...) + _GL_ARG_NONNULL ((2))); +_GL_CXXALIAS_RPL (openat, int, + (int fd, char const *file, int flags, /* mode_t mode */ ...)); +# else +# if !@HAVE_OPENAT@ +_GL_FUNCDECL_SYS (openat, int, + (int fd, char const *file, int flags, /* mode_t mode */ ...) + _GL_ARG_NONNULL ((2))); +# endif +_GL_CXXALIAS_SYS (openat, int, + (int fd, char const *file, int flags, /* mode_t mode */ ...)); +# endif +_GL_CXXALIASWARN (openat); +#elif defined GNULIB_POSIXCHECK +# undef openat +# if HAVE_RAW_DECL_OPENAT +_GL_WARN_ON_USE (openat, "openat is not portable - " + "use gnulib module openat for portability"); +# endif +#endif + + +/* Fix up the FD_* macros, only known to be missing on mingw. */ + +#ifndef FD_CLOEXEC +# define FD_CLOEXEC 1 +#endif + +/* Fix up the supported F_* macros. Intentionally leave other F_* + macros undefined. Only known to be missing on mingw. */ + +#ifndef F_DUPFD_CLOEXEC +# define F_DUPFD_CLOEXEC 0x40000000 +/* Witness variable: 1 if gnulib defined F_DUPFD_CLOEXEC, 0 otherwise. */ +# define GNULIB_defined_F_DUPFD_CLOEXEC 1 +#else +# define GNULIB_defined_F_DUPFD_CLOEXEC 0 +#endif + +#ifndef F_DUPFD +# define F_DUPFD 1 +#endif + +#ifndef F_GETFD +# define F_GETFD 2 +#endif + +/* Fix up the O_* macros. */ + +/* AIX 7.1 with XL C 12.1 defines O_CLOEXEC, O_NOFOLLOW, and O_TTY_INIT + to values outside 'int' range, so omit these misdefinitions. + But avoid namespace pollution on non-AIX systems. */ +#ifdef _AIX +# include <limits.h> +# if defined O_CLOEXEC && ! (INT_MIN <= O_CLOEXEC && O_CLOEXEC <= INT_MAX) +# undef O_CLOEXEC +# endif +# if defined O_NOFOLLOW && ! (INT_MIN <= O_NOFOLLOW && O_NOFOLLOW <= INT_MAX) +# undef O_NOFOLLOW +# endif +# if defined O_TTY_INIT && ! (INT_MIN <= O_TTY_INIT && O_TTY_INIT <= INT_MAX) +# undef O_TTY_INIT +# endif +#endif + +#if !defined O_DIRECT && defined O_DIRECTIO +/* Tru64 spells it 'O_DIRECTIO'. */ +# define O_DIRECT O_DIRECTIO +#endif + +#if !defined O_CLOEXEC && defined O_NOINHERIT +/* Mingw spells it 'O_NOINHERIT'. */ +# define O_CLOEXEC O_NOINHERIT +#endif + +#ifndef O_CLOEXEC +# define O_CLOEXEC 0x40000000 /* Try to not collide with system O_* flags. */ +# define GNULIB_defined_O_CLOEXEC 1 +#else +# define GNULIB_defined_O_CLOEXEC 0 +#endif + +#ifndef O_DIRECT +# define O_DIRECT 0 +#endif + +#ifndef O_DIRECTORY +# define O_DIRECTORY 0 +#endif + +#ifndef O_DSYNC +# define O_DSYNC 0 +#endif + +#ifndef O_EXEC +# define O_EXEC O_RDONLY /* This is often close enough in older systems. */ +#endif + +#ifndef O_IGNORE_CTTY +# define O_IGNORE_CTTY 0 +#endif + +#ifndef O_NDELAY +# define O_NDELAY 0 +#endif + +#ifndef O_NOATIME +# define O_NOATIME 0 +#endif + +#ifndef O_NONBLOCK +# define O_NONBLOCK O_NDELAY +#endif + +/* If the gnulib module 'nonblocking' is in use, guarantee a working non-zero + value of O_NONBLOCK. Otherwise, O_NONBLOCK is defined (above) to O_NDELAY + or to 0 as fallback. */ +#if @GNULIB_NONBLOCKING@ +# if O_NONBLOCK +# define GNULIB_defined_O_NONBLOCK 0 +# else +# define GNULIB_defined_O_NONBLOCK 1 +# undef O_NONBLOCK +# define O_NONBLOCK 0x40000000 +# endif +#endif + +#ifndef O_NOCTTY +# define O_NOCTTY 0 +#endif + +#ifndef O_NOFOLLOW +# define O_NOFOLLOW 0 +#endif + +#ifndef O_NOLINK +# define O_NOLINK 0 +#endif + +#ifndef O_NOLINKS +# define O_NOLINKS 0 +#endif + +#ifndef O_NOTRANS +# define O_NOTRANS 0 +#endif + +#ifndef O_RSYNC +# define O_RSYNC 0 +#endif + +#ifndef O_SEARCH +# define O_SEARCH O_RDONLY /* This is often close enough in older systems. */ +#endif + +#ifndef O_SYNC +# define O_SYNC 0 +#endif + +#ifndef O_TTY_INIT +# define O_TTY_INIT 0 +#endif + +#if ~O_ACCMODE & (O_RDONLY | O_WRONLY | O_RDWR | O_EXEC | O_SEARCH) +# undef O_ACCMODE +# define O_ACCMODE (O_RDONLY | O_WRONLY | O_RDWR | O_EXEC | O_SEARCH) +#endif + +/* For systems that distinguish between text and binary I/O. + O_BINARY is usually declared in fcntl.h */ +#if !defined O_BINARY && defined _O_BINARY + /* For MSC-compatible compilers. */ +# define O_BINARY _O_BINARY +# define O_TEXT _O_TEXT +#endif + +#if defined __BEOS__ || defined __HAIKU__ + /* BeOS 5 and Haiku have O_BINARY and O_TEXT, but they have no effect. */ +# undef O_BINARY +# undef O_TEXT +#endif + +#ifndef O_BINARY +# define O_BINARY 0 +# define O_TEXT 0 +#endif + +/* Fix up the AT_* macros. */ + +/* Work around a bug in Solaris 9 and 10: AT_FDCWD is positive. Its + value exceeds INT_MAX, so its use as an int doesn't conform to the + C standard, and GCC and Sun C complain in some cases. If the bug + is present, undef AT_FDCWD here, so it can be redefined below. */ +#if 0 < AT_FDCWD && AT_FDCWD == 0xffd19553 +# undef AT_FDCWD +#endif + +/* Use the same bit pattern as Solaris 9, but with the proper + signedness. The bit pattern is important, in case this actually is + Solaris with the above workaround. */ +#ifndef AT_FDCWD +# define AT_FDCWD (-3041965) +#endif + +/* Use the same values as Solaris 9. This shouldn't matter, but + there's no real reason to differ. */ +#ifndef AT_SYMLINK_NOFOLLOW +# define AT_SYMLINK_NOFOLLOW 4096 +#endif + +#ifndef AT_REMOVEDIR +# define AT_REMOVEDIR 1 +#endif + +/* Solaris 9 lacks these two, so just pick unique values. */ +#ifndef AT_SYMLINK_FOLLOW +# define AT_SYMLINK_FOLLOW 2 +#endif + +#ifndef AT_EACCESS +# define AT_EACCESS 4 +#endif + + +#endif /* _@GUARD_PREFIX@_FCNTL_H */ +#endif /* _@GUARD_PREFIX@_FCNTL_H */ +#endif diff --git a/lib/filename.h b/lib/filename.h new file mode 100644 index 0000000..a7ad7e7 --- /dev/null +++ b/lib/filename.h @@ -0,0 +1,54 @@ +/* Basic filename support macros. + Copyright (C) 2001-2004, 2007-2018 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +#ifndef _FILENAME_H +#define _FILENAME_H + +#ifdef __cplusplus +extern "C" { +#endif + + +/* Pathname support. + ISSLASH(C) tests whether C is a directory separator character. + IS_ABSOLUTE_PATH(P) tests whether P is an absolute path. If it is not, + it may be concatenated to a directory pathname. + IS_PATH_WITH_DIR(P) tests whether P contains a directory specification. + */ +#if defined _WIN32 || defined __CYGWIN__ || defined __EMX__ || defined __DJGPP__ + /* Native Windows, Cygwin, OS/2, DOS */ +# define ISSLASH(C) ((C) == '/' || (C) == '\\') +# define HAS_DEVICE(P) \ + ((((P)[0] >= 'A' && (P)[0] <= 'Z') || ((P)[0] >= 'a' && (P)[0] <= 'z')) \ + && (P)[1] == ':') +# define IS_ABSOLUTE_PATH(P) (ISSLASH ((P)[0]) || HAS_DEVICE (P)) +# define IS_PATH_WITH_DIR(P) \ + (strchr (P, '/') != NULL || strchr (P, '\\') != NULL || HAS_DEVICE (P)) +# define FILE_SYSTEM_PREFIX_LEN(P) (HAS_DEVICE (P) ? 2 : 0) +#else + /* Unix */ +# define ISSLASH(C) ((C) == '/') +# define IS_ABSOLUTE_PATH(P) ISSLASH ((P)[0]) +# define IS_PATH_WITH_DIR(P) (strchr (P, '/') != NULL) +# define FILE_SYSTEM_PREFIX_LEN(P) 0 +#endif + + +#ifdef __cplusplus +} +#endif + +#endif /* _FILENAME_H */ diff --git a/lib/fpending.c b/lib/fpending.c new file mode 100644 index 0000000..de370d4 --- /dev/null +++ b/lib/fpending.c @@ -0,0 +1,62 @@ +/* fpending.c -- return the number of pending output bytes on a stream + Copyright (C) 2000, 2004, 2006-2007, 2009-2018 Free Software Foundation, + Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +/* Written by Jim Meyering. */ + +#include <config.h> + +/* Specification. */ +#include "fpending.h" + +#include "stdio-impl.h" + +/* This file is not used on systems that already have the __fpending function, + namely glibc >= 2.2, Solaris >= 7, Android API >= 23. */ + +/* Return the number of pending (aka buffered, unflushed) + bytes on the stream, FP, that is open for writing. */ +size_t +__fpending (FILE *fp) +{ + /* Most systems provide FILE as a struct and the necessary bitmask in + <stdio.h>, because they need it for implementing getc() and putc() as + fast macros. */ +#if defined _IO_EOF_SEEN || defined _IO_ftrylockfile || __GNU_LIBRARY__ == 1 + /* GNU libc, BeOS, Haiku, Linux libc5 */ + return fp->_IO_write_ptr - fp->_IO_write_base; +#elif defined __sferror || defined __DragonFly__ || defined __ANDROID__ + /* FreeBSD, NetBSD, OpenBSD, DragonFly, Mac OS X, Cygwin, Minix 3, Android */ + return fp->_p - fp->_bf._base; +#elif defined __EMX__ /* emx+gcc */ + return fp->_ptr - fp->_buffer; +#elif defined __minix /* Minix */ + return fp_->_ptr - fp_->_buf; +#elif defined _IOERR /* AIX, HP-UX, IRIX, OSF/1, Solaris, OpenServer, mingw, MSVC, NonStop Kernel, OpenVMS */ + return (fp_->_ptr ? fp_->_ptr - fp_->_base : 0); +#elif defined __UCLIBC__ /* uClibc */ + return (fp->__modeflags & __FLAG_WRITING ? fp->__bufpos - fp->__bufstart : 0); +#elif defined __QNX__ /* QNX */ + return (fp->_Mode & 0x2000 /*_MWRITE*/ ? fp->_Next - fp->_Buf : 0); +#elif defined __MINT__ /* Atari FreeMiNT */ + return fp->__bufp - fp->__buffer; +#elif defined EPLAN9 /* Plan9 */ + return fp->wp - fp->buf; +#else +# error "Please port gnulib fpending.c to your platform!" + return 1; +#endif +} diff --git a/lib/fpending.h b/lib/fpending.h new file mode 100644 index 0000000..2ff26bc --- /dev/null +++ b/lib/fpending.h @@ -0,0 +1,29 @@ +/* Declare __fpending. + + Copyright (C) 2000, 2003, 2005-2006, 2009-2018 Free Software Foundation, + Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. + + Written by Jim Meyering. */ + +#include <stddef.h> +#include <stdio.h> +#if HAVE_STDIO_EXT_H +# include <stdio_ext.h> +#endif + +#if !HAVE_DECL___FPENDING +size_t __fpending (FILE *) _GL_ATTRIBUTE_PURE; +#endif diff --git a/lib/fstat.c b/lib/fstat.c new file mode 100644 index 0000000..41ee761 --- /dev/null +++ b/lib/fstat.c @@ -0,0 +1,90 @@ +/* fstat() replacement. + Copyright (C) 2011-2018 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +/* If the user's config.h happens to include <sys/stat.h>, let it include only + the system's <sys/stat.h> here, so that orig_fstat doesn't recurse to + rpl_fstat. */ +#define __need_system_sys_stat_h +#include <config.h> + +/* Get the original definition of fstat. It might be defined as a macro. */ +#include <sys/types.h> +#include <sys/stat.h> +#undef __need_system_sys_stat_h + +#if defined _WIN32 && ! defined __CYGWIN__ +# define WINDOWS_NATIVE +#endif + +#if !defined WINDOWS_NATIVE + +static int +orig_fstat (int fd, struct stat *buf) +{ + return fstat (fd, buf); +} + +#endif + +/* Specification. */ +/* Write "sys/stat.h" here, not <sys/stat.h>, otherwise OSF/1 5.1 DTK cc + eliminates this include because of the preliminary #include <sys/stat.h> + above. */ +#include "sys/stat.h" + +#include "stat-time.h" + +#include <errno.h> +#include <unistd.h> +#ifdef WINDOWS_NATIVE +# define WIN32_LEAN_AND_MEAN +# include <windows.h> +# if GNULIB_MSVC_NOTHROW +# include "msvc-nothrow.h" +# else +# include <io.h> +# endif +# include "stat-w32.h" +#endif + +int +rpl_fstat (int fd, struct stat *buf) +{ +#if REPLACE_FCHDIR && REPLACE_OPEN_DIRECTORY + /* Handle the case when rpl_open() used a dummy file descriptor to work + around an open() that can't normally visit directories. */ + const char *name = _gl_directory_name (fd); + if (name != NULL) + return stat (name, buf); +#endif + +#ifdef WINDOWS_NATIVE + /* Fill the fields ourselves, because the original fstat function returns + values for st_atime, st_mtime, st_ctime that depend on the current time + zone. See + <https://lists.gnu.org/r/bug-gnulib/2017-04/msg00134.html> */ + HANDLE h = (HANDLE) _get_osfhandle (fd); + + if (h == INVALID_HANDLE_VALUE) + { + errno = EBADF; + return -1; + } + return _gl_fstat_by_handle (h, NULL, buf); +#else + return stat_time_normalize (orig_fstat (fd, buf), buf); +#endif +} diff --git a/lib/fwriting.c b/lib/fwriting.c new file mode 100644 index 0000000..4ffa12d --- /dev/null +++ b/lib/fwriting.c @@ -0,0 +1,67 @@ +/* Retrieve information about a FILE stream. + Copyright (C) 2007-2018 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +#include <config.h> + +/* Specification. */ +#include "fwriting.h" + +#include "stdio-impl.h" + +/* This file is not used on systems that have the __fwritable function, + namely glibc >= 2.2, Solaris >= 7, Android API >= 29, musl libc. */ + +bool +fwriting (FILE *fp) +{ + /* Most systems provide FILE as a struct and the necessary bitmask in + <stdio.h>, because they need it for implementing getc() and putc() as + fast macros. */ +#if defined _IO_EOF_SEEN || defined _IO_ftrylockfile || __GNU_LIBRARY__ == 1 + /* GNU libc, BeOS, Haiku, Linux libc5 */ + return (fp->_flags & (_IO_NO_READS | _IO_CURRENTLY_PUTTING)) != 0; +#elif defined __sferror || defined __DragonFly__ || defined __ANDROID__ + /* FreeBSD, NetBSD, OpenBSD, DragonFly, Mac OS X, Cygwin, Minix 3, Android */ + return (fp_->_flags & __SWR) != 0; +#elif defined __EMX__ /* emx+gcc */ + return (fp->_flags & _IOWRT) != 0; +#elif defined __minix /* Minix */ + return (fp->_flags & _IOWRITING) != 0; +#elif defined _IOERR /* AIX, HP-UX, IRIX, OSF/1, Solaris, OpenServer, mingw, MSVC, NonStop Kernel, OpenVMS */ + return (fp_->_flag & _IOWRT) != 0; +#elif defined __UCLIBC__ /* uClibc */ + return (fp->__modeflags & __FLAG_WRITING) != 0; +#elif defined __QNX__ /* QNX */ + return ((fp->_Mode & 0x1 /* _MOPENR */) == 0 + || (fp->_Mode & 0x2000 /* _MWRITE */) != 0); +#elif defined __MINT__ /* Atari FreeMiNT */ + if (!fp->__mode.__read) + return 1; + if (!fp->__mode.__write) + return 0; +# ifdef _IO_CURRENTLY_PUTTING /* Flag added on 2009-02-28 */ + return (fp->__flags & _IO_CURRENTLY_PUTTING) != 0; +# else + return (fp->__buffer < fp->__put_limit /*|| fp->__bufp == fp->__get_limit ??*/); +# endif +#elif defined EPLAN9 /* Plan9 */ + if (fp->state == 0 /* CLOSED */ || fp->state == 3 /* RD */) + return 0; + return (fp->state == 4 /* WR */ && (fp->bufl == 0 || fp->wp < fp->rp)); +#else +# error "Please port gnulib fwriting.c to your platform!" +#endif +} diff --git a/lib/fwriting.h b/lib/fwriting.h new file mode 100644 index 0000000..42f7e94 --- /dev/null +++ b/lib/fwriting.h @@ -0,0 +1,53 @@ +/* Retrieve information about a FILE stream. + Copyright (C) 2007, 2009-2018 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +#include <stdbool.h> +#include <stdio.h> + +/* Return true if the stream STREAM is opened write-only or + append-only, or if the last operation on the stream was a write + operation. Return false if the stream is opened read-only, or if + it supports reading and there is no current write operation (such + as fputc). + + freading and fwriting will never both be true. If STREAM supports + both reads and writes, then: + - both freading and fwriting might be false when the stream is first + opened, after read encounters EOF, or after fflush, + - freading might be false or true and fwriting might be false + after repositioning (such as fseek, fsetpos, or rewind), + depending on the underlying implementation. + + STREAM must not be wide-character oriented. */ + +#if HAVE___FWRITING /* glibc >= 2.2, Solaris >= 7, Android API >= 29, musl libc */ + +# include <stdio_ext.h> +# define fwriting(stream) (__fwriting (stream) != 0) + +#else + +# ifdef __cplusplus +extern "C" { +# endif + +extern bool fwriting (FILE *stream) _GL_ATTRIBUTE_PURE; + +# ifdef __cplusplus +} +# endif + +#endif diff --git a/lib/get-permissions.c b/lib/get-permissions.c new file mode 100644 index 0000000..3b98451 --- /dev/null +++ b/lib/get-permissions.c @@ -0,0 +1,291 @@ +/* Get permissions of a file. -*- coding: utf-8 -*- + + Copyright (C) 2002-2003, 2005-2018 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. + + Written by Paul Eggert, Andreas Grünbacher, and Bruno Haible. */ + +#include <config.h> + +#include <string.h> +#include "acl.h" + +#include "acl-internal.h" + +/* Read the permissions of a file into CTX. If DESC is a valid file descriptor, + use file descriptor operations, else use filename based operations on NAME. + MODE is the file mode obtained from a previous stat call. + Return 0 if successful. Return -1 and set errno upon failure. */ + +int +get_permissions (const char *name, int desc, mode_t mode, + struct permission_context *ctx) +{ + memset (ctx, 0, sizeof *ctx); + ctx->mode = mode; + +#if USE_ACL && HAVE_ACL_GET_FILE + /* POSIX 1003.1e (draft 17 -- abandoned) specific version. */ + /* Linux, FreeBSD, Mac OS X, IRIX, Tru64, Cygwin >= 2.5 */ +# if !HAVE_ACL_TYPE_EXTENDED + /* Linux, FreeBSD, IRIX, Tru64, Cygwin >= 2.5 */ + + if (HAVE_ACL_GET_FD && desc != -1) + ctx->acl = acl_get_fd (desc); + else + ctx->acl = acl_get_file (name, ACL_TYPE_ACCESS); + if (ctx->acl == NULL) + return acl_errno_valid (errno) ? -1 : 0; + + /* With POSIX ACLs, a file cannot have "no" acl; a file without + extended permissions has a "minimal" acl which is equivalent to the + file mode. */ + + if (S_ISDIR (mode)) + { + ctx->default_acl = acl_get_file (name, ACL_TYPE_DEFAULT); + if (ctx->default_acl == NULL) + return -1; + } + +# if HAVE_ACL_TYPE_NFS4 /* FreeBSD */ + + /* TODO (see set_permissions). */ + +# endif + +# else /* HAVE_ACL_TYPE_EXTENDED */ + /* Mac OS X */ + + /* On Mac OS X, acl_get_file (name, ACL_TYPE_ACCESS) + and acl_get_file (name, ACL_TYPE_DEFAULT) + always return NULL / EINVAL. You have to use + acl_get_file (name, ACL_TYPE_EXTENDED) + or acl_get_fd (open (name, ...)) + to retrieve an ACL. + On the other hand, + acl_set_file (name, ACL_TYPE_ACCESS, acl) + and acl_set_file (name, ACL_TYPE_DEFAULT, acl) + have the same effect as + acl_set_file (name, ACL_TYPE_EXTENDED, acl): + Each of these calls sets the file's ACL. */ + + if (HAVE_ACL_GET_FD && desc != -1) + ctx->acl = acl_get_fd (desc); + else + ctx->acl = acl_get_file (name, ACL_TYPE_EXTENDED); + if (ctx->acl == NULL) + return acl_errno_valid (errno) ? -1 : 0; + +# endif + +#elif USE_ACL && defined GETACL /* Solaris, Cygwin, not HP-UX */ + + /* Solaris 2.5 through Solaris 10, Cygwin, and contemporaneous versions + of Unixware. The acl() call returns the access and default ACL both + at once. */ +# ifdef ACE_GETACL + /* Solaris also has a different variant of ACLs, used in ZFS and NFSv4 + file systems (whereas the other ones are used in UFS file systems). + There is an API + pathconf (name, _PC_ACL_ENABLED) + fpathconf (desc, _PC_ACL_ENABLED) + that allows us to determine which of the two kinds of ACLs is supported + for the given file. But some file systems may implement this call + incorrectly, so better not use it. + When fetching the source ACL, we simply fetch both ACL types. + When setting the destination ACL, we try either ACL types, assuming + that the kernel will translate the ACL from one form to the other. + (See in <http://docs.sun.com/app/docs/doc/819-2241/6n4huc7ia?l=en&a=view> + the description of ENOTSUP.) */ + for (;;) + { + int ret; + + if (desc != -1) + ret = facl (desc, ACE_GETACLCNT, 0, NULL); + else + ret = acl (name, ACE_GETACLCNT, 0, NULL); + if (ret < 0) + { + if (errno == ENOSYS || errno == EINVAL) + ret = 0; + else + return -1; + } + ctx->ace_count = ret; + + if (ctx->ace_count == 0) + break; + + ctx->ace_entries = (ace_t *) malloc (ctx->ace_count * sizeof (ace_t)); + if (ctx->ace_entries == NULL) + { + errno = ENOMEM; + return -1; + } + + if (desc != -1) + ret = facl (desc, ACE_GETACL, ctx->ace_count, ctx->ace_entries); + else + ret = acl (name, ACE_GETACL, ctx->ace_count, ctx->ace_entries); + if (ret < 0) + { + if (errno == ENOSYS || errno == EINVAL) + { + free (ctx->ace_entries); + ctx->ace_entries = NULL; + ctx->ace_count = 0; + break; + } + else + return -1; + } + if (ret <= ctx->ace_count) + { + ctx->ace_count = ret; + break; + } + /* Huh? The number of ACL entries has increased since the last call. + Repeat. */ + free (ctx->ace_entries); + ctx->ace_entries = NULL; + } +# endif + + for (;;) + { + int ret; + + if (desc != -1) + ret = facl (desc, GETACLCNT, 0, NULL); + else + ret = acl (name, GETACLCNT, 0, NULL); + if (ret < 0) + { + if (errno == ENOSYS || errno == ENOTSUP || errno == EOPNOTSUPP) + ret = 0; + else + return -1; + } + ctx->count = ret; + + if (ctx->count == 0) + break; + + ctx->entries = (aclent_t *) malloc (ctx->count * sizeof (aclent_t)); + if (ctx->entries == NULL) + { + errno = ENOMEM; + return -1; + } + + if (desc != -1) + ret = facl (desc, GETACL, ctx->count, ctx->entries); + else + ret = acl (name, GETACL, ctx->count, ctx->entries); + if (ret < 0) + { + if (errno == ENOSYS || errno == ENOTSUP || errno == EOPNOTSUPP) + { + free (ctx->entries); + ctx->entries = NULL; + ctx->count = 0; + break; + } + else + return -1; + } + if (ret <= ctx->count) + { + ctx->count = ret; + break; + } + /* Huh? The number of ACL entries has increased since the last call. + Repeat. */ + free (ctx->entries); + ctx->entries = NULL; + } + +#elif USE_ACL && HAVE_GETACL /* HP-UX */ + + { + int ret; + + if (desc != -1) + ret = fgetacl (desc, NACLENTRIES, ctx->entries); + else + ret = getacl (name, NACLENTRIES, ctx->entries); + if (ret < 0) + { + if (errno == ENOSYS || errno == EOPNOTSUPP || errno == ENOTSUP) + ret = 0; + else + return -1; + } + else if (ret > NACLENTRIES) + /* If NACLENTRIES cannot be trusted, use dynamic memory allocation. */ + abort (); + ctx->count = ret; + +# if HAVE_ACLV_H + ret = acl ((char *) name, ACL_GET, NACLVENTRIES, ctx->aclv_entries); + if (ret < 0) + { + if (errno == ENOSYS || errno == EOPNOTSUPP || errno == EINVAL) + ret = 0; + else + return -2; + } + else if (ret > NACLVENTRIES) + /* If NACLVENTRIES cannot be trusted, use dynamic memory allocation. */ + abort (); + ctx->aclv_count = ret; +# endif + } + +#elif USE_ACL && HAVE_ACLX_GET && ACL_AIX_WIP /* AIX */ + + /* TODO (see set_permissions). */ + +#elif USE_ACL && HAVE_STATACL /* older AIX */ + + { + int ret; + if (desc != -1) + ret = fstatacl (desc, STX_NORMAL, &ctx->u.a, sizeof ctx->u); + else + ret = statacl ((char *) name, STX_NORMAL, &ctx->u.a, sizeof ctx->u); + if (ret == 0) + ctx->have_u = true; + } + +#elif USE_ACL && HAVE_ACLSORT /* NonStop Kernel */ + + { + int ret = acl ((char *) name, ACL_GET, NACLENTRIES, ctx->entries); + if (ret < 0) + return -1; + else if (ret > NACLENTRIES) + /* If NACLENTRIES cannot be trusted, use dynamic memory allocation. */ + abort (); + ctx->count = ret; + } + +#endif + + return 0; + +} diff --git a/lib/getdelim.c b/lib/getdelim.c new file mode 100644 index 0000000..cc1e9fe --- /dev/null +++ b/lib/getdelim.c @@ -0,0 +1,147 @@ +/* getdelim.c --- Implementation of replacement getdelim function. + Copyright (C) 1994, 1996-1998, 2001, 2003, 2005-2018 Free Software + Foundation, Inc. + + This program is free software; you can redistribute it and/or + modify it under the terms of the GNU General Public License as + published by the Free Software Foundation; either version 3, or (at + your option) any later version. + + This program is distributed in the hope that it will be useful, but + WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, see <https://www.gnu.org/licenses/>. */ + +/* Ported from glibc by Simon Josefsson. */ + +/* Don't use __attribute__ __nonnull__ in this compilation unit. Otherwise gcc + optimizes away the lineptr == NULL || n == NULL || fp == NULL tests below. */ +#define _GL_ARG_NONNULL(params) + +#include <config.h> + +#include <stdio.h> + +#include <limits.h> +#include <stdint.h> +#include <stdlib.h> +#include <errno.h> + +#ifndef SSIZE_MAX +# define SSIZE_MAX ((ssize_t) (SIZE_MAX / 2)) +#endif + +#if USE_UNLOCKED_IO +# include "unlocked-io.h" +# define getc_maybe_unlocked(fp) getc(fp) +#elif !HAVE_FLOCKFILE || !HAVE_FUNLOCKFILE || !HAVE_DECL_GETC_UNLOCKED +# undef flockfile +# undef funlockfile +# define flockfile(x) ((void) 0) +# define funlockfile(x) ((void) 0) +# define getc_maybe_unlocked(fp) getc(fp) +#else +# define getc_maybe_unlocked(fp) getc_unlocked(fp) +#endif + +static void +alloc_failed (void) +{ +#if defined _WIN32 && ! defined __CYGWIN__ + /* Avoid errno problem without using the realloc module; see: + https://lists.gnu.org/r/bug-gnulib/2016-08/msg00025.html */ + errno = ENOMEM; +#endif +} + +/* Read up to (and including) a DELIMITER from FP into *LINEPTR (and + NUL-terminate it). *LINEPTR is a pointer returned from malloc (or + NULL), pointing to *N characters of space. It is realloc'ed as + necessary. Returns the number of characters read (not including + the null terminator), or -1 on error or EOF. */ + +ssize_t +getdelim (char **lineptr, size_t *n, int delimiter, FILE *fp) +{ + ssize_t result; + size_t cur_len = 0; + + if (lineptr == NULL || n == NULL || fp == NULL) + { + errno = EINVAL; + return -1; + } + + flockfile (fp); + + if (*lineptr == NULL || *n == 0) + { + char *new_lineptr; + *n = 120; + new_lineptr = (char *) realloc (*lineptr, *n); + if (new_lineptr == NULL) + { + alloc_failed (); + result = -1; + goto unlock_return; + } + *lineptr = new_lineptr; + } + + for (;;) + { + int i; + + i = getc_maybe_unlocked (fp); + if (i == EOF) + { + result = -1; + break; + } + + /* Make enough space for len+1 (for final NUL) bytes. */ + if (cur_len + 1 >= *n) + { + size_t needed_max = + SSIZE_MAX < SIZE_MAX ? (size_t) SSIZE_MAX + 1 : SIZE_MAX; + size_t needed = 2 * *n + 1; /* Be generous. */ + char *new_lineptr; + + if (needed_max < needed) + needed = needed_max; + if (cur_len + 1 >= needed) + { + result = -1; + errno = EOVERFLOW; + goto unlock_return; + } + + new_lineptr = (char *) realloc (*lineptr, needed); + if (new_lineptr == NULL) + { + alloc_failed (); + result = -1; + goto unlock_return; + } + + *lineptr = new_lineptr; + *n = needed; + } + + (*lineptr)[cur_len] = i; + cur_len++; + + if (i == delimiter) + break; + } + (*lineptr)[cur_len] = '\0'; + result = cur_len ? cur_len : result; + + unlock_return: + funlockfile (fp); /* doesn't set errno */ + + return result; +} diff --git a/lib/getfilecon.c b/lib/getfilecon.c new file mode 100644 index 0000000..3bf254e --- /dev/null +++ b/lib/getfilecon.c @@ -0,0 +1,87 @@ +/* wrap getfilecon, lgetfilecon, and fgetfilecon + Copyright (C) 2009-2018 Free Software Foundation, Inc. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3, or (at your option) + any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, see <https://www.gnu.org/licenses/>. */ + +/* written by Jim Meyering */ + +#include <config.h> + +#include <selinux/selinux.h> + +#include <sys/types.h> +#include <errno.h> +#include <string.h> + +/* FIXME: remove this once there is an errno-gnu module + that guarantees the definition of ENODATA. */ +#ifndef ENODATA +# define ENODATA ENOTSUP +#endif + +#undef getfilecon +#undef lgetfilecon +#undef fgetfilecon +int getfilecon (char const *file, security_context_t *con); +int lgetfilecon (char const *file, security_context_t *con); +int fgetfilecon (int fd, security_context_t *con); + +/* getfilecon, lgetfilecon, and fgetfilecon can all misbehave, be it + via an old version of libselinux where these would return 0 and set the + result context to NULL, or via a modern kernel+lib operating on a file + from a disk whose attributes were set by a kernel from around 2006. + In that latter case, the functions return a length of 10 for the + "unlabeled" context. Map both failures to a return value of -1, and + set errno to ENOTSUP in the first case, and ENODATA in the latter. */ + +static int +map_to_failure (int ret, security_context_t *con) +{ + if (ret == 0) + { + errno = ENOTSUP; + return -1; + } + + if (ret == 10 && strcmp (*con, "unlabeled") == 0) + { + freecon (*con); + *con = NULL; + errno = ENODATA; + return -1; + } + + return ret; +} + +int +rpl_getfilecon (char const *file, security_context_t *con) +{ + int ret = getfilecon (file, con); + return map_to_failure (ret, con); +} + +int +rpl_lgetfilecon (char const *file, security_context_t *con) +{ + int ret = lgetfilecon (file, con); + return map_to_failure (ret, con); +} + +int +rpl_fgetfilecon (int fd, security_context_t *con) +{ + int ret = fgetfilecon (fd, con); + return map_to_failure (ret, con); +} diff --git a/lib/getopt-cdefs.in.h b/lib/getopt-cdefs.in.h new file mode 100644 index 0000000..12b5a87 --- /dev/null +++ b/lib/getopt-cdefs.in.h @@ -0,0 +1,67 @@ +/* getopt-on-non-glibc compatibility macros. + Copyright (C) 1989-2018 Free Software Foundation, Inc. + This file is part of gnulib. + Unlike most of the getopt implementation, it is NOT shared + with the GNU C Library. + + This file is free software; you can redistribute it and/or modify it + under the terms of the GNU General Public License as + published by the Free Software Foundation; either version 3 of + the License, or (at your option) any later version. + + This file is distributed in the hope that it will be useful, but + WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + General Public License for more details. + + You should have received a copy of the GNU General Public + License along with gnulib; if not, see + <https://www.gnu.org/licenses/>. */ + +#ifndef _GETOPT_CDEFS_H +#define _GETOPT_CDEFS_H 1 + +/* This header should not be used directly; include getopt.h or + unistd.h instead. It does not have a protective #error, because + the guard macro for getopt.h in gnulib is not fixed. */ + +/* getopt-core.h and getopt-ext.h are shared with GNU libc, and expect + a number of the internal macros supplied to GNU libc's headers by + sys/cdefs.h. Provide fallback definitions for all of them. */ +#if @HAVE_SYS_CDEFS_H@ +# include <sys/cdefs.h> +#endif + +#ifndef __BEGIN_DECLS +# ifdef __cplusplus +# define __BEGIN_DECLS extern "C" { +# else +# define __BEGIN_DECLS /* nothing */ +# endif +#endif +#ifndef __END_DECLS +# ifdef __cplusplus +# define __END_DECLS } +# else +# define __END_DECLS /* nothing */ +# endif +#endif + +#ifndef __GNUC_PREREQ +# if defined __GNUC__ && defined __GNUC_VERSION__ +# define __GNUC_PREREQ(maj, min) \ + ((__GNUC__ << 16) + __GNUC_MINOR__ >= ((maj) << 16) + (min)) +# else +# define __GNUC_PREREQ(maj, min) 0 +# endif +#endif + +#ifndef __THROW +# if defined __cplusplus && __GNUC_PREREQ (2,8) +# define __THROW throw () +# else +# define __THROW +# endif +#endif + +#endif /* _GETOPT_CDEFS_H */ diff --git a/lib/getopt-core.h b/lib/getopt-core.h new file mode 100644 index 0000000..e51b6c7 --- /dev/null +++ b/lib/getopt-core.h @@ -0,0 +1,96 @@ +/* Declarations for getopt (basic, portable features only). + Copyright (C) 1989-2018 Free Software Foundation, Inc. + This file is part of the GNU C Library and is also part of gnulib. + Patches to this file should be submitted to both projects. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU General Public + License as published by the Free Software Foundation; either + version 3 of the License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + General Public License for more details. + + You should have received a copy of the GNU General Public + License along with the GNU C Library; if not, see + <https://www.gnu.org/licenses/>. */ + +#ifndef _GETOPT_CORE_H +#define _GETOPT_CORE_H 1 + +/* This header should not be used directly; include getopt.h or + unistd.h instead. Unlike most bits headers, it does not have + a protective #error, because the guard macro for getopt.h in + gnulib is not fixed. */ + +__BEGIN_DECLS + +/* For communication from 'getopt' to the caller. + When 'getopt' finds an option that takes an argument, + the argument value is returned here. + Also, when 'ordering' is RETURN_IN_ORDER, + each non-option ARGV-element is returned here. */ + +extern char *optarg; + +/* Index in ARGV of the next element to be scanned. + This is used for communication to and from the caller + and for communication between successive calls to 'getopt'. + + On entry to 'getopt', zero means this is the first call; initialize. + + When 'getopt' returns -1, this is the index of the first of the + non-option elements that the caller should itself scan. + + Otherwise, 'optind' communicates from one call to the next + how much of ARGV has been scanned so far. */ + +extern int optind; + +/* Callers store zero here to inhibit the error message 'getopt' prints + for unrecognized options. */ + +extern int opterr; + +/* Set to an option character which was unrecognized. */ + +extern int optopt; + +/* Get definitions and prototypes for functions to process the + arguments in ARGV (ARGC of them, minus the program name) for + options given in OPTS. + + Return the option character from OPTS just read. Return -1 when + there are no more options. For unrecognized options, or options + missing arguments, 'optopt' is set to the option letter, and '?' is + returned. + + The OPTS string is a list of characters which are recognized option + letters, optionally followed by colons, specifying that that letter + takes an argument, to be placed in 'optarg'. + + If a letter in OPTS is followed by two colons, its argument is + optional. This behavior is specific to the GNU 'getopt'. + + The argument '--' causes premature termination of argument + scanning, explicitly telling 'getopt' that there are no more + options. + + If OPTS begins with '-', then non-option arguments are treated as + arguments to the option '\1'. This behavior is specific to the GNU + 'getopt'. If OPTS begins with '+', or POSIXLY_CORRECT is set in + the environment, then do not permute arguments. + + For standards compliance, the 'argv' argument has the type + char *const *, but this is inaccurate; if argument permutation is + enabled, the argv array (not the strings it points to) must be + writable. */ + +extern int getopt (int ___argc, char *const *___argv, const char *__shortopts) + __THROW _GL_ARG_NONNULL ((2, 3)); + +__END_DECLS + +#endif /* _GETOPT_CORE_H */ diff --git a/lib/getopt-ext.h b/lib/getopt-ext.h new file mode 100644 index 0000000..cb41206 --- /dev/null +++ b/lib/getopt-ext.h @@ -0,0 +1,77 @@ +/* Declarations for getopt (GNU extensions). + Copyright (C) 1989-2018 Free Software Foundation, Inc. + This file is part of the GNU C Library and is also part of gnulib. + Patches to this file should be submitted to both projects. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU General Public + License as published by the Free Software Foundation; either + version 3 of the License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + General Public License for more details. + + You should have received a copy of the GNU General Public + License along with the GNU C Library; if not, see + <https://www.gnu.org/licenses/>. */ + +#ifndef _GETOPT_EXT_H +#define _GETOPT_EXT_H 1 + +/* This header should not be used directly; include getopt.h instead. + Unlike most bits headers, it does not have a protective #error, + because the guard macro for getopt.h in gnulib is not fixed. */ + +__BEGIN_DECLS + +/* Describe the long-named options requested by the application. + The LONG_OPTIONS argument to getopt_long or getopt_long_only is a vector + of 'struct option' terminated by an element containing a name which is + zero. + + The field 'has_arg' is: + no_argument (or 0) if the option does not take an argument, + required_argument (or 1) if the option requires an argument, + optional_argument (or 2) if the option takes an optional argument. + + If the field 'flag' is not NULL, it points to a variable that is set + to the value given in the field 'val' when the option is found, but + left unchanged if the option is not found. + + To have a long-named option do something other than set an 'int' to + a compiled-in constant, such as set a value from 'optarg', set the + option's 'flag' field to zero and its 'val' field to a nonzero + value (the equivalent single-letter option character, if there is + one). For long options that have a zero 'flag' field, 'getopt' + returns the contents of the 'val' field. */ + +struct option +{ + const char *name; + /* has_arg can't be an enum because some compilers complain about + type mismatches in all the code that assumes it is an int. */ + int has_arg; + int *flag; + int val; +}; + +/* Names for the values of the 'has_arg' field of 'struct option'. */ + +#define no_argument 0 +#define required_argument 1 +#define optional_argument 2 + +extern int getopt_long (int ___argc, char *__getopt_argv_const *___argv, + const char *__shortopts, + const struct option *__longopts, int *__longind) + __THROW _GL_ARG_NONNULL ((2, 3)); +extern int getopt_long_only (int ___argc, char *__getopt_argv_const *___argv, + const char *__shortopts, + const struct option *__longopts, int *__longind) + __THROW _GL_ARG_NONNULL ((2, 3)); + +__END_DECLS + +#endif /* _GETOPT_EXT_H */ diff --git a/lib/getopt-pfx-core.h b/lib/getopt-pfx-core.h new file mode 100644 index 0000000..c62f9e2 --- /dev/null +++ b/lib/getopt-pfx-core.h @@ -0,0 +1,59 @@ +/* getopt (basic, portable features) gnulib wrapper header. + Copyright (C) 1989-2018 Free Software Foundation, Inc. + This file is part of gnulib. + Unlike most of the getopt implementation, it is NOT shared + with the GNU C Library. + + This file is free software; you can redistribute it and/or modify it + under the terms of the GNU General Public License as + published by the Free Software Foundation; either version 3 of + the License, or (at your option) any later version. + + This file is distributed in the hope that it will be useful, but + WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + General Public License for more details. + + You should have received a copy of the GNU General Public + License along with gnulib; if not, see + <https://www.gnu.org/licenses/>. */ + +#ifndef _GETOPT_PFX_CORE_H +#define _GETOPT_PFX_CORE_H 1 + +/* This header should not be used directly; include getopt.h or + unistd.h instead. It does not have a protective #error, because + the guard macro for getopt.h in gnulib is not fixed. */ + +/* Standalone applications should #define __GETOPT_PREFIX to an + identifier that prefixes the external functions and variables + defined in getopt-core.h and getopt-ext.h. Systematically + rename identifiers so that they do not collide with the system + functions and variables. Renaming avoids problems with some + compilers and linkers. */ +#ifdef __GETOPT_PREFIX +# ifndef __GETOPT_ID +# define __GETOPT_CONCAT(x, y) x ## y +# define __GETOPT_XCONCAT(x, y) __GETOPT_CONCAT (x, y) +# define __GETOPT_ID(y) __GETOPT_XCONCAT (__GETOPT_PREFIX, y) +# endif +# undef getopt +# undef optarg +# undef opterr +# undef optind +# undef optopt +# define getopt __GETOPT_ID (getopt) +# define optarg __GETOPT_ID (optarg) +# define opterr __GETOPT_ID (opterr) +# define optind __GETOPT_ID (optind) +# define optopt __GETOPT_ID (optopt) + +/* The system's getopt.h may have already included getopt-core.h to + declare the unprefixed identifiers. Undef _GETOPT_CORE_H so that + getopt-core.h declares them with prefixes. */ +# undef _GETOPT_CORE_H +#endif + +#include <getopt-core.h> + +#endif /* _GETOPT_PFX_CORE_H */ diff --git a/lib/getopt-pfx-ext.h b/lib/getopt-pfx-ext.h new file mode 100644 index 0000000..753f129 --- /dev/null +++ b/lib/getopt-pfx-ext.h @@ -0,0 +1,71 @@ +/* getopt (GNU extensions) gnulib wrapper header. + Copyright (C) 1989-2018 Free Software Foundation, Inc. + This file is part of gnulib. + Unlike most of the getopt implementation, it is NOT shared + with the GNU C Library. + + This file is free software; you can redistribute it and/or modify it + under the terms of the GNU General Public License as + published by the Free Software Foundation; either version 3 of + the License, or (at your option) any later version. + + This file is distributed in the hope that it will be useful, but + WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + General Public License for more details. + + You should have received a copy of the GNU General Public + License along with gnulib; if not, see + <https://www.gnu.org/licenses/>. */ + +#ifndef _GETOPT_PFX_EXT_H +#define _GETOPT_PFX_EXT_H 1 + +/* This header should not be used directly; include getopt.h instead. + It does not have a protective #error, because the guard macro for + getopt.h in gnulib is not fixed. */ + +/* Standalone applications should #define __GETOPT_PREFIX to an + identifier that prefixes the external functions and variables + defined in getopt-core.h and getopt-ext.h. Systematically + rename identifiers so that they do not collide with the system + functions and variables. Renaming avoids problems with some + compilers and linkers. */ +#ifdef __GETOPT_PREFIX +# ifndef __GETOPT_ID +# define __GETOPT_CONCAT(x, y) x ## y +# define __GETOPT_XCONCAT(x, y) __GETOPT_CONCAT (x, y) +# define __GETOPT_ID(y) __GETOPT_XCONCAT (__GETOPT_PREFIX, y) +# endif +# undef getopt_long +# undef getopt_long_only +# undef option +# undef _getopt_internal +# define getopt_long __GETOPT_ID (getopt_long) +# define getopt_long_only __GETOPT_ID (getopt_long_only) +# define option __GETOPT_ID (option) +# define _getopt_internal __GETOPT_ID (getopt_internal) + +/* The system's getopt.h may have already included getopt-ext.h to + declare the unprefixed identifiers. Undef _GETOPT_EXT_H so that + getopt-ext.h declares them with prefixes. */ +# undef _GETOPT_EXT_H +#endif + +/* Standalone applications get correct prototypes for getopt_long and + getopt_long_only; they declare "char **argv". For backward + compatibility with old applications, if __GETOPT_PREFIX is not + defined, we supply GNU-libc-compatible, but incorrect, prototypes + using "char *const *argv". (GNU libc is stuck with the incorrect + prototypes, as they are baked into older versions of LSB.) */ +#ifndef __getopt_argv_const +# if defined __GETOPT_PREFIX +# define __getopt_argv_const /* empty */ +# else +# define __getopt_argv_const const +# endif +#endif + +#include <getopt-ext.h> + +#endif /* _GETOPT_PFX_EXT_H */ diff --git a/lib/getopt.c b/lib/getopt.c new file mode 100644 index 0000000..11e36ee --- /dev/null +++ b/lib/getopt.c @@ -0,0 +1,811 @@ +/* Getopt for GNU. + Copyright (C) 1987-2018 Free Software Foundation, Inc. + This file is part of the GNU C Library and is also part of gnulib. + Patches to this file should be submitted to both projects. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU General Public + License as published by the Free Software Foundation; either + version 3 of the License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + General Public License for more details. + + You should have received a copy of the GNU General Public + License along with the GNU C Library; if not, see + <https://www.gnu.org/licenses/>. */ + +#ifndef _LIBC +# include <config.h> +#endif + +#include "getopt.h" + +#include <stdio.h> +#include <stdlib.h> +#include <string.h> +#include <unistd.h> + +#ifdef _LIBC +/* When used as part of glibc, error printing must be done differently + for standards compliance. getopt is not a cancellation point, so + it must not call functions that are, and it is specified by an + older standard than stdio locking, so it must not refer to + functions in the "user namespace" related to stdio locking. + Finally, it must use glibc's internal message translation so that + the messages are looked up in the proper text domain. */ +# include <libintl.h> +# define fprintf __fxprintf_nocancel +# define flockfile(fp) _IO_flockfile (fp) +# define funlockfile(fp) _IO_funlockfile (fp) +#else +# include "gettext.h" +# define _(msgid) gettext (msgid) +/* When used standalone, flockfile and funlockfile might not be + available. */ +# if (!defined _POSIX_THREAD_SAFE_FUNCTIONS \ + || (defined _WIN32 && ! defined __CYGWIN__)) +# define flockfile(fp) /* nop */ +# define funlockfile(fp) /* nop */ +# endif +/* When used standalone, do not attempt to use alloca. */ +# define __libc_use_alloca(size) 0 +# undef alloca +# define alloca(size) (abort (), (void *)0) +#endif + +/* This implementation of 'getopt' has three modes for handling + options interspersed with non-option arguments. It can stop + scanning for options at the first non-option argument encountered, + as POSIX specifies. It can continue scanning for options after the + first non-option argument, but permute 'argv' as it goes so that, + after 'getopt' is done, all the options precede all the non-option + arguments and 'optind' points to the first non-option argument. + Or, it can report non-option arguments as if they were arguments to + the option character '\x01'. + + The default behavior of 'getopt_long' is to permute the argument list. + When this implementation is used standalone, the default behavior of + 'getopt' is to stop at the first non-option argument, but when it is + used as part of GNU libc it also permutes the argument list. In both + cases, setting the environment variable POSIXLY_CORRECT to any value + disables permutation. + + If the first character of the OPTSTRING argument to 'getopt' or + 'getopt_long' is '+', both functions will stop at the first + non-option argument. If it is '-', both functions will report + non-option arguments as arguments to the option character '\x01'. */ + +#include "getopt_int.h" + +/* For communication from 'getopt' to the caller. + When 'getopt' finds an option that takes an argument, + the argument value is returned here. + Also, when 'ordering' is RETURN_IN_ORDER, + each non-option ARGV-element is returned here. */ + +char *optarg; + +/* Index in ARGV of the next element to be scanned. + This is used for communication to and from the caller + and for communication between successive calls to 'getopt'. + + On entry to 'getopt', zero means this is the first call; initialize. + + When 'getopt' returns -1, this is the index of the first of the + non-option elements that the caller should itself scan. + + Otherwise, 'optind' communicates from one call to the next + how much of ARGV has been scanned so far. */ + +/* 1003.2 says this must be 1 before any call. */ +int optind = 1; + +/* Callers store zero here to inhibit the error message + for unrecognized options. */ + +int opterr = 1; + +/* Set to an option character which was unrecognized. + This must be initialized on some systems to avoid linking in the + system's own getopt implementation. */ + +int optopt = '?'; + +/* Keep a global copy of all internal members of getopt_data. */ + +static struct _getopt_data getopt_data; + +/* Exchange two adjacent subsequences of ARGV. + One subsequence is elements [first_nonopt,last_nonopt) + which contains all the non-options that have been skipped so far. + The other is elements [last_nonopt,optind), which contains all + the options processed since those non-options were skipped. + + 'first_nonopt' and 'last_nonopt' are relocated so that they describe + the new indices of the non-options in ARGV after they are moved. */ + +static void +exchange (char **argv, struct _getopt_data *d) +{ + int bottom = d->__first_nonopt; + int middle = d->__last_nonopt; + int top = d->optind; + char *tem; + + /* Exchange the shorter segment with the far end of the longer segment. + That puts the shorter segment into the right place. + It leaves the longer segment in the right place overall, + but it consists of two parts that need to be swapped next. */ + + while (top > middle && middle > bottom) + { + if (top - middle > middle - bottom) + { + /* Bottom segment is the short one. */ + int len = middle - bottom; + int i; + + /* Swap it with the top part of the top segment. */ + for (i = 0; i < len; i++) + { + tem = argv[bottom + i]; + argv[bottom + i] = argv[top - (middle - bottom) + i]; + argv[top - (middle - bottom) + i] = tem; + } + /* Exclude the moved bottom segment from further swapping. */ + top -= len; + } + else + { + /* Top segment is the short one. */ + int len = top - middle; + int i; + + /* Swap it with the bottom part of the bottom segment. */ + for (i = 0; i < len; i++) + { + tem = argv[bottom + i]; + argv[bottom + i] = argv[middle + i]; + argv[middle + i] = tem; + } + /* Exclude the moved top segment from further swapping. */ + bottom += len; + } + } + + /* Update records for the slots the non-options now occupy. */ + + d->__first_nonopt += (d->optind - d->__last_nonopt); + d->__last_nonopt = d->optind; +} + +/* Process the argument starting with d->__nextchar as a long option. + d->optind should *not* have been advanced over this argument. + + If the value returned is -1, it was not actually a long option, the + state is unchanged, and the argument should be processed as a set + of short options (this can only happen when long_only is true). + Otherwise, the option (and its argument, if any) have been consumed + and the return value is the value to return from _getopt_internal_r. */ +static int +process_long_option (int argc, char **argv, const char *optstring, + const struct option *longopts, int *longind, + int long_only, struct _getopt_data *d, + int print_errors, const char *prefix) +{ + char *nameend; + size_t namelen; + const struct option *p; + const struct option *pfound = NULL; + int n_options; + int option_index; + + for (nameend = d->__nextchar; *nameend && *nameend != '='; nameend++) + /* Do nothing. */ ; + namelen = nameend - d->__nextchar; + + /* First look for an exact match, counting the options as a side + effect. */ + for (p = longopts, n_options = 0; p->name; p++, n_options++) + if (!strncmp (p->name, d->__nextchar, namelen) + && namelen == strlen (p->name)) + { + /* Exact match found. */ + pfound = p; + option_index = n_options; + break; + } + + if (pfound == NULL) + { + /* Didn't find an exact match, so look for abbreviations. */ + unsigned char *ambig_set = NULL; + int ambig_malloced = 0; + int ambig_fallback = 0; + int indfound = -1; + + for (p = longopts, option_index = 0; p->name; p++, option_index++) + if (!strncmp (p->name, d->__nextchar, namelen)) + { + if (pfound == NULL) + { + /* First nonexact match found. */ + pfound = p; + indfound = option_index; + } + else if (long_only + || pfound->has_arg != p->has_arg + || pfound->flag != p->flag + || pfound->val != p->val) + { + /* Second or later nonexact match found. */ + if (!ambig_fallback) + { + if (!print_errors) + /* Don't waste effort tracking the ambig set if + we're not going to print it anyway. */ + ambig_fallback = 1; + else if (!ambig_set) + { + if (__libc_use_alloca (n_options)) + ambig_set = alloca (n_options); + else if ((ambig_set = malloc (n_options)) == NULL) + /* Fall back to simpler error message. */ + ambig_fallback = 1; + else + ambig_malloced = 1; + + if (ambig_set) + { + memset (ambig_set, 0, n_options); + ambig_set[indfound] = 1; + } + } + if (ambig_set) + ambig_set[option_index] = 1; + } + } + } + + if (ambig_set || ambig_fallback) + { + if (print_errors) + { + if (ambig_fallback) + fprintf (stderr, _("%s: option '%s%s' is ambiguous\n"), + argv[0], prefix, d->__nextchar); + else + { + flockfile (stderr); + fprintf (stderr, + _("%s: option '%s%s' is ambiguous; possibilities:"), + argv[0], prefix, d->__nextchar); + + for (option_index = 0; option_index < n_options; option_index++) + if (ambig_set[option_index]) + fprintf (stderr, " '%s%s'", + prefix, longopts[option_index].name); + + /* This must use 'fprintf' even though it's only + printing a single character, so that it goes through + __fxprintf_nocancel when compiled as part of glibc. */ + fprintf (stderr, "\n"); + funlockfile (stderr); + } + } + if (ambig_malloced) + free (ambig_set); + d->__nextchar += strlen (d->__nextchar); + d->optind++; + d->optopt = 0; + return '?'; + } + + option_index = indfound; + } + + if (pfound == NULL) + { + /* Can't find it as a long option. If this is not getopt_long_only, + or the option starts with '--' or is not a valid short option, + then it's an error. */ + if (!long_only || argv[d->optind][1] == '-' + || strchr (optstring, *d->__nextchar) == NULL) + { + if (print_errors) + fprintf (stderr, _("%s: unrecognized option '%s%s'\n"), + argv[0], prefix, d->__nextchar); + + d->__nextchar = NULL; + d->optind++; + d->optopt = 0; + return '?'; + } + + /* Otherwise interpret it as a short option. */ + return -1; + } + + /* We have found a matching long option. Consume it. */ + d->optind++; + d->__nextchar = NULL; + if (*nameend) + { + /* Don't test has_arg with >, because some C compilers don't + allow it to be used on enums. */ + if (pfound->has_arg) + d->optarg = nameend + 1; + else + { + if (print_errors) + fprintf (stderr, + _("%s: option '%s%s' doesn't allow an argument\n"), + argv[0], prefix, pfound->name); + + d->optopt = pfound->val; + return '?'; + } + } + else if (pfound->has_arg == 1) + { + if (d->optind < argc) + d->optarg = argv[d->optind++]; + else + { + if (print_errors) + fprintf (stderr, + _("%s: option '%s%s' requires an argument\n"), + argv[0], prefix, pfound->name); + + d->optopt = pfound->val; + return optstring[0] == ':' ? ':' : '?'; + } + } + + if (longind != NULL) + *longind = option_index; + if (pfound->flag) + { + *(pfound->flag) = pfound->val; + return 0; + } + return pfound->val; +} + +/* Initialize internal data upon the first call to getopt. */ + +static const char * +_getopt_initialize (int argc _GL_UNUSED, + char **argv _GL_UNUSED, const char *optstring, + struct _getopt_data *d, int posixly_correct) +{ + /* Start processing options with ARGV-element 1 (since ARGV-element 0 + is the program name); the sequence of previously skipped + non-option ARGV-elements is empty. */ + if (d->optind == 0) + d->optind = 1; + + d->__first_nonopt = d->__last_nonopt = d->optind; + d->__nextchar = NULL; + + /* Determine how to handle the ordering of options and nonoptions. */ + if (optstring[0] == '-') + { + d->__ordering = RETURN_IN_ORDER; + ++optstring; + } + else if (optstring[0] == '+') + { + d->__ordering = REQUIRE_ORDER; + ++optstring; + } + else if (posixly_correct || !!getenv ("POSIXLY_CORRECT")) + d->__ordering = REQUIRE_ORDER; + else + d->__ordering = PERMUTE; + + d->__initialized = 1; + return optstring; +} + +/* Scan elements of ARGV (whose length is ARGC) for option characters + given in OPTSTRING. + + If an element of ARGV starts with '-', and is not exactly "-" or "--", + then it is an option element. The characters of this element + (aside from the initial '-') are option characters. If 'getopt' + is called repeatedly, it returns successively each of the option characters + from each of the option elements. + + If 'getopt' finds another option character, it returns that character, + updating 'optind' and 'nextchar' so that the next call to 'getopt' can + resume the scan with the following option character or ARGV-element. + + If there are no more option characters, 'getopt' returns -1. + Then 'optind' is the index in ARGV of the first ARGV-element + that is not an option. (The ARGV-elements have been permuted + so that those that are not options now come last.) + + OPTSTRING is a string containing the legitimate option characters. + If an option character is seen that is not listed in OPTSTRING, + return '?' after printing an error message. If you set 'opterr' to + zero, the error message is suppressed but we still return '?'. + + If a char in OPTSTRING is followed by a colon, that means it wants an arg, + so the following text in the same ARGV-element, or the text of the following + ARGV-element, is returned in 'optarg'. Two colons mean an option that + wants an optional arg; if there is text in the current ARGV-element, + it is returned in 'optarg', otherwise 'optarg' is set to zero. + + If OPTSTRING starts with '-' or '+', it requests different methods of + handling the non-option ARGV-elements. + See the comments about RETURN_IN_ORDER and REQUIRE_ORDER, above. + + Long-named options begin with '--' instead of '-'. + Their names may be abbreviated as long as the abbreviation is unique + or is an exact match for some defined option. If they have an + argument, it follows the option name in the same ARGV-element, separated + from the option name by a '=', or else the in next ARGV-element. + When 'getopt' finds a long-named option, it returns 0 if that option's + 'flag' field is nonzero, the value of the option's 'val' field + if the 'flag' field is zero. + + The elements of ARGV aren't really const, because we permute them. + But we pretend they're const in the prototype to be compatible + with other systems. + + LONGOPTS is a vector of 'struct option' terminated by an + element containing a name which is zero. + + LONGIND returns the index in LONGOPT of the long-named option found. + It is only valid when a long-named option has been found by the most + recent call. + + If LONG_ONLY is nonzero, '-' as well as '--' can introduce + long-named options. */ + +int +_getopt_internal_r (int argc, char **argv, const char *optstring, + const struct option *longopts, int *longind, + int long_only, struct _getopt_data *d, int posixly_correct) +{ + int print_errors = d->opterr; + + if (argc < 1) + return -1; + + d->optarg = NULL; + + if (d->optind == 0 || !d->__initialized) + optstring = _getopt_initialize (argc, argv, optstring, d, posixly_correct); + else if (optstring[0] == '-' || optstring[0] == '+') + optstring++; + + if (optstring[0] == ':') + print_errors = 0; + + /* Test whether ARGV[optind] points to a non-option argument. */ +#define NONOPTION_P (argv[d->optind][0] != '-' || argv[d->optind][1] == '\0') + + if (d->__nextchar == NULL || *d->__nextchar == '\0') + { + /* Advance to the next ARGV-element. */ + + /* Give FIRST_NONOPT & LAST_NONOPT rational values if OPTIND has been + moved back by the user (who may also have changed the arguments). */ + if (d->__last_nonopt > d->optind) + d->__last_nonopt = d->optind; + if (d->__first_nonopt > d->optind) + d->__first_nonopt = d->optind; + + if (d->__ordering == PERMUTE) + { + /* If we have just processed some options following some non-options, + exchange them so that the options come first. */ + + if (d->__first_nonopt != d->__last_nonopt + && d->__last_nonopt != d->optind) + exchange (argv, d); + else if (d->__last_nonopt != d->optind) + d->__first_nonopt = d->optind; + + /* Skip any additional non-options + and extend the range of non-options previously skipped. */ + + while (d->optind < argc && NONOPTION_P) + d->optind++; + d->__last_nonopt = d->optind; + } + + /* The special ARGV-element '--' means premature end of options. + Skip it like a null option, + then exchange with previous non-options as if it were an option, + then skip everything else like a non-option. */ + + if (d->optind != argc && !strcmp (argv[d->optind], "--")) + { + d->optind++; + + if (d->__first_nonopt != d->__last_nonopt + && d->__last_nonopt != d->optind) + exchange (argv, d); + else if (d->__first_nonopt == d->__last_nonopt) + d->__first_nonopt = d->optind; + d->__last_nonopt = argc; + + d->optind = argc; + } + + /* If we have done all the ARGV-elements, stop the scan + and back over any non-options that we skipped and permuted. */ + + if (d->optind == argc) + { + /* Set the next-arg-index to point at the non-options + that we previously skipped, so the caller will digest them. */ + if (d->__first_nonopt != d->__last_nonopt) + d->optind = d->__first_nonopt; + return -1; + } + + /* If we have come to a non-option and did not permute it, + either stop the scan or describe it to the caller and pass it by. */ + + if (NONOPTION_P) + { + if (d->__ordering == REQUIRE_ORDER) + return -1; + d->optarg = argv[d->optind++]; + return 1; + } + + /* We have found another option-ARGV-element. + Check whether it might be a long option. */ + if (longopts) + { + if (argv[d->optind][1] == '-') + { + /* "--foo" is always a long option. The special option + "--" was handled above. */ + d->__nextchar = argv[d->optind] + 2; + return process_long_option (argc, argv, optstring, longopts, + longind, long_only, d, + print_errors, "--"); + } + + /* If long_only and the ARGV-element has the form "-f", + where f is a valid short option, don't consider it an + abbreviated form of a long option that starts with f. + Otherwise there would be no way to give the -f short + option. + + On the other hand, if there's a long option "fubar" and + the ARGV-element is "-fu", do consider that an + abbreviation of the long option, just like "--fu", and + not "-f" with arg "u". + + This distinction seems to be the most useful approach. */ + if (long_only && (argv[d->optind][2] + || !strchr (optstring, argv[d->optind][1]))) + { + int code; + d->__nextchar = argv[d->optind] + 1; + code = process_long_option (argc, argv, optstring, longopts, + longind, long_only, d, + print_errors, "-"); + if (code != -1) + return code; + } + } + + /* It is not a long option. Skip the initial punctuation. */ + d->__nextchar = argv[d->optind] + 1; + } + + /* Look at and handle the next short option-character. */ + + { + char c = *d->__nextchar++; + const char *temp = strchr (optstring, c); + + /* Increment 'optind' when we start to process its last character. */ + if (*d->__nextchar == '\0') + ++d->optind; + + if (temp == NULL || c == ':' || c == ';') + { + if (print_errors) + fprintf (stderr, _("%s: invalid option -- '%c'\n"), argv[0], c); + d->optopt = c; + return '?'; + } + + /* Convenience. Treat POSIX -W foo same as long option --foo */ + if (temp[0] == 'W' && temp[1] == ';' && longopts != NULL) + { + /* This is an option that requires an argument. */ + if (*d->__nextchar != '\0') + d->optarg = d->__nextchar; + else if (d->optind == argc) + { + if (print_errors) + fprintf (stderr, + _("%s: option requires an argument -- '%c'\n"), + argv[0], c); + + d->optopt = c; + if (optstring[0] == ':') + c = ':'; + else + c = '?'; + return c; + } + else + d->optarg = argv[d->optind]; + + d->__nextchar = d->optarg; + d->optarg = NULL; + return process_long_option (argc, argv, optstring, longopts, longind, + 0 /* long_only */, d, print_errors, "-W "); + } + if (temp[1] == ':') + { + if (temp[2] == ':') + { + /* This is an option that accepts an argument optionally. */ + if (*d->__nextchar != '\0') + { + d->optarg = d->__nextchar; + d->optind++; + } + else + d->optarg = NULL; + d->__nextchar = NULL; + } + else + { + /* This is an option that requires an argument. */ + if (*d->__nextchar != '\0') + { + d->optarg = d->__nextchar; + /* If we end this ARGV-element by taking the rest as an arg, + we must advance to the next element now. */ + d->optind++; + } + else if (d->optind == argc) + { + if (print_errors) + fprintf (stderr, + _("%s: option requires an argument -- '%c'\n"), + argv[0], c); + + d->optopt = c; + if (optstring[0] == ':') + c = ':'; + else + c = '?'; + } + else + /* We already incremented 'optind' once; + increment it again when taking next ARGV-elt as argument. */ + d->optarg = argv[d->optind++]; + d->__nextchar = NULL; + } + } + return c; + } +} + +int +_getopt_internal (int argc, char **argv, const char *optstring, + const struct option *longopts, int *longind, int long_only, + int posixly_correct) +{ + int result; + + getopt_data.optind = optind; + getopt_data.opterr = opterr; + + result = _getopt_internal_r (argc, argv, optstring, longopts, + longind, long_only, &getopt_data, + posixly_correct); + + optind = getopt_data.optind; + optarg = getopt_data.optarg; + optopt = getopt_data.optopt; + + return result; +} + +/* glibc gets a LSB-compliant getopt and a POSIX-complaint __posix_getopt. + Standalone applications just get a POSIX-compliant getopt. + POSIX and LSB both require these functions to take 'char *const *argv' + even though this is incorrect (because of the permutation). */ +#define GETOPT_ENTRY(NAME, POSIXLY_CORRECT) \ + int \ + NAME (int argc, char *const *argv, const char *optstring) \ + { \ + return _getopt_internal (argc, (char **)argv, optstring, \ + 0, 0, 0, POSIXLY_CORRECT); \ + } + +#ifdef _LIBC +GETOPT_ENTRY(getopt, 0) +GETOPT_ENTRY(__posix_getopt, 1) +#else +GETOPT_ENTRY(getopt, 1) +#endif + + +#ifdef TEST + +/* Compile with -DTEST to make an executable for use in testing + the above definition of 'getopt'. */ + +int +main (int argc, char **argv) +{ + int c; + int digit_optind = 0; + + while (1) + { + int this_option_optind = optind ? optind : 1; + + c = getopt (argc, argv, "abc:d:0123456789"); + if (c == -1) + break; + + switch (c) + { + case '0': + case '1': + case '2': + case '3': + case '4': + case '5': + case '6': + case '7': + case '8': + case '9': + if (digit_optind != 0 && digit_optind != this_option_optind) + printf ("digits occur in two different argv-elements.\n"); + digit_optind = this_option_optind; + printf ("option %c\n", c); + break; + + case 'a': + printf ("option a\n"); + break; + + case 'b': + printf ("option b\n"); + break; + + case 'c': + printf ("option c with value '%s'\n", optarg); + break; + + case '?': + break; + + default: + printf ("?? getopt returned character code 0%o ??\n", c); + } + } + + if (optind < argc) + { + printf ("non-option ARGV-elements: "); + while (optind < argc) + printf ("%s ", argv[optind++]); + printf ("\n"); + } + + exit (0); +} + +#endif /* TEST */ diff --git a/lib/getopt.in.h b/lib/getopt.in.h new file mode 100644 index 0000000..5fb58dd --- /dev/null +++ b/lib/getopt.in.h @@ -0,0 +1,61 @@ +/* Declarations for getopt. + Copyright (C) 1989-2018 Free Software Foundation, Inc. + This file is part of gnulib. + Unlike most of the getopt implementation, it is NOT shared + with the GNU C Library, which supplies a different version of + this file. + + This file is free software; you can redistribute it and/or modify it + under the terms of the GNU General Public License as + published by the Free Software Foundation; either version 3 of + the License, or (at your option) any later version. + + This file is distributed in the hope that it will be useful, but + WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + General Public License for more details. + + You should have received a copy of the GNU General Public + License along with gnulib; if not, see <https://www.gnu.org/licenses/>. */ + +#ifndef _@GUARD_PREFIX@_GETOPT_H + +#if __GNUC__ >= 3 +@PRAGMA_SYSTEM_HEADER@ +#endif +@PRAGMA_COLUMNS@ + +/* The include_next requires a split double-inclusion guard. We must + also inform the replacement unistd.h to not recursively use + <getopt.h>; our definitions will be present soon enough. */ +#if @HAVE_GETOPT_H@ +# define _GL_SYSTEM_GETOPT +# @INCLUDE_NEXT@ @NEXT_GETOPT_H@ +# undef _GL_SYSTEM_GETOPT +#endif + +#define _@GUARD_PREFIX@_GETOPT_H 1 + +/* Standalone applications should #define __GETOPT_PREFIX to an + identifier that prefixes the external functions and variables + defined in getopt-core.h and getopt-ext.h. When this happens, + include the headers that might declare getopt so that they will not + cause confusion if included after this file (if the system had + <getopt.h>, we have already included it). */ +#if defined __GETOPT_PREFIX +# if !@HAVE_GETOPT_H@ +# define __need_system_stdlib_h +# include <stdlib.h> +# undef __need_system_stdlib_h +# include <stdio.h> +# include <unistd.h> +# endif +#endif + +/* The definition of _GL_ARG_NONNULL is copied here. */ + +#include <getopt-cdefs.h> +#include <getopt-pfx-core.h> +#include <getopt-pfx-ext.h> + +#endif /* _@GUARD_PREFIX@_GETOPT_H */ diff --git a/lib/getopt1.c b/lib/getopt1.c new file mode 100644 index 0000000..9c7fff4 --- /dev/null +++ b/lib/getopt1.c @@ -0,0 +1,159 @@ +/* getopt_long and getopt_long_only entry points for GNU getopt. + Copyright (C) 1987-2018 Free Software Foundation, Inc. + This file is part of the GNU C Library and is also part of gnulib. + Patches to this file should be submitted to both projects. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU General Public + License as published by the Free Software Foundation; either + version 3 of the License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + General Public License for more details. + + You should have received a copy of the GNU General Public + License along with the GNU C Library; if not, see + <https://www.gnu.org/licenses/>. */ + +#ifndef _LIBC +# include <config.h> +#endif + +#include "getopt.h" +#include "getopt_int.h" + +int +getopt_long (int argc, char *__getopt_argv_const *argv, const char *options, + const struct option *long_options, int *opt_index) +{ + return _getopt_internal (argc, (char **) argv, options, long_options, + opt_index, 0, 0); +} + +int +_getopt_long_r (int argc, char **argv, const char *options, + const struct option *long_options, int *opt_index, + struct _getopt_data *d) +{ + return _getopt_internal_r (argc, argv, options, long_options, opt_index, + 0, d, 0); +} + +/* Like getopt_long, but '-' as well as '--' can indicate a long option. + If an option that starts with '-' (not '--') doesn't match a long option, + but does match a short option, it is parsed as a short option + instead. */ + +int +getopt_long_only (int argc, char *__getopt_argv_const *argv, + const char *options, + const struct option *long_options, int *opt_index) +{ + return _getopt_internal (argc, (char **) argv, options, long_options, + opt_index, 1, 0); +} + +int +_getopt_long_only_r (int argc, char **argv, const char *options, + const struct option *long_options, int *opt_index, + struct _getopt_data *d) +{ + return _getopt_internal_r (argc, argv, options, long_options, opt_index, + 1, d, 0); +} + + +#ifdef TEST + +#include <stdio.h> +#include <stdlib.h> + +int +main (int argc, char **argv) +{ + int c; + int digit_optind = 0; + + while (1) + { + int this_option_optind = optind ? optind : 1; + int option_index = 0; + static const struct option long_options[] = + { + {"add", 1, 0, 0}, + {"append", 0, 0, 0}, + {"delete", 1, 0, 0}, + {"verbose", 0, 0, 0}, + {"create", 0, 0, 0}, + {"file", 1, 0, 0}, + {0, 0, 0, 0} + }; + + c = getopt_long (argc, argv, "abc:d:0123456789", + long_options, &option_index); + if (c == -1) + break; + + switch (c) + { + case 0: + printf ("option %s", long_options[option_index].name); + if (optarg) + printf (" with arg %s", optarg); + printf ("\n"); + break; + + case '0': + case '1': + case '2': + case '3': + case '4': + case '5': + case '6': + case '7': + case '8': + case '9': + if (digit_optind != 0 && digit_optind != this_option_optind) + printf ("digits occur in two different argv-elements.\n"); + digit_optind = this_option_optind; + printf ("option %c\n", c); + break; + + case 'a': + printf ("option a\n"); + break; + + case 'b': + printf ("option b\n"); + break; + + case 'c': + printf ("option c with value '%s'\n", optarg); + break; + + case 'd': + printf ("option d with value '%s'\n", optarg); + break; + + case '?': + break; + + default: + printf ("?? getopt returned character code 0%o ??\n", c); + } + } + + if (optind < argc) + { + printf ("non-option ARGV-elements: "); + while (optind < argc) + printf ("%s ", argv[optind++]); + printf ("\n"); + } + + exit (0); +} + +#endif /* TEST */ diff --git a/lib/getopt_int.h b/lib/getopt_int.h new file mode 100644 index 0000000..b0e9a6d --- /dev/null +++ b/lib/getopt_int.h @@ -0,0 +1,118 @@ +/* Internal declarations for getopt. + Copyright (C) 1989-2018 Free Software Foundation, Inc. + This file is part of the GNU C Library and is also part of gnulib. + Patches to this file should be submitted to both projects. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU General Public + License as published by the Free Software Foundation; either + version 3 of the License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + General Public License for more details. + + You should have received a copy of the GNU General Public + License along with the GNU C Library; if not, see + <https://www.gnu.org/licenses/>. */ + +#ifndef _GETOPT_INT_H +#define _GETOPT_INT_H 1 + +#include <getopt.h> + +extern int _getopt_internal (int ___argc, char **___argv, + const char *__shortopts, + const struct option *__longopts, int *__longind, + int __long_only, int __posixly_correct); + + +/* Reentrant versions which can handle parsing multiple argument + vectors at the same time. */ + +/* Describe how to deal with options that follow non-option ARGV-elements. + + REQUIRE_ORDER means don't recognize them as options; stop option + processing when the first non-option is seen. This is what POSIX + specifies should happen. + + PERMUTE means permute the contents of ARGV as we scan, so that + eventually all the non-options are at the end. This allows options + to be given in any order, even with programs that were not written + to expect this. + + RETURN_IN_ORDER is an option available to programs that were + written to expect options and other ARGV-elements in any order + and that care about the ordering of the two. We describe each + non-option ARGV-element as if it were the argument of an option + with character code 1. + + The special argument '--' forces an end of option-scanning regardless + of the value of 'ordering'. In the case of RETURN_IN_ORDER, only + '--' can cause 'getopt' to return -1 with 'optind' != ARGC. */ + +enum __ord + { + REQUIRE_ORDER, PERMUTE, RETURN_IN_ORDER + }; + +/* Data type for reentrant functions. */ +struct _getopt_data +{ + /* These have exactly the same meaning as the corresponding global + variables, except that they are used for the reentrant + versions of getopt. */ + int optind; + int opterr; + int optopt; + char *optarg; + + /* Internal members. */ + + /* True if the internal members have been initialized. */ + int __initialized; + + /* The next char to be scanned in the option-element + in which the last option character we returned was found. + This allows us to pick up the scan where we left off. + + If this is zero, or a null string, it means resume the scan + by advancing to the next ARGV-element. */ + char *__nextchar; + + /* See __ord above. */ + enum __ord __ordering; + + /* Handle permutation of arguments. */ + + /* Describe the part of ARGV that contains non-options that have + been skipped. 'first_nonopt' is the index in ARGV of the first + of them; 'last_nonopt' is the index after the last of them. */ + + int __first_nonopt; + int __last_nonopt; +}; + +/* The initializer is necessary to set OPTIND and OPTERR to their + default values and to clear the initialization flag. */ +#define _GETOPT_DATA_INITIALIZER { 1, 1 } + +extern int _getopt_internal_r (int ___argc, char **___argv, + const char *__shortopts, + const struct option *__longopts, int *__longind, + int __long_only, struct _getopt_data *__data, + int __posixly_correct); + +extern int _getopt_long_r (int ___argc, char **___argv, + const char *__shortopts, + const struct option *__longopts, int *__longind, + struct _getopt_data *__data); + +extern int _getopt_long_only_r (int ___argc, char **___argv, + const char *__shortopts, + const struct option *__longopts, + int *__longind, + struct _getopt_data *__data); + +#endif /* getopt_int.h */ diff --git a/lib/getprogname.c b/lib/getprogname.c new file mode 100644 index 0000000..58da144 --- /dev/null +++ b/lib/getprogname.c @@ -0,0 +1,255 @@ +/* Program name management. + Copyright (C) 2016-2018 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +#include <config.h> + +/* Specification. */ +#include "getprogname.h" + +#include <errno.h> /* get program_invocation_name declaration */ +#include <stdlib.h> /* get __argv declaration */ + +#ifdef _AIX +# include <unistd.h> +# include <procinfo.h> +# include <string.h> +#endif + +#ifdef __MVS__ +# ifndef _OPEN_SYS +# define _OPEN_SYS +# endif +# include <string.h> +# include <sys/ps.h> +#endif + +#ifdef __hpux +# include <unistd.h> +# include <sys/param.h> +# include <sys/pstat.h> +# include <string.h> +#endif + +#ifdef __sgi +# include <string.h> +# include <unistd.h> +# include <stdio.h> +# include <fcntl.h> +# include <sys/procfs.h> +#endif + +#include "dirname.h" + +#ifndef HAVE_GETPROGNAME /* not Mac OS X, FreeBSD, NetBSD, OpenBSD >= 5.4, Cygwin */ +char const * +getprogname (void) +{ +# if HAVE_DECL_PROGRAM_INVOCATION_SHORT_NAME /* glibc, BeOS */ + /* https://www.gnu.org/software/libc/manual/html_node/Error-Messages.html */ + return program_invocation_short_name; +# elif HAVE_DECL_PROGRAM_INVOCATION_NAME /* glibc, BeOS */ + /* https://www.gnu.org/software/libc/manual/html_node/Error-Messages.html */ + return last_component (program_invocation_name); +# elif HAVE_GETEXECNAME /* Solaris */ + /* https://docs.oracle.com/cd/E19253-01/816-5168/6mbb3hrb1/index.html */ + const char *p = getexecname (); + if (!p) + p = "?"; + return last_component (p); +# elif HAVE_DECL___ARGV /* mingw, MSVC */ + /* https://msdn.microsoft.com/en-us/library/dn727674.aspx */ + const char *p = __argv && __argv[0] ? __argv[0] : "?"; + return last_component (p); +# elif HAVE_VAR___PROGNAME /* OpenBSD, QNX */ + /* https://man.openbsd.org/style.9 */ + /* http://www.qnx.de/developers/docs/6.5.0/index.jsp?topic=%2Fcom.qnx.doc.neutrino_lib_ref%2Fp%2F__progname.html */ + /* Be careful to declare this only when we absolutely need it + (OpenBSD 5.1), rather than when it's available. Otherwise, + its mere declaration makes program_invocation_short_name + malfunction (have zero length) with Fedora 25's glibc. */ + extern char *__progname; + const char *p = __progname; + return p && p[0] ? p : "?"; +# elif _AIX /* AIX */ + /* Idea by Bastien ROUCARIÈS, + https://lists.gnu.org/r/bug-gnulib/2010-12/msg00095.html + Reference: https://www.ibm.com/support/knowledgecenter/en/ssw_aix_61/com.ibm.aix.basetrf1/getprocs.htm + */ + static char *p; + static int first = 1; + if (first) + { + first = 0; + pid_t pid = getpid (); + struct procentry64 procs; + p = (0 < getprocs64 (&procs, sizeof procs, NULL, 0, &pid, 1) + ? strdup (procs.pi_comm) + : NULL); + if (!p) + p = "?"; + } + return p; +# elif defined __hpux + static char *p; + static int first = 1; + if (first) + { + first = 0; + pid_t pid = getpid (); + struct pst_status status; + if (pstat_getproc (&status, sizeof status, 0, pid) > 0) + { + char *ucomm = status.pst_ucomm; + char *cmd = status.pst_cmd; + if (strlen (ucomm) < PST_UCOMMLEN - 1) + p = ucomm; + else + { + /* ucomm is truncated to length PST_UCOMMLEN - 1. + Look at cmd instead. */ + char *space = strchr (cmd, ' '); + if (space != NULL) + *space = '\0'; + p = strrchr (cmd, '/'); + if (p != NULL) + p++; + else + p = cmd; + if (strlen (p) > PST_UCOMMLEN - 1 + && memcmp (p, ucomm, PST_UCOMMLEN - 1) == 0) + /* p is less truncated than ucomm. */ + ; + else + p = ucomm; + } + p = strdup (p); + } + else + { +# if !defined __LP64__ + /* Support for 32-bit programs running in 64-bit HP-UX. + The documented way to do this is to use the same source code + as above, but in a compilation unit where '#define _PSTAT64 1' + is in effect. I prefer a single compilation unit; the struct + size and the offsets are not going to change. */ + char status64[1216]; + if (__pstat_getproc64 (status64, sizeof status64, 0, pid) > 0) + { + char *ucomm = status64 + 288; + char *cmd = status64 + 168; + if (strlen (ucomm) < PST_UCOMMLEN - 1) + p = ucomm; + else + { + /* ucomm is truncated to length PST_UCOMMLEN - 1. + Look at cmd instead. */ + char *space = strchr (cmd, ' '); + if (space != NULL) + *space = '\0'; + p = strrchr (cmd, '/'); + if (p != NULL) + p++; + else + p = cmd; + if (strlen (p) > PST_UCOMMLEN - 1 + && memcmp (p, ucomm, PST_UCOMMLEN - 1) == 0) + /* p is less truncated than ucomm. */ + ; + else + p = ucomm; + } + p = strdup (p); + } + else +# endif + p = NULL; + } + if (!p) + p = "?"; + } + return p; +# elif __MVS__ /* z/OS */ + /* https://www.ibm.com/support/knowledgecenter/SSLTBW_2.1.0/com.ibm.zos.v2r1.bpxbd00/rtwgetp.htm */ + static char *p = "?"; + static int first = 1; + if (first) + { + pid_t pid = getpid (); + int token; + W_PSPROC buf; + first = 0; + memset (&buf, 0, sizeof(buf)); + buf.ps_cmdptr = (char *) malloc (buf.ps_cmdlen = PS_CMDBLEN_LONG); + buf.ps_conttyptr = (char *) malloc (buf.ps_conttylen = PS_CONTTYBLEN); + buf.ps_pathptr = (char *) malloc (buf.ps_pathlen = PS_PATHBLEN); + if (buf.ps_cmdptr && buf.ps_conttyptr && buf.ps_pathptr) + { + for (token = 0; token >= 0; + token = w_getpsent (token, &buf, sizeof(buf))) + { + if (token > 0 && buf.ps_pid == pid) + { + char *s = strdup (last_component (buf.ps_pathptr)); + if (s) + p = s; + break; + } + } + } + free (buf.ps_cmdptr); + free (buf.ps_conttyptr); + free (buf.ps_pathptr); + } + return p; +# elif defined __sgi /* IRIX */ + char filename[50]; + int fd; + + sprintf (filename, "/proc/pinfo/%d", (int) getpid ()); + fd = open (filename, O_RDONLY); + if (0 <= fd) + { + prpsinfo_t buf; + int ioctl_ok = 0 <= ioctl (fd, PIOCPSINFO, &buf); + close (fd); + if (ioctl_ok) + { + char *name = buf.pr_fname; + size_t namesize = sizeof buf.pr_fname; + char *namenul = memchr (name, '\0', namesize); + size_t namelen = namenul ? namenul - name : namesize; + char *namecopy = malloc (namelen + 1); + if (namecopy) + { + namecopy[namelen] = 0; + return memcpy (namecopy, name, namelen); + } + } + } + return NULL; +# else +# error "getprogname module not ported to this OS" +# endif +} + +#endif + +/* + * Hey Emacs! + * Local Variables: + * coding: utf-8 + * End: + */ diff --git a/lib/getprogname.h b/lib/getprogname.h new file mode 100644 index 0000000..ee6688f --- /dev/null +++ b/lib/getprogname.h @@ -0,0 +1,40 @@ +/* Program name management. + Copyright (C) 2016-2018 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +#ifndef _GL_GETPROGNAME_H +#define _GL_GETPROGNAME_H + +#include <stdlib.h> + +#ifdef __cplusplus +extern "C" { +#endif + +/* Return the base name of the executing program. + On native Windows this will usually end in ".exe" or ".EXE". */ +#ifndef HAVE_GETPROGNAME +extern char const *getprogname (void) +# ifdef HAVE_DECL_PROGRAM_INVOCATION_NAME + _GL_ATTRIBUTE_PURE +# endif + ; +#endif + +#ifdef __cplusplus +} +#endif + +#endif diff --git a/lib/gettext.h b/lib/gettext.h new file mode 100644 index 0000000..8eef008 --- /dev/null +++ b/lib/gettext.h @@ -0,0 +1,294 @@ +/* Convenience header for conditional use of GNU <libintl.h>. + Copyright (C) 1995-1998, 2000-2002, 2004-2006, 2009-2018 Free Software + Foundation, Inc. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3, or (at your option) + any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License along + with this program; if not, see <https://www.gnu.org/licenses/>. */ + +#ifndef _LIBGETTEXT_H +#define _LIBGETTEXT_H 1 + +/* NLS can be disabled through the configure --disable-nls option + or through "#define ENABLE NLS 0" before including this file. */ +#if defined ENABLE_NLS && ENABLE_NLS + +/* Get declarations of GNU message catalog functions. */ +# include <libintl.h> + +/* You can set the DEFAULT_TEXT_DOMAIN macro to specify the domain used by + the gettext() and ngettext() macros. This is an alternative to calling + textdomain(), and is useful for libraries. */ +# ifdef DEFAULT_TEXT_DOMAIN +# undef gettext +# define gettext(Msgid) \ + dgettext (DEFAULT_TEXT_DOMAIN, Msgid) +# undef ngettext +# define ngettext(Msgid1, Msgid2, N) \ + dngettext (DEFAULT_TEXT_DOMAIN, Msgid1, Msgid2, N) +# endif + +#else + +/* Solaris /usr/include/locale.h includes /usr/include/libintl.h, which + chokes if dcgettext is defined as a macro. So include it now, to make + later inclusions of <locale.h> a NOP. We don't include <libintl.h> + as well because people using "gettext.h" will not include <libintl.h>, + and also including <libintl.h> would fail on SunOS 4, whereas <locale.h> + is OK. */ +#if defined(__sun) +# include <locale.h> +#endif + +/* Many header files from the libstdc++ coming with g++ 3.3 or newer include + <libintl.h>, which chokes if dcgettext is defined as a macro. So include + it now, to make later inclusions of <libintl.h> a NOP. */ +#if defined(__cplusplus) && defined(__GNUG__) && (__GNUC__ >= 3) +# include <cstdlib> +# if (__GLIBC__ >= 2 && !defined __UCLIBC__) || _GLIBCXX_HAVE_LIBINTL_H +# include <libintl.h> +# endif +#endif + +/* Disabled NLS. + The casts to 'const char *' serve the purpose of producing warnings + for invalid uses of the value returned from these functions. + On pre-ANSI systems without 'const', the config.h file is supposed to + contain "#define const". */ +# undef gettext +# define gettext(Msgid) ((const char *) (Msgid)) +# undef dgettext +# define dgettext(Domainname, Msgid) ((void) (Domainname), gettext (Msgid)) +# undef dcgettext +# define dcgettext(Domainname, Msgid, Category) \ + ((void) (Category), dgettext (Domainname, Msgid)) +# undef ngettext +# define ngettext(Msgid1, Msgid2, N) \ + ((N) == 1 \ + ? ((void) (Msgid2), (const char *) (Msgid1)) \ + : ((void) (Msgid1), (const char *) (Msgid2))) +# undef dngettext +# define dngettext(Domainname, Msgid1, Msgid2, N) \ + ((void) (Domainname), ngettext (Msgid1, Msgid2, N)) +# undef dcngettext +# define dcngettext(Domainname, Msgid1, Msgid2, N, Category) \ + ((void) (Category), dngettext (Domainname, Msgid1, Msgid2, N)) +# undef textdomain +# define textdomain(Domainname) ((const char *) (Domainname)) +# undef bindtextdomain +# define bindtextdomain(Domainname, Dirname) \ + ((void) (Domainname), (const char *) (Dirname)) +# undef bind_textdomain_codeset +# define bind_textdomain_codeset(Domainname, Codeset) \ + ((void) (Domainname), (const char *) (Codeset)) + +#endif + +/* Prefer gnulib's setlocale override over libintl's setlocale override. */ +#ifdef GNULIB_defined_setlocale +# undef setlocale +# define setlocale rpl_setlocale +#endif + +/* A pseudo function call that serves as a marker for the automated + extraction of messages, but does not call gettext(). The run-time + translation is done at a different place in the code. + The argument, String, should be a literal string. Concatenated strings + and other string expressions won't work. + The macro's expansion is not parenthesized, so that it is suitable as + initializer for static 'char[]' or 'const char[]' variables. */ +#define gettext_noop(String) String + +/* The separator between msgctxt and msgid in a .mo file. */ +#define GETTEXT_CONTEXT_GLUE "\004" + +/* Pseudo function calls, taking a MSGCTXT and a MSGID instead of just a + MSGID. MSGCTXT and MSGID must be string literals. MSGCTXT should be + short and rarely need to change. + The letter 'p' stands for 'particular' or 'special'. */ +#ifdef DEFAULT_TEXT_DOMAIN +# define pgettext(Msgctxt, Msgid) \ + pgettext_aux (DEFAULT_TEXT_DOMAIN, Msgctxt GETTEXT_CONTEXT_GLUE Msgid, Msgid, LC_MESSAGES) +#else +# define pgettext(Msgctxt, Msgid) \ + pgettext_aux (NULL, Msgctxt GETTEXT_CONTEXT_GLUE Msgid, Msgid, LC_MESSAGES) +#endif +#define dpgettext(Domainname, Msgctxt, Msgid) \ + pgettext_aux (Domainname, Msgctxt GETTEXT_CONTEXT_GLUE Msgid, Msgid, LC_MESSAGES) +#define dcpgettext(Domainname, Msgctxt, Msgid, Category) \ + pgettext_aux (Domainname, Msgctxt GETTEXT_CONTEXT_GLUE Msgid, Msgid, Category) +#ifdef DEFAULT_TEXT_DOMAIN +# define npgettext(Msgctxt, Msgid, MsgidPlural, N) \ + npgettext_aux (DEFAULT_TEXT_DOMAIN, Msgctxt GETTEXT_CONTEXT_GLUE Msgid, Msgid, MsgidPlural, N, LC_MESSAGES) +#else +# define npgettext(Msgctxt, Msgid, MsgidPlural, N) \ + npgettext_aux (NULL, Msgctxt GETTEXT_CONTEXT_GLUE Msgid, Msgid, MsgidPlural, N, LC_MESSAGES) +#endif +#define dnpgettext(Domainname, Msgctxt, Msgid, MsgidPlural, N) \ + npgettext_aux (Domainname, Msgctxt GETTEXT_CONTEXT_GLUE Msgid, Msgid, MsgidPlural, N, LC_MESSAGES) +#define dcnpgettext(Domainname, Msgctxt, Msgid, MsgidPlural, N, Category) \ + npgettext_aux (Domainname, Msgctxt GETTEXT_CONTEXT_GLUE Msgid, Msgid, MsgidPlural, N, Category) + +#ifdef __GNUC__ +__inline +#else +#ifdef __cplusplus +inline +#endif +#endif +static const char * +pgettext_aux (const char *domain, + const char *msg_ctxt_id, const char *msgid, + int category) +{ + const char *translation = dcgettext (domain, msg_ctxt_id, category); + if (translation == msg_ctxt_id) + return msgid; + else + return translation; +} + +#ifdef __GNUC__ +__inline +#else +#ifdef __cplusplus +inline +#endif +#endif +static const char * +npgettext_aux (const char *domain, + const char *msg_ctxt_id, const char *msgid, + const char *msgid_plural, unsigned long int n, + int category) +{ + const char *translation = + dcngettext (domain, msg_ctxt_id, msgid_plural, n, category); + if (translation == msg_ctxt_id || translation == msgid_plural) + return (n == 1 ? msgid : msgid_plural); + else + return translation; +} + +/* The same thing extended for non-constant arguments. Here MSGCTXT and MSGID + can be arbitrary expressions. But for string literals these macros are + less efficient than those above. */ + +#include <string.h> + +#if (((__GNUC__ >= 3 || __GNUG__ >= 2) && !defined __STRICT_ANSI__) \ + /* || (__STDC_VERSION__ == 199901L && !defined __HP_cc) + || (__STDC_VERSION__ >= 201112L && !defined __STDC_NO_VLA__) */ ) +# define _LIBGETTEXT_HAVE_VARIABLE_SIZE_ARRAYS 1 +#else +# define _LIBGETTEXT_HAVE_VARIABLE_SIZE_ARRAYS 0 +#endif + +#if !_LIBGETTEXT_HAVE_VARIABLE_SIZE_ARRAYS +#include <stdlib.h> +#endif + +#define pgettext_expr(Msgctxt, Msgid) \ + dcpgettext_expr (NULL, Msgctxt, Msgid, LC_MESSAGES) +#define dpgettext_expr(Domainname, Msgctxt, Msgid) \ + dcpgettext_expr (Domainname, Msgctxt, Msgid, LC_MESSAGES) + +#ifdef __GNUC__ +__inline +#else +#ifdef __cplusplus +inline +#endif +#endif +static const char * +dcpgettext_expr (const char *domain, + const char *msgctxt, const char *msgid, + int category) +{ + size_t msgctxt_len = strlen (msgctxt) + 1; + size_t msgid_len = strlen (msgid) + 1; + const char *translation; +#if _LIBGETTEXT_HAVE_VARIABLE_SIZE_ARRAYS + char msg_ctxt_id[msgctxt_len + msgid_len]; +#else + char buf[1024]; + char *msg_ctxt_id = + (msgctxt_len + msgid_len <= sizeof (buf) + ? buf + : (char *) malloc (msgctxt_len + msgid_len)); + if (msg_ctxt_id != NULL) +#endif + { + int found_translation; + memcpy (msg_ctxt_id, msgctxt, msgctxt_len - 1); + msg_ctxt_id[msgctxt_len - 1] = '\004'; + memcpy (msg_ctxt_id + msgctxt_len, msgid, msgid_len); + translation = dcgettext (domain, msg_ctxt_id, category); + found_translation = (translation != msg_ctxt_id); +#if !_LIBGETTEXT_HAVE_VARIABLE_SIZE_ARRAYS + if (msg_ctxt_id != buf) + free (msg_ctxt_id); +#endif + if (found_translation) + return translation; + } + return msgid; +} + +#define npgettext_expr(Msgctxt, Msgid, MsgidPlural, N) \ + dcnpgettext_expr (NULL, Msgctxt, Msgid, MsgidPlural, N, LC_MESSAGES) +#define dnpgettext_expr(Domainname, Msgctxt, Msgid, MsgidPlural, N) \ + dcnpgettext_expr (Domainname, Msgctxt, Msgid, MsgidPlural, N, LC_MESSAGES) + +#ifdef __GNUC__ +__inline +#else +#ifdef __cplusplus +inline +#endif +#endif +static const char * +dcnpgettext_expr (const char *domain, + const char *msgctxt, const char *msgid, + const char *msgid_plural, unsigned long int n, + int category) +{ + size_t msgctxt_len = strlen (msgctxt) + 1; + size_t msgid_len = strlen (msgid) + 1; + const char *translation; +#if _LIBGETTEXT_HAVE_VARIABLE_SIZE_ARRAYS + char msg_ctxt_id[msgctxt_len + msgid_len]; +#else + char buf[1024]; + char *msg_ctxt_id = + (msgctxt_len + msgid_len <= sizeof (buf) + ? buf + : (char *) malloc (msgctxt_len + msgid_len)); + if (msg_ctxt_id != NULL) +#endif + { + int found_translation; + memcpy (msg_ctxt_id, msgctxt, msgctxt_len - 1); + msg_ctxt_id[msgctxt_len - 1] = '\004'; + memcpy (msg_ctxt_id + msgctxt_len, msgid, msgid_len); + translation = dcngettext (domain, msg_ctxt_id, msgid_plural, n, category); + found_translation = !(translation == msg_ctxt_id || translation == msgid_plural); +#if !_LIBGETTEXT_HAVE_VARIABLE_SIZE_ARRAYS + if (msg_ctxt_id != buf) + free (msg_ctxt_id); +#endif + if (found_translation) + return translation; + } + return (n == 1 ? msgid : msgid_plural); +} + +#endif /* _LIBGETTEXT_H */ diff --git a/lib/gettimeofday.c b/lib/gettimeofday.c new file mode 100644 index 0000000..1bd50fa --- /dev/null +++ b/lib/gettimeofday.c @@ -0,0 +1,150 @@ +/* Provide gettimeofday for systems that don't have it or for which it's broken. + + Copyright (C) 2001-2003, 2005-2007, 2009-2018 Free Software Foundation, Inc. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3, or (at your option) + any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, see <https://www.gnu.org/licenses/>. */ + +/* written by Jim Meyering */ + +#include <config.h> + +/* Specification. */ +#include <sys/time.h> + +#include <time.h> + +#if defined _WIN32 && ! defined __CYGWIN__ +# define WINDOWS_NATIVE +# include <windows.h> +#endif + +#include "localtime-buffer.h" + +#ifdef WINDOWS_NATIVE + +/* Avoid warnings from gcc -Wcast-function-type. */ +# define GetProcAddress \ + (void *) GetProcAddress + +/* GetSystemTimePreciseAsFileTime was introduced only in Windows 8. */ +typedef void (WINAPI * GetSystemTimePreciseAsFileTimeFuncType) (FILETIME *lpTime); +static GetSystemTimePreciseAsFileTimeFuncType GetSystemTimePreciseAsFileTimeFunc = NULL; +static BOOL initialized = FALSE; + +static void +initialize (void) +{ + HMODULE kernel32 = LoadLibrary ("kernel32.dll"); + if (kernel32 != NULL) + { + GetSystemTimePreciseAsFileTimeFunc = + (GetSystemTimePreciseAsFileTimeFuncType) GetProcAddress (kernel32, "GetSystemTimePreciseAsFileTime"); + } + initialized = TRUE; +} + +#endif + +/* This is a wrapper for gettimeofday. It is used only on systems + that lack this function, or whose implementation of this function + causes problems. + Work around the bug in some systems whereby gettimeofday clobbers + the static buffer that localtime uses for its return value. The + gettimeofday function from Mac OS X 10.0.4 (i.e., Darwin 1.3.7) has + this problem. */ + +int +gettimeofday (struct timeval *restrict tv, void *restrict tz) +{ +#undef gettimeofday +#ifdef WINDOWS_NATIVE + + /* On native Windows, there are two ways to get the current time: + GetSystemTimeAsFileTime + <https://msdn.microsoft.com/en-us/library/ms724397.aspx> + or + GetSystemTimePreciseAsFileTime + <https://msdn.microsoft.com/en-us/library/hh706895.aspx>. + GetSystemTimeAsFileTime produces values that jump by increments of + 15.627 milliseconds (!) on average. + Whereas GetSystemTimePreciseAsFileTime values usually jump by 1 or 2 + microseconds. + More discussion on this topic: + <http://www.windowstimestamp.com/description>. */ + FILETIME current_time; + + if (!initialized) + initialize (); + if (GetSystemTimePreciseAsFileTimeFunc != NULL) + GetSystemTimePreciseAsFileTimeFunc (¤t_time); + else + GetSystemTimeAsFileTime (¤t_time); + + /* Convert from FILETIME to 'struct timeval'. */ + /* FILETIME: <https://msdn.microsoft.com/en-us/library/ms724284.aspx> */ + ULONGLONG since_1601 = + ((ULONGLONG) current_time.dwHighDateTime << 32) + | (ULONGLONG) current_time.dwLowDateTime; + /* Between 1601-01-01 and 1970-01-01 there were 280 normal years and 89 leap + years, in total 134774 days. */ + ULONGLONG since_1970 = + since_1601 - (ULONGLONG) 134774 * (ULONGLONG) 86400 * (ULONGLONG) 10000000; + ULONGLONG microseconds_since_1970 = since_1970 / (ULONGLONG) 10; + tv->tv_sec = microseconds_since_1970 / (ULONGLONG) 1000000; + tv->tv_usec = microseconds_since_1970 % (ULONGLONG) 1000000; + + return 0; + +#else + +# if HAVE_GETTIMEOFDAY +# if GETTIMEOFDAY_CLOBBERS_LOCALTIME + /* Save and restore the contents of the buffer used for localtime's + result around the call to gettimeofday. */ + struct tm save = *localtime_buffer_addr; +# endif + +# if defined timeval /* 'struct timeval' overridden by gnulib? */ +# undef timeval + struct timeval otv; + int result = gettimeofday (&otv, (struct timezone *) tz); + if (result == 0) + { + tv->tv_sec = otv.tv_sec; + tv->tv_usec = otv.tv_usec; + } +# else + int result = gettimeofday (tv, (struct timezone *) tz); +# endif + +# if GETTIMEOFDAY_CLOBBERS_LOCALTIME + *localtime_buffer_addr = save; +# endif + + return result; + +# else + +# if !defined OK_TO_USE_1S_CLOCK +# error "Only 1-second nominal clock resolution found. Is that intended?" \ + "If so, compile with the -DOK_TO_USE_1S_CLOCK option." +# endif + tv->tv_sec = time (NULL); + tv->tv_usec = 0; + + return 0; + +# endif +#endif +} diff --git a/lib/glthread/lock.c b/lib/glthread/lock.c new file mode 100644 index 0000000..f166d7c --- /dev/null +++ b/lib/glthread/lock.c @@ -0,0 +1,1221 @@ +/* Locking in multithreaded situations. + Copyright (C) 2005-2018 Free Software Foundation, Inc. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3, or (at your option) + any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, see <https://www.gnu.org/licenses/>. */ + +/* Written by Bruno Haible <bruno@clisp.org>, 2005. + Based on GCC's gthr-posix.h, gthr-posix95.h, gthr-solaris.h, + gthr-win32.h. */ + +#include <config.h> + +#include "glthread/lock.h" + +/* ========================================================================= */ + +#if USE_POSIX_THREADS + +/* -------------------------- gl_lock_t datatype -------------------------- */ + +/* ------------------------- gl_rwlock_t datatype ------------------------- */ + +# if HAVE_PTHREAD_RWLOCK && (HAVE_PTHREAD_RWLOCK_RDLOCK_PREFER_WRITER || (defined PTHREAD_RWLOCK_WRITER_NONRECURSIVE_INITIALIZER_NP && (__GNU_LIBRARY__ > 1))) + +# ifdef PTHREAD_RWLOCK_INITIALIZER + +# if !HAVE_PTHREAD_RWLOCK_RDLOCK_PREFER_WRITER + /* glibc with bug https://sourceware.org/bugzilla/show_bug.cgi?id=13701 */ + +int +glthread_rwlock_init_for_glibc (pthread_rwlock_t *lock) +{ + pthread_rwlockattr_t attributes; + int err; + + err = pthread_rwlockattr_init (&attributes); + if (err != 0) + return err; + /* Note: PTHREAD_RWLOCK_PREFER_WRITER_NONRECURSIVE_NP is the only value that + causes the writer to be preferred. PTHREAD_RWLOCK_PREFER_WRITER_NP does not + do this; see + http://man7.org/linux/man-pages/man3/pthread_rwlockattr_setkind_np.3.html */ + err = pthread_rwlockattr_setkind_np (&attributes, + PTHREAD_RWLOCK_PREFER_WRITER_NONRECURSIVE_NP); + if (err == 0) + err = pthread_rwlock_init(lock, &attributes); + /* pthread_rwlockattr_destroy always returns 0. It cannot influence the + return value. */ + pthread_rwlockattr_destroy (&attributes); + return err; +} + +# endif +# else + +int +glthread_rwlock_init_multithreaded (gl_rwlock_t *lock) +{ + int err; + + err = pthread_rwlock_init (&lock->rwlock, NULL); + if (err != 0) + return err; + lock->initialized = 1; + return 0; +} + +int +glthread_rwlock_rdlock_multithreaded (gl_rwlock_t *lock) +{ + if (!lock->initialized) + { + int err; + + err = pthread_mutex_lock (&lock->guard); + if (err != 0) + return err; + if (!lock->initialized) + { + err = glthread_rwlock_init_multithreaded (lock); + if (err != 0) + { + pthread_mutex_unlock (&lock->guard); + return err; + } + } + err = pthread_mutex_unlock (&lock->guard); + if (err != 0) + return err; + } + return pthread_rwlock_rdlock (&lock->rwlock); +} + +int +glthread_rwlock_wrlock_multithreaded (gl_rwlock_t *lock) +{ + if (!lock->initialized) + { + int err; + + err = pthread_mutex_lock (&lock->guard); + if (err != 0) + return err; + if (!lock->initialized) + { + err = glthread_rwlock_init_multithreaded (lock); + if (err != 0) + { + pthread_mutex_unlock (&lock->guard); + return err; + } + } + err = pthread_mutex_unlock (&lock->guard); + if (err != 0) + return err; + } + return pthread_rwlock_wrlock (&lock->rwlock); +} + +int +glthread_rwlock_unlock_multithreaded (gl_rwlock_t *lock) +{ + if (!lock->initialized) + return EINVAL; + return pthread_rwlock_unlock (&lock->rwlock); +} + +int +glthread_rwlock_destroy_multithreaded (gl_rwlock_t *lock) +{ + int err; + + if (!lock->initialized) + return EINVAL; + err = pthread_rwlock_destroy (&lock->rwlock); + if (err != 0) + return err; + lock->initialized = 0; + return 0; +} + +# endif + +# else + +int +glthread_rwlock_init_multithreaded (gl_rwlock_t *lock) +{ + int err; + + err = pthread_mutex_init (&lock->lock, NULL); + if (err != 0) + return err; + err = pthread_cond_init (&lock->waiting_readers, NULL); + if (err != 0) + return err; + err = pthread_cond_init (&lock->waiting_writers, NULL); + if (err != 0) + return err; + lock->waiting_writers_count = 0; + lock->runcount = 0; + return 0; +} + +int +glthread_rwlock_rdlock_multithreaded (gl_rwlock_t *lock) +{ + int err; + + err = pthread_mutex_lock (&lock->lock); + if (err != 0) + return err; + /* Test whether only readers are currently running, and whether the runcount + field will not overflow, and whether no writer is waiting. The latter + condition is because POSIX recommends that "write locks shall take + precedence over read locks", to avoid "writer starvation". */ + while (!(lock->runcount + 1 > 0 && lock->waiting_writers_count == 0)) + { + /* This thread has to wait for a while. Enqueue it among the + waiting_readers. */ + err = pthread_cond_wait (&lock->waiting_readers, &lock->lock); + if (err != 0) + { + pthread_mutex_unlock (&lock->lock); + return err; + } + } + lock->runcount++; + return pthread_mutex_unlock (&lock->lock); +} + +int +glthread_rwlock_wrlock_multithreaded (gl_rwlock_t *lock) +{ + int err; + + err = pthread_mutex_lock (&lock->lock); + if (err != 0) + return err; + /* Test whether no readers or writers are currently running. */ + while (!(lock->runcount == 0)) + { + /* This thread has to wait for a while. Enqueue it among the + waiting_writers. */ + lock->waiting_writers_count++; + err = pthread_cond_wait (&lock->waiting_writers, &lock->lock); + if (err != 0) + { + lock->waiting_writers_count--; + pthread_mutex_unlock (&lock->lock); + return err; + } + lock->waiting_writers_count--; + } + lock->runcount--; /* runcount becomes -1 */ + return pthread_mutex_unlock (&lock->lock); +} + +int +glthread_rwlock_unlock_multithreaded (gl_rwlock_t *lock) +{ + int err; + + err = pthread_mutex_lock (&lock->lock); + if (err != 0) + return err; + if (lock->runcount < 0) + { + /* Drop a writer lock. */ + if (!(lock->runcount == -1)) + { + pthread_mutex_unlock (&lock->lock); + return EINVAL; + } + lock->runcount = 0; + } + else + { + /* Drop a reader lock. */ + if (!(lock->runcount > 0)) + { + pthread_mutex_unlock (&lock->lock); + return EINVAL; + } + lock->runcount--; + } + if (lock->runcount == 0) + { + /* POSIX recommends that "write locks shall take precedence over read + locks", to avoid "writer starvation". */ + if (lock->waiting_writers_count > 0) + { + /* Wake up one of the waiting writers. */ + err = pthread_cond_signal (&lock->waiting_writers); + if (err != 0) + { + pthread_mutex_unlock (&lock->lock); + return err; + } + } + else + { + /* Wake up all waiting readers. */ + err = pthread_cond_broadcast (&lock->waiting_readers); + if (err != 0) + { + pthread_mutex_unlock (&lock->lock); + return err; + } + } + } + return pthread_mutex_unlock (&lock->lock); +} + +int +glthread_rwlock_destroy_multithreaded (gl_rwlock_t *lock) +{ + int err; + + err = pthread_mutex_destroy (&lock->lock); + if (err != 0) + return err; + err = pthread_cond_destroy (&lock->waiting_readers); + if (err != 0) + return err; + err = pthread_cond_destroy (&lock->waiting_writers); + if (err != 0) + return err; + return 0; +} + +# endif + +/* --------------------- gl_recursive_lock_t datatype --------------------- */ + +# if HAVE_PTHREAD_MUTEX_RECURSIVE + +# if defined PTHREAD_RECURSIVE_MUTEX_INITIALIZER || defined PTHREAD_RECURSIVE_MUTEX_INITIALIZER_NP + +int +glthread_recursive_lock_init_multithreaded (gl_recursive_lock_t *lock) +{ + pthread_mutexattr_t attributes; + int err; + + err = pthread_mutexattr_init (&attributes); + if (err != 0) + return err; + err = pthread_mutexattr_settype (&attributes, PTHREAD_MUTEX_RECURSIVE); + if (err != 0) + { + pthread_mutexattr_destroy (&attributes); + return err; + } + err = pthread_mutex_init (lock, &attributes); + if (err != 0) + { + pthread_mutexattr_destroy (&attributes); + return err; + } + err = pthread_mutexattr_destroy (&attributes); + if (err != 0) + return err; + return 0; +} + +# else + +int +glthread_recursive_lock_init_multithreaded (gl_recursive_lock_t *lock) +{ + pthread_mutexattr_t attributes; + int err; + + err = pthread_mutexattr_init (&attributes); + if (err != 0) + return err; + err = pthread_mutexattr_settype (&attributes, PTHREAD_MUTEX_RECURSIVE); + if (err != 0) + { + pthread_mutexattr_destroy (&attributes); + return err; + } + err = pthread_mutex_init (&lock->recmutex, &attributes); + if (err != 0) + { + pthread_mutexattr_destroy (&attributes); + return err; + } + err = pthread_mutexattr_destroy (&attributes); + if (err != 0) + return err; + lock->initialized = 1; + return 0; +} + +int +glthread_recursive_lock_lock_multithreaded (gl_recursive_lock_t *lock) +{ + if (!lock->initialized) + { + int err; + + err = pthread_mutex_lock (&lock->guard); + if (err != 0) + return err; + if (!lock->initialized) + { + err = glthread_recursive_lock_init_multithreaded (lock); + if (err != 0) + { + pthread_mutex_unlock (&lock->guard); + return err; + } + } + err = pthread_mutex_unlock (&lock->guard); + if (err != 0) + return err; + } + return pthread_mutex_lock (&lock->recmutex); +} + +int +glthread_recursive_lock_unlock_multithreaded (gl_recursive_lock_t *lock) +{ + if (!lock->initialized) + return EINVAL; + return pthread_mutex_unlock (&lock->recmutex); +} + +int +glthread_recursive_lock_destroy_multithreaded (gl_recursive_lock_t *lock) +{ + int err; + + if (!lock->initialized) + return EINVAL; + err = pthread_mutex_destroy (&lock->recmutex); + if (err != 0) + return err; + lock->initialized = 0; + return 0; +} + +# endif + +# else + +int +glthread_recursive_lock_init_multithreaded (gl_recursive_lock_t *lock) +{ + int err; + + err = pthread_mutex_init (&lock->mutex, NULL); + if (err != 0) + return err; + lock->owner = (pthread_t) 0; + lock->depth = 0; + return 0; +} + +int +glthread_recursive_lock_lock_multithreaded (gl_recursive_lock_t *lock) +{ + pthread_t self = pthread_self (); + if (lock->owner != self) + { + int err; + + err = pthread_mutex_lock (&lock->mutex); + if (err != 0) + return err; + lock->owner = self; + } + if (++(lock->depth) == 0) /* wraparound? */ + { + lock->depth--; + return EAGAIN; + } + return 0; +} + +int +glthread_recursive_lock_unlock_multithreaded (gl_recursive_lock_t *lock) +{ + if (lock->owner != pthread_self ()) + return EPERM; + if (lock->depth == 0) + return EINVAL; + if (--(lock->depth) == 0) + { + lock->owner = (pthread_t) 0; + return pthread_mutex_unlock (&lock->mutex); + } + else + return 0; +} + +int +glthread_recursive_lock_destroy_multithreaded (gl_recursive_lock_t *lock) +{ + if (lock->owner != (pthread_t) 0) + return EBUSY; + return pthread_mutex_destroy (&lock->mutex); +} + +# endif + +/* -------------------------- gl_once_t datatype -------------------------- */ + +static const pthread_once_t fresh_once = PTHREAD_ONCE_INIT; + +int +glthread_once_singlethreaded (pthread_once_t *once_control) +{ + /* We don't know whether pthread_once_t is an integer type, a floating-point + type, a pointer type, or a structure type. */ + char *firstbyte = (char *)once_control; + if (*firstbyte == *(const char *)&fresh_once) + { + /* First time use of once_control. Invert the first byte. */ + *firstbyte = ~ *(const char *)&fresh_once; + return 1; + } + else + return 0; +} + +#endif + +/* ========================================================================= */ + +#if USE_PTH_THREADS + +/* Use the GNU Pth threads library. */ + +/* -------------------------- gl_lock_t datatype -------------------------- */ + +/* ------------------------- gl_rwlock_t datatype ------------------------- */ + +# if !HAVE_PTH_RWLOCK_ACQUIRE_PREFER_WRITER + +int +glthread_rwlock_init_multithreaded (gl_rwlock_t *lock) +{ + if (!pth_mutex_init (&lock->lock)) + return errno; + if (!pth_cond_init (&lock->waiting_readers)) + return errno; + if (!pth_cond_init (&lock->waiting_writers)) + return errno; + lock->waiting_writers_count = 0; + lock->runcount = 0; + lock->initialized = 1; + return 0; +} + +int +glthread_rwlock_rdlock_multithreaded (gl_rwlock_t *lock) +{ + if (!lock->initialized) + glthread_rwlock_init_multithreaded (lock); + if (!pth_mutex_acquire (&lock->lock, 0, NULL)) + return errno; + /* Test whether only readers are currently running, and whether the runcount + field will not overflow, and whether no writer is waiting. The latter + condition is because POSIX recommends that "write locks shall take + precedence over read locks", to avoid "writer starvation". */ + while (!(lock->runcount + 1 > 0 && lock->waiting_writers_count == 0)) + { + /* This thread has to wait for a while. Enqueue it among the + waiting_readers. */ + if (!pth_cond_await (&lock->waiting_readers, &lock->lock, NULL)) + { + int err = errno; + pth_mutex_release (&lock->lock); + return err; + } + } + lock->runcount++; + return (!pth_mutex_release (&lock->lock) ? errno : 0); +} + +int +glthread_rwlock_wrlock_multithreaded (gl_rwlock_t *lock) +{ + if (!lock->initialized) + glthread_rwlock_init_multithreaded (lock); + if (!pth_mutex_acquire (&lock->lock, 0, NULL)) + return errno; + /* Test whether no readers or writers are currently running. */ + while (!(lock->runcount == 0)) + { + /* This thread has to wait for a while. Enqueue it among the + waiting_writers. */ + lock->waiting_writers_count++; + if (!pth_cond_await (&lock->waiting_writers, &lock->lock, NULL)) + { + int err = errno; + lock->waiting_writers_count--; + pth_mutex_release (&lock->lock); + return err; + } + lock->waiting_writers_count--; + } + lock->runcount--; /* runcount becomes -1 */ + return (!pth_mutex_release (&lock->lock) ? errno : 0); +} + +int +glthread_rwlock_unlock_multithreaded (gl_rwlock_t *lock) +{ + int err; + + if (!lock->initialized) + return EINVAL; + if (!pth_mutex_acquire (&lock->lock, 0, NULL)) + return errno; + if (lock->runcount < 0) + { + /* Drop a writer lock. */ + if (!(lock->runcount == -1)) + { + pth_mutex_release (&lock->lock); + return EINVAL; + } + lock->runcount = 0; + } + else + { + /* Drop a reader lock. */ + if (!(lock->runcount > 0)) + { + pth_mutex_release (&lock->lock); + return EINVAL; + } + lock->runcount--; + } + if (lock->runcount == 0) + { + /* POSIX recommends that "write locks shall take precedence over read + locks", to avoid "writer starvation". */ + if (lock->waiting_writers_count > 0) + { + /* Wake up one of the waiting writers. */ + if (!pth_cond_notify (&lock->waiting_writers, FALSE)) + { + int err = errno; + pth_mutex_release (&lock->lock); + return err; + } + } + else + { + /* Wake up all waiting readers. */ + if (!pth_cond_notify (&lock->waiting_readers, TRUE)) + { + int err = errno; + pth_mutex_release (&lock->lock); + return err; + } + } + } + return (!pth_mutex_release (&lock->lock) ? errno : 0); +} + +int +glthread_rwlock_destroy_multithreaded (gl_rwlock_t *lock) +{ + lock->initialized = 0; + return 0; +} + +# endif + +/* --------------------- gl_recursive_lock_t datatype --------------------- */ + +/* -------------------------- gl_once_t datatype -------------------------- */ + +static void +glthread_once_call (void *arg) +{ + void (**gl_once_temp_addr) (void) = (void (**) (void)) arg; + void (*initfunction) (void) = *gl_once_temp_addr; + initfunction (); +} + +int +glthread_once_multithreaded (pth_once_t *once_control, void (*initfunction) (void)) +{ + void (*temp) (void) = initfunction; + return (!pth_once (once_control, glthread_once_call, &temp) ? errno : 0); +} + +int +glthread_once_singlethreaded (pth_once_t *once_control) +{ + /* We know that pth_once_t is an integer type. */ + if (*once_control == PTH_ONCE_INIT) + { + /* First time use of once_control. Invert the marker. */ + *once_control = ~ PTH_ONCE_INIT; + return 1; + } + else + return 0; +} + +#endif + +/* ========================================================================= */ + +#if USE_SOLARIS_THREADS + +/* Use the old Solaris threads library. */ + +/* -------------------------- gl_lock_t datatype -------------------------- */ + +/* ------------------------- gl_rwlock_t datatype ------------------------- */ + +/* --------------------- gl_recursive_lock_t datatype --------------------- */ + +int +glthread_recursive_lock_init_multithreaded (gl_recursive_lock_t *lock) +{ + int err; + + err = mutex_init (&lock->mutex, USYNC_THREAD, NULL); + if (err != 0) + return err; + lock->owner = (thread_t) 0; + lock->depth = 0; + return 0; +} + +int +glthread_recursive_lock_lock_multithreaded (gl_recursive_lock_t *lock) +{ + thread_t self = thr_self (); + if (lock->owner != self) + { + int err; + + err = mutex_lock (&lock->mutex); + if (err != 0) + return err; + lock->owner = self; + } + if (++(lock->depth) == 0) /* wraparound? */ + { + lock->depth--; + return EAGAIN; + } + return 0; +} + +int +glthread_recursive_lock_unlock_multithreaded (gl_recursive_lock_t *lock) +{ + if (lock->owner != thr_self ()) + return EPERM; + if (lock->depth == 0) + return EINVAL; + if (--(lock->depth) == 0) + { + lock->owner = (thread_t) 0; + return mutex_unlock (&lock->mutex); + } + else + return 0; +} + +int +glthread_recursive_lock_destroy_multithreaded (gl_recursive_lock_t *lock) +{ + if (lock->owner != (thread_t) 0) + return EBUSY; + return mutex_destroy (&lock->mutex); +} + +/* -------------------------- gl_once_t datatype -------------------------- */ + +int +glthread_once_multithreaded (gl_once_t *once_control, void (*initfunction) (void)) +{ + if (!once_control->inited) + { + int err; + + /* Use the mutex to guarantee that if another thread is already calling + the initfunction, this thread waits until it's finished. */ + err = mutex_lock (&once_control->mutex); + if (err != 0) + return err; + if (!once_control->inited) + { + once_control->inited = 1; + initfunction (); + } + return mutex_unlock (&once_control->mutex); + } + else + return 0; +} + +int +glthread_once_singlethreaded (gl_once_t *once_control) +{ + /* We know that gl_once_t contains an integer type. */ + if (!once_control->inited) + { + /* First time use of once_control. Invert the marker. */ + once_control->inited = ~ 0; + return 1; + } + else + return 0; +} + +#endif + +/* ========================================================================= */ + +#if USE_WINDOWS_THREADS + +/* -------------------------- gl_lock_t datatype -------------------------- */ + +void +glthread_lock_init_func (gl_lock_t *lock) +{ + InitializeCriticalSection (&lock->lock); + lock->guard.done = 1; +} + +int +glthread_lock_lock_func (gl_lock_t *lock) +{ + if (!lock->guard.done) + { + if (InterlockedIncrement (&lock->guard.started) == 0) + /* This thread is the first one to need this lock. Initialize it. */ + glthread_lock_init (lock); + else + /* Yield the CPU while waiting for another thread to finish + initializing this lock. */ + while (!lock->guard.done) + Sleep (0); + } + EnterCriticalSection (&lock->lock); + return 0; +} + +int +glthread_lock_unlock_func (gl_lock_t *lock) +{ + if (!lock->guard.done) + return EINVAL; + LeaveCriticalSection (&lock->lock); + return 0; +} + +int +glthread_lock_destroy_func (gl_lock_t *lock) +{ + if (!lock->guard.done) + return EINVAL; + DeleteCriticalSection (&lock->lock); + lock->guard.done = 0; + return 0; +} + +/* ------------------------- gl_rwlock_t datatype ------------------------- */ + +/* In this file, the waitqueues are implemented as circular arrays. */ +#define gl_waitqueue_t gl_carray_waitqueue_t + +static void +gl_waitqueue_init (gl_waitqueue_t *wq) +{ + wq->array = NULL; + wq->count = 0; + wq->alloc = 0; + wq->offset = 0; +} + +/* Enqueues the current thread, represented by an event, in a wait queue. + Returns INVALID_HANDLE_VALUE if an allocation failure occurs. */ +static HANDLE +gl_waitqueue_add (gl_waitqueue_t *wq) +{ + HANDLE event; + unsigned int index; + + if (wq->count == wq->alloc) + { + unsigned int new_alloc = 2 * wq->alloc + 1; + HANDLE *new_array = + (HANDLE *) realloc (wq->array, new_alloc * sizeof (HANDLE)); + if (new_array == NULL) + /* No more memory. */ + return INVALID_HANDLE_VALUE; + /* Now is a good opportunity to rotate the array so that its contents + starts at offset 0. */ + if (wq->offset > 0) + { + unsigned int old_count = wq->count; + unsigned int old_alloc = wq->alloc; + unsigned int old_offset = wq->offset; + unsigned int i; + if (old_offset + old_count > old_alloc) + { + unsigned int limit = old_offset + old_count - old_alloc; + for (i = 0; i < limit; i++) + new_array[old_alloc + i] = new_array[i]; + } + for (i = 0; i < old_count; i++) + new_array[i] = new_array[old_offset + i]; + wq->offset = 0; + } + wq->array = new_array; + wq->alloc = new_alloc; + } + /* Whether the created event is a manual-reset one or an auto-reset one, + does not matter, since we will wait on it only once. */ + event = CreateEvent (NULL, TRUE, FALSE, NULL); + if (event == INVALID_HANDLE_VALUE) + /* No way to allocate an event. */ + return INVALID_HANDLE_VALUE; + index = wq->offset + wq->count; + if (index >= wq->alloc) + index -= wq->alloc; + wq->array[index] = event; + wq->count++; + return event; +} + +/* Notifies the first thread from a wait queue and dequeues it. */ +static void +gl_waitqueue_notify_first (gl_waitqueue_t *wq) +{ + SetEvent (wq->array[wq->offset + 0]); + wq->offset++; + wq->count--; + if (wq->count == 0 || wq->offset == wq->alloc) + wq->offset = 0; +} + +/* Notifies all threads from a wait queue and dequeues them all. */ +static void +gl_waitqueue_notify_all (gl_waitqueue_t *wq) +{ + unsigned int i; + + for (i = 0; i < wq->count; i++) + { + unsigned int index = wq->offset + i; + if (index >= wq->alloc) + index -= wq->alloc; + SetEvent (wq->array[index]); + } + wq->count = 0; + wq->offset = 0; +} + +void +glthread_rwlock_init_func (gl_rwlock_t *lock) +{ + InitializeCriticalSection (&lock->lock); + gl_waitqueue_init (&lock->waiting_readers); + gl_waitqueue_init (&lock->waiting_writers); + lock->runcount = 0; + lock->guard.done = 1; +} + +int +glthread_rwlock_rdlock_func (gl_rwlock_t *lock) +{ + if (!lock->guard.done) + { + if (InterlockedIncrement (&lock->guard.started) == 0) + /* This thread is the first one to need this lock. Initialize it. */ + glthread_rwlock_init (lock); + else + /* Yield the CPU while waiting for another thread to finish + initializing this lock. */ + while (!lock->guard.done) + Sleep (0); + } + EnterCriticalSection (&lock->lock); + /* Test whether only readers are currently running, and whether the runcount + field will not overflow, and whether no writer is waiting. The latter + condition is because POSIX recommends that "write locks shall take + precedence over read locks", to avoid "writer starvation". */ + if (!(lock->runcount + 1 > 0 && lock->waiting_writers.count == 0)) + { + /* This thread has to wait for a while. Enqueue it among the + waiting_readers. */ + HANDLE event = gl_waitqueue_add (&lock->waiting_readers); + if (event != INVALID_HANDLE_VALUE) + { + DWORD result; + LeaveCriticalSection (&lock->lock); + /* Wait until another thread signals this event. */ + result = WaitForSingleObject (event, INFINITE); + if (result == WAIT_FAILED || result == WAIT_TIMEOUT) + abort (); + CloseHandle (event); + /* The thread which signalled the event already did the bookkeeping: + removed us from the waiting_readers, incremented lock->runcount. */ + if (!(lock->runcount > 0)) + abort (); + return 0; + } + else + { + /* Allocation failure. Weird. */ + do + { + LeaveCriticalSection (&lock->lock); + Sleep (1); + EnterCriticalSection (&lock->lock); + } + while (!(lock->runcount + 1 > 0)); + } + } + lock->runcount++; + LeaveCriticalSection (&lock->lock); + return 0; +} + +int +glthread_rwlock_wrlock_func (gl_rwlock_t *lock) +{ + if (!lock->guard.done) + { + if (InterlockedIncrement (&lock->guard.started) == 0) + /* This thread is the first one to need this lock. Initialize it. */ + glthread_rwlock_init (lock); + else + /* Yield the CPU while waiting for another thread to finish + initializing this lock. */ + while (!lock->guard.done) + Sleep (0); + } + EnterCriticalSection (&lock->lock); + /* Test whether no readers or writers are currently running. */ + if (!(lock->runcount == 0)) + { + /* This thread has to wait for a while. Enqueue it among the + waiting_writers. */ + HANDLE event = gl_waitqueue_add (&lock->waiting_writers); + if (event != INVALID_HANDLE_VALUE) + { + DWORD result; + LeaveCriticalSection (&lock->lock); + /* Wait until another thread signals this event. */ + result = WaitForSingleObject (event, INFINITE); + if (result == WAIT_FAILED || result == WAIT_TIMEOUT) + abort (); + CloseHandle (event); + /* The thread which signalled the event already did the bookkeeping: + removed us from the waiting_writers, set lock->runcount = -1. */ + if (!(lock->runcount == -1)) + abort (); + return 0; + } + else + { + /* Allocation failure. Weird. */ + do + { + LeaveCriticalSection (&lock->lock); + Sleep (1); + EnterCriticalSection (&lock->lock); + } + while (!(lock->runcount == 0)); + } + } + lock->runcount--; /* runcount becomes -1 */ + LeaveCriticalSection (&lock->lock); + return 0; +} + +int +glthread_rwlock_unlock_func (gl_rwlock_t *lock) +{ + if (!lock->guard.done) + return EINVAL; + EnterCriticalSection (&lock->lock); + if (lock->runcount < 0) + { + /* Drop a writer lock. */ + if (!(lock->runcount == -1)) + abort (); + lock->runcount = 0; + } + else + { + /* Drop a reader lock. */ + if (!(lock->runcount > 0)) + { + LeaveCriticalSection (&lock->lock); + return EPERM; + } + lock->runcount--; + } + if (lock->runcount == 0) + { + /* POSIX recommends that "write locks shall take precedence over read + locks", to avoid "writer starvation". */ + if (lock->waiting_writers.count > 0) + { + /* Wake up one of the waiting writers. */ + lock->runcount--; + gl_waitqueue_notify_first (&lock->waiting_writers); + } + else + { + /* Wake up all waiting readers. */ + lock->runcount += lock->waiting_readers.count; + gl_waitqueue_notify_all (&lock->waiting_readers); + } + } + LeaveCriticalSection (&lock->lock); + return 0; +} + +int +glthread_rwlock_destroy_func (gl_rwlock_t *lock) +{ + if (!lock->guard.done) + return EINVAL; + if (lock->runcount != 0) + return EBUSY; + DeleteCriticalSection (&lock->lock); + if (lock->waiting_readers.array != NULL) + free (lock->waiting_readers.array); + if (lock->waiting_writers.array != NULL) + free (lock->waiting_writers.array); + lock->guard.done = 0; + return 0; +} + +/* --------------------- gl_recursive_lock_t datatype --------------------- */ + +void +glthread_recursive_lock_init_func (gl_recursive_lock_t *lock) +{ + lock->owner = 0; + lock->depth = 0; + InitializeCriticalSection (&lock->lock); + lock->guard.done = 1; +} + +int +glthread_recursive_lock_lock_func (gl_recursive_lock_t *lock) +{ + if (!lock->guard.done) + { + if (InterlockedIncrement (&lock->guard.started) == 0) + /* This thread is the first one to need this lock. Initialize it. */ + glthread_recursive_lock_init (lock); + else + /* Yield the CPU while waiting for another thread to finish + initializing this lock. */ + while (!lock->guard.done) + Sleep (0); + } + { + DWORD self = GetCurrentThreadId (); + if (lock->owner != self) + { + EnterCriticalSection (&lock->lock); + lock->owner = self; + } + if (++(lock->depth) == 0) /* wraparound? */ + { + lock->depth--; + return EAGAIN; + } + } + return 0; +} + +int +glthread_recursive_lock_unlock_func (gl_recursive_lock_t *lock) +{ + if (lock->owner != GetCurrentThreadId ()) + return EPERM; + if (lock->depth == 0) + return EINVAL; + if (--(lock->depth) == 0) + { + lock->owner = 0; + LeaveCriticalSection (&lock->lock); + } + return 0; +} + +int +glthread_recursive_lock_destroy_func (gl_recursive_lock_t *lock) +{ + if (lock->owner != 0) + return EBUSY; + DeleteCriticalSection (&lock->lock); + lock->guard.done = 0; + return 0; +} + +/* -------------------------- gl_once_t datatype -------------------------- */ + +void +glthread_once_func (gl_once_t *once_control, void (*initfunction) (void)) +{ + if (once_control->inited <= 0) + { + if (InterlockedIncrement (&once_control->started) == 0) + { + /* This thread is the first one to come to this once_control. */ + InitializeCriticalSection (&once_control->lock); + EnterCriticalSection (&once_control->lock); + once_control->inited = 0; + initfunction (); + once_control->inited = 1; + LeaveCriticalSection (&once_control->lock); + } + else + { + /* Undo last operation. */ + InterlockedDecrement (&once_control->started); + /* Some other thread has already started the initialization. + Yield the CPU while waiting for the other thread to finish + initializing and taking the lock. */ + while (once_control->inited < 0) + Sleep (0); + if (once_control->inited <= 0) + { + /* Take the lock. This blocks until the other thread has + finished calling the initfunction. */ + EnterCriticalSection (&once_control->lock); + LeaveCriticalSection (&once_control->lock); + if (!(once_control->inited > 0)) + abort (); + } + } + } +} + +#endif + +/* ========================================================================= */ diff --git a/lib/glthread/lock.h b/lib/glthread/lock.h new file mode 100644 index 0000000..30aa730 --- /dev/null +++ b/lib/glthread/lock.h @@ -0,0 +1,985 @@ +/* Locking in multithreaded situations. + Copyright (C) 2005-2018 Free Software Foundation, Inc. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3, or (at your option) + any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, see <https://www.gnu.org/licenses/>. */ + +/* Written by Bruno Haible <bruno@clisp.org>, 2005. + Based on GCC's gthr-posix.h, gthr-posix95.h, gthr-solaris.h, + gthr-win32.h. */ + +/* This file contains locking primitives for use with a given thread library. + It does not contain primitives for creating threads or for other + synchronization primitives. + + Normal (non-recursive) locks: + Type: gl_lock_t + Declaration: gl_lock_define(extern, name) + Initializer: gl_lock_define_initialized(, name) + Initialization: gl_lock_init (name); + Taking the lock: gl_lock_lock (name); + Releasing the lock: gl_lock_unlock (name); + De-initialization: gl_lock_destroy (name); + Equivalent functions with control of error handling: + Initialization: err = glthread_lock_init (&name); + Taking the lock: err = glthread_lock_lock (&name); + Releasing the lock: err = glthread_lock_unlock (&name); + De-initialization: err = glthread_lock_destroy (&name); + + Read-Write (non-recursive) locks: + Type: gl_rwlock_t + Declaration: gl_rwlock_define(extern, name) + Initializer: gl_rwlock_define_initialized(, name) + Initialization: gl_rwlock_init (name); + Taking the lock: gl_rwlock_rdlock (name); + gl_rwlock_wrlock (name); + Releasing the lock: gl_rwlock_unlock (name); + De-initialization: gl_rwlock_destroy (name); + Equivalent functions with control of error handling: + Initialization: err = glthread_rwlock_init (&name); + Taking the lock: err = glthread_rwlock_rdlock (&name); + err = glthread_rwlock_wrlock (&name); + Releasing the lock: err = glthread_rwlock_unlock (&name); + De-initialization: err = glthread_rwlock_destroy (&name); + + Recursive locks: + Type: gl_recursive_lock_t + Declaration: gl_recursive_lock_define(extern, name) + Initializer: gl_recursive_lock_define_initialized(, name) + Initialization: gl_recursive_lock_init (name); + Taking the lock: gl_recursive_lock_lock (name); + Releasing the lock: gl_recursive_lock_unlock (name); + De-initialization: gl_recursive_lock_destroy (name); + Equivalent functions with control of error handling: + Initialization: err = glthread_recursive_lock_init (&name); + Taking the lock: err = glthread_recursive_lock_lock (&name); + Releasing the lock: err = glthread_recursive_lock_unlock (&name); + De-initialization: err = glthread_recursive_lock_destroy (&name); + + Once-only execution: + Type: gl_once_t + Initializer: gl_once_define(extern, name) + Execution: gl_once (name, initfunction); + Equivalent functions with control of error handling: + Execution: err = glthread_once (&name, initfunction); +*/ + + +#ifndef _LOCK_H +#define _LOCK_H + +#include <errno.h> +#include <stdlib.h> + +/* ========================================================================= */ + +#if USE_POSIX_THREADS + +/* Use the POSIX threads library. */ + +# include <pthread.h> + +# ifdef __cplusplus +extern "C" { +# endif + +# if PTHREAD_IN_USE_DETECTION_HARD + +/* The pthread_in_use() detection needs to be done at runtime. */ +# define pthread_in_use() \ + glthread_in_use () +extern int glthread_in_use (void); + +# endif + +# if USE_POSIX_THREADS_WEAK + +/* Use weak references to the POSIX threads library. */ + +/* Weak references avoid dragging in external libraries if the other parts + of the program don't use them. Here we use them, because we don't want + every program that uses libintl to depend on libpthread. This assumes + that libpthread would not be loaded after libintl; i.e. if libintl is + loaded first, by an executable that does not depend on libpthread, and + then a module is dynamically loaded that depends on libpthread, libintl + will not be multithread-safe. */ + +/* The way to test at runtime whether libpthread is present is to test + whether a function pointer's value, such as &pthread_mutex_init, is + non-NULL. However, some versions of GCC have a bug through which, in + PIC mode, &foo != NULL always evaluates to true if there is a direct + call to foo(...) in the same function. To avoid this, we test the + address of a function in libpthread that we don't use. */ + +# pragma weak pthread_mutex_init +# pragma weak pthread_mutex_lock +# pragma weak pthread_mutex_unlock +# pragma weak pthread_mutex_destroy +# pragma weak pthread_rwlock_init +# pragma weak pthread_rwlock_rdlock +# pragma weak pthread_rwlock_wrlock +# pragma weak pthread_rwlock_unlock +# pragma weak pthread_rwlock_destroy +# pragma weak pthread_once +# pragma weak pthread_cond_init +# pragma weak pthread_cond_wait +# pragma weak pthread_cond_signal +# pragma weak pthread_cond_broadcast +# pragma weak pthread_cond_destroy +# pragma weak pthread_mutexattr_init +# pragma weak pthread_mutexattr_settype +# pragma weak pthread_mutexattr_destroy +# pragma weak pthread_rwlockattr_init +# if __GNU_LIBRARY__ > 1 +# pragma weak pthread_rwlockattr_setkind_np +# endif +# pragma weak pthread_rwlockattr_destroy +# ifndef pthread_self +# pragma weak pthread_self +# endif + +# if !PTHREAD_IN_USE_DETECTION_HARD + /* Considering all platforms with USE_POSIX_THREADS_WEAK, only few symbols + can be used to determine whether libpthread is in use. These are: + pthread_mutexattr_gettype + pthread_rwlockattr_destroy + pthread_rwlockattr_init + */ +# pragma weak pthread_mutexattr_gettype +# define pthread_in_use() (pthread_mutexattr_gettype != NULL) +# endif + +# else + +# if !PTHREAD_IN_USE_DETECTION_HARD +# define pthread_in_use() 1 +# endif + +# endif + +/* -------------------------- gl_lock_t datatype -------------------------- */ + +typedef pthread_mutex_t gl_lock_t; +# define gl_lock_define(STORAGECLASS, NAME) \ + STORAGECLASS pthread_mutex_t NAME; +# define gl_lock_define_initialized(STORAGECLASS, NAME) \ + STORAGECLASS pthread_mutex_t NAME = gl_lock_initializer; +# define gl_lock_initializer \ + PTHREAD_MUTEX_INITIALIZER +# define glthread_lock_init(LOCK) \ + (pthread_in_use () ? pthread_mutex_init (LOCK, NULL) : 0) +# define glthread_lock_lock(LOCK) \ + (pthread_in_use () ? pthread_mutex_lock (LOCK) : 0) +# define glthread_lock_unlock(LOCK) \ + (pthread_in_use () ? pthread_mutex_unlock (LOCK) : 0) +# define glthread_lock_destroy(LOCK) \ + (pthread_in_use () ? pthread_mutex_destroy (LOCK) : 0) + +/* ------------------------- gl_rwlock_t datatype ------------------------- */ + +# if HAVE_PTHREAD_RWLOCK && (HAVE_PTHREAD_RWLOCK_RDLOCK_PREFER_WRITER || (defined PTHREAD_RWLOCK_WRITER_NONRECURSIVE_INITIALIZER_NP && (__GNU_LIBRARY__ > 1))) + +# ifdef PTHREAD_RWLOCK_INITIALIZER + +typedef pthread_rwlock_t gl_rwlock_t; +# define gl_rwlock_define(STORAGECLASS, NAME) \ + STORAGECLASS pthread_rwlock_t NAME; +# define gl_rwlock_define_initialized(STORAGECLASS, NAME) \ + STORAGECLASS pthread_rwlock_t NAME = gl_rwlock_initializer; +# if HAVE_PTHREAD_RWLOCK_RDLOCK_PREFER_WRITER +# define gl_rwlock_initializer \ + PTHREAD_RWLOCK_INITIALIZER +# define glthread_rwlock_init(LOCK) \ + (pthread_in_use () ? pthread_rwlock_init (LOCK, NULL) : 0) +# else /* glibc with bug https://sourceware.org/bugzilla/show_bug.cgi?id=13701 */ +# define gl_rwlock_initializer \ + PTHREAD_RWLOCK_WRITER_NONRECURSIVE_INITIALIZER_NP +# define glthread_rwlock_init(LOCK) \ + (pthread_in_use () ? glthread_rwlock_init_for_glibc (LOCK) : 0) +extern int glthread_rwlock_init_for_glibc (pthread_rwlock_t *lock); +# endif +# define glthread_rwlock_rdlock(LOCK) \ + (pthread_in_use () ? pthread_rwlock_rdlock (LOCK) : 0) +# define glthread_rwlock_wrlock(LOCK) \ + (pthread_in_use () ? pthread_rwlock_wrlock (LOCK) : 0) +# define glthread_rwlock_unlock(LOCK) \ + (pthread_in_use () ? pthread_rwlock_unlock (LOCK) : 0) +# define glthread_rwlock_destroy(LOCK) \ + (pthread_in_use () ? pthread_rwlock_destroy (LOCK) : 0) + +# else + +typedef struct + { + int initialized; + pthread_mutex_t guard; /* protects the initialization */ + pthread_rwlock_t rwlock; /* read-write lock */ + } + gl_rwlock_t; +# define gl_rwlock_define(STORAGECLASS, NAME) \ + STORAGECLASS gl_rwlock_t NAME; +# define gl_rwlock_define_initialized(STORAGECLASS, NAME) \ + STORAGECLASS gl_rwlock_t NAME = gl_rwlock_initializer; +# define gl_rwlock_initializer \ + { 0, PTHREAD_MUTEX_INITIALIZER } +# define glthread_rwlock_init(LOCK) \ + (pthread_in_use () ? glthread_rwlock_init_multithreaded (LOCK) : 0) +# define glthread_rwlock_rdlock(LOCK) \ + (pthread_in_use () ? glthread_rwlock_rdlock_multithreaded (LOCK) : 0) +# define glthread_rwlock_wrlock(LOCK) \ + (pthread_in_use () ? glthread_rwlock_wrlock_multithreaded (LOCK) : 0) +# define glthread_rwlock_unlock(LOCK) \ + (pthread_in_use () ? glthread_rwlock_unlock_multithreaded (LOCK) : 0) +# define glthread_rwlock_destroy(LOCK) \ + (pthread_in_use () ? glthread_rwlock_destroy_multithreaded (LOCK) : 0) +extern int glthread_rwlock_init_multithreaded (gl_rwlock_t *lock); +extern int glthread_rwlock_rdlock_multithreaded (gl_rwlock_t *lock); +extern int glthread_rwlock_wrlock_multithreaded (gl_rwlock_t *lock); +extern int glthread_rwlock_unlock_multithreaded (gl_rwlock_t *lock); +extern int glthread_rwlock_destroy_multithreaded (gl_rwlock_t *lock); + +# endif + +# else + +typedef struct + { + pthread_mutex_t lock; /* protects the remaining fields */ + pthread_cond_t waiting_readers; /* waiting readers */ + pthread_cond_t waiting_writers; /* waiting writers */ + unsigned int waiting_writers_count; /* number of waiting writers */ + int runcount; /* number of readers running, or -1 when a writer runs */ + } + gl_rwlock_t; +# define gl_rwlock_define(STORAGECLASS, NAME) \ + STORAGECLASS gl_rwlock_t NAME; +# define gl_rwlock_define_initialized(STORAGECLASS, NAME) \ + STORAGECLASS gl_rwlock_t NAME = gl_rwlock_initializer; +# define gl_rwlock_initializer \ + { PTHREAD_MUTEX_INITIALIZER, PTHREAD_COND_INITIALIZER, PTHREAD_COND_INITIALIZER, 0, 0 } +# define glthread_rwlock_init(LOCK) \ + (pthread_in_use () ? glthread_rwlock_init_multithreaded (LOCK) : 0) +# define glthread_rwlock_rdlock(LOCK) \ + (pthread_in_use () ? glthread_rwlock_rdlock_multithreaded (LOCK) : 0) +# define glthread_rwlock_wrlock(LOCK) \ + (pthread_in_use () ? glthread_rwlock_wrlock_multithreaded (LOCK) : 0) +# define glthread_rwlock_unlock(LOCK) \ + (pthread_in_use () ? glthread_rwlock_unlock_multithreaded (LOCK) : 0) +# define glthread_rwlock_destroy(LOCK) \ + (pthread_in_use () ? glthread_rwlock_destroy_multithreaded (LOCK) : 0) +extern int glthread_rwlock_init_multithreaded (gl_rwlock_t *lock); +extern int glthread_rwlock_rdlock_multithreaded (gl_rwlock_t *lock); +extern int glthread_rwlock_wrlock_multithreaded (gl_rwlock_t *lock); +extern int glthread_rwlock_unlock_multithreaded (gl_rwlock_t *lock); +extern int glthread_rwlock_destroy_multithreaded (gl_rwlock_t *lock); + +# endif + +/* --------------------- gl_recursive_lock_t datatype --------------------- */ + +# if HAVE_PTHREAD_MUTEX_RECURSIVE + +# if defined PTHREAD_RECURSIVE_MUTEX_INITIALIZER || defined PTHREAD_RECURSIVE_MUTEX_INITIALIZER_NP + +typedef pthread_mutex_t gl_recursive_lock_t; +# define gl_recursive_lock_define(STORAGECLASS, NAME) \ + STORAGECLASS pthread_mutex_t NAME; +# define gl_recursive_lock_define_initialized(STORAGECLASS, NAME) \ + STORAGECLASS pthread_mutex_t NAME = gl_recursive_lock_initializer; +# ifdef PTHREAD_RECURSIVE_MUTEX_INITIALIZER +# define gl_recursive_lock_initializer \ + PTHREAD_RECURSIVE_MUTEX_INITIALIZER +# else +# define gl_recursive_lock_initializer \ + PTHREAD_RECURSIVE_MUTEX_INITIALIZER_NP +# endif +# define glthread_recursive_lock_init(LOCK) \ + (pthread_in_use () ? glthread_recursive_lock_init_multithreaded (LOCK) : 0) +# define glthread_recursive_lock_lock(LOCK) \ + (pthread_in_use () ? pthread_mutex_lock (LOCK) : 0) +# define glthread_recursive_lock_unlock(LOCK) \ + (pthread_in_use () ? pthread_mutex_unlock (LOCK) : 0) +# define glthread_recursive_lock_destroy(LOCK) \ + (pthread_in_use () ? pthread_mutex_destroy (LOCK) : 0) +extern int glthread_recursive_lock_init_multithreaded (gl_recursive_lock_t *lock); + +# else + +typedef struct + { + pthread_mutex_t recmutex; /* recursive mutex */ + pthread_mutex_t guard; /* protects the initialization */ + int initialized; + } + gl_recursive_lock_t; +# define gl_recursive_lock_define(STORAGECLASS, NAME) \ + STORAGECLASS gl_recursive_lock_t NAME; +# define gl_recursive_lock_define_initialized(STORAGECLASS, NAME) \ + STORAGECLASS gl_recursive_lock_t NAME = gl_recursive_lock_initializer; +# define gl_recursive_lock_initializer \ + { PTHREAD_MUTEX_INITIALIZER, PTHREAD_MUTEX_INITIALIZER, 0 } +# define glthread_recursive_lock_init(LOCK) \ + (pthread_in_use () ? glthread_recursive_lock_init_multithreaded (LOCK) : 0) +# define glthread_recursive_lock_lock(LOCK) \ + (pthread_in_use () ? glthread_recursive_lock_lock_multithreaded (LOCK) : 0) +# define glthread_recursive_lock_unlock(LOCK) \ + (pthread_in_use () ? glthread_recursive_lock_unlock_multithreaded (LOCK) : 0) +# define glthread_recursive_lock_destroy(LOCK) \ + (pthread_in_use () ? glthread_recursive_lock_destroy_multithreaded (LOCK) : 0) +extern int glthread_recursive_lock_init_multithreaded (gl_recursive_lock_t *lock); +extern int glthread_recursive_lock_lock_multithreaded (gl_recursive_lock_t *lock); +extern int glthread_recursive_lock_unlock_multithreaded (gl_recursive_lock_t *lock); +extern int glthread_recursive_lock_destroy_multithreaded (gl_recursive_lock_t *lock); + +# endif + +# else + +/* Old versions of POSIX threads on Solaris did not have recursive locks. + We have to implement them ourselves. */ + +typedef struct + { + pthread_mutex_t mutex; + pthread_t owner; + unsigned long depth; + } + gl_recursive_lock_t; +# define gl_recursive_lock_define(STORAGECLASS, NAME) \ + STORAGECLASS gl_recursive_lock_t NAME; +# define gl_recursive_lock_define_initialized(STORAGECLASS, NAME) \ + STORAGECLASS gl_recursive_lock_t NAME = gl_recursive_lock_initializer; +# define gl_recursive_lock_initializer \ + { PTHREAD_MUTEX_INITIALIZER, (pthread_t) 0, 0 } +# define glthread_recursive_lock_init(LOCK) \ + (pthread_in_use () ? glthread_recursive_lock_init_multithreaded (LOCK) : 0) +# define glthread_recursive_lock_lock(LOCK) \ + (pthread_in_use () ? glthread_recursive_lock_lock_multithreaded (LOCK) : 0) +# define glthread_recursive_lock_unlock(LOCK) \ + (pthread_in_use () ? glthread_recursive_lock_unlock_multithreaded (LOCK) : 0) +# define glthread_recursive_lock_destroy(LOCK) \ + (pthread_in_use () ? glthread_recursive_lock_destroy_multithreaded (LOCK) : 0) +extern int glthread_recursive_lock_init_multithreaded (gl_recursive_lock_t *lock); +extern int glthread_recursive_lock_lock_multithreaded (gl_recursive_lock_t *lock); +extern int glthread_recursive_lock_unlock_multithreaded (gl_recursive_lock_t *lock); +extern int glthread_recursive_lock_destroy_multithreaded (gl_recursive_lock_t *lock); + +# endif + +/* -------------------------- gl_once_t datatype -------------------------- */ + +typedef pthread_once_t gl_once_t; +# define gl_once_define(STORAGECLASS, NAME) \ + STORAGECLASS pthread_once_t NAME = PTHREAD_ONCE_INIT; +# define glthread_once(ONCE_CONTROL, INITFUNCTION) \ + (pthread_in_use () \ + ? pthread_once (ONCE_CONTROL, INITFUNCTION) \ + : (glthread_once_singlethreaded (ONCE_CONTROL) ? (INITFUNCTION (), 0) : 0)) +extern int glthread_once_singlethreaded (pthread_once_t *once_control); + +# ifdef __cplusplus +} +# endif + +#endif + +/* ========================================================================= */ + +#if USE_PTH_THREADS + +/* Use the GNU Pth threads library. */ + +# include <pth.h> + +# ifdef __cplusplus +extern "C" { +# endif + +# if USE_PTH_THREADS_WEAK + +/* Use weak references to the GNU Pth threads library. */ + +# pragma weak pth_mutex_init +# pragma weak pth_mutex_acquire +# pragma weak pth_mutex_release +# pragma weak pth_rwlock_init +# pragma weak pth_rwlock_acquire +# pragma weak pth_rwlock_release +# pragma weak pth_once + +# pragma weak pth_cancel +# define pth_in_use() (pth_cancel != NULL) + +# else + +# define pth_in_use() 1 + +# endif + +/* -------------------------- gl_lock_t datatype -------------------------- */ + +typedef pth_mutex_t gl_lock_t; +# define gl_lock_define(STORAGECLASS, NAME) \ + STORAGECLASS pth_mutex_t NAME; +# define gl_lock_define_initialized(STORAGECLASS, NAME) \ + STORAGECLASS pth_mutex_t NAME = gl_lock_initializer; +# define gl_lock_initializer \ + PTH_MUTEX_INIT +# define glthread_lock_init(LOCK) \ + (pth_in_use () && !pth_mutex_init (LOCK) ? errno : 0) +# define glthread_lock_lock(LOCK) \ + (pth_in_use () && !pth_mutex_acquire (LOCK, 0, NULL) ? errno : 0) +# define glthread_lock_unlock(LOCK) \ + (pth_in_use () && !pth_mutex_release (LOCK) ? errno : 0) +# define glthread_lock_destroy(LOCK) \ + ((void)(LOCK), 0) + +/* ------------------------- gl_rwlock_t datatype ------------------------- */ + +/* Pth pth_rwlock_acquire always prefers readers. No autoconf test so far. */ +# if HAVE_PTH_RWLOCK_ACQUIRE_PREFER_WRITER + +typedef pth_rwlock_t gl_rwlock_t; +# define gl_rwlock_define(STORAGECLASS, NAME) \ + STORAGECLASS pth_rwlock_t NAME; +# define gl_rwlock_define_initialized(STORAGECLASS, NAME) \ + STORAGECLASS pth_rwlock_t NAME = gl_rwlock_initializer; +# define gl_rwlock_initializer \ + PTH_RWLOCK_INIT +# define glthread_rwlock_init(LOCK) \ + (pth_in_use () && !pth_rwlock_init (LOCK) ? errno : 0) +# define glthread_rwlock_rdlock(LOCK) \ + (pth_in_use () && !pth_rwlock_acquire (LOCK, PTH_RWLOCK_RD, 0, NULL) ? errno : 0) +# define glthread_rwlock_wrlock(LOCK) \ + (pth_in_use () && !pth_rwlock_acquire (LOCK, PTH_RWLOCK_RW, 0, NULL) ? errno : 0) +# define glthread_rwlock_unlock(LOCK) \ + (pth_in_use () && !pth_rwlock_release (LOCK) ? errno : 0) +# define glthread_rwlock_destroy(LOCK) \ + ((void)(LOCK), 0) + +# else + +typedef struct + { + int initialized; + pth_mutex_t lock; /* protects the remaining fields */ + pth_cond_t waiting_readers; /* waiting readers */ + pth_cond_t waiting_writers; /* waiting writers */ + unsigned int waiting_writers_count; /* number of waiting writers */ + int runcount; /* number of readers running, or -1 when a writer runs */ + } + gl_rwlock_t; +# define gl_rwlock_define(STORAGECLASS, NAME) \ + STORAGECLASS gl_rwlock_t NAME; +# define gl_rwlock_define_initialized(STORAGECLASS, NAME) \ + STORAGECLASS gl_rwlock_t NAME = gl_rwlock_initializer; +# define gl_rwlock_initializer \ + { 0 } +# define glthread_rwlock_init(LOCK) \ + (pth_in_use () ? glthread_rwlock_init_multithreaded (LOCK) : 0) +# define glthread_rwlock_rdlock(LOCK) \ + (pth_in_use () ? glthread_rwlock_rdlock_multithreaded (LOCK) : 0) +# define glthread_rwlock_wrlock(LOCK) \ + (pth_in_use () ? glthread_rwlock_wrlock_multithreaded (LOCK) : 0) +# define glthread_rwlock_unlock(LOCK) \ + (pth_in_use () ? glthread_rwlock_unlock_multithreaded (LOCK) : 0) +# define glthread_rwlock_destroy(LOCK) \ + (pth_in_use () ? glthread_rwlock_destroy_multithreaded (LOCK) : 0) +extern int glthread_rwlock_init_multithreaded (gl_rwlock_t *lock); +extern int glthread_rwlock_rdlock_multithreaded (gl_rwlock_t *lock); +extern int glthread_rwlock_wrlock_multithreaded (gl_rwlock_t *lock); +extern int glthread_rwlock_unlock_multithreaded (gl_rwlock_t *lock); +extern int glthread_rwlock_destroy_multithreaded (gl_rwlock_t *lock); + +# endif + +/* --------------------- gl_recursive_lock_t datatype --------------------- */ + +/* In Pth, mutexes are recursive by default. */ +typedef pth_mutex_t gl_recursive_lock_t; +# define gl_recursive_lock_define(STORAGECLASS, NAME) \ + STORAGECLASS pth_mutex_t NAME; +# define gl_recursive_lock_define_initialized(STORAGECLASS, NAME) \ + STORAGECLASS pth_mutex_t NAME = gl_recursive_lock_initializer; +# define gl_recursive_lock_initializer \ + PTH_MUTEX_INIT +# define glthread_recursive_lock_init(LOCK) \ + (pth_in_use () && !pth_mutex_init (LOCK) ? errno : 0) +# define glthread_recursive_lock_lock(LOCK) \ + (pth_in_use () && !pth_mutex_acquire (LOCK, 0, NULL) ? errno : 0) +# define glthread_recursive_lock_unlock(LOCK) \ + (pth_in_use () && !pth_mutex_release (LOCK) ? errno : 0) +# define glthread_recursive_lock_destroy(LOCK) \ + ((void)(LOCK), 0) + +/* -------------------------- gl_once_t datatype -------------------------- */ + +typedef pth_once_t gl_once_t; +# define gl_once_define(STORAGECLASS, NAME) \ + STORAGECLASS pth_once_t NAME = PTH_ONCE_INIT; +# define glthread_once(ONCE_CONTROL, INITFUNCTION) \ + (pth_in_use () \ + ? glthread_once_multithreaded (ONCE_CONTROL, INITFUNCTION) \ + : (glthread_once_singlethreaded (ONCE_CONTROL) ? (INITFUNCTION (), 0) : 0)) +extern int glthread_once_multithreaded (pth_once_t *once_control, void (*initfunction) (void)); +extern int glthread_once_singlethreaded (pth_once_t *once_control); + +# ifdef __cplusplus +} +# endif + +#endif + +/* ========================================================================= */ + +#if USE_SOLARIS_THREADS + +/* Use the old Solaris threads library. */ + +# include <thread.h> +# include <synch.h> + +# ifdef __cplusplus +extern "C" { +# endif + +# if USE_SOLARIS_THREADS_WEAK + +/* Use weak references to the old Solaris threads library. */ + +# pragma weak mutex_init +# pragma weak mutex_lock +# pragma weak mutex_unlock +# pragma weak mutex_destroy +# pragma weak rwlock_init +# pragma weak rw_rdlock +# pragma weak rw_wrlock +# pragma weak rw_unlock +# pragma weak rwlock_destroy +# pragma weak thr_self + +# pragma weak thr_suspend +# define thread_in_use() (thr_suspend != NULL) + +# else + +# define thread_in_use() 1 + +# endif + +/* -------------------------- gl_lock_t datatype -------------------------- */ + +typedef mutex_t gl_lock_t; +# define gl_lock_define(STORAGECLASS, NAME) \ + STORAGECLASS mutex_t NAME; +# define gl_lock_define_initialized(STORAGECLASS, NAME) \ + STORAGECLASS mutex_t NAME = gl_lock_initializer; +# define gl_lock_initializer \ + DEFAULTMUTEX +# define glthread_lock_init(LOCK) \ + (thread_in_use () ? mutex_init (LOCK, USYNC_THREAD, NULL) : 0) +# define glthread_lock_lock(LOCK) \ + (thread_in_use () ? mutex_lock (LOCK) : 0) +# define glthread_lock_unlock(LOCK) \ + (thread_in_use () ? mutex_unlock (LOCK) : 0) +# define glthread_lock_destroy(LOCK) \ + (thread_in_use () ? mutex_destroy (LOCK) : 0) + +/* ------------------------- gl_rwlock_t datatype ------------------------- */ + +typedef rwlock_t gl_rwlock_t; +# define gl_rwlock_define(STORAGECLASS, NAME) \ + STORAGECLASS rwlock_t NAME; +# define gl_rwlock_define_initialized(STORAGECLASS, NAME) \ + STORAGECLASS rwlock_t NAME = gl_rwlock_initializer; +# define gl_rwlock_initializer \ + DEFAULTRWLOCK +# define glthread_rwlock_init(LOCK) \ + (thread_in_use () ? rwlock_init (LOCK, USYNC_THREAD, NULL) : 0) +# define glthread_rwlock_rdlock(LOCK) \ + (thread_in_use () ? rw_rdlock (LOCK) : 0) +# define glthread_rwlock_wrlock(LOCK) \ + (thread_in_use () ? rw_wrlock (LOCK) : 0) +# define glthread_rwlock_unlock(LOCK) \ + (thread_in_use () ? rw_unlock (LOCK) : 0) +# define glthread_rwlock_destroy(LOCK) \ + (thread_in_use () ? rwlock_destroy (LOCK) : 0) + +/* --------------------- gl_recursive_lock_t datatype --------------------- */ + +/* Old Solaris threads did not have recursive locks. + We have to implement them ourselves. */ + +typedef struct + { + mutex_t mutex; + thread_t owner; + unsigned long depth; + } + gl_recursive_lock_t; +# define gl_recursive_lock_define(STORAGECLASS, NAME) \ + STORAGECLASS gl_recursive_lock_t NAME; +# define gl_recursive_lock_define_initialized(STORAGECLASS, NAME) \ + STORAGECLASS gl_recursive_lock_t NAME = gl_recursive_lock_initializer; +# define gl_recursive_lock_initializer \ + { DEFAULTMUTEX, (thread_t) 0, 0 } +# define glthread_recursive_lock_init(LOCK) \ + (thread_in_use () ? glthread_recursive_lock_init_multithreaded (LOCK) : 0) +# define glthread_recursive_lock_lock(LOCK) \ + (thread_in_use () ? glthread_recursive_lock_lock_multithreaded (LOCK) : 0) +# define glthread_recursive_lock_unlock(LOCK) \ + (thread_in_use () ? glthread_recursive_lock_unlock_multithreaded (LOCK) : 0) +# define glthread_recursive_lock_destroy(LOCK) \ + (thread_in_use () ? glthread_recursive_lock_destroy_multithreaded (LOCK) : 0) +extern int glthread_recursive_lock_init_multithreaded (gl_recursive_lock_t *lock); +extern int glthread_recursive_lock_lock_multithreaded (gl_recursive_lock_t *lock); +extern int glthread_recursive_lock_unlock_multithreaded (gl_recursive_lock_t *lock); +extern int glthread_recursive_lock_destroy_multithreaded (gl_recursive_lock_t *lock); + +/* -------------------------- gl_once_t datatype -------------------------- */ + +typedef struct + { + volatile int inited; + mutex_t mutex; + } + gl_once_t; +# define gl_once_define(STORAGECLASS, NAME) \ + STORAGECLASS gl_once_t NAME = { 0, DEFAULTMUTEX }; +# define glthread_once(ONCE_CONTROL, INITFUNCTION) \ + (thread_in_use () \ + ? glthread_once_multithreaded (ONCE_CONTROL, INITFUNCTION) \ + : (glthread_once_singlethreaded (ONCE_CONTROL) ? (INITFUNCTION (), 0) : 0)) +extern int glthread_once_multithreaded (gl_once_t *once_control, void (*initfunction) (void)); +extern int glthread_once_singlethreaded (gl_once_t *once_control); + +# ifdef __cplusplus +} +# endif + +#endif + +/* ========================================================================= */ + +#if USE_WINDOWS_THREADS + +# define WIN32_LEAN_AND_MEAN /* avoid including junk */ +# include <windows.h> + +# ifdef __cplusplus +extern "C" { +# endif + +/* We can use CRITICAL_SECTION directly, rather than the native Windows Event, + Mutex, Semaphore types, because + - we need only to synchronize inside a single process (address space), + not inter-process locking, + - we don't need to support trylock operations. (TryEnterCriticalSection + does not work on Windows 95/98/ME. Packages that need trylock usually + define their own mutex type.) */ + +/* There is no way to statically initialize a CRITICAL_SECTION. It needs + to be done lazily, once only. For this we need spinlocks. */ + +typedef struct { volatile int done; volatile long started; } gl_spinlock_t; + +/* -------------------------- gl_lock_t datatype -------------------------- */ + +typedef struct + { + gl_spinlock_t guard; /* protects the initialization */ + CRITICAL_SECTION lock; + } + gl_lock_t; +# define gl_lock_define(STORAGECLASS, NAME) \ + STORAGECLASS gl_lock_t NAME; +# define gl_lock_define_initialized(STORAGECLASS, NAME) \ + STORAGECLASS gl_lock_t NAME = gl_lock_initializer; +# define gl_lock_initializer \ + { { 0, -1 } } +# define glthread_lock_init(LOCK) \ + (glthread_lock_init_func (LOCK), 0) +# define glthread_lock_lock(LOCK) \ + glthread_lock_lock_func (LOCK) +# define glthread_lock_unlock(LOCK) \ + glthread_lock_unlock_func (LOCK) +# define glthread_lock_destroy(LOCK) \ + glthread_lock_destroy_func (LOCK) +extern void glthread_lock_init_func (gl_lock_t *lock); +extern int glthread_lock_lock_func (gl_lock_t *lock); +extern int glthread_lock_unlock_func (gl_lock_t *lock); +extern int glthread_lock_destroy_func (gl_lock_t *lock); + +/* ------------------------- gl_rwlock_t datatype ------------------------- */ + +/* It is impossible to implement read-write locks using plain locks, without + introducing an extra thread dedicated to managing read-write locks. + Therefore here we need to use the low-level Event type. */ + +typedef struct + { + HANDLE *array; /* array of waiting threads, each represented by an event */ + unsigned int count; /* number of waiting threads */ + unsigned int alloc; /* length of allocated array */ + unsigned int offset; /* index of first waiting thread in array */ + } + gl_carray_waitqueue_t; +typedef struct + { + gl_spinlock_t guard; /* protects the initialization */ + CRITICAL_SECTION lock; /* protects the remaining fields */ + gl_carray_waitqueue_t waiting_readers; /* waiting readers */ + gl_carray_waitqueue_t waiting_writers; /* waiting writers */ + int runcount; /* number of readers running, or -1 when a writer runs */ + } + gl_rwlock_t; +# define gl_rwlock_define(STORAGECLASS, NAME) \ + STORAGECLASS gl_rwlock_t NAME; +# define gl_rwlock_define_initialized(STORAGECLASS, NAME) \ + STORAGECLASS gl_rwlock_t NAME = gl_rwlock_initializer; +# define gl_rwlock_initializer \ + { { 0, -1 } } +# define glthread_rwlock_init(LOCK) \ + (glthread_rwlock_init_func (LOCK), 0) +# define glthread_rwlock_rdlock(LOCK) \ + glthread_rwlock_rdlock_func (LOCK) +# define glthread_rwlock_wrlock(LOCK) \ + glthread_rwlock_wrlock_func (LOCK) +# define glthread_rwlock_unlock(LOCK) \ + glthread_rwlock_unlock_func (LOCK) +# define glthread_rwlock_destroy(LOCK) \ + glthread_rwlock_destroy_func (LOCK) +extern void glthread_rwlock_init_func (gl_rwlock_t *lock); +extern int glthread_rwlock_rdlock_func (gl_rwlock_t *lock); +extern int glthread_rwlock_wrlock_func (gl_rwlock_t *lock); +extern int glthread_rwlock_unlock_func (gl_rwlock_t *lock); +extern int glthread_rwlock_destroy_func (gl_rwlock_t *lock); + +/* --------------------- gl_recursive_lock_t datatype --------------------- */ + +/* The native Windows documentation says that CRITICAL_SECTION already + implements a recursive lock. But we need not rely on it: It's easy to + implement a recursive lock without this assumption. */ + +typedef struct + { + gl_spinlock_t guard; /* protects the initialization */ + DWORD owner; + unsigned long depth; + CRITICAL_SECTION lock; + } + gl_recursive_lock_t; +# define gl_recursive_lock_define(STORAGECLASS, NAME) \ + STORAGECLASS gl_recursive_lock_t NAME; +# define gl_recursive_lock_define_initialized(STORAGECLASS, NAME) \ + STORAGECLASS gl_recursive_lock_t NAME = gl_recursive_lock_initializer; +# define gl_recursive_lock_initializer \ + { { 0, -1 }, 0, 0 } +# define glthread_recursive_lock_init(LOCK) \ + (glthread_recursive_lock_init_func (LOCK), 0) +# define glthread_recursive_lock_lock(LOCK) \ + glthread_recursive_lock_lock_func (LOCK) +# define glthread_recursive_lock_unlock(LOCK) \ + glthread_recursive_lock_unlock_func (LOCK) +# define glthread_recursive_lock_destroy(LOCK) \ + glthread_recursive_lock_destroy_func (LOCK) +extern void glthread_recursive_lock_init_func (gl_recursive_lock_t *lock); +extern int glthread_recursive_lock_lock_func (gl_recursive_lock_t *lock); +extern int glthread_recursive_lock_unlock_func (gl_recursive_lock_t *lock); +extern int glthread_recursive_lock_destroy_func (gl_recursive_lock_t *lock); + +/* -------------------------- gl_once_t datatype -------------------------- */ + +typedef struct + { + volatile int inited; + volatile long started; + CRITICAL_SECTION lock; + } + gl_once_t; +# define gl_once_define(STORAGECLASS, NAME) \ + STORAGECLASS gl_once_t NAME = { -1, -1 }; +# define glthread_once(ONCE_CONTROL, INITFUNCTION) \ + (glthread_once_func (ONCE_CONTROL, INITFUNCTION), 0) +extern void glthread_once_func (gl_once_t *once_control, void (*initfunction) (void)); + +# ifdef __cplusplus +} +# endif + +#endif + +/* ========================================================================= */ + +#if !(USE_POSIX_THREADS || USE_PTH_THREADS || USE_SOLARIS_THREADS || USE_WINDOWS_THREADS) + +/* Provide dummy implementation if threads are not supported. */ + +/* -------------------------- gl_lock_t datatype -------------------------- */ + +typedef int gl_lock_t; +# define gl_lock_define(STORAGECLASS, NAME) +# define gl_lock_define_initialized(STORAGECLASS, NAME) +# define glthread_lock_init(NAME) 0 +# define glthread_lock_lock(NAME) 0 +# define glthread_lock_unlock(NAME) 0 +# define glthread_lock_destroy(NAME) 0 + +/* ------------------------- gl_rwlock_t datatype ------------------------- */ + +typedef int gl_rwlock_t; +# define gl_rwlock_define(STORAGECLASS, NAME) +# define gl_rwlock_define_initialized(STORAGECLASS, NAME) +# define glthread_rwlock_init(NAME) 0 +# define glthread_rwlock_rdlock(NAME) 0 +# define glthread_rwlock_wrlock(NAME) 0 +# define glthread_rwlock_unlock(NAME) 0 +# define glthread_rwlock_destroy(NAME) 0 + +/* --------------------- gl_recursive_lock_t datatype --------------------- */ + +typedef int gl_recursive_lock_t; +# define gl_recursive_lock_define(STORAGECLASS, NAME) +# define gl_recursive_lock_define_initialized(STORAGECLASS, NAME) +# define glthread_recursive_lock_init(NAME) 0 +# define glthread_recursive_lock_lock(NAME) 0 +# define glthread_recursive_lock_unlock(NAME) 0 +# define glthread_recursive_lock_destroy(NAME) 0 + +/* -------------------------- gl_once_t datatype -------------------------- */ + +typedef int gl_once_t; +# define gl_once_define(STORAGECLASS, NAME) \ + STORAGECLASS gl_once_t NAME = 0; +# define glthread_once(ONCE_CONTROL, INITFUNCTION) \ + (*(ONCE_CONTROL) == 0 ? (*(ONCE_CONTROL) = ~ 0, INITFUNCTION (), 0) : 0) + +#endif + +/* ========================================================================= */ + +/* Macros with built-in error handling. */ + +/* -------------------------- gl_lock_t datatype -------------------------- */ + +#define gl_lock_init(NAME) \ + do \ + { \ + if (glthread_lock_init (&NAME)) \ + abort (); \ + } \ + while (0) +#define gl_lock_lock(NAME) \ + do \ + { \ + if (glthread_lock_lock (&NAME)) \ + abort (); \ + } \ + while (0) +#define gl_lock_unlock(NAME) \ + do \ + { \ + if (glthread_lock_unlock (&NAME)) \ + abort (); \ + } \ + while (0) +#define gl_lock_destroy(NAME) \ + do \ + { \ + if (glthread_lock_destroy (&NAME)) \ + abort (); \ + } \ + while (0) + +/* ------------------------- gl_rwlock_t datatype ------------------------- */ + +#define gl_rwlock_init(NAME) \ + do \ + { \ + if (glthread_rwlock_init (&NAME)) \ + abort (); \ + } \ + while (0) +#define gl_rwlock_rdlock(NAME) \ + do \ + { \ + if (glthread_rwlock_rdlock (&NAME)) \ + abort (); \ + } \ + while (0) +#define gl_rwlock_wrlock(NAME) \ + do \ + { \ + if (glthread_rwlock_wrlock (&NAME)) \ + abort (); \ + } \ + while (0) +#define gl_rwlock_unlock(NAME) \ + do \ + { \ + if (glthread_rwlock_unlock (&NAME)) \ + abort (); \ + } \ + while (0) +#define gl_rwlock_destroy(NAME) \ + do \ + { \ + if (glthread_rwlock_destroy (&NAME)) \ + abort (); \ + } \ + while (0) + +/* --------------------- gl_recursive_lock_t datatype --------------------- */ + +#define gl_recursive_lock_init(NAME) \ + do \ + { \ + if (glthread_recursive_lock_init (&NAME)) \ + abort (); \ + } \ + while (0) +#define gl_recursive_lock_lock(NAME) \ + do \ + { \ + if (glthread_recursive_lock_lock (&NAME)) \ + abort (); \ + } \ + while (0) +#define gl_recursive_lock_unlock(NAME) \ + do \ + { \ + if (glthread_recursive_lock_unlock (&NAME)) \ + abort (); \ + } \ + while (0) +#define gl_recursive_lock_destroy(NAME) \ + do \ + { \ + if (glthread_recursive_lock_destroy (&NAME)) \ + abort (); \ + } \ + while (0) + +/* -------------------------- gl_once_t datatype -------------------------- */ + +#define gl_once(NAME, INITFUNCTION) \ + do \ + { \ + if (glthread_once (&NAME, INITFUNCTION)) \ + abort (); \ + } \ + while (0) + +/* ========================================================================= */ + +#endif /* _LOCK_H */ diff --git a/lib/glthread/threadlib.c b/lib/glthread/threadlib.c new file mode 100644 index 0000000..e8b484a --- /dev/null +++ b/lib/glthread/threadlib.c @@ -0,0 +1,73 @@ +/* Multithreading primitives. + Copyright (C) 2005-2018 Free Software Foundation, Inc. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3, or (at your option) + any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, see <https://www.gnu.org/licenses/>. */ + +/* Written by Bruno Haible <bruno@clisp.org>, 2005. */ + +#include <config.h> + +/* ========================================================================= */ + +#if USE_POSIX_THREADS + +/* Use the POSIX threads library. */ + +# include <pthread.h> +# include <stdlib.h> + +# if PTHREAD_IN_USE_DETECTION_HARD + +/* The function to be executed by a dummy thread. */ +static void * +dummy_thread_func (void *arg) +{ + return arg; +} + +int +glthread_in_use (void) +{ + static int tested; + static int result; /* 1: linked with -lpthread, 0: only with libc */ + + if (!tested) + { + pthread_t thread; + + if (pthread_create (&thread, NULL, dummy_thread_func, NULL) != 0) + /* Thread creation failed. */ + result = 0; + else + { + /* Thread creation works. */ + void *retval; + if (pthread_join (thread, &retval) != 0) + abort (); + result = 1; + } + tested = 1; + } + return result; +} + +# endif + +#endif + +/* ========================================================================= */ + +/* This declaration is solely to ensure that after preprocessing + this file is never empty. */ +typedef int dummy; diff --git a/lib/gnulib.mk b/lib/gnulib.mk new file mode 100644 index 0000000..bedc672 --- /dev/null +++ b/lib/gnulib.mk @@ -0,0 +1,2215 @@ +## DO NOT EDIT! GENERATED AUTOMATICALLY! +## Process this file with automake to produce Makefile.in. +# Copyright (C) 2002-2018 Free Software Foundation, Inc. +# +# This file is free software; you can redistribute it and/or modify +# it under the terms of the GNU General Public License as published by +# the Free Software Foundation; either version 3 of the License, or +# (at your option) any later version. +# +# This file is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with this file. If not, see <https://www.gnu.org/licenses/>. +# +# As a special exception to the GNU General Public License, +# this file may be distributed as part of a program that +# contains a configuration script generated by Autoconf, under +# the same distribution terms as the rest of that program. +# +# Generated by gnulib-tool. +# Reproduce by: +# gnulib-tool --import --local-dir=gl \ +# --lib=libsed \ +# --source-base=lib \ +# --m4-base=m4 \ +# --doc-base=doc \ +# --tests-base=gnulib-tests \ +# --aux-dir=build-aux \ +# --with-tests \ +# --makefile-name=gnulib.mk \ +# --no-conditional-dependencies \ +# --no-libtool \ +# --macro-prefix=gl \ +# --avoid=lock-tests \ +# acl \ +# alloca \ +# binary-io \ +# btowc \ +# c-ctype \ +# closeout \ +# dfa \ +# extensions \ +# fdl \ +# fwriting \ +# getdelim \ +# getopt \ +# gettext-h \ +# git-version-gen \ +# gitlog-to-changelog \ +# ignore-value \ +# localcharset \ +# manywarnings \ +# mbrlen \ +# mbrtowc \ +# mbsinit \ +# memchr \ +# memrchr \ +# mkostemp \ +# non-recursive-gnulib-prefix-hack \ +# obstack \ +# progname \ +# readme-release \ +# regex \ +# rename \ +# selinux-h \ +# ssize_t \ +# stat-macros \ +# stdalign \ +# stdbool \ +# strerror \ +# strverscmp \ +# threadlib \ +# unlocked-io \ +# update-copyright \ +# verify \ +# version-etc-fsf \ +# wcrtomb \ +# wctob \ +# xalloc + + +MOSTLYCLEANFILES += lib/core lib/*.stackdump +# No GNU Make output. + +noinst_LIBRARIES += lib/libsed.a + +lib_libsed_a_SOURCES = +lib_libsed_a_LIBADD = $(gl_LIBOBJS) +lib_libsed_a_DEPENDENCIES = $(gl_LIBOBJS) +EXTRA_lib_libsed_a_SOURCES = + +## begin gnulib module absolute-header + +# Use this preprocessor expression to decide whether #include_next works. +# Do not rely on a 'configure'-time test for this, since the expression +# might appear in an installed header, which is used by some other compiler. +HAVE_INCLUDE_NEXT = lib/(__GNUC__ lib/|| lib/60000000 lib/<= lib/__DECC_VER) + +## end gnulib module absolute-header + +## begin gnulib module acl + +lib_libsed_a_SOURCES += lib/copy-acl.c lib/set-acl.c + +## end gnulib module acl + +## begin gnulib module acl-permissions + +lib_libsed_a_SOURCES += lib/acl-errno-valid.c lib/acl-internal.c lib/get-permissions.c lib/set-permissions.c + +EXTRA_DIST += lib/acl-internal.h lib/acl.h lib/acl_entries.c + +EXTRA_lib_libsed_a_SOURCES += lib/acl_entries.c + +## end gnulib module acl-permissions + +## begin gnulib module alignof + + +EXTRA_DIST += lib/alignof.h + +## end gnulib module alignof + +## begin gnulib module alloca + + +lib_libsed_a_LIBADD += @ALLOCA@ +lib_libsed_a_DEPENDENCIES += @ALLOCA@ +EXTRA_DIST += lib/alloca.c + +EXTRA_lib_libsed_a_SOURCES += lib/alloca.c + +## end gnulib module alloca + +## begin gnulib module alloca-opt + +BUILT_SOURCES += $(ALLOCA_H) + +# We need the following in order to create <alloca.h> when the system +# doesn't have one that works with the given compiler. +if GL_GENERATE_ALLOCA_H +lib/alloca.h: lib/alloca.in.h $(top_builddir)/config.status + $(AM_V_GEN)rm -f $@-t $@ && \ + { echo '/* DO NOT EDIT! GENERATED AUTOMATICALLY! */'; \ + cat $(top_srcdir)/lib/alloca.in.h; \ + } > $@-t && \ + mv -f $@-t $@ +else +lib/alloca.h: $(top_builddir)/config.status + rm -f $@ +endif +MOSTLYCLEANFILES += lib/alloca.h lib/alloca.h-t + +EXTRA_DIST += lib/alloca.in.h + +## end gnulib module alloca-opt + +## begin gnulib module announce-gen + + +EXTRA_DIST += $(top_srcdir)/build-aux/announce-gen + +## end gnulib module announce-gen + +## begin gnulib module binary-io + +lib_libsed_a_SOURCES += lib/binary-io.h lib/binary-io.c + +## end gnulib module binary-io + +## begin gnulib module btowc + + +EXTRA_DIST += lib/btowc.c + +EXTRA_lib_libsed_a_SOURCES += lib/btowc.c + +## end gnulib module btowc + +## begin gnulib module c-ctype + +lib_libsed_a_SOURCES += lib/c-ctype.h lib/c-ctype.c + +## end gnulib module c-ctype + +## begin gnulib module c-strcase + +lib_libsed_a_SOURCES += lib/c-strcase.h lib/c-strcasecmp.c lib/c-strncasecmp.c + +## end gnulib module c-strcase + +## begin gnulib module c-strcaseeq + + +EXTRA_DIST += lib/c-strcaseeq.h + +## end gnulib module c-strcaseeq + +## begin gnulib module canonicalize-lgpl + + +EXTRA_DIST += lib/canonicalize-lgpl.c + +EXTRA_lib_libsed_a_SOURCES += lib/canonicalize-lgpl.c + +## end gnulib module canonicalize-lgpl + +## begin gnulib module close-stream + +lib_libsed_a_SOURCES += lib/close-stream.c + +EXTRA_DIST += lib/close-stream.h + +## end gnulib module close-stream + +## begin gnulib module closeout + +lib_libsed_a_SOURCES += lib/closeout.c + +EXTRA_DIST += lib/closeout.h + +## end gnulib module closeout + +## begin gnulib module ctype + +BUILT_SOURCES += lib/ctype.h + +# We need the following in order to create <ctype.h> when the system +# doesn't have one that works with the given compiler. +lib/ctype.h: lib/ctype.in.h $(top_builddir)/config.status $(CXXDEFS_H) $(WARN_ON_USE_H) + $(AM_V_GEN)rm -f $@-t $@ && \ + { echo '/* DO NOT EDIT! GENERATED AUTOMATICALLY! */'; \ + sed -e 's|@''GUARD_PREFIX''@|GL|g' \ + -e 's|@''INCLUDE_NEXT''@|$(INCLUDE_NEXT)|g' \ + -e 's|@''PRAGMA_SYSTEM_HEADER''@|@PRAGMA_SYSTEM_HEADER@|g' \ + -e 's|@''PRAGMA_COLUMNS''@|@PRAGMA_COLUMNS@|g' \ + -e 's|@''NEXT_CTYPE_H''@|$(NEXT_CTYPE_H)|g' \ + -e 's/@''GNULIB_ISBLANK''@/$(GNULIB_ISBLANK)/g' \ + -e 's/@''HAVE_ISBLANK''@/$(HAVE_ISBLANK)/g' \ + -e '/definitions of _GL_FUNCDECL_RPL/r $(CXXDEFS_H)' \ + -e '/definition of _GL_WARN_ON_USE/r $(WARN_ON_USE_H)' \ + < $(top_srcdir)/lib/ctype.in.h; \ + } > $@-t && \ + mv $@-t $@ +MOSTLYCLEANFILES += lib/ctype.h lib/ctype.h-t + +EXTRA_DIST += lib/ctype.in.h + +## end gnulib module ctype + +## begin gnulib module dfa + +lib_libsed_a_SOURCES += lib/dfa.c lib/localeinfo.c + +EXTRA_DIST += lib/dfa.h lib/localeinfo.h + +## end gnulib module dfa + +## begin gnulib module dirname-lgpl + +lib_libsed_a_SOURCES += lib/dirname-lgpl.c lib/basename-lgpl.c lib/stripslash.c + +EXTRA_DIST += lib/dirname.h + +## end gnulib module dirname-lgpl + +## begin gnulib module do-release-commit-and-tag + + +EXTRA_DIST += $(top_srcdir)/build-aux/do-release-commit-and-tag + +## end gnulib module do-release-commit-and-tag + +## begin gnulib module dosname + + +EXTRA_DIST += lib/dosname.h + +## end gnulib module dosname + +## begin gnulib module errno + +BUILT_SOURCES += $(ERRNO_H) + +# We need the following in order to create <errno.h> when the system +# doesn't have one that is POSIX compliant. +if GL_GENERATE_ERRNO_H +lib/errno.h: lib/errno.in.h $(top_builddir)/config.status + $(AM_V_GEN)rm -f $@-t $@ && \ + { echo '/* DO NOT EDIT! GENERATED AUTOMATICALLY! */' && \ + sed -e 's|@''GUARD_PREFIX''@|GL|g' \ + -e 's|@''INCLUDE_NEXT''@|$(INCLUDE_NEXT)|g' \ + -e 's|@''PRAGMA_SYSTEM_HEADER''@|@PRAGMA_SYSTEM_HEADER@|g' \ + -e 's|@''PRAGMA_COLUMNS''@|@PRAGMA_COLUMNS@|g' \ + -e 's|@''NEXT_ERRNO_H''@|$(NEXT_ERRNO_H)|g' \ + -e 's|@''EMULTIHOP_HIDDEN''@|$(EMULTIHOP_HIDDEN)|g' \ + -e 's|@''EMULTIHOP_VALUE''@|$(EMULTIHOP_VALUE)|g' \ + -e 's|@''ENOLINK_HIDDEN''@|$(ENOLINK_HIDDEN)|g' \ + -e 's|@''ENOLINK_VALUE''@|$(ENOLINK_VALUE)|g' \ + -e 's|@''EOVERFLOW_HIDDEN''@|$(EOVERFLOW_HIDDEN)|g' \ + -e 's|@''EOVERFLOW_VALUE''@|$(EOVERFLOW_VALUE)|g' \ + < $(top_srcdir)/lib/errno.in.h; \ + } > $@-t && \ + mv $@-t $@ +else +lib/errno.h: $(top_builddir)/config.status + rm -f $@ +endif +MOSTLYCLEANFILES += lib/errno.h lib/errno.h-t + +EXTRA_DIST += lib/errno.in.h + +## end gnulib module errno + +## begin gnulib module error + + +EXTRA_DIST += lib/error.c lib/error.h + +EXTRA_lib_libsed_a_SOURCES += lib/error.c + +## end gnulib module error + +## begin gnulib module exitfail + +lib_libsed_a_SOURCES += lib/exitfail.c + +EXTRA_DIST += lib/exitfail.h + +## end gnulib module exitfail + +## begin gnulib module fcntl-h + +BUILT_SOURCES += lib/fcntl.h + +# We need the following in order to create <fcntl.h> when the system +# doesn't have one that works with the given compiler. +lib/fcntl.h: lib/fcntl.in.h $(top_builddir)/config.status $(CXXDEFS_H) $(ARG_NONNULL_H) $(WARN_ON_USE_H) + $(AM_V_GEN)rm -f $@-t $@ && \ + { echo '/* DO NOT EDIT! GENERATED AUTOMATICALLY! */'; \ + sed -e 's|@''GUARD_PREFIX''@|GL|g' \ + -e 's|@''INCLUDE_NEXT''@|$(INCLUDE_NEXT)|g' \ + -e 's|@''PRAGMA_SYSTEM_HEADER''@|@PRAGMA_SYSTEM_HEADER@|g' \ + -e 's|@''PRAGMA_COLUMNS''@|@PRAGMA_COLUMNS@|g' \ + -e 's|@''NEXT_FCNTL_H''@|$(NEXT_FCNTL_H)|g' \ + -e 's/@''GNULIB_FCNTL''@/$(GNULIB_FCNTL)/g' \ + -e 's/@''GNULIB_NONBLOCKING''@/$(GNULIB_NONBLOCKING)/g' \ + -e 's/@''GNULIB_OPEN''@/$(GNULIB_OPEN)/g' \ + -e 's/@''GNULIB_OPENAT''@/$(GNULIB_OPENAT)/g' \ + -e 's|@''HAVE_FCNTL''@|$(HAVE_FCNTL)|g' \ + -e 's|@''HAVE_OPENAT''@|$(HAVE_OPENAT)|g' \ + -e 's|@''REPLACE_FCNTL''@|$(REPLACE_FCNTL)|g' \ + -e 's|@''REPLACE_OPEN''@|$(REPLACE_OPEN)|g' \ + -e 's|@''REPLACE_OPENAT''@|$(REPLACE_OPENAT)|g' \ + -e '/definitions of _GL_FUNCDECL_RPL/r $(CXXDEFS_H)' \ + -e '/definition of _GL_ARG_NONNULL/r $(ARG_NONNULL_H)' \ + -e '/definition of _GL_WARN_ON_USE/r $(WARN_ON_USE_H)' \ + < $(top_srcdir)/lib/fcntl.in.h; \ + } > $@-t && \ + mv $@-t $@ +MOSTLYCLEANFILES += lib/fcntl.h lib/fcntl.h-t + +EXTRA_DIST += lib/fcntl.in.h + +## end gnulib module fcntl-h + +## begin gnulib module filename + + +EXTRA_DIST += lib/filename.h + +## end gnulib module filename + +## begin gnulib module fpending + + +EXTRA_DIST += lib/fpending.c lib/fpending.h lib/stdio-impl.h + +EXTRA_lib_libsed_a_SOURCES += lib/fpending.c + +## end gnulib module fpending + +## begin gnulib module fstat + + +EXTRA_DIST += lib/fstat.c lib/stat-w32.c lib/stat-w32.h + +EXTRA_lib_libsed_a_SOURCES += lib/fstat.c lib/stat-w32.c + +## end gnulib module fstat + +## begin gnulib module fwriting + + +EXTRA_DIST += lib/fwriting.c lib/fwriting.h lib/stdio-impl.h + +EXTRA_lib_libsed_a_SOURCES += lib/fwriting.c + +## end gnulib module fwriting + +## begin gnulib module gendocs + + +EXTRA_DIST += $(top_srcdir)/build-aux/gendocs.sh + +## end gnulib module gendocs + +## begin gnulib module getdelim + + +EXTRA_DIST += lib/getdelim.c + +EXTRA_lib_libsed_a_SOURCES += lib/getdelim.c + +## end gnulib module getdelim + +## begin gnulib module getopt-posix + +BUILT_SOURCES += $(GETOPT_H) $(GETOPT_CDEFS_H) + +# We need the following in order to create <getopt.h> when the system +# doesn't have one that works with the given compiler. +lib/getopt.h: lib/getopt.in.h $(top_builddir)/config.status $(ARG_NONNULL_H) + $(AM_V_GEN)rm -f $@-t $@ && \ + { echo '/* DO NOT EDIT! GENERATED AUTOMATICALLY! */'; \ + sed -e 's|@''GUARD_PREFIX''@|GL|g' \ + -e 's|@''HAVE_GETOPT_H''@|$(HAVE_GETOPT_H)|g' \ + -e 's|@''INCLUDE_NEXT''@|$(INCLUDE_NEXT)|g' \ + -e 's|@''PRAGMA_SYSTEM_HEADER''@|@PRAGMA_SYSTEM_HEADER@|g' \ + -e 's|@''PRAGMA_COLUMNS''@|@PRAGMA_COLUMNS@|g' \ + -e 's|@''NEXT_GETOPT_H''@|$(NEXT_GETOPT_H)|g' \ + -e '/definition of _GL_ARG_NONNULL/r $(ARG_NONNULL_H)' \ + < $(top_srcdir)/lib/getopt.in.h; \ + } > $@-t && \ + mv -f $@-t $@ + +lib/getopt-cdefs.h: lib/getopt-cdefs.in.h $(top_builddir)/config.status + $(AM_V_GEN)rm -f $@-t $@ && \ + { echo '/* DO NOT EDIT! GENERATED AUTOMATICALLY! */'; \ + sed -e 's|@''HAVE_SYS_CDEFS_H''@|$(HAVE_SYS_CDEFS_H)|g' \ + < $(top_srcdir)/lib/getopt-cdefs.in.h; \ + } > $@-t && \ + mv -f $@-t $@ + +MOSTLYCLEANFILES += lib/getopt.h lib/getopt.h-t lib/getopt-cdefs.h lib/getopt-cdefs.h-t + +EXTRA_DIST += lib/getopt-cdefs.in.h lib/getopt-core.h lib/getopt-ext.h lib/getopt-pfx-core.h lib/getopt-pfx-ext.h lib/getopt.c lib/getopt.in.h lib/getopt1.c lib/getopt_int.h + +EXTRA_lib_libsed_a_SOURCES += lib/getopt.c lib/getopt1.c + +## end gnulib module getopt-posix + +## begin gnulib module getprogname + +lib_libsed_a_SOURCES += lib/getprogname.h lib/getprogname.c + +## end gnulib module getprogname + +## begin gnulib module gettext-h + +lib_libsed_a_SOURCES += lib/gettext.h + +## end gnulib module gettext-h + +## begin gnulib module gettimeofday + + +EXTRA_DIST += lib/gettimeofday.c + +EXTRA_lib_libsed_a_SOURCES += lib/gettimeofday.c + +## end gnulib module gettimeofday + +## begin gnulib module git-version-gen + + +EXTRA_DIST += $(top_srcdir)/build-aux/git-version-gen + +## end gnulib module git-version-gen + +## begin gnulib module gitlog-to-changelog + + +EXTRA_DIST += $(top_srcdir)/build-aux/gitlog-to-changelog + +## end gnulib module gitlog-to-changelog + +## begin gnulib module gnu-web-doc-update + + +EXTRA_DIST += $(top_srcdir)/build-aux/gnu-web-doc-update + +## end gnulib module gnu-web-doc-update + +## begin gnulib module gnumakefile + +EXTRA_DIST += $(top_srcdir)/GNUmakefile +distclean-local: clean-GNUmakefile +clean-GNUmakefile: + test '$(srcdir)' = . || rm -f $(top_builddir)/GNUmakefile + +## end gnulib module gnumakefile + +## begin gnulib module gnupload + + +EXTRA_DIST += $(top_srcdir)/build-aux/gnupload + +## end gnulib module gnupload + +## begin gnulib module hard-locale + +lib_libsed_a_SOURCES += lib/hard-locale.c + +EXTRA_DIST += lib/hard-locale.h + +## end gnulib module hard-locale + +## begin gnulib module havelib + + +EXTRA_DIST += $(top_srcdir)/build-aux/config.rpath + +## end gnulib module havelib + +## begin gnulib module ignore-value + + +EXTRA_DIST += lib/ignore-value.h + +## end gnulib module ignore-value + +## begin gnulib module intprops + + +EXTRA_DIST += lib/intprops.h + +## end gnulib module intprops + +## begin gnulib module isblank + + +EXTRA_DIST += lib/isblank.c + +EXTRA_lib_libsed_a_SOURCES += lib/isblank.c + +## end gnulib module isblank + +## begin gnulib module langinfo + +BUILT_SOURCES += lib/langinfo.h + +# We need the following in order to create an empty placeholder for +# <langinfo.h> when the system doesn't have one. +lib/langinfo.h: lib/langinfo.in.h $(top_builddir)/config.status $(CXXDEFS_H) $(WARN_ON_USE_H) + $(AM_V_GEN)rm -f $@-t $@ && \ + { echo '/* DO NOT EDIT! GENERATED AUTOMATICALLY! */'; \ + sed -e 's|@''GUARD_PREFIX''@|GL|g' \ + -e 's|@''HAVE_LANGINFO_H''@|$(HAVE_LANGINFO_H)|g' \ + -e 's|@''INCLUDE_NEXT''@|$(INCLUDE_NEXT)|g' \ + -e 's|@''PRAGMA_SYSTEM_HEADER''@|@PRAGMA_SYSTEM_HEADER@|g' \ + -e 's|@''PRAGMA_COLUMNS''@|@PRAGMA_COLUMNS@|g' \ + -e 's|@''NEXT_LANGINFO_H''@|$(NEXT_LANGINFO_H)|g' \ + -e 's/@''GNULIB_NL_LANGINFO''@/$(GNULIB_NL_LANGINFO)/g' \ + -e 's|@''HAVE_LANGINFO_CODESET''@|$(HAVE_LANGINFO_CODESET)|g' \ + -e 's|@''HAVE_LANGINFO_T_FMT_AMPM''@|$(HAVE_LANGINFO_T_FMT_AMPM)|g' \ + -e 's|@''HAVE_LANGINFO_ALTMON''@|$(HAVE_LANGINFO_ALTMON)|g' \ + -e 's|@''HAVE_LANGINFO_ERA''@|$(HAVE_LANGINFO_ERA)|g' \ + -e 's|@''HAVE_LANGINFO_YESEXPR''@|$(HAVE_LANGINFO_YESEXPR)|g' \ + -e 's|@''HAVE_NL_LANGINFO''@|$(HAVE_NL_LANGINFO)|g' \ + -e 's|@''REPLACE_NL_LANGINFO''@|$(REPLACE_NL_LANGINFO)|g' \ + -e '/definitions of _GL_FUNCDECL_RPL/r $(CXXDEFS_H)' \ + -e '/definition of _GL_WARN_ON_USE/r $(WARN_ON_USE_H)' \ + < $(top_srcdir)/lib/langinfo.in.h; \ + } > $@-t && \ + mv $@-t $@ +MOSTLYCLEANFILES += lib/langinfo.h lib/langinfo.h-t + +EXTRA_DIST += lib/langinfo.in.h + +## end gnulib module langinfo + +## begin gnulib module libc-config + + +EXTRA_DIST += lib/cdefs.h lib/libc-config.h + +## end gnulib module libc-config + +## begin gnulib module limits-h + +BUILT_SOURCES += $(LIMITS_H) + +# We need the following in order to create <limits.h> when the system +# doesn't have one that is compatible with GNU. +if GL_GENERATE_LIMITS_H +lib/limits.h: lib/limits.in.h $(top_builddir)/config.status + $(AM_V_GEN)rm -f $@-t $@ && \ + { echo '/* DO NOT EDIT! GENERATED AUTOMATICALLY! */' && \ + sed -e 's|@''GUARD_PREFIX''@|GL|g' \ + -e 's|@''INCLUDE_NEXT''@|$(INCLUDE_NEXT)|g' \ + -e 's|@''PRAGMA_SYSTEM_HEADER''@|@PRAGMA_SYSTEM_HEADER@|g' \ + -e 's|@''PRAGMA_COLUMNS''@|@PRAGMA_COLUMNS@|g' \ + -e 's|@''NEXT_LIMITS_H''@|$(NEXT_LIMITS_H)|g' \ + < $(top_srcdir)/lib/limits.in.h; \ + } > $@-t && \ + mv $@-t $@ +else +lib/limits.h: $(top_builddir)/config.status + rm -f $@ +endif +MOSTLYCLEANFILES += lib/limits.h lib/limits.h-t + +EXTRA_DIST += lib/limits.in.h + +## end gnulib module limits-h + +## begin gnulib module localcharset + +lib_libsed_a_SOURCES += lib/localcharset.c + +EXTRA_DIST += lib/localcharset.h + +## end gnulib module localcharset + +## begin gnulib module locale + +BUILT_SOURCES += lib/locale.h + +# We need the following in order to create <locale.h> when the system +# doesn't have one that provides all definitions. +lib/locale.h: lib/locale.in.h $(top_builddir)/config.status $(CXXDEFS_H) $(ARG_NONNULL_H) $(WARN_ON_USE_H) + $(AM_V_GEN)rm -f $@-t $@ && \ + { echo '/* DO NOT EDIT! GENERATED AUTOMATICALLY! */' && \ + sed -e 's|@''GUARD_PREFIX''@|GL|g' \ + -e 's|@''INCLUDE_NEXT''@|$(INCLUDE_NEXT)|g' \ + -e 's|@''PRAGMA_SYSTEM_HEADER''@|@PRAGMA_SYSTEM_HEADER@|g' \ + -e 's|@''PRAGMA_COLUMNS''@|@PRAGMA_COLUMNS@|g' \ + -e 's|@''NEXT_LOCALE_H''@|$(NEXT_LOCALE_H)|g' \ + -e 's/@''GNULIB_LOCALECONV''@/$(GNULIB_LOCALECONV)/g' \ + -e 's/@''GNULIB_SETLOCALE''@/$(GNULIB_SETLOCALE)/g' \ + -e 's/@''GNULIB_DUPLOCALE''@/$(GNULIB_DUPLOCALE)/g' \ + -e 's/@''GNULIB_LOCALENAME''@/$(GNULIB_LOCALENAME)/g' \ + -e 's|@''HAVE_NEWLOCALE''@|$(HAVE_NEWLOCALE)|g' \ + -e 's|@''HAVE_DUPLOCALE''@|$(HAVE_DUPLOCALE)|g' \ + -e 's|@''HAVE_FREELOCALE''@|$(HAVE_FREELOCALE)|g' \ + -e 's|@''HAVE_XLOCALE_H''@|$(HAVE_XLOCALE_H)|g' \ + -e 's|@''REPLACE_LOCALECONV''@|$(REPLACE_LOCALECONV)|g' \ + -e 's|@''REPLACE_SETLOCALE''@|$(REPLACE_SETLOCALE)|g' \ + -e 's|@''REPLACE_NEWLOCALE''@|$(REPLACE_NEWLOCALE)|g' \ + -e 's|@''REPLACE_DUPLOCALE''@|$(REPLACE_DUPLOCALE)|g' \ + -e 's|@''REPLACE_FREELOCALE''@|$(REPLACE_FREELOCALE)|g' \ + -e 's|@''REPLACE_STRUCT_LCONV''@|$(REPLACE_STRUCT_LCONV)|g' \ + -e '/definitions of _GL_FUNCDECL_RPL/r $(CXXDEFS_H)' \ + -e '/definition of _GL_ARG_NONNULL/r $(ARG_NONNULL_H)' \ + -e '/definition of _GL_WARN_ON_USE/r $(WARN_ON_USE_H)' \ + < $(top_srcdir)/lib/locale.in.h; \ + } > $@-t && \ + mv $@-t $@ +MOSTLYCLEANFILES += lib/locale.h lib/locale.h-t + +EXTRA_DIST += lib/locale.in.h + +## end gnulib module locale + +## begin gnulib module localeconv + + +EXTRA_DIST += lib/localeconv.c + +EXTRA_lib_libsed_a_SOURCES += lib/localeconv.c + +## end gnulib module localeconv + +## begin gnulib module localtime-buffer + + +EXTRA_DIST += lib/localtime-buffer.c lib/localtime-buffer.h + +EXTRA_lib_libsed_a_SOURCES += lib/localtime-buffer.c + +## end gnulib module localtime-buffer + +## begin gnulib module lock + +lib_libsed_a_SOURCES += lib/glthread/lock.h lib/glthread/lock.c + +## end gnulib module lock + +## begin gnulib module lstat + + +EXTRA_DIST += lib/lstat.c + +EXTRA_lib_libsed_a_SOURCES += lib/lstat.c + +## end gnulib module lstat + +## begin gnulib module maintainer-makefile + +EXTRA_DIST += $(top_srcdir)/maint.mk + +## end gnulib module maintainer-makefile + +## begin gnulib module malloc-posix + + +EXTRA_DIST += lib/malloc.c + +EXTRA_lib_libsed_a_SOURCES += lib/malloc.c + +## end gnulib module malloc-posix + +## begin gnulib module malloca + +lib_libsed_a_SOURCES += lib/malloca.c + +EXTRA_DIST += lib/malloca.h + +## end gnulib module malloca + +## begin gnulib module mbrlen + + +EXTRA_DIST += lib/mbrlen.c + +EXTRA_lib_libsed_a_SOURCES += lib/mbrlen.c + +## end gnulib module mbrlen + +## begin gnulib module mbrtowc + + +EXTRA_DIST += lib/mbrtowc.c + +EXTRA_lib_libsed_a_SOURCES += lib/mbrtowc.c + +## end gnulib module mbrtowc + +## begin gnulib module mbsinit + + +EXTRA_DIST += lib/mbsinit.c + +EXTRA_lib_libsed_a_SOURCES += lib/mbsinit.c + +## end gnulib module mbsinit + +## begin gnulib module mbtowc + + +EXTRA_DIST += lib/mbtowc-impl.h lib/mbtowc.c + +EXTRA_lib_libsed_a_SOURCES += lib/mbtowc.c + +## end gnulib module mbtowc + +## begin gnulib module memchr + + +EXTRA_DIST += lib/memchr.c lib/memchr.valgrind + +EXTRA_lib_libsed_a_SOURCES += lib/memchr.c + +## end gnulib module memchr + +## begin gnulib module memrchr + + +EXTRA_DIST += lib/memrchr.c + +EXTRA_lib_libsed_a_SOURCES += lib/memrchr.c + +## end gnulib module memrchr + +## begin gnulib module minmax + +lib_libsed_a_SOURCES += lib/minmax.h + +## end gnulib module minmax + +## begin gnulib module mkdir + + +EXTRA_DIST += lib/mkdir.c + +EXTRA_lib_libsed_a_SOURCES += lib/mkdir.c + +## end gnulib module mkdir + +## begin gnulib module mkostemp + + +EXTRA_DIST += lib/mkostemp.c + +EXTRA_lib_libsed_a_SOURCES += lib/mkostemp.c + +## end gnulib module mkostemp + +## begin gnulib module msvc-inval + + +EXTRA_DIST += lib/msvc-inval.c lib/msvc-inval.h + +EXTRA_lib_libsed_a_SOURCES += lib/msvc-inval.c + +## end gnulib module msvc-inval + +## begin gnulib module msvc-nothrow + + +EXTRA_DIST += lib/msvc-nothrow.c lib/msvc-nothrow.h + +EXTRA_lib_libsed_a_SOURCES += lib/msvc-nothrow.c + +## end gnulib module msvc-nothrow + +## begin gnulib module nl_langinfo + + +EXTRA_DIST += lib/nl_langinfo.c + +EXTRA_lib_libsed_a_SOURCES += lib/nl_langinfo.c + +## end gnulib module nl_langinfo + +## begin gnulib module non-recursive-gnulib-prefix-hack + + +EXTRA_DIST += $(top_srcdir)/build-aux/prefix-gnulib-mk + +## end gnulib module non-recursive-gnulib-prefix-hack + +## begin gnulib module obstack + + +EXTRA_DIST += lib/obstack.c lib/obstack.h + +EXTRA_lib_libsed_a_SOURCES += lib/obstack.c + +## end gnulib module obstack + +## begin gnulib module pathmax + + +EXTRA_DIST += lib/pathmax.h + +## end gnulib module pathmax + +## begin gnulib module progname + +lib_libsed_a_SOURCES += lib/progname.h lib/progname.c + +## end gnulib module progname + +## begin gnulib module qcopy-acl + +lib_libsed_a_SOURCES += lib/qcopy-acl.c + +## end gnulib module qcopy-acl + +## begin gnulib module qset-acl + +lib_libsed_a_SOURCES += lib/qset-acl.c + +## end gnulib module qset-acl + +## begin gnulib module quote + + +EXTRA_DIST += lib/quote.h + +## end gnulib module quote + +## begin gnulib module quotearg + +lib_libsed_a_SOURCES += lib/quotearg.c + +EXTRA_DIST += lib/quote.h lib/quotearg.h + +## end gnulib module quotearg + +## begin gnulib module readlink + + +EXTRA_DIST += lib/readlink.c + +EXTRA_lib_libsed_a_SOURCES += lib/readlink.c + +## end gnulib module readlink + +## begin gnulib module regex + + +EXTRA_DIST += lib/regcomp.c lib/regex.c lib/regex.h lib/regex_internal.c lib/regex_internal.h lib/regexec.c + +EXTRA_lib_libsed_a_SOURCES += lib/regcomp.c lib/regex.c lib/regex_internal.c lib/regexec.c + +## end gnulib module regex + +## begin gnulib module rename + + +EXTRA_DIST += lib/rename.c + +EXTRA_lib_libsed_a_SOURCES += lib/rename.c + +## end gnulib module rename + +## begin gnulib module rmdir + + +EXTRA_DIST += lib/rmdir.c + +EXTRA_lib_libsed_a_SOURCES += lib/rmdir.c + +## end gnulib module rmdir + +## begin gnulib module same-inode + + +EXTRA_DIST += lib/same-inode.h + +## end gnulib module same-inode + +## begin gnulib module selinux-h + +lib_libsed_a_SOURCES += lib/se-context.in.h lib/se-selinux.in.h lib/se-context.c lib/se-selinux.c + +BUILT_SOURCES += lib/selinux/selinux.h +lib/selinux/selinux.h: lib/se-selinux.in.h $(top_builddir)/config.status $(UNUSED_PARAMETER_H) + $(AM_V_at)$(MKDIR_P) lib/selinux + $(AM_V_GEN)rm -f $@-t $@ && \ + { echo '/* DO NOT EDIT! GENERATED AUTOMATICALLY! */' && \ + sed -e 's|@''GUARD_PREFIX''@|GL|g' \ + -e 's|@''INCLUDE_NEXT''@|$(INCLUDE_NEXT)|g' \ + -e 's|@''PRAGMA_SYSTEM_HEADER''@|@PRAGMA_SYSTEM_HEADER@|g' \ + -e 's|@''PRAGMA_COLUMNS''@|@PRAGMA_COLUMNS@|g' \ + -e 's|@''NEXT_SELINUX_SELINUX_H''@|$(NEXT_SELINUX_SELINUX_H)|g' \ + -e '/definition of _GL_UNUSED_PARAMETER/r $(UNUSED_PARAMETER_H)' \ + < $(top_srcdir)/lib/se-selinux.in.h; \ + } > $@-t && \ + chmod a-x $@-t && \ + mv $@-t $@ +MOSTLYCLEANFILES += lib/selinux/selinux.h lib/selinux/selinux.h-t + +BUILT_SOURCES += $(SELINUX_CONTEXT_H) +if GL_GENERATE_SELINUX_CONTEXT_H +lib/selinux/context.h: lib/se-context.in.h $(top_builddir)/config.status $(UNUSED_PARAMETER_H) + $(AM_V_at)$(MKDIR_P) lib/selinux + $(AM_V_GEN)rm -f $@-t $@ && \ + { echo '/* DO NOT EDIT! GENERATED AUTOMATICALLY! */' && \ + sed -e '/definition of _GL_UNUSED_PARAMETER/r $(UNUSED_PARAMETER_H)' \ + < $(top_srcdir)/lib/se-context.in.h; \ + } > $@-t && \ + chmod a-x $@-t && \ + mv $@-t $@ +else +lib/selinux/context.h: $(top_builddir)/config.status + rm -f $@ +endif +MOSTLYCLEANFILES += lib/selinux/context.h lib/selinux/context.h-t +MOSTLYCLEANDIRS += lib/selinux + +EXTRA_DIST += lib/getfilecon.c + +EXTRA_lib_libsed_a_SOURCES += lib/getfilecon.c + +## end gnulib module selinux-h + +## begin gnulib module snippet/_Noreturn + +# Because this Makefile snippet defines a variable used by other +# gnulib Makefile snippets, it must be present in all makefiles that +# need it. This is ensured by the applicability 'all' defined above. + +_NORETURN_H=$(top_srcdir)/lib/_Noreturn.h + +EXTRA_DIST += lib/_Noreturn.h + +## end gnulib module snippet/_Noreturn + +## begin gnulib module snippet/arg-nonnull + +# Because this Makefile snippet defines a variable used by other +# gnulib Makefile snippets, it must be present in all makefiles that +# need it. This is ensured by the applicability 'all' defined above. + +ARG_NONNULL_H=$(top_srcdir)/lib/arg-nonnull.h + +EXTRA_DIST += lib/arg-nonnull.h + +## end gnulib module snippet/arg-nonnull + +## begin gnulib module snippet/c++defs + +# Because this Makefile snippet defines a variable used by other +# gnulib Makefile snippets, it must be present in all makefiles that +# need it. This is ensured by the applicability 'all' defined above. + +CXXDEFS_H=$(top_srcdir)/lib/c++defs.h + +EXTRA_DIST += lib/c++defs.h + +## end gnulib module snippet/c++defs + +## begin gnulib module snippet/unused-parameter + +# Because this Makefile snippet defines a variable used by other +# gnulib Makefile snippets, it must be present in all makefiles that +# need it. This is ensured by the applicability 'all' defined above. + +UNUSED_PARAMETER_H=$(top_srcdir)/lib/unused-parameter.h + +EXTRA_DIST += lib/unused-parameter.h + +## end gnulib module snippet/unused-parameter + +## begin gnulib module snippet/warn-on-use + +# Because this Makefile snippet defines a variable used by other +# gnulib Makefile snippets, it must be present in all makefiles that +# need it. This is ensured by the applicability 'all' defined above. + +WARN_ON_USE_H=$(top_srcdir)/lib/warn-on-use.h + +EXTRA_DIST += lib/warn-on-use.h + +## end gnulib module snippet/warn-on-use + +## begin gnulib module stat + + +EXTRA_DIST += lib/stat-w32.c lib/stat-w32.h lib/stat.c + +EXTRA_lib_libsed_a_SOURCES += lib/stat-w32.c lib/stat.c + +## end gnulib module stat + +## begin gnulib module stat-macros + + +EXTRA_DIST += lib/stat-macros.h + +## end gnulib module stat-macros + +## begin gnulib module stat-time + +lib_libsed_a_SOURCES += lib/stat-time.c + +EXTRA_DIST += lib/stat-time.h + +## end gnulib module stat-time + +## begin gnulib module stdalign + +BUILT_SOURCES += $(STDALIGN_H) + +# We need the following in order to create <stdalign.h> when the system +# doesn't have one that works. +if GL_GENERATE_STDALIGN_H +lib/stdalign.h: lib/stdalign.in.h $(top_builddir)/config.status + $(AM_V_GEN)rm -f $@-t $@ && \ + { echo '/* DO NOT EDIT! GENERATED AUTOMATICALLY! */'; \ + cat $(top_srcdir)/lib/stdalign.in.h; \ + } > $@-t && \ + mv $@-t $@ +else +lib/stdalign.h: $(top_builddir)/config.status + rm -f $@ +endif +MOSTLYCLEANFILES += lib/stdalign.h lib/stdalign.h-t + +EXTRA_DIST += lib/stdalign.in.h + +## end gnulib module stdalign + +## begin gnulib module stdarg + +BUILT_SOURCES += $(STDARG_H) + +# We need the following in order to create <stdarg.h> when the system +# doesn't have one that works with the given compiler. +if GL_GENERATE_STDARG_H +lib/stdarg.h: lib/stdarg.in.h $(top_builddir)/config.status + $(AM_V_GEN)rm -f $@-t $@ && \ + { echo '/* DO NOT EDIT! GENERATED AUTOMATICALLY! */' && \ + sed -e 's|@''GUARD_PREFIX''@|GL|g' \ + -e 's|@''INCLUDE_NEXT''@|$(INCLUDE_NEXT)|g' \ + -e 's|@''PRAGMA_SYSTEM_HEADER''@|@PRAGMA_SYSTEM_HEADER@|g' \ + -e 's|@''PRAGMA_COLUMNS''@|@PRAGMA_COLUMNS@|g' \ + -e 's|@''NEXT_STDARG_H''@|$(NEXT_STDARG_H)|g' \ + < $(top_srcdir)/lib/stdarg.in.h; \ + } > $@-t && \ + mv $@-t $@ +else +lib/stdarg.h: $(top_builddir)/config.status + rm -f $@ +endif +MOSTLYCLEANFILES += lib/stdarg.h lib/stdarg.h-t + +EXTRA_DIST += lib/stdarg.in.h + +## end gnulib module stdarg + +## begin gnulib module stdbool + +BUILT_SOURCES += $(STDBOOL_H) + +# We need the following in order to create <stdbool.h> when the system +# doesn't have one that works. +if GL_GENERATE_STDBOOL_H +lib/stdbool.h: lib/stdbool.in.h $(top_builddir)/config.status + $(AM_V_GEN)rm -f $@-t $@ && \ + { echo '/* DO NOT EDIT! GENERATED AUTOMATICALLY! */'; \ + sed -e 's/@''HAVE__BOOL''@/$(HAVE__BOOL)/g' < $(top_srcdir)/lib/stdbool.in.h; \ + } > $@-t && \ + mv $@-t $@ +else +lib/stdbool.h: $(top_builddir)/config.status + rm -f $@ +endif +MOSTLYCLEANFILES += lib/stdbool.h lib/stdbool.h-t + +EXTRA_DIST += lib/stdbool.in.h + +## end gnulib module stdbool + +## begin gnulib module stddef + +BUILT_SOURCES += $(STDDEF_H) + +# We need the following in order to create <stddef.h> when the system +# doesn't have one that works with the given compiler. +if GL_GENERATE_STDDEF_H +lib/stddef.h: lib/stddef.in.h $(top_builddir)/config.status + $(AM_V_GEN)rm -f $@-t $@ && \ + { echo '/* DO NOT EDIT! GENERATED AUTOMATICALLY! */' && \ + sed -e 's|@''GUARD_PREFIX''@|GL|g' \ + -e 's|@''INCLUDE_NEXT''@|$(INCLUDE_NEXT)|g' \ + -e 's|@''PRAGMA_SYSTEM_HEADER''@|@PRAGMA_SYSTEM_HEADER@|g' \ + -e 's|@''PRAGMA_COLUMNS''@|@PRAGMA_COLUMNS@|g' \ + -e 's|@''NEXT_STDDEF_H''@|$(NEXT_STDDEF_H)|g' \ + -e 's|@''HAVE_MAX_ALIGN_T''@|$(HAVE_MAX_ALIGN_T)|g' \ + -e 's|@''HAVE_WCHAR_T''@|$(HAVE_WCHAR_T)|g' \ + -e 's|@''REPLACE_NULL''@|$(REPLACE_NULL)|g' \ + < $(top_srcdir)/lib/stddef.in.h; \ + } > $@-t && \ + mv $@-t $@ +else +lib/stddef.h: $(top_builddir)/config.status + rm -f $@ +endif +MOSTLYCLEANFILES += lib/stddef.h lib/stddef.h-t + +EXTRA_DIST += lib/stddef.in.h + +## end gnulib module stddef + +## begin gnulib module stdint + +BUILT_SOURCES += $(STDINT_H) + +# We need the following in order to create <stdint.h> when the system +# doesn't have one that works with the given compiler. +if GL_GENERATE_STDINT_H +lib/stdint.h: lib/stdint.in.h $(top_builddir)/config.status + $(AM_V_GEN)rm -f $@-t $@ && \ + { echo '/* DO NOT EDIT! GENERATED AUTOMATICALLY! */'; \ + sed -e 's|@''GUARD_PREFIX''@|GL|g' \ + -e 's/@''HAVE_STDINT_H''@/$(HAVE_STDINT_H)/g' \ + -e 's|@''INCLUDE_NEXT''@|$(INCLUDE_NEXT)|g' \ + -e 's|@''PRAGMA_SYSTEM_HEADER''@|@PRAGMA_SYSTEM_HEADER@|g' \ + -e 's|@''PRAGMA_COLUMNS''@|@PRAGMA_COLUMNS@|g' \ + -e 's|@''NEXT_STDINT_H''@|$(NEXT_STDINT_H)|g' \ + -e 's/@''HAVE_C99_STDINT_H''@/$(HAVE_C99_STDINT_H)/g' \ + -e 's/@''HAVE_SYS_TYPES_H''@/$(HAVE_SYS_TYPES_H)/g' \ + -e 's/@''HAVE_INTTYPES_H''@/$(HAVE_INTTYPES_H)/g' \ + -e 's/@''HAVE_SYS_INTTYPES_H''@/$(HAVE_SYS_INTTYPES_H)/g' \ + -e 's/@''HAVE_SYS_BITYPES_H''@/$(HAVE_SYS_BITYPES_H)/g' \ + -e 's/@''HAVE_WCHAR_H''@/$(HAVE_WCHAR_H)/g' \ + -e 's/@''HAVE_LONG_LONG_INT''@/$(HAVE_LONG_LONG_INT)/g' \ + -e 's/@''HAVE_UNSIGNED_LONG_LONG_INT''@/$(HAVE_UNSIGNED_LONG_LONG_INT)/g' \ + -e 's/@''APPLE_UNIVERSAL_BUILD''@/$(APPLE_UNIVERSAL_BUILD)/g' \ + -e 's/@''BITSIZEOF_PTRDIFF_T''@/$(BITSIZEOF_PTRDIFF_T)/g' \ + -e 's/@''PTRDIFF_T_SUFFIX''@/$(PTRDIFF_T_SUFFIX)/g' \ + -e 's/@''BITSIZEOF_SIG_ATOMIC_T''@/$(BITSIZEOF_SIG_ATOMIC_T)/g' \ + -e 's/@''HAVE_SIGNED_SIG_ATOMIC_T''@/$(HAVE_SIGNED_SIG_ATOMIC_T)/g' \ + -e 's/@''SIG_ATOMIC_T_SUFFIX''@/$(SIG_ATOMIC_T_SUFFIX)/g' \ + -e 's/@''BITSIZEOF_SIZE_T''@/$(BITSIZEOF_SIZE_T)/g' \ + -e 's/@''SIZE_T_SUFFIX''@/$(SIZE_T_SUFFIX)/g' \ + -e 's/@''BITSIZEOF_WCHAR_T''@/$(BITSIZEOF_WCHAR_T)/g' \ + -e 's/@''HAVE_SIGNED_WCHAR_T''@/$(HAVE_SIGNED_WCHAR_T)/g' \ + -e 's/@''WCHAR_T_SUFFIX''@/$(WCHAR_T_SUFFIX)/g' \ + -e 's/@''BITSIZEOF_WINT_T''@/$(BITSIZEOF_WINT_T)/g' \ + -e 's/@''HAVE_SIGNED_WINT_T''@/$(HAVE_SIGNED_WINT_T)/g' \ + -e 's/@''WINT_T_SUFFIX''@/$(WINT_T_SUFFIX)/g' \ + -e 's/@''GNULIB_OVERRIDES_WINT_T''@/$(GNULIB_OVERRIDES_WINT_T)/g' \ + < $(top_srcdir)/lib/stdint.in.h; \ + } > $@-t && \ + mv $@-t $@ +else +lib/stdint.h: $(top_builddir)/config.status + rm -f $@ +endif +MOSTLYCLEANFILES += lib/stdint.h lib/stdint.h-t + +EXTRA_DIST += lib/stdint.in.h + +## end gnulib module stdint + +## begin gnulib module stdio + +BUILT_SOURCES += lib/stdio.h + +# We need the following in order to create <stdio.h> when the system +# doesn't have one that works with the given compiler. +lib/stdio.h: lib/stdio.in.h $(top_builddir)/config.status $(CXXDEFS_H) $(ARG_NONNULL_H) $(WARN_ON_USE_H) + $(AM_V_GEN)rm -f $@-t $@ && \ + { echo '/* DO NOT EDIT! GENERATED AUTOMATICALLY! */' && \ + sed -e 's|@''GUARD_PREFIX''@|GL|g' \ + -e 's|@''INCLUDE_NEXT''@|$(INCLUDE_NEXT)|g' \ + -e 's|@''PRAGMA_SYSTEM_HEADER''@|@PRAGMA_SYSTEM_HEADER@|g' \ + -e 's|@''PRAGMA_COLUMNS''@|@PRAGMA_COLUMNS@|g' \ + -e 's|@''NEXT_STDIO_H''@|$(NEXT_STDIO_H)|g' \ + -e 's/@''GNULIB_DPRINTF''@/$(GNULIB_DPRINTF)/g' \ + -e 's/@''GNULIB_FCLOSE''@/$(GNULIB_FCLOSE)/g' \ + -e 's/@''GNULIB_FDOPEN''@/$(GNULIB_FDOPEN)/g' \ + -e 's/@''GNULIB_FFLUSH''@/$(GNULIB_FFLUSH)/g' \ + -e 's/@''GNULIB_FGETC''@/$(GNULIB_FGETC)/g' \ + -e 's/@''GNULIB_FGETS''@/$(GNULIB_FGETS)/g' \ + -e 's/@''GNULIB_FOPEN''@/$(GNULIB_FOPEN)/g' \ + -e 's/@''GNULIB_FPRINTF''@/$(GNULIB_FPRINTF)/g' \ + -e 's/@''GNULIB_FPRINTF_POSIX''@/$(GNULIB_FPRINTF_POSIX)/g' \ + -e 's/@''GNULIB_FPURGE''@/$(GNULIB_FPURGE)/g' \ + -e 's/@''GNULIB_FPUTC''@/$(GNULIB_FPUTC)/g' \ + -e 's/@''GNULIB_FPUTS''@/$(GNULIB_FPUTS)/g' \ + -e 's/@''GNULIB_FREAD''@/$(GNULIB_FREAD)/g' \ + -e 's/@''GNULIB_FREOPEN''@/$(GNULIB_FREOPEN)/g' \ + -e 's/@''GNULIB_FSCANF''@/$(GNULIB_FSCANF)/g' \ + -e 's/@''GNULIB_FSEEK''@/$(GNULIB_FSEEK)/g' \ + -e 's/@''GNULIB_FSEEKO''@/$(GNULIB_FSEEKO)/g' \ + -e 's/@''GNULIB_FTELL''@/$(GNULIB_FTELL)/g' \ + -e 's/@''GNULIB_FTELLO''@/$(GNULIB_FTELLO)/g' \ + -e 's/@''GNULIB_FWRITE''@/$(GNULIB_FWRITE)/g' \ + -e 's/@''GNULIB_GETC''@/$(GNULIB_GETC)/g' \ + -e 's/@''GNULIB_GETCHAR''@/$(GNULIB_GETCHAR)/g' \ + -e 's/@''GNULIB_GETDELIM''@/$(GNULIB_GETDELIM)/g' \ + -e 's/@''GNULIB_GETLINE''@/$(GNULIB_GETLINE)/g' \ + -e 's/@''GNULIB_OBSTACK_PRINTF''@/$(GNULIB_OBSTACK_PRINTF)/g' \ + -e 's/@''GNULIB_OBSTACK_PRINTF_POSIX''@/$(GNULIB_OBSTACK_PRINTF_POSIX)/g' \ + -e 's/@''GNULIB_PCLOSE''@/$(GNULIB_PCLOSE)/g' \ + -e 's/@''GNULIB_PERROR''@/$(GNULIB_PERROR)/g' \ + -e 's/@''GNULIB_POPEN''@/$(GNULIB_POPEN)/g' \ + -e 's/@''GNULIB_PRINTF''@/$(GNULIB_PRINTF)/g' \ + -e 's/@''GNULIB_PRINTF_POSIX''@/$(GNULIB_PRINTF_POSIX)/g' \ + -e 's/@''GNULIB_PUTC''@/$(GNULIB_PUTC)/g' \ + -e 's/@''GNULIB_PUTCHAR''@/$(GNULIB_PUTCHAR)/g' \ + -e 's/@''GNULIB_PUTS''@/$(GNULIB_PUTS)/g' \ + -e 's/@''GNULIB_REMOVE''@/$(GNULIB_REMOVE)/g' \ + -e 's/@''GNULIB_RENAME''@/$(GNULIB_RENAME)/g' \ + -e 's/@''GNULIB_RENAMEAT''@/$(GNULIB_RENAMEAT)/g' \ + -e 's/@''GNULIB_SCANF''@/$(GNULIB_SCANF)/g' \ + -e 's/@''GNULIB_SNPRINTF''@/$(GNULIB_SNPRINTF)/g' \ + -e 's/@''GNULIB_SPRINTF_POSIX''@/$(GNULIB_SPRINTF_POSIX)/g' \ + -e 's/@''GNULIB_STDIO_H_NONBLOCKING''@/$(GNULIB_STDIO_H_NONBLOCKING)/g' \ + -e 's/@''GNULIB_STDIO_H_SIGPIPE''@/$(GNULIB_STDIO_H_SIGPIPE)/g' \ + -e 's/@''GNULIB_TMPFILE''@/$(GNULIB_TMPFILE)/g' \ + -e 's/@''GNULIB_VASPRINTF''@/$(GNULIB_VASPRINTF)/g' \ + -e 's/@''GNULIB_VDPRINTF''@/$(GNULIB_VDPRINTF)/g' \ + -e 's/@''GNULIB_VFPRINTF''@/$(GNULIB_VFPRINTF)/g' \ + -e 's/@''GNULIB_VFPRINTF_POSIX''@/$(GNULIB_VFPRINTF_POSIX)/g' \ + -e 's/@''GNULIB_VFSCANF''@/$(GNULIB_VFSCANF)/g' \ + -e 's/@''GNULIB_VSCANF''@/$(GNULIB_VSCANF)/g' \ + -e 's/@''GNULIB_VPRINTF''@/$(GNULIB_VPRINTF)/g' \ + -e 's/@''GNULIB_VPRINTF_POSIX''@/$(GNULIB_VPRINTF_POSIX)/g' \ + -e 's/@''GNULIB_VSNPRINTF''@/$(GNULIB_VSNPRINTF)/g' \ + -e 's/@''GNULIB_VSPRINTF_POSIX''@/$(GNULIB_VSPRINTF_POSIX)/g' \ + < $(top_srcdir)/lib/stdio.in.h | \ + sed -e 's|@''HAVE_DECL_FPURGE''@|$(HAVE_DECL_FPURGE)|g' \ + -e 's|@''HAVE_DECL_FSEEKO''@|$(HAVE_DECL_FSEEKO)|g' \ + -e 's|@''HAVE_DECL_FTELLO''@|$(HAVE_DECL_FTELLO)|g' \ + -e 's|@''HAVE_DECL_GETDELIM''@|$(HAVE_DECL_GETDELIM)|g' \ + -e 's|@''HAVE_DECL_GETLINE''@|$(HAVE_DECL_GETLINE)|g' \ + -e 's|@''HAVE_DECL_OBSTACK_PRINTF''@|$(HAVE_DECL_OBSTACK_PRINTF)|g' \ + -e 's|@''HAVE_DECL_SNPRINTF''@|$(HAVE_DECL_SNPRINTF)|g' \ + -e 's|@''HAVE_DECL_VSNPRINTF''@|$(HAVE_DECL_VSNPRINTF)|g' \ + -e 's|@''HAVE_DPRINTF''@|$(HAVE_DPRINTF)|g' \ + -e 's|@''HAVE_FSEEKO''@|$(HAVE_FSEEKO)|g' \ + -e 's|@''HAVE_FTELLO''@|$(HAVE_FTELLO)|g' \ + -e 's|@''HAVE_PCLOSE''@|$(HAVE_PCLOSE)|g' \ + -e 's|@''HAVE_POPEN''@|$(HAVE_POPEN)|g' \ + -e 's|@''HAVE_RENAMEAT''@|$(HAVE_RENAMEAT)|g' \ + -e 's|@''HAVE_VASPRINTF''@|$(HAVE_VASPRINTF)|g' \ + -e 's|@''HAVE_VDPRINTF''@|$(HAVE_VDPRINTF)|g' \ + -e 's|@''REPLACE_DPRINTF''@|$(REPLACE_DPRINTF)|g' \ + -e 's|@''REPLACE_FCLOSE''@|$(REPLACE_FCLOSE)|g' \ + -e 's|@''REPLACE_FDOPEN''@|$(REPLACE_FDOPEN)|g' \ + -e 's|@''REPLACE_FFLUSH''@|$(REPLACE_FFLUSH)|g' \ + -e 's|@''REPLACE_FOPEN''@|$(REPLACE_FOPEN)|g' \ + -e 's|@''REPLACE_FPRINTF''@|$(REPLACE_FPRINTF)|g' \ + -e 's|@''REPLACE_FPURGE''@|$(REPLACE_FPURGE)|g' \ + -e 's|@''REPLACE_FREOPEN''@|$(REPLACE_FREOPEN)|g' \ + -e 's|@''REPLACE_FSEEK''@|$(REPLACE_FSEEK)|g' \ + -e 's|@''REPLACE_FSEEKO''@|$(REPLACE_FSEEKO)|g' \ + -e 's|@''REPLACE_FTELL''@|$(REPLACE_FTELL)|g' \ + -e 's|@''REPLACE_FTELLO''@|$(REPLACE_FTELLO)|g' \ + -e 's|@''REPLACE_GETDELIM''@|$(REPLACE_GETDELIM)|g' \ + -e 's|@''REPLACE_GETLINE''@|$(REPLACE_GETLINE)|g' \ + -e 's|@''REPLACE_OBSTACK_PRINTF''@|$(REPLACE_OBSTACK_PRINTF)|g' \ + -e 's|@''REPLACE_PERROR''@|$(REPLACE_PERROR)|g' \ + -e 's|@''REPLACE_POPEN''@|$(REPLACE_POPEN)|g' \ + -e 's|@''REPLACE_PRINTF''@|$(REPLACE_PRINTF)|g' \ + -e 's|@''REPLACE_REMOVE''@|$(REPLACE_REMOVE)|g' \ + -e 's|@''REPLACE_RENAME''@|$(REPLACE_RENAME)|g' \ + -e 's|@''REPLACE_RENAMEAT''@|$(REPLACE_RENAMEAT)|g' \ + -e 's|@''REPLACE_SNPRINTF''@|$(REPLACE_SNPRINTF)|g' \ + -e 's|@''REPLACE_SPRINTF''@|$(REPLACE_SPRINTF)|g' \ + -e 's|@''REPLACE_STDIO_READ_FUNCS''@|$(REPLACE_STDIO_READ_FUNCS)|g' \ + -e 's|@''REPLACE_STDIO_WRITE_FUNCS''@|$(REPLACE_STDIO_WRITE_FUNCS)|g' \ + -e 's|@''REPLACE_TMPFILE''@|$(REPLACE_TMPFILE)|g' \ + -e 's|@''REPLACE_VASPRINTF''@|$(REPLACE_VASPRINTF)|g' \ + -e 's|@''REPLACE_VDPRINTF''@|$(REPLACE_VDPRINTF)|g' \ + -e 's|@''REPLACE_VFPRINTF''@|$(REPLACE_VFPRINTF)|g' \ + -e 's|@''REPLACE_VPRINTF''@|$(REPLACE_VPRINTF)|g' \ + -e 's|@''REPLACE_VSNPRINTF''@|$(REPLACE_VSNPRINTF)|g' \ + -e 's|@''REPLACE_VSPRINTF''@|$(REPLACE_VSPRINTF)|g' \ + -e 's|@''ASM_SYMBOL_PREFIX''@|$(ASM_SYMBOL_PREFIX)|g' \ + -e '/definitions of _GL_FUNCDECL_RPL/r $(CXXDEFS_H)' \ + -e '/definition of _GL_ARG_NONNULL/r $(ARG_NONNULL_H)' \ + -e '/definition of _GL_WARN_ON_USE/r $(WARN_ON_USE_H)'; \ + } > $@-t && \ + mv $@-t $@ +MOSTLYCLEANFILES += lib/stdio.h lib/stdio.h-t + +EXTRA_DIST += lib/stdio.in.h + +## end gnulib module stdio + +## begin gnulib module stdlib + +BUILT_SOURCES += lib/stdlib.h + +# We need the following in order to create <stdlib.h> when the system +# doesn't have one that works with the given compiler. +lib/stdlib.h: lib/stdlib.in.h $(top_builddir)/config.status $(CXXDEFS_H) \ + $(_NORETURN_H) $(ARG_NONNULL_H) $(WARN_ON_USE_H) + $(AM_V_GEN)rm -f $@-t $@ && \ + { echo '/* DO NOT EDIT! GENERATED AUTOMATICALLY! */' && \ + sed -e 's|@''GUARD_PREFIX''@|GL|g' \ + -e 's|@''INCLUDE_NEXT''@|$(INCLUDE_NEXT)|g' \ + -e 's|@''PRAGMA_SYSTEM_HEADER''@|@PRAGMA_SYSTEM_HEADER@|g' \ + -e 's|@''PRAGMA_COLUMNS''@|@PRAGMA_COLUMNS@|g' \ + -e 's|@''NEXT_STDLIB_H''@|$(NEXT_STDLIB_H)|g' \ + -e 's/@''GNULIB__EXIT''@/$(GNULIB__EXIT)/g' \ + -e 's/@''GNULIB_ATOLL''@/$(GNULIB_ATOLL)/g' \ + -e 's/@''GNULIB_CALLOC_POSIX''@/$(GNULIB_CALLOC_POSIX)/g' \ + -e 's/@''GNULIB_CANONICALIZE_FILE_NAME''@/$(GNULIB_CANONICALIZE_FILE_NAME)/g' \ + -e 's/@''GNULIB_GETLOADAVG''@/$(GNULIB_GETLOADAVG)/g' \ + -e 's/@''GNULIB_GETSUBOPT''@/$(GNULIB_GETSUBOPT)/g' \ + -e 's/@''GNULIB_GRANTPT''@/$(GNULIB_GRANTPT)/g' \ + -e 's/@''GNULIB_MALLOC_POSIX''@/$(GNULIB_MALLOC_POSIX)/g' \ + -e 's/@''GNULIB_MBTOWC''@/$(GNULIB_MBTOWC)/g' \ + -e 's/@''GNULIB_MKDTEMP''@/$(GNULIB_MKDTEMP)/g' \ + -e 's/@''GNULIB_MKOSTEMP''@/$(GNULIB_MKOSTEMP)/g' \ + -e 's/@''GNULIB_MKOSTEMPS''@/$(GNULIB_MKOSTEMPS)/g' \ + -e 's/@''GNULIB_MKSTEMP''@/$(GNULIB_MKSTEMP)/g' \ + -e 's/@''GNULIB_MKSTEMPS''@/$(GNULIB_MKSTEMPS)/g' \ + -e 's/@''GNULIB_POSIX_OPENPT''@/$(GNULIB_POSIX_OPENPT)/g' \ + -e 's/@''GNULIB_PTSNAME''@/$(GNULIB_PTSNAME)/g' \ + -e 's/@''GNULIB_PTSNAME_R''@/$(GNULIB_PTSNAME_R)/g' \ + -e 's/@''GNULIB_PUTENV''@/$(GNULIB_PUTENV)/g' \ + -e 's/@''GNULIB_QSORT_R''@/$(GNULIB_QSORT_R)/g' \ + -e 's/@''GNULIB_RANDOM''@/$(GNULIB_RANDOM)/g' \ + -e 's/@''GNULIB_RANDOM_R''@/$(GNULIB_RANDOM_R)/g' \ + -e 's/@''GNULIB_REALLOC_POSIX''@/$(GNULIB_REALLOC_POSIX)/g' \ + -e 's/@''GNULIB_REALLOCARRAY''@/$(GNULIB_REALLOCARRAY)/g' \ + -e 's/@''GNULIB_REALPATH''@/$(GNULIB_REALPATH)/g' \ + -e 's/@''GNULIB_RPMATCH''@/$(GNULIB_RPMATCH)/g' \ + -e 's/@''GNULIB_SECURE_GETENV''@/$(GNULIB_SECURE_GETENV)/g' \ + -e 's/@''GNULIB_SETENV''@/$(GNULIB_SETENV)/g' \ + -e 's/@''GNULIB_STRTOD''@/$(GNULIB_STRTOD)/g' \ + -e 's/@''GNULIB_STRTOLL''@/$(GNULIB_STRTOLL)/g' \ + -e 's/@''GNULIB_STRTOULL''@/$(GNULIB_STRTOULL)/g' \ + -e 's/@''GNULIB_SYSTEM_POSIX''@/$(GNULIB_SYSTEM_POSIX)/g' \ + -e 's/@''GNULIB_UNLOCKPT''@/$(GNULIB_UNLOCKPT)/g' \ + -e 's/@''GNULIB_UNSETENV''@/$(GNULIB_UNSETENV)/g' \ + -e 's/@''GNULIB_WCTOMB''@/$(GNULIB_WCTOMB)/g' \ + < $(top_srcdir)/lib/stdlib.in.h | \ + sed -e 's|@''HAVE__EXIT''@|$(HAVE__EXIT)|g' \ + -e 's|@''HAVE_ATOLL''@|$(HAVE_ATOLL)|g' \ + -e 's|@''HAVE_CANONICALIZE_FILE_NAME''@|$(HAVE_CANONICALIZE_FILE_NAME)|g' \ + -e 's|@''HAVE_DECL_GETLOADAVG''@|$(HAVE_DECL_GETLOADAVG)|g' \ + -e 's|@''HAVE_GETSUBOPT''@|$(HAVE_GETSUBOPT)|g' \ + -e 's|@''HAVE_GRANTPT''@|$(HAVE_GRANTPT)|g' \ + -e 's|@''HAVE_DECL_INITSTATE''@|$(HAVE_DECL_INITSTATE)|g' \ + -e 's|@''HAVE_MKDTEMP''@|$(HAVE_MKDTEMP)|g' \ + -e 's|@''HAVE_MKOSTEMP''@|$(HAVE_MKOSTEMP)|g' \ + -e 's|@''HAVE_MKOSTEMPS''@|$(HAVE_MKOSTEMPS)|g' \ + -e 's|@''HAVE_MKSTEMP''@|$(HAVE_MKSTEMP)|g' \ + -e 's|@''HAVE_MKSTEMPS''@|$(HAVE_MKSTEMPS)|g' \ + -e 's|@''HAVE_POSIX_OPENPT''@|$(HAVE_POSIX_OPENPT)|g' \ + -e 's|@''HAVE_PTSNAME''@|$(HAVE_PTSNAME)|g' \ + -e 's|@''HAVE_PTSNAME_R''@|$(HAVE_PTSNAME_R)|g' \ + -e 's|@''HAVE_QSORT_R''@|$(HAVE_QSORT_R)|g' \ + -e 's|@''HAVE_RANDOM''@|$(HAVE_RANDOM)|g' \ + -e 's|@''HAVE_RANDOM_H''@|$(HAVE_RANDOM_H)|g' \ + -e 's|@''HAVE_RANDOM_R''@|$(HAVE_RANDOM_R)|g' \ + -e 's|@''HAVE_REALLOCARRAY''@|$(HAVE_REALLOCARRAY)|g' \ + -e 's|@''HAVE_REALPATH''@|$(HAVE_REALPATH)|g' \ + -e 's|@''HAVE_RPMATCH''@|$(HAVE_RPMATCH)|g' \ + -e 's|@''HAVE_SECURE_GETENV''@|$(HAVE_SECURE_GETENV)|g' \ + -e 's|@''HAVE_DECL_SETENV''@|$(HAVE_DECL_SETENV)|g' \ + -e 's|@''HAVE_DECL_SETSTATE''@|$(HAVE_DECL_SETSTATE)|g' \ + -e 's|@''HAVE_STRTOD''@|$(HAVE_STRTOD)|g' \ + -e 's|@''HAVE_STRTOLL''@|$(HAVE_STRTOLL)|g' \ + -e 's|@''HAVE_STRTOULL''@|$(HAVE_STRTOULL)|g' \ + -e 's|@''HAVE_STRUCT_RANDOM_DATA''@|$(HAVE_STRUCT_RANDOM_DATA)|g' \ + -e 's|@''HAVE_SYS_LOADAVG_H''@|$(HAVE_SYS_LOADAVG_H)|g' \ + -e 's|@''HAVE_UNLOCKPT''@|$(HAVE_UNLOCKPT)|g' \ + -e 's|@''HAVE_DECL_UNSETENV''@|$(HAVE_DECL_UNSETENV)|g' \ + -e 's|@''REPLACE_CALLOC''@|$(REPLACE_CALLOC)|g' \ + -e 's|@''REPLACE_CANONICALIZE_FILE_NAME''@|$(REPLACE_CANONICALIZE_FILE_NAME)|g' \ + -e 's|@''REPLACE_MALLOC''@|$(REPLACE_MALLOC)|g' \ + -e 's|@''REPLACE_MBTOWC''@|$(REPLACE_MBTOWC)|g' \ + -e 's|@''REPLACE_MKSTEMP''@|$(REPLACE_MKSTEMP)|g' \ + -e 's|@''REPLACE_PTSNAME''@|$(REPLACE_PTSNAME)|g' \ + -e 's|@''REPLACE_PTSNAME_R''@|$(REPLACE_PTSNAME_R)|g' \ + -e 's|@''REPLACE_PUTENV''@|$(REPLACE_PUTENV)|g' \ + -e 's|@''REPLACE_QSORT_R''@|$(REPLACE_QSORT_R)|g' \ + -e 's|@''REPLACE_RANDOM_R''@|$(REPLACE_RANDOM_R)|g' \ + -e 's|@''REPLACE_REALLOC''@|$(REPLACE_REALLOC)|g' \ + -e 's|@''REPLACE_REALPATH''@|$(REPLACE_REALPATH)|g' \ + -e 's|@''REPLACE_SETENV''@|$(REPLACE_SETENV)|g' \ + -e 's|@''REPLACE_STRTOD''@|$(REPLACE_STRTOD)|g' \ + -e 's|@''REPLACE_UNSETENV''@|$(REPLACE_UNSETENV)|g' \ + -e 's|@''REPLACE_WCTOMB''@|$(REPLACE_WCTOMB)|g' \ + -e '/definitions of _GL_FUNCDECL_RPL/r $(CXXDEFS_H)' \ + -e '/definition of _Noreturn/r $(_NORETURN_H)' \ + -e '/definition of _GL_ARG_NONNULL/r $(ARG_NONNULL_H)' \ + -e '/definition of _GL_WARN_ON_USE/r $(WARN_ON_USE_H)'; \ + } > $@-t && \ + mv $@-t $@ +MOSTLYCLEANFILES += lib/stdlib.h lib/stdlib.h-t + +EXTRA_DIST += lib/stdlib.in.h + +## end gnulib module stdlib + +## begin gnulib module streq + + +EXTRA_DIST += lib/streq.h + +## end gnulib module streq + +## begin gnulib module strerror + + +EXTRA_DIST += lib/strerror.c + +EXTRA_lib_libsed_a_SOURCES += lib/strerror.c + +## end gnulib module strerror + +## begin gnulib module strerror-override + + +EXTRA_DIST += lib/strerror-override.c lib/strerror-override.h + +EXTRA_lib_libsed_a_SOURCES += lib/strerror-override.c + +## end gnulib module strerror-override + +## begin gnulib module string + +BUILT_SOURCES += lib/string.h + +# We need the following in order to create <string.h> when the system +# doesn't have one that works with the given compiler. +lib/string.h: lib/string.in.h $(top_builddir)/config.status $(CXXDEFS_H) $(ARG_NONNULL_H) $(WARN_ON_USE_H) + $(AM_V_GEN)rm -f $@-t $@ && \ + { echo '/* DO NOT EDIT! GENERATED AUTOMATICALLY! */' && \ + sed -e 's|@''GUARD_PREFIX''@|GL|g' \ + -e 's|@''INCLUDE_NEXT''@|$(INCLUDE_NEXT)|g' \ + -e 's|@''PRAGMA_SYSTEM_HEADER''@|@PRAGMA_SYSTEM_HEADER@|g' \ + -e 's|@''PRAGMA_COLUMNS''@|@PRAGMA_COLUMNS@|g' \ + -e 's|@''NEXT_STRING_H''@|$(NEXT_STRING_H)|g' \ + -e 's/@''GNULIB_EXPLICIT_BZERO''@/$(GNULIB_EXPLICIT_BZERO)/g' \ + -e 's/@''GNULIB_FFSL''@/$(GNULIB_FFSL)/g' \ + -e 's/@''GNULIB_FFSLL''@/$(GNULIB_FFSLL)/g' \ + -e 's/@''GNULIB_MBSLEN''@/$(GNULIB_MBSLEN)/g' \ + -e 's/@''GNULIB_MBSNLEN''@/$(GNULIB_MBSNLEN)/g' \ + -e 's/@''GNULIB_MBSCHR''@/$(GNULIB_MBSCHR)/g' \ + -e 's/@''GNULIB_MBSRCHR''@/$(GNULIB_MBSRCHR)/g' \ + -e 's/@''GNULIB_MBSSTR''@/$(GNULIB_MBSSTR)/g' \ + -e 's/@''GNULIB_MBSCASECMP''@/$(GNULIB_MBSCASECMP)/g' \ + -e 's/@''GNULIB_MBSNCASECMP''@/$(GNULIB_MBSNCASECMP)/g' \ + -e 's/@''GNULIB_MBSPCASECMP''@/$(GNULIB_MBSPCASECMP)/g' \ + -e 's/@''GNULIB_MBSCASESTR''@/$(GNULIB_MBSCASESTR)/g' \ + -e 's/@''GNULIB_MBSCSPN''@/$(GNULIB_MBSCSPN)/g' \ + -e 's/@''GNULIB_MBSPBRK''@/$(GNULIB_MBSPBRK)/g' \ + -e 's/@''GNULIB_MBSSPN''@/$(GNULIB_MBSSPN)/g' \ + -e 's/@''GNULIB_MBSSEP''@/$(GNULIB_MBSSEP)/g' \ + -e 's/@''GNULIB_MBSTOK_R''@/$(GNULIB_MBSTOK_R)/g' \ + -e 's/@''GNULIB_MEMCHR''@/$(GNULIB_MEMCHR)/g' \ + -e 's/@''GNULIB_MEMMEM''@/$(GNULIB_MEMMEM)/g' \ + -e 's/@''GNULIB_MEMPCPY''@/$(GNULIB_MEMPCPY)/g' \ + -e 's/@''GNULIB_MEMRCHR''@/$(GNULIB_MEMRCHR)/g' \ + -e 's/@''GNULIB_RAWMEMCHR''@/$(GNULIB_RAWMEMCHR)/g' \ + -e 's/@''GNULIB_STPCPY''@/$(GNULIB_STPCPY)/g' \ + -e 's/@''GNULIB_STPNCPY''@/$(GNULIB_STPNCPY)/g' \ + -e 's/@''GNULIB_STRCHRNUL''@/$(GNULIB_STRCHRNUL)/g' \ + -e 's/@''GNULIB_STRDUP''@/$(GNULIB_STRDUP)/g' \ + -e 's/@''GNULIB_STRNCAT''@/$(GNULIB_STRNCAT)/g' \ + -e 's/@''GNULIB_STRNDUP''@/$(GNULIB_STRNDUP)/g' \ + -e 's/@''GNULIB_STRNLEN''@/$(GNULIB_STRNLEN)/g' \ + -e 's/@''GNULIB_STRPBRK''@/$(GNULIB_STRPBRK)/g' \ + -e 's/@''GNULIB_STRSEP''@/$(GNULIB_STRSEP)/g' \ + -e 's/@''GNULIB_STRSTR''@/$(GNULIB_STRSTR)/g' \ + -e 's/@''GNULIB_STRCASESTR''@/$(GNULIB_STRCASESTR)/g' \ + -e 's/@''GNULIB_STRTOK_R''@/$(GNULIB_STRTOK_R)/g' \ + -e 's/@''GNULIB_STRERROR''@/$(GNULIB_STRERROR)/g' \ + -e 's/@''GNULIB_STRERROR_R''@/$(GNULIB_STRERROR_R)/g' \ + -e 's/@''GNULIB_STRSIGNAL''@/$(GNULIB_STRSIGNAL)/g' \ + -e 's/@''GNULIB_STRVERSCMP''@/$(GNULIB_STRVERSCMP)/g' \ + < $(top_srcdir)/lib/string.in.h | \ + sed -e 's|@''HAVE_EXPLICIT_BZERO''@|$(HAVE_EXPLICIT_BZERO)|g' \ + -e 's|@''HAVE_FFSL''@|$(HAVE_FFSL)|g' \ + -e 's|@''HAVE_FFSLL''@|$(HAVE_FFSLL)|g' \ + -e 's|@''HAVE_MBSLEN''@|$(HAVE_MBSLEN)|g' \ + -e 's|@''HAVE_MEMCHR''@|$(HAVE_MEMCHR)|g' \ + -e 's|@''HAVE_DECL_MEMMEM''@|$(HAVE_DECL_MEMMEM)|g' \ + -e 's|@''HAVE_MEMPCPY''@|$(HAVE_MEMPCPY)|g' \ + -e 's|@''HAVE_DECL_MEMRCHR''@|$(HAVE_DECL_MEMRCHR)|g' \ + -e 's|@''HAVE_RAWMEMCHR''@|$(HAVE_RAWMEMCHR)|g' \ + -e 's|@''HAVE_STPCPY''@|$(HAVE_STPCPY)|g' \ + -e 's|@''HAVE_STPNCPY''@|$(HAVE_STPNCPY)|g' \ + -e 's|@''HAVE_STRCHRNUL''@|$(HAVE_STRCHRNUL)|g' \ + -e 's|@''HAVE_DECL_STRDUP''@|$(HAVE_DECL_STRDUP)|g' \ + -e 's|@''HAVE_DECL_STRNDUP''@|$(HAVE_DECL_STRNDUP)|g' \ + -e 's|@''HAVE_DECL_STRNLEN''@|$(HAVE_DECL_STRNLEN)|g' \ + -e 's|@''HAVE_STRPBRK''@|$(HAVE_STRPBRK)|g' \ + -e 's|@''HAVE_STRSEP''@|$(HAVE_STRSEP)|g' \ + -e 's|@''HAVE_STRCASESTR''@|$(HAVE_STRCASESTR)|g' \ + -e 's|@''HAVE_DECL_STRTOK_R''@|$(HAVE_DECL_STRTOK_R)|g' \ + -e 's|@''HAVE_DECL_STRERROR_R''@|$(HAVE_DECL_STRERROR_R)|g' \ + -e 's|@''HAVE_DECL_STRSIGNAL''@|$(HAVE_DECL_STRSIGNAL)|g' \ + -e 's|@''HAVE_STRVERSCMP''@|$(HAVE_STRVERSCMP)|g' \ + -e 's|@''REPLACE_MEMCHR''@|$(REPLACE_MEMCHR)|g' \ + -e 's|@''REPLACE_MEMMEM''@|$(REPLACE_MEMMEM)|g' \ + -e 's|@''REPLACE_STPNCPY''@|$(REPLACE_STPNCPY)|g' \ + -e 's|@''REPLACE_STRCHRNUL''@|$(REPLACE_STRCHRNUL)|g' \ + -e 's|@''REPLACE_STRDUP''@|$(REPLACE_STRDUP)|g' \ + -e 's|@''REPLACE_STRNCAT''@|$(REPLACE_STRNCAT)|g' \ + -e 's|@''REPLACE_STRNDUP''@|$(REPLACE_STRNDUP)|g' \ + -e 's|@''REPLACE_STRNLEN''@|$(REPLACE_STRNLEN)|g' \ + -e 's|@''REPLACE_STRSTR''@|$(REPLACE_STRSTR)|g' \ + -e 's|@''REPLACE_STRCASESTR''@|$(REPLACE_STRCASESTR)|g' \ + -e 's|@''REPLACE_STRTOK_R''@|$(REPLACE_STRTOK_R)|g' \ + -e 's|@''REPLACE_STRERROR''@|$(REPLACE_STRERROR)|g' \ + -e 's|@''REPLACE_STRERROR_R''@|$(REPLACE_STRERROR_R)|g' \ + -e 's|@''REPLACE_STRSIGNAL''@|$(REPLACE_STRSIGNAL)|g' \ + -e 's|@''UNDEFINE_STRTOK_R''@|$(UNDEFINE_STRTOK_R)|g' \ + -e '/definitions of _GL_FUNCDECL_RPL/r $(CXXDEFS_H)' \ + -e '/definition of _GL_ARG_NONNULL/r $(ARG_NONNULL_H)' \ + -e '/definition of _GL_WARN_ON_USE/r $(WARN_ON_USE_H)'; \ + < $(top_srcdir)/lib/string.in.h; \ + } > $@-t && \ + mv $@-t $@ +MOSTLYCLEANFILES += lib/string.h lib/string.h-t + +EXTRA_DIST += lib/string.in.h + +## end gnulib module string + +## begin gnulib module strverscmp + + +EXTRA_DIST += lib/strverscmp.c + +EXTRA_lib_libsed_a_SOURCES += lib/strverscmp.c + +## end gnulib module strverscmp + +## begin gnulib module sys_stat + +BUILT_SOURCES += lib/sys/stat.h + +# We need the following in order to create <sys/stat.h> when the system +# has one that is incomplete. +lib/sys/stat.h: lib/sys_stat.in.h $(top_builddir)/config.status $(CXXDEFS_H) $(ARG_NONNULL_H) $(WARN_ON_USE_H) + $(AM_V_at)$(MKDIR_P) lib/sys + $(AM_V_GEN)rm -f $@-t $@ && \ + { echo '/* DO NOT EDIT! GENERATED AUTOMATICALLY! */'; \ + sed -e 's|@''GUARD_PREFIX''@|GL|g' \ + -e 's|@''INCLUDE_NEXT''@|$(INCLUDE_NEXT)|g' \ + -e 's|@''PRAGMA_SYSTEM_HEADER''@|@PRAGMA_SYSTEM_HEADER@|g' \ + -e 's|@''PRAGMA_COLUMNS''@|@PRAGMA_COLUMNS@|g' \ + -e 's|@''NEXT_SYS_STAT_H''@|$(NEXT_SYS_STAT_H)|g' \ + -e 's|@''WINDOWS_64_BIT_ST_SIZE''@|$(WINDOWS_64_BIT_ST_SIZE)|g' \ + -e 's|@''WINDOWS_STAT_TIMESPEC''@|$(WINDOWS_STAT_TIMESPEC)|g' \ + -e 's/@''GNULIB_FCHMODAT''@/$(GNULIB_FCHMODAT)/g' \ + -e 's/@''GNULIB_FSTAT''@/$(GNULIB_FSTAT)/g' \ + -e 's/@''GNULIB_FSTATAT''@/$(GNULIB_FSTATAT)/g' \ + -e 's/@''GNULIB_FUTIMENS''@/$(GNULIB_FUTIMENS)/g' \ + -e 's/@''GNULIB_LCHMOD''@/$(GNULIB_LCHMOD)/g' \ + -e 's/@''GNULIB_LSTAT''@/$(GNULIB_LSTAT)/g' \ + -e 's/@''GNULIB_MKDIRAT''@/$(GNULIB_MKDIRAT)/g' \ + -e 's/@''GNULIB_MKFIFO''@/$(GNULIB_MKFIFO)/g' \ + -e 's/@''GNULIB_MKFIFOAT''@/$(GNULIB_MKFIFOAT)/g' \ + -e 's/@''GNULIB_MKNOD''@/$(GNULIB_MKNOD)/g' \ + -e 's/@''GNULIB_MKNODAT''@/$(GNULIB_MKNODAT)/g' \ + -e 's/@''GNULIB_STAT''@/$(GNULIB_STAT)/g' \ + -e 's/@''GNULIB_UTIMENSAT''@/$(GNULIB_UTIMENSAT)/g' \ + -e 's/@''GNULIB_OVERRIDES_STRUCT_STAT''@/$(GNULIB_OVERRIDES_STRUCT_STAT)/g' \ + -e 's|@''HAVE_FCHMODAT''@|$(HAVE_FCHMODAT)|g' \ + -e 's|@''HAVE_FSTATAT''@|$(HAVE_FSTATAT)|g' \ + -e 's|@''HAVE_FUTIMENS''@|$(HAVE_FUTIMENS)|g' \ + -e 's|@''HAVE_LCHMOD''@|$(HAVE_LCHMOD)|g' \ + -e 's|@''HAVE_LSTAT''@|$(HAVE_LSTAT)|g' \ + -e 's|@''HAVE_MKDIRAT''@|$(HAVE_MKDIRAT)|g' \ + -e 's|@''HAVE_MKFIFO''@|$(HAVE_MKFIFO)|g' \ + -e 's|@''HAVE_MKFIFOAT''@|$(HAVE_MKFIFOAT)|g' \ + -e 's|@''HAVE_MKNOD''@|$(HAVE_MKNOD)|g' \ + -e 's|@''HAVE_MKNODAT''@|$(HAVE_MKNODAT)|g' \ + -e 's|@''HAVE_UTIMENSAT''@|$(HAVE_UTIMENSAT)|g' \ + -e 's|@''REPLACE_FSTAT''@|$(REPLACE_FSTAT)|g' \ + -e 's|@''REPLACE_FSTATAT''@|$(REPLACE_FSTATAT)|g' \ + -e 's|@''REPLACE_FUTIMENS''@|$(REPLACE_FUTIMENS)|g' \ + -e 's|@''REPLACE_LSTAT''@|$(REPLACE_LSTAT)|g' \ + -e 's|@''REPLACE_MKDIR''@|$(REPLACE_MKDIR)|g' \ + -e 's|@''REPLACE_MKFIFO''@|$(REPLACE_MKFIFO)|g' \ + -e 's|@''REPLACE_MKNOD''@|$(REPLACE_MKNOD)|g' \ + -e 's|@''REPLACE_STAT''@|$(REPLACE_STAT)|g' \ + -e 's|@''REPLACE_UTIMENSAT''@|$(REPLACE_UTIMENSAT)|g' \ + -e '/definitions of _GL_FUNCDECL_RPL/r $(CXXDEFS_H)' \ + -e '/definition of _GL_ARG_NONNULL/r $(ARG_NONNULL_H)' \ + -e '/definition of _GL_WARN_ON_USE/r $(WARN_ON_USE_H)' \ + < $(top_srcdir)/lib/sys_stat.in.h; \ + } > $@-t && \ + mv $@-t $@ +MOSTLYCLEANFILES += lib/sys/stat.h lib/sys/stat.h-t +MOSTLYCLEANDIRS += lib/sys + +EXTRA_DIST += lib/sys_stat.in.h + +## end gnulib module sys_stat + +## begin gnulib module sys_time + +BUILT_SOURCES += lib/sys/time.h + +# We need the following in order to create <sys/time.h> when the system +# doesn't have one that works with the given compiler. +lib/sys/time.h: lib/sys_time.in.h $(top_builddir)/config.status $(CXXDEFS_H) $(ARG_NONNULL_H) $(WARN_ON_USE_H) + $(AM_V_at)$(MKDIR_P) lib/sys + $(AM_V_GEN)rm -f $@-t $@ && \ + { echo '/* DO NOT EDIT! GENERATED AUTOMATICALLY! */'; \ + sed -e 's|@''GUARD_PREFIX''@|GL|g' \ + -e 's/@''HAVE_SYS_TIME_H''@/$(HAVE_SYS_TIME_H)/g' \ + -e 's|@''INCLUDE_NEXT''@|$(INCLUDE_NEXT)|g' \ + -e 's|@''PRAGMA_SYSTEM_HEADER''@|@PRAGMA_SYSTEM_HEADER@|g' \ + -e 's|@''PRAGMA_COLUMNS''@|@PRAGMA_COLUMNS@|g' \ + -e 's|@''NEXT_SYS_TIME_H''@|$(NEXT_SYS_TIME_H)|g' \ + -e 's/@''GNULIB_GETTIMEOFDAY''@/$(GNULIB_GETTIMEOFDAY)/g' \ + -e 's|@''HAVE_WINSOCK2_H''@|$(HAVE_WINSOCK2_H)|g' \ + -e 's/@''HAVE_GETTIMEOFDAY''@/$(HAVE_GETTIMEOFDAY)/g' \ + -e 's/@''HAVE_STRUCT_TIMEVAL''@/$(HAVE_STRUCT_TIMEVAL)/g' \ + -e 's/@''REPLACE_GETTIMEOFDAY''@/$(REPLACE_GETTIMEOFDAY)/g' \ + -e 's/@''REPLACE_STRUCT_TIMEVAL''@/$(REPLACE_STRUCT_TIMEVAL)/g' \ + -e '/definitions of _GL_FUNCDECL_RPL/r $(CXXDEFS_H)' \ + -e '/definition of _GL_ARG_NONNULL/r $(ARG_NONNULL_H)' \ + -e '/definition of _GL_WARN_ON_USE/r $(WARN_ON_USE_H)' \ + < $(top_srcdir)/lib/sys_time.in.h; \ + } > $@-t && \ + mv $@-t $@ +MOSTLYCLEANFILES += lib/sys/time.h lib/sys/time.h-t + +EXTRA_DIST += lib/sys_time.in.h + +## end gnulib module sys_time + +## begin gnulib module sys_types + +BUILT_SOURCES += lib/sys/types.h + +# We need the following in order to create <sys/types.h> when the system +# doesn't have one that works with the given compiler. +lib/sys/types.h: lib/sys_types.in.h $(top_builddir)/config.status + $(AM_V_at)$(MKDIR_P) lib/sys + $(AM_V_GEN)rm -f $@-t $@ && \ + { echo '/* DO NOT EDIT! GENERATED AUTOMATICALLY! */'; \ + sed -e 's|@''GUARD_PREFIX''@|GL|g' \ + -e 's|@''INCLUDE_NEXT''@|$(INCLUDE_NEXT)|g' \ + -e 's|@''PRAGMA_SYSTEM_HEADER''@|@PRAGMA_SYSTEM_HEADER@|g' \ + -e 's|@''PRAGMA_COLUMNS''@|@PRAGMA_COLUMNS@|g' \ + -e 's|@''NEXT_SYS_TYPES_H''@|$(NEXT_SYS_TYPES_H)|g' \ + -e 's|@''WINDOWS_64_BIT_OFF_T''@|$(WINDOWS_64_BIT_OFF_T)|g' \ + -e 's|@''WINDOWS_STAT_INODES''@|$(WINDOWS_STAT_INODES)|g' \ + < $(top_srcdir)/lib/sys_types.in.h; \ + } > $@-t && \ + mv $@-t $@ +MOSTLYCLEANFILES += lib/sys/types.h lib/sys/types.h-t + +EXTRA_DIST += lib/sys_types.in.h + +## end gnulib module sys_types + +## begin gnulib module tempname + +lib_libsed_a_SOURCES += lib/tempname.c + +EXTRA_DIST += lib/tempname.h + +## end gnulib module tempname + +## begin gnulib module threadlib + +lib_libsed_a_SOURCES += lib/glthread/threadlib.c + +EXTRA_DIST += $(top_srcdir)/build-aux/config.rpath + +## end gnulib module threadlib + +## begin gnulib module time + +BUILT_SOURCES += lib/time.h + +# We need the following in order to create <time.h> when the system +# doesn't have one that works with the given compiler. +lib/time.h: lib/time.in.h $(top_builddir)/config.status $(CXXDEFS_H) $(ARG_NONNULL_H) $(WARN_ON_USE_H) + $(AM_V_GEN)rm -f $@-t $@ && \ + { echo '/* DO NOT EDIT! GENERATED AUTOMATICALLY! */' && \ + sed -e 's|@''GUARD_PREFIX''@|GL|g' \ + -e 's|@''INCLUDE_NEXT''@|$(INCLUDE_NEXT)|g' \ + -e 's|@''PRAGMA_SYSTEM_HEADER''@|@PRAGMA_SYSTEM_HEADER@|g' \ + -e 's|@''PRAGMA_COLUMNS''@|@PRAGMA_COLUMNS@|g' \ + -e 's|@''NEXT_TIME_H''@|$(NEXT_TIME_H)|g' \ + -e 's/@''GNULIB_CTIME''@/$(GNULIB_CTIME)/g' \ + -e 's/@''GNULIB_LOCALTIME''@/$(GNULIB_LOCALTIME)/g' \ + -e 's/@''GNULIB_MKTIME''@/$(GNULIB_MKTIME)/g' \ + -e 's/@''GNULIB_NANOSLEEP''@/$(GNULIB_NANOSLEEP)/g' \ + -e 's/@''GNULIB_STRFTIME''@/$(GNULIB_STRFTIME)/g' \ + -e 's/@''GNULIB_STRPTIME''@/$(GNULIB_STRPTIME)/g' \ + -e 's/@''GNULIB_TIMEGM''@/$(GNULIB_TIMEGM)/g' \ + -e 's/@''GNULIB_TIME_R''@/$(GNULIB_TIME_R)/g' \ + -e 's/@''GNULIB_TIME_RZ''@/$(GNULIB_TIME_RZ)/g' \ + -e 's/@''GNULIB_TZSET''@/$(GNULIB_TZSET)/g' \ + -e 's|@''HAVE_DECL_LOCALTIME_R''@|$(HAVE_DECL_LOCALTIME_R)|g' \ + -e 's|@''HAVE_NANOSLEEP''@|$(HAVE_NANOSLEEP)|g' \ + -e 's|@''HAVE_STRPTIME''@|$(HAVE_STRPTIME)|g' \ + -e 's|@''HAVE_TIMEGM''@|$(HAVE_TIMEGM)|g' \ + -e 's|@''HAVE_TIMEZONE_T''@|$(HAVE_TIMEZONE_T)|g' \ + -e 's|@''HAVE_TZSET''@|$(HAVE_TZSET)|g' \ + -e 's|@''REPLACE_CTIME''@|$(REPLACE_CTIME)|g' \ + -e 's|@''REPLACE_GMTIME''@|$(REPLACE_GMTIME)|g' \ + -e 's|@''REPLACE_LOCALTIME''@|$(REPLACE_LOCALTIME)|g' \ + -e 's|@''REPLACE_LOCALTIME_R''@|$(REPLACE_LOCALTIME_R)|g' \ + -e 's|@''REPLACE_MKTIME''@|$(REPLACE_MKTIME)|g' \ + -e 's|@''REPLACE_NANOSLEEP''@|$(REPLACE_NANOSLEEP)|g' \ + -e 's|@''REPLACE_STRFTIME''@|$(REPLACE_STRFTIME)|g' \ + -e 's|@''REPLACE_TIMEGM''@|$(REPLACE_TIMEGM)|g' \ + -e 's|@''REPLACE_TZSET''@|$(REPLACE_TZSET)|g' \ + -e 's|@''PTHREAD_H_DEFINES_STRUCT_TIMESPEC''@|$(PTHREAD_H_DEFINES_STRUCT_TIMESPEC)|g' \ + -e 's|@''SYS_TIME_H_DEFINES_STRUCT_TIMESPEC''@|$(SYS_TIME_H_DEFINES_STRUCT_TIMESPEC)|g' \ + -e 's|@''TIME_H_DEFINES_STRUCT_TIMESPEC''@|$(TIME_H_DEFINES_STRUCT_TIMESPEC)|g' \ + -e 's|@''UNISTD_H_DEFINES_STRUCT_TIMESPEC''@|$(UNISTD_H_DEFINES_STRUCT_TIMESPEC)|g' \ + -e '/definitions of _GL_FUNCDECL_RPL/r $(CXXDEFS_H)' \ + -e '/definition of _GL_ARG_NONNULL/r $(ARG_NONNULL_H)' \ + -e '/definition of _GL_WARN_ON_USE/r $(WARN_ON_USE_H)' \ + < $(top_srcdir)/lib/time.in.h; \ + } > $@-t && \ + mv $@-t $@ +MOSTLYCLEANFILES += lib/time.h lib/time.h-t + +EXTRA_DIST += lib/time.in.h + +## end gnulib module time + +## begin gnulib module unistd + +BUILT_SOURCES += lib/unistd.h +lib_libsed_a_SOURCES += lib/unistd.c + +# We need the following in order to create an empty placeholder for +# <unistd.h> when the system doesn't have one. +lib/unistd.h: lib/unistd.in.h $(top_builddir)/config.status $(CXXDEFS_H) $(ARG_NONNULL_H) $(WARN_ON_USE_H) + $(AM_V_GEN)rm -f $@-t $@ && \ + { echo '/* DO NOT EDIT! GENERATED AUTOMATICALLY! */'; \ + sed -e 's|@''GUARD_PREFIX''@|GL|g' \ + -e 's|@''HAVE_UNISTD_H''@|$(HAVE_UNISTD_H)|g' \ + -e 's|@''INCLUDE_NEXT''@|$(INCLUDE_NEXT)|g' \ + -e 's|@''PRAGMA_SYSTEM_HEADER''@|@PRAGMA_SYSTEM_HEADER@|g' \ + -e 's|@''PRAGMA_COLUMNS''@|@PRAGMA_COLUMNS@|g' \ + -e 's|@''NEXT_UNISTD_H''@|$(NEXT_UNISTD_H)|g' \ + -e 's|@''WINDOWS_64_BIT_OFF_T''@|$(WINDOWS_64_BIT_OFF_T)|g' \ + -e 's/@''GNULIB_CHDIR''@/$(GNULIB_CHDIR)/g' \ + -e 's/@''GNULIB_CHOWN''@/$(GNULIB_CHOWN)/g' \ + -e 's/@''GNULIB_CLOSE''@/$(GNULIB_CLOSE)/g' \ + -e 's/@''GNULIB_DUP''@/$(GNULIB_DUP)/g' \ + -e 's/@''GNULIB_DUP2''@/$(GNULIB_DUP2)/g' \ + -e 's/@''GNULIB_DUP3''@/$(GNULIB_DUP3)/g' \ + -e 's/@''GNULIB_ENVIRON''@/$(GNULIB_ENVIRON)/g' \ + -e 's/@''GNULIB_EUIDACCESS''@/$(GNULIB_EUIDACCESS)/g' \ + -e 's/@''GNULIB_FACCESSAT''@/$(GNULIB_FACCESSAT)/g' \ + -e 's/@''GNULIB_FCHDIR''@/$(GNULIB_FCHDIR)/g' \ + -e 's/@''GNULIB_FCHOWNAT''@/$(GNULIB_FCHOWNAT)/g' \ + -e 's/@''GNULIB_FDATASYNC''@/$(GNULIB_FDATASYNC)/g' \ + -e 's/@''GNULIB_FSYNC''@/$(GNULIB_FSYNC)/g' \ + -e 's/@''GNULIB_FTRUNCATE''@/$(GNULIB_FTRUNCATE)/g' \ + -e 's/@''GNULIB_GETCWD''@/$(GNULIB_GETCWD)/g' \ + -e 's/@''GNULIB_GETDOMAINNAME''@/$(GNULIB_GETDOMAINNAME)/g' \ + -e 's/@''GNULIB_GETDTABLESIZE''@/$(GNULIB_GETDTABLESIZE)/g' \ + -e 's/@''GNULIB_GETGROUPS''@/$(GNULIB_GETGROUPS)/g' \ + -e 's/@''GNULIB_GETHOSTNAME''@/$(GNULIB_GETHOSTNAME)/g' \ + -e 's/@''GNULIB_GETLOGIN''@/$(GNULIB_GETLOGIN)/g' \ + -e 's/@''GNULIB_GETLOGIN_R''@/$(GNULIB_GETLOGIN_R)/g' \ + -e 's/@''GNULIB_GETPAGESIZE''@/$(GNULIB_GETPAGESIZE)/g' \ + -e 's/@''GNULIB_GETPASS''@/$(GNULIB_GETPASS)/g' \ + -e 's/@''GNULIB_GETUSERSHELL''@/$(GNULIB_GETUSERSHELL)/g' \ + -e 's/@''GNULIB_GROUP_MEMBER''@/$(GNULIB_GROUP_MEMBER)/g' \ + -e 's/@''GNULIB_ISATTY''@/$(GNULIB_ISATTY)/g' \ + -e 's/@''GNULIB_LCHOWN''@/$(GNULIB_LCHOWN)/g' \ + -e 's/@''GNULIB_LINK''@/$(GNULIB_LINK)/g' \ + -e 's/@''GNULIB_LINKAT''@/$(GNULIB_LINKAT)/g' \ + -e 's/@''GNULIB_LSEEK''@/$(GNULIB_LSEEK)/g' \ + -e 's/@''GNULIB_PIPE''@/$(GNULIB_PIPE)/g' \ + -e 's/@''GNULIB_PIPE2''@/$(GNULIB_PIPE2)/g' \ + -e 's/@''GNULIB_PREAD''@/$(GNULIB_PREAD)/g' \ + -e 's/@''GNULIB_PWRITE''@/$(GNULIB_PWRITE)/g' \ + -e 's/@''GNULIB_READ''@/$(GNULIB_READ)/g' \ + -e 's/@''GNULIB_READLINK''@/$(GNULIB_READLINK)/g' \ + -e 's/@''GNULIB_READLINKAT''@/$(GNULIB_READLINKAT)/g' \ + -e 's/@''GNULIB_RMDIR''@/$(GNULIB_RMDIR)/g' \ + -e 's/@''GNULIB_SETHOSTNAME''@/$(GNULIB_SETHOSTNAME)/g' \ + -e 's/@''GNULIB_SLEEP''@/$(GNULIB_SLEEP)/g' \ + -e 's/@''GNULIB_SYMLINK''@/$(GNULIB_SYMLINK)/g' \ + -e 's/@''GNULIB_SYMLINKAT''@/$(GNULIB_SYMLINKAT)/g' \ + -e 's/@''GNULIB_TRUNCATE''@/$(GNULIB_TRUNCATE)/g' \ + -e 's/@''GNULIB_TTYNAME_R''@/$(GNULIB_TTYNAME_R)/g' \ + -e 's/@''GNULIB_UNISTD_H_GETOPT''@/0$(GNULIB_GL_UNISTD_H_GETOPT)/g' \ + -e 's/@''GNULIB_UNISTD_H_NONBLOCKING''@/$(GNULIB_UNISTD_H_NONBLOCKING)/g' \ + -e 's/@''GNULIB_UNISTD_H_SIGPIPE''@/$(GNULIB_UNISTD_H_SIGPIPE)/g' \ + -e 's/@''GNULIB_UNLINK''@/$(GNULIB_UNLINK)/g' \ + -e 's/@''GNULIB_UNLINKAT''@/$(GNULIB_UNLINKAT)/g' \ + -e 's/@''GNULIB_USLEEP''@/$(GNULIB_USLEEP)/g' \ + -e 's/@''GNULIB_WRITE''@/$(GNULIB_WRITE)/g' \ + < $(top_srcdir)/lib/unistd.in.h | \ + sed -e 's|@''HAVE_CHOWN''@|$(HAVE_CHOWN)|g' \ + -e 's|@''HAVE_DUP2''@|$(HAVE_DUP2)|g' \ + -e 's|@''HAVE_DUP3''@|$(HAVE_DUP3)|g' \ + -e 's|@''HAVE_EUIDACCESS''@|$(HAVE_EUIDACCESS)|g' \ + -e 's|@''HAVE_FACCESSAT''@|$(HAVE_FACCESSAT)|g' \ + -e 's|@''HAVE_FCHDIR''@|$(HAVE_FCHDIR)|g' \ + -e 's|@''HAVE_FCHOWNAT''@|$(HAVE_FCHOWNAT)|g' \ + -e 's|@''HAVE_FDATASYNC''@|$(HAVE_FDATASYNC)|g' \ + -e 's|@''HAVE_FSYNC''@|$(HAVE_FSYNC)|g' \ + -e 's|@''HAVE_FTRUNCATE''@|$(HAVE_FTRUNCATE)|g' \ + -e 's|@''HAVE_GETDTABLESIZE''@|$(HAVE_GETDTABLESIZE)|g' \ + -e 's|@''HAVE_GETGROUPS''@|$(HAVE_GETGROUPS)|g' \ + -e 's|@''HAVE_GETHOSTNAME''@|$(HAVE_GETHOSTNAME)|g' \ + -e 's|@''HAVE_GETPAGESIZE''@|$(HAVE_GETPAGESIZE)|g' \ + -e 's|@''HAVE_GETPASS''@|$(HAVE_GETPASS)|g' \ + -e 's|@''HAVE_GROUP_MEMBER''@|$(HAVE_GROUP_MEMBER)|g' \ + -e 's|@''HAVE_LCHOWN''@|$(HAVE_LCHOWN)|g' \ + -e 's|@''HAVE_LINK''@|$(HAVE_LINK)|g' \ + -e 's|@''HAVE_LINKAT''@|$(HAVE_LINKAT)|g' \ + -e 's|@''HAVE_PIPE''@|$(HAVE_PIPE)|g' \ + -e 's|@''HAVE_PIPE2''@|$(HAVE_PIPE2)|g' \ + -e 's|@''HAVE_PREAD''@|$(HAVE_PREAD)|g' \ + -e 's|@''HAVE_PWRITE''@|$(HAVE_PWRITE)|g' \ + -e 's|@''HAVE_READLINK''@|$(HAVE_READLINK)|g' \ + -e 's|@''HAVE_READLINKAT''@|$(HAVE_READLINKAT)|g' \ + -e 's|@''HAVE_SETHOSTNAME''@|$(HAVE_SETHOSTNAME)|g' \ + -e 's|@''HAVE_SLEEP''@|$(HAVE_SLEEP)|g' \ + -e 's|@''HAVE_SYMLINK''@|$(HAVE_SYMLINK)|g' \ + -e 's|@''HAVE_SYMLINKAT''@|$(HAVE_SYMLINKAT)|g' \ + -e 's|@''HAVE_UNLINKAT''@|$(HAVE_UNLINKAT)|g' \ + -e 's|@''HAVE_USLEEP''@|$(HAVE_USLEEP)|g' \ + -e 's|@''HAVE_DECL_ENVIRON''@|$(HAVE_DECL_ENVIRON)|g' \ + -e 's|@''HAVE_DECL_FCHDIR''@|$(HAVE_DECL_FCHDIR)|g' \ + -e 's|@''HAVE_DECL_FDATASYNC''@|$(HAVE_DECL_FDATASYNC)|g' \ + -e 's|@''HAVE_DECL_GETDOMAINNAME''@|$(HAVE_DECL_GETDOMAINNAME)|g' \ + -e 's|@''HAVE_DECL_GETLOGIN''@|$(HAVE_DECL_GETLOGIN)|g' \ + -e 's|@''HAVE_DECL_GETLOGIN_R''@|$(HAVE_DECL_GETLOGIN_R)|g' \ + -e 's|@''HAVE_DECL_GETPAGESIZE''@|$(HAVE_DECL_GETPAGESIZE)|g' \ + -e 's|@''HAVE_DECL_GETUSERSHELL''@|$(HAVE_DECL_GETUSERSHELL)|g' \ + -e 's|@''HAVE_DECL_SETHOSTNAME''@|$(HAVE_DECL_SETHOSTNAME)|g' \ + -e 's|@''HAVE_DECL_TRUNCATE''@|$(HAVE_DECL_TRUNCATE)|g' \ + -e 's|@''HAVE_DECL_TTYNAME_R''@|$(HAVE_DECL_TTYNAME_R)|g' \ + -e 's|@''HAVE_OS_H''@|$(HAVE_OS_H)|g' \ + -e 's|@''HAVE_SYS_PARAM_H''@|$(HAVE_SYS_PARAM_H)|g' \ + | \ + sed -e 's|@''REPLACE_CHOWN''@|$(REPLACE_CHOWN)|g' \ + -e 's|@''REPLACE_CLOSE''@|$(REPLACE_CLOSE)|g' \ + -e 's|@''REPLACE_DUP''@|$(REPLACE_DUP)|g' \ + -e 's|@''REPLACE_DUP2''@|$(REPLACE_DUP2)|g' \ + -e 's|@''REPLACE_FACCESSAT''@|$(REPLACE_FACCESSAT)|g' \ + -e 's|@''REPLACE_FCHOWNAT''@|$(REPLACE_FCHOWNAT)|g' \ + -e 's|@''REPLACE_FTRUNCATE''@|$(REPLACE_FTRUNCATE)|g' \ + -e 's|@''REPLACE_GETCWD''@|$(REPLACE_GETCWD)|g' \ + -e 's|@''REPLACE_GETDOMAINNAME''@|$(REPLACE_GETDOMAINNAME)|g' \ + -e 's|@''REPLACE_GETDTABLESIZE''@|$(REPLACE_GETDTABLESIZE)|g' \ + -e 's|@''REPLACE_GETLOGIN_R''@|$(REPLACE_GETLOGIN_R)|g' \ + -e 's|@''REPLACE_GETGROUPS''@|$(REPLACE_GETGROUPS)|g' \ + -e 's|@''REPLACE_GETPAGESIZE''@|$(REPLACE_GETPAGESIZE)|g' \ + -e 's|@''REPLACE_GETPASS''@|$(REPLACE_GETPASS)|g' \ + -e 's|@''REPLACE_ISATTY''@|$(REPLACE_ISATTY)|g' \ + -e 's|@''REPLACE_LCHOWN''@|$(REPLACE_LCHOWN)|g' \ + -e 's|@''REPLACE_LINK''@|$(REPLACE_LINK)|g' \ + -e 's|@''REPLACE_LINKAT''@|$(REPLACE_LINKAT)|g' \ + -e 's|@''REPLACE_LSEEK''@|$(REPLACE_LSEEK)|g' \ + -e 's|@''REPLACE_PREAD''@|$(REPLACE_PREAD)|g' \ + -e 's|@''REPLACE_PWRITE''@|$(REPLACE_PWRITE)|g' \ + -e 's|@''REPLACE_READ''@|$(REPLACE_READ)|g' \ + -e 's|@''REPLACE_READLINK''@|$(REPLACE_READLINK)|g' \ + -e 's|@''REPLACE_READLINKAT''@|$(REPLACE_READLINKAT)|g' \ + -e 's|@''REPLACE_RMDIR''@|$(REPLACE_RMDIR)|g' \ + -e 's|@''REPLACE_SLEEP''@|$(REPLACE_SLEEP)|g' \ + -e 's|@''REPLACE_SYMLINK''@|$(REPLACE_SYMLINK)|g' \ + -e 's|@''REPLACE_SYMLINKAT''@|$(REPLACE_SYMLINKAT)|g' \ + -e 's|@''REPLACE_TRUNCATE''@|$(REPLACE_TRUNCATE)|g' \ + -e 's|@''REPLACE_TTYNAME_R''@|$(REPLACE_TTYNAME_R)|g' \ + -e 's|@''REPLACE_UNLINK''@|$(REPLACE_UNLINK)|g' \ + -e 's|@''REPLACE_UNLINKAT''@|$(REPLACE_UNLINKAT)|g' \ + -e 's|@''REPLACE_USLEEP''@|$(REPLACE_USLEEP)|g' \ + -e 's|@''REPLACE_WRITE''@|$(REPLACE_WRITE)|g' \ + -e 's|@''UNISTD_H_HAVE_WINSOCK2_H''@|$(UNISTD_H_HAVE_WINSOCK2_H)|g' \ + -e 's|@''UNISTD_H_HAVE_WINSOCK2_H_AND_USE_SOCKETS''@|$(UNISTD_H_HAVE_WINSOCK2_H_AND_USE_SOCKETS)|g' \ + -e '/definitions of _GL_FUNCDECL_RPL/r $(CXXDEFS_H)' \ + -e '/definition of _GL_ARG_NONNULL/r $(ARG_NONNULL_H)' \ + -e '/definition of _GL_WARN_ON_USE/r $(WARN_ON_USE_H)'; \ + } > $@-t && \ + mv $@-t $@ +MOSTLYCLEANFILES += lib/unistd.h lib/unistd.h-t + +EXTRA_DIST += lib/unistd.in.h + +## end gnulib module unistd + +## begin gnulib module unlocked-io + + +EXTRA_DIST += lib/unlocked-io.h + +## end gnulib module unlocked-io + +## begin gnulib module update-copyright + + +EXTRA_DIST += $(top_srcdir)/build-aux/update-copyright + +## end gnulib module update-copyright + +## begin gnulib module useless-if-before-free + + +EXTRA_DIST += $(top_srcdir)/build-aux/useless-if-before-free + +## end gnulib module useless-if-before-free + +## begin gnulib module vc-list-files + + +EXTRA_DIST += $(top_srcdir)/build-aux/vc-list-files + +## end gnulib module vc-list-files + +## begin gnulib module verify + + +EXTRA_DIST += lib/verify.h + +## end gnulib module verify + +## begin gnulib module version-etc + +lib_libsed_a_SOURCES += lib/version-etc.h lib/version-etc.c + +## end gnulib module version-etc + +## begin gnulib module version-etc-fsf + +lib_libsed_a_SOURCES += lib/version-etc-fsf.c + +## end gnulib module version-etc-fsf + +## begin gnulib module wchar + +BUILT_SOURCES += lib/wchar.h + +# We need the following in order to create <wchar.h> when the system +# version does not work standalone. +lib/wchar.h: lib/wchar.in.h $(top_builddir)/config.status $(CXXDEFS_H) $(ARG_NONNULL_H) $(WARN_ON_USE_H) + $(AM_V_GEN)rm -f $@-t $@ && \ + { echo '/* DO NOT EDIT! GENERATED AUTOMATICALLY! */'; \ + sed -e 's|@''GUARD_PREFIX''@|GL|g' \ + -e 's|@''INCLUDE_NEXT''@|$(INCLUDE_NEXT)|g' \ + -e 's|@''PRAGMA_SYSTEM_HEADER''@|@PRAGMA_SYSTEM_HEADER@|g' \ + -e 's|@''PRAGMA_COLUMNS''@|@PRAGMA_COLUMNS@|g' \ + -e 's|@''HAVE_FEATURES_H''@|$(HAVE_FEATURES_H)|g' \ + -e 's|@''NEXT_WCHAR_H''@|$(NEXT_WCHAR_H)|g' \ + -e 's|@''HAVE_WCHAR_H''@|$(HAVE_WCHAR_H)|g' \ + -e 's/@''HAVE_CRTDEFS_H''@/$(HAVE_CRTDEFS_H)/g' \ + -e 's/@''GNULIB_OVERRIDES_WINT_T''@/$(GNULIB_OVERRIDES_WINT_T)/g' \ + -e 's/@''GNULIB_BTOWC''@/$(GNULIB_BTOWC)/g' \ + -e 's/@''GNULIB_WCTOB''@/$(GNULIB_WCTOB)/g' \ + -e 's/@''GNULIB_MBSINIT''@/$(GNULIB_MBSINIT)/g' \ + -e 's/@''GNULIB_MBRTOWC''@/$(GNULIB_MBRTOWC)/g' \ + -e 's/@''GNULIB_MBRLEN''@/$(GNULIB_MBRLEN)/g' \ + -e 's/@''GNULIB_MBSRTOWCS''@/$(GNULIB_MBSRTOWCS)/g' \ + -e 's/@''GNULIB_MBSNRTOWCS''@/$(GNULIB_MBSNRTOWCS)/g' \ + -e 's/@''GNULIB_WCRTOMB''@/$(GNULIB_WCRTOMB)/g' \ + -e 's/@''GNULIB_WCSRTOMBS''@/$(GNULIB_WCSRTOMBS)/g' \ + -e 's/@''GNULIB_WCSNRTOMBS''@/$(GNULIB_WCSNRTOMBS)/g' \ + -e 's/@''GNULIB_WCWIDTH''@/$(GNULIB_WCWIDTH)/g' \ + -e 's/@''GNULIB_WMEMCHR''@/$(GNULIB_WMEMCHR)/g' \ + -e 's/@''GNULIB_WMEMCMP''@/$(GNULIB_WMEMCMP)/g' \ + -e 's/@''GNULIB_WMEMCPY''@/$(GNULIB_WMEMCPY)/g' \ + -e 's/@''GNULIB_WMEMMOVE''@/$(GNULIB_WMEMMOVE)/g' \ + -e 's/@''GNULIB_WMEMSET''@/$(GNULIB_WMEMSET)/g' \ + -e 's/@''GNULIB_WCSLEN''@/$(GNULIB_WCSLEN)/g' \ + -e 's/@''GNULIB_WCSNLEN''@/$(GNULIB_WCSNLEN)/g' \ + -e 's/@''GNULIB_WCSCPY''@/$(GNULIB_WCSCPY)/g' \ + -e 's/@''GNULIB_WCPCPY''@/$(GNULIB_WCPCPY)/g' \ + -e 's/@''GNULIB_WCSNCPY''@/$(GNULIB_WCSNCPY)/g' \ + -e 's/@''GNULIB_WCPNCPY''@/$(GNULIB_WCPNCPY)/g' \ + -e 's/@''GNULIB_WCSCAT''@/$(GNULIB_WCSCAT)/g' \ + -e 's/@''GNULIB_WCSNCAT''@/$(GNULIB_WCSNCAT)/g' \ + -e 's/@''GNULIB_WCSCMP''@/$(GNULIB_WCSCMP)/g' \ + -e 's/@''GNULIB_WCSNCMP''@/$(GNULIB_WCSNCMP)/g' \ + -e 's/@''GNULIB_WCSCASECMP''@/$(GNULIB_WCSCASECMP)/g' \ + -e 's/@''GNULIB_WCSNCASECMP''@/$(GNULIB_WCSNCASECMP)/g' \ + -e 's/@''GNULIB_WCSCOLL''@/$(GNULIB_WCSCOLL)/g' \ + -e 's/@''GNULIB_WCSXFRM''@/$(GNULIB_WCSXFRM)/g' \ + -e 's/@''GNULIB_WCSDUP''@/$(GNULIB_WCSDUP)/g' \ + -e 's/@''GNULIB_WCSCHR''@/$(GNULIB_WCSCHR)/g' \ + -e 's/@''GNULIB_WCSRCHR''@/$(GNULIB_WCSRCHR)/g' \ + -e 's/@''GNULIB_WCSCSPN''@/$(GNULIB_WCSCSPN)/g' \ + -e 's/@''GNULIB_WCSSPN''@/$(GNULIB_WCSSPN)/g' \ + -e 's/@''GNULIB_WCSPBRK''@/$(GNULIB_WCSPBRK)/g' \ + -e 's/@''GNULIB_WCSSTR''@/$(GNULIB_WCSSTR)/g' \ + -e 's/@''GNULIB_WCSTOK''@/$(GNULIB_WCSTOK)/g' \ + -e 's/@''GNULIB_WCSWIDTH''@/$(GNULIB_WCSWIDTH)/g' \ + -e 's/@''GNULIB_WCSFTIME''@/$(GNULIB_WCSFTIME)/g' \ + < $(top_srcdir)/lib/wchar.in.h | \ + sed -e 's|@''HAVE_WINT_T''@|$(HAVE_WINT_T)|g' \ + -e 's|@''HAVE_BTOWC''@|$(HAVE_BTOWC)|g' \ + -e 's|@''HAVE_MBSINIT''@|$(HAVE_MBSINIT)|g' \ + -e 's|@''HAVE_MBRTOWC''@|$(HAVE_MBRTOWC)|g' \ + -e 's|@''HAVE_MBRLEN''@|$(HAVE_MBRLEN)|g' \ + -e 's|@''HAVE_MBSRTOWCS''@|$(HAVE_MBSRTOWCS)|g' \ + -e 's|@''HAVE_MBSNRTOWCS''@|$(HAVE_MBSNRTOWCS)|g' \ + -e 's|@''HAVE_WCRTOMB''@|$(HAVE_WCRTOMB)|g' \ + -e 's|@''HAVE_WCSRTOMBS''@|$(HAVE_WCSRTOMBS)|g' \ + -e 's|@''HAVE_WCSNRTOMBS''@|$(HAVE_WCSNRTOMBS)|g' \ + -e 's|@''HAVE_WMEMCHR''@|$(HAVE_WMEMCHR)|g' \ + -e 's|@''HAVE_WMEMCMP''@|$(HAVE_WMEMCMP)|g' \ + -e 's|@''HAVE_WMEMCPY''@|$(HAVE_WMEMCPY)|g' \ + -e 's|@''HAVE_WMEMMOVE''@|$(HAVE_WMEMMOVE)|g' \ + -e 's|@''HAVE_WMEMSET''@|$(HAVE_WMEMSET)|g' \ + -e 's|@''HAVE_WCSLEN''@|$(HAVE_WCSLEN)|g' \ + -e 's|@''HAVE_WCSNLEN''@|$(HAVE_WCSNLEN)|g' \ + -e 's|@''HAVE_WCSCPY''@|$(HAVE_WCSCPY)|g' \ + -e 's|@''HAVE_WCPCPY''@|$(HAVE_WCPCPY)|g' \ + -e 's|@''HAVE_WCSNCPY''@|$(HAVE_WCSNCPY)|g' \ + -e 's|@''HAVE_WCPNCPY''@|$(HAVE_WCPNCPY)|g' \ + -e 's|@''HAVE_WCSCAT''@|$(HAVE_WCSCAT)|g' \ + -e 's|@''HAVE_WCSNCAT''@|$(HAVE_WCSNCAT)|g' \ + -e 's|@''HAVE_WCSCMP''@|$(HAVE_WCSCMP)|g' \ + -e 's|@''HAVE_WCSNCMP''@|$(HAVE_WCSNCMP)|g' \ + -e 's|@''HAVE_WCSCASECMP''@|$(HAVE_WCSCASECMP)|g' \ + -e 's|@''HAVE_WCSNCASECMP''@|$(HAVE_WCSNCASECMP)|g' \ + -e 's|@''HAVE_WCSCOLL''@|$(HAVE_WCSCOLL)|g' \ + -e 's|@''HAVE_WCSXFRM''@|$(HAVE_WCSXFRM)|g' \ + -e 's|@''HAVE_WCSDUP''@|$(HAVE_WCSDUP)|g' \ + -e 's|@''HAVE_WCSCHR''@|$(HAVE_WCSCHR)|g' \ + -e 's|@''HAVE_WCSRCHR''@|$(HAVE_WCSRCHR)|g' \ + -e 's|@''HAVE_WCSCSPN''@|$(HAVE_WCSCSPN)|g' \ + -e 's|@''HAVE_WCSSPN''@|$(HAVE_WCSSPN)|g' \ + -e 's|@''HAVE_WCSPBRK''@|$(HAVE_WCSPBRK)|g' \ + -e 's|@''HAVE_WCSSTR''@|$(HAVE_WCSSTR)|g' \ + -e 's|@''HAVE_WCSTOK''@|$(HAVE_WCSTOK)|g' \ + -e 's|@''HAVE_WCSWIDTH''@|$(HAVE_WCSWIDTH)|g' \ + -e 's|@''HAVE_WCSFTIME''@|$(HAVE_WCSFTIME)|g' \ + -e 's|@''HAVE_DECL_WCTOB''@|$(HAVE_DECL_WCTOB)|g' \ + -e 's|@''HAVE_DECL_WCWIDTH''@|$(HAVE_DECL_WCWIDTH)|g' \ + | \ + sed -e 's|@''REPLACE_MBSTATE_T''@|$(REPLACE_MBSTATE_T)|g' \ + -e 's|@''REPLACE_BTOWC''@|$(REPLACE_BTOWC)|g' \ + -e 's|@''REPLACE_WCTOB''@|$(REPLACE_WCTOB)|g' \ + -e 's|@''REPLACE_MBSINIT''@|$(REPLACE_MBSINIT)|g' \ + -e 's|@''REPLACE_MBRTOWC''@|$(REPLACE_MBRTOWC)|g' \ + -e 's|@''REPLACE_MBRLEN''@|$(REPLACE_MBRLEN)|g' \ + -e 's|@''REPLACE_MBSRTOWCS''@|$(REPLACE_MBSRTOWCS)|g' \ + -e 's|@''REPLACE_MBSNRTOWCS''@|$(REPLACE_MBSNRTOWCS)|g' \ + -e 's|@''REPLACE_WCRTOMB''@|$(REPLACE_WCRTOMB)|g' \ + -e 's|@''REPLACE_WCSRTOMBS''@|$(REPLACE_WCSRTOMBS)|g' \ + -e 's|@''REPLACE_WCSNRTOMBS''@|$(REPLACE_WCSNRTOMBS)|g' \ + -e 's|@''REPLACE_WCWIDTH''@|$(REPLACE_WCWIDTH)|g' \ + -e 's|@''REPLACE_WCSWIDTH''@|$(REPLACE_WCSWIDTH)|g' \ + -e 's|@''REPLACE_WCSFTIME''@|$(REPLACE_WCSFTIME)|g' \ + -e '/definitions of _GL_FUNCDECL_RPL/r $(CXXDEFS_H)' \ + -e '/definition of _GL_ARG_NONNULL/r $(ARG_NONNULL_H)' \ + -e '/definition of _GL_WARN_ON_USE/r $(WARN_ON_USE_H)'; \ + } > $@-t && \ + mv $@-t $@ +MOSTLYCLEANFILES += lib/wchar.h lib/wchar.h-t + +EXTRA_DIST += lib/wchar.in.h + +## end gnulib module wchar + +## begin gnulib module wcrtomb + + +EXTRA_DIST += lib/wcrtomb.c + +EXTRA_lib_libsed_a_SOURCES += lib/wcrtomb.c + +## end gnulib module wcrtomb + +## begin gnulib module wctob + + +EXTRA_DIST += lib/wctob.c + +EXTRA_lib_libsed_a_SOURCES += lib/wctob.c + +## end gnulib module wctob + +## begin gnulib module wctomb + + +EXTRA_DIST += lib/wctomb-impl.h lib/wctomb.c + +EXTRA_lib_libsed_a_SOURCES += lib/wctomb.c + +## end gnulib module wctomb + +## begin gnulib module wctype-h + +BUILT_SOURCES += lib/wctype.h +lib_libsed_a_SOURCES += lib/wctype-h.c + +# We need the following in order to create <wctype.h> when the system +# doesn't have one that works with the given compiler. +lib/wctype.h: lib/wctype.in.h $(top_builddir)/config.status $(CXXDEFS_H) $(WARN_ON_USE_H) + $(AM_V_GEN)rm -f $@-t $@ && \ + { echo '/* DO NOT EDIT! GENERATED AUTOMATICALLY! */'; \ + sed -e 's|@''GUARD_PREFIX''@|GL|g' \ + -e 's/@''HAVE_WCTYPE_H''@/$(HAVE_WCTYPE_H)/g' \ + -e 's|@''INCLUDE_NEXT''@|$(INCLUDE_NEXT)|g' \ + -e 's|@''PRAGMA_SYSTEM_HEADER''@|@PRAGMA_SYSTEM_HEADER@|g' \ + -e 's|@''PRAGMA_COLUMNS''@|@PRAGMA_COLUMNS@|g' \ + -e 's|@''NEXT_WCTYPE_H''@|$(NEXT_WCTYPE_H)|g' \ + -e 's/@''HAVE_CRTDEFS_H''@/$(HAVE_CRTDEFS_H)/g' \ + -e 's/@''GNULIB_OVERRIDES_WINT_T''@/$(GNULIB_OVERRIDES_WINT_T)/g' \ + -e 's/@''GNULIB_ISWBLANK''@/$(GNULIB_ISWBLANK)/g' \ + -e 's/@''GNULIB_WCTYPE''@/$(GNULIB_WCTYPE)/g' \ + -e 's/@''GNULIB_ISWCTYPE''@/$(GNULIB_ISWCTYPE)/g' \ + -e 's/@''GNULIB_WCTRANS''@/$(GNULIB_WCTRANS)/g' \ + -e 's/@''GNULIB_TOWCTRANS''@/$(GNULIB_TOWCTRANS)/g' \ + -e 's/@''HAVE_ISWBLANK''@/$(HAVE_ISWBLANK)/g' \ + -e 's/@''HAVE_ISWCNTRL''@/$(HAVE_ISWCNTRL)/g' \ + -e 's/@''HAVE_WCTYPE_T''@/$(HAVE_WCTYPE_T)/g' \ + -e 's/@''HAVE_WCTRANS_T''@/$(HAVE_WCTRANS_T)/g' \ + -e 's/@''HAVE_WINT_T''@/$(HAVE_WINT_T)/g' \ + -e 's/@''REPLACE_ISWBLANK''@/$(REPLACE_ISWBLANK)/g' \ + -e 's/@''REPLACE_ISWCNTRL''@/$(REPLACE_ISWCNTRL)/g' \ + -e 's/@''REPLACE_TOWLOWER''@/$(REPLACE_TOWLOWER)/g' \ + -e '/definitions of _GL_FUNCDECL_RPL/r $(CXXDEFS_H)' \ + -e '/definition of _GL_WARN_ON_USE/r $(WARN_ON_USE_H)' \ + < $(top_srcdir)/lib/wctype.in.h; \ + } > $@-t && \ + mv $@-t $@ +MOSTLYCLEANFILES += lib/wctype.h lib/wctype.h-t + +EXTRA_DIST += lib/wctype.in.h + +## end gnulib module wctype-h + +## begin gnulib module xalloc + +lib_libsed_a_SOURCES += lib/xmalloc.c + +EXTRA_DIST += lib/xalloc.h + +## end gnulib module xalloc + +## begin gnulib module xalloc-die + +lib_libsed_a_SOURCES += lib/xalloc-die.c + +## end gnulib module xalloc-die + +## begin gnulib module xalloc-oversized + + +EXTRA_DIST += lib/xalloc-oversized.h + +## end gnulib module xalloc-oversized + + +mostlyclean-local: mostlyclean-generic + @for dir in '' $(MOSTLYCLEANDIRS); do \ + if test -n "$$dir" && test -d $$dir; then \ + echo "rmdir $$dir"; rmdir $$dir; \ + fi; \ + done; \ + : diff --git a/lib/hard-locale.c b/lib/hard-locale.c new file mode 100644 index 0000000..49bc6ca --- /dev/null +++ b/lib/hard-locale.c @@ -0,0 +1,72 @@ +/* hard-locale.c -- Determine whether a locale is hard. + + Copyright (C) 1997-1999, 2002-2004, 2006-2007, 2009-2018 Free Software + Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +#include <config.h> + +#include "hard-locale.h" + +#include <locale.h> +#include <stdlib.h> +#include <string.h> + +#ifdef __GLIBC__ +# define GLIBC_VERSION __GLIBC__ +#elif defined __UCLIBC__ +# define GLIBC_VERSION 2 +#else +# define GLIBC_VERSION 0 +#endif + +/* Return true if the current CATEGORY locale is hard, i.e. if you + can't get away with assuming traditional C or POSIX behavior. */ +bool +hard_locale (int category) +{ + bool hard = true; + char const *p = setlocale (category, NULL); + + if (p) + { + if (2 <= GLIBC_VERSION) + { + if (strcmp (p, "C") == 0 || strcmp (p, "POSIX") == 0) + hard = false; + } + else + { + char *locale = strdup (p); + if (locale) + { + /* Temporarily set the locale to the "C" and "POSIX" locales + to find their names, so that we can determine whether one + or the other is the caller's locale. */ + if (((p = setlocale (category, "C")) + && strcmp (p, locale) == 0) + || ((p = setlocale (category, "POSIX")) + && strcmp (p, locale) == 0)) + hard = false; + + /* Restore the caller's locale. */ + setlocale (category, locale); + free (locale); + } + } + } + + return hard; +} diff --git a/lib/hard-locale.h b/lib/hard-locale.h new file mode 100644 index 0000000..22eecc5 --- /dev/null +++ b/lib/hard-locale.h @@ -0,0 +1,25 @@ +/* Determine whether a locale is hard. + + Copyright (C) 1999, 2003-2004, 2009-2018 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +#ifndef HARD_LOCALE_H_ +# define HARD_LOCALE_H_ 1 + +# include <stdbool.h> + +bool hard_locale (int); + +#endif /* HARD_LOCALE_H_ */ diff --git a/lib/ignore-value.h b/lib/ignore-value.h new file mode 100644 index 0000000..95eac1c --- /dev/null +++ b/lib/ignore-value.h @@ -0,0 +1,50 @@ +/* ignore a function return without a compiler warning. -*- coding: utf-8 -*- + + Copyright (C) 2008-2018 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +/* Written by Jim Meyering, Eric Blake and Pádraig Brady. */ + +/* Use "ignore_value" to avoid a warning when using a function declared with + gcc's warn_unused_result attribute, but for which you really do want to + ignore the result. Traditionally, people have used a "(void)" cast to + indicate that a function's return value is deliberately unused. However, + if the function is declared with __attribute__((warn_unused_result)), + gcc issues a warning even with the cast. + + Caution: most of the time, you really should heed gcc's warning, and + check the return value. However, in those exceptional cases in which + you're sure you know what you're doing, use this function. + + For the record, here's one of the ignorable warnings: + "copy.c:233: warning: ignoring return value of 'fchown', + declared with attribute warn_unused_result". */ + +#ifndef _GL_IGNORE_VALUE_H +#define _GL_IGNORE_VALUE_H + +/* Normally casting an expression to void discards its value, but GCC + versions 3.4 and newer have __attribute__ ((__warn_unused_result__)) + which may cause unwanted diagnostics in that case. Use __typeof__ + and __extension__ to work around the problem, if the workaround is + known to be needed. */ +#if 3 < __GNUC__ + (4 <= __GNUC_MINOR__) +# define ignore_value(x) \ + (__extension__ ({ __typeof__ (x) __x = (x); (void) __x; })) +#else +# define ignore_value(x) ((void) (x)) +#endif + +#endif diff --git a/lib/intprops.h b/lib/intprops.h new file mode 100644 index 0000000..cdaf658 --- /dev/null +++ b/lib/intprops.h @@ -0,0 +1,455 @@ +/* intprops.h -- properties of integer types + + Copyright (C) 2001-2018 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify it + under the terms of the GNU General Public License as published + by the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +/* Written by Paul Eggert. */ + +#ifndef _GL_INTPROPS_H +#define _GL_INTPROPS_H + +#include <limits.h> + +/* Return a value with the common real type of E and V and the value of V. + Do not evaluate E. */ +#define _GL_INT_CONVERT(e, v) ((1 ? 0 : (e)) + (v)) + +/* Act like _GL_INT_CONVERT (E, -V) but work around a bug in IRIX 6.5 cc; see + <https://lists.gnu.org/r/bug-gnulib/2011-05/msg00406.html>. */ +#define _GL_INT_NEGATE_CONVERT(e, v) ((1 ? 0 : (e)) - (v)) + +/* The extra casts in the following macros work around compiler bugs, + e.g., in Cray C 5.0.3.0. */ + +/* True if the arithmetic type T is an integer type. bool counts as + an integer. */ +#define TYPE_IS_INTEGER(t) ((t) 1.5 == 1) + +/* True if the real type T is signed. */ +#define TYPE_SIGNED(t) (! ((t) 0 < (t) -1)) + +/* Return 1 if the real expression E, after promotion, has a + signed or floating type. Do not evaluate E. */ +#define EXPR_SIGNED(e) (_GL_INT_NEGATE_CONVERT (e, 1) < 0) + + +/* Minimum and maximum values for integer types and expressions. */ + +/* The width in bits of the integer type or expression T. + Do not evaluate T. + Padding bits are not supported; this is checked at compile-time below. */ +#define TYPE_WIDTH(t) (sizeof (t) * CHAR_BIT) + +/* The maximum and minimum values for the integer type T. */ +#define TYPE_MINIMUM(t) ((t) ~ TYPE_MAXIMUM (t)) +#define TYPE_MAXIMUM(t) \ + ((t) (! TYPE_SIGNED (t) \ + ? (t) -1 \ + : ((((t) 1 << (TYPE_WIDTH (t) - 2)) - 1) * 2 + 1))) + +/* The maximum and minimum values for the type of the expression E, + after integer promotion. E is not evaluated. */ +#define _GL_INT_MINIMUM(e) \ + (EXPR_SIGNED (e) \ + ? ~ _GL_SIGNED_INT_MAXIMUM (e) \ + : _GL_INT_CONVERT (e, 0)) +#define _GL_INT_MAXIMUM(e) \ + (EXPR_SIGNED (e) \ + ? _GL_SIGNED_INT_MAXIMUM (e) \ + : _GL_INT_NEGATE_CONVERT (e, 1)) +#define _GL_SIGNED_INT_MAXIMUM(e) \ + (((_GL_INT_CONVERT (e, 1) << (TYPE_WIDTH ((e) + 0) - 2)) - 1) * 2 + 1) + +/* Work around OpenVMS incompatibility with C99. */ +#if !defined LLONG_MAX && defined __INT64_MAX +# define LLONG_MAX __INT64_MAX +# define LLONG_MIN __INT64_MIN +#endif + +/* This include file assumes that signed types are two's complement without + padding bits; the above macros have undefined behavior otherwise. + If this is a problem for you, please let us know how to fix it for your host. + This assumption is tested by the intprops-tests module. */ + +/* Does the __typeof__ keyword work? This could be done by + 'configure', but for now it's easier to do it by hand. */ +#if (2 <= __GNUC__ \ + || (1210 <= __IBMC__ && defined __IBM__TYPEOF__) \ + || (0x5110 <= __SUNPRO_C && !__STDC__)) +# define _GL_HAVE___TYPEOF__ 1 +#else +# define _GL_HAVE___TYPEOF__ 0 +#endif + +/* Return 1 if the integer type or expression T might be signed. Return 0 + if it is definitely unsigned. This macro does not evaluate its argument, + and expands to an integer constant expression. */ +#if _GL_HAVE___TYPEOF__ +# define _GL_SIGNED_TYPE_OR_EXPR(t) TYPE_SIGNED (__typeof__ (t)) +#else +# define _GL_SIGNED_TYPE_OR_EXPR(t) 1 +#endif + +/* Bound on length of the string representing an unsigned integer + value representable in B bits. log10 (2.0) < 146/485. The + smallest value of B where this bound is not tight is 2621. */ +#define INT_BITS_STRLEN_BOUND(b) (((b) * 146 + 484) / 485) + +/* Bound on length of the string representing an integer type or expression T. + Subtract 1 for the sign bit if T is signed, and then add 1 more for + a minus sign if needed. + + Because _GL_SIGNED_TYPE_OR_EXPR sometimes returns 0 when its argument is + signed, this macro may overestimate the true bound by one byte when + applied to unsigned types of size 2, 4, 16, ... bytes. */ +#define INT_STRLEN_BOUND(t) \ + (INT_BITS_STRLEN_BOUND (TYPE_WIDTH (t) - _GL_SIGNED_TYPE_OR_EXPR (t)) \ + + _GL_SIGNED_TYPE_OR_EXPR (t)) + +/* Bound on buffer size needed to represent an integer type or expression T, + including the terminating null. */ +#define INT_BUFSIZE_BOUND(t) (INT_STRLEN_BOUND (t) + 1) + + +/* Range overflow checks. + + The INT_<op>_RANGE_OVERFLOW macros return 1 if the corresponding C + operators might not yield numerically correct answers due to + arithmetic overflow. They do not rely on undefined or + implementation-defined behavior. Their implementations are simple + and straightforward, but they are a bit harder to use than the + INT_<op>_OVERFLOW macros described below. + + Example usage: + + long int i = ...; + long int j = ...; + if (INT_MULTIPLY_RANGE_OVERFLOW (i, j, LONG_MIN, LONG_MAX)) + printf ("multiply would overflow"); + else + printf ("product is %ld", i * j); + + Restrictions on *_RANGE_OVERFLOW macros: + + These macros do not check for all possible numerical problems or + undefined or unspecified behavior: they do not check for division + by zero, for bad shift counts, or for shifting negative numbers. + + These macros may evaluate their arguments zero or multiple times, + so the arguments should not have side effects. The arithmetic + arguments (including the MIN and MAX arguments) must be of the same + integer type after the usual arithmetic conversions, and the type + must have minimum value MIN and maximum MAX. Unsigned types should + use a zero MIN of the proper type. + + These macros are tuned for constant MIN and MAX. For commutative + operations such as A + B, they are also tuned for constant B. */ + +/* Return 1 if A + B would overflow in [MIN,MAX] arithmetic. + See above for restrictions. */ +#define INT_ADD_RANGE_OVERFLOW(a, b, min, max) \ + ((b) < 0 \ + ? (a) < (min) - (b) \ + : (max) - (b) < (a)) + +/* Return 1 if A - B would overflow in [MIN,MAX] arithmetic. + See above for restrictions. */ +#define INT_SUBTRACT_RANGE_OVERFLOW(a, b, min, max) \ + ((b) < 0 \ + ? (max) + (b) < (a) \ + : (a) < (min) + (b)) + +/* Return 1 if - A would overflow in [MIN,MAX] arithmetic. + See above for restrictions. */ +#define INT_NEGATE_RANGE_OVERFLOW(a, min, max) \ + ((min) < 0 \ + ? (a) < - (max) \ + : 0 < (a)) + +/* Return 1 if A * B would overflow in [MIN,MAX] arithmetic. + See above for restrictions. Avoid && and || as they tickle + bugs in Sun C 5.11 2010/08/13 and other compilers; see + <https://lists.gnu.org/r/bug-gnulib/2011-05/msg00401.html>. */ +#define INT_MULTIPLY_RANGE_OVERFLOW(a, b, min, max) \ + ((b) < 0 \ + ? ((a) < 0 \ + ? (a) < (max) / (b) \ + : (b) == -1 \ + ? 0 \ + : (min) / (b) < (a)) \ + : (b) == 0 \ + ? 0 \ + : ((a) < 0 \ + ? (a) < (min) / (b) \ + : (max) / (b) < (a))) + +/* Return 1 if A / B would overflow in [MIN,MAX] arithmetic. + See above for restrictions. Do not check for division by zero. */ +#define INT_DIVIDE_RANGE_OVERFLOW(a, b, min, max) \ + ((min) < 0 && (b) == -1 && (a) < - (max)) + +/* Return 1 if A % B would overflow in [MIN,MAX] arithmetic. + See above for restrictions. Do not check for division by zero. + Mathematically, % should never overflow, but on x86-like hosts + INT_MIN % -1 traps, and the C standard permits this, so treat this + as an overflow too. */ +#define INT_REMAINDER_RANGE_OVERFLOW(a, b, min, max) \ + INT_DIVIDE_RANGE_OVERFLOW (a, b, min, max) + +/* Return 1 if A << B would overflow in [MIN,MAX] arithmetic. + See above for restrictions. Here, MIN and MAX are for A only, and B need + not be of the same type as the other arguments. The C standard says that + behavior is undefined for shifts unless 0 <= B < wordwidth, and that when + A is negative then A << B has undefined behavior and A >> B has + implementation-defined behavior, but do not check these other + restrictions. */ +#define INT_LEFT_SHIFT_RANGE_OVERFLOW(a, b, min, max) \ + ((a) < 0 \ + ? (a) < (min) >> (b) \ + : (max) >> (b) < (a)) + +/* True if __builtin_add_overflow (A, B, P) works when P is non-null. */ +#if 5 <= __GNUC__ && !defined __ICC +# define _GL_HAS_BUILTIN_OVERFLOW 1 +#else +# define _GL_HAS_BUILTIN_OVERFLOW 0 +#endif + +/* True if __builtin_add_overflow_p (A, B, C) works. */ +#define _GL_HAS_BUILTIN_OVERFLOW_P (7 <= __GNUC__) + +/* The _GL*_OVERFLOW macros have the same restrictions as the + *_RANGE_OVERFLOW macros, except that they do not assume that operands + (e.g., A and B) have the same type as MIN and MAX. Instead, they assume + that the result (e.g., A + B) has that type. */ +#if _GL_HAS_BUILTIN_OVERFLOW_P +# define _GL_ADD_OVERFLOW(a, b, min, max) \ + __builtin_add_overflow_p (a, b, (__typeof__ ((a) + (b))) 0) +# define _GL_SUBTRACT_OVERFLOW(a, b, min, max) \ + __builtin_sub_overflow_p (a, b, (__typeof__ ((a) - (b))) 0) +# define _GL_MULTIPLY_OVERFLOW(a, b, min, max) \ + __builtin_mul_overflow_p (a, b, (__typeof__ ((a) * (b))) 0) +#else +# define _GL_ADD_OVERFLOW(a, b, min, max) \ + ((min) < 0 ? INT_ADD_RANGE_OVERFLOW (a, b, min, max) \ + : (a) < 0 ? (b) <= (a) + (b) \ + : (b) < 0 ? (a) <= (a) + (b) \ + : (a) + (b) < (b)) +# define _GL_SUBTRACT_OVERFLOW(a, b, min, max) \ + ((min) < 0 ? INT_SUBTRACT_RANGE_OVERFLOW (a, b, min, max) \ + : (a) < 0 ? 1 \ + : (b) < 0 ? (a) - (b) <= (a) \ + : (a) < (b)) +# define _GL_MULTIPLY_OVERFLOW(a, b, min, max) \ + (((min) == 0 && (((a) < 0 && 0 < (b)) || ((b) < 0 && 0 < (a)))) \ + || INT_MULTIPLY_RANGE_OVERFLOW (a, b, min, max)) +#endif +#define _GL_DIVIDE_OVERFLOW(a, b, min, max) \ + ((min) < 0 ? (b) == _GL_INT_NEGATE_CONVERT (min, 1) && (a) < - (max) \ + : (a) < 0 ? (b) <= (a) + (b) - 1 \ + : (b) < 0 && (a) + (b) <= (a)) +#define _GL_REMAINDER_OVERFLOW(a, b, min, max) \ + ((min) < 0 ? (b) == _GL_INT_NEGATE_CONVERT (min, 1) && (a) < - (max) \ + : (a) < 0 ? (a) % (b) != ((max) - (b) + 1) % (b) \ + : (b) < 0 && ! _GL_UNSIGNED_NEG_MULTIPLE (a, b, max)) + +/* Return a nonzero value if A is a mathematical multiple of B, where + A is unsigned, B is negative, and MAX is the maximum value of A's + type. A's type must be the same as (A % B)'s type. Normally (A % + -B == 0) suffices, but things get tricky if -B would overflow. */ +#define _GL_UNSIGNED_NEG_MULTIPLE(a, b, max) \ + (((b) < -_GL_SIGNED_INT_MAXIMUM (b) \ + ? (_GL_SIGNED_INT_MAXIMUM (b) == (max) \ + ? (a) \ + : (a) % (_GL_INT_CONVERT (a, _GL_SIGNED_INT_MAXIMUM (b)) + 1)) \ + : (a) % - (b)) \ + == 0) + +/* Check for integer overflow, and report low order bits of answer. + + The INT_<op>_OVERFLOW macros return 1 if the corresponding C operators + might not yield numerically correct answers due to arithmetic overflow. + The INT_<op>_WRAPV macros also store the low-order bits of the answer. + These macros work correctly on all known practical hosts, and do not rely + on undefined behavior due to signed arithmetic overflow. + + Example usage, assuming A and B are long int: + + if (INT_MULTIPLY_OVERFLOW (a, b)) + printf ("result would overflow\n"); + else + printf ("result is %ld (no overflow)\n", a * b); + + Example usage with WRAPV flavor: + + long int result; + bool overflow = INT_MULTIPLY_WRAPV (a, b, &result); + printf ("result is %ld (%s)\n", result, + overflow ? "after overflow" : "no overflow"); + + Restrictions on these macros: + + These macros do not check for all possible numerical problems or + undefined or unspecified behavior: they do not check for division + by zero, for bad shift counts, or for shifting negative numbers. + + These macros may evaluate their arguments zero or multiple times, so the + arguments should not have side effects. + + The WRAPV macros are not constant expressions. They support only + +, binary -, and *. The result type must be signed. + + These macros are tuned for their last argument being a constant. + + Return 1 if the integer expressions A * B, A - B, -A, A * B, A / B, + A % B, and A << B would overflow, respectively. */ + +#define INT_ADD_OVERFLOW(a, b) \ + _GL_BINARY_OP_OVERFLOW (a, b, _GL_ADD_OVERFLOW) +#define INT_SUBTRACT_OVERFLOW(a, b) \ + _GL_BINARY_OP_OVERFLOW (a, b, _GL_SUBTRACT_OVERFLOW) +#if _GL_HAS_BUILTIN_OVERFLOW_P +# define INT_NEGATE_OVERFLOW(a) INT_SUBTRACT_OVERFLOW (0, a) +#else +# define INT_NEGATE_OVERFLOW(a) \ + INT_NEGATE_RANGE_OVERFLOW (a, _GL_INT_MINIMUM (a), _GL_INT_MAXIMUM (a)) +#endif +#define INT_MULTIPLY_OVERFLOW(a, b) \ + _GL_BINARY_OP_OVERFLOW (a, b, _GL_MULTIPLY_OVERFLOW) +#define INT_DIVIDE_OVERFLOW(a, b) \ + _GL_BINARY_OP_OVERFLOW (a, b, _GL_DIVIDE_OVERFLOW) +#define INT_REMAINDER_OVERFLOW(a, b) \ + _GL_BINARY_OP_OVERFLOW (a, b, _GL_REMAINDER_OVERFLOW) +#define INT_LEFT_SHIFT_OVERFLOW(a, b) \ + INT_LEFT_SHIFT_RANGE_OVERFLOW (a, b, \ + _GL_INT_MINIMUM (a), _GL_INT_MAXIMUM (a)) + +/* Return 1 if the expression A <op> B would overflow, + where OP_RESULT_OVERFLOW (A, B, MIN, MAX) does the actual test, + assuming MIN and MAX are the minimum and maximum for the result type. + Arguments should be free of side effects. */ +#define _GL_BINARY_OP_OVERFLOW(a, b, op_result_overflow) \ + op_result_overflow (a, b, \ + _GL_INT_MINIMUM (_GL_INT_CONVERT (a, b)), \ + _GL_INT_MAXIMUM (_GL_INT_CONVERT (a, b))) + +/* Store the low-order bits of A + B, A - B, A * B, respectively, into *R. + Return 1 if the result overflows. See above for restrictions. */ +#define INT_ADD_WRAPV(a, b, r) \ + _GL_INT_OP_WRAPV (a, b, r, +, __builtin_add_overflow, INT_ADD_OVERFLOW) +#define INT_SUBTRACT_WRAPV(a, b, r) \ + _GL_INT_OP_WRAPV (a, b, r, -, __builtin_sub_overflow, INT_SUBTRACT_OVERFLOW) +#define INT_MULTIPLY_WRAPV(a, b, r) \ + _GL_INT_OP_WRAPV (a, b, r, *, __builtin_mul_overflow, INT_MULTIPLY_OVERFLOW) + +/* Nonzero if this compiler has GCC bug 68193 or Clang bug 25390. See: + https://gcc.gnu.org/bugzilla/show_bug.cgi?id=68193 + https://llvm.org/bugs/show_bug.cgi?id=25390 + For now, assume all versions of GCC-like compilers generate bogus + warnings for _Generic. This matters only for older compilers that + lack __builtin_add_overflow. */ +#if __GNUC__ +# define _GL__GENERIC_BOGUS 1 +#else +# define _GL__GENERIC_BOGUS 0 +#endif + +/* Store the low-order bits of A <op> B into *R, where OP specifies + the operation. BUILTIN is the builtin operation, and OVERFLOW the + overflow predicate. Return 1 if the result overflows. See above + for restrictions. */ +#if _GL_HAS_BUILTIN_OVERFLOW +# define _GL_INT_OP_WRAPV(a, b, r, op, builtin, overflow) builtin (a, b, r) +#elif 201112 <= __STDC_VERSION__ && !_GL__GENERIC_BOGUS +# define _GL_INT_OP_WRAPV(a, b, r, op, builtin, overflow) \ + (_Generic \ + (*(r), \ + signed char: \ + _GL_INT_OP_CALC (a, b, r, op, overflow, unsigned int, \ + signed char, SCHAR_MIN, SCHAR_MAX), \ + short int: \ + _GL_INT_OP_CALC (a, b, r, op, overflow, unsigned int, \ + short int, SHRT_MIN, SHRT_MAX), \ + int: \ + _GL_INT_OP_CALC (a, b, r, op, overflow, unsigned int, \ + int, INT_MIN, INT_MAX), \ + long int: \ + _GL_INT_OP_CALC (a, b, r, op, overflow, unsigned long int, \ + long int, LONG_MIN, LONG_MAX), \ + long long int: \ + _GL_INT_OP_CALC (a, b, r, op, overflow, unsigned long long int, \ + long long int, LLONG_MIN, LLONG_MAX))) +#else +# define _GL_INT_OP_WRAPV(a, b, r, op, builtin, overflow) \ + (sizeof *(r) == sizeof (signed char) \ + ? _GL_INT_OP_CALC (a, b, r, op, overflow, unsigned int, \ + signed char, SCHAR_MIN, SCHAR_MAX) \ + : sizeof *(r) == sizeof (short int) \ + ? _GL_INT_OP_CALC (a, b, r, op, overflow, unsigned int, \ + short int, SHRT_MIN, SHRT_MAX) \ + : sizeof *(r) == sizeof (int) \ + ? _GL_INT_OP_CALC (a, b, r, op, overflow, unsigned int, \ + int, INT_MIN, INT_MAX) \ + : _GL_INT_OP_WRAPV_LONGISH(a, b, r, op, overflow)) +# ifdef LLONG_MAX +# define _GL_INT_OP_WRAPV_LONGISH(a, b, r, op, overflow) \ + (sizeof *(r) == sizeof (long int) \ + ? _GL_INT_OP_CALC (a, b, r, op, overflow, unsigned long int, \ + long int, LONG_MIN, LONG_MAX) \ + : _GL_INT_OP_CALC (a, b, r, op, overflow, unsigned long long int, \ + long long int, LLONG_MIN, LLONG_MAX)) +# else +# define _GL_INT_OP_WRAPV_LONGISH(a, b, r, op, overflow) \ + _GL_INT_OP_CALC (a, b, r, op, overflow, unsigned long int, \ + long int, LONG_MIN, LONG_MAX) +# endif +#endif + +/* Store the low-order bits of A <op> B into *R, where the operation + is given by OP. Use the unsigned type UT for calculation to avoid + overflow problems. *R's type is T, with extrema TMIN and TMAX. + T must be a signed integer type. Return 1 if the result overflows. */ +#define _GL_INT_OP_CALC(a, b, r, op, overflow, ut, t, tmin, tmax) \ + (sizeof ((a) op (b)) < sizeof (t) \ + ? _GL_INT_OP_CALC1 ((t) (a), (t) (b), r, op, overflow, ut, t, tmin, tmax) \ + : _GL_INT_OP_CALC1 (a, b, r, op, overflow, ut, t, tmin, tmax)) +#define _GL_INT_OP_CALC1(a, b, r, op, overflow, ut, t, tmin, tmax) \ + ((overflow (a, b) \ + || (EXPR_SIGNED ((a) op (b)) && ((a) op (b)) < (tmin)) \ + || (tmax) < ((a) op (b))) \ + ? (*(r) = _GL_INT_OP_WRAPV_VIA_UNSIGNED (a, b, op, ut, t), 1) \ + : (*(r) = _GL_INT_OP_WRAPV_VIA_UNSIGNED (a, b, op, ut, t), 0)) + +/* Return the low-order bits of A <op> B, where the operation is given + by OP. Use the unsigned type UT for calculation to avoid undefined + behavior on signed integer overflow, and convert the result to type T. + UT is at least as wide as T and is no narrower than unsigned int, + T is two's complement, and there is no padding or trap representations. + Assume that converting UT to T yields the low-order bits, as is + done in all known two's-complement C compilers. E.g., see: + https://gcc.gnu.org/onlinedocs/gcc/Integers-implementation.html + + According to the C standard, converting UT to T yields an + implementation-defined result or signal for values outside T's + range. However, code that works around this theoretical problem + runs afoul of a compiler bug in Oracle Studio 12.3 x86. See: + https://lists.gnu.org/r/bug-gnulib/2017-04/msg00049.html + As the compiler bug is real, don't try to work around the + theoretical problem. */ + +#define _GL_INT_OP_WRAPV_VIA_UNSIGNED(a, b, op, ut, t) \ + ((t) ((ut) (a) op (ut) (b))) + +#endif /* _GL_INTPROPS_H */ diff --git a/lib/isblank.c b/lib/isblank.c new file mode 100644 index 0000000..df4b4aa --- /dev/null +++ b/lib/isblank.c @@ -0,0 +1,33 @@ +/* Test whether a character is a blank. + + Copyright (C) 2009-2018 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +#include <config.h> + +/* Specification. */ +#include <ctype.h> + +int +isblank (int c) +{ + /* On all known platforms, in all predefined locales, isblank(c) is likely + equivalent with (c == ' ' || c == '\t'). Look at the glibc definition + (in glibc/localedata/locales/i18n): The "blank" characters are '\t', ' ', + U+1680, U+180E, U+2000..U+2006, U+2008..U+200A, U+205F, U+3000, and none + except the first two is present in a common 8-bit encoding. Therefore + the substitute for other platforms is not more complicated than this. */ + return (c == ' ' || c == '\t'); +} diff --git a/lib/langinfo.in.h b/lib/langinfo.in.h new file mode 100644 index 0000000..96c237a --- /dev/null +++ b/lib/langinfo.in.h @@ -0,0 +1,222 @@ +/* Substitute for and wrapper around <langinfo.h>. + Copyright (C) 2009-2018 Free Software Foundation, Inc. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3, or (at your option) + any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, see <https://www.gnu.org/licenses/>. */ + +/* + * POSIX <langinfo.h> for platforms that lack it or have an incomplete one. + * <http://www.opengroup.org/onlinepubs/9699919799/basedefs/langinfo.h.html> + */ + +#ifndef _@GUARD_PREFIX@_LANGINFO_H + +#if __GNUC__ >= 3 +@PRAGMA_SYSTEM_HEADER@ +#endif +@PRAGMA_COLUMNS@ + +/* The include_next requires a split double-inclusion guard. */ +#if @HAVE_LANGINFO_H@ +# @INCLUDE_NEXT@ @NEXT_LANGINFO_H@ +#endif + +#ifndef _@GUARD_PREFIX@_LANGINFO_H +#define _@GUARD_PREFIX@_LANGINFO_H + + +#if !@HAVE_LANGINFO_H@ + +/* A platform that lacks <langinfo.h>. */ + +/* Assume that it also lacks <nl_types.h> and the nl_item type. */ +# if !GNULIB_defined_nl_item +typedef int nl_item; +# define GNULIB_defined_nl_item 1 +# endif + +/* nl_langinfo items of the LC_CTYPE category */ +# define CODESET 10000 +/* nl_langinfo items of the LC_NUMERIC category */ +# define RADIXCHAR 10001 +# define DECIMAL_POINT RADIXCHAR +# define THOUSEP 10002 +# define THOUSANDS_SEP THOUSEP +# define GROUPING 10114 +/* nl_langinfo items of the LC_TIME category */ +# define D_T_FMT 10003 +# define D_FMT 10004 +# define T_FMT 10005 +# define T_FMT_AMPM 10006 +# define AM_STR 10007 +# define PM_STR 10008 +# define DAY_1 10009 +# define DAY_2 (DAY_1 + 1) +# define DAY_3 (DAY_1 + 2) +# define DAY_4 (DAY_1 + 3) +# define DAY_5 (DAY_1 + 4) +# define DAY_6 (DAY_1 + 5) +# define DAY_7 (DAY_1 + 6) +# define ABDAY_1 10016 +# define ABDAY_2 (ABDAY_1 + 1) +# define ABDAY_3 (ABDAY_1 + 2) +# define ABDAY_4 (ABDAY_1 + 3) +# define ABDAY_5 (ABDAY_1 + 4) +# define ABDAY_6 (ABDAY_1 + 5) +# define ABDAY_7 (ABDAY_1 + 6) +# define MON_1 10023 +# define MON_2 (MON_1 + 1) +# define MON_3 (MON_1 + 2) +# define MON_4 (MON_1 + 3) +# define MON_5 (MON_1 + 4) +# define MON_6 (MON_1 + 5) +# define MON_7 (MON_1 + 6) +# define MON_8 (MON_1 + 7) +# define MON_9 (MON_1 + 8) +# define MON_10 (MON_1 + 9) +# define MON_11 (MON_1 + 10) +# define MON_12 (MON_1 + 11) +# define ALTMON_1 10200 +# define ALTMON_2 (ALTMON_1 + 1) +# define ALTMON_3 (ALTMON_1 + 2) +# define ALTMON_4 (ALTMON_1 + 3) +# define ALTMON_5 (ALTMON_1 + 4) +# define ALTMON_6 (ALTMON_1 + 5) +# define ALTMON_7 (ALTMON_1 + 6) +# define ALTMON_8 (ALTMON_1 + 7) +# define ALTMON_9 (ALTMON_1 + 8) +# define ALTMON_10 (ALTMON_1 + 9) +# define ALTMON_11 (ALTMON_1 + 10) +# define ALTMON_12 (ALTMON_1 + 11) +# define ABMON_1 10035 +# define ABMON_2 (ABMON_1 + 1) +# define ABMON_3 (ABMON_1 + 2) +# define ABMON_4 (ABMON_1 + 3) +# define ABMON_5 (ABMON_1 + 4) +# define ABMON_6 (ABMON_1 + 5) +# define ABMON_7 (ABMON_1 + 6) +# define ABMON_8 (ABMON_1 + 7) +# define ABMON_9 (ABMON_1 + 8) +# define ABMON_10 (ABMON_1 + 9) +# define ABMON_11 (ABMON_1 + 10) +# define ABMON_12 (ABMON_1 + 11) +# define ERA 10047 +# define ERA_D_FMT 10048 +# define ERA_D_T_FMT 10049 +# define ERA_T_FMT 10050 +# define ALT_DIGITS 10051 +/* nl_langinfo items of the LC_MONETARY category */ +# define CRNCYSTR 10052 +# define CURRENCY_SYMBOL CRNCYSTR +# define INT_CURR_SYMBOL 10100 +# define MON_DECIMAL_POINT 10101 +# define MON_THOUSANDS_SEP 10102 +# define MON_GROUPING 10103 +# define POSITIVE_SIGN 10104 +# define NEGATIVE_SIGN 10105 +# define FRAC_DIGITS 10106 +# define INT_FRAC_DIGITS 10107 +# define P_CS_PRECEDES 10108 +# define N_CS_PRECEDES 10109 +# define P_SEP_BY_SPACE 10110 +# define N_SEP_BY_SPACE 10111 +# define P_SIGN_POSN 10112 +# define N_SIGN_POSN 10113 +/* nl_langinfo items of the LC_MESSAGES category */ +# define YESEXPR 10053 +# define NOEXPR 10054 + +#else + +/* A platform that has <langinfo.h>. */ + +# if !@HAVE_LANGINFO_CODESET@ +# define CODESET 10000 +# define GNULIB_defined_CODESET 1 +# endif + +# if !@HAVE_LANGINFO_T_FMT_AMPM@ +# define T_FMT_AMPM 10006 +# define GNULIB_defined_T_FMT_AMPM 1 +# endif + +# if !@HAVE_LANGINFO_ALTMON@ +# define ALTMON_1 10200 +# define ALTMON_2 (ALTMON_1 + 1) +# define ALTMON_3 (ALTMON_1 + 2) +# define ALTMON_4 (ALTMON_1 + 3) +# define ALTMON_5 (ALTMON_1 + 4) +# define ALTMON_6 (ALTMON_1 + 5) +# define ALTMON_7 (ALTMON_1 + 6) +# define ALTMON_8 (ALTMON_1 + 7) +# define ALTMON_9 (ALTMON_1 + 8) +# define ALTMON_10 (ALTMON_1 + 9) +# define ALTMON_11 (ALTMON_1 + 10) +# define ALTMON_12 (ALTMON_1 + 11) +# define GNULIB_defined_ALTMON 1 +# endif + +# if !@HAVE_LANGINFO_ERA@ +# define ERA 10047 +# define ERA_D_FMT 10048 +# define ERA_D_T_FMT 10049 +# define ERA_T_FMT 10050 +# define ALT_DIGITS 10051 +# define GNULIB_defined_ERA 1 +# endif + +# if !@HAVE_LANGINFO_YESEXPR@ +# define YESEXPR 10053 +# define NOEXPR 10054 +# define GNULIB_defined_YESEXPR 1 +# endif + +#endif + +/* The definitions of _GL_FUNCDECL_RPL etc. are copied here. */ + +/* The definition of _GL_WARN_ON_USE is copied here. */ + +/* Declare overridden functions. */ + + +/* Return a piece of locale dependent information. + Note: The difference between nl_langinfo (CODESET) and locale_charset () + is that the latter normalizes the encoding names to GNU conventions. */ + +#if @GNULIB_NL_LANGINFO@ +# if @REPLACE_NL_LANGINFO@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef nl_langinfo +# define nl_langinfo rpl_nl_langinfo +# endif +_GL_FUNCDECL_RPL (nl_langinfo, char *, (nl_item item)); +_GL_CXXALIAS_RPL (nl_langinfo, char *, (nl_item item)); +# else +# if !@HAVE_NL_LANGINFO@ +_GL_FUNCDECL_SYS (nl_langinfo, char *, (nl_item item)); +# endif +_GL_CXXALIAS_SYS (nl_langinfo, char *, (nl_item item)); +# endif +_GL_CXXALIASWARN (nl_langinfo); +#elif defined GNULIB_POSIXCHECK +# undef nl_langinfo +# if HAVE_RAW_DECL_NL_LANGINFO +_GL_WARN_ON_USE (nl_langinfo, "nl_langinfo is not portable - " + "use gnulib module nl_langinfo for portability"); +# endif +#endif + + +#endif /* _@GUARD_PREFIX@_LANGINFO_H */ +#endif /* _@GUARD_PREFIX@_LANGINFO_H */ diff --git a/lib/libc-config.h b/lib/libc-config.h new file mode 100644 index 0000000..d7b4093 --- /dev/null +++ b/lib/libc-config.h @@ -0,0 +1,174 @@ +/* System definitions for code taken from the GNU C Library + + Copyright 2017-2018 Free Software Foundation, Inc. + + This program is free software; you can redistribute it and/or + modify it under the terms of the GNU General Public + License as published by the Free Software Foundation; either + version 3 of the License, or (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + General Public License for more details. + + You should have received a copy of the GNU General Public + License along with this program; if not, see + <https://www.gnu.org/licenses/>. */ + +/* Written by Paul Eggert. */ + +/* This is intended to be a good-enough substitute for glibc system + macros like those defined in <sys/cdefs.h>, so that Gnulib code + shared with glibc can do this as the first #include: + + #ifndef _LIBC + # include <libc-config.h> + #endif + + When compiled as part of glibc this is a no-op; when compiled as + part of Gnulib this includes Gnulib's <config.h> and defines macros + that glibc library code would normally assume. */ + +#include <config.h> + +/* On glibc this includes <features.h> and <sys/cdefs.h> and #defines + _FEATURES_H, __WORDSIZE, and __set_errno. On FreeBSD 11 it + includes <sys/cdefs.h> which defines __nonnull. Elsewhere it + is harmless. */ +#include <errno.h> + +/* From glibc <errno.h>. */ +#ifndef __set_errno +# define __set_errno(val) (errno = (val)) +#endif + +/* From glibc <features.h>. */ + +#ifndef __GNUC_PREREQ +# if defined __GNUC__ && defined __GNUC_MINOR__ +# define __GNUC_PREREQ(maj, min) ((maj) < __GNUC__ + ((min) <= __GNUC_MINOR__)) +# else +# define __GNUC_PREREQ(maj, min) 0 +# endif +#endif + +#ifndef __glibc_clang_prereq +# if defined __clang_major__ && defined __clang_minor__ +# define __glibc_clang_prereq(maj, min) \ + ((maj) < __clang_major__ + ((min) <= __clang_minor__)) +# else +# define __glibc_clang_prereq(maj, min) 0 +# endif +#endif + + +/* Prepare to include <cdefs.h>, which is our copy of glibc + <sys/cdefs.h>. */ + +/* Define _FEATURES_H so that <cdefs.h> does not include <features.h>. */ +#ifndef _FEATURES_H +# define _FEATURES_H 1 +#endif +/* Define __WORDSIZE so that <cdefs.h> does not attempt to include + nonexistent files. Make it a syntax error, since Gnulib does not + use __WORDSIZE now, and if Gnulib uses it later the syntax error + will let us know that __WORDSIZE needs configuring. */ +#ifndef __WORDSIZE +# define __WORDSIZE %%% +#endif +/* Undef the macros unconditionally defined by our copy of glibc + <sys/cdefs.h>, so that they do not clash with any system-defined + versions. */ +#undef _SYS_CDEFS_H +#undef __ASMNAME +#undef __ASMNAME2 +#undef __BEGIN_DECLS +#undef __CONCAT +#undef __END_DECLS +#undef __HAVE_GENERIC_SELECTION +#undef __LDBL_COMPAT +#undef __LDBL_REDIR +#undef __LDBL_REDIR1 +#undef __LDBL_REDIR1_DECL +#undef __LDBL_REDIR1_NTH +#undef __LDBL_REDIR_DECL +#undef __LDBL_REDIR_NTH +#undef __LEAF +#undef __LEAF_ATTR +#undef __NTH +#undef __NTHNL +#undef __P +#undef __PMT +#undef __REDIRECT +#undef __REDIRECT_LDBL +#undef __REDIRECT_NTH +#undef __REDIRECT_NTHNL +#undef __REDIRECT_NTH_LDBL +#undef __STRING +#undef __THROW +#undef __THROWNL +#undef __always_inline +#undef __attribute__ +#undef __attribute_alloc_size__ +#undef __attribute_artificial__ +#undef __attribute_const__ +#undef __attribute_deprecated__ +#undef __attribute_deprecated_msg__ +#undef __attribute_format_arg__ +#undef __attribute_format_strfmon__ +#undef __attribute_malloc__ +#undef __attribute_noinline__ +#undef __attribute_nonstring__ +#undef __attribute_pure__ +#undef __attribute_used__ +#undef __attribute_warn_unused_result__ +#undef __bos +#undef __bos0 +#undef __errordecl +#undef __extension__ +#undef __extern_always_inline +#undef __extern_inline +#undef __flexarr +#undef __fortify_function +#undef __glibc_c99_flexarr_available +#undef __glibc_clang_has_extension +#undef __glibc_likely +#undef __glibc_macro_warning +#undef __glibc_macro_warning1 +#undef __glibc_unlikely +#undef __inline +#undef __ptr_t +#undef __restrict +#undef __restrict_arr +#undef __va_arg_pack +#undef __va_arg_pack_len +#undef __warnattr +#undef __warndecl + +/* Include our copy of glibc <sys/cdefs.h>. */ +#include <cdefs.h> + +/* <cdefs.h> __inline is too pessimistic for non-GCC. */ +#undef __inline +#ifndef HAVE___INLINE +# if 199901 <= __STDC_VERSION__ || defined inline +# define __inline inline +# else +# define __inline +# endif +#endif + + +/* A substitute for glibc <libc-symbols.h>, good enough for Gnulib. */ +#define attribute_hidden +#define libc_hidden_proto(name, ...) +#define libc_hidden_def(name) +#define libc_hidden_weak(name) +#define libc_hidden_ver(local, name) +#define strong_alias(name, aliasname) +#define weak_alias(name, aliasname) + +/* A substitute for glibc <shlib-compat.h>, good enough for Gnulib. */ +#define SHLIB_COMPAT(lib, introduced, obsoleted) 0 +#define versioned_symbol(lib, local, symbol, version) diff --git a/lib/limits.in.h b/lib/limits.in.h new file mode 100644 index 0000000..89d7195 --- /dev/null +++ b/lib/limits.in.h @@ -0,0 +1,104 @@ +/* A GNU-like <limits.h>. + + Copyright 2016-2018 Free Software Foundation, Inc. + + This program is free software; you can redistribute it and/or + modify it under the terms of the GNU General Public License + as published by the Free Software Foundation; either version 3, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, see <https://www.gnu.org/licenses/>. */ + +#ifndef _@GUARD_PREFIX@_LIMITS_H + +#if __GNUC__ >= 3 +@PRAGMA_SYSTEM_HEADER@ +#endif +@PRAGMA_COLUMNS@ + +/* The include_next requires a split double-inclusion guard. */ +#@INCLUDE_NEXT@ @NEXT_LIMITS_H@ + +#ifndef _@GUARD_PREFIX@_LIMITS_H +#define _@GUARD_PREFIX@_LIMITS_H + +#ifndef LLONG_MIN +# if defined LONG_LONG_MIN /* HP-UX 11.31 */ +# define LLONG_MIN LONG_LONG_MIN +# elif defined LONGLONG_MIN /* IRIX 6.5 */ +# define LLONG_MIN LONGLONG_MIN +# elif defined __GNUC__ +# define LLONG_MIN (- __LONG_LONG_MAX__ - 1LL) +# endif +#endif +#ifndef LLONG_MAX +# if defined LONG_LONG_MAX /* HP-UX 11.31 */ +# define LLONG_MAX LONG_LONG_MAX +# elif defined LONGLONG_MAX /* IRIX 6.5 */ +# define LLONG_MAX LONGLONG_MAX +# elif defined __GNUC__ +# define LLONG_MAX __LONG_LONG_MAX__ +# endif +#endif +#ifndef ULLONG_MAX +# if defined ULONG_LONG_MAX /* HP-UX 11.31 */ +# define ULLONG_MAX ULONG_LONG_MAX +# elif defined ULONGLONG_MAX /* IRIX 6.5 */ +# define ULLONG_MAX ULONGLONG_MAX +# elif defined __GNUC__ +# define ULLONG_MAX (__LONG_LONG_MAX__ * 2ULL + 1ULL) +# endif +#endif + +/* The number of usable bits in an unsigned or signed integer type + with minimum value MIN and maximum value MAX, as an int expression + suitable in #if. Cover all known practical hosts. This + implementation exploits the fact that MAX is 1 less than a power of + 2, and merely counts the number of 1 bits in MAX; "COBn" means + "count the number of 1 bits in the low-order n bits"). */ +#define _GL_INTEGER_WIDTH(min, max) (((min) < 0) + _GL_COB128 (max)) +#define _GL_COB128(n) (_GL_COB64 ((n) >> 31 >> 31 >> 2) + _GL_COB64 (n)) +#define _GL_COB64(n) (_GL_COB32 ((n) >> 31 >> 1) + _GL_COB32 (n)) +#define _GL_COB32(n) (_GL_COB16 ((n) >> 16) + _GL_COB16 (n)) +#define _GL_COB16(n) (_GL_COB8 ((n) >> 8) + _GL_COB8 (n)) +#define _GL_COB8(n) (_GL_COB4 ((n) >> 4) + _GL_COB4 (n)) +#define _GL_COB4(n) (!!((n) & 8) + !!((n) & 4) + !!((n) & 2) + !!((n) & 1)) + +#ifndef WORD_BIT +/* Assume 'int' is 32 bits wide. */ +# define WORD_BIT 32 +#endif +#ifndef LONG_BIT +/* Assume 'long' is 32 or 64 bits wide. */ +# if LONG_MAX == INT_MAX +# define LONG_BIT 32 +# else +# define LONG_BIT 64 +# endif +#endif + +/* Macros specified by ISO/IEC TS 18661-1:2014. */ + +#if (! defined ULLONG_WIDTH \ + && (defined _GNU_SOURCE || defined __STDC_WANT_IEC_60559_BFP_EXT__)) +# define CHAR_WIDTH _GL_INTEGER_WIDTH (CHAR_MIN, CHAR_MAX) +# define SCHAR_WIDTH _GL_INTEGER_WIDTH (SCHAR_MIN, SCHAR_MAX) +# define UCHAR_WIDTH _GL_INTEGER_WIDTH (0, UCHAR_MAX) +# define SHRT_WIDTH _GL_INTEGER_WIDTH (SHRT_MIN, SHRT_MAX) +# define USHRT_WIDTH _GL_INTEGER_WIDTH (0, USHRT_MAX) +# define INT_WIDTH _GL_INTEGER_WIDTH (INT_MIN, INT_MAX) +# define UINT_WIDTH _GL_INTEGER_WIDTH (0, UINT_MAX) +# define LONG_WIDTH _GL_INTEGER_WIDTH (LONG_MIN, LONG_MAX) +# define ULONG_WIDTH _GL_INTEGER_WIDTH (0, ULONG_MAX) +# define LLONG_WIDTH _GL_INTEGER_WIDTH (LLONG_MIN, LLONG_MAX) +# define ULLONG_WIDTH _GL_INTEGER_WIDTH (0, ULLONG_MAX) +#endif /* !ULLONG_WIDTH && (_GNU_SOURCE || __STDC_WANT_IEC_60559_BFP_EXT__) */ + +#endif /* _@GUARD_PREFIX@_LIMITS_H */ +#endif /* _@GUARD_PREFIX@_LIMITS_H */ diff --git a/lib/local.mk b/lib/local.mk new file mode 100644 index 0000000..1ef14b4 --- /dev/null +++ b/lib/local.mk @@ -0,0 +1,17 @@ +# Copyright 1997-2018 Free Software Foundation, Inc. +# +# This program is free software; you can redistribute it and/or modify +# it under the terms of the GNU General Public License as published by +# the Free Software Foundation; either version 3, or (at your option) +# any later version. +# +# This program is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with this program. If not, see <https://www.gnu.org/licenses/>. +# + +include lib/gnulib.mk diff --git a/lib/localcharset.c b/lib/localcharset.c new file mode 100644 index 0000000..58c5718 --- /dev/null +++ b/lib/localcharset.c @@ -0,0 +1,996 @@ +/* Determine a canonical name for the current locale's character encoding. + + Copyright (C) 2000-2006, 2008-2018 Free Software Foundation, Inc. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3, or (at your option) + any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License along + with this program; if not, see <https://www.gnu.org/licenses/>. */ + +/* Written by Bruno Haible <bruno@clisp.org>. */ + +#include <config.h> + +/* Specification. */ +#include "localcharset.h" + +#include <stddef.h> +#include <stdio.h> +#include <string.h> +#include <stdlib.h> + +#if defined __APPLE__ && defined __MACH__ && HAVE_LANGINFO_CODESET +# define DARWIN7 /* Darwin 7 or newer, i.e. Mac OS X 10.3 or newer */ +#endif + +#if defined _WIN32 && !defined __CYGWIN__ +# define WINDOWS_NATIVE +# include <locale.h> +#endif + +#if defined __EMX__ +/* Assume EMX program runs on OS/2, even if compiled under DOS. */ +# ifndef OS2 +# define OS2 +# endif +#endif + +#if !defined WINDOWS_NATIVE +# if HAVE_LANGINFO_CODESET +# include <langinfo.h> +# else +# if 0 /* see comment regarding use of setlocale(), below */ +# include <locale.h> +# endif +# endif +# ifdef __CYGWIN__ +# define WIN32_LEAN_AND_MEAN +# include <windows.h> +# endif +#elif defined WINDOWS_NATIVE +# define WIN32_LEAN_AND_MEAN +# include <windows.h> +#endif +#if defined OS2 +# define INCL_DOS +# include <os2.h> +#endif + +/* For MB_CUR_MAX_L */ +#if defined DARWIN7 +# include <xlocale.h> +#endif + + +#if HAVE_LANGINFO_CODESET || defined WINDOWS_NATIVE || defined OS2 + +/* On these platforms, we use a mapping from non-canonical encoding name + to GNU canonical encoding name. */ + +/* With glibc-2.1 or newer, we don't need any canonicalization, + because glibc has iconv and both glibc and libiconv support all + GNU canonical names directly. */ +# if !((defined __GNU_LIBRARY__ && __GLIBC__ >= 2) || defined __UCLIBC__) + +struct table_entry +{ + const char alias[11+1]; + const char canonical[11+1]; +}; + +/* Table of platform-dependent mappings, sorted in ascending order. */ +static const struct table_entry alias_table[] = + { +# if defined __FreeBSD__ /* FreeBSD */ + /*{ "ARMSCII-8", "ARMSCII-8" },*/ + { "Big5", "BIG5" }, + { "C", "ASCII" }, + /*{ "CP1131", "CP1131" },*/ + /*{ "CP1251", "CP1251" },*/ + /*{ "CP866", "CP866" },*/ + /*{ "GB18030", "GB18030" },*/ + /*{ "GB2312", "GB2312" },*/ + /*{ "GBK", "GBK" },*/ + /*{ "ISCII-DEV", "?" },*/ + { "ISO8859-1", "ISO-8859-1" }, + { "ISO8859-13", "ISO-8859-13" }, + { "ISO8859-15", "ISO-8859-15" }, + { "ISO8859-2", "ISO-8859-2" }, + { "ISO8859-5", "ISO-8859-5" }, + { "ISO8859-7", "ISO-8859-7" }, + { "ISO8859-9", "ISO-8859-9" }, + /*{ "KOI8-R", "KOI8-R" },*/ + /*{ "KOI8-U", "KOI8-U" },*/ + { "SJIS", "SHIFT_JIS" }, + { "US-ASCII", "ASCII" }, + { "eucCN", "GB2312" }, + { "eucJP", "EUC-JP" }, + { "eucKR", "EUC-KR" } +# define alias_table_defined +# endif +# if defined __NetBSD__ /* NetBSD */ + { "646", "ASCII" }, + /*{ "ARMSCII-8", "ARMSCII-8" },*/ + /*{ "BIG5", "BIG5" },*/ + { "Big5-HKSCS", "BIG5-HKSCS" }, + /*{ "CP1251", "CP1251" },*/ + /*{ "CP866", "CP866" },*/ + /*{ "GB18030", "GB18030" },*/ + /*{ "GB2312", "GB2312" },*/ + { "ISO8859-1", "ISO-8859-1" }, + { "ISO8859-13", "ISO-8859-13" }, + { "ISO8859-15", "ISO-8859-15" }, + { "ISO8859-2", "ISO-8859-2" }, + { "ISO8859-4", "ISO-8859-4" }, + { "ISO8859-5", "ISO-8859-5" }, + { "ISO8859-7", "ISO-8859-7" }, + /*{ "KOI8-R", "KOI8-R" },*/ + /*{ "KOI8-U", "KOI8-U" },*/ + /*{ "PT154", "PT154" },*/ + { "SJIS", "SHIFT_JIS" }, + { "eucCN", "GB2312" }, + { "eucJP", "EUC-JP" }, + { "eucKR", "EUC-KR" }, + { "eucTW", "EUC-TW" } +# define alias_table_defined +# endif +# if defined __OpenBSD__ /* OpenBSD */ + { "646", "ASCII" }, + { "ISO8859-1", "ISO-8859-1" }, + { "ISO8859-13", "ISO-8859-13" }, + { "ISO8859-15", "ISO-8859-15" }, + { "ISO8859-2", "ISO-8859-2" }, + { "ISO8859-4", "ISO-8859-4" }, + { "ISO8859-5", "ISO-8859-5" }, + { "ISO8859-7", "ISO-8859-7" } +# define alias_table_defined +# endif +# if defined __APPLE__ && defined __MACH__ /* Mac OS X */ + /* Darwin 7.5 has nl_langinfo(CODESET), but sometimes its value is + useless: + - It returns the empty string when LANG is set to a locale of the + form ll_CC, although ll_CC/LC_CTYPE is a symlink to an UTF-8 + LC_CTYPE file. + - The environment variables LANG, LC_CTYPE, LC_ALL are not set by + the system; nl_langinfo(CODESET) returns "US-ASCII" in this case. + - The documentation says: + "... all code that calls BSD system routines should ensure + that the const *char parameters of these routines are in UTF-8 + encoding. All BSD system functions expect their string + parameters to be in UTF-8 encoding and nothing else." + It also says + "An additional caveat is that string parameters for files, + paths, and other file-system entities must be in canonical + UTF-8. In a canonical UTF-8 Unicode string, all decomposable + characters are decomposed ..." + but this is not true: You can pass non-decomposed UTF-8 strings + to file system functions, and it is the OS which will convert + them to decomposed UTF-8 before accessing the file system. + - The Apple Terminal application displays UTF-8 by default. + - However, other applications are free to use different encodings: + - xterm uses ISO-8859-1 by default. + - TextEdit uses MacRoman by default. + We prefer UTF-8 over decomposed UTF-8-MAC because one should + minimize the use of decomposed Unicode. Unfortunately, through the + Darwin file system, decomposed UTF-8 strings are leaked into user + space nevertheless. + Then there are also the locales with encodings other than US-ASCII + and UTF-8. These locales can be occasionally useful to users (e.g. + when grepping through ISO-8859-1 encoded text files), when all their + file names are in US-ASCII. + */ + { "ARMSCII-8", "ARMSCII-8" }, + { "Big5", "BIG5" }, + { "Big5HKSCS", "BIG5-HKSCS" }, + { "CP1131", "CP1131" }, + { "CP1251", "CP1251" }, + { "CP866", "CP866" }, + { "CP949", "CP949" }, + { "GB18030", "GB18030" }, + { "GB2312", "GB2312" }, + { "GBK", "GBK" }, + /*{ "ISCII-DEV", "?" },*/ + { "ISO8859-1", "ISO-8859-1" }, + { "ISO8859-13", "ISO-8859-13" }, + { "ISO8859-15", "ISO-8859-15" }, + { "ISO8859-2", "ISO-8859-2" }, + { "ISO8859-4", "ISO-8859-4" }, + { "ISO8859-5", "ISO-8859-5" }, + { "ISO8859-7", "ISO-8859-7" }, + { "ISO8859-9", "ISO-8859-9" }, + { "KOI8-R", "KOI8-R" }, + { "KOI8-U", "KOI8-U" }, + { "PT154", "PT154" }, + { "SJIS", "SHIFT_JIS" }, + { "eucCN", "GB2312" }, + { "eucJP", "EUC-JP" }, + { "eucKR", "EUC-KR" } +# define alias_table_defined +# endif +# if defined _AIX /* AIX */ + /*{ "GBK", "GBK" },*/ + { "IBM-1046", "CP1046" }, + { "IBM-1124", "CP1124" }, + { "IBM-1129", "CP1129" }, + { "IBM-1252", "CP1252" }, + { "IBM-850", "CP850" }, + { "IBM-856", "CP856" }, + { "IBM-921", "ISO-8859-13" }, + { "IBM-922", "CP922" }, + { "IBM-932", "CP932" }, + { "IBM-943", "CP943" }, + { "IBM-eucCN", "GB2312" }, + { "IBM-eucJP", "EUC-JP" }, + { "IBM-eucKR", "EUC-KR" }, + { "IBM-eucTW", "EUC-TW" }, + { "ISO8859-1", "ISO-8859-1" }, + { "ISO8859-15", "ISO-8859-15" }, + { "ISO8859-2", "ISO-8859-2" }, + { "ISO8859-5", "ISO-8859-5" }, + { "ISO8859-6", "ISO-8859-6" }, + { "ISO8859-7", "ISO-8859-7" }, + { "ISO8859-8", "ISO-8859-8" }, + { "ISO8859-9", "ISO-8859-9" }, + { "TIS-620", "TIS-620" }, + /*{ "UTF-8", "UTF-8" },*/ + { "big5", "BIG5" } +# define alias_table_defined +# endif +# if defined __hpux /* HP-UX */ + { "SJIS", "SHIFT_JIS" }, + { "arabic8", "HP-ARABIC8" }, + { "big5", "BIG5" }, + { "cp1251", "CP1251" }, + { "eucJP", "EUC-JP" }, + { "eucKR", "EUC-KR" }, + { "eucTW", "EUC-TW" }, + { "gb18030", "GB18030" }, + { "greek8", "HP-GREEK8" }, + { "hebrew8", "HP-HEBREW8" }, + { "hkbig5", "BIG5-HKSCS" }, + { "hp15CN", "GB2312" }, + { "iso88591", "ISO-8859-1" }, + { "iso885913", "ISO-8859-13" }, + { "iso885915", "ISO-8859-15" }, + { "iso88592", "ISO-8859-2" }, + { "iso88594", "ISO-8859-4" }, + { "iso88595", "ISO-8859-5" }, + { "iso88596", "ISO-8859-6" }, + { "iso88597", "ISO-8859-7" }, + { "iso88598", "ISO-8859-8" }, + { "iso88599", "ISO-8859-9" }, + { "kana8", "HP-KANA8" }, + { "koi8r", "KOI8-R" }, + { "roman8", "HP-ROMAN8" }, + { "tis620", "TIS-620" }, + { "turkish8", "HP-TURKISH8" }, + { "utf8", "UTF-8" } +# define alias_table_defined +# endif +# if defined __sgi /* IRIX */ + { "ISO8859-1", "ISO-8859-1" }, + { "ISO8859-15", "ISO-8859-15" }, + { "ISO8859-2", "ISO-8859-2" }, + { "ISO8859-5", "ISO-8859-5" }, + { "ISO8859-7", "ISO-8859-7" }, + { "ISO8859-9", "ISO-8859-9" }, + { "eucCN", "GB2312" }, + { "eucJP", "EUC-JP" }, + { "eucKR", "EUC-KR" }, + { "eucTW", "EUC-TW" } +# define alias_table_defined +# endif +# if defined __osf__ /* OSF/1 */ + /*{ "GBK", "GBK" },*/ + { "ISO8859-1", "ISO-8859-1" }, + { "ISO8859-15", "ISO-8859-15" }, + { "ISO8859-2", "ISO-8859-2" }, + { "ISO8859-4", "ISO-8859-4" }, + { "ISO8859-5", "ISO-8859-5" }, + { "ISO8859-7", "ISO-8859-7" }, + { "ISO8859-8", "ISO-8859-8" }, + { "ISO8859-9", "ISO-8859-9" }, + { "KSC5601", "CP949" }, + { "SJIS", "SHIFT_JIS" }, + { "TACTIS", "TIS-620" }, + /*{ "UTF-8", "UTF-8" },*/ + { "big5", "BIG5" }, + { "cp850", "CP850" }, + { "dechanyu", "DEC-HANYU" }, + { "dechanzi", "GB2312" }, + { "deckanji", "DEC-KANJI" }, + { "deckorean", "EUC-KR" }, + { "eucJP", "EUC-JP" }, + { "eucKR", "EUC-KR" }, + { "eucTW", "EUC-TW" }, + { "sdeckanji", "EUC-JP" } +# define alias_table_defined +# endif +# if defined __sun /* Solaris */ + { "5601", "EUC-KR" }, + { "646", "ASCII" }, + /*{ "BIG5", "BIG5" },*/ + { "Big5-HKSCS", "BIG5-HKSCS" }, + { "GB18030", "GB18030" }, + /*{ "GBK", "GBK" },*/ + { "ISO8859-1", "ISO-8859-1" }, + { "ISO8859-11", "TIS-620" }, + { "ISO8859-13", "ISO-8859-13" }, + { "ISO8859-15", "ISO-8859-15" }, + { "ISO8859-2", "ISO-8859-2" }, + { "ISO8859-3", "ISO-8859-3" }, + { "ISO8859-4", "ISO-8859-4" }, + { "ISO8859-5", "ISO-8859-5" }, + { "ISO8859-6", "ISO-8859-6" }, + { "ISO8859-7", "ISO-8859-7" }, + { "ISO8859-8", "ISO-8859-8" }, + { "ISO8859-9", "ISO-8859-9" }, + { "PCK", "SHIFT_JIS" }, + { "TIS620.2533", "TIS-620" }, + /*{ "UTF-8", "UTF-8" },*/ + { "ansi-1251", "CP1251" }, + { "cns11643", "EUC-TW" }, + { "eucJP", "EUC-JP" }, + { "gb2312", "GB2312" }, + { "koi8-r", "KOI8-R" } +# define alias_table_defined +# endif +# if defined __minix /* Minix */ + { "646", "ASCII" } +# define alias_table_defined +# endif +# if defined WINDOWS_NATIVE || defined __CYGWIN__ /* Windows */ + { "CP1361", "JOHAB" }, + { "CP20127", "ASCII" }, + { "CP20866", "KOI8-R" }, + { "CP20936", "GB2312" }, + { "CP21866", "KOI8-RU" }, + { "CP28591", "ISO-8859-1" }, + { "CP28592", "ISO-8859-2" }, + { "CP28593", "ISO-8859-3" }, + { "CP28594", "ISO-8859-4" }, + { "CP28595", "ISO-8859-5" }, + { "CP28596", "ISO-8859-6" }, + { "CP28597", "ISO-8859-7" }, + { "CP28598", "ISO-8859-8" }, + { "CP28599", "ISO-8859-9" }, + { "CP28605", "ISO-8859-15" }, + { "CP38598", "ISO-8859-8" }, + { "CP51932", "EUC-JP" }, + { "CP51936", "GB2312" }, + { "CP51949", "EUC-KR" }, + { "CP51950", "EUC-TW" }, + { "CP54936", "GB18030" }, + { "CP65001", "UTF-8" }, + { "CP936", "GBK" } +# define alias_table_defined +# endif +# if defined OS2 /* OS/2 */ + /* The list of encodings is taken from "List of OS/2 Codepages" + by Alex Taylor: + <http://altsan.org/os2/toolkits/uls/index.html#codepages>. + See also "IBM Globalization - Code page identifiers": + <https://www-01.ibm.com/software/globalization/cp/cp_cpgid.html>. */ + { "CP1089", "ISO-8859-6" }, + { "CP1208", "UTF-8" }, + { "CP1381", "GB2312" }, + { "CP1386", "GBK" }, + { "CP3372", "EUC-JP" }, + { "CP813", "ISO-8859-7" }, + { "CP819", "ISO-8859-1" }, + { "CP878", "KOI8-R" }, + { "CP912", "ISO-8859-2" }, + { "CP913", "ISO-8859-3" }, + { "CP914", "ISO-8859-4" }, + { "CP915", "ISO-8859-5" }, + { "CP916", "ISO-8859-8" }, + { "CP920", "ISO-8859-9" }, + { "CP921", "ISO-8859-13" }, + { "CP923", "ISO-8859-15" }, + { "CP954", "EUC-JP" }, + { "CP964", "EUC-TW" }, + { "CP970", "EUC-KR" } +# define alias_table_defined +# endif +# if defined VMS /* OpenVMS */ + /* The list of encodings is taken from the OpenVMS 7.3-1 documentation + "Compaq C Run-Time Library Reference Manual for OpenVMS systems" + section 10.7 "Handling Different Character Sets". */ + { "DECHANYU", "DEC-HANYU" }, + { "DECHANZI", "GB2312" }, + { "DECKANJI", "DEC-KANJI" }, + { "DECKOREAN", "EUC-KR" }, + { "ISO8859-1", "ISO-8859-1" }, + { "ISO8859-2", "ISO-8859-2" }, + { "ISO8859-5", "ISO-8859-5" }, + { "ISO8859-7", "ISO-8859-7" }, + { "ISO8859-8", "ISO-8859-8" }, + { "ISO8859-9", "ISO-8859-9" }, + { "SDECKANJI", "EUC-JP" }, + { "SJIS", "SHIFT_JIS" }, + { "eucJP", "EUC-JP" }, + { "eucTW", "EUC-TW" } +# define alias_table_defined +# endif +# ifndef alias_table_defined + /* Just a dummy entry, to avoid a C syntax error. */ + { "", "" } +# endif + }; + +# endif + +#else + +/* On these platforms, we use a mapping from locale name to GNU canonical + encoding name. */ + +struct table_entry +{ + const char locale[17+1]; + const char canonical[11+1]; +}; + +/* Table of platform-dependent mappings, sorted in ascending order. */ +static const struct table_entry locale_table[] = + { +# if defined __FreeBSD__ /* FreeBSD 4.2 */ + { "cs_CZ.ISO_8859-2", "ISO-8859-2" }, + { "da_DK.DIS_8859-15", "ISO-8859-15" }, + { "da_DK.ISO_8859-1", "ISO-8859-1" }, + { "de_AT.DIS_8859-15", "ISO-8859-15" }, + { "de_AT.ISO_8859-1", "ISO-8859-1" }, + { "de_CH.DIS_8859-15", "ISO-8859-15" }, + { "de_CH.ISO_8859-1", "ISO-8859-1" }, + { "de_DE.DIS_8859-15", "ISO-8859-15" }, + { "de_DE.ISO_8859-1", "ISO-8859-1" }, + { "en_AU.DIS_8859-15", "ISO-8859-15" }, + { "en_AU.ISO_8859-1", "ISO-8859-1" }, + { "en_CA.DIS_8859-15", "ISO-8859-15" }, + { "en_CA.ISO_8859-1", "ISO-8859-1" }, + { "en_GB.DIS_8859-15", "ISO-8859-15" }, + { "en_GB.ISO_8859-1", "ISO-8859-1" }, + { "en_US.DIS_8859-15", "ISO-8859-15" }, + { "en_US.ISO_8859-1", "ISO-8859-1" }, + { "es_ES.DIS_8859-15", "ISO-8859-15" }, + { "es_ES.ISO_8859-1", "ISO-8859-1" }, + { "fi_FI.DIS_8859-15", "ISO-8859-15" }, + { "fi_FI.ISO_8859-1", "ISO-8859-1" }, + { "fr_BE.DIS_8859-15", "ISO-8859-15" }, + { "fr_BE.ISO_8859-1", "ISO-8859-1" }, + { "fr_CA.DIS_8859-15", "ISO-8859-15" }, + { "fr_CA.ISO_8859-1", "ISO-8859-1" }, + { "fr_CH.DIS_8859-15", "ISO-8859-15" }, + { "fr_CH.ISO_8859-1", "ISO-8859-1" }, + { "fr_FR.DIS_8859-15", "ISO-8859-15" }, + { "fr_FR.ISO_8859-1", "ISO-8859-1" }, + { "hr_HR.ISO_8859-2", "ISO-8859-2" }, + { "hu_HU.ISO_8859-2", "ISO-8859-2" }, + { "is_IS.DIS_8859-15", "ISO-8859-15" }, + { "is_IS.ISO_8859-1", "ISO-8859-1" }, + { "it_CH.DIS_8859-15", "ISO-8859-15" }, + { "it_CH.ISO_8859-1", "ISO-8859-1" }, + { "it_IT.DIS_8859-15", "ISO-8859-15" }, + { "it_IT.ISO_8859-1", "ISO-8859-1" }, + { "ja_JP.EUC", "EUC-JP" }, + { "ja_JP.SJIS", "SHIFT_JIS" }, + { "ja_JP.Shift_JIS", "SHIFT_JIS" }, + { "ko_KR.EUC", "EUC-KR" }, + { "la_LN.ASCII", "ASCII" }, + { "la_LN.DIS_8859-15", "ISO-8859-15" }, + { "la_LN.ISO_8859-1", "ISO-8859-1" }, + { "la_LN.ISO_8859-2", "ISO-8859-2" }, + { "la_LN.ISO_8859-4", "ISO-8859-4" }, + { "lt_LN.ASCII", "ASCII" }, + { "lt_LN.DIS_8859-15", "ISO-8859-15" }, + { "lt_LN.ISO_8859-1", "ISO-8859-1" }, + { "lt_LN.ISO_8859-2", "ISO-8859-2" }, + { "lt_LT.ISO_8859-4", "ISO-8859-4" }, + { "nl_BE.DIS_8859-15", "ISO-8859-15" }, + { "nl_BE.ISO_8859-1", "ISO-8859-1" }, + { "nl_NL.DIS_8859-15", "ISO-8859-15" }, + { "nl_NL.ISO_8859-1", "ISO-8859-1" }, + { "no_NO.DIS_8859-15", "ISO-8859-15" }, + { "no_NO.ISO_8859-1", "ISO-8859-1" }, + { "pl_PL.ISO_8859-2", "ISO-8859-2" }, + { "pt_PT.DIS_8859-15", "ISO-8859-15" }, + { "pt_PT.ISO_8859-1", "ISO-8859-1" }, + { "ru_RU.CP866", "CP866" }, + { "ru_RU.ISO_8859-5", "ISO-8859-5" }, + { "ru_RU.KOI8-R", "KOI8-R" }, + { "ru_SU.CP866", "CP866" }, + { "ru_SU.ISO_8859-5", "ISO-8859-5" }, + { "ru_SU.KOI8-R", "KOI8-R" }, + { "sl_SI.ISO_8859-2", "ISO-8859-2" }, + { "sv_SE.DIS_8859-15", "ISO-8859-15" }, + { "sv_SE.ISO_8859-1", "ISO-8859-1" }, + { "uk_UA.KOI8-U", "KOI8-U" }, + { "zh_CN.EUC", "GB2312" }, + { "zh_TW.BIG5", "BIG5" }, + { "zh_TW.Big5", "BIG5" } +# define locale_table_defined +# endif +# if defined __DJGPP__ /* DOS / DJGPP 2.03 */ + /* The encodings given here may not all be correct. + If you find that the encoding given for your language and + country is not the one your DOS machine actually uses, just + correct it in this file, and send a mail to + Juan Manuel Guerrero <juan.guerrero@gmx.de> + and <bug-gnulib@gnu.org>. */ + { "C", "ASCII" }, + { "ar", "CP864" }, + { "ar_AE", "CP864" }, + { "ar_DZ", "CP864" }, + { "ar_EG", "CP864" }, + { "ar_IQ", "CP864" }, + { "ar_IR", "CP864" }, + { "ar_JO", "CP864" }, + { "ar_KW", "CP864" }, + { "ar_MA", "CP864" }, + { "ar_OM", "CP864" }, + { "ar_QA", "CP864" }, + { "ar_SA", "CP864" }, + { "ar_SY", "CP864" }, + { "be", "CP866" }, + { "be_BE", "CP866" }, + { "bg", "CP866" }, /* not CP855 ?? */ + { "bg_BG", "CP866" }, /* not CP855 ?? */ + { "ca", "CP850" }, + { "ca_ES", "CP850" }, + { "cs", "CP852" }, + { "cs_CZ", "CP852" }, + { "da", "CP865" }, /* not CP850 ?? */ + { "da_DK", "CP865" }, /* not CP850 ?? */ + { "de", "CP850" }, + { "de_AT", "CP850" }, + { "de_CH", "CP850" }, + { "de_DE", "CP850" }, + { "el", "CP869" }, + { "el_GR", "CP869" }, + { "en", "CP850" }, + { "en_AU", "CP850" }, /* not CP437 ?? */ + { "en_CA", "CP850" }, + { "en_GB", "CP850" }, + { "en_NZ", "CP437" }, + { "en_US", "CP437" }, + { "en_ZA", "CP850" }, /* not CP437 ?? */ + { "eo", "CP850" }, + { "eo_EO", "CP850" }, + { "es", "CP850" }, + { "es_AR", "CP850" }, + { "es_BO", "CP850" }, + { "es_CL", "CP850" }, + { "es_CO", "CP850" }, + { "es_CR", "CP850" }, + { "es_CU", "CP850" }, + { "es_DO", "CP850" }, + { "es_EC", "CP850" }, + { "es_ES", "CP850" }, + { "es_GT", "CP850" }, + { "es_HN", "CP850" }, + { "es_MX", "CP850" }, + { "es_NI", "CP850" }, + { "es_PA", "CP850" }, + { "es_PE", "CP850" }, + { "es_PY", "CP850" }, + { "es_SV", "CP850" }, + { "es_UY", "CP850" }, + { "es_VE", "CP850" }, + { "et", "CP850" }, + { "et_EE", "CP850" }, + { "eu", "CP850" }, + { "eu_ES", "CP850" }, + { "fi", "CP850" }, + { "fi_FI", "CP850" }, + { "fr", "CP850" }, + { "fr_BE", "CP850" }, + { "fr_CA", "CP850" }, + { "fr_CH", "CP850" }, + { "fr_FR", "CP850" }, + { "ga", "CP850" }, + { "ga_IE", "CP850" }, + { "gd", "CP850" }, + { "gd_GB", "CP850" }, + { "gl", "CP850" }, + { "gl_ES", "CP850" }, + { "he", "CP862" }, + { "he_IL", "CP862" }, + { "hr", "CP852" }, + { "hr_HR", "CP852" }, + { "hu", "CP852" }, + { "hu_HU", "CP852" }, + { "id", "CP850" }, /* not CP437 ?? */ + { "id_ID", "CP850" }, /* not CP437 ?? */ + { "is", "CP861" }, /* not CP850 ?? */ + { "is_IS", "CP861" }, /* not CP850 ?? */ + { "it", "CP850" }, + { "it_CH", "CP850" }, + { "it_IT", "CP850" }, + { "ja", "CP932" }, + { "ja_JP", "CP932" }, + { "kr", "CP949" }, /* not CP934 ?? */ + { "kr_KR", "CP949" }, /* not CP934 ?? */ + { "lt", "CP775" }, + { "lt_LT", "CP775" }, + { "lv", "CP775" }, + { "lv_LV", "CP775" }, + { "mk", "CP866" }, /* not CP855 ?? */ + { "mk_MK", "CP866" }, /* not CP855 ?? */ + { "mt", "CP850" }, + { "mt_MT", "CP850" }, + { "nb", "CP865" }, /* not CP850 ?? */ + { "nb_NO", "CP865" }, /* not CP850 ?? */ + { "nl", "CP850" }, + { "nl_BE", "CP850" }, + { "nl_NL", "CP850" }, + { "nn", "CP865" }, /* not CP850 ?? */ + { "nn_NO", "CP865" }, /* not CP850 ?? */ + { "no", "CP865" }, /* not CP850 ?? */ + { "no_NO", "CP865" }, /* not CP850 ?? */ + { "pl", "CP852" }, + { "pl_PL", "CP852" }, + { "pt", "CP850" }, + { "pt_BR", "CP850" }, + { "pt_PT", "CP850" }, + { "ro", "CP852" }, + { "ro_RO", "CP852" }, + { "ru", "CP866" }, + { "ru_RU", "CP866" }, + { "sk", "CP852" }, + { "sk_SK", "CP852" }, + { "sl", "CP852" }, + { "sl_SI", "CP852" }, + { "sq", "CP852" }, + { "sq_AL", "CP852" }, + { "sr", "CP852" }, /* CP852 or CP866 or CP855 ?? */ + { "sr_CS", "CP852" }, /* CP852 or CP866 or CP855 ?? */ + { "sr_YU", "CP852" }, /* CP852 or CP866 or CP855 ?? */ + { "sv", "CP850" }, + { "sv_SE", "CP850" }, + { "th", "CP874" }, + { "th_TH", "CP874" }, + { "tr", "CP857" }, + { "tr_TR", "CP857" }, + { "uk", "CP1125" }, + { "uk_UA", "CP1125" }, + { "zh_CN", "GBK" }, + { "zh_TW", "CP950" } /* not CP938 ?? */ +# define locale_table_defined +# endif +# ifndef locale_table_defined + /* Just a dummy entry, to avoid a C syntax error. */ + { "", "" } +# endif + }; + +#endif + + +/* Determine the current locale's character encoding, and canonicalize it + into one of the canonical names listed in localcharset.h. + The result must not be freed; it is statically allocated. + If the canonical name cannot be determined, the result is a non-canonical + name. */ + +#ifdef STATIC +STATIC +#endif +const char * +locale_charset (void) +{ + const char *codeset; + +#if HAVE_LANGINFO_CODESET || defined WINDOWS_NATIVE || defined OS2 + +# if HAVE_LANGINFO_CODESET + + /* Most systems support nl_langinfo (CODESET) nowadays. */ + codeset = nl_langinfo (CODESET); + +# ifdef __CYGWIN__ + /* Cygwin < 1.7 does not have locales. nl_langinfo (CODESET) always + returns "US-ASCII". Return the suffix of the locale name from the + environment variables (if present) or the codepage as a number. */ + if (codeset != NULL && strcmp (codeset, "US-ASCII") == 0) + { + const char *locale; + static char buf[2 + 10 + 1]; + + locale = getenv ("LC_ALL"); + if (locale == NULL || locale[0] == '\0') + { + locale = getenv ("LC_CTYPE"); + if (locale == NULL || locale[0] == '\0') + locale = getenv ("LANG"); + } + if (locale != NULL && locale[0] != '\0') + { + /* If the locale name contains an encoding after the dot, return + it. */ + const char *dot = strchr (locale, '.'); + + if (dot != NULL) + { + const char *modifier; + + dot++; + /* Look for the possible @... trailer and remove it, if any. */ + modifier = strchr (dot, '@'); + if (modifier == NULL) + return dot; + if (modifier - dot < sizeof (buf)) + { + memcpy (buf, dot, modifier - dot); + buf [modifier - dot] = '\0'; + return buf; + } + } + } + + /* The Windows API has a function returning the locale's codepage as a + number: GetACP(). This encoding is used by Cygwin, unless the user + has set the environment variable CYGWIN=codepage:oem (which very few + people do). + Output directed to console windows needs to be converted (to + GetOEMCP() if the console is using a raster font, or to + GetConsoleOutputCP() if it is using a TrueType font). Cygwin does + this conversion transparently (see winsup/cygwin/fhandler_console.cc), + converting to GetConsoleOutputCP(). This leads to correct results, + except when SetConsoleOutputCP has been called and a raster font is + in use. */ + sprintf (buf, "CP%u", GetACP ()); + codeset = buf; + } +# endif + + if (codeset == NULL) + /* The canonical name cannot be determined. */ + codeset = ""; + +# elif defined WINDOWS_NATIVE + + static char buf[2 + 10 + 1]; + + /* The Windows API has a function returning the locale's codepage as + a number, but the value doesn't change according to what the + 'setlocale' call specified. So we use it as a last resort, in + case the string returned by 'setlocale' doesn't specify the + codepage. */ + char *current_locale = setlocale (LC_ALL, NULL); + char *pdot; + + /* If they set different locales for different categories, + 'setlocale' will return a semi-colon separated list of locale + values. To make sure we use the correct one, we choose LC_CTYPE. */ + if (strchr (current_locale, ';')) + current_locale = setlocale (LC_CTYPE, NULL); + + pdot = strrchr (current_locale, '.'); + if (pdot && 2 + strlen (pdot + 1) + 1 <= sizeof (buf)) + sprintf (buf, "CP%s", pdot + 1); + else + { + /* The Windows API has a function returning the locale's codepage as a + number: GetACP(). + When the output goes to a console window, it needs to be provided in + GetOEMCP() encoding if the console is using a raster font, or in + GetConsoleOutputCP() encoding if it is using a TrueType font. + But in GUI programs and for output sent to files and pipes, GetACP() + encoding is the best bet. */ + sprintf (buf, "CP%u", GetACP ()); + } + codeset = buf; + +# elif defined OS2 + + const char *locale; + static char buf[2 + 10 + 1]; + ULONG cp[3]; + ULONG cplen; + + codeset = NULL; + + /* Allow user to override the codeset, as set in the operating system, + with standard language environment variables. */ + locale = getenv ("LC_ALL"); + if (locale == NULL || locale[0] == '\0') + { + locale = getenv ("LC_CTYPE"); + if (locale == NULL || locale[0] == '\0') + locale = getenv ("LANG"); + } + if (locale != NULL && locale[0] != '\0') + { + /* If the locale name contains an encoding after the dot, return it. */ + const char *dot = strchr (locale, '.'); + + if (dot != NULL) + { + const char *modifier; + + dot++; + /* Look for the possible @... trailer and remove it, if any. */ + modifier = strchr (dot, '@'); + if (modifier == NULL) + return dot; + if (modifier - dot < sizeof (buf)) + { + memcpy (buf, dot, modifier - dot); + buf [modifier - dot] = '\0'; + return buf; + } + } + + /* For the POSIX locale, don't use the system's codepage. */ + if (strcmp (locale, "C") == 0 || strcmp (locale, "POSIX") == 0) + codeset = ""; + } + + if (codeset == NULL) + { + /* OS/2 has a function returning the locale's codepage as a number. */ + if (DosQueryCp (sizeof (cp), cp, &cplen)) + codeset = ""; + else + { + sprintf (buf, "CP%u", cp[0]); + codeset = buf; + } + } + +# else + +# error "Add code for other platforms here." + +# endif + + /* Resolve alias. */ + { +# ifdef alias_table_defined + /* On some platforms, UTF-8 locales are the most frequently used ones. + Speed up the common case and slow down the less common cases by + testing for this case first. */ +# if defined __OpenBSD__ || (defined __APPLE__ && defined __MACH__) || defined __sun || defined __CYGWIN__ + if (strcmp (codeset, "UTF-8") == 0) + goto done_table_lookup; + else +# endif + { + const struct table_entry * const table = alias_table; + size_t const table_size = + sizeof (alias_table) / sizeof (struct table_entry); + /* The table is sorted. Perform a binary search. */ + size_t hi = table_size; + size_t lo = 0; + while (lo < hi) + { + /* Invariant: + for i < lo, strcmp (table[i].alias, codeset) < 0, + for i >= hi, strcmp (table[i].alias, codeset) > 0. */ + size_t mid = (hi + lo) >> 1; /* >= lo, < hi */ + int cmp = strcmp (table[mid].alias, codeset); + if (cmp < 0) + lo = mid + 1; + else if (cmp > 0) + hi = mid; + else + { + /* Found an i with + strcmp (table[i].alias, codeset) == 0. */ + codeset = table[mid].canonical; + goto done_table_lookup; + } + } + } + if (0) + done_table_lookup: ; + else +# endif + { + /* Did not find it in the table. */ + /* On Mac OS X, all modern locales use the UTF-8 encoding. + BeOS and Haiku have a single locale, and it has UTF-8 encoding. */ +# if (defined __APPLE__ && defined __MACH__) || defined __BEOS__ || defined __HAIKU__ + codeset = "UTF-8"; +# else + /* Don't return an empty string. GNU libc and GNU libiconv interpret + the empty string as denoting "the locale's character encoding", + thus GNU libiconv would call this function a second time. */ + if (codeset[0] == '\0') + codeset = "ASCII"; +# endif + } + } + +#else + + /* On old systems which lack it, use setlocale or getenv. */ + const char *locale = NULL; + + /* But most old systems don't have a complete set of locales. Some + (like DJGPP) have only the C locale. Therefore we don't use setlocale + here; it would return "C" when it doesn't support the locale name the + user has set. */ +# if 0 + locale = setlocale (LC_CTYPE, NULL); +# endif + if (locale == NULL || locale[0] == '\0') + { + locale = getenv ("LC_ALL"); + if (locale == NULL || locale[0] == '\0') + { + locale = getenv ("LC_CTYPE"); + if (locale == NULL || locale[0] == '\0') + locale = getenv ("LANG"); + if (locale == NULL) + locale = ""; + } + } + + /* Map locale name to canonical encoding name. */ + { +# ifdef locale_table_defined + const struct table_entry * const table = locale_table; + size_t const table_size = + sizeof (locale_table) / sizeof (struct table_entry); + /* The table is sorted. Perform a binary search. */ + size_t hi = table_size; + size_t lo = 0; + while (lo < hi) + { + /* Invariant: + for i < lo, strcmp (table[i].locale, locale) < 0, + for i >= hi, strcmp (table[i].locale, locale) > 0. */ + size_t mid = (hi + lo) >> 1; /* >= lo, < hi */ + int cmp = strcmp (table[mid].locale, locale); + if (cmp < 0) + lo = mid + 1; + else if (cmp > 0) + hi = mid; + else + { + /* Found an i with + strcmp (table[i].locale, locale) == 0. */ + codeset = table[mid].canonical; + goto done_table_lookup; + } + } + if (0) + done_table_lookup: ; + else +# endif + { + /* Did not find it in the table. */ + /* On Mac OS X, all modern locales use the UTF-8 encoding. + BeOS and Haiku have a single locale, and it has UTF-8 encoding. */ +# if (defined __APPLE__ && defined __MACH__) || defined __BEOS__ || defined __HAIKU__ + codeset = "UTF-8"; +# else + /* The canonical name cannot be determined. */ + /* Don't return an empty string. GNU libc and GNU libiconv interpret + the empty string as denoting "the locale's character encoding", + thus GNU libiconv would call this function a second time. */ + codeset = "ASCII"; +# endif + } + } + +#endif + +#ifdef DARWIN7 + /* Mac OS X sets MB_CUR_MAX to 1 when LC_ALL=C, and "UTF-8" + (the default codeset) does not work when MB_CUR_MAX is 1. */ + if (strcmp (codeset, "UTF-8") == 0 && MB_CUR_MAX_L (uselocale (NULL)) <= 1) + codeset = "ASCII"; +#endif + + return codeset; +} diff --git a/lib/localcharset.h b/lib/localcharset.h new file mode 100644 index 0000000..e4ba296 --- /dev/null +++ b/lib/localcharset.h @@ -0,0 +1,134 @@ +/* Determine a canonical name for the current locale's character encoding. + Copyright (C) 2000-2003, 2009-2018 Free Software Foundation, Inc. + This file is part of the GNU CHARSET Library. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3, or (at your option) + any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License along + with this program; if not, see <https://www.gnu.org/licenses/>. */ + +#ifndef _LOCALCHARSET_H +#define _LOCALCHARSET_H + + +#ifdef __cplusplus +extern "C" { +#endif + + +/* Determine the current locale's character encoding, and canonicalize it + into one of the canonical names listed below. + The result must not be freed; it is statically allocated. + If the canonical name cannot be determined, the result is a non-canonical + name. */ +extern const char * locale_charset (void); + +/* About GNU canonical names for character encodings: + + Every canonical name must be supported by GNU libiconv. Support by GNU libc + is also desirable. + + The name is case insensitive. Usually an upper case MIME charset name is + preferred. + + The current list of these GNU canonical names is: + + name MIME? used by which systems + (darwin = Mac OS X, windows = native Windows) + + ASCII, ANSI_X3.4-1968 glibc solaris freebsd netbsd darwin minix cygwin + ISO-8859-1 Y glibc aix hpux irix osf solaris freebsd netbsd openbsd darwin cygwin + ISO-8859-2 Y glibc aix hpux irix osf solaris freebsd netbsd openbsd darwin cygwin + ISO-8859-3 Y glibc solaris cygwin + ISO-8859-4 Y hpux osf solaris freebsd netbsd openbsd darwin + ISO-8859-5 Y glibc aix hpux irix osf solaris freebsd netbsd openbsd darwin cygwin + ISO-8859-6 Y glibc aix hpux solaris cygwin + ISO-8859-7 Y glibc aix hpux irix osf solaris freebsd netbsd openbsd darwin cygwin + ISO-8859-8 Y glibc aix hpux osf solaris cygwin + ISO-8859-9 Y glibc aix hpux irix osf solaris freebsd darwin cygwin + ISO-8859-13 glibc hpux solaris freebsd netbsd openbsd darwin cygwin + ISO-8859-14 glibc cygwin + ISO-8859-15 glibc aix irix osf solaris freebsd netbsd openbsd darwin cygwin + KOI8-R Y glibc hpux solaris freebsd netbsd openbsd darwin + KOI8-U Y glibc freebsd netbsd openbsd darwin cygwin + KOI8-T glibc + CP437 dos + CP775 dos + CP850 aix osf dos + CP852 dos + CP855 dos + CP856 aix + CP857 dos + CP861 dos + CP862 dos + CP864 dos + CP865 dos + CP866 freebsd netbsd openbsd darwin dos + CP869 dos + CP874 windows dos + CP922 aix + CP932 aix cygwin windows dos + CP943 aix + CP949 osf darwin windows dos + CP950 windows dos + CP1046 aix + CP1124 aix + CP1125 dos + CP1129 aix + CP1131 freebsd darwin + CP1250 windows + CP1251 glibc hpux solaris freebsd netbsd openbsd darwin cygwin windows + CP1252 aix windows + CP1253 windows + CP1254 windows + CP1255 glibc windows + CP1256 windows + CP1257 windows + GB2312 Y glibc aix hpux irix solaris freebsd netbsd darwin cygwin + EUC-JP Y glibc aix hpux irix osf solaris freebsd netbsd darwin cygwin + EUC-KR Y glibc aix hpux irix osf solaris freebsd netbsd darwin cygwin + EUC-TW glibc aix hpux irix osf solaris netbsd + BIG5 Y glibc aix hpux osf solaris freebsd netbsd darwin cygwin + BIG5-HKSCS glibc hpux solaris netbsd darwin + GBK glibc aix osf solaris freebsd darwin cygwin windows dos + GB18030 glibc hpux solaris freebsd netbsd darwin + SHIFT_JIS Y hpux osf solaris freebsd netbsd darwin + JOHAB glibc solaris windows + TIS-620 glibc aix hpux osf solaris cygwin + VISCII Y glibc + TCVN5712-1 glibc + ARMSCII-8 glibc freebsd netbsd darwin + GEORGIAN-PS glibc cygwin + PT154 glibc netbsd cygwin + HP-ROMAN8 hpux + HP-ARABIC8 hpux + HP-GREEK8 hpux + HP-HEBREW8 hpux + HP-TURKISH8 hpux + HP-KANA8 hpux + DEC-KANJI osf + DEC-HANYU osf + UTF-8 Y glibc aix hpux osf solaris netbsd darwin cygwin + + Note: Names which are not marked as being a MIME name should not be used in + Internet protocols for information interchange (mail, news, etc.). + + Note: ASCII and ANSI_X3.4-1968 are synonymous canonical names. Applications + must understand both names and treat them as equivalent. + */ + + +#ifdef __cplusplus +} +#endif + + +#endif /* _LOCALCHARSET_H */ diff --git a/lib/locale.in.h b/lib/locale.in.h new file mode 100644 index 0000000..123df7a --- /dev/null +++ b/lib/locale.in.h @@ -0,0 +1,272 @@ +/* A POSIX <locale.h>. + Copyright (C) 2007-2018 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +#if __GNUC__ >= 3 +@PRAGMA_SYSTEM_HEADER@ +#endif +@PRAGMA_COLUMNS@ + +#if (defined _WIN32 && !defined __CYGWIN__ && defined __need_locale_t) \ + || defined _GL_ALREADY_INCLUDING_LOCALE_H + +/* Special invocation convention: + - Inside mingw header files, + - To handle Solaris header files (through Solaris 10) when combined + with gettext's libintl.h. */ + +#@INCLUDE_NEXT@ @NEXT_LOCALE_H@ + +#else +/* Normal invocation convention. */ + +#ifndef _@GUARD_PREFIX@_LOCALE_H + +#define _GL_ALREADY_INCLUDING_LOCALE_H + +/* The include_next requires a split double-inclusion guard. */ +#@INCLUDE_NEXT@ @NEXT_LOCALE_H@ + +#undef _GL_ALREADY_INCLUDING_LOCALE_H + +#ifndef _@GUARD_PREFIX@_LOCALE_H +#define _@GUARD_PREFIX@_LOCALE_H + +/* NetBSD 5.0 mis-defines NULL. */ +#include <stddef.h> + +/* Mac OS X 10.5 defines the locale_t type in <xlocale.h>. */ +#if @HAVE_XLOCALE_H@ +# include <xlocale.h> +#endif + +/* The definitions of _GL_FUNCDECL_RPL etc. are copied here. */ + +/* The definition of _GL_ARG_NONNULL is copied here. */ + +/* The definition of _GL_WARN_ON_USE is copied here. */ + +/* The LC_MESSAGES locale category is specified in POSIX, but not in ISO C. + On systems that don't define it, use the same value as GNU libintl. */ +#if !defined LC_MESSAGES +# define LC_MESSAGES 1729 +#endif + +/* Bionic libc's 'struct lconv' is just a dummy. */ +#if @REPLACE_STRUCT_LCONV@ +# define lconv rpl_lconv +struct lconv +{ + /* All 'char *' are actually 'const char *'. */ + + /* Members that depend on the LC_NUMERIC category of the locale. See + <http://pubs.opengroup.org/onlinepubs/9699919799/basedefs/V1_chap07.html#tag_07_03_04> */ + + /* Symbol used as decimal point. */ + char *decimal_point; + /* Symbol used to separate groups of digits to the left of the decimal + point. */ + char *thousands_sep; + /* Definition of the size of groups of digits to the left of the decimal + point. */ + char *grouping; + + /* Members that depend on the LC_MONETARY category of the locale. See + <http://pubs.opengroup.org/onlinepubs/9699919799/basedefs/V1_chap07.html#tag_07_03_03> */ + + /* Symbol used as decimal point. */ + char *mon_decimal_point; + /* Symbol used to separate groups of digits to the left of the decimal + point. */ + char *mon_thousands_sep; + /* Definition of the size of groups of digits to the left of the decimal + point. */ + char *mon_grouping; + /* Sign used to indicate a value >= 0. */ + char *positive_sign; + /* Sign used to indicate a value < 0. */ + char *negative_sign; + + /* For formatting local currency. */ + /* Currency symbol (3 characters) followed by separator (1 character). */ + char *currency_symbol; + /* Number of digits after the decimal point. */ + char frac_digits; + /* For values >= 0: 1 if the currency symbol precedes the number, 0 if it + comes after the number. */ + char p_cs_precedes; + /* For values >= 0: Position of the sign. */ + char p_sign_posn; + /* For values >= 0: Placement of spaces between currency symbol, sign, and + number. */ + char p_sep_by_space; + /* For values < 0: 1 if the currency symbol precedes the number, 0 if it + comes after the number. */ + char n_cs_precedes; + /* For values < 0: Position of the sign. */ + char n_sign_posn; + /* For values < 0: Placement of spaces between currency symbol, sign, and + number. */ + char n_sep_by_space; + + /* For formatting international currency. */ + /* Currency symbol (3 characters) followed by separator (1 character). */ + char *int_curr_symbol; + /* Number of digits after the decimal point. */ + char int_frac_digits; + /* For values >= 0: 1 if the currency symbol precedes the number, 0 if it + comes after the number. */ + char int_p_cs_precedes; + /* For values >= 0: Position of the sign. */ + char int_p_sign_posn; + /* For values >= 0: Placement of spaces between currency symbol, sign, and + number. */ + char int_p_sep_by_space; + /* For values < 0: 1 if the currency symbol precedes the number, 0 if it + comes after the number. */ + char int_n_cs_precedes; + /* For values < 0: Position of the sign. */ + char int_n_sign_posn; + /* For values < 0: Placement of spaces between currency symbol, sign, and + number. */ + char int_n_sep_by_space; +}; +#endif + +#if @GNULIB_LOCALECONV@ +# if @REPLACE_LOCALECONV@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef localeconv +# define localeconv rpl_localeconv +# endif +_GL_FUNCDECL_RPL (localeconv, struct lconv *, (void)); +_GL_CXXALIAS_RPL (localeconv, struct lconv *, (void)); +# else +_GL_CXXALIAS_SYS (localeconv, struct lconv *, (void)); +# endif +_GL_CXXALIASWARN (localeconv); +#elif @REPLACE_STRUCT_LCONV@ +# undef localeconv +# define localeconv localeconv_used_without_requesting_gnulib_module_localeconv +#elif defined GNULIB_POSIXCHECK +# undef localeconv +# if HAVE_RAW_DECL_LOCALECONV +_GL_WARN_ON_USE (localeconv, + "localeconv returns too few information on some platforms - " + "use gnulib module localeconv for portability"); +# endif +#endif + +#if @GNULIB_SETLOCALE@ +# if @REPLACE_SETLOCALE@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef setlocale +# define setlocale rpl_setlocale +# define GNULIB_defined_setlocale 1 +# endif +_GL_FUNCDECL_RPL (setlocale, char *, (int category, const char *locale)); +_GL_CXXALIAS_RPL (setlocale, char *, (int category, const char *locale)); +# else +_GL_CXXALIAS_SYS (setlocale, char *, (int category, const char *locale)); +# endif +_GL_CXXALIASWARN (setlocale); +#elif defined GNULIB_POSIXCHECK +# undef setlocale +# if HAVE_RAW_DECL_SETLOCALE +_GL_WARN_ON_USE (setlocale, "setlocale works differently on native Windows - " + "use gnulib module setlocale for portability"); +# endif +#endif + +#if /*@GNULIB_NEWLOCALE@ ||*/ (@GNULIB_LOCALENAME@ && @HAVE_NEWLOCALE@) +# if @REPLACE_NEWLOCALE@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef newlocale +# define newlocale rpl_newlocale +# define GNULIB_defined_newlocale 1 +# endif +_GL_FUNCDECL_RPL (newlocale, locale_t, + (int category_mask, const char *name, locale_t base) + _GL_ARG_NONNULL ((2))); +_GL_CXXALIAS_RPL (newlocale, locale_t, + (int category_mask, const char *name, locale_t base)); +# else +# if @HAVE_NEWLOCALE@ +_GL_CXXALIAS_SYS (newlocale, locale_t, + (int category_mask, const char *name, locale_t base)); +# endif +# endif +# if @HAVE_NEWLOCALE@ +_GL_CXXALIASWARN (newlocale); +# endif +#elif defined GNULIB_POSIXCHECK +# undef newlocale +# if HAVE_RAW_DECL_NEWLOCALE +_GL_WARN_ON_USE (newlocale, "newlocale is not portable"); +# endif +#endif + +#if @GNULIB_DUPLOCALE@ || (@GNULIB_LOCALENAME@ && @HAVE_DUPLOCALE@) +# if @REPLACE_DUPLOCALE@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef duplocale +# define duplocale rpl_duplocale +# define GNULIB_defined_duplocale 1 +# endif +_GL_FUNCDECL_RPL (duplocale, locale_t, (locale_t locale) _GL_ARG_NONNULL ((1))); +_GL_CXXALIAS_RPL (duplocale, locale_t, (locale_t locale)); +# else +# if @HAVE_DUPLOCALE@ +_GL_CXXALIAS_SYS (duplocale, locale_t, (locale_t locale)); +# endif +# endif +# if @HAVE_DUPLOCALE@ +_GL_CXXALIASWARN (duplocale); +# endif +#elif defined GNULIB_POSIXCHECK +# undef duplocale +# if HAVE_RAW_DECL_DUPLOCALE +_GL_WARN_ON_USE (duplocale, "duplocale is buggy on some glibc systems - " + "use gnulib module duplocale for portability"); +# endif +#endif + +#if /*@GNULIB_FREELOCALE@ ||*/ (@GNULIB_LOCALENAME@ && @HAVE_FREELOCALE@) +# if @REPLACE_FREELOCALE@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef freelocale +# define freelocale rpl_freelocale +# define GNULIB_defined_freelocale 1 +# endif +_GL_FUNCDECL_RPL (freelocale, void, (locale_t locale) _GL_ARG_NONNULL ((1))); +_GL_CXXALIAS_RPL (freelocale, void, (locale_t locale)); +# else +# if @HAVE_FREELOCALE@ +_GL_CXXALIAS_SYS (freelocale, void, (locale_t locale)); +# endif +# endif +# if @HAVE_FREELOCALE@ +_GL_CXXALIASWARN (freelocale); +# endif +#elif defined GNULIB_POSIXCHECK +# undef freelocale +# if HAVE_RAW_DECL_FREELOCALE +_GL_WARN_ON_USE (freelocale, "freelocale is not portable"); +# endif +#endif + +#endif /* _@GUARD_PREFIX@_LOCALE_H */ +#endif /* _@GUARD_PREFIX@_LOCALE_H */ +#endif /* !(__need_locale_t || _GL_ALREADY_INCLUDING_LOCALE_H) */ diff --git a/lib/localeconv.c b/lib/localeconv.c new file mode 100644 index 0000000..87cface --- /dev/null +++ b/lib/localeconv.c @@ -0,0 +1,103 @@ +/* Query locale dependent information for formatting numbers. + Copyright (C) 2012-2018 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +#include <config.h> + +/* Specification. */ +#include <locale.h> + +#if HAVE_STRUCT_LCONV_DECIMAL_POINT + +/* Override for platforms where 'struct lconv' lacks the int_p_*, int_n_* + members. */ + +struct lconv * +localeconv (void) +{ + static struct lconv result; +# undef lconv +# undef localeconv + struct lconv *sys_result = localeconv (); + + result.decimal_point = sys_result->decimal_point; + result.thousands_sep = sys_result->thousands_sep; + result.grouping = sys_result->grouping; + result.mon_decimal_point = sys_result->mon_decimal_point; + result.mon_thousands_sep = sys_result->mon_thousands_sep; + result.mon_grouping = sys_result->mon_grouping; + result.positive_sign = sys_result->positive_sign; + result.negative_sign = sys_result->negative_sign; + result.currency_symbol = sys_result->currency_symbol; + result.frac_digits = sys_result->frac_digits; + result.p_cs_precedes = sys_result->p_cs_precedes; + result.p_sign_posn = sys_result->p_sign_posn; + result.p_sep_by_space = sys_result->p_sep_by_space; + result.n_cs_precedes = sys_result->n_cs_precedes; + result.n_sign_posn = sys_result->n_sign_posn; + result.n_sep_by_space = sys_result->n_sep_by_space; + result.int_curr_symbol = sys_result->int_curr_symbol; + result.int_frac_digits = sys_result->int_frac_digits; + result.int_p_cs_precedes = sys_result->p_cs_precedes; + result.int_p_sign_posn = sys_result->p_sign_posn; + result.int_p_sep_by_space = sys_result->p_sep_by_space; + result.int_n_cs_precedes = sys_result->n_cs_precedes; + result.int_n_sign_posn = sys_result->n_sign_posn; + result.int_n_sep_by_space = sys_result->n_sep_by_space; + + return &result; +} + +#else + +/* Override for platforms where 'struct lconv' is a dummy. */ + +# include <limits.h> + +struct lconv * +localeconv (void) +{ + static /*const*/ struct lconv result = + { + /* decimal_point */ ".", + /* thousands_sep */ "", + /* grouping */ "", + /* mon_decimal_point */ "", + /* mon_thousands_sep */ "", + /* mon_grouping */ "", + /* positive_sign */ "", + /* negative_sign */ "", + /* currency_symbol */ "", + /* frac_digits */ CHAR_MAX, + /* p_cs_precedes */ CHAR_MAX, + /* p_sign_posn */ CHAR_MAX, + /* p_sep_by_space */ CHAR_MAX, + /* n_cs_precedes */ CHAR_MAX, + /* n_sign_posn */ CHAR_MAX, + /* n_sep_by_space */ CHAR_MAX, + /* int_curr_symbol */ "", + /* int_frac_digits */ CHAR_MAX, + /* int_p_cs_precedes */ CHAR_MAX, + /* int_p_sign_posn */ CHAR_MAX, + /* int_p_sep_by_space */ CHAR_MAX, + /* int_n_cs_precedes */ CHAR_MAX, + /* int_n_sign_posn */ CHAR_MAX, + /* int_n_sep_by_space */ CHAR_MAX + }; + + return &result; +} + +#endif diff --git a/lib/localeinfo.c b/lib/localeinfo.c new file mode 100644 index 0000000..4fdeeb4 --- /dev/null +++ b/lib/localeinfo.c @@ -0,0 +1,113 @@ +/* locale information + + Copyright 2016-2018 Free Software Foundation, Inc. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3, or (at your option) + any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, write to the Free Software + Foundation, Inc., 51 Franklin Street - Fifth Floor, Boston, MA + 02110-1301, USA. */ + +/* Written by Paul Eggert. */ + +#include <config.h> + +#include <localeinfo.h> + +#include <verify.h> + +#include <limits.h> +#include <locale.h> +#include <stdlib.h> +#include <string.h> +#include <wctype.h> + +/* The sbclen implementation relies on this. */ +verify (MB_LEN_MAX <= SCHAR_MAX); + +/* Return true if the locale uses UTF-8. */ + +static bool +is_using_utf8 (void) +{ + wchar_t wc; + mbstate_t mbs = {0}; + return mbrtowc (&wc, "\xc4\x80", 2, &mbs) == 2 && wc == 0x100; +} + +/* Initialize *LOCALEINFO from the current locale. */ + +void +init_localeinfo (struct localeinfo *localeinfo) +{ + int i; + + localeinfo->multibyte = MB_CUR_MAX > 1; + localeinfo->using_utf8 = is_using_utf8 (); + + for (i = CHAR_MIN; i <= CHAR_MAX; i++) + { + char c = i; + unsigned char uc = i; + mbstate_t s = {0}; + wchar_t wc; + size_t len = mbrtowc (&wc, &c, 1, &s); + localeinfo->sbclen[uc] = len <= 1 ? 1 : - (int) - len; + localeinfo->sbctowc[uc] = len <= 1 ? wc : WEOF; + } +} + +/* The set of wchar_t values C such that there's a useful locale + somewhere where C != towupper (C) && C != towlower (towupper (C)). + For example, 0x00B5 (U+00B5 MICRO SIGN) is in this table, because + towupper (0x00B5) == 0x039C (U+039C GREEK CAPITAL LETTER MU), and + towlower (0x039C) == 0x03BC (U+03BC GREEK SMALL LETTER MU). */ +static short const lonesome_lower[] = + { + 0x00B5, 0x0131, 0x017F, 0x01C5, 0x01C8, 0x01CB, 0x01F2, 0x0345, + 0x03C2, 0x03D0, 0x03D1, 0x03D5, 0x03D6, 0x03F0, 0x03F1, + + /* U+03F2 GREEK LUNATE SIGMA SYMBOL lacks a specific uppercase + counterpart in locales predating Unicode 4.0.0 (April 2003). */ + 0x03F2, + + 0x03F5, 0x1E9B, 0x1FBE, + }; + +/* Verify that the worst case fits. This is 1 for towupper, 1 for + towlower, and 1 for each entry in LONESOME_LOWER. */ +verify (1 + 1 + sizeof lonesome_lower / sizeof *lonesome_lower + <= CASE_FOLDED_BUFSIZE); + +/* Find the characters equal to C after case-folding, other than C + itself, and store them into FOLDED. Return the number of characters + stored; this is zero if C is WEOF. */ + +int +case_folded_counterparts (wint_t c, wchar_t folded[CASE_FOLDED_BUFSIZE]) +{ + int i; + int n = 0; + wint_t uc = towupper (c); + wint_t lc = towlower (uc); + if (uc != c) + folded[n++] = uc; + if (lc != uc && lc != c && towupper (lc) == uc) + folded[n++] = lc; + for (i = 0; i < sizeof lonesome_lower / sizeof *lonesome_lower; i++) + { + wint_t li = lonesome_lower[i]; + if (li != lc && li != uc && li != c && towupper (li) == uc) + folded[n++] = li; + } + return n; +} diff --git a/lib/localeinfo.h b/lib/localeinfo.h new file mode 100644 index 0000000..fcd5311 --- /dev/null +++ b/lib/localeinfo.h @@ -0,0 +1,54 @@ +/* locale information + + Copyright 2016-2018 Free Software Foundation, Inc. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3, or (at your option) + any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, write to the Free Software + Foundation, Inc., 51 Franklin Street - Fifth Floor, Boston, MA + 02110-1301, USA. */ + +/* Written by Paul Eggert. */ + +#include <limits.h> +#include <stdbool.h> +#include <wchar.h> + +struct localeinfo +{ + /* MB_CUR_MAX > 1. */ + bool multibyte; + + /* The locale uses UTF-8. */ + bool using_utf8; + + /* An array indexed by byte values B that contains 1 if B is a + single-byte character, -1 if B is an encoding error, and -2 if B + is the leading byte of a multibyte character that contains more + than one byte. */ + signed char sbclen[UCHAR_MAX + 1]; + + /* An array indexed by byte values B that contains the corresponding + wide character (if any) for B if sbclen[B] == 1. WEOF means the + byte is not a valid single-byte character, i.e., sbclen[B] == -1 + or -2. */ + wint_t sbctowc[UCHAR_MAX + 1]; +}; + +extern void init_localeinfo (struct localeinfo *); + +/* Maximum number of characters that can be the case-folded + counterparts of a single character, not counting the character + itself. This is a generous upper bound. */ +enum { CASE_FOLDED_BUFSIZE = 32 }; + +extern int case_folded_counterparts (wint_t, wchar_t[CASE_FOLDED_BUFSIZE]); diff --git a/lib/localtime-buffer.c b/lib/localtime-buffer.c new file mode 100644 index 0000000..910c8bc --- /dev/null +++ b/lib/localtime-buffer.c @@ -0,0 +1,60 @@ +/* Provide access to the last buffer returned by localtime() or gmtime(). + + Copyright (C) 2001-2003, 2005-2007, 2009-2018 Free Software Foundation, Inc. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3, or (at your option) + any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, see <https://www.gnu.org/licenses/>. */ + +/* written by Jim Meyering */ + +#include <config.h> + +/* Specification. */ +#include "localtime-buffer.h" + +#if GETTIMEOFDAY_CLOBBERS_LOCALTIME || TZSET_CLOBBERS_LOCALTIME + +static struct tm tm_zero_buffer; +struct tm *localtime_buffer_addr = &tm_zero_buffer; + +/* This is a wrapper for localtime. + + On the first call, record the address of the static buffer that + localtime uses for its result. */ + +struct tm * +rpl_localtime (time_t const *timep) +#undef localtime +{ + struct tm *tm = localtime (timep); + + if (localtime_buffer_addr == &tm_zero_buffer) + localtime_buffer_addr = tm; + + return tm; +} + +/* Same as above, since gmtime and localtime use the same buffer. */ +struct tm * +rpl_gmtime (time_t const *timep) +#undef gmtime +{ + struct tm *tm = gmtime (timep); + + if (localtime_buffer_addr == &tm_zero_buffer) + localtime_buffer_addr = tm; + + return tm; +} + +#endif diff --git a/lib/localtime-buffer.h b/lib/localtime-buffer.h new file mode 100644 index 0000000..f381ff0 --- /dev/null +++ b/lib/localtime-buffer.h @@ -0,0 +1,27 @@ +/* Provide access to the last buffer returned by localtime() or gmtime(). + + Copyright (C) 2001-2003, 2005-2007, 2009-2018 Free Software Foundation, Inc. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3, or (at your option) + any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, see <https://www.gnu.org/licenses/>. */ + +/* written by Jim Meyering */ + +#include <time.h> + +#if GETTIMEOFDAY_CLOBBERS_LOCALTIME || TZSET_CLOBBERS_LOCALTIME + +/* The address of the last buffer returned by localtime() or gmtime(). */ +extern struct tm *localtime_buffer_addr; + +#endif diff --git a/lib/lstat.c b/lib/lstat.c new file mode 100644 index 0000000..5873bbd --- /dev/null +++ b/lib/lstat.c @@ -0,0 +1,100 @@ +/* Work around a bug of lstat on some systems + + Copyright (C) 1997-2006, 2008-2018 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +/* written by Jim Meyering */ + +/* If the user's config.h happens to include <sys/stat.h>, let it include only + the system's <sys/stat.h> here, so that orig_lstat doesn't recurse to + rpl_lstat. */ +#define __need_system_sys_stat_h +#include <config.h> + +#if !HAVE_LSTAT +/* On systems that lack symlinks, our replacement <sys/stat.h> already + defined lstat as stat, so there is nothing further to do other than + avoid an empty file. */ +typedef int dummy; +#else /* HAVE_LSTAT */ + +/* Get the original definition of lstat. It might be defined as a macro. */ +# include <sys/types.h> +# include <sys/stat.h> +# undef __need_system_sys_stat_h + +static int +orig_lstat (const char *filename, struct stat *buf) +{ + return lstat (filename, buf); +} + +/* Specification. */ +/* Write "sys/stat.h" here, not <sys/stat.h>, otherwise OSF/1 5.1 DTK cc + eliminates this include because of the preliminary #include <sys/stat.h> + above. */ +# include "sys/stat.h" + +# include "stat-time.h" + +# include <string.h> +# include <errno.h> + +/* lstat works differently on Linux and Solaris systems. POSIX (see + "pathname resolution" in the glossary) requires that programs like + 'ls' take into consideration the fact that FILE has a trailing slash + when FILE is a symbolic link. On Linux and Solaris 10 systems, the + lstat function already has the desired semantics (in treating + 'lstat ("symlink/", sbuf)' just like 'lstat ("symlink/.", sbuf)', + but on Solaris 9 and earlier it does not. + + If FILE has a trailing slash and specifies a symbolic link, + then use stat() to get more info on the referent of FILE. + If the referent is a non-directory, then set errno to ENOTDIR + and return -1. Otherwise, return stat's result. */ + +int +rpl_lstat (const char *file, struct stat *sbuf) +{ + int result = orig_lstat (file, sbuf); + + /* This replacement file can blindly check against '/' rather than + using the ISSLASH macro, because all platforms with '\\' either + lack symlinks (mingw) or have working lstat (cygwin) and thus do + not compile this file. 0 len should have already been filtered + out above, with a failure return of ENOENT. */ + if (result == 0) + { + if (S_ISDIR (sbuf->st_mode) || file[strlen (file) - 1] != '/') + result = stat_time_normalize (result, sbuf); + else + { + /* At this point, a trailing slash is permitted only on + symlink-to-dir; but it should have found information on the + directory, not the symlink. Call 'stat' to get info about the + link's referent. Our replacement stat guarantees valid results, + even if the symlink is not pointing to a directory. */ + if (!S_ISLNK (sbuf->st_mode)) + { + errno = ENOTDIR; + return -1; + } + result = stat (file, sbuf); + } + } + return result; +} + +#endif /* HAVE_LSTAT */ diff --git a/lib/malloc.c b/lib/malloc.c new file mode 100644 index 0000000..471c3ab --- /dev/null +++ b/lib/malloc.c @@ -0,0 +1,56 @@ +/* malloc() function that is glibc compatible. + + Copyright (C) 1997-1998, 2006-2007, 2009-2018 Free Software Foundation, Inc. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3, or (at your option) + any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, see <https://www.gnu.org/licenses/>. */ + +/* written by Jim Meyering and Bruno Haible */ + +#define _GL_USE_STDLIB_ALLOC 1 +#include <config.h> +/* Only the AC_FUNC_MALLOC macro defines 'malloc' already in config.h. */ +#ifdef malloc +# define NEED_MALLOC_GNU 1 +# undef malloc +/* Whereas the gnulib module 'malloc-gnu' defines HAVE_MALLOC_GNU. */ +#elif GNULIB_MALLOC_GNU && !HAVE_MALLOC_GNU +# define NEED_MALLOC_GNU 1 +#endif + +#include <stdlib.h> + +#include <errno.h> + +/* Allocate an N-byte block of memory from the heap. + If N is zero, allocate a 1-byte block. */ + +void * +rpl_malloc (size_t n) +{ + void *result; + +#if NEED_MALLOC_GNU + if (n == 0) + n = 1; +#endif + + result = malloc (n); + +#if !HAVE_MALLOC_POSIX + if (result == NULL) + errno = ENOMEM; +#endif + + return result; +} diff --git a/lib/malloca.c b/lib/malloca.c new file mode 100644 index 0000000..c54e1e0 --- /dev/null +++ b/lib/malloca.c @@ -0,0 +1,105 @@ +/* Safe automatic memory allocation. + Copyright (C) 2003, 2006-2007, 2009-2018 Free Software Foundation, Inc. + Written by Bruno Haible <bruno@clisp.org>, 2003, 2018. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3, or (at your option) + any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, see <https://www.gnu.org/licenses/>. */ + +#define _GL_USE_STDLIB_ALLOC 1 +#include <config.h> + +/* Specification. */ +#include "malloca.h" + +#include "verify.h" + +/* The speed critical point in this file is freea() applied to an alloca() + result: it must be fast, to match the speed of alloca(). The speed of + mmalloca() and freea() in the other case are not critical, because they + are only invoked for big memory sizes. + Here we use a bit in the address as an indicator, an idea by OndÅ™ej BÃlka. + malloca() can return three types of pointers: + - Pointers ≡ 0 mod 2*sa_alignment_max come from stack allocation. + - Pointers ≡ sa_alignment_max mod 2*sa_alignment_max come from heap + allocation. + - NULL comes from a failed heap allocation. */ + +/* Type for holding very small pointer differences. */ +typedef unsigned char small_t; +/* Verify that it is wide enough. */ +verify (2 * sa_alignment_max - 1 <= (small_t) -1); + +void * +mmalloca (size_t n) +{ +#if HAVE_ALLOCA + /* Allocate one more word, used to determine the address to pass to freea(), + and room for the alignment ≡ sa_alignment_max mod 2*sa_alignment_max. */ + size_t nplus = n + sizeof (small_t) + 2 * sa_alignment_max - 1; + + if (nplus >= n) + { + char *mem = (char *) malloc (nplus); + + if (mem != NULL) + { + char *p = + (char *)((((uintptr_t)mem + sizeof (small_t) + sa_alignment_max - 1) + & ~(uintptr_t)(2 * sa_alignment_max - 1)) + + sa_alignment_max); + /* Here p >= mem + sizeof (small_t), + and p <= mem + sizeof (small_t) + 2 * sa_alignment_max - 1 + hence p + n <= mem + nplus. + So, the memory range [p, p+n) lies in the allocated memory range + [mem, mem + nplus). */ + ((small_t *) p)[-1] = p - mem; + /* p ≡ sa_alignment_max mod 2*sa_alignment_max. */ + return p; + } + } + /* Out of memory. */ + return NULL; +#else +# if !MALLOC_0_IS_NONNULL + if (n == 0) + n = 1; +# endif + return malloc (n); +#endif +} + +#if HAVE_ALLOCA +void +freea (void *p) +{ + /* Check argument. */ + if ((uintptr_t) p & (sa_alignment_max - 1)) + { + /* p was not the result of a malloca() call. Invalid argument. */ + abort (); + } + /* Determine whether p was a non-NULL pointer returned by mmalloca(). */ + if ((uintptr_t) p & sa_alignment_max) + { + void *mem = (char *) p - ((small_t *) p)[-1]; + free (mem); + } +} +#endif + +/* + * Hey Emacs! + * Local Variables: + * coding: utf-8 + * End: + */ diff --git a/lib/malloca.h b/lib/malloca.h new file mode 100644 index 0000000..1e47813 --- /dev/null +++ b/lib/malloca.h @@ -0,0 +1,127 @@ +/* Safe automatic memory allocation. + Copyright (C) 2003-2007, 2009-2018 Free Software Foundation, Inc. + Written by Bruno Haible <bruno@clisp.org>, 2003. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3, or (at your option) + any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, see <https://www.gnu.org/licenses/>. */ + +#ifndef _MALLOCA_H +#define _MALLOCA_H + +#include <alloca.h> +#include <stddef.h> +#include <stdlib.h> +#include <stdint.h> + +#include "xalloc-oversized.h" + + +#ifdef __cplusplus +extern "C" { +#endif + + +/* safe_alloca(N) is equivalent to alloca(N) when it is safe to call + alloca(N); otherwise it returns NULL. It either returns N bytes of + memory allocated on the stack, that lasts until the function returns, + or NULL. + Use of safe_alloca should be avoided: + - inside arguments of function calls - undefined behaviour, + - in inline functions - the allocation may actually last until the + calling function returns. +*/ +#if HAVE_ALLOCA +/* The OS usually guarantees only one guard page at the bottom of the stack, + and a page size can be as small as 4096 bytes. So we cannot safely + allocate anything larger than 4096 bytes. Also care for the possibility + of a few compiler-allocated temporary stack slots. + This must be a macro, not a function. */ +# define safe_alloca(N) ((N) < 4032 ? alloca (N) : NULL) +#else +# define safe_alloca(N) ((void) (N), NULL) +#endif + +/* malloca(N) is a safe variant of alloca(N). It allocates N bytes of + memory allocated on the stack, that must be freed using freea() before + the function returns. Upon failure, it returns NULL. */ +#if HAVE_ALLOCA +# define malloca(N) \ + ((N) < 4032 - (2 * sa_alignment_max - 1) \ + ? (void *) (((uintptr_t) (char *) alloca ((N) + 2 * sa_alignment_max - 1) \ + + (2 * sa_alignment_max - 1)) \ + & ~(uintptr_t)(2 * sa_alignment_max - 1)) \ + : mmalloca (N)) +#else +# define malloca(N) \ + mmalloca (N) +#endif +extern void * mmalloca (size_t n); + +/* Free a block of memory allocated through malloca(). */ +#if HAVE_ALLOCA +extern void freea (void *p); +#else +# define freea free +#endif + +/* nmalloca(N,S) is an overflow-safe variant of malloca (N * S). + It allocates an array of N objects, each with S bytes of memory, + on the stack. S must be positive and N must be nonnegative. + The array must be freed using freea() before the function returns. */ +#define nmalloca(n, s) (xalloc_oversized (n, s) ? NULL : malloca ((n) * (s))) + + +#ifdef __cplusplus +} +#endif + + +/* ------------------- Auxiliary, non-public definitions ------------------- */ + +/* Determine the alignment of a type at compile time. */ +#if defined __GNUC__ || defined __IBM__ALIGNOF__ +# define sa_alignof __alignof__ +#elif defined __cplusplus + template <class type> struct sa_alignof_helper { char __slot1; type __slot2; }; +# define sa_alignof(type) offsetof (sa_alignof_helper<type>, __slot2) +#elif defined __hpux + /* Work around a HP-UX 10.20 cc bug with enums constants defined as offsetof + values. */ +# define sa_alignof(type) (sizeof (type) <= 4 ? 4 : 8) +#elif defined _AIX + /* Work around an AIX 3.2.5 xlc bug with enums constants defined as offsetof + values. */ +# define sa_alignof(type) (sizeof (type) <= 4 ? 4 : 8) +#else +# define sa_alignof(type) offsetof (struct { char __slot1; type __slot2; }, __slot2) +#endif + +enum +{ +/* The desired alignment of memory allocations is the maximum alignment + among all elementary types. */ + sa_alignment_long = sa_alignof (long), + sa_alignment_double = sa_alignof (double), +#if HAVE_LONG_LONG_INT + sa_alignment_longlong = sa_alignof (long long), +#endif + sa_alignment_longdouble = sa_alignof (long double), + sa_alignment_max = ((sa_alignment_long - 1) | (sa_alignment_double - 1) +#if HAVE_LONG_LONG_INT + | (sa_alignment_longlong - 1) +#endif + | (sa_alignment_longdouble - 1) + ) + 1 +}; + +#endif /* _MALLOCA_H */ diff --git a/lib/mbrlen.c b/lib/mbrlen.c new file mode 100644 index 0000000..2812a2e --- /dev/null +++ b/lib/mbrlen.c @@ -0,0 +1,32 @@ +/* Recognize multibyte character. + Copyright (C) 1999-2000, 2008-2018 Free Software Foundation, Inc. + Written by Bruno Haible <bruno@clisp.org>, 2008. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +#include <config.h> + +/* Specification. */ +#include <wchar.h> + + +static mbstate_t internal_state; + +size_t +mbrlen (const char *s, size_t n, mbstate_t *ps) +{ + if (ps == NULL) + ps = &internal_state; + return mbrtowc (NULL, s, n, ps); +} diff --git a/lib/mbrtowc.c b/lib/mbrtowc.c new file mode 100644 index 0000000..2f6df28 --- /dev/null +++ b/lib/mbrtowc.c @@ -0,0 +1,458 @@ +/* Convert multibyte character to wide character. + Copyright (C) 1999-2002, 2005-2018 Free Software Foundation, Inc. + Written by Bruno Haible <bruno@clisp.org>, 2008. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +#include <config.h> + +/* Specification. */ +#include <wchar.h> + +#if C_LOCALE_MAYBE_EILSEQ +# include "hard-locale.h" +# include <locale.h> +#endif + +#if GNULIB_defined_mbstate_t +/* Implement mbrtowc() on top of mbtowc(). */ + +# include <errno.h> +# include <stdlib.h> + +# include "localcharset.h" +# include "streq.h" +# include "verify.h" + +# ifndef FALLTHROUGH +# if __GNUC__ < 7 +# define FALLTHROUGH ((void) 0) +# else +# define FALLTHROUGH __attribute__ ((__fallthrough__)) +# endif +# endif + +/* Returns a classification of special values of the encoding of the current + locale. */ +typedef enum { + enc_other, /* other */ + enc_utf8, /* UTF-8 */ + enc_eucjp, /* EUC-JP */ + enc_94, /* EUC-KR, GB2312, BIG5 */ + enc_euctw, /* EUC-TW */ + enc_gb18030, /* GB18030 */ + enc_sjis /* SJIS */ +} enc_t; +static inline enc_t +locale_enc (void) +{ + const char *encoding = locale_charset (); + if (STREQ_OPT (encoding, "UTF-8", 'U', 'T', 'F', '-', '8', 0, 0, 0, 0)) + return enc_utf8; + if (STREQ_OPT (encoding, "EUC-JP", 'E', 'U', 'C', '-', 'J', 'P', 0, 0, 0)) + return enc_eucjp; + if (STREQ_OPT (encoding, "EUC-KR", 'E', 'U', 'C', '-', 'K', 'R', 0, 0, 0) + || STREQ_OPT (encoding, "GB2312", 'G', 'B', '2', '3', '1', '2', 0, 0, 0) + || STREQ_OPT (encoding, "BIG5", 'B', 'I', 'G', '5', 0, 0, 0, 0, 0)) + return enc_94; + if (STREQ_OPT (encoding, "EUC-TW", 'E', 'U', 'C', '-', 'T', 'W', 0, 0, 0)) + return enc_euctw; + if (STREQ_OPT (encoding, "GB18030", 'G', 'B', '1', '8', '0', '3', '0', 0, 0)) + return enc_gb18030; + if (STREQ_OPT (encoding, "SJIS", 'S', 'J', 'I', 'S', 0, 0, 0, 0, 0)) + return enc_sjis; + return enc_other; +} + +#if GNULIB_WCHAR_SINGLE +/* When we know that the locale does not change, provide a speedup by + caching the value of locale_enc. */ +static int cached_locale_enc = -1; +static inline enc_t +locale_enc_cached (void) +{ + if (cached_locale_enc < 0) + cached_locale_enc = locale_enc (); + return cached_locale_enc; +} +#else +/* By default, don't make assumptions, hence no caching. */ +# define locale_enc_cached locale_enc +#endif + +verify (sizeof (mbstate_t) >= 4); + +static char internal_state[4]; + +size_t +mbrtowc (wchar_t *pwc, const char *s, size_t n, mbstate_t *ps) +{ + char *pstate = (char *)ps; + + if (s == NULL) + { + pwc = NULL; + s = ""; + n = 1; + } + + if (n == 0) + return (size_t)(-2); + + /* Here n > 0. */ + + if (pstate == NULL) + pstate = internal_state; + + { + size_t nstate = pstate[0]; + char buf[4]; + const char *p; + size_t m; + + switch (nstate) + { + case 0: + p = s; + m = n; + break; + case 3: + buf[2] = pstate[3]; + FALLTHROUGH; + case 2: + buf[1] = pstate[2]; + FALLTHROUGH; + case 1: + buf[0] = pstate[1]; + p = buf; + m = nstate; + buf[m++] = s[0]; + if (n >= 2 && m < 4) + { + buf[m++] = s[1]; + if (n >= 3 && m < 4) + buf[m++] = s[2]; + } + break; + default: + errno = EINVAL; + return (size_t)(-1); + } + + /* Here m > 0. */ + +# if __GLIBC__ || defined __UCLIBC__ + /* Work around bug <https://sourceware.org/bugzilla/show_bug.cgi?id=9674> */ + mbtowc (NULL, NULL, 0); +# endif + { + int res = mbtowc (pwc, p, m); + + if (res >= 0) + { + if (pwc != NULL && ((*pwc == 0) != (res == 0))) + abort (); + if (nstate >= (res > 0 ? res : 1)) + abort (); + res -= nstate; + pstate[0] = 0; + return res; + } + + /* mbtowc does not distinguish between invalid and incomplete multibyte + sequences. But mbrtowc needs to make this distinction. + There are two possible approaches: + - Use iconv() and its return value. + - Use built-in knowledge about the possible encodings. + Given the low quality of implementation of iconv() on the systems that + lack mbrtowc(), we use the second approach. + The possible encodings are: + - 8-bit encodings, + - EUC-JP, EUC-KR, GB2312, EUC-TW, BIG5, GB18030, SJIS, + - UTF-8. + Use specialized code for each. */ + if (m >= 4 || m >= MB_CUR_MAX) + goto invalid; + /* Here MB_CUR_MAX > 1 and 0 < m < 4. */ + switch (locale_enc_cached ()) + { + case enc_utf8: /* UTF-8 */ + { + /* Cf. unistr/u8-mblen.c. */ + unsigned char c = (unsigned char) p[0]; + + if (c >= 0xc2) + { + if (c < 0xe0) + { + if (m == 1) + goto incomplete; + } + else if (c < 0xf0) + { + if (m == 1) + goto incomplete; + if (m == 2) + { + unsigned char c2 = (unsigned char) p[1]; + + if ((c2 ^ 0x80) < 0x40 + && (c >= 0xe1 || c2 >= 0xa0) + && (c != 0xed || c2 < 0xa0)) + goto incomplete; + } + } + else if (c <= 0xf4) + { + if (m == 1) + goto incomplete; + else /* m == 2 || m == 3 */ + { + unsigned char c2 = (unsigned char) p[1]; + + if ((c2 ^ 0x80) < 0x40 + && (c >= 0xf1 || c2 >= 0x90) + && (c < 0xf4 || (c == 0xf4 && c2 < 0x90))) + { + if (m == 2) + goto incomplete; + else /* m == 3 */ + { + unsigned char c3 = (unsigned char) p[2]; + + if ((c3 ^ 0x80) < 0x40) + goto incomplete; + } + } + } + } + } + goto invalid; + } + + /* As a reference for this code, you can use the GNU libiconv + implementation. Look for uses of the RET_TOOFEW macro. */ + + case enc_eucjp: /* EUC-JP */ + { + if (m == 1) + { + unsigned char c = (unsigned char) p[0]; + + if ((c >= 0xa1 && c < 0xff) || c == 0x8e || c == 0x8f) + goto incomplete; + } + if (m == 2) + { + unsigned char c = (unsigned char) p[0]; + + if (c == 0x8f) + { + unsigned char c2 = (unsigned char) p[1]; + + if (c2 >= 0xa1 && c2 < 0xff) + goto incomplete; + } + } + goto invalid; + } + + case enc_94: /* EUC-KR, GB2312, BIG5 */ + { + if (m == 1) + { + unsigned char c = (unsigned char) p[0]; + + if (c >= 0xa1 && c < 0xff) + goto incomplete; + } + goto invalid; + } + + case enc_euctw: /* EUC-TW */ + { + if (m == 1) + { + unsigned char c = (unsigned char) p[0]; + + if ((c >= 0xa1 && c < 0xff) || c == 0x8e) + goto incomplete; + } + else /* m == 2 || m == 3 */ + { + unsigned char c = (unsigned char) p[0]; + + if (c == 0x8e) + goto incomplete; + } + goto invalid; + } + + case enc_gb18030: /* GB18030 */ + { + if (m == 1) + { + unsigned char c = (unsigned char) p[0]; + + if ((c >= 0x90 && c <= 0xe3) || (c >= 0xf8 && c <= 0xfe)) + goto incomplete; + } + else /* m == 2 || m == 3 */ + { + unsigned char c = (unsigned char) p[0]; + + if (c >= 0x90 && c <= 0xe3) + { + unsigned char c2 = (unsigned char) p[1]; + + if (c2 >= 0x30 && c2 <= 0x39) + { + if (m == 2) + goto incomplete; + else /* m == 3 */ + { + unsigned char c3 = (unsigned char) p[2]; + + if (c3 >= 0x81 && c3 <= 0xfe) + goto incomplete; + } + } + } + } + goto invalid; + } + + case enc_sjis: /* SJIS */ + { + if (m == 1) + { + unsigned char c = (unsigned char) p[0]; + + if ((c >= 0x81 && c <= 0x9f) || (c >= 0xe0 && c <= 0xea) + || (c >= 0xf0 && c <= 0xf9)) + goto incomplete; + } + goto invalid; + } + + default: + /* An unknown multibyte encoding. */ + goto incomplete; + } + + incomplete: + { + size_t k = nstate; + /* Here 0 <= k < m < 4. */ + pstate[++k] = s[0]; + if (k < m) + { + pstate[++k] = s[1]; + if (k < m) + pstate[++k] = s[2]; + } + if (k != m) + abort (); + } + pstate[0] = m; + return (size_t)(-2); + + invalid: + errno = EILSEQ; + /* The conversion state is undefined, says POSIX. */ + return (size_t)(-1); + } + } +} + +#else +/* Override the system's mbrtowc() function. */ + +# undef mbrtowc + +size_t +rpl_mbrtowc (wchar_t *pwc, const char *s, size_t n, mbstate_t *ps) +{ + size_t ret; + wchar_t wc; + +# if MBRTOWC_NULL_ARG2_BUG || MBRTOWC_RETVAL_BUG || MBRTOWC_EMPTY_INPUT_BUG + if (s == NULL) + { + pwc = NULL; + s = ""; + n = 1; + } +# endif + +# if MBRTOWC_EMPTY_INPUT_BUG + if (n == 0) + return (size_t) -2; +# endif + + if (! pwc) + pwc = &wc; + +# if MBRTOWC_RETVAL_BUG + { + static mbstate_t internal_state; + + /* Override mbrtowc's internal state. We cannot call mbsinit() on the + hidden internal state, but we can call it on our variable. */ + if (ps == NULL) + ps = &internal_state; + + if (!mbsinit (ps)) + { + /* Parse the rest of the multibyte character byte for byte. */ + size_t count = 0; + for (; n > 0; s++, n--) + { + ret = mbrtowc (&wc, s, 1, ps); + + if (ret == (size_t)(-1)) + return (size_t)(-1); + count++; + if (ret != (size_t)(-2)) + { + /* The multibyte character has been completed. */ + *pwc = wc; + return (wc == 0 ? 0 : count); + } + } + return (size_t)(-2); + } + } +# endif + + ret = mbrtowc (pwc, s, n, ps); + +# if MBRTOWC_NUL_RETVAL_BUG + if (ret < (size_t) -2 && !*pwc) + return 0; +# endif + +# if C_LOCALE_MAYBE_EILSEQ + if ((size_t) -2 <= ret && n != 0 && ! hard_locale (LC_CTYPE)) + { + unsigned char uc = *s; + *pwc = uc; + return 1; + } +# endif + + return ret; +} + +#endif diff --git a/lib/mbsinit.c b/lib/mbsinit.c new file mode 100644 index 0000000..e23a5ad --- /dev/null +++ b/lib/mbsinit.c @@ -0,0 +1,73 @@ +/* Test for initial conversion state. + Copyright (C) 2008-2018 Free Software Foundation, Inc. + Written by Bruno Haible <bruno@clisp.org>, 2008. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +#include <config.h> + +/* Specification. */ +#include <wchar.h> + +#include "verify.h" + +#if GNULIB_defined_mbstate_t + +/* Platforms that lack mbsinit() also lack mbrlen(), mbrtowc(), mbsrtowcs() + and wcrtomb(), wcsrtombs(). + We assume that + - sizeof (mbstate_t) >= 4, + - only stateless encodings are supported (such as UTF-8 and EUC-JP, but + not ISO-2022 variants), + - for each encoding, the number of bytes for a wide character is <= 4. + (This maximum is attained for UTF-8, GB18030, EUC-TW.) + We define the meaning of mbstate_t as follows: + - In mb -> wc direction, mbstate_t's first byte contains the number of + buffered bytes (in the range 0..3), followed by up to 3 buffered bytes. + See mbrtowc.c. + - In wc -> mb direction, mbstate_t contains no information. In other + words, it is always in the initial state. */ + +verify (sizeof (mbstate_t) >= 4); + +int +mbsinit (const mbstate_t *ps) +{ + const char *pstate = (const char *)ps; + + return pstate == NULL || pstate[0] == 0; +} + +#else + +int +mbsinit (const mbstate_t *ps) +{ +# if defined _WIN32 && !defined __CYGWIN__ + /* Native Windows. */ +# ifdef __MINGW32__ + /* On mingw, 'mbstate_t' is defined as 'int'. */ + return ps == NULL || *ps == 0; +# else + /* MSVC defines 'mbstate_t' as an 8-byte struct; the first 4-bytes matter. */ + return ps == NULL || *(const unsigned int *)ps == 0; +# endif +# else + /* Minix, HP-UX 11.00, Solaris 2.6, Interix, ... */ + /* Maybe this definition works, maybe not... */ + return ps == NULL || *(const char *)ps == 0; +# endif +} + +#endif diff --git a/lib/mbtowc-impl.h b/lib/mbtowc-impl.h new file mode 100644 index 0000000..633832f --- /dev/null +++ b/lib/mbtowc-impl.h @@ -0,0 +1,44 @@ +/* Convert multibyte character to wide character. + Copyright (C) 2011-2018 Free Software Foundation, Inc. + Written by Bruno Haible <bruno@clisp.org>, 2011. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +/* We don't need a static internal state, because the encoding is not state + dependent, and when mbrtowc returns (size_t)(-2). we throw the result + away. */ + +int +mbtowc (wchar_t *pwc, const char *s, size_t n) +{ + if (s == NULL) + return 0; + else + { + mbstate_t state; + wchar_t wc; + size_t result; + + memset (&state, 0, sizeof (mbstate_t)); + result = mbrtowc (&wc, s, n, &state); + if (result == (size_t)-1 || result == (size_t)-2) + { + errno = EILSEQ; + return -1; + } + if (pwc != NULL) + *pwc = wc; + return (wc == 0 ? 0 : result); + } +} diff --git a/lib/mbtowc.c b/lib/mbtowc.c new file mode 100644 index 0000000..a212b22 --- /dev/null +++ b/lib/mbtowc.c @@ -0,0 +1,26 @@ +/* Convert multibyte character to wide character. + Copyright (C) 2011-2018 Free Software Foundation, Inc. + Written by Bruno Haible <bruno@clisp.org>, 2011. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +#include <config.h> + +#include <stdlib.h> + +#include <errno.h> +#include <string.h> +#include <wchar.h> + +#include "mbtowc-impl.h" diff --git a/lib/memchr.c b/lib/memchr.c new file mode 100644 index 0000000..0b47546 --- /dev/null +++ b/lib/memchr.c @@ -0,0 +1,172 @@ +/* Copyright (C) 1991, 1993, 1996-1997, 1999-2000, 2003-2004, 2006, 2008-2018 + Free Software Foundation, Inc. + + Based on strlen implementation by Torbjorn Granlund (tege@sics.se), + with help from Dan Sahlin (dan@sics.se) and + commentary by Jim Blandy (jimb@ai.mit.edu); + adaptation to memchr suggested by Dick Karpinski (dick@cca.ucsf.edu), + and implemented by Roland McGrath (roland@ai.mit.edu). + +NOTE: The canonical source of this file is maintained with the GNU C Library. +Bugs can be reported to bug-glibc@prep.ai.mit.edu. + +This program is free software: you can redistribute it and/or modify it +under the terms of the GNU General Public License as published by the +Free Software Foundation; either version 3 of the License, or any +later version. + +This program is distributed in the hope that it will be useful, +but WITHOUT ANY WARRANTY; without even the implied warranty of +MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +GNU General Public License for more details. + +You should have received a copy of the GNU General Public License +along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +#ifndef _LIBC +# include <config.h> +#endif + +#include <string.h> + +#include <stddef.h> + +#if defined _LIBC +# include <memcopy.h> +#else +# define reg_char char +#endif + +#include <limits.h> + +#if HAVE_BP_SYM_H || defined _LIBC +# include <bp-sym.h> +#else +# define BP_SYM(sym) sym +#endif + +#undef __memchr +#ifdef _LIBC +# undef memchr +#endif + +#ifndef weak_alias +# define __memchr memchr +#endif + +/* Search no more than N bytes of S for C. */ +void * +__memchr (void const *s, int c_in, size_t n) +{ + /* On 32-bit hardware, choosing longword to be a 32-bit unsigned + long instead of a 64-bit uintmax_t tends to give better + performance. On 64-bit hardware, unsigned long is generally 64 + bits already. Change this typedef to experiment with + performance. */ + typedef unsigned long int longword; + + const unsigned char *char_ptr; + const longword *longword_ptr; + longword repeated_one; + longword repeated_c; + unsigned reg_char c; + + c = (unsigned char) c_in; + + /* Handle the first few bytes by reading one byte at a time. + Do this until CHAR_PTR is aligned on a longword boundary. */ + for (char_ptr = (const unsigned char *) s; + n > 0 && (size_t) char_ptr % sizeof (longword) != 0; + --n, ++char_ptr) + if (*char_ptr == c) + return (void *) char_ptr; + + longword_ptr = (const longword *) char_ptr; + + /* All these elucidatory comments refer to 4-byte longwords, + but the theory applies equally well to any size longwords. */ + + /* Compute auxiliary longword values: + repeated_one is a value which has a 1 in every byte. + repeated_c has c in every byte. */ + repeated_one = 0x01010101; + repeated_c = c | (c << 8); + repeated_c |= repeated_c << 16; + if (0xffffffffU < (longword) -1) + { + repeated_one |= repeated_one << 31 << 1; + repeated_c |= repeated_c << 31 << 1; + if (8 < sizeof (longword)) + { + size_t i; + + for (i = 64; i < sizeof (longword) * 8; i *= 2) + { + repeated_one |= repeated_one << i; + repeated_c |= repeated_c << i; + } + } + } + + /* Instead of the traditional loop which tests each byte, we will test a + longword at a time. The tricky part is testing if *any of the four* + bytes in the longword in question are equal to c. We first use an xor + with repeated_c. This reduces the task to testing whether *any of the + four* bytes in longword1 is zero. + + We compute tmp = + ((longword1 - repeated_one) & ~longword1) & (repeated_one << 7). + That is, we perform the following operations: + 1. Subtract repeated_one. + 2. & ~longword1. + 3. & a mask consisting of 0x80 in every byte. + Consider what happens in each byte: + - If a byte of longword1 is zero, step 1 and 2 transform it into 0xff, + and step 3 transforms it into 0x80. A carry can also be propagated + to more significant bytes. + - If a byte of longword1 is nonzero, let its lowest 1 bit be at + position k (0 <= k <= 7); so the lowest k bits are 0. After step 1, + the byte ends in a single bit of value 0 and k bits of value 1. + After step 2, the result is just k bits of value 1: 2^k - 1. After + step 3, the result is 0. And no carry is produced. + So, if longword1 has only non-zero bytes, tmp is zero. + Whereas if longword1 has a zero byte, call j the position of the least + significant zero byte. Then the result has a zero at positions 0, ..., + j-1 and a 0x80 at position j. We cannot predict the result at the more + significant bytes (positions j+1..3), but it does not matter since we + already have a non-zero bit at position 8*j+7. + + So, the test whether any byte in longword1 is zero is equivalent to + testing whether tmp is nonzero. */ + + while (n >= sizeof (longword)) + { + longword longword1 = *longword_ptr ^ repeated_c; + + if ((((longword1 - repeated_one) & ~longword1) + & (repeated_one << 7)) != 0) + break; + longword_ptr++; + n -= sizeof (longword); + } + + char_ptr = (const unsigned char *) longword_ptr; + + /* At this point, we know that either n < sizeof (longword), or one of the + sizeof (longword) bytes starting at char_ptr is == c. On little-endian + machines, we could determine the first such byte without any further + memory accesses, just by looking at the tmp result from the last loop + iteration. But this does not work on big-endian machines. Choose code + that works in both cases. */ + + for (; n > 0; --n, ++char_ptr) + { + if (*char_ptr == c) + return (void *) char_ptr; + } + + return NULL; +} +#ifdef weak_alias +weak_alias (__memchr, BP_SYM (memchr)) +#endif diff --git a/lib/memchr.valgrind b/lib/memchr.valgrind new file mode 100644 index 0000000..60f247e --- /dev/null +++ b/lib/memchr.valgrind @@ -0,0 +1,14 @@ +# Suppress a valgrind message about use of uninitialized memory in memchr(). +# POSIX states that when the character is found, memchr must not read extra +# bytes in an overestimated length (for example, where memchr is used to +# implement strnlen). However, we use a safe word read to provide a speedup. +{ + memchr-value4 + Memcheck:Value4 + fun:rpl_memchr +} +{ + memchr-value8 + Memcheck:Value8 + fun:rpl_memchr +} diff --git a/lib/memrchr.c b/lib/memrchr.c new file mode 100644 index 0000000..2efc7cb --- /dev/null +++ b/lib/memrchr.c @@ -0,0 +1,161 @@ +/* memrchr -- find the last occurrence of a byte in a memory block + + Copyright (C) 1991, 1993, 1996-1997, 1999-2000, 2003-2018 Free Software + Foundation, Inc. + + Based on strlen implementation by Torbjorn Granlund (tege@sics.se), + with help from Dan Sahlin (dan@sics.se) and + commentary by Jim Blandy (jimb@ai.mit.edu); + adaptation to memchr suggested by Dick Karpinski (dick@cca.ucsf.edu), + and implemented by Roland McGrath (roland@ai.mit.edu). + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +#if defined _LIBC +# include <memcopy.h> +#else +# include <config.h> +# define reg_char char +#endif + +#include <string.h> +#include <limits.h> + +#undef __memrchr +#ifdef _LIBC +# undef memrchr +#endif + +#ifndef weak_alias +# define __memrchr memrchr +#endif + +/* Search no more than N bytes of S for C. */ +void * +__memrchr (void const *s, int c_in, size_t n) +{ + /* On 32-bit hardware, choosing longword to be a 32-bit unsigned + long instead of a 64-bit uintmax_t tends to give better + performance. On 64-bit hardware, unsigned long is generally 64 + bits already. Change this typedef to experiment with + performance. */ + typedef unsigned long int longword; + + const unsigned char *char_ptr; + const longword *longword_ptr; + longword repeated_one; + longword repeated_c; + unsigned reg_char c; + + c = (unsigned char) c_in; + + /* Handle the last few bytes by reading one byte at a time. + Do this until CHAR_PTR is aligned on a longword boundary. */ + for (char_ptr = (const unsigned char *) s + n; + n > 0 && (size_t) char_ptr % sizeof (longword) != 0; + --n) + if (*--char_ptr == c) + return (void *) char_ptr; + + longword_ptr = (const void *) char_ptr; + + /* All these elucidatory comments refer to 4-byte longwords, + but the theory applies equally well to any size longwords. */ + + /* Compute auxiliary longword values: + repeated_one is a value which has a 1 in every byte. + repeated_c has c in every byte. */ + repeated_one = 0x01010101; + repeated_c = c | (c << 8); + repeated_c |= repeated_c << 16; + if (0xffffffffU < (longword) -1) + { + repeated_one |= repeated_one << 31 << 1; + repeated_c |= repeated_c << 31 << 1; + if (8 < sizeof (longword)) + { + size_t i; + + for (i = 64; i < sizeof (longword) * 8; i *= 2) + { + repeated_one |= repeated_one << i; + repeated_c |= repeated_c << i; + } + } + } + + /* Instead of the traditional loop which tests each byte, we will test a + longword at a time. The tricky part is testing if *any of the four* + bytes in the longword in question are equal to c. We first use an xor + with repeated_c. This reduces the task to testing whether *any of the + four* bytes in longword1 is zero. + + We compute tmp = + ((longword1 - repeated_one) & ~longword1) & (repeated_one << 7). + That is, we perform the following operations: + 1. Subtract repeated_one. + 2. & ~longword1. + 3. & a mask consisting of 0x80 in every byte. + Consider what happens in each byte: + - If a byte of longword1 is zero, step 1 and 2 transform it into 0xff, + and step 3 transforms it into 0x80. A carry can also be propagated + to more significant bytes. + - If a byte of longword1 is nonzero, let its lowest 1 bit be at + position k (0 <= k <= 7); so the lowest k bits are 0. After step 1, + the byte ends in a single bit of value 0 and k bits of value 1. + After step 2, the result is just k bits of value 1: 2^k - 1. After + step 3, the result is 0. And no carry is produced. + So, if longword1 has only non-zero bytes, tmp is zero. + Whereas if longword1 has a zero byte, call j the position of the least + significant zero byte. Then the result has a zero at positions 0, ..., + j-1 and a 0x80 at position j. We cannot predict the result at the more + significant bytes (positions j+1..3), but it does not matter since we + already have a non-zero bit at position 8*j+7. + + So, the test whether any byte in longword1 is zero is equivalent to + testing whether tmp is nonzero. */ + + while (n >= sizeof (longword)) + { + longword longword1 = *--longword_ptr ^ repeated_c; + + if ((((longword1 - repeated_one) & ~longword1) + & (repeated_one << 7)) != 0) + { + longword_ptr++; + break; + } + n -= sizeof (longword); + } + + char_ptr = (const unsigned char *) longword_ptr; + + /* At this point, we know that either n < sizeof (longword), or one of the + sizeof (longword) bytes starting at char_ptr is == c. On little-endian + machines, we could determine the first such byte without any further + memory accesses, just by looking at the tmp result from the last loop + iteration. But this does not work on big-endian machines. Choose code + that works in both cases. */ + + while (n-- > 0) + { + if (*--char_ptr == c) + return (void *) char_ptr; + } + + return NULL; +} +#ifdef weak_alias +weak_alias (__memrchr, memrchr) +#endif diff --git a/lib/minmax.h b/lib/minmax.h new file mode 100644 index 0000000..33a5305 --- /dev/null +++ b/lib/minmax.h @@ -0,0 +1,60 @@ +/* MIN, MAX macros. + Copyright (C) 1995, 1998, 2001, 2003, 2005, 2009-2018 Free Software + Foundation, Inc. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3, or (at your option) + any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, see <https://www.gnu.org/licenses/>. */ + +#ifndef _MINMAX_H +#define _MINMAX_H + +/* Note: MIN, MAX are also defined in <sys/param.h> on some systems + (glibc, IRIX, HP-UX, OSF/1). Therefore you might get warnings about + MIN, MAX macro redefinitions on some systems; the workaround is to + #include this file as the last one among the #include list. */ + +/* Before we define the following symbols we get the <limits.h> file + since otherwise we get redefinitions on some systems if <limits.h> is + included after this file. Likewise for <sys/param.h>. + If more than one of these system headers define MIN and MAX, pick just + one of the headers (because the definitions most likely are the same). */ +#if HAVE_MINMAX_IN_LIMITS_H +# include <limits.h> +#elif HAVE_MINMAX_IN_SYS_PARAM_H +# include <sys/param.h> +#endif + +/* Note: MIN and MAX should be used with two arguments of the + same type. They might not return the minimum and maximum of their two + arguments, if the arguments have different types or have unusual + floating-point values. For example, on a typical host with 32-bit 'int', + 64-bit 'long long', and 64-bit IEEE 754 'double' types: + + MAX (-1, 2147483648) returns 4294967295. + MAX (9007199254740992.0, 9007199254740993) returns 9007199254740992.0. + MAX (NaN, 0.0) returns 0.0. + MAX (+0.0, -0.0) returns -0.0. + + and in each case the answer is in some sense bogus. */ + +/* MAX(a,b) returns the maximum of A and B. */ +#ifndef MAX +# define MAX(a,b) ((a) > (b) ? (a) : (b)) +#endif + +/* MIN(a,b) returns the minimum of A and B. */ +#ifndef MIN +# define MIN(a,b) ((a) < (b) ? (a) : (b)) +#endif + +#endif /* _MINMAX_H */ diff --git a/lib/mkdir.c b/lib/mkdir.c new file mode 100644 index 0000000..37d5a0f --- /dev/null +++ b/lib/mkdir.c @@ -0,0 +1,93 @@ +/* On some systems, mkdir ("foo/", 0700) fails because of the trailing + slash. On those systems, this wrapper removes the trailing slash. + + Copyright (C) 2001, 2003, 2006, 2008-2018 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +/* written by Jim Meyering */ + +#include <config.h> + +/* Specification. */ +#include <sys/stat.h> + +#include <errno.h> +#include <stdio.h> +#include <stdlib.h> +#include <string.h> + +#include "dirname.h" + +/* Disable the definition of mkdir to rpl_mkdir (from the <sys/stat.h> + substitute) in this file. Otherwise, we'd get an endless recursion. */ +#undef mkdir + +/* mingw's _mkdir() function has 1 argument, but we pass 2 arguments. + Additionally, it declares _mkdir (and depending on compile flags, an + alias mkdir), only in the nonstandard includes <direct.h> and <io.h>, + which are included in the <sys/stat.h> override. */ +#if defined _WIN32 && ! defined __CYGWIN__ +# define mkdir(name,mode) _mkdir (name) +# define maybe_unused _GL_UNUSED +#else +# define maybe_unused /* empty */ +#endif + +/* This function is required at least for NetBSD 1.5.2. */ + +int +rpl_mkdir (char const *dir, mode_t mode maybe_unused) +{ + int ret_val; + char *tmp_dir; + size_t len = strlen (dir); + + if (len && dir[len - 1] == '/') + { + tmp_dir = strdup (dir); + if (!tmp_dir) + { + /* Rather than rely on strdup-posix, we set errno ourselves. */ + errno = ENOMEM; + return -1; + } + strip_trailing_slashes (tmp_dir); + } + else + { + tmp_dir = (char *) dir; + } +#if FUNC_MKDIR_DOT_BUG + /* Additionally, cygwin 1.5 mistakenly creates a directory "d/./". */ + { + char *last = last_component (tmp_dir); + if (*last == '.' && (last[1] == '\0' + || (last[1] == '.' && last[2] == '\0'))) + { + struct stat st; + if (stat (tmp_dir, &st) == 0) + errno = EEXIST; + return -1; + } + } +#endif /* FUNC_MKDIR_DOT_BUG */ + + ret_val = mkdir (tmp_dir, mode); + + if (tmp_dir != dir) + free (tmp_dir); + + return ret_val; +} diff --git a/lib/mkostemp.c b/lib/mkostemp.c new file mode 100644 index 0000000..df9ecf8 --- /dev/null +++ b/lib/mkostemp.c @@ -0,0 +1,46 @@ +/* Copyright (C) 1998-1999, 2001, 2005-2007, 2009-2018 Free Software + Foundation, Inc. + This file is derived from the one in the GNU C Library. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +#if !_LIBC +# include <config.h> +#endif + +#include <stdlib.h> + +#if !_LIBC +# include "tempname.h" +# define __gen_tempname gen_tempname +# ifndef __GTFILE +# define __GT_FILE GT_FILE +# endif +#endif + +#include <stdio.h> + +#ifndef __GT_FILE +# define __GT_FILE 0 +#endif + +/* Generate a unique temporary file name from XTEMPLATE. + The last six characters of XTEMPLATE must be "XXXXXX"; + they are replaced with a string that makes the file name unique. + Then open the file and return a fd. */ +int +mkostemp (char *xtemplate, int flags) +{ + return __gen_tempname (xtemplate, 0, flags, __GT_FILE); +} diff --git a/lib/msvc-inval.c b/lib/msvc-inval.c new file mode 100644 index 0000000..0f19a09 --- /dev/null +++ b/lib/msvc-inval.c @@ -0,0 +1,129 @@ +/* Invalid parameter handler for MSVC runtime libraries. + Copyright (C) 2011-2018 Free Software Foundation, Inc. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3, or (at your option) + any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License along + with this program; if not, see <https://www.gnu.org/licenses/>. */ + +#include <config.h> + +/* Specification. */ +#include "msvc-inval.h" + +#if HAVE_MSVC_INVALID_PARAMETER_HANDLER \ + && !(MSVC_INVALID_PARAMETER_HANDLING == SANE_LIBRARY_HANDLING) + +/* Get _invalid_parameter_handler type and _set_invalid_parameter_handler + declaration. */ +# include <stdlib.h> + +# if MSVC_INVALID_PARAMETER_HANDLING == DEFAULT_HANDLING + +static void __cdecl +gl_msvc_invalid_parameter_handler (const wchar_t *expression, + const wchar_t *function, + const wchar_t *file, + unsigned int line, + uintptr_t dummy) +{ +} + +# else + +/* Get declarations of the native Windows API functions. */ +# define WIN32_LEAN_AND_MEAN +# include <windows.h> + +# if defined _MSC_VER + +static void __cdecl +gl_msvc_invalid_parameter_handler (const wchar_t *expression, + const wchar_t *function, + const wchar_t *file, + unsigned int line, + uintptr_t dummy) +{ + RaiseException (STATUS_GNULIB_INVALID_PARAMETER, 0, 0, NULL); +} + +# else + +/* An index to thread-local storage. */ +static DWORD tls_index; +static int tls_initialized /* = 0 */; + +/* Used as a fallback only. */ +static struct gl_msvc_inval_per_thread not_per_thread; + +struct gl_msvc_inval_per_thread * +gl_msvc_inval_current (void) +{ + if (!tls_initialized) + { + tls_index = TlsAlloc (); + tls_initialized = 1; + } + if (tls_index == TLS_OUT_OF_INDEXES) + /* TlsAlloc had failed. */ + return ¬_per_thread; + else + { + struct gl_msvc_inval_per_thread *pointer = + (struct gl_msvc_inval_per_thread *) TlsGetValue (tls_index); + if (pointer == NULL) + { + /* First call. Allocate a new 'struct gl_msvc_inval_per_thread'. */ + pointer = + (struct gl_msvc_inval_per_thread *) + malloc (sizeof (struct gl_msvc_inval_per_thread)); + if (pointer == NULL) + /* Could not allocate memory. Use the global storage. */ + pointer = ¬_per_thread; + TlsSetValue (tls_index, pointer); + } + return pointer; + } +} + +static void __cdecl +gl_msvc_invalid_parameter_handler (const wchar_t *expression, + const wchar_t *function, + const wchar_t *file, + unsigned int line, + uintptr_t dummy) +{ + struct gl_msvc_inval_per_thread *current = gl_msvc_inval_current (); + if (current->restart_valid) + longjmp (current->restart, 1); + else + /* An invalid parameter notification from outside the gnulib code. + Give the caller a chance to intervene. */ + RaiseException (STATUS_GNULIB_INVALID_PARAMETER, 0, 0, NULL); +} + +# endif + +# endif + +static int gl_msvc_inval_initialized /* = 0 */; + +void +gl_msvc_inval_ensure_handler (void) +{ + if (gl_msvc_inval_initialized == 0) + { + _set_invalid_parameter_handler (gl_msvc_invalid_parameter_handler); + gl_msvc_inval_initialized = 1; + } +} + +#endif diff --git a/lib/msvc-inval.h b/lib/msvc-inval.h new file mode 100644 index 0000000..8e7ff30 --- /dev/null +++ b/lib/msvc-inval.h @@ -0,0 +1,222 @@ +/* Invalid parameter handler for MSVC runtime libraries. + Copyright (C) 2011-2018 Free Software Foundation, Inc. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3, or (at your option) + any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License along + with this program; if not, see <https://www.gnu.org/licenses/>. */ + +#ifndef _MSVC_INVAL_H +#define _MSVC_INVAL_H + +/* With MSVC runtime libraries with the "invalid parameter handler" concept, + functions like fprintf(), dup2(), or close() crash when the caller passes + an invalid argument. But POSIX wants error codes (such as EINVAL or EBADF) + instead. + This file defines macros that turn such an invalid parameter notification + into a non-local exit. An error code can then be produced at the target + of this exit. You can thus write code like + + TRY_MSVC_INVAL + { + <Code that can trigger an invalid parameter notification + but does not do 'return', 'break', 'continue', nor 'goto'.> + } + CATCH_MSVC_INVAL + { + <Code that handles an invalid parameter notification + but does not do 'return', 'break', 'continue', nor 'goto'.> + } + DONE_MSVC_INVAL; + + This entire block expands to a single statement. + + The handling of invalid parameters can be done in three ways: + + * The default way, which is reasonable for programs (not libraries): + AC_DEFINE([MSVC_INVALID_PARAMETER_HANDLING], [DEFAULT_HANDLING]) + + * The way for libraries that make "hairy" calls (like close(-1), or + fclose(fp) where fileno(fp) is closed, or simply getdtablesize()): + AC_DEFINE([MSVC_INVALID_PARAMETER_HANDLING], [HAIRY_LIBRARY_HANDLING]) + + * The way for libraries that make no "hairy" calls: + AC_DEFINE([MSVC_INVALID_PARAMETER_HANDLING], [SANE_LIBRARY_HANDLING]) + */ + +#define DEFAULT_HANDLING 0 +#define HAIRY_LIBRARY_HANDLING 1 +#define SANE_LIBRARY_HANDLING 2 + +#if HAVE_MSVC_INVALID_PARAMETER_HANDLER \ + && !(MSVC_INVALID_PARAMETER_HANDLING == SANE_LIBRARY_HANDLING) +/* A native Windows platform with the "invalid parameter handler" concept, + and either DEFAULT_HANDLING or HAIRY_LIBRARY_HANDLING. */ + +# if MSVC_INVALID_PARAMETER_HANDLING == DEFAULT_HANDLING +/* Default handling. */ + +# ifdef __cplusplus +extern "C" { +# endif + +/* Ensure that the invalid parameter handler in installed that just returns. + Because we assume no other part of the program installs a different + invalid parameter handler, this solution is multithread-safe. */ +extern void gl_msvc_inval_ensure_handler (void); + +# ifdef __cplusplus +} +# endif + +# define TRY_MSVC_INVAL \ + do \ + { \ + gl_msvc_inval_ensure_handler (); \ + if (1) +# define CATCH_MSVC_INVAL \ + else +# define DONE_MSVC_INVAL \ + } \ + while (0) + +# else +/* Handling for hairy libraries. */ + +# include <excpt.h> + +/* Gnulib can define its own status codes, as described in the page + "Raising Software Exceptions" on microsoft.com + <https://msdn.microsoft.com/en-us/library/het71c37.aspx>. + Our status codes are composed of + - 0xE0000000, mandatory for all user-defined status codes, + - 0x474E550, a API identifier ("GNU"), + - 0, 1, 2, ..., used to distinguish different status codes from the + same API. */ +# define STATUS_GNULIB_INVALID_PARAMETER (0xE0000000 + 0x474E550 + 0) + +# if defined _MSC_VER +/* A compiler that supports __try/__except, as described in the page + "try-except statement" on microsoft.com + <https://msdn.microsoft.com/en-us/library/s58ftw19.aspx>. + With __try/__except, we can use the multithread-safe exception handling. */ + +# ifdef __cplusplus +extern "C" { +# endif + +/* Ensure that the invalid parameter handler in installed that raises a + software exception with code STATUS_GNULIB_INVALID_PARAMETER. + Because we assume no other part of the program installs a different + invalid parameter handler, this solution is multithread-safe. */ +extern void gl_msvc_inval_ensure_handler (void); + +# ifdef __cplusplus +} +# endif + +# define TRY_MSVC_INVAL \ + do \ + { \ + gl_msvc_inval_ensure_handler (); \ + __try +# define CATCH_MSVC_INVAL \ + __except (GetExceptionCode () == STATUS_GNULIB_INVALID_PARAMETER \ + ? EXCEPTION_EXECUTE_HANDLER \ + : EXCEPTION_CONTINUE_SEARCH) +# define DONE_MSVC_INVAL \ + } \ + while (0) + +# else +/* Any compiler. + We can only use setjmp/longjmp. */ + +# include <setjmp.h> + +# ifdef __cplusplus +extern "C" { +# endif + +struct gl_msvc_inval_per_thread +{ + /* The restart that will resume execution at the code between + CATCH_MSVC_INVAL and DONE_MSVC_INVAL. It is enabled only between + TRY_MSVC_INVAL and CATCH_MSVC_INVAL. */ + jmp_buf restart; + + /* Tells whether the contents of restart is valid. */ + int restart_valid; +}; + +/* Ensure that the invalid parameter handler in installed that passes + control to the gl_msvc_inval_restart if it is valid, or raises a + software exception with code STATUS_GNULIB_INVALID_PARAMETER otherwise. + Because we assume no other part of the program installs a different + invalid parameter handler, this solution is multithread-safe. */ +extern void gl_msvc_inval_ensure_handler (void); + +/* Return a pointer to the per-thread data for the current thread. */ +extern struct gl_msvc_inval_per_thread *gl_msvc_inval_current (void); + +# ifdef __cplusplus +} +# endif + +# define TRY_MSVC_INVAL \ + do \ + { \ + struct gl_msvc_inval_per_thread *msvc_inval_current; \ + gl_msvc_inval_ensure_handler (); \ + msvc_inval_current = gl_msvc_inval_current (); \ + /* First, initialize gl_msvc_inval_restart. */ \ + if (setjmp (msvc_inval_current->restart) == 0) \ + { \ + /* Then, mark it as valid. */ \ + msvc_inval_current->restart_valid = 1; +# define CATCH_MSVC_INVAL \ + /* Execution completed. \ + Mark gl_msvc_inval_restart as invalid. */ \ + msvc_inval_current->restart_valid = 0; \ + } \ + else \ + { \ + /* Execution triggered an invalid parameter notification. \ + Mark gl_msvc_inval_restart as invalid. */ \ + msvc_inval_current->restart_valid = 0; +# define DONE_MSVC_INVAL \ + } \ + } \ + while (0) + +# endif + +# endif + +#else +/* A platform that does not need to the invalid parameter handler, + or when SANE_LIBRARY_HANDLING is desired. */ + +/* The braces here avoid GCC warnings like + "warning: suggest explicit braces to avoid ambiguous 'else'". */ +# define TRY_MSVC_INVAL \ + do \ + { \ + if (1) +# define CATCH_MSVC_INVAL \ + else +# define DONE_MSVC_INVAL \ + } \ + while (0) + +#endif + +#endif /* _MSVC_INVAL_H */ diff --git a/lib/msvc-nothrow.c b/lib/msvc-nothrow.c new file mode 100644 index 0000000..785733e --- /dev/null +++ b/lib/msvc-nothrow.c @@ -0,0 +1,51 @@ +/* Wrappers that don't throw invalid parameter notifications + with MSVC runtime libraries. + Copyright (C) 2011-2018 Free Software Foundation, Inc. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3, or (at your option) + any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License along + with this program; if not, see <https://www.gnu.org/licenses/>. */ + +#include <config.h> + +/* Specification. */ +#include "msvc-nothrow.h" + +/* Get declarations of the native Windows API functions. */ +#define WIN32_LEAN_AND_MEAN +#include <windows.h> + +#if HAVE_MSVC_INVALID_PARAMETER_HANDLER +# include "msvc-inval.h" +#endif + +#undef _get_osfhandle + +#if HAVE_MSVC_INVALID_PARAMETER_HANDLER +intptr_t +_gl_nothrow_get_osfhandle (int fd) +{ + intptr_t result; + + TRY_MSVC_INVAL + { + result = _get_osfhandle (fd); + } + CATCH_MSVC_INVAL + { + result = (intptr_t) INVALID_HANDLE_VALUE; + } + DONE_MSVC_INVAL; + + return result; +} +#endif diff --git a/lib/msvc-nothrow.h b/lib/msvc-nothrow.h new file mode 100644 index 0000000..a9671c3 --- /dev/null +++ b/lib/msvc-nothrow.h @@ -0,0 +1,43 @@ +/* Wrappers that don't throw invalid parameter notifications + with MSVC runtime libraries. + Copyright (C) 2011-2018 Free Software Foundation, Inc. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3, or (at your option) + any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License along + with this program; if not, see <https://www.gnu.org/licenses/>. */ + +#ifndef _MSVC_NOTHROW_H +#define _MSVC_NOTHROW_H + +/* With MSVC runtime libraries with the "invalid parameter handler" concept, + functions like fprintf(), dup2(), or close() crash when the caller passes + an invalid argument. But POSIX wants error codes (such as EINVAL or EBADF) + instead. + This file defines wrappers that turn such an invalid parameter notification + into an error code. */ + +#if defined _WIN32 && ! defined __CYGWIN__ + +/* Get original declaration of _get_osfhandle. */ +# include <io.h> + +# if HAVE_MSVC_INVALID_PARAMETER_HANDLER + +/* Override _get_osfhandle. */ +extern intptr_t _gl_nothrow_get_osfhandle (int fd); +# define _get_osfhandle _gl_nothrow_get_osfhandle + +# endif + +#endif + +#endif /* _MSVC_NOTHROW_H */ diff --git a/lib/nl_langinfo.c b/lib/nl_langinfo.c new file mode 100644 index 0000000..ea26730 --- /dev/null +++ b/lib/nl_langinfo.c @@ -0,0 +1,366 @@ +/* nl_langinfo() replacement: query locale dependent information. + + Copyright (C) 2007-2018 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +#include <config.h> + +/* Specification. */ +#include <langinfo.h> + +#include <locale.h> +#include <string.h> +#if defined _WIN32 && ! defined __CYGWIN__ +# define WIN32_LEAN_AND_MEAN /* avoid including junk */ +# include <windows.h> +# include <stdio.h> +#endif + +#if !REPLACE_NL_LANGINFO || GNULIB_defined_CODESET +/* Return the codeset of the current locale, if this is easily deducible. + Otherwise, return "". */ +static char * +ctype_codeset (void) +{ + static char buf[2 + 10 + 1]; + char const *locale = setlocale (LC_CTYPE, NULL); + char *codeset = buf; + size_t codesetlen; + codeset[0] = '\0'; + + if (locale && locale[0]) + { + /* If the locale name contains an encoding after the dot, return it. */ + char *dot = strchr (locale, '.'); + + if (dot) + { + /* Look for the possible @... trailer and remove it, if any. */ + char *codeset_start = dot + 1; + char const *modifier = strchr (codeset_start, '@'); + + if (! modifier) + codeset = codeset_start; + else + { + codesetlen = modifier - codeset_start; + if (codesetlen < sizeof buf) + { + codeset = memcpy (buf, codeset_start, codesetlen); + codeset[codesetlen] = '\0'; + } + } + } + } + +# if defined _WIN32 && ! defined __CYGWIN__ + /* If setlocale is successful, it returns the number of the + codepage, as a string. Otherwise, fall back on Windows API + GetACP, which returns the locale's codepage as a number (although + this doesn't change according to what the 'setlocale' call specified). + Either way, prepend "CP" to make it a valid codeset name. */ + codesetlen = strlen (codeset); + if (0 < codesetlen && codesetlen < sizeof buf - 2) + memmove (buf + 2, codeset, codesetlen + 1); + else + sprintf (buf + 2, "%u", GetACP ()); + codeset = memcpy (buf, "CP", 2); +# endif + return codeset; +} +#endif + + +#if REPLACE_NL_LANGINFO + +/* Override nl_langinfo with support for added nl_item values. */ + +# undef nl_langinfo + +char * +rpl_nl_langinfo (nl_item item) +{ + switch (item) + { +# if GNULIB_defined_CODESET + case CODESET: + return ctype_codeset (); +# endif +# if GNULIB_defined_T_FMT_AMPM + case T_FMT_AMPM: + return (char *) "%I:%M:%S %p"; +# endif +# if GNULIB_defined_ALTMON + case ALTMON_1: + case ALTMON_2: + case ALTMON_3: + case ALTMON_4: + case ALTMON_5: + case ALTMON_6: + case ALTMON_7: + case ALTMON_8: + case ALTMON_9: + case ALTMON_10: + case ALTMON_11: + case ALTMON_12: + /* We don't ship the appropriate localizations with gnulib. Therefore, + treat ALTMON_i like MON_i. */ + item = item - ALTMON_1 + MON_1; + break; +# endif +# if GNULIB_defined_ERA + case ERA: + /* The format is not standardized. In glibc it is a sequence of strings + of the form "direction:offset:start_date:end_date:era_name:era_format" + with an empty string at the end. */ + return (char *) ""; + case ERA_D_FMT: + /* The %Ex conversion in strftime behaves like %x if the locale does not + have an alternative time format. */ + item = D_FMT; + break; + case ERA_D_T_FMT: + /* The %Ec conversion in strftime behaves like %c if the locale does not + have an alternative time format. */ + item = D_T_FMT; + break; + case ERA_T_FMT: + /* The %EX conversion in strftime behaves like %X if the locale does not + have an alternative time format. */ + item = T_FMT; + break; + case ALT_DIGITS: + /* The format is not standardized. In glibc it is a sequence of 10 + strings, appended in memory. */ + return (char *) "\0\0\0\0\0\0\0\0\0\0"; +# endif +# if GNULIB_defined_YESEXPR || !FUNC_NL_LANGINFO_YESEXPR_WORKS + case YESEXPR: + return (char *) "^[yY]"; + case NOEXPR: + return (char *) "^[nN]"; +# endif + default: + break; + } + return nl_langinfo (item); +} + +#else + +/* Provide nl_langinfo from scratch, either for native MS-Windows, or + for old Unix platforms without locales, such as Linux libc5 or + BeOS. */ + +# include <time.h> + +char * +nl_langinfo (nl_item item) +{ + static char nlbuf[100]; + struct tm tmm = { 0 }; + + switch (item) + { + /* nl_langinfo items of the LC_CTYPE category */ + case CODESET: + { + char *codeset = ctype_codeset (); + if (*codeset) + return codeset; + } +# ifdef __BEOS__ + return (char *) "UTF-8"; +# else + return (char *) "ISO-8859-1"; +# endif + /* nl_langinfo items of the LC_NUMERIC category */ + case RADIXCHAR: + return localeconv () ->decimal_point; + case THOUSEP: + return localeconv () ->thousands_sep; +# ifdef GROUPING + case GROUPING: + return localeconv () ->grouping; +# endif + /* nl_langinfo items of the LC_TIME category. + TODO: Really use the locale. */ + case D_T_FMT: + case ERA_D_T_FMT: + return (char *) "%a %b %e %H:%M:%S %Y"; + case D_FMT: + case ERA_D_FMT: + return (char *) "%m/%d/%y"; + case T_FMT: + case ERA_T_FMT: + return (char *) "%H:%M:%S"; + case T_FMT_AMPM: + return (char *) "%I:%M:%S %p"; + case AM_STR: + if (!strftime (nlbuf, sizeof nlbuf, "%p", &tmm)) + return (char *) "AM"; + return nlbuf; + case PM_STR: + tmm.tm_hour = 12; + if (!strftime (nlbuf, sizeof nlbuf, "%p", &tmm)) + return (char *) "PM"; + return nlbuf; + case DAY_1: + case DAY_2: + case DAY_3: + case DAY_4: + case DAY_5: + case DAY_6: + case DAY_7: + { + static char const days[][sizeof "Wednesday"] = { + "Sunday", "Monday", "Tuesday", "Wednesday", "Thursday", + "Friday", "Saturday" + }; + tmm.tm_wday = item - DAY_1; + if (!strftime (nlbuf, sizeof nlbuf, "%A", &tmm)) + return (char *) days[item - DAY_1]; + return nlbuf; + } + case ABDAY_1: + case ABDAY_2: + case ABDAY_3: + case ABDAY_4: + case ABDAY_5: + case ABDAY_6: + case ABDAY_7: + { + static char const abdays[][sizeof "Sun"] = { + "Sun", "Mon", "Tue", "Wed", "Thu", "Fri", "Sat" + }; + tmm.tm_wday = item - ABDAY_1; + if (!strftime (nlbuf, sizeof nlbuf, "%a", &tmm)) + return (char *) abdays[item - ABDAY_1]; + return nlbuf; + } + { + static char const months[][sizeof "September"] = { + "January", "February", "March", "April", "May", "June", "July", + "September", "October", "November", "December" + }; + case MON_1: + case MON_2: + case MON_3: + case MON_4: + case MON_5: + case MON_6: + case MON_7: + case MON_8: + case MON_9: + case MON_10: + case MON_11: + case MON_12: + tmm.tm_mon = item - MON_1; + if (!strftime (nlbuf, sizeof nlbuf, "%B", &tmm)) + return (char *) months[item - MON_1]; + return nlbuf; + case ALTMON_1: + case ALTMON_2: + case ALTMON_3: + case ALTMON_4: + case ALTMON_5: + case ALTMON_6: + case ALTMON_7: + case ALTMON_8: + case ALTMON_9: + case ALTMON_10: + case ALTMON_11: + case ALTMON_12: + tmm.tm_mon = item - ALTMON_1; + /* The platforms without nl_langinfo() don't support strftime with %OB. + We don't even need to try. */ + #if 0 + if (!strftime (nlbuf, sizeof nlbuf, "%OB", &tmm)) + #endif + if (!strftime (nlbuf, sizeof nlbuf, "%B", &tmm)) + return (char *) months[item - ALTMON_1]; + return nlbuf; + } + case ABMON_1: + case ABMON_2: + case ABMON_3: + case ABMON_4: + case ABMON_5: + case ABMON_6: + case ABMON_7: + case ABMON_8: + case ABMON_9: + case ABMON_10: + case ABMON_11: + case ABMON_12: + { + static char const abmonths[][sizeof "Jan"] = { + "Jan", "Feb", "Mar", "Apr", "May", "Jun", "Jul", + "Sep", "Oct", "Nov", "Dec" + }; + tmm.tm_mon = item - ABMON_1; + if (!strftime (nlbuf, sizeof nlbuf, "%b", &tmm)) + return (char *) abmonths[item - ABMON_1]; + return nlbuf; + } + case ERA: + return (char *) ""; + case ALT_DIGITS: + return (char *) "\0\0\0\0\0\0\0\0\0\0"; + /* nl_langinfo items of the LC_MONETARY category. */ + case CRNCYSTR: + return localeconv () ->currency_symbol; +# ifdef INT_CURR_SYMBOL + case INT_CURR_SYMBOL: + return localeconv () ->int_curr_symbol; + case MON_DECIMAL_POINT: + return localeconv () ->mon_decimal_point; + case MON_THOUSANDS_SEP: + return localeconv () ->mon_thousands_sep; + case MON_GROUPING: + return localeconv () ->mon_grouping; + case POSITIVE_SIGN: + return localeconv () ->positive_sign; + case NEGATIVE_SIGN: + return localeconv () ->negative_sign; + case FRAC_DIGITS: + return & localeconv () ->frac_digits; + case INT_FRAC_DIGITS: + return & localeconv () ->int_frac_digits; + case P_CS_PRECEDES: + return & localeconv () ->p_cs_precedes; + case N_CS_PRECEDES: + return & localeconv () ->n_cs_precedes; + case P_SEP_BY_SPACE: + return & localeconv () ->p_sep_by_space; + case N_SEP_BY_SPACE: + return & localeconv () ->n_sep_by_space; + case P_SIGN_POSN: + return & localeconv () ->p_sign_posn; + case N_SIGN_POSN: + return & localeconv () ->n_sign_posn; +# endif + /* nl_langinfo items of the LC_MESSAGES category + TODO: Really use the locale. */ + case YESEXPR: + return (char *) "^[yY]"; + case NOEXPR: + return (char *) "^[nN]"; + default: + return (char *) ""; + } +} + +#endif diff --git a/lib/obstack.c b/lib/obstack.c new file mode 100644 index 0000000..02be6cf --- /dev/null +++ b/lib/obstack.c @@ -0,0 +1,354 @@ +/* obstack.c - subroutines used implicitly by object stack macros + Copyright (C) 1988-2018 Free Software Foundation, Inc. + This file is part of the GNU C Library. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU General Public + License as published by the Free Software Foundation; either + version 3 of the License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + General Public License for more details. + + You should have received a copy of the GNU General Public + License along with the GNU C Library; if not, see + <https://www.gnu.org/licenses/>. */ + + +#ifdef _LIBC +# include <obstack.h> +#else +# include <config.h> +# include "obstack.h" +#endif + +/* NOTE BEFORE MODIFYING THIS FILE: _OBSTACK_INTERFACE_VERSION in + obstack.h must be incremented whenever callers compiled using an old + obstack.h can no longer properly call the functions in this file. */ + +/* Comment out all this code if we are using the GNU C Library, and are not + actually compiling the library itself, and the installed library + supports the same library interface we do. This code is part of the GNU + C Library, but also included in many other GNU distributions. Compiling + and linking in this code is a waste when using the GNU C library + (especially if it is a shared library). Rather than having every GNU + program understand 'configure --with-gnu-libc' and omit the object + files, it is simpler to just do this in the source for each such file. */ +#if !defined _LIBC && defined __GNU_LIBRARY__ && __GNU_LIBRARY__ > 1 +# include <gnu-versions.h> +# if (_GNU_OBSTACK_INTERFACE_VERSION == _OBSTACK_INTERFACE_VERSION \ + || (_GNU_OBSTACK_INTERFACE_VERSION == 1 \ + && _OBSTACK_INTERFACE_VERSION == 2 \ + && defined SIZEOF_INT && defined SIZEOF_SIZE_T \ + && SIZEOF_INT == SIZEOF_SIZE_T)) +# define _OBSTACK_ELIDE_CODE +# endif +#endif + +#ifndef _OBSTACK_ELIDE_CODE +/* If GCC, or if an oddball (testing?) host that #defines __alignof__, + use the already-supplied __alignof__. Otherwise, this must be Gnulib + (as glibc assumes GCC); defer to Gnulib's alignof_type. */ +# if !defined __GNUC__ && !defined __alignof__ +# include <alignof.h> +# define __alignof__(type) alignof_type (type) +# endif +# include <stdlib.h> +# include <stdint.h> + +# ifndef MAX +# define MAX(a,b) ((a) > (b) ? (a) : (b)) +# endif + +/* Determine default alignment. */ + +/* If malloc were really smart, it would round addresses to DEFAULT_ALIGNMENT. + But in fact it might be less smart and round addresses to as much as + DEFAULT_ROUNDING. So we prepare for it to do that. + + DEFAULT_ALIGNMENT cannot be an enum constant; see gnulib's alignof.h. */ +#define DEFAULT_ALIGNMENT MAX (__alignof__ (long double), \ + MAX (__alignof__ (uintmax_t), \ + __alignof__ (void *))) +#define DEFAULT_ROUNDING MAX (sizeof (long double), \ + MAX (sizeof (uintmax_t), \ + sizeof (void *))) + +/* Call functions with either the traditional malloc/free calling + interface, or the mmalloc/mfree interface (that adds an extra first + argument), based on the value of use_extra_arg. */ + +static void * +call_chunkfun (struct obstack *h, size_t size) +{ + if (h->use_extra_arg) + return h->chunkfun.extra (h->extra_arg, size); + else + return h->chunkfun.plain (size); +} + +static void +call_freefun (struct obstack *h, void *old_chunk) +{ + if (h->use_extra_arg) + h->freefun.extra (h->extra_arg, old_chunk); + else + h->freefun.plain (old_chunk); +} + + +/* Initialize an obstack H for use. Specify chunk size SIZE (0 means default). + Objects start on multiples of ALIGNMENT (0 means use default). + + Return nonzero if successful, calls obstack_alloc_failed_handler if + allocation fails. */ + +static int +_obstack_begin_worker (struct obstack *h, + _OBSTACK_SIZE_T size, _OBSTACK_SIZE_T alignment) +{ + struct _obstack_chunk *chunk; /* points to new chunk */ + + if (alignment == 0) + alignment = DEFAULT_ALIGNMENT; + if (size == 0) + /* Default size is what GNU malloc can fit in a 4096-byte block. */ + { + /* 12 is sizeof (mhead) and 4 is EXTRA from GNU malloc. + Use the values for range checking, because if range checking is off, + the extra bytes won't be missed terribly, but if range checking is on + and we used a larger request, a whole extra 4096 bytes would be + allocated. + + These number are irrelevant to the new GNU malloc. I suspect it is + less sensitive to the size of the request. */ + int extra = ((((12 + DEFAULT_ROUNDING - 1) & ~(DEFAULT_ROUNDING - 1)) + + 4 + DEFAULT_ROUNDING - 1) + & ~(DEFAULT_ROUNDING - 1)); + size = 4096 - extra; + } + + h->chunk_size = size; + h->alignment_mask = alignment - 1; + + chunk = h->chunk = call_chunkfun (h, h->chunk_size); + if (!chunk) + (*obstack_alloc_failed_handler) (); + h->next_free = h->object_base = __PTR_ALIGN ((char *) chunk, chunk->contents, + alignment - 1); + h->chunk_limit = chunk->limit = (char *) chunk + h->chunk_size; + chunk->prev = 0; + /* The initial chunk now contains no empty object. */ + h->maybe_empty_object = 0; + h->alloc_failed = 0; + return 1; +} + +int +_obstack_begin (struct obstack *h, + _OBSTACK_SIZE_T size, _OBSTACK_SIZE_T alignment, + void *(*chunkfun) (size_t), + void (*freefun) (void *)) +{ + h->chunkfun.plain = chunkfun; + h->freefun.plain = freefun; + h->use_extra_arg = 0; + return _obstack_begin_worker (h, size, alignment); +} + +int +_obstack_begin_1 (struct obstack *h, + _OBSTACK_SIZE_T size, _OBSTACK_SIZE_T alignment, + void *(*chunkfun) (void *, size_t), + void (*freefun) (void *, void *), + void *arg) +{ + h->chunkfun.extra = chunkfun; + h->freefun.extra = freefun; + h->extra_arg = arg; + h->use_extra_arg = 1; + return _obstack_begin_worker (h, size, alignment); +} + +/* Allocate a new current chunk for the obstack *H + on the assumption that LENGTH bytes need to be added + to the current object, or a new object of length LENGTH allocated. + Copies any partial object from the end of the old chunk + to the beginning of the new one. */ + +void +_obstack_newchunk (struct obstack *h, _OBSTACK_SIZE_T length) +{ + struct _obstack_chunk *old_chunk = h->chunk; + struct _obstack_chunk *new_chunk = 0; + size_t obj_size = h->next_free - h->object_base; + char *object_base; + + /* Compute size for new chunk. */ + size_t sum1 = obj_size + length; + size_t sum2 = sum1 + h->alignment_mask; + size_t new_size = sum2 + (obj_size >> 3) + 100; + if (new_size < sum2) + new_size = sum2; + if (new_size < h->chunk_size) + new_size = h->chunk_size; + + /* Allocate and initialize the new chunk. */ + if (obj_size <= sum1 && sum1 <= sum2) + new_chunk = call_chunkfun (h, new_size); + if (!new_chunk) + (*obstack_alloc_failed_handler)(); + h->chunk = new_chunk; + new_chunk->prev = old_chunk; + new_chunk->limit = h->chunk_limit = (char *) new_chunk + new_size; + + /* Compute an aligned object_base in the new chunk */ + object_base = + __PTR_ALIGN ((char *) new_chunk, new_chunk->contents, h->alignment_mask); + + /* Move the existing object to the new chunk. */ + memcpy (object_base, h->object_base, obj_size); + + /* If the object just copied was the only data in OLD_CHUNK, + free that chunk and remove it from the chain. + But not if that chunk might contain an empty object. */ + if (!h->maybe_empty_object + && (h->object_base + == __PTR_ALIGN ((char *) old_chunk, old_chunk->contents, + h->alignment_mask))) + { + new_chunk->prev = old_chunk->prev; + call_freefun (h, old_chunk); + } + + h->object_base = object_base; + h->next_free = h->object_base + obj_size; + /* The new chunk certainly contains no empty object yet. */ + h->maybe_empty_object = 0; +} + +/* Return nonzero if object OBJ has been allocated from obstack H. + This is here for debugging. + If you use it in a program, you are probably losing. */ + +/* Suppress -Wmissing-prototypes warning. We don't want to declare this in + obstack.h because it is just for debugging. */ +int _obstack_allocated_p (struct obstack *h, void *obj) __attribute_pure__; + +int +_obstack_allocated_p (struct obstack *h, void *obj) +{ + struct _obstack_chunk *lp; /* below addr of any objects in this chunk */ + struct _obstack_chunk *plp; /* point to previous chunk if any */ + + lp = (h)->chunk; + /* We use >= rather than > since the object cannot be exactly at + the beginning of the chunk but might be an empty object exactly + at the end of an adjacent chunk. */ + while (lp != 0 && ((void *) lp >= obj || (void *) (lp)->limit < obj)) + { + plp = lp->prev; + lp = plp; + } + return lp != 0; +} + +/* Free objects in obstack H, including OBJ and everything allocate + more recently than OBJ. If OBJ is zero, free everything in H. */ + +void +_obstack_free (struct obstack *h, void *obj) +{ + struct _obstack_chunk *lp; /* below addr of any objects in this chunk */ + struct _obstack_chunk *plp; /* point to previous chunk if any */ + + lp = h->chunk; + /* We use >= because there cannot be an object at the beginning of a chunk. + But there can be an empty object at that address + at the end of another chunk. */ + while (lp != 0 && ((void *) lp >= obj || (void *) (lp)->limit < obj)) + { + plp = lp->prev; + call_freefun (h, lp); + lp = plp; + /* If we switch chunks, we can't tell whether the new current + chunk contains an empty object, so assume that it may. */ + h->maybe_empty_object = 1; + } + if (lp) + { + h->object_base = h->next_free = (char *) (obj); + h->chunk_limit = lp->limit; + h->chunk = lp; + } + else if (obj != 0) + /* obj is not in any of the chunks! */ + abort (); +} + +_OBSTACK_SIZE_T +_obstack_memory_used (struct obstack *h) +{ + struct _obstack_chunk *lp; + _OBSTACK_SIZE_T nbytes = 0; + + for (lp = h->chunk; lp != 0; lp = lp->prev) + { + nbytes += lp->limit - (char *) lp; + } + return nbytes; +} + +# ifndef _OBSTACK_NO_ERROR_HANDLER +/* Define the error handler. */ +# include <stdio.h> + +/* Exit value used when 'print_and_abort' is used. */ +# ifdef _LIBC +int obstack_exit_failure = EXIT_FAILURE; +# else +# include "exitfail.h" +# define obstack_exit_failure exit_failure +# endif + +# ifdef _LIBC +# include <libintl.h> +# else +# include "gettext.h" +# endif +# ifndef _ +# define _(msgid) gettext (msgid) +# endif + +# ifdef _LIBC +# include <libio/iolibio.h> +# endif + +static _Noreturn void +print_and_abort (void) +{ + /* Don't change any of these strings. Yes, it would be possible to add + the newline to the string and use fputs or so. But this must not + happen because the "memory exhausted" message appears in other places + like this and the translation should be reused instead of creating + a very similar string which requires a separate translation. */ +# ifdef _LIBC + (void) __fxprintf (NULL, "%s\n", _("memory exhausted")); +# else + fprintf (stderr, "%s\n", _("memory exhausted")); +# endif + exit (obstack_exit_failure); +} + +/* The functions allocating more room by calling 'obstack_chunk_alloc' + jump to the handler pointed to by 'obstack_alloc_failed_handler'. + This can be set to a user defined function which should either + abort gracefully or use longjump - but shouldn't return. This + variable by default points to the internal function + 'print_and_abort'. */ +__attribute_noreturn__ void (*obstack_alloc_failed_handler) (void) + = print_and_abort; +# endif /* !_OBSTACK_NO_ERROR_HANDLER */ +#endif /* !_OBSTACK_ELIDE_CODE */ diff --git a/lib/obstack.h b/lib/obstack.h new file mode 100644 index 0000000..cff4205 --- /dev/null +++ b/lib/obstack.h @@ -0,0 +1,546 @@ +/* obstack.h - object stack macros + Copyright (C) 1988-2018 Free Software Foundation, Inc. + This file is part of the GNU C Library. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU General Public + License as published by the Free Software Foundation; either + version 3 of the License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + General Public License for more details. + + You should have received a copy of the GNU General Public + License along with the GNU C Library; if not, see + <https://www.gnu.org/licenses/>. */ + +/* Summary: + + All the apparent functions defined here are macros. The idea + is that you would use these pre-tested macros to solve a + very specific set of problems, and they would run fast. + Caution: no side-effects in arguments please!! They may be + evaluated MANY times!! + + These macros operate a stack of objects. Each object starts life + small, and may grow to maturity. (Consider building a word syllable + by syllable.) An object can move while it is growing. Once it has + been "finished" it never changes address again. So the "top of the + stack" is typically an immature growing object, while the rest of the + stack is of mature, fixed size and fixed address objects. + + These routines grab large chunks of memory, using a function you + supply, called 'obstack_chunk_alloc'. On occasion, they free chunks, + by calling 'obstack_chunk_free'. You must define them and declare + them before using any obstack macros. + + Each independent stack is represented by a 'struct obstack'. + Each of the obstack macros expects a pointer to such a structure + as the first argument. + + One motivation for this package is the problem of growing char strings + in symbol tables. Unless you are "fascist pig with a read-only mind" + --Gosper's immortal quote from HAKMEM item 154, out of context--you + would not like to put any arbitrary upper limit on the length of your + symbols. + + In practice this often means you will build many short symbols and a + few long symbols. At the time you are reading a symbol you don't know + how long it is. One traditional method is to read a symbol into a + buffer, realloc()ating the buffer every time you try to read a symbol + that is longer than the buffer. This is beaut, but you still will + want to copy the symbol from the buffer to a more permanent + symbol-table entry say about half the time. + + With obstacks, you can work differently. Use one obstack for all symbol + names. As you read a symbol, grow the name in the obstack gradually. + When the name is complete, finalize it. Then, if the symbol exists already, + free the newly read name. + + The way we do this is to take a large chunk, allocating memory from + low addresses. When you want to build a symbol in the chunk you just + add chars above the current "high water mark" in the chunk. When you + have finished adding chars, because you got to the end of the symbol, + you know how long the chars are, and you can create a new object. + Mostly the chars will not burst over the highest address of the chunk, + because you would typically expect a chunk to be (say) 100 times as + long as an average object. + + In case that isn't clear, when we have enough chars to make up + the object, THEY ARE ALREADY CONTIGUOUS IN THE CHUNK (guaranteed) + so we just point to it where it lies. No moving of chars is + needed and this is the second win: potentially long strings need + never be explicitly shuffled. Once an object is formed, it does not + change its address during its lifetime. + + When the chars burst over a chunk boundary, we allocate a larger + chunk, and then copy the partly formed object from the end of the old + chunk to the beginning of the new larger chunk. We then carry on + accreting characters to the end of the object as we normally would. + + A special macro is provided to add a single char at a time to a + growing object. This allows the use of register variables, which + break the ordinary 'growth' macro. + + Summary: + We allocate large chunks. + We carve out one object at a time from the current chunk. + Once carved, an object never moves. + We are free to append data of any size to the currently + growing object. + Exactly one object is growing in an obstack at any one time. + You can run one obstack per control block. + You may have as many control blocks as you dare. + Because of the way we do it, you can "unwind" an obstack + back to a previous state. (You may remove objects much + as you would with a stack.) + */ + + +/* Don't do the contents of this file more than once. */ + +#ifndef _OBSTACK_H +#define _OBSTACK_H 1 + +#ifndef _OBSTACK_INTERFACE_VERSION +# define _OBSTACK_INTERFACE_VERSION 2 +#endif + +#include <stddef.h> /* For size_t and ptrdiff_t. */ +#include <string.h> /* For __GNU_LIBRARY__, and memcpy. */ + +#if __STDC_VERSION__ < 199901L || defined __HP_cc +# define __FLEXIBLE_ARRAY_MEMBER 1 +#else +# define __FLEXIBLE_ARRAY_MEMBER +#endif + +#if _OBSTACK_INTERFACE_VERSION == 1 +/* For binary compatibility with obstack version 1, which used "int" + and "long" for these two types. */ +# define _OBSTACK_SIZE_T unsigned int +# define _CHUNK_SIZE_T unsigned long +# define _OBSTACK_CAST(type, expr) ((type) (expr)) +#else +/* Version 2 with sane types, especially for 64-bit hosts. */ +# define _OBSTACK_SIZE_T size_t +# define _CHUNK_SIZE_T size_t +# define _OBSTACK_CAST(type, expr) (expr) +#endif + +/* If B is the base of an object addressed by P, return the result of + aligning P to the next multiple of A + 1. B and P must be of type + char *. A + 1 must be a power of 2. */ + +#define __BPTR_ALIGN(B, P, A) ((B) + (((P) - (B) + (A)) & ~(A))) + +/* Similar to __BPTR_ALIGN (B, P, A), except optimize the common case + where pointers can be converted to integers, aligned as integers, + and converted back again. If ptrdiff_t is narrower than a + pointer (e.g., the AS/400), play it safe and compute the alignment + relative to B. Otherwise, use the faster strategy of computing the + alignment relative to 0. */ + +#define __PTR_ALIGN(B, P, A) \ + __BPTR_ALIGN (sizeof (ptrdiff_t) < sizeof (void *) ? (B) : (char *) 0, \ + P, A) + +#ifndef __attribute_pure__ +# define __attribute_pure__ _GL_ATTRIBUTE_PURE +#endif + +/* Not the same as _Noreturn, since it also works with function pointers. */ +#ifndef __attribute_noreturn__ +# if 2 < __GNUC__ + (8 <= __GNUC_MINOR__) || 0x5110 <= __SUNPRO_C +# define __attribute_noreturn__ __attribute__ ((__noreturn__)) +# else +# define __attribute_noreturn__ +# endif +#endif + +#ifdef __cplusplus +extern "C" { +#endif + +struct _obstack_chunk /* Lives at front of each chunk. */ +{ + char *limit; /* 1 past end of this chunk */ + struct _obstack_chunk *prev; /* address of prior chunk or NULL */ + char contents[__FLEXIBLE_ARRAY_MEMBER]; /* objects begin here */ +}; + +struct obstack /* control current object in current chunk */ +{ + _CHUNK_SIZE_T chunk_size; /* preferred size to allocate chunks in */ + struct _obstack_chunk *chunk; /* address of current struct obstack_chunk */ + char *object_base; /* address of object we are building */ + char *next_free; /* where to add next char to current object */ + char *chunk_limit; /* address of char after current chunk */ + union + { + _OBSTACK_SIZE_T i; + void *p; + } temp; /* Temporary for some macros. */ + _OBSTACK_SIZE_T alignment_mask; /* Mask of alignment for each object. */ + + /* These prototypes vary based on 'use_extra_arg'. */ + union + { + void *(*plain) (size_t); + void *(*extra) (void *, size_t); + } chunkfun; + union + { + void (*plain) (void *); + void (*extra) (void *, void *); + } freefun; + + void *extra_arg; /* first arg for chunk alloc/dealloc funcs */ + unsigned use_extra_arg : 1; /* chunk alloc/dealloc funcs take extra arg */ + unsigned maybe_empty_object : 1; /* There is a possibility that the current + chunk contains a zero-length object. This + prevents freeing the chunk if we allocate + a bigger chunk to replace it. */ + unsigned alloc_failed : 1; /* No longer used, as we now call the failed + handler on error, but retained for binary + compatibility. */ +}; + +/* Declare the external functions we use; they are in obstack.c. */ + +extern void _obstack_newchunk (struct obstack *, _OBSTACK_SIZE_T); +extern void _obstack_free (struct obstack *, void *); +extern int _obstack_begin (struct obstack *, + _OBSTACK_SIZE_T, _OBSTACK_SIZE_T, + void *(*) (size_t), void (*) (void *)); +extern int _obstack_begin_1 (struct obstack *, + _OBSTACK_SIZE_T, _OBSTACK_SIZE_T, + void *(*) (void *, size_t), + void (*) (void *, void *), void *); +extern _OBSTACK_SIZE_T _obstack_memory_used (struct obstack *) + __attribute_pure__; + + +/* Error handler called when 'obstack_chunk_alloc' failed to allocate + more memory. This can be set to a user defined function which + should either abort gracefully or use longjump - but shouldn't + return. The default action is to print a message and abort. */ +extern __attribute_noreturn__ void (*obstack_alloc_failed_handler) (void); + +/* Exit value used when 'print_and_abort' is used. */ +extern int obstack_exit_failure; + +/* Pointer to beginning of object being allocated or to be allocated next. + Note that this might not be the final address of the object + because a new chunk might be needed to hold the final size. */ + +#define obstack_base(h) ((void *) (h)->object_base) + +/* Size for allocating ordinary chunks. */ + +#define obstack_chunk_size(h) ((h)->chunk_size) + +/* Pointer to next byte not yet allocated in current chunk. */ + +#define obstack_next_free(h) ((void *) (h)->next_free) + +/* Mask specifying low bits that should be clear in address of an object. */ + +#define obstack_alignment_mask(h) ((h)->alignment_mask) + +/* To prevent prototype warnings provide complete argument list. */ +#define obstack_init(h) \ + _obstack_begin ((h), 0, 0, \ + _OBSTACK_CAST (void *(*) (size_t), obstack_chunk_alloc), \ + _OBSTACK_CAST (void (*) (void *), obstack_chunk_free)) + +#define obstack_begin(h, size) \ + _obstack_begin ((h), (size), 0, \ + _OBSTACK_CAST (void *(*) (size_t), obstack_chunk_alloc), \ + _OBSTACK_CAST (void (*) (void *), obstack_chunk_free)) + +#define obstack_specify_allocation(h, size, alignment, chunkfun, freefun) \ + _obstack_begin ((h), (size), (alignment), \ + _OBSTACK_CAST (void *(*) (size_t), chunkfun), \ + _OBSTACK_CAST (void (*) (void *), freefun)) + +#define obstack_specify_allocation_with_arg(h, size, alignment, chunkfun, freefun, arg) \ + _obstack_begin_1 ((h), (size), (alignment), \ + _OBSTACK_CAST (void *(*) (void *, size_t), chunkfun), \ + _OBSTACK_CAST (void (*) (void *, void *), freefun), arg) + +#define obstack_chunkfun(h, newchunkfun) \ + ((void) ((h)->chunkfun.extra = (void *(*) (void *, size_t)) (newchunkfun))) + +#define obstack_freefun(h, newfreefun) \ + ((void) ((h)->freefun.extra = (void *(*) (void *, void *)) (newfreefun))) + +#define obstack_1grow_fast(h, achar) ((void) (*((h)->next_free)++ = (achar))) + +#define obstack_blank_fast(h, n) ((void) ((h)->next_free += (n))) + +#define obstack_memory_used(h) _obstack_memory_used (h) + +#if defined __GNUC__ +# if !defined __GNUC_MINOR__ || __GNUC__ * 1000 + __GNUC_MINOR__ < 2008 +# define __extension__ +# endif + +/* For GNU C, if not -traditional, + we can define these macros to compute all args only once + without using a global variable. + Also, we can avoid using the 'temp' slot, to make faster code. */ + +# define obstack_object_size(OBSTACK) \ + __extension__ \ + ({ struct obstack const *__o = (OBSTACK); \ + (_OBSTACK_SIZE_T) (__o->next_free - __o->object_base); }) + +/* The local variable is named __o1 to avoid a shadowed variable + warning when invoked from other obstack macros. */ +# define obstack_room(OBSTACK) \ + __extension__ \ + ({ struct obstack const *__o1 = (OBSTACK); \ + (_OBSTACK_SIZE_T) (__o1->chunk_limit - __o1->next_free); }) + +# define obstack_make_room(OBSTACK, length) \ + __extension__ \ + ({ struct obstack *__o = (OBSTACK); \ + _OBSTACK_SIZE_T __len = (length); \ + if (obstack_room (__o) < __len) \ + _obstack_newchunk (__o, __len); \ + (void) 0; }) + +# define obstack_empty_p(OBSTACK) \ + __extension__ \ + ({ struct obstack const *__o = (OBSTACK); \ + (__o->chunk->prev == 0 \ + && __o->next_free == __PTR_ALIGN ((char *) __o->chunk, \ + __o->chunk->contents, \ + __o->alignment_mask)); }) + +# define obstack_grow(OBSTACK, where, length) \ + __extension__ \ + ({ struct obstack *__o = (OBSTACK); \ + _OBSTACK_SIZE_T __len = (length); \ + if (obstack_room (__o) < __len) \ + _obstack_newchunk (__o, __len); \ + memcpy (__o->next_free, where, __len); \ + __o->next_free += __len; \ + (void) 0; }) + +# define obstack_grow0(OBSTACK, where, length) \ + __extension__ \ + ({ struct obstack *__o = (OBSTACK); \ + _OBSTACK_SIZE_T __len = (length); \ + if (obstack_room (__o) < __len + 1) \ + _obstack_newchunk (__o, __len + 1); \ + memcpy (__o->next_free, where, __len); \ + __o->next_free += __len; \ + *(__o->next_free)++ = 0; \ + (void) 0; }) + +# define obstack_1grow(OBSTACK, datum) \ + __extension__ \ + ({ struct obstack *__o = (OBSTACK); \ + if (obstack_room (__o) < 1) \ + _obstack_newchunk (__o, 1); \ + obstack_1grow_fast (__o, datum); }) + +/* These assume that the obstack alignment is good enough for pointers + or ints, and that the data added so far to the current object + shares that much alignment. */ + +# define obstack_ptr_grow(OBSTACK, datum) \ + __extension__ \ + ({ struct obstack *__o = (OBSTACK); \ + if (obstack_room (__o) < sizeof (void *)) \ + _obstack_newchunk (__o, sizeof (void *)); \ + obstack_ptr_grow_fast (__o, datum); }) + +# define obstack_int_grow(OBSTACK, datum) \ + __extension__ \ + ({ struct obstack *__o = (OBSTACK); \ + if (obstack_room (__o) < sizeof (int)) \ + _obstack_newchunk (__o, sizeof (int)); \ + obstack_int_grow_fast (__o, datum); }) + +# define obstack_ptr_grow_fast(OBSTACK, aptr) \ + __extension__ \ + ({ struct obstack *__o1 = (OBSTACK); \ + void *__p1 = __o1->next_free; \ + *(const void **) __p1 = (aptr); \ + __o1->next_free += sizeof (const void *); \ + (void) 0; }) + +# define obstack_int_grow_fast(OBSTACK, aint) \ + __extension__ \ + ({ struct obstack *__o1 = (OBSTACK); \ + void *__p1 = __o1->next_free; \ + *(int *) __p1 = (aint); \ + __o1->next_free += sizeof (int); \ + (void) 0; }) + +# define obstack_blank(OBSTACK, length) \ + __extension__ \ + ({ struct obstack *__o = (OBSTACK); \ + _OBSTACK_SIZE_T __len = (length); \ + if (obstack_room (__o) < __len) \ + _obstack_newchunk (__o, __len); \ + obstack_blank_fast (__o, __len); }) + +# define obstack_alloc(OBSTACK, length) \ + __extension__ \ + ({ struct obstack *__h = (OBSTACK); \ + obstack_blank (__h, (length)); \ + obstack_finish (__h); }) + +# define obstack_copy(OBSTACK, where, length) \ + __extension__ \ + ({ struct obstack *__h = (OBSTACK); \ + obstack_grow (__h, (where), (length)); \ + obstack_finish (__h); }) + +# define obstack_copy0(OBSTACK, where, length) \ + __extension__ \ + ({ struct obstack *__h = (OBSTACK); \ + obstack_grow0 (__h, (where), (length)); \ + obstack_finish (__h); }) + +/* The local variable is named __o1 to avoid a shadowed variable + warning when invoked from other obstack macros, typically obstack_free. */ +# define obstack_finish(OBSTACK) \ + __extension__ \ + ({ struct obstack *__o1 = (OBSTACK); \ + void *__value = (void *) __o1->object_base; \ + if (__o1->next_free == __value) \ + __o1->maybe_empty_object = 1; \ + __o1->next_free \ + = __PTR_ALIGN (__o1->object_base, __o1->next_free, \ + __o1->alignment_mask); \ + if ((size_t) (__o1->next_free - (char *) __o1->chunk) \ + > (size_t) (__o1->chunk_limit - (char *) __o1->chunk)) \ + __o1->next_free = __o1->chunk_limit; \ + __o1->object_base = __o1->next_free; \ + __value; }) + +# define obstack_free(OBSTACK, OBJ) \ + __extension__ \ + ({ struct obstack *__o = (OBSTACK); \ + void *__obj = (void *) (OBJ); \ + if (__obj > (void *) __o->chunk && __obj < (void *) __o->chunk_limit) \ + __o->next_free = __o->object_base = (char *) __obj; \ + else \ + _obstack_free (__o, __obj); }) + +#else /* not __GNUC__ */ + +# define obstack_object_size(h) \ + ((_OBSTACK_SIZE_T) ((h)->next_free - (h)->object_base)) + +# define obstack_room(h) \ + ((_OBSTACK_SIZE_T) ((h)->chunk_limit - (h)->next_free)) + +# define obstack_empty_p(h) \ + ((h)->chunk->prev == 0 \ + && (h)->next_free == __PTR_ALIGN ((char *) (h)->chunk, \ + (h)->chunk->contents, \ + (h)->alignment_mask)) + +/* Note that the call to _obstack_newchunk is enclosed in (..., 0) + so that we can avoid having void expressions + in the arms of the conditional expression. + Casting the third operand to void was tried before, + but some compilers won't accept it. */ + +# define obstack_make_room(h, length) \ + ((h)->temp.i = (length), \ + ((obstack_room (h) < (h)->temp.i) \ + ? (_obstack_newchunk (h, (h)->temp.i), 0) : 0), \ + (void) 0) + +# define obstack_grow(h, where, length) \ + ((h)->temp.i = (length), \ + ((obstack_room (h) < (h)->temp.i) \ + ? (_obstack_newchunk ((h), (h)->temp.i), 0) : 0), \ + memcpy ((h)->next_free, where, (h)->temp.i), \ + (h)->next_free += (h)->temp.i, \ + (void) 0) + +# define obstack_grow0(h, where, length) \ + ((h)->temp.i = (length), \ + ((obstack_room (h) < (h)->temp.i + 1) \ + ? (_obstack_newchunk ((h), (h)->temp.i + 1), 0) : 0), \ + memcpy ((h)->next_free, where, (h)->temp.i), \ + (h)->next_free += (h)->temp.i, \ + *((h)->next_free)++ = 0, \ + (void) 0) + +# define obstack_1grow(h, datum) \ + (((obstack_room (h) < 1) \ + ? (_obstack_newchunk ((h), 1), 0) : 0), \ + obstack_1grow_fast (h, datum)) + +# define obstack_ptr_grow(h, datum) \ + (((obstack_room (h) < sizeof (char *)) \ + ? (_obstack_newchunk ((h), sizeof (char *)), 0) : 0), \ + obstack_ptr_grow_fast (h, datum)) + +# define obstack_int_grow(h, datum) \ + (((obstack_room (h) < sizeof (int)) \ + ? (_obstack_newchunk ((h), sizeof (int)), 0) : 0), \ + obstack_int_grow_fast (h, datum)) + +# define obstack_ptr_grow_fast(h, aptr) \ + (((const void **) ((h)->next_free += sizeof (void *)))[-1] = (aptr), \ + (void) 0) + +# define obstack_int_grow_fast(h, aint) \ + (((int *) ((h)->next_free += sizeof (int)))[-1] = (aint), \ + (void) 0) + +# define obstack_blank(h, length) \ + ((h)->temp.i = (length), \ + ((obstack_room (h) < (h)->temp.i) \ + ? (_obstack_newchunk ((h), (h)->temp.i), 0) : 0), \ + obstack_blank_fast (h, (h)->temp.i)) + +# define obstack_alloc(h, length) \ + (obstack_blank ((h), (length)), obstack_finish ((h))) + +# define obstack_copy(h, where, length) \ + (obstack_grow ((h), (where), (length)), obstack_finish ((h))) + +# define obstack_copy0(h, where, length) \ + (obstack_grow0 ((h), (where), (length)), obstack_finish ((h))) + +# define obstack_finish(h) \ + (((h)->next_free == (h)->object_base \ + ? (((h)->maybe_empty_object = 1), 0) \ + : 0), \ + (h)->temp.p = (h)->object_base, \ + (h)->next_free \ + = __PTR_ALIGN ((h)->object_base, (h)->next_free, \ + (h)->alignment_mask), \ + (((size_t) ((h)->next_free - (char *) (h)->chunk) \ + > (size_t) ((h)->chunk_limit - (char *) (h)->chunk)) \ + ? ((h)->next_free = (h)->chunk_limit) : 0), \ + (h)->object_base = (h)->next_free, \ + (h)->temp.p) + +# define obstack_free(h, obj) \ + ((h)->temp.p = (void *) (obj), \ + (((h)->temp.p > (void *) (h)->chunk \ + && (h)->temp.p < (void *) (h)->chunk_limit) \ + ? (void) ((h)->next_free = (h)->object_base = (char *) (h)->temp.p) \ + : _obstack_free ((h), (h)->temp.p))) + +#endif /* not __GNUC__ */ + +#ifdef __cplusplus +} /* C++ */ +#endif + +#endif /* _OBSTACK_H */ diff --git a/lib/pathmax.h b/lib/pathmax.h new file mode 100644 index 0000000..7052e6a --- /dev/null +++ b/lib/pathmax.h @@ -0,0 +1,83 @@ +/* Define PATH_MAX somehow. Requires sys/types.h. + Copyright (C) 1992, 1999, 2001, 2003, 2005, 2009-2018 Free Software + Foundation, Inc. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3, or (at your option) + any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, see <https://www.gnu.org/licenses/>. */ + +#ifndef _PATHMAX_H +# define _PATHMAX_H + +/* POSIX:2008 defines PATH_MAX to be the maximum number of bytes in a filename, + including the terminating NUL byte. + <http://pubs.opengroup.org/onlinepubs/9699919799/basedefs/limits.h.html> + PATH_MAX is not defined on systems which have no limit on filename length, + such as GNU/Hurd. + + This file does *not* define PATH_MAX always. Programs that use this file + can handle the GNU/Hurd case in several ways: + - Either with a package-wide handling, or with a per-file handling, + - Either through a + #ifdef PATH_MAX + or through a fallback like + #ifndef PATH_MAX + # define PATH_MAX 8192 + #endif + or through a fallback like + #ifndef PATH_MAX + # define PATH_MAX pathconf ("/", _PC_PATH_MAX) + #endif + */ + +# include <unistd.h> + +# include <limits.h> + +# ifndef _POSIX_PATH_MAX +# define _POSIX_PATH_MAX 256 +# endif + +/* Don't include sys/param.h if it already has been. */ +# if defined HAVE_SYS_PARAM_H && !defined PATH_MAX && !defined MAXPATHLEN +# include <sys/param.h> +# endif + +# if !defined PATH_MAX && defined MAXPATHLEN +# define PATH_MAX MAXPATHLEN +# endif + +# ifdef __hpux +/* On HP-UX, PATH_MAX designates the maximum number of bytes in a filename, + *not* including the terminating NUL byte, and is set to 1023. + Additionally, when _XOPEN_SOURCE is defined to 500 or more, PATH_MAX is + not defined at all any more. */ +# undef PATH_MAX +# define PATH_MAX 1024 +# endif + +# if defined _WIN32 && ! defined __CYGWIN__ +/* The page "Naming Files, Paths, and Namespaces" on msdn.microsoft.com, + section "Maximum Path Length Limitation", + <https://msdn.microsoft.com/en-us/library/aa365247.aspx#maxpath> + explains that the maximum size of a filename, including the terminating + NUL byte, is 260 = 3 + 256 + 1. + This is the same value as + - FILENAME_MAX in <stdio.h>, + - _MAX_PATH in <stdlib.h>, + - MAX_PATH in <windef.h>. + Undefine the original value, because mingw's <limits.h> gets it wrong. */ +# undef PATH_MAX +# define PATH_MAX 260 +# endif + +#endif /* _PATHMAX_H */ diff --git a/lib/progname.c b/lib/progname.c new file mode 100644 index 0000000..382f503 --- /dev/null +++ b/lib/progname.c @@ -0,0 +1,92 @@ +/* Program name management. + Copyright (C) 2001-2003, 2005-2018 Free Software Foundation, Inc. + Written by Bruno Haible <bruno@clisp.org>, 2001. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + + +#include <config.h> + +/* Specification. */ +#undef ENABLE_RELOCATABLE /* avoid defining set_program_name as a macro */ +#include "progname.h" + +#include <errno.h> /* get program_invocation_name declaration */ +#include <stdio.h> +#include <stdlib.h> +#include <string.h> + + +/* String containing name the program is called with. + To be initialized by main(). */ +const char *program_name = NULL; + +/* Set program_name, based on argv[0]. + argv0 must be a string allocated with indefinite extent, and must not be + modified after this call. */ +void +set_program_name (const char *argv0) +{ + /* libtool creates a temporary executable whose name is sometimes prefixed + with "lt-" (depends on the platform). It also makes argv[0] absolute. + But the name of the temporary executable is a detail that should not be + visible to the end user and to the test suite. + Remove this "<dirname>/.libs/" or "<dirname>/.libs/lt-" prefix here. */ + const char *slash; + const char *base; + + /* Sanity check. POSIX requires the invoking process to pass a non-NULL + argv[0]. */ + if (argv0 == NULL) + { + /* It's a bug in the invoking program. Help diagnosing it. */ + fputs ("A NULL argv[0] was passed through an exec system call.\n", + stderr); + abort (); + } + + slash = strrchr (argv0, '/'); + base = (slash != NULL ? slash + 1 : argv0); + if (base - argv0 >= 7 && strncmp (base - 7, "/.libs/", 7) == 0) + { + argv0 = base; + if (strncmp (base, "lt-", 3) == 0) + { + argv0 = base + 3; + /* On glibc systems, remove the "lt-" prefix from the variable + program_invocation_short_name. */ +#if HAVE_DECL_PROGRAM_INVOCATION_SHORT_NAME + program_invocation_short_name = (char *) argv0; +#endif + } + } + + /* But don't strip off a leading <dirname>/ in general, because when the user + runs + /some/hidden/place/bin/cp foo foo + he should get the error message + /some/hidden/place/bin/cp: `foo' and `foo' are the same file + not + cp: `foo' and `foo' are the same file + */ + + program_name = argv0; + + /* On glibc systems, the error() function comes from libc and uses the + variable program_invocation_name, not program_name. So set this variable + as well. */ +#if HAVE_DECL_PROGRAM_INVOCATION_NAME + program_invocation_name = (char *) argv0; +#endif +} diff --git a/lib/progname.h b/lib/progname.h new file mode 100644 index 0000000..adc6b01 --- /dev/null +++ b/lib/progname.h @@ -0,0 +1,62 @@ +/* Program name management. + Copyright (C) 2001-2004, 2006, 2009-2018 Free Software Foundation, Inc. + Written by Bruno Haible <bruno@clisp.org>, 2001. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +#ifndef _PROGNAME_H +#define _PROGNAME_H + +/* Programs using this file should do the following in main(): + set_program_name (argv[0]); + */ + + +#ifdef __cplusplus +extern "C" { +#endif + + +/* String containing name the program is called with. */ +extern const char *program_name; + +/* Set program_name, based on argv[0]. + argv0 must be a string allocated with indefinite extent, and must not be + modified after this call. */ +extern void set_program_name (const char *argv0); + +#if ENABLE_RELOCATABLE + +/* Set program_name, based on argv[0], and original installation prefix and + directory, for relocatability. */ +extern void set_program_name_and_installdir (const char *argv0, + const char *orig_installprefix, + const char *orig_installdir); +#undef set_program_name +#define set_program_name(ARG0) \ + set_program_name_and_installdir (ARG0, INSTALLPREFIX, INSTALLDIR) + +/* Return the full pathname of the current executable, based on the earlier + call to set_program_name_and_installdir. Return NULL if unknown. */ +extern char *get_full_program_name (void); + +#endif + + +#ifdef __cplusplus +} +#endif + + +#endif /* _PROGNAME_H */ diff --git a/lib/qcopy-acl.c b/lib/qcopy-acl.c new file mode 100644 index 0000000..ee482ef --- /dev/null +++ b/lib/qcopy-acl.c @@ -0,0 +1,51 @@ +/* Copy access control list from one file to another. -*- coding: utf-8 -*- + + Copyright (C) 2002-2003, 2005-2018 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. + + Written by Paul Eggert, Andreas Grünbacher, and Bruno Haible. */ + +#include <config.h> + +#include "acl.h" + +#include "acl-internal.h" + + +/* Copy access control lists from one file to another. If SOURCE_DESC is + a valid file descriptor, use file descriptor operations, else use + filename based operations on SRC_NAME. Likewise for DEST_DESC and + DST_NAME. + If access control lists are not available, fchmod the target file to + MODE. Also sets the non-permission bits of the destination file + (S_ISUID, S_ISGID, S_ISVTX) to those from MODE if any are set. + Return 0 if successful. + Return -2 and set errno for an error relating to the source file. + Return -1 and set errno for an error relating to the destination file. */ + +int +qcopy_acl (const char *src_name, int source_desc, const char *dst_name, + int dest_desc, mode_t mode) +{ + struct permission_context ctx; + int ret; + + ret = get_permissions (src_name, source_desc, mode, &ctx); + if (ret != 0) + return -2; + ret = set_permissions (&ctx, dst_name, dest_desc); + free_permission_context (&ctx); + return ret; +} diff --git a/lib/qset-acl.c b/lib/qset-acl.c new file mode 100644 index 0000000..e2d4cc5 --- /dev/null +++ b/lib/qset-acl.c @@ -0,0 +1,49 @@ +/* qset-acl.c - set access control list equivalent to a mode + + Copyright (C) 2002-2003, 2005-2018 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. + + Written by Paul Eggert and Andreas Gruenbacher, and Bruno Haible. */ + +#include <config.h> + +#define ACL_INTERNAL_INLINE _GL_EXTERN_INLINE + +#include <string.h> +#include "acl.h" + +#include "acl-internal.h" + + +/* Set the access control lists of a file. If DESC is a valid file + descriptor, use file descriptor operations where available, else use + filename based operations on NAME. If access control lists are not + available, fchmod the target file to MODE. Also sets the + non-permission bits of the destination file (S_ISUID, S_ISGID, S_ISVTX) + to those from MODE if any are set. + Return 0 if successful. Return -1 and set errno upon failure. */ + +int +qset_acl (char const *name, int desc, mode_t mode) +{ + struct permission_context ctx; + int ret; + + memset (&ctx, 0, sizeof ctx); + ctx.mode = mode; + ret = set_permissions (&ctx, name, desc); + free_permission_context (&ctx); + return ret; +} diff --git a/lib/quote.h b/lib/quote.h new file mode 100644 index 0000000..eedc283 --- /dev/null +++ b/lib/quote.h @@ -0,0 +1,46 @@ +/* quote.h - prototypes for quote.c + + Copyright (C) 1998-2001, 2003, 2009-2018 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +#ifndef QUOTE_H_ +# define QUOTE_H_ 1 + +# include <stddef.h> + +/* The quoting options used by quote_n and quote. Its type is incomplete, + so it's useful only in expressions like '"e_quoting_options'. */ +extern struct quoting_options quote_quoting_options; + +/* Return an unambiguous printable representation of ARG (of size + ARGSIZE), allocated in slot N, suitable for diagnostics. If + ARGSIZE is SIZE_MAX, use the string length of the argument for + ARGSIZE. */ +char const *quote_n_mem (int n, char const *arg, size_t argsize); + +/* Return an unambiguous printable representation of ARG (of size + ARGSIZE), suitable for diagnostics. If ARGSIZE is SIZE_MAX, use + the string length of the argument for ARGSIZE. */ +char const *quote_mem (char const *arg, size_t argsize); + +/* Return an unambiguous printable representation of ARG, allocated in + slot N, suitable for diagnostics. */ +char const *quote_n (int n, char const *arg); + +/* Return an unambiguous printable representation of ARG, suitable for + diagnostics. */ +char const *quote (char const *arg); + +#endif /* !QUOTE_H_ */ diff --git a/lib/quotearg.c b/lib/quotearg.c new file mode 100644 index 0000000..fe68dca --- /dev/null +++ b/lib/quotearg.c @@ -0,0 +1,1089 @@ +/* quotearg.c - quote arguments for output + + Copyright (C) 1998-2002, 2004-2018 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +/* Written by Paul Eggert <eggert@twinsun.com> */ + +/* Without this pragma, gcc 4.7.0 20111124 mistakenly suggests that + the quoting_options_from_style function might be candidate for + attribute 'pure' */ +#if (__GNUC__ == 4 && 6 <= __GNUC_MINOR__) || 4 < __GNUC__ +# pragma GCC diagnostic ignored "-Wsuggest-attribute=pure" +#endif + +#include <config.h> + +#include "quotearg.h" +#include "quote.h" + +#include "minmax.h" +#include "xalloc.h" +#include "c-strcaseeq.h" +#include "localcharset.h" + +#include <ctype.h> +#include <errno.h> +#include <limits.h> +#include <stdbool.h> +#include <stdint.h> +#include <stdlib.h> +#include <string.h> +#include <wchar.h> +#include <wctype.h> + +#include "gettext.h" +#define _(msgid) gettext (msgid) +#define N_(msgid) msgid + +#ifndef SIZE_MAX +# define SIZE_MAX ((size_t) -1) +#endif + +#define INT_BITS (sizeof (int) * CHAR_BIT) + +#ifndef FALLTHROUGH +# if __GNUC__ < 7 +# define FALLTHROUGH ((void) 0) +# else +# define FALLTHROUGH __attribute__ ((__fallthrough__)) +# endif +#endif + +struct quoting_options +{ + /* Basic quoting style. */ + enum quoting_style style; + + /* Additional flags. Bitwise combination of enum quoting_flags. */ + int flags; + + /* Quote the characters indicated by this bit vector even if the + quoting style would not normally require them to be quoted. */ + unsigned int quote_these_too[(UCHAR_MAX / INT_BITS) + 1]; + + /* The left quote for custom_quoting_style. */ + char const *left_quote; + + /* The right quote for custom_quoting_style. */ + char const *right_quote; +}; + +/* Names of quoting styles. */ +char const *const quoting_style_args[] = +{ + "literal", + "shell", + "shell-always", + "shell-escape", + "shell-escape-always", + "c", + "c-maybe", + "escape", + "locale", + "clocale", + 0 +}; + +/* Correspondences to quoting style names. */ +enum quoting_style const quoting_style_vals[] = +{ + literal_quoting_style, + shell_quoting_style, + shell_always_quoting_style, + shell_escape_quoting_style, + shell_escape_always_quoting_style, + c_quoting_style, + c_maybe_quoting_style, + escape_quoting_style, + locale_quoting_style, + clocale_quoting_style +}; + +/* The default quoting options. */ +static struct quoting_options default_quoting_options; + +/* Allocate a new set of quoting options, with contents initially identical + to O if O is not null, or to the default if O is null. + It is the caller's responsibility to free the result. */ +struct quoting_options * +clone_quoting_options (struct quoting_options *o) +{ + int e = errno; + struct quoting_options *p = xmemdup (o ? o : &default_quoting_options, + sizeof *o); + errno = e; + return p; +} + +/* Get the value of O's quoting style. If O is null, use the default. */ +enum quoting_style +get_quoting_style (struct quoting_options const *o) +{ + return (o ? o : &default_quoting_options)->style; +} + +/* In O (or in the default if O is null), + set the value of the quoting style to S. */ +void +set_quoting_style (struct quoting_options *o, enum quoting_style s) +{ + (o ? o : &default_quoting_options)->style = s; +} + +/* In O (or in the default if O is null), + set the value of the quoting options for character C to I. + Return the old value. Currently, the only values defined for I are + 0 (the default) and 1 (which means to quote the character even if + it would not otherwise be quoted). */ +int +set_char_quoting (struct quoting_options *o, char c, int i) +{ + unsigned char uc = c; + unsigned int *p = + (o ? o : &default_quoting_options)->quote_these_too + uc / INT_BITS; + int shift = uc % INT_BITS; + int r = (*p >> shift) & 1; + *p ^= ((i & 1) ^ r) << shift; + return r; +} + +/* In O (or in the default if O is null), + set the value of the quoting options flag to I, which can be a + bitwise combination of enum quoting_flags, or 0 for default + behavior. Return the old value. */ +int +set_quoting_flags (struct quoting_options *o, int i) +{ + int r; + if (!o) + o = &default_quoting_options; + r = o->flags; + o->flags = i; + return r; +} + +void +set_custom_quoting (struct quoting_options *o, + char const *left_quote, char const *right_quote) +{ + if (!o) + o = &default_quoting_options; + o->style = custom_quoting_style; + if (!left_quote || !right_quote) + abort (); + o->left_quote = left_quote; + o->right_quote = right_quote; +} + +/* Return quoting options for STYLE, with no extra quoting. */ +static struct quoting_options /* NOT PURE!! */ +quoting_options_from_style (enum quoting_style style) +{ + struct quoting_options o = { literal_quoting_style, 0, { 0 }, NULL, NULL }; + if (style == custom_quoting_style) + abort (); + o.style = style; + return o; +} + +/* MSGID approximates a quotation mark. Return its translation if it + has one; otherwise, return either it or "\"", depending on S. + + S is either clocale_quoting_style or locale_quoting_style. */ +static char const * +gettext_quote (char const *msgid, enum quoting_style s) +{ + char const *translation = _(msgid); + char const *locale_code; + + if (translation != msgid) + return translation; + + /* For UTF-8 and GB-18030, use single quotes U+2018 and U+2019. + Here is a list of other locales that include U+2018 and U+2019: + + ISO-8859-7 0xA1 KOI8-T 0x91 + CP869 0x8B CP874 0x91 + CP932 0x81 0x65 CP936 0xA1 0xAE + CP949 0xA1 0xAE CP950 0xA1 0xA5 + CP1250 0x91 CP1251 0x91 + CP1252 0x91 CP1253 0x91 + CP1254 0x91 CP1255 0x91 + CP1256 0x91 CP1257 0x91 + EUC-JP 0xA1 0xC6 EUC-KR 0xA1 0xAE + EUC-TW 0xA1 0xE4 BIG5 0xA1 0xA5 + BIG5-HKSCS 0xA1 0xA5 EUC-CN 0xA1 0xAE + GBK 0xA1 0xAE Georgian-PS 0x91 + PT154 0x91 + + None of these is still in wide use; using iconv is overkill. */ + locale_code = locale_charset (); + if (STRCASEEQ (locale_code, "UTF-8", 'U','T','F','-','8',0,0,0,0)) + return msgid[0] == '`' ? "\xe2\x80\x98": "\xe2\x80\x99"; + if (STRCASEEQ (locale_code, "GB18030", 'G','B','1','8','0','3','0',0,0)) + return msgid[0] == '`' ? "\xa1\ae": "\xa1\xaf"; + + return (s == clocale_quoting_style ? "\"" : "'"); +} + +/* Place into buffer BUFFER (of size BUFFERSIZE) a quoted version of + argument ARG (of size ARGSIZE), using QUOTING_STYLE, FLAGS, and + QUOTE_THESE_TOO to control quoting. + Terminate the output with a null character, and return the written + size of the output, not counting the terminating null. + If BUFFERSIZE is too small to store the output string, return the + value that would have been returned had BUFFERSIZE been large enough. + If ARGSIZE is SIZE_MAX, use the string length of the argument for ARGSIZE. + + This function acts like quotearg_buffer (BUFFER, BUFFERSIZE, ARG, + ARGSIZE, O), except it breaks O into its component pieces and is + not careful about errno. */ + +static size_t +quotearg_buffer_restyled (char *buffer, size_t buffersize, + char const *arg, size_t argsize, + enum quoting_style quoting_style, int flags, + unsigned int const *quote_these_too, + char const *left_quote, + char const *right_quote) +{ + size_t i; + size_t len = 0; + size_t orig_buffersize = 0; + char const *quote_string = 0; + size_t quote_string_len = 0; + bool backslash_escapes = false; + bool unibyte_locale = MB_CUR_MAX == 1; + bool elide_outer_quotes = (flags & QA_ELIDE_OUTER_QUOTES) != 0; + bool pending_shell_escape_end = false; + bool encountered_single_quote = false; + bool all_c_and_shell_quote_compat = true; + +#define STORE(c) \ + do \ + { \ + if (len < buffersize) \ + buffer[len] = (c); \ + len++; \ + } \ + while (0) + +#define START_ESC() \ + do \ + { \ + if (elide_outer_quotes) \ + goto force_outer_quoting_style; \ + escaping = true; \ + if (quoting_style == shell_always_quoting_style \ + && ! pending_shell_escape_end) \ + { \ + STORE ('\''); \ + STORE ('$'); \ + STORE ('\''); \ + pending_shell_escape_end = true; \ + } \ + STORE ('\\'); \ + } \ + while (0) + +#define END_ESC() \ + do \ + { \ + if (pending_shell_escape_end && ! escaping) \ + { \ + STORE ('\''); \ + STORE ('\''); \ + pending_shell_escape_end = false; \ + } \ + } \ + while (0) + + process_input: + + switch (quoting_style) + { + case c_maybe_quoting_style: + quoting_style = c_quoting_style; + elide_outer_quotes = true; + FALLTHROUGH; + case c_quoting_style: + if (!elide_outer_quotes) + STORE ('"'); + backslash_escapes = true; + quote_string = "\""; + quote_string_len = 1; + break; + + case escape_quoting_style: + backslash_escapes = true; + elide_outer_quotes = false; + break; + + case locale_quoting_style: + case clocale_quoting_style: + case custom_quoting_style: + { + if (quoting_style != custom_quoting_style) + { + /* TRANSLATORS: + Get translations for open and closing quotation marks. + The message catalog should translate "`" to a left + quotation mark suitable for the locale, and similarly for + "'". For example, a French Unicode local should translate + these to U+00AB (LEFT-POINTING DOUBLE ANGLE + QUOTATION MARK), and U+00BB (RIGHT-POINTING DOUBLE ANGLE + QUOTATION MARK), respectively. + + If the catalog has no translation, we will try to + use Unicode U+2018 (LEFT SINGLE QUOTATION MARK) and + Unicode U+2019 (RIGHT SINGLE QUOTATION MARK). If the + current locale is not Unicode, locale_quoting_style + will quote 'like this', and clocale_quoting_style will + quote "like this". You should always include translations + for "`" and "'" even if U+2018 and U+2019 are appropriate + for your locale. + + If you don't know what to put here, please see + <https://en.wikipedia.org/wiki/Quotation_marks_in_other_languages> + and use glyphs suitable for your language. */ + left_quote = gettext_quote (N_("`"), quoting_style); + right_quote = gettext_quote (N_("'"), quoting_style); + } + if (!elide_outer_quotes) + for (quote_string = left_quote; *quote_string; quote_string++) + STORE (*quote_string); + backslash_escapes = true; + quote_string = right_quote; + quote_string_len = strlen (quote_string); + } + break; + + case shell_escape_quoting_style: + backslash_escapes = true; + FALLTHROUGH; + case shell_quoting_style: + elide_outer_quotes = true; + FALLTHROUGH; + case shell_escape_always_quoting_style: + if (!elide_outer_quotes) + backslash_escapes = true; + FALLTHROUGH; + case shell_always_quoting_style: + quoting_style = shell_always_quoting_style; + if (!elide_outer_quotes) + STORE ('\''); + quote_string = "'"; + quote_string_len = 1; + break; + + case literal_quoting_style: + elide_outer_quotes = false; + break; + + default: + abort (); + } + + for (i = 0; ! (argsize == SIZE_MAX ? arg[i] == '\0' : i == argsize); i++) + { + unsigned char c; + unsigned char esc; + bool is_right_quote = false; + bool escaping = false; + bool c_and_shell_quote_compat = false; + + if (backslash_escapes + && quoting_style != shell_always_quoting_style + && quote_string_len + && (i + quote_string_len + <= (argsize == SIZE_MAX && 1 < quote_string_len + /* Use strlen only if we must: when argsize is SIZE_MAX, + and when the quote string is more than 1 byte long. + If we do call strlen, save the result. */ + ? (argsize = strlen (arg)) : argsize)) + && memcmp (arg + i, quote_string, quote_string_len) == 0) + { + if (elide_outer_quotes) + goto force_outer_quoting_style; + is_right_quote = true; + } + + c = arg[i]; + switch (c) + { + case '\0': + if (backslash_escapes) + { + START_ESC (); + /* If quote_string were to begin with digits, we'd need to + test for the end of the arg as well. However, it's + hard to imagine any locale that would use digits in + quotes, and set_custom_quoting is documented not to + accept them. Use only a single \0 with shell-escape + as currently digits are not printed within $'...' */ + if (quoting_style != shell_always_quoting_style + && i + 1 < argsize && '0' <= arg[i + 1] && arg[i + 1] <= '9') + { + STORE ('0'); + STORE ('0'); + } + c = '0'; + /* We don't have to worry that this last '0' will be + backslash-escaped because, again, quote_string should + not start with it and because quote_these_too is + documented as not accepting it. */ + } + else if (flags & QA_ELIDE_NULL_BYTES) + continue; + break; + + case '?': + switch (quoting_style) + { + case shell_always_quoting_style: + if (elide_outer_quotes) + goto force_outer_quoting_style; + break; + + case c_quoting_style: + if ((flags & QA_SPLIT_TRIGRAPHS) + && i + 2 < argsize && arg[i + 1] == '?') + switch (arg[i + 2]) + { + case '!': case '\'': + case '(': case ')': case '-': case '/': + case '<': case '=': case '>': + /* Escape the second '?' in what would otherwise be + a trigraph. */ + if (elide_outer_quotes) + goto force_outer_quoting_style; + c = arg[i + 2]; + i += 2; + STORE ('?'); + STORE ('"'); + STORE ('"'); + STORE ('?'); + break; + + default: + break; + } + break; + + default: + break; + } + break; + + case '\a': esc = 'a'; goto c_escape; + case '\b': esc = 'b'; goto c_escape; + case '\f': esc = 'f'; goto c_escape; + case '\n': esc = 'n'; goto c_and_shell_escape; + case '\r': esc = 'r'; goto c_and_shell_escape; + case '\t': esc = 't'; goto c_and_shell_escape; + case '\v': esc = 'v'; goto c_escape; + case '\\': esc = c; + /* Never need to escape '\' in shell case. */ + if (quoting_style == shell_always_quoting_style) + { + if (elide_outer_quotes) + goto force_outer_quoting_style; + goto store_c; + } + + /* No need to escape the escape if we are trying to elide + outer quotes and nothing else is problematic. */ + if (backslash_escapes && elide_outer_quotes && quote_string_len) + goto store_c; + + c_and_shell_escape: + if (quoting_style == shell_always_quoting_style + && elide_outer_quotes) + goto force_outer_quoting_style; + /* fall through */ + c_escape: + if (backslash_escapes) + { + c = esc; + goto store_escape; + } + break; + + case '{': case '}': /* sometimes special if isolated */ + if (! (argsize == SIZE_MAX ? arg[1] == '\0' : argsize == 1)) + break; + FALLTHROUGH; + case '#': case '~': + if (i != 0) + break; + FALLTHROUGH; + case ' ': + c_and_shell_quote_compat = true; + FALLTHROUGH; + case '!': /* special in bash */ + case '"': case '$': case '&': + case '(': case ')': case '*': case ';': + case '<': + case '=': /* sometimes special in 0th or (with "set -k") later args */ + case '>': case '[': + case '^': /* special in old /bin/sh, e.g. SunOS 4.1.4 */ + case '`': case '|': + /* A shell special character. In theory, '$' and '`' could + be the first bytes of multibyte characters, which means + we should check them with mbrtowc, but in practice this + doesn't happen so it's not worth worrying about. */ + if (quoting_style == shell_always_quoting_style + && elide_outer_quotes) + goto force_outer_quoting_style; + break; + + case '\'': + encountered_single_quote = true; + c_and_shell_quote_compat = true; + if (quoting_style == shell_always_quoting_style) + { + if (elide_outer_quotes) + goto force_outer_quoting_style; + + if (buffersize && ! orig_buffersize) + { + /* Just scan string to see if supports a more concise + representation, rather than writing a longer string + but returning the length of the more concise form. */ + orig_buffersize = buffersize; + buffersize = 0; + } + + STORE ('\''); + STORE ('\\'); + STORE ('\''); + pending_shell_escape_end = false; + } + break; + + case '%': case '+': case ',': case '-': case '.': case '/': + case '0': case '1': case '2': case '3': case '4': case '5': + case '6': case '7': case '8': case '9': case ':': + case 'A': case 'B': case 'C': case 'D': case 'E': case 'F': + case 'G': case 'H': case 'I': case 'J': case 'K': case 'L': + case 'M': case 'N': case 'O': case 'P': case 'Q': case 'R': + case 'S': case 'T': case 'U': case 'V': case 'W': case 'X': + case 'Y': case 'Z': case ']': case '_': case 'a': case 'b': + case 'c': case 'd': case 'e': case 'f': case 'g': case 'h': + case 'i': case 'j': case 'k': case 'l': case 'm': case 'n': + case 'o': case 'p': case 'q': case 'r': case 's': case 't': + case 'u': case 'v': case 'w': case 'x': case 'y': case 'z': + /* These characters don't cause problems, no matter what the + quoting style is. They cannot start multibyte sequences. + A digit or a special letter would cause trouble if it + appeared at the beginning of quote_string because we'd then + escape by prepending a backslash. However, it's hard to + imagine any locale that would use digits or letters as + quotes, and set_custom_quoting is documented not to accept + them. Also, a digit or a special letter would cause + trouble if it appeared in quote_these_too, but that's also + documented as not accepting them. */ + c_and_shell_quote_compat = true; + break; + + default: + /* If we have a multibyte sequence, copy it until we reach + its end, find an error, or come back to the initial shift + state. For C-like styles, if the sequence has + unprintable characters, escape the whole sequence, since + we can't easily escape single characters within it. */ + { + /* Length of multibyte sequence found so far. */ + size_t m; + + bool printable; + + if (unibyte_locale) + { + m = 1; + printable = isprint (c) != 0; + } + else + { + mbstate_t mbstate; + memset (&mbstate, 0, sizeof mbstate); + + m = 0; + printable = true; + if (argsize == SIZE_MAX) + argsize = strlen (arg); + + do + { + wchar_t w; + size_t bytes = mbrtowc (&w, &arg[i + m], + argsize - (i + m), &mbstate); + if (bytes == 0) + break; + else if (bytes == (size_t) -1) + { + printable = false; + break; + } + else if (bytes == (size_t) -2) + { + printable = false; + while (i + m < argsize && arg[i + m]) + m++; + break; + } + else + { + /* Work around a bug with older shells that "see" a '\' + that is really the 2nd byte of a multibyte character. + In practice the problem is limited to ASCII + chars >= '@' that are shell special chars. */ + if ('[' == 0x5b && elide_outer_quotes + && quoting_style == shell_always_quoting_style) + { + size_t j; + for (j = 1; j < bytes; j++) + switch (arg[i + m + j]) + { + case '[': case '\\': case '^': + case '`': case '|': + goto force_outer_quoting_style; + + default: + break; + } + } + + if (! iswprint (w)) + printable = false; + m += bytes; + } + } + while (! mbsinit (&mbstate)); + } + + c_and_shell_quote_compat = printable; + + if (1 < m || (backslash_escapes && ! printable)) + { + /* Output a multibyte sequence, or an escaped + unprintable unibyte character. */ + size_t ilim = i + m; + + for (;;) + { + if (backslash_escapes && ! printable) + { + START_ESC (); + STORE ('0' + (c >> 6)); + STORE ('0' + ((c >> 3) & 7)); + c = '0' + (c & 7); + } + else if (is_right_quote) + { + STORE ('\\'); + is_right_quote = false; + } + if (ilim <= i + 1) + break; + END_ESC (); + STORE (c); + c = arg[++i]; + } + + goto store_c; + } + } + } + + if (! (((backslash_escapes && quoting_style != shell_always_quoting_style) + || elide_outer_quotes) + && quote_these_too + && quote_these_too[c / INT_BITS] >> (c % INT_BITS) & 1) + && !is_right_quote) + goto store_c; + + store_escape: + START_ESC (); + + store_c: + END_ESC (); + STORE (c); + + if (! c_and_shell_quote_compat) + all_c_and_shell_quote_compat = false; + } + + if (len == 0 && quoting_style == shell_always_quoting_style + && elide_outer_quotes) + goto force_outer_quoting_style; + + /* Single shell quotes (') are commonly enough used as an apostrophe, + that we attempt to minimize the quoting in this case. Note itʼs + better to use the apostrophe modifier "\u02BC" if possible, as that + renders better and works with the word match regex \W+ etc. */ + if (quoting_style == shell_always_quoting_style && ! elide_outer_quotes + && encountered_single_quote) + { + if (all_c_and_shell_quote_compat) + return quotearg_buffer_restyled (buffer, orig_buffersize, arg, argsize, + c_quoting_style, + flags, quote_these_too, + left_quote, right_quote); + else if (! buffersize && orig_buffersize) + { + /* Disable read-only scan, and reprocess to write quoted string. */ + buffersize = orig_buffersize; + len = 0; + goto process_input; + } + } + + if (quote_string && !elide_outer_quotes) + for (; *quote_string; quote_string++) + STORE (*quote_string); + + if (len < buffersize) + buffer[len] = '\0'; + return len; + + force_outer_quoting_style: + /* Don't reuse quote_these_too, since the addition of outer quotes + sufficiently quotes the specified characters. */ + if (quoting_style == shell_always_quoting_style && backslash_escapes) + quoting_style = shell_escape_always_quoting_style; + return quotearg_buffer_restyled (buffer, buffersize, arg, argsize, + quoting_style, + flags & ~QA_ELIDE_OUTER_QUOTES, NULL, + left_quote, right_quote); +} + +/* Place into buffer BUFFER (of size BUFFERSIZE) a quoted version of + argument ARG (of size ARGSIZE), using O to control quoting. + If O is null, use the default. + Terminate the output with a null character, and return the written + size of the output, not counting the terminating null. + If BUFFERSIZE is too small to store the output string, return the + value that would have been returned had BUFFERSIZE been large enough. + If ARGSIZE is SIZE_MAX, use the string length of the argument for + ARGSIZE. */ +size_t +quotearg_buffer (char *buffer, size_t buffersize, + char const *arg, size_t argsize, + struct quoting_options const *o) +{ + struct quoting_options const *p = o ? o : &default_quoting_options; + int e = errno; + size_t r = quotearg_buffer_restyled (buffer, buffersize, arg, argsize, + p->style, p->flags, p->quote_these_too, + p->left_quote, p->right_quote); + errno = e; + return r; +} + +/* Equivalent to quotearg_alloc (ARG, ARGSIZE, NULL, O). */ +char * +quotearg_alloc (char const *arg, size_t argsize, + struct quoting_options const *o) +{ + return quotearg_alloc_mem (arg, argsize, NULL, o); +} + +/* Like quotearg_buffer (..., ARG, ARGSIZE, O), except return newly + allocated storage containing the quoted string, and store the + resulting size into *SIZE, if non-NULL. The result can contain + embedded null bytes only if ARGSIZE is not SIZE_MAX, SIZE is not + NULL, and set_quoting_flags has not set the null byte elision + flag. */ +char * +quotearg_alloc_mem (char const *arg, size_t argsize, size_t *size, + struct quoting_options const *o) +{ + struct quoting_options const *p = o ? o : &default_quoting_options; + int e = errno; + /* Elide embedded null bytes if we can't return a size. */ + int flags = p->flags | (size ? 0 : QA_ELIDE_NULL_BYTES); + size_t bufsize = quotearg_buffer_restyled (0, 0, arg, argsize, p->style, + flags, p->quote_these_too, + p->left_quote, + p->right_quote) + 1; + char *buf = xcharalloc (bufsize); + quotearg_buffer_restyled (buf, bufsize, arg, argsize, p->style, flags, + p->quote_these_too, + p->left_quote, p->right_quote); + errno = e; + if (size) + *size = bufsize - 1; + return buf; +} + +/* A storage slot with size and pointer to a value. */ +struct slotvec +{ + size_t size; + char *val; +}; + +/* Preallocate a slot 0 buffer, so that the caller can always quote + one small component of a "memory exhausted" message in slot 0. */ +static char slot0[256]; +static int nslots = 1; +static struct slotvec slotvec0 = {sizeof slot0, slot0}; +static struct slotvec *slotvec = &slotvec0; + +void +quotearg_free (void) +{ + struct slotvec *sv = slotvec; + int i; + for (i = 1; i < nslots; i++) + free (sv[i].val); + if (sv[0].val != slot0) + { + free (sv[0].val); + slotvec0.size = sizeof slot0; + slotvec0.val = slot0; + } + if (sv != &slotvec0) + { + free (sv); + slotvec = &slotvec0; + } + nslots = 1; +} + +/* Use storage slot N to return a quoted version of argument ARG. + ARG is of size ARGSIZE, but if that is SIZE_MAX, ARG is a + null-terminated string. + OPTIONS specifies the quoting options. + The returned value points to static storage that can be + reused by the next call to this function with the same value of N. + N must be nonnegative. N is deliberately declared with type "int" + to allow for future extensions (using negative values). */ +static char * +quotearg_n_options (int n, char const *arg, size_t argsize, + struct quoting_options const *options) +{ + int e = errno; + + struct slotvec *sv = slotvec; + + if (n < 0) + abort (); + + if (nslots <= n) + { + bool preallocated = (sv == &slotvec0); + int nmax = MIN (INT_MAX, MIN (PTRDIFF_MAX, SIZE_MAX) / sizeof *sv) - 1; + + if (nmax < n) + xalloc_die (); + + slotvec = sv = xrealloc (preallocated ? NULL : sv, (n + 1) * sizeof *sv); + if (preallocated) + *sv = slotvec0; + memset (sv + nslots, 0, (n + 1 - nslots) * sizeof *sv); + nslots = n + 1; + } + + { + size_t size = sv[n].size; + char *val = sv[n].val; + /* Elide embedded null bytes since we don't return a size. */ + int flags = options->flags | QA_ELIDE_NULL_BYTES; + size_t qsize = quotearg_buffer_restyled (val, size, arg, argsize, + options->style, flags, + options->quote_these_too, + options->left_quote, + options->right_quote); + + if (size <= qsize) + { + sv[n].size = size = qsize + 1; + if (val != slot0) + free (val); + sv[n].val = val = xcharalloc (size); + quotearg_buffer_restyled (val, size, arg, argsize, options->style, + flags, options->quote_these_too, + options->left_quote, + options->right_quote); + } + + errno = e; + return val; + } +} + +char * +quotearg_n (int n, char const *arg) +{ + return quotearg_n_options (n, arg, SIZE_MAX, &default_quoting_options); +} + +char * +quotearg_n_mem (int n, char const *arg, size_t argsize) +{ + return quotearg_n_options (n, arg, argsize, &default_quoting_options); +} + +char * +quotearg (char const *arg) +{ + return quotearg_n (0, arg); +} + +char * +quotearg_mem (char const *arg, size_t argsize) +{ + return quotearg_n_mem (0, arg, argsize); +} + +char * +quotearg_n_style (int n, enum quoting_style s, char const *arg) +{ + struct quoting_options const o = quoting_options_from_style (s); + return quotearg_n_options (n, arg, SIZE_MAX, &o); +} + +char * +quotearg_n_style_mem (int n, enum quoting_style s, + char const *arg, size_t argsize) +{ + struct quoting_options const o = quoting_options_from_style (s); + return quotearg_n_options (n, arg, argsize, &o); +} + +char * +quotearg_style (enum quoting_style s, char const *arg) +{ + return quotearg_n_style (0, s, arg); +} + +char * +quotearg_style_mem (enum quoting_style s, char const *arg, size_t argsize) +{ + return quotearg_n_style_mem (0, s, arg, argsize); +} + +char * +quotearg_char_mem (char const *arg, size_t argsize, char ch) +{ + struct quoting_options options; + options = default_quoting_options; + set_char_quoting (&options, ch, 1); + return quotearg_n_options (0, arg, argsize, &options); +} + +char * +quotearg_char (char const *arg, char ch) +{ + return quotearg_char_mem (arg, SIZE_MAX, ch); +} + +char * +quotearg_colon (char const *arg) +{ + return quotearg_char (arg, ':'); +} + +char * +quotearg_colon_mem (char const *arg, size_t argsize) +{ + return quotearg_char_mem (arg, argsize, ':'); +} + +char * +quotearg_n_style_colon (int n, enum quoting_style s, char const *arg) +{ + struct quoting_options options; + options = quoting_options_from_style (s); + set_char_quoting (&options, ':', 1); + return quotearg_n_options (n, arg, SIZE_MAX, &options); +} + +char * +quotearg_n_custom (int n, char const *left_quote, + char const *right_quote, char const *arg) +{ + return quotearg_n_custom_mem (n, left_quote, right_quote, arg, + SIZE_MAX); +} + +char * +quotearg_n_custom_mem (int n, char const *left_quote, + char const *right_quote, + char const *arg, size_t argsize) +{ + struct quoting_options o = default_quoting_options; + set_custom_quoting (&o, left_quote, right_quote); + return quotearg_n_options (n, arg, argsize, &o); +} + +char * +quotearg_custom (char const *left_quote, char const *right_quote, + char const *arg) +{ + return quotearg_n_custom (0, left_quote, right_quote, arg); +} + +char * +quotearg_custom_mem (char const *left_quote, char const *right_quote, + char const *arg, size_t argsize) +{ + return quotearg_n_custom_mem (0, left_quote, right_quote, arg, + argsize); +} + + +/* The quoting option used by the functions of quote.h. */ +struct quoting_options quote_quoting_options = + { + locale_quoting_style, + 0, + { 0 }, + NULL, NULL + }; + +char const * +quote_n_mem (int n, char const *arg, size_t argsize) +{ + return quotearg_n_options (n, arg, argsize, "e_quoting_options); +} + +char const * +quote_mem (char const *arg, size_t argsize) +{ + return quote_n_mem (0, arg, argsize); +} + +char const * +quote_n (int n, char const *arg) +{ + return quote_n_mem (n, arg, SIZE_MAX); +} + +char const * +quote (char const *arg) +{ + return quote_n (0, arg); +} + +/* + * Hey Emacs! + * Local Variables: + * coding: utf-8 + * End: + */ diff --git a/lib/quotearg.h b/lib/quotearg.h new file mode 100644 index 0000000..0584c56 --- /dev/null +++ b/lib/quotearg.h @@ -0,0 +1,425 @@ +/* quotearg.h - quote arguments for output + + Copyright (C) 1998-2002, 2004, 2006, 2008-2018 Free Software Foundation, + Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +/* Written by Paul Eggert <eggert@twinsun.com> */ + +#ifndef QUOTEARG_H_ +# define QUOTEARG_H_ 1 + +# include <stddef.h> + +/* Basic quoting styles. For each style, an example is given on the + input strings "simple", "\0 \t\n'\"\033?""?/\\", and "a:b", using + quotearg_buffer, quotearg_mem, and quotearg_colon_mem with that + style and the default flags and quoted characters. Note that the + examples are shown here as valid C strings rather than what + displays on a terminal (with "??/" as a trigraph for "\\"). */ +enum quoting_style + { + /* Output names as-is (ls --quoting-style=literal). Can result in + embedded null bytes if QA_ELIDE_NULL_BYTES is not in + effect. + + quotearg_buffer: + "simple", "\0 \t\n'\"\033??/\\", "a:b" + quotearg: + "simple", " \t\n'\"\033??/\\", "a:b" + quotearg_colon: + "simple", " \t\n'\"\033??/\\", "a:b" + */ + literal_quoting_style, + + /* Quote names for the shell if they contain shell metacharacters + or would cause ambiguous output (ls --quoting-style=shell). + Can result in embedded null bytes if QA_ELIDE_NULL_BYTES is not + in effect. + + quotearg_buffer: + "simple", "'\0 \t\n'\\''\"\033??/\\'", "a:b" + quotearg: + "simple", "' \t\n'\\''\"\033??/\\'", "a:b" + quotearg_colon: + "simple", "' \t\n'\\''\"\033??/\\'", "'a:b'" + */ + shell_quoting_style, + + /* Quote names for the shell, even if they would normally not + require quoting (ls --quoting-style=shell-always). Can result + in embedded null bytes if QA_ELIDE_NULL_BYTES is not in effect. + Behaves like shell_quoting_style if QA_ELIDE_OUTER_QUOTES is in + effect. + + quotearg_buffer: + "'simple'", "'\0 \t\n'\\''\"\033??/\\'", "'a:b'" + quotearg: + "'simple'", "' \t\n'\\''\"\033??/\\'", "'a:b'" + quotearg_colon: + "'simple'", "' \t\n'\\''\"\033??/\\'", "'a:b'" + */ + shell_always_quoting_style, + + /* Quote names for the shell if they contain shell metacharacters + or other problematic characters (ls --quoting-style=shell-escape). + Non printable characters are quoted using the $'...' syntax, + which originated in ksh93 and is widely supported by most shells, + and proposed for inclusion in POSIX. + + quotearg_buffer: + "simple", "''$'\\0'' '$'\\t\\n'\\''\"'$'\\033''??/\\'", "a:b" + quotearg: + "simple", "''$'\\0'' '$'\\t\\n'\\''\"'$'\\033''??/\\'", "a:b" + quotearg_colon: + "simple", "''$'\\0'' '$'\\t\\n'\\''\"'$'\\033''??/\\'", "'a:b'" + */ + shell_escape_quoting_style, + + /* Quote names for the shell even if they would normally not + require quoting (ls --quoting-style=shell-escape). + Non printable characters are quoted using the $'...' syntax, + which originated in ksh93 and is widely supported by most shells, + and proposed for inclusion in POSIX. Behaves like + shell_escape_quoting_style if QA_ELIDE_OUTER_QUOTES is in effect. + + quotearg_buffer: + "simple", "''$'\\0'' '$'\\t\\n'\\''\"'$'\\033''??/\'", "a:b" + quotearg: + "simple", "''$'\\0'' '$'\\t\\n'\\''\"'$'\\033''??/\'", "a:b" + quotearg_colon: + "simple", "''$'\\0'' '$'\\t\\n'\\''\"'$'\\033''??/\'", "'a:b'" + */ + shell_escape_always_quoting_style, + + /* Quote names as for a C language string (ls --quoting-style=c). + Behaves like c_maybe_quoting_style if QA_ELIDE_OUTER_QUOTES is + in effect. Split into consecutive strings if + QA_SPLIT_TRIGRAPHS. + + quotearg_buffer: + "\"simple\"", "\"\\0 \\t\\n'\\\"\\033??/\\\\\"", "\"a:b\"" + quotearg: + "\"simple\"", "\"\\0 \\t\\n'\\\"\\033??/\\\\\"", "\"a:b\"" + quotearg_colon: + "\"simple\"", "\"\\0 \\t\\n'\\\"\\033??/\\\\\"", "\"a\\:b\"" + */ + c_quoting_style, + + /* Like c_quoting_style except omit the surrounding double-quote + characters if no quoted characters are encountered. + + quotearg_buffer: + "simple", "\"\\0 \\t\\n'\\\"\\033??/\\\\\"", "a:b" + quotearg: + "simple", "\"\\0 \\t\\n'\\\"\\033??/\\\\\"", "a:b" + quotearg_colon: + "simple", "\"\\0 \\t\\n'\\\"\\033??/\\\\\"", "\"a:b\"" + */ + c_maybe_quoting_style, + + /* Like c_quoting_style except always omit the surrounding + double-quote characters and ignore QA_SPLIT_TRIGRAPHS + (ls --quoting-style=escape). + + quotearg_buffer: + "simple", "\\0 \\t\\n'\"\\033??/\\\\", "a:b" + quotearg: + "simple", "\\0 \\t\\n'\"\\033??/\\\\", "a:b" + quotearg_colon: + "simple", "\\0 \\t\\n'\"\\033??/\\\\", "a\\:b" + */ + escape_quoting_style, + + /* Like clocale_quoting_style, but use single quotes in the + default C locale or if the program does not use gettext + (ls --quoting-style=locale). For UTF-8 locales, quote + characters will use Unicode. + + LC_MESSAGES=C + quotearg_buffer: + "`simple'", "`\\0 \\t\\n\\'\"\\033??/\\\\'", "`a:b'" + quotearg: + "`simple'", "`\\0 \\t\\n\\'\"\\033??/\\\\'", "`a:b'" + quotearg_colon: + "`simple'", "`\\0 \\t\\n\\'\"\\033??/\\\\'", "`a\\:b'" + + LC_MESSAGES=pt_PT.utf8 + quotearg_buffer: + "\302\253simple\302\273", + "\302\253\\0 \\t\\n'\"\\033??/\\\\\302\253", "\302\253a:b\302\273" + quotearg: + "\302\253simple\302\273", + "\302\253\\0 \\t\\n'\"\\033??/\\\\\302\253", "\302\253a:b\302\273" + quotearg_colon: + "\302\253simple\302\273", + "\302\253\\0 \\t\\n'\"\\033??/\\\\\302\253", "\302\253a\\:b\302\273" + */ + locale_quoting_style, + + /* Like c_quoting_style except use quotation marks appropriate for + the locale and ignore QA_SPLIT_TRIGRAPHS + (ls --quoting-style=clocale). + + LC_MESSAGES=C + quotearg_buffer: + "\"simple\"", "\"\\0 \\t\\n'\\\"\\033??/\\\\\"", "\"a:b\"" + quotearg: + "\"simple\"", "\"\\0 \\t\\n'\\\"\\033??/\\\\\"", "\"a:b\"" + quotearg_colon: + "\"simple\"", "\"\\0 \\t\\n'\\\"\\033??/\\\\\"", "\"a\\:b\"" + + LC_MESSAGES=pt_PT.utf8 + quotearg_buffer: + "\302\253simple\302\273", + "\302\253\\0 \\t\\n'\"\\033??/\\\\\302\253", "\302\253a:b\302\273" + quotearg: + "\302\253simple\302\273", + "\302\253\\0 \\t\\n'\"\\033??/\\\\\302\253", "\302\253a:b\302\273" + quotearg_colon: + "\302\253simple\302\273", + "\302\253\\0 \\t\\n'\"\\033??/\\\\\302\253", "\302\253a\\:b\302\273" + */ + clocale_quoting_style, + + /* Like clocale_quoting_style except use the custom quotation marks + set by set_custom_quoting. If custom quotation marks are not + set, the behavior is undefined. + + left_quote = right_quote = "'" + quotearg_buffer: + "'simple'", "'\\0 \\t\\n\\'\"\\033??/\\\\'", "'a:b'" + quotearg: + "'simple'", "'\\0 \\t\\n\\'\"\\033??/\\\\'", "'a:b'" + quotearg_colon: + "'simple'", "'\\0 \\t\\n\\'\"\\033??/\\\\'", "'a\\:b'" + + left_quote = "(" and right_quote = ")" + quotearg_buffer: + "(simple)", "(\\0 \\t\\n'\"\\033??/\\\\)", "(a:b)" + quotearg: + "(simple)", "(\\0 \\t\\n'\"\\033??/\\\\)", "(a:b)" + quotearg_colon: + "(simple)", "(\\0 \\t\\n'\"\\033??/\\\\)", "(a\\:b)" + + left_quote = ":" and right_quote = " " + quotearg_buffer: + ":simple ", ":\\0\\ \\t\\n'\"\\033??/\\\\ ", ":a:b " + quotearg: + ":simple ", ":\\0\\ \\t\\n'\"\\033??/\\\\ ", ":a:b " + quotearg_colon: + ":simple ", ":\\0\\ \\t\\n'\"\\033??/\\\\ ", ":a\\:b " + + left_quote = "\"'" and right_quote = "'\"" + Notice that this is treated as a single level of quotes or two + levels where the outer quote need not be escaped within the inner + quotes. For two levels where the outer quote must be escaped + within the inner quotes, you must use separate quotearg + invocations. + quotearg_buffer: + "\"'simple'\"", "\"'\\0 \\t\\n\\'\"\\033??/\\\\'\"", "\"'a:b'\"" + quotearg: + "\"'simple'\"", "\"'\\0 \\t\\n\\'\"\\033??/\\\\'\"", "\"'a:b'\"" + quotearg_colon: + "\"'simple'\"", "\"'\\0 \\t\\n\\'\"\\033??/\\\\'\"", "\"'a\\:b'\"" + */ + custom_quoting_style + }; + +/* Flags for use in set_quoting_flags. */ +enum quoting_flags + { + /* Always elide null bytes from styles that do not quote them, + even when the length of the result is available to the + caller. */ + QA_ELIDE_NULL_BYTES = 0x01, + + /* Omit the surrounding quote characters if no escaped characters + are encountered. Note that if no other character needs + escaping, then neither does the escape character. */ + QA_ELIDE_OUTER_QUOTES = 0x02, + + /* In the c_quoting_style and c_maybe_quoting_style, split ANSI + trigraph sequences into concatenated strings (for example, + "?""?/" rather than "??/", which could be confused with + "\\"). */ + QA_SPLIT_TRIGRAPHS = 0x04 + }; + +/* For now, --quoting-style=literal is the default, but this may change. */ +# ifndef DEFAULT_QUOTING_STYLE +# define DEFAULT_QUOTING_STYLE literal_quoting_style +# endif + +/* Names of quoting styles and their corresponding values. */ +extern char const *const quoting_style_args[]; +extern enum quoting_style const quoting_style_vals[]; + +struct quoting_options; + +/* The functions listed below set and use a hidden variable + that contains the default quoting style options. */ + +/* Allocate a new set of quoting options, with contents initially identical + to O if O is not null, or to the default if O is null. + It is the caller's responsibility to free the result. */ +struct quoting_options *clone_quoting_options (struct quoting_options *o); + +/* Get the value of O's quoting style. If O is null, use the default. */ +enum quoting_style get_quoting_style (struct quoting_options const *o); + +/* In O (or in the default if O is null), + set the value of the quoting style to S. */ +void set_quoting_style (struct quoting_options *o, enum quoting_style s); + +/* In O (or in the default if O is null), + set the value of the quoting options for character C to I. + Return the old value. Currently, the only values defined for I are + 0 (the default) and 1 (which means to quote the character even if + it would not otherwise be quoted). C must never be a digit or a + letter that has special meaning after a backslash (for example, "\t" + for tab). */ +int set_char_quoting (struct quoting_options *o, char c, int i); + +/* In O (or in the default if O is null), + set the value of the quoting options flag to I, which can be a + bitwise combination of enum quoting_flags, or 0 for default + behavior. Return the old value. */ +int set_quoting_flags (struct quoting_options *o, int i); + +/* In O (or in the default if O is null), + set the value of the quoting style to custom_quoting_style, + set the left quote to LEFT_QUOTE, and set the right quote to + RIGHT_QUOTE. Each of LEFT_QUOTE and RIGHT_QUOTE must be + null-terminated and can be the empty string. Because backslashes are + used for escaping, it does not make sense for RIGHT_QUOTE to contain + a backslash. RIGHT_QUOTE must not begin with a digit or a letter + that has special meaning after a backslash (for example, "\t" for + tab). */ +void set_custom_quoting (struct quoting_options *o, + char const *left_quote, + char const *right_quote); + +/* Place into buffer BUFFER (of size BUFFERSIZE) a quoted version of + argument ARG (of size ARGSIZE), using O to control quoting. + If O is null, use the default. + Terminate the output with a null character, and return the written + size of the output, not counting the terminating null. + If BUFFERSIZE is too small to store the output string, return the + value that would have been returned had BUFFERSIZE been large enough. + If ARGSIZE is -1, use the string length of the argument for ARGSIZE. + On output, BUFFER might contain embedded null bytes if ARGSIZE was + not -1, the style of O does not use backslash escapes, and the + flags of O do not request elision of null bytes.*/ +size_t quotearg_buffer (char *buffer, size_t buffersize, + char const *arg, size_t argsize, + struct quoting_options const *o); + +/* Like quotearg_buffer, except return the result in a newly allocated + buffer. It is the caller's responsibility to free the result. The + result will not contain embedded null bytes. */ +char *quotearg_alloc (char const *arg, size_t argsize, + struct quoting_options const *o); + +/* Like quotearg_alloc, except that the length of the result, + excluding the terminating null byte, is stored into SIZE if it is + non-NULL. The result might contain embedded null bytes if ARGSIZE + was not -1, SIZE was not NULL, the style of O does not use + backslash escapes, and the flags of O do not request elision of + null bytes.*/ +char *quotearg_alloc_mem (char const *arg, size_t argsize, + size_t *size, struct quoting_options const *o); + +/* Use storage slot N to return a quoted version of the string ARG. + Use the default quoting options. + The returned value points to static storage that can be + reused by the next call to this function with the same value of N. + N must be nonnegative. The output of all functions in the + quotearg_n family are guaranteed to not contain embedded null + bytes.*/ +char *quotearg_n (int n, char const *arg); + +/* Equivalent to quotearg_n (0, ARG). */ +char *quotearg (char const *arg); + +/* Use storage slot N to return a quoted version of the argument ARG + of size ARGSIZE. This is like quotearg_n (N, ARG), except it can + quote null bytes. */ +char *quotearg_n_mem (int n, char const *arg, size_t argsize); + +/* Equivalent to quotearg_n_mem (0, ARG, ARGSIZE). */ +char *quotearg_mem (char const *arg, size_t argsize); + +/* Use style S and storage slot N to return a quoted version of the string ARG. + This is like quotearg_n (N, ARG), except that it uses S with no other + options to specify the quoting method. */ +char *quotearg_n_style (int n, enum quoting_style s, char const *arg); + +/* Use style S and storage slot N to return a quoted version of the + argument ARG of size ARGSIZE. This is like quotearg_n_style + (N, S, ARG), except it can quote null bytes. */ +char *quotearg_n_style_mem (int n, enum quoting_style s, + char const *arg, size_t argsize); + +/* Equivalent to quotearg_n_style (0, S, ARG). */ +char *quotearg_style (enum quoting_style s, char const *arg); + +/* Equivalent to quotearg_n_style_mem (0, S, ARG, ARGSIZE). */ +char *quotearg_style_mem (enum quoting_style s, + char const *arg, size_t argsize); + +/* Like quotearg (ARG), except also quote any instances of CH. + See set_char_quoting for a description of acceptable CH values. */ +char *quotearg_char (char const *arg, char ch); + +/* Like quotearg_char (ARG, CH), except it can quote null bytes. */ +char *quotearg_char_mem (char const *arg, size_t argsize, char ch); + +/* Equivalent to quotearg_char (ARG, ':'). */ +char *quotearg_colon (char const *arg); + +/* Like quotearg_colon (ARG), except it can quote null bytes. */ +char *quotearg_colon_mem (char const *arg, size_t argsize); + +/* Like quotearg_n_style, except with ':' quoting enabled. */ +char *quotearg_n_style_colon (int n, enum quoting_style s, char const *arg); + +/* Like quotearg_n_style (N, S, ARG) but with S as custom_quoting_style + with left quote as LEFT_QUOTE and right quote as RIGHT_QUOTE. See + set_custom_quoting for a description of acceptable LEFT_QUOTE and + RIGHT_QUOTE values. */ +char *quotearg_n_custom (int n, char const *left_quote, + char const *right_quote, char const *arg); + +/* Like quotearg_n_custom (N, LEFT_QUOTE, RIGHT_QUOTE, ARG) except it + can quote null bytes. */ +char *quotearg_n_custom_mem (int n, char const *left_quote, + char const *right_quote, + char const *arg, size_t argsize); + +/* Equivalent to quotearg_n_custom (0, LEFT_QUOTE, RIGHT_QUOTE, ARG). */ +char *quotearg_custom (char const *left_quote, char const *right_quote, + char const *arg); + +/* Equivalent to quotearg_n_custom_mem (0, LEFT_QUOTE, RIGHT_QUOTE, ARG, + ARGSIZE). */ +char *quotearg_custom_mem (char const *left_quote, + char const *right_quote, + char const *arg, size_t argsize); + +/* Free any dynamically allocated memory. */ +void quotearg_free (void); + +#endif /* !QUOTEARG_H_ */ diff --git a/lib/readlink.c b/lib/readlink.c new file mode 100644 index 0000000..924e00d --- /dev/null +++ b/lib/readlink.c @@ -0,0 +1,74 @@ +/* Stub for readlink(). + Copyright (C) 2003-2007, 2009-2018 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +#include <config.h> + +/* Specification. */ +#include <unistd.h> + +#include <errno.h> +#include <string.h> +#include <sys/stat.h> + +#if !HAVE_READLINK + +/* readlink() substitute for systems that don't have a readlink() function, + such as DJGPP 2.03 and mingw32. */ + +ssize_t +readlink (const char *name, char *buf _GL_UNUSED, + size_t bufsize _GL_UNUSED) +{ + struct stat statbuf; + + /* In general we should use lstat() here, not stat(). But on platforms + without symbolic links, lstat() - if it exists - would be equivalent to + stat(), therefore we can use stat(). This saves us a configure check. */ + if (stat (name, &statbuf) >= 0) + errno = EINVAL; + return -1; +} + +#else /* HAVE_READLINK */ + +# undef readlink + +/* readlink() wrapper that uses correct types, for systems like cygwin + 1.5.x where readlink returns int, and which rejects trailing slash, + for Solaris 9. */ + +ssize_t +rpl_readlink (const char *name, char *buf, size_t bufsize) +{ +# if READLINK_TRAILING_SLASH_BUG + size_t len = strlen (name); + if (len && name[len - 1] == '/') + { + /* Even if name without the slash is a symlink to a directory, + both lstat() and stat() must resolve the trailing slash to + the directory rather than the symlink. We can therefore + safely use stat() to distinguish between EINVAL and + ENOTDIR/ENOENT, avoiding extra overhead of rpl_lstat(). */ + struct stat st; + if (stat (name, &st) == 0) + errno = EINVAL; + return -1; + } +# endif /* READLINK_TRAILING_SLASH_BUG */ + return readlink (name, buf, bufsize); +} + +#endif /* HAVE_READLINK */ diff --git a/lib/regcomp.c b/lib/regcomp.c new file mode 100644 index 0000000..0b05a63 --- /dev/null +++ b/lib/regcomp.c @@ -0,0 +1,3946 @@ +/* Extended regular expression matching and search library. + Copyright (C) 2002-2018 Free Software Foundation, Inc. + This file is part of the GNU C Library. + Contributed by Isamu Hasegawa <isamu@yamato.ibm.com>. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU General Public + License as published by the Free Software Foundation; either + version 3 of the License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + General Public License for more details. + + You should have received a copy of the GNU General Public + License along with the GNU C Library; if not, see + <https://www.gnu.org/licenses/>. */ + +#ifdef _LIBC +# include <locale/weight.h> +#endif + +static reg_errcode_t re_compile_internal (regex_t *preg, const char * pattern, + size_t length, reg_syntax_t syntax); +static void re_compile_fastmap_iter (regex_t *bufp, + const re_dfastate_t *init_state, + char *fastmap); +static reg_errcode_t init_dfa (re_dfa_t *dfa, size_t pat_len); +#ifdef RE_ENABLE_I18N +static void free_charset (re_charset_t *cset); +#endif /* RE_ENABLE_I18N */ +static void free_workarea_compile (regex_t *preg); +static reg_errcode_t create_initial_state (re_dfa_t *dfa); +#ifdef RE_ENABLE_I18N +static void optimize_utf8 (re_dfa_t *dfa); +#endif +static reg_errcode_t analyze (regex_t *preg); +static reg_errcode_t preorder (bin_tree_t *root, + reg_errcode_t (fn (void *, bin_tree_t *)), + void *extra); +static reg_errcode_t postorder (bin_tree_t *root, + reg_errcode_t (fn (void *, bin_tree_t *)), + void *extra); +static reg_errcode_t optimize_subexps (void *extra, bin_tree_t *node); +static reg_errcode_t lower_subexps (void *extra, bin_tree_t *node); +static bin_tree_t *lower_subexp (reg_errcode_t *err, regex_t *preg, + bin_tree_t *node); +static reg_errcode_t calc_first (void *extra, bin_tree_t *node); +static reg_errcode_t calc_next (void *extra, bin_tree_t *node); +static reg_errcode_t link_nfa_nodes (void *extra, bin_tree_t *node); +static Idx duplicate_node (re_dfa_t *dfa, Idx org_idx, unsigned int constraint); +static Idx search_duplicated_node (const re_dfa_t *dfa, Idx org_node, + unsigned int constraint); +static reg_errcode_t calc_eclosure (re_dfa_t *dfa); +static reg_errcode_t calc_eclosure_iter (re_node_set *new_set, re_dfa_t *dfa, + Idx node, bool root); +static reg_errcode_t calc_inveclosure (re_dfa_t *dfa); +static Idx fetch_number (re_string_t *input, re_token_t *token, + reg_syntax_t syntax); +static int peek_token (re_token_t *token, re_string_t *input, + reg_syntax_t syntax); +static bin_tree_t *parse (re_string_t *regexp, regex_t *preg, + reg_syntax_t syntax, reg_errcode_t *err); +static bin_tree_t *parse_reg_exp (re_string_t *regexp, regex_t *preg, + re_token_t *token, reg_syntax_t syntax, + Idx nest, reg_errcode_t *err); +static bin_tree_t *parse_branch (re_string_t *regexp, regex_t *preg, + re_token_t *token, reg_syntax_t syntax, + Idx nest, reg_errcode_t *err); +static bin_tree_t *parse_expression (re_string_t *regexp, regex_t *preg, + re_token_t *token, reg_syntax_t syntax, + Idx nest, reg_errcode_t *err); +static bin_tree_t *parse_sub_exp (re_string_t *regexp, regex_t *preg, + re_token_t *token, reg_syntax_t syntax, + Idx nest, reg_errcode_t *err); +static bin_tree_t *parse_dup_op (bin_tree_t *dup_elem, re_string_t *regexp, + re_dfa_t *dfa, re_token_t *token, + reg_syntax_t syntax, reg_errcode_t *err); +static bin_tree_t *parse_bracket_exp (re_string_t *regexp, re_dfa_t *dfa, + re_token_t *token, reg_syntax_t syntax, + reg_errcode_t *err); +static reg_errcode_t parse_bracket_element (bracket_elem_t *elem, + re_string_t *regexp, + re_token_t *token, int token_len, + re_dfa_t *dfa, + reg_syntax_t syntax, + bool accept_hyphen); +static reg_errcode_t parse_bracket_symbol (bracket_elem_t *elem, + re_string_t *regexp, + re_token_t *token); +#ifdef RE_ENABLE_I18N +static reg_errcode_t build_equiv_class (bitset_t sbcset, + re_charset_t *mbcset, + Idx *equiv_class_alloc, + const unsigned char *name); +static reg_errcode_t build_charclass (RE_TRANSLATE_TYPE trans, + bitset_t sbcset, + re_charset_t *mbcset, + Idx *char_class_alloc, + const char *class_name, + reg_syntax_t syntax); +#else /* not RE_ENABLE_I18N */ +static reg_errcode_t build_equiv_class (bitset_t sbcset, + const unsigned char *name); +static reg_errcode_t build_charclass (RE_TRANSLATE_TYPE trans, + bitset_t sbcset, + const char *class_name, + reg_syntax_t syntax); +#endif /* not RE_ENABLE_I18N */ +static bin_tree_t *build_charclass_op (re_dfa_t *dfa, + RE_TRANSLATE_TYPE trans, + const char *class_name, + const char *extra, + bool non_match, reg_errcode_t *err); +static bin_tree_t *create_tree (re_dfa_t *dfa, + bin_tree_t *left, bin_tree_t *right, + re_token_type_t type); +static bin_tree_t *create_token_tree (re_dfa_t *dfa, + bin_tree_t *left, bin_tree_t *right, + const re_token_t *token); +static bin_tree_t *duplicate_tree (const bin_tree_t *src, re_dfa_t *dfa); +static void free_token (re_token_t *node); +static reg_errcode_t free_tree (void *extra, bin_tree_t *node); +static reg_errcode_t mark_opt_subexp (void *extra, bin_tree_t *node); + +/* This table gives an error message for each of the error codes listed + in regex.h. Obviously the order here has to be same as there. + POSIX doesn't require that we do anything for REG_NOERROR, + but why not be nice? */ + +static const char __re_error_msgid[] = + { +#define REG_NOERROR_IDX 0 + gettext_noop ("Success") /* REG_NOERROR */ + "\0" +#define REG_NOMATCH_IDX (REG_NOERROR_IDX + sizeof "Success") + gettext_noop ("No match") /* REG_NOMATCH */ + "\0" +#define REG_BADPAT_IDX (REG_NOMATCH_IDX + sizeof "No match") + gettext_noop ("Invalid regular expression") /* REG_BADPAT */ + "\0" +#define REG_ECOLLATE_IDX (REG_BADPAT_IDX + sizeof "Invalid regular expression") + gettext_noop ("Invalid collation character") /* REG_ECOLLATE */ + "\0" +#define REG_ECTYPE_IDX (REG_ECOLLATE_IDX + sizeof "Invalid collation character") + gettext_noop ("Invalid character class name") /* REG_ECTYPE */ + "\0" +#define REG_EESCAPE_IDX (REG_ECTYPE_IDX + sizeof "Invalid character class name") + gettext_noop ("Trailing backslash") /* REG_EESCAPE */ + "\0" +#define REG_ESUBREG_IDX (REG_EESCAPE_IDX + sizeof "Trailing backslash") + gettext_noop ("Invalid back reference") /* REG_ESUBREG */ + "\0" +#define REG_EBRACK_IDX (REG_ESUBREG_IDX + sizeof "Invalid back reference") + gettext_noop ("Unmatched [, [^, [:, [., or [=") /* REG_EBRACK */ + "\0" +#define REG_EPAREN_IDX (REG_EBRACK_IDX + sizeof "Unmatched [, [^, [:, [., or [=") + gettext_noop ("Unmatched ( or \\(") /* REG_EPAREN */ + "\0" +#define REG_EBRACE_IDX (REG_EPAREN_IDX + sizeof "Unmatched ( or \\(") + gettext_noop ("Unmatched \\{") /* REG_EBRACE */ + "\0" +#define REG_BADBR_IDX (REG_EBRACE_IDX + sizeof "Unmatched \\{") + gettext_noop ("Invalid content of \\{\\}") /* REG_BADBR */ + "\0" +#define REG_ERANGE_IDX (REG_BADBR_IDX + sizeof "Invalid content of \\{\\}") + gettext_noop ("Invalid range end") /* REG_ERANGE */ + "\0" +#define REG_ESPACE_IDX (REG_ERANGE_IDX + sizeof "Invalid range end") + gettext_noop ("Memory exhausted") /* REG_ESPACE */ + "\0" +#define REG_BADRPT_IDX (REG_ESPACE_IDX + sizeof "Memory exhausted") + gettext_noop ("Invalid preceding regular expression") /* REG_BADRPT */ + "\0" +#define REG_EEND_IDX (REG_BADRPT_IDX + sizeof "Invalid preceding regular expression") + gettext_noop ("Premature end of regular expression") /* REG_EEND */ + "\0" +#define REG_ESIZE_IDX (REG_EEND_IDX + sizeof "Premature end of regular expression") + gettext_noop ("Regular expression too big") /* REG_ESIZE */ + "\0" +#define REG_ERPAREN_IDX (REG_ESIZE_IDX + sizeof "Regular expression too big") + gettext_noop ("Unmatched ) or \\)") /* REG_ERPAREN */ + }; + +static const size_t __re_error_msgid_idx[] = + { + REG_NOERROR_IDX, + REG_NOMATCH_IDX, + REG_BADPAT_IDX, + REG_ECOLLATE_IDX, + REG_ECTYPE_IDX, + REG_EESCAPE_IDX, + REG_ESUBREG_IDX, + REG_EBRACK_IDX, + REG_EPAREN_IDX, + REG_EBRACE_IDX, + REG_BADBR_IDX, + REG_ERANGE_IDX, + REG_ESPACE_IDX, + REG_BADRPT_IDX, + REG_EEND_IDX, + REG_ESIZE_IDX, + REG_ERPAREN_IDX + }; + +/* Entry points for GNU code. */ + +/* re_compile_pattern is the GNU regular expression compiler: it + compiles PATTERN (of length LENGTH) and puts the result in BUFP. + Returns 0 if the pattern was valid, otherwise an error string. + + Assumes the 'allocated' (and perhaps 'buffer') and 'translate' fields + are set in BUFP on entry. */ + +const char * +re_compile_pattern (const char *pattern, size_t length, + struct re_pattern_buffer *bufp) +{ + reg_errcode_t ret; + + /* And GNU code determines whether or not to get register information + by passing null for the REGS argument to re_match, etc., not by + setting no_sub, unless RE_NO_SUB is set. */ + bufp->no_sub = !!(re_syntax_options & RE_NO_SUB); + + /* Match anchors at newline. */ + bufp->newline_anchor = 1; + + ret = re_compile_internal (bufp, pattern, length, re_syntax_options); + + if (!ret) + return NULL; + return gettext (__re_error_msgid + __re_error_msgid_idx[(int) ret]); +} +#ifdef _LIBC +weak_alias (__re_compile_pattern, re_compile_pattern) +#endif + +/* Set by 're_set_syntax' to the current regexp syntax to recognize. Can + also be assigned to arbitrarily: each pattern buffer stores its own + syntax, so it can be changed between regex compilations. */ +/* This has no initializer because initialized variables in Emacs + become read-only after dumping. */ +reg_syntax_t re_syntax_options; + + +/* Specify the precise syntax of regexps for compilation. This provides + for compatibility for various utilities which historically have + different, incompatible syntaxes. + + The argument SYNTAX is a bit mask comprised of the various bits + defined in regex.h. We return the old syntax. */ + +reg_syntax_t +re_set_syntax (reg_syntax_t syntax) +{ + reg_syntax_t ret = re_syntax_options; + + re_syntax_options = syntax; + return ret; +} +#ifdef _LIBC +weak_alias (__re_set_syntax, re_set_syntax) +#endif + +int +re_compile_fastmap (struct re_pattern_buffer *bufp) +{ + re_dfa_t *dfa = bufp->buffer; + char *fastmap = bufp->fastmap; + + memset (fastmap, '\0', sizeof (char) * SBC_MAX); + re_compile_fastmap_iter (bufp, dfa->init_state, fastmap); + if (dfa->init_state != dfa->init_state_word) + re_compile_fastmap_iter (bufp, dfa->init_state_word, fastmap); + if (dfa->init_state != dfa->init_state_nl) + re_compile_fastmap_iter (bufp, dfa->init_state_nl, fastmap); + if (dfa->init_state != dfa->init_state_begbuf) + re_compile_fastmap_iter (bufp, dfa->init_state_begbuf, fastmap); + bufp->fastmap_accurate = 1; + return 0; +} +#ifdef _LIBC +weak_alias (__re_compile_fastmap, re_compile_fastmap) +#endif + +static inline void +__attribute__ ((always_inline)) +re_set_fastmap (char *fastmap, bool icase, int ch) +{ + fastmap[ch] = 1; + if (icase) + fastmap[tolower (ch)] = 1; +} + +/* Helper function for re_compile_fastmap. + Compile fastmap for the initial_state INIT_STATE. */ + +static void +re_compile_fastmap_iter (regex_t *bufp, const re_dfastate_t *init_state, + char *fastmap) +{ + re_dfa_t *dfa = bufp->buffer; + Idx node_cnt; + bool icase = (dfa->mb_cur_max == 1 && (bufp->syntax & RE_ICASE)); + for (node_cnt = 0; node_cnt < init_state->nodes.nelem; ++node_cnt) + { + Idx node = init_state->nodes.elems[node_cnt]; + re_token_type_t type = dfa->nodes[node].type; + + if (type == CHARACTER) + { + re_set_fastmap (fastmap, icase, dfa->nodes[node].opr.c); +#ifdef RE_ENABLE_I18N + if ((bufp->syntax & RE_ICASE) && dfa->mb_cur_max > 1) + { + unsigned char buf[MB_LEN_MAX]; + unsigned char *p; + wchar_t wc; + mbstate_t state; + + p = buf; + *p++ = dfa->nodes[node].opr.c; + while (++node < dfa->nodes_len + && dfa->nodes[node].type == CHARACTER + && dfa->nodes[node].mb_partial) + *p++ = dfa->nodes[node].opr.c; + memset (&state, '\0', sizeof (state)); + if (__mbrtowc (&wc, (const char *) buf, p - buf, + &state) == p - buf + && (__wcrtomb ((char *) buf, __towlower (wc), &state) + != (size_t) -1)) + re_set_fastmap (fastmap, false, buf[0]); + } +#endif + } + else if (type == SIMPLE_BRACKET) + { + int i, ch; + for (i = 0, ch = 0; i < BITSET_WORDS; ++i) + { + int j; + bitset_word_t w = dfa->nodes[node].opr.sbcset[i]; + for (j = 0; j < BITSET_WORD_BITS; ++j, ++ch) + if (w & ((bitset_word_t) 1 << j)) + re_set_fastmap (fastmap, icase, ch); + } + } +#ifdef RE_ENABLE_I18N + else if (type == COMPLEX_BRACKET) + { + re_charset_t *cset = dfa->nodes[node].opr.mbcset; + Idx i; + +# ifdef _LIBC + /* See if we have to try all bytes which start multiple collation + elements. + e.g. In da_DK, we want to catch 'a' since "aa" is a valid + collation element, and don't catch 'b' since 'b' is + the only collation element which starts from 'b' (and + it is caught by SIMPLE_BRACKET). */ + if (_NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES) != 0 + && (cset->ncoll_syms || cset->nranges)) + { + const int32_t *table = (const int32_t *) + _NL_CURRENT (LC_COLLATE, _NL_COLLATE_TABLEMB); + for (i = 0; i < SBC_MAX; ++i) + if (table[i] < 0) + re_set_fastmap (fastmap, icase, i); + } +# endif /* _LIBC */ + + /* See if we have to start the match at all multibyte characters, + i.e. where we would not find an invalid sequence. This only + applies to multibyte character sets; for single byte character + sets, the SIMPLE_BRACKET again suffices. */ + if (dfa->mb_cur_max > 1 + && (cset->nchar_classes || cset->non_match || cset->nranges +# ifdef _LIBC + || cset->nequiv_classes +# endif /* _LIBC */ + )) + { + unsigned char c = 0; + do + { + mbstate_t mbs; + memset (&mbs, 0, sizeof (mbs)); + if (__mbrtowc (NULL, (char *) &c, 1, &mbs) == (size_t) -2) + re_set_fastmap (fastmap, false, (int) c); + } + while (++c != 0); + } + + else + { + /* ... Else catch all bytes which can start the mbchars. */ + for (i = 0; i < cset->nmbchars; ++i) + { + char buf[256]; + mbstate_t state; + memset (&state, '\0', sizeof (state)); + if (__wcrtomb (buf, cset->mbchars[i], &state) != (size_t) -1) + re_set_fastmap (fastmap, icase, *(unsigned char *) buf); + if ((bufp->syntax & RE_ICASE) && dfa->mb_cur_max > 1) + { + if (__wcrtomb (buf, __towlower (cset->mbchars[i]), &state) + != (size_t) -1) + re_set_fastmap (fastmap, false, *(unsigned char *) buf); + } + } + } + } +#endif /* RE_ENABLE_I18N */ + else if (type == OP_PERIOD +#ifdef RE_ENABLE_I18N + || type == OP_UTF8_PERIOD +#endif /* RE_ENABLE_I18N */ + || type == END_OF_RE) + { + memset (fastmap, '\1', sizeof (char) * SBC_MAX); + if (type == END_OF_RE) + bufp->can_be_null = 1; + return; + } + } +} + +/* Entry point for POSIX code. */ +/* regcomp takes a regular expression as a string and compiles it. + + PREG is a regex_t *. We do not expect any fields to be initialized, + since POSIX says we shouldn't. Thus, we set + + 'buffer' to the compiled pattern; + 'used' to the length of the compiled pattern; + 'syntax' to RE_SYNTAX_POSIX_EXTENDED if the + REG_EXTENDED bit in CFLAGS is set; otherwise, to + RE_SYNTAX_POSIX_BASIC; + 'newline_anchor' to REG_NEWLINE being set in CFLAGS; + 'fastmap' to an allocated space for the fastmap; + 'fastmap_accurate' to zero; + 're_nsub' to the number of subexpressions in PATTERN. + + PATTERN is the address of the pattern string. + + CFLAGS is a series of bits which affect compilation. + + If REG_EXTENDED is set, we use POSIX extended syntax; otherwise, we + use POSIX basic syntax. + + If REG_NEWLINE is set, then . and [^...] don't match newline. + Also, regexec will try a match beginning after every newline. + + If REG_ICASE is set, then we considers upper- and lowercase + versions of letters to be equivalent when matching. + + If REG_NOSUB is set, then when PREG is passed to regexec, that + routine will report only success or failure, and nothing about the + registers. + + It returns 0 if it succeeds, nonzero if it doesn't. (See regex.h for + the return codes and their meanings.) */ + +int +regcomp (regex_t *_Restrict_ preg, const char *_Restrict_ pattern, int cflags) +{ + reg_errcode_t ret; + reg_syntax_t syntax = ((cflags & REG_EXTENDED) ? RE_SYNTAX_POSIX_EXTENDED + : RE_SYNTAX_POSIX_BASIC); + + preg->buffer = NULL; + preg->allocated = 0; + preg->used = 0; + + /* Try to allocate space for the fastmap. */ + preg->fastmap = re_malloc (char, SBC_MAX); + if (__glibc_unlikely (preg->fastmap == NULL)) + return REG_ESPACE; + + syntax |= (cflags & REG_ICASE) ? RE_ICASE : 0; + + /* If REG_NEWLINE is set, newlines are treated differently. */ + if (cflags & REG_NEWLINE) + { /* REG_NEWLINE implies neither . nor [^...] match newline. */ + syntax &= ~RE_DOT_NEWLINE; + syntax |= RE_HAT_LISTS_NOT_NEWLINE; + /* It also changes the matching behavior. */ + preg->newline_anchor = 1; + } + else + preg->newline_anchor = 0; + preg->no_sub = !!(cflags & REG_NOSUB); + preg->translate = NULL; + + ret = re_compile_internal (preg, pattern, strlen (pattern), syntax); + + /* POSIX doesn't distinguish between an unmatched open-group and an + unmatched close-group: both are REG_EPAREN. */ + if (ret == REG_ERPAREN) + ret = REG_EPAREN; + + /* We have already checked preg->fastmap != NULL. */ + if (__glibc_likely (ret == REG_NOERROR)) + /* Compute the fastmap now, since regexec cannot modify the pattern + buffer. This function never fails in this implementation. */ + (void) re_compile_fastmap (preg); + else + { + /* Some error occurred while compiling the expression. */ + re_free (preg->fastmap); + preg->fastmap = NULL; + } + + return (int) ret; +} +#ifdef _LIBC +libc_hidden_def (__regcomp) +weak_alias (__regcomp, regcomp) +#endif + +/* Returns a message corresponding to an error code, ERRCODE, returned + from either regcomp or regexec. We don't use PREG here. */ + +size_t +regerror (int errcode, const regex_t *_Restrict_ preg, char *_Restrict_ errbuf, + size_t errbuf_size) +{ + const char *msg; + size_t msg_size; + int nerrcodes = sizeof __re_error_msgid_idx / sizeof __re_error_msgid_idx[0]; + + if (__glibc_unlikely (errcode < 0 || errcode >= nerrcodes)) + /* Only error codes returned by the rest of the code should be passed + to this routine. If we are given anything else, or if other regex + code generates an invalid error code, then the program has a bug. + Dump core so we can fix it. */ + abort (); + + msg = gettext (__re_error_msgid + __re_error_msgid_idx[errcode]); + + msg_size = strlen (msg) + 1; /* Includes the null. */ + + if (__glibc_likely (errbuf_size != 0)) + { + size_t cpy_size = msg_size; + if (__glibc_unlikely (msg_size > errbuf_size)) + { + cpy_size = errbuf_size - 1; + errbuf[cpy_size] = '\0'; + } + memcpy (errbuf, msg, cpy_size); + } + + return msg_size; +} +#ifdef _LIBC +weak_alias (__regerror, regerror) +#endif + + +#ifdef RE_ENABLE_I18N +/* This static array is used for the map to single-byte characters when + UTF-8 is used. Otherwise we would allocate memory just to initialize + it the same all the time. UTF-8 is the preferred encoding so this is + a worthwhile optimization. */ +static const bitset_t utf8_sb_map = +{ + /* Set the first 128 bits. */ +# if defined __GNUC__ && !defined __STRICT_ANSI__ + [0 ... 0x80 / BITSET_WORD_BITS - 1] = BITSET_WORD_MAX +# else +# if 4 * BITSET_WORD_BITS < ASCII_CHARS +# error "bitset_word_t is narrower than 32 bits" +# elif 3 * BITSET_WORD_BITS < ASCII_CHARS + BITSET_WORD_MAX, BITSET_WORD_MAX, BITSET_WORD_MAX, +# elif 2 * BITSET_WORD_BITS < ASCII_CHARS + BITSET_WORD_MAX, BITSET_WORD_MAX, +# elif 1 * BITSET_WORD_BITS < ASCII_CHARS + BITSET_WORD_MAX, +# endif + (BITSET_WORD_MAX + >> (SBC_MAX % BITSET_WORD_BITS == 0 + ? 0 + : BITSET_WORD_BITS - SBC_MAX % BITSET_WORD_BITS)) +# endif +}; +#endif + + +static void +free_dfa_content (re_dfa_t *dfa) +{ + Idx i, j; + + if (dfa->nodes) + for (i = 0; i < dfa->nodes_len; ++i) + free_token (dfa->nodes + i); + re_free (dfa->nexts); + for (i = 0; i < dfa->nodes_len; ++i) + { + if (dfa->eclosures != NULL) + re_node_set_free (dfa->eclosures + i); + if (dfa->inveclosures != NULL) + re_node_set_free (dfa->inveclosures + i); + if (dfa->edests != NULL) + re_node_set_free (dfa->edests + i); + } + re_free (dfa->edests); + re_free (dfa->eclosures); + re_free (dfa->inveclosures); + re_free (dfa->nodes); + + if (dfa->state_table) + for (i = 0; i <= dfa->state_hash_mask; ++i) + { + struct re_state_table_entry *entry = dfa->state_table + i; + for (j = 0; j < entry->num; ++j) + { + re_dfastate_t *state = entry->array[j]; + free_state (state); + } + re_free (entry->array); + } + re_free (dfa->state_table); +#ifdef RE_ENABLE_I18N + if (dfa->sb_char != utf8_sb_map) + re_free (dfa->sb_char); +#endif + re_free (dfa->subexp_map); +#ifdef DEBUG + re_free (dfa->re_str); +#endif + + re_free (dfa); +} + + +/* Free dynamically allocated space used by PREG. */ + +void +regfree (regex_t *preg) +{ + re_dfa_t *dfa = preg->buffer; + if (__glibc_likely (dfa != NULL)) + { + lock_fini (dfa->lock); + free_dfa_content (dfa); + } + preg->buffer = NULL; + preg->allocated = 0; + + re_free (preg->fastmap); + preg->fastmap = NULL; + + re_free (preg->translate); + preg->translate = NULL; +} +#ifdef _LIBC +libc_hidden_def (__regfree) +weak_alias (__regfree, regfree) +#endif + +/* Entry points compatible with 4.2 BSD regex library. We don't define + them unless specifically requested. */ + +#if defined _REGEX_RE_COMP || defined _LIBC + +/* BSD has one and only one pattern buffer. */ +static struct re_pattern_buffer re_comp_buf; + +char * +# ifdef _LIBC +/* Make these definitions weak in libc, so POSIX programs can redefine + these names if they don't use our functions, and still use + regcomp/regexec above without link errors. */ +weak_function +# endif +re_comp (const char *s) +{ + reg_errcode_t ret; + char *fastmap; + + if (!s) + { + if (!re_comp_buf.buffer) + return gettext ("No previous regular expression"); + return 0; + } + + if (re_comp_buf.buffer) + { + fastmap = re_comp_buf.fastmap; + re_comp_buf.fastmap = NULL; + __regfree (&re_comp_buf); + memset (&re_comp_buf, '\0', sizeof (re_comp_buf)); + re_comp_buf.fastmap = fastmap; + } + + if (re_comp_buf.fastmap == NULL) + { + re_comp_buf.fastmap = re_malloc (char, SBC_MAX); + if (re_comp_buf.fastmap == NULL) + return (char *) gettext (__re_error_msgid + + __re_error_msgid_idx[(int) REG_ESPACE]); + } + + /* Since 're_exec' always passes NULL for the 'regs' argument, we + don't need to initialize the pattern buffer fields which affect it. */ + + /* Match anchors at newlines. */ + re_comp_buf.newline_anchor = 1; + + ret = re_compile_internal (&re_comp_buf, s, strlen (s), re_syntax_options); + + if (!ret) + return NULL; + + /* Yes, we're discarding 'const' here if !HAVE_LIBINTL. */ + return (char *) gettext (__re_error_msgid + __re_error_msgid_idx[(int) ret]); +} + +#ifdef _LIBC +libc_freeres_fn (free_mem) +{ + __regfree (&re_comp_buf); +} +#endif + +#endif /* _REGEX_RE_COMP */ + +/* Internal entry point. + Compile the regular expression PATTERN, whose length is LENGTH. + SYNTAX indicate regular expression's syntax. */ + +static reg_errcode_t +re_compile_internal (regex_t *preg, const char * pattern, size_t length, + reg_syntax_t syntax) +{ + reg_errcode_t err = REG_NOERROR; + re_dfa_t *dfa; + re_string_t regexp; + + /* Initialize the pattern buffer. */ + preg->fastmap_accurate = 0; + preg->syntax = syntax; + preg->not_bol = preg->not_eol = 0; + preg->used = 0; + preg->re_nsub = 0; + preg->can_be_null = 0; + preg->regs_allocated = REGS_UNALLOCATED; + + /* Initialize the dfa. */ + dfa = preg->buffer; + if (__glibc_unlikely (preg->allocated < sizeof (re_dfa_t))) + { + /* If zero allocated, but buffer is non-null, try to realloc + enough space. This loses if buffer's address is bogus, but + that is the user's responsibility. If ->buffer is NULL this + is a simple allocation. */ + dfa = re_realloc (preg->buffer, re_dfa_t, 1); + if (dfa == NULL) + return REG_ESPACE; + preg->allocated = sizeof (re_dfa_t); + preg->buffer = dfa; + } + preg->used = sizeof (re_dfa_t); + + err = init_dfa (dfa, length); + if (__glibc_unlikely (err == REG_NOERROR && lock_init (dfa->lock) != 0)) + err = REG_ESPACE; + if (__glibc_unlikely (err != REG_NOERROR)) + { + free_dfa_content (dfa); + preg->buffer = NULL; + preg->allocated = 0; + return err; + } +#ifdef DEBUG + /* Note: length+1 will not overflow since it is checked in init_dfa. */ + dfa->re_str = re_malloc (char, length + 1); + strncpy (dfa->re_str, pattern, length + 1); +#endif + + err = re_string_construct (®exp, pattern, length, preg->translate, + (syntax & RE_ICASE) != 0, dfa); + if (__glibc_unlikely (err != REG_NOERROR)) + { + re_compile_internal_free_return: + free_workarea_compile (preg); + re_string_destruct (®exp); + lock_fini (dfa->lock); + free_dfa_content (dfa); + preg->buffer = NULL; + preg->allocated = 0; + return err; + } + + /* Parse the regular expression, and build a structure tree. */ + preg->re_nsub = 0; + dfa->str_tree = parse (®exp, preg, syntax, &err); + if (__glibc_unlikely (dfa->str_tree == NULL)) + goto re_compile_internal_free_return; + + /* Analyze the tree and create the nfa. */ + err = analyze (preg); + if (__glibc_unlikely (err != REG_NOERROR)) + goto re_compile_internal_free_return; + +#ifdef RE_ENABLE_I18N + /* If possible, do searching in single byte encoding to speed things up. */ + if (dfa->is_utf8 && !(syntax & RE_ICASE) && preg->translate == NULL) + optimize_utf8 (dfa); +#endif + + /* Then create the initial state of the dfa. */ + err = create_initial_state (dfa); + + /* Release work areas. */ + free_workarea_compile (preg); + re_string_destruct (®exp); + + if (__glibc_unlikely (err != REG_NOERROR)) + { + lock_fini (dfa->lock); + free_dfa_content (dfa); + preg->buffer = NULL; + preg->allocated = 0; + } + + return err; +} + +/* Initialize DFA. We use the length of the regular expression PAT_LEN + as the initial length of some arrays. */ + +static reg_errcode_t +init_dfa (re_dfa_t *dfa, size_t pat_len) +{ + __re_size_t table_size; +#ifndef _LIBC + const char *codeset_name; +#endif +#ifdef RE_ENABLE_I18N + size_t max_i18n_object_size = MAX (sizeof (wchar_t), sizeof (wctype_t)); +#else + size_t max_i18n_object_size = 0; +#endif + size_t max_object_size = + MAX (sizeof (struct re_state_table_entry), + MAX (sizeof (re_token_t), + MAX (sizeof (re_node_set), + MAX (sizeof (regmatch_t), + max_i18n_object_size)))); + + memset (dfa, '\0', sizeof (re_dfa_t)); + + /* Force allocation of str_tree_storage the first time. */ + dfa->str_tree_storage_idx = BIN_TREE_STORAGE_SIZE; + + /* Avoid overflows. The extra "/ 2" is for the table_size doubling + calculation below, and for similar doubling calculations + elsewhere. And it's <= rather than <, because some of the + doubling calculations add 1 afterwards. */ + if (__glibc_unlikely (MIN (IDX_MAX, SIZE_MAX / max_object_size) / 2 + <= pat_len)) + return REG_ESPACE; + + dfa->nodes_alloc = pat_len + 1; + dfa->nodes = re_malloc (re_token_t, dfa->nodes_alloc); + + /* table_size = 2 ^ ceil(log pat_len) */ + for (table_size = 1; ; table_size <<= 1) + if (table_size > pat_len) + break; + + dfa->state_table = calloc (sizeof (struct re_state_table_entry), table_size); + dfa->state_hash_mask = table_size - 1; + + dfa->mb_cur_max = MB_CUR_MAX; +#ifdef _LIBC + if (dfa->mb_cur_max == 6 + && strcmp (_NL_CURRENT (LC_CTYPE, _NL_CTYPE_CODESET_NAME), "UTF-8") == 0) + dfa->is_utf8 = 1; + dfa->map_notascii = (_NL_CURRENT_WORD (LC_CTYPE, _NL_CTYPE_MAP_TO_NONASCII) + != 0); +#else + codeset_name = nl_langinfo (CODESET); + if ((codeset_name[0] == 'U' || codeset_name[0] == 'u') + && (codeset_name[1] == 'T' || codeset_name[1] == 't') + && (codeset_name[2] == 'F' || codeset_name[2] == 'f') + && strcmp (codeset_name + 3 + (codeset_name[3] == '-'), "8") == 0) + dfa->is_utf8 = 1; + + /* We check exhaustively in the loop below if this charset is a + superset of ASCII. */ + dfa->map_notascii = 0; +#endif + +#ifdef RE_ENABLE_I18N + if (dfa->mb_cur_max > 1) + { + if (dfa->is_utf8) + dfa->sb_char = (re_bitset_ptr_t) utf8_sb_map; + else + { + int i, j, ch; + + dfa->sb_char = (re_bitset_ptr_t) calloc (sizeof (bitset_t), 1); + if (__glibc_unlikely (dfa->sb_char == NULL)) + return REG_ESPACE; + + /* Set the bits corresponding to single byte chars. */ + for (i = 0, ch = 0; i < BITSET_WORDS; ++i) + for (j = 0; j < BITSET_WORD_BITS; ++j, ++ch) + { + wint_t wch = __btowc (ch); + if (wch != WEOF) + dfa->sb_char[i] |= (bitset_word_t) 1 << j; +# ifndef _LIBC + if (isascii (ch) && wch != ch) + dfa->map_notascii = 1; +# endif + } + } + } +#endif + + if (__glibc_unlikely (dfa->nodes == NULL || dfa->state_table == NULL)) + return REG_ESPACE; + return REG_NOERROR; +} + +/* Initialize WORD_CHAR table, which indicate which character is + "word". In this case "word" means that it is the word construction + character used by some operators like "\<", "\>", etc. */ + +static void +init_word_char (re_dfa_t *dfa) +{ + int i = 0; + int j; + int ch = 0; + dfa->word_ops_used = 1; + if (__glibc_likely (dfa->map_notascii == 0)) + { + /* Avoid uint32_t and uint64_t as some non-GCC platforms lack + them, an issue when this code is used in Gnulib. */ + bitset_word_t bits0 = 0x00000000; + bitset_word_t bits1 = 0x03ff0000; + bitset_word_t bits2 = 0x87fffffe; + bitset_word_t bits3 = 0x07fffffe; + if (BITSET_WORD_BITS == 64) + { + /* Pacify gcc -Woverflow on 32-bit platformns. */ + dfa->word_char[0] = bits1 << 31 << 1 | bits0; + dfa->word_char[1] = bits3 << 31 << 1 | bits2; + i = 2; + } + else if (BITSET_WORD_BITS == 32) + { + dfa->word_char[0] = bits0; + dfa->word_char[1] = bits1; + dfa->word_char[2] = bits2; + dfa->word_char[3] = bits3; + i = 4; + } + else + goto general_case; + ch = 128; + + if (__glibc_likely (dfa->is_utf8)) + { + memset (&dfa->word_char[i], '\0', (SBC_MAX - ch) / 8); + return; + } + } + + general_case: + for (; i < BITSET_WORDS; ++i) + for (j = 0; j < BITSET_WORD_BITS; ++j, ++ch) + if (isalnum (ch) || ch == '_') + dfa->word_char[i] |= (bitset_word_t) 1 << j; +} + +/* Free the work area which are only used while compiling. */ + +static void +free_workarea_compile (regex_t *preg) +{ + re_dfa_t *dfa = preg->buffer; + bin_tree_storage_t *storage, *next; + for (storage = dfa->str_tree_storage; storage; storage = next) + { + next = storage->next; + re_free (storage); + } + dfa->str_tree_storage = NULL; + dfa->str_tree_storage_idx = BIN_TREE_STORAGE_SIZE; + dfa->str_tree = NULL; + re_free (dfa->org_indices); + dfa->org_indices = NULL; +} + +/* Create initial states for all contexts. */ + +static reg_errcode_t +create_initial_state (re_dfa_t *dfa) +{ + Idx first, i; + reg_errcode_t err; + re_node_set init_nodes; + + /* Initial states have the epsilon closure of the node which is + the first node of the regular expression. */ + first = dfa->str_tree->first->node_idx; + dfa->init_node = first; + err = re_node_set_init_copy (&init_nodes, dfa->eclosures + first); + if (__glibc_unlikely (err != REG_NOERROR)) + return err; + + /* The back-references which are in initial states can epsilon transit, + since in this case all of the subexpressions can be null. + Then we add epsilon closures of the nodes which are the next nodes of + the back-references. */ + if (dfa->nbackref > 0) + for (i = 0; i < init_nodes.nelem; ++i) + { + Idx node_idx = init_nodes.elems[i]; + re_token_type_t type = dfa->nodes[node_idx].type; + + Idx clexp_idx; + if (type != OP_BACK_REF) + continue; + for (clexp_idx = 0; clexp_idx < init_nodes.nelem; ++clexp_idx) + { + re_token_t *clexp_node; + clexp_node = dfa->nodes + init_nodes.elems[clexp_idx]; + if (clexp_node->type == OP_CLOSE_SUBEXP + && clexp_node->opr.idx == dfa->nodes[node_idx].opr.idx) + break; + } + if (clexp_idx == init_nodes.nelem) + continue; + + if (type == OP_BACK_REF) + { + Idx dest_idx = dfa->edests[node_idx].elems[0]; + if (!re_node_set_contains (&init_nodes, dest_idx)) + { + reg_errcode_t merge_err + = re_node_set_merge (&init_nodes, dfa->eclosures + dest_idx); + if (merge_err != REG_NOERROR) + return merge_err; + i = 0; + } + } + } + + /* It must be the first time to invoke acquire_state. */ + dfa->init_state = re_acquire_state_context (&err, dfa, &init_nodes, 0); + /* We don't check ERR here, since the initial state must not be NULL. */ + if (__glibc_unlikely (dfa->init_state == NULL)) + return err; + if (dfa->init_state->has_constraint) + { + dfa->init_state_word = re_acquire_state_context (&err, dfa, &init_nodes, + CONTEXT_WORD); + dfa->init_state_nl = re_acquire_state_context (&err, dfa, &init_nodes, + CONTEXT_NEWLINE); + dfa->init_state_begbuf = re_acquire_state_context (&err, dfa, + &init_nodes, + CONTEXT_NEWLINE + | CONTEXT_BEGBUF); + if (__glibc_unlikely (dfa->init_state_word == NULL + || dfa->init_state_nl == NULL + || dfa->init_state_begbuf == NULL)) + return err; + } + else + dfa->init_state_word = dfa->init_state_nl + = dfa->init_state_begbuf = dfa->init_state; + + re_node_set_free (&init_nodes); + return REG_NOERROR; +} + +#ifdef RE_ENABLE_I18N +/* If it is possible to do searching in single byte encoding instead of UTF-8 + to speed things up, set dfa->mb_cur_max to 1, clear is_utf8 and change + DFA nodes where needed. */ + +static void +optimize_utf8 (re_dfa_t *dfa) +{ + Idx node; + int i; + bool mb_chars = false; + bool has_period = false; + + for (node = 0; node < dfa->nodes_len; ++node) + switch (dfa->nodes[node].type) + { + case CHARACTER: + if (dfa->nodes[node].opr.c >= ASCII_CHARS) + mb_chars = true; + break; + case ANCHOR: + switch (dfa->nodes[node].opr.ctx_type) + { + case LINE_FIRST: + case LINE_LAST: + case BUF_FIRST: + case BUF_LAST: + break; + default: + /* Word anchors etc. cannot be handled. It's okay to test + opr.ctx_type since constraints (for all DFA nodes) are + created by ORing one or more opr.ctx_type values. */ + return; + } + break; + case OP_PERIOD: + has_period = true; + break; + case OP_BACK_REF: + case OP_ALT: + case END_OF_RE: + case OP_DUP_ASTERISK: + case OP_OPEN_SUBEXP: + case OP_CLOSE_SUBEXP: + break; + case COMPLEX_BRACKET: + return; + case SIMPLE_BRACKET: + /* Just double check. */ + { + int rshift = (ASCII_CHARS % BITSET_WORD_BITS == 0 + ? 0 + : BITSET_WORD_BITS - ASCII_CHARS % BITSET_WORD_BITS); + for (i = ASCII_CHARS / BITSET_WORD_BITS; i < BITSET_WORDS; ++i) + { + if (dfa->nodes[node].opr.sbcset[i] >> rshift != 0) + return; + rshift = 0; + } + } + break; + default: + abort (); + } + + if (mb_chars || has_period) + for (node = 0; node < dfa->nodes_len; ++node) + { + if (dfa->nodes[node].type == CHARACTER + && dfa->nodes[node].opr.c >= ASCII_CHARS) + dfa->nodes[node].mb_partial = 0; + else if (dfa->nodes[node].type == OP_PERIOD) + dfa->nodes[node].type = OP_UTF8_PERIOD; + } + + /* The search can be in single byte locale. */ + dfa->mb_cur_max = 1; + dfa->is_utf8 = 0; + dfa->has_mb_node = dfa->nbackref > 0 || has_period; +} +#endif + +/* Analyze the structure tree, and calculate "first", "next", "edest", + "eclosure", and "inveclosure". */ + +static reg_errcode_t +analyze (regex_t *preg) +{ + re_dfa_t *dfa = preg->buffer; + reg_errcode_t ret; + + /* Allocate arrays. */ + dfa->nexts = re_malloc (Idx, dfa->nodes_alloc); + dfa->org_indices = re_malloc (Idx, dfa->nodes_alloc); + dfa->edests = re_malloc (re_node_set, dfa->nodes_alloc); + dfa->eclosures = re_malloc (re_node_set, dfa->nodes_alloc); + if (__glibc_unlikely (dfa->nexts == NULL || dfa->org_indices == NULL + || dfa->edests == NULL || dfa->eclosures == NULL)) + return REG_ESPACE; + + dfa->subexp_map = re_malloc (Idx, preg->re_nsub); + if (dfa->subexp_map != NULL) + { + Idx i; + for (i = 0; i < preg->re_nsub; i++) + dfa->subexp_map[i] = i; + preorder (dfa->str_tree, optimize_subexps, dfa); + for (i = 0; i < preg->re_nsub; i++) + if (dfa->subexp_map[i] != i) + break; + if (i == preg->re_nsub) + { + re_free (dfa->subexp_map); + dfa->subexp_map = NULL; + } + } + + ret = postorder (dfa->str_tree, lower_subexps, preg); + if (__glibc_unlikely (ret != REG_NOERROR)) + return ret; + ret = postorder (dfa->str_tree, calc_first, dfa); + if (__glibc_unlikely (ret != REG_NOERROR)) + return ret; + preorder (dfa->str_tree, calc_next, dfa); + ret = preorder (dfa->str_tree, link_nfa_nodes, dfa); + if (__glibc_unlikely (ret != REG_NOERROR)) + return ret; + ret = calc_eclosure (dfa); + if (__glibc_unlikely (ret != REG_NOERROR)) + return ret; + + /* We only need this during the prune_impossible_nodes pass in regexec.c; + skip it if p_i_n will not run, as calc_inveclosure can be quadratic. */ + if ((!preg->no_sub && preg->re_nsub > 0 && dfa->has_plural_match) + || dfa->nbackref) + { + dfa->inveclosures = re_malloc (re_node_set, dfa->nodes_len); + if (__glibc_unlikely (dfa->inveclosures == NULL)) + return REG_ESPACE; + ret = calc_inveclosure (dfa); + } + + return ret; +} + +/* Our parse trees are very unbalanced, so we cannot use a stack to + implement parse tree visits. Instead, we use parent pointers and + some hairy code in these two functions. */ +static reg_errcode_t +postorder (bin_tree_t *root, reg_errcode_t (fn (void *, bin_tree_t *)), + void *extra) +{ + bin_tree_t *node, *prev; + + for (node = root; ; ) + { + /* Descend down the tree, preferably to the left (or to the right + if that's the only child). */ + while (node->left || node->right) + if (node->left) + node = node->left; + else + node = node->right; + + do + { + reg_errcode_t err = fn (extra, node); + if (__glibc_unlikely (err != REG_NOERROR)) + return err; + if (node->parent == NULL) + return REG_NOERROR; + prev = node; + node = node->parent; + } + /* Go up while we have a node that is reached from the right. */ + while (node->right == prev || node->right == NULL); + node = node->right; + } +} + +static reg_errcode_t +preorder (bin_tree_t *root, reg_errcode_t (fn (void *, bin_tree_t *)), + void *extra) +{ + bin_tree_t *node; + + for (node = root; ; ) + { + reg_errcode_t err = fn (extra, node); + if (__glibc_unlikely (err != REG_NOERROR)) + return err; + + /* Go to the left node, or up and to the right. */ + if (node->left) + node = node->left; + else + { + bin_tree_t *prev = NULL; + while (node->right == prev || node->right == NULL) + { + prev = node; + node = node->parent; + if (!node) + return REG_NOERROR; + } + node = node->right; + } + } +} + +/* Optimization pass: if a SUBEXP is entirely contained, strip it and tell + re_search_internal to map the inner one's opr.idx to this one's. Adjust + backreferences as well. Requires a preorder visit. */ +static reg_errcode_t +optimize_subexps (void *extra, bin_tree_t *node) +{ + re_dfa_t *dfa = (re_dfa_t *) extra; + + if (node->token.type == OP_BACK_REF && dfa->subexp_map) + { + int idx = node->token.opr.idx; + node->token.opr.idx = dfa->subexp_map[idx]; + dfa->used_bkref_map |= 1 << node->token.opr.idx; + } + + else if (node->token.type == SUBEXP + && node->left && node->left->token.type == SUBEXP) + { + Idx other_idx = node->left->token.opr.idx; + + node->left = node->left->left; + if (node->left) + node->left->parent = node; + + dfa->subexp_map[other_idx] = dfa->subexp_map[node->token.opr.idx]; + if (other_idx < BITSET_WORD_BITS) + dfa->used_bkref_map &= ~((bitset_word_t) 1 << other_idx); + } + + return REG_NOERROR; +} + +/* Lowering pass: Turn each SUBEXP node into the appropriate concatenation + of OP_OPEN_SUBEXP, the body of the SUBEXP (if any) and OP_CLOSE_SUBEXP. */ +static reg_errcode_t +lower_subexps (void *extra, bin_tree_t *node) +{ + regex_t *preg = (regex_t *) extra; + reg_errcode_t err = REG_NOERROR; + + if (node->left && node->left->token.type == SUBEXP) + { + node->left = lower_subexp (&err, preg, node->left); + if (node->left) + node->left->parent = node; + } + if (node->right && node->right->token.type == SUBEXP) + { + node->right = lower_subexp (&err, preg, node->right); + if (node->right) + node->right->parent = node; + } + + return err; +} + +static bin_tree_t * +lower_subexp (reg_errcode_t *err, regex_t *preg, bin_tree_t *node) +{ + re_dfa_t *dfa = preg->buffer; + bin_tree_t *body = node->left; + bin_tree_t *op, *cls, *tree1, *tree; + + if (preg->no_sub + /* We do not optimize empty subexpressions, because otherwise we may + have bad CONCAT nodes with NULL children. This is obviously not + very common, so we do not lose much. An example that triggers + this case is the sed "script" /\(\)/x. */ + && node->left != NULL + && (node->token.opr.idx >= BITSET_WORD_BITS + || !(dfa->used_bkref_map + & ((bitset_word_t) 1 << node->token.opr.idx)))) + return node->left; + + /* Convert the SUBEXP node to the concatenation of an + OP_OPEN_SUBEXP, the contents, and an OP_CLOSE_SUBEXP. */ + op = create_tree (dfa, NULL, NULL, OP_OPEN_SUBEXP); + cls = create_tree (dfa, NULL, NULL, OP_CLOSE_SUBEXP); + tree1 = body ? create_tree (dfa, body, cls, CONCAT) : cls; + tree = create_tree (dfa, op, tree1, CONCAT); + if (__glibc_unlikely (tree == NULL || tree1 == NULL + || op == NULL || cls == NULL)) + { + *err = REG_ESPACE; + return NULL; + } + + op->token.opr.idx = cls->token.opr.idx = node->token.opr.idx; + op->token.opt_subexp = cls->token.opt_subexp = node->token.opt_subexp; + return tree; +} + +/* Pass 1 in building the NFA: compute FIRST and create unlinked automaton + nodes. Requires a postorder visit. */ +static reg_errcode_t +calc_first (void *extra, bin_tree_t *node) +{ + re_dfa_t *dfa = (re_dfa_t *) extra; + if (node->token.type == CONCAT) + { + node->first = node->left->first; + node->node_idx = node->left->node_idx; + } + else + { + node->first = node; + node->node_idx = re_dfa_add_node (dfa, node->token); + if (__glibc_unlikely (node->node_idx == -1)) + return REG_ESPACE; + if (node->token.type == ANCHOR) + dfa->nodes[node->node_idx].constraint = node->token.opr.ctx_type; + } + return REG_NOERROR; +} + +/* Pass 2: compute NEXT on the tree. Preorder visit. */ +static reg_errcode_t +calc_next (void *extra, bin_tree_t *node) +{ + switch (node->token.type) + { + case OP_DUP_ASTERISK: + node->left->next = node; + break; + case CONCAT: + node->left->next = node->right->first; + node->right->next = node->next; + break; + default: + if (node->left) + node->left->next = node->next; + if (node->right) + node->right->next = node->next; + break; + } + return REG_NOERROR; +} + +/* Pass 3: link all DFA nodes to their NEXT node (any order will do). */ +static reg_errcode_t +link_nfa_nodes (void *extra, bin_tree_t *node) +{ + re_dfa_t *dfa = (re_dfa_t *) extra; + Idx idx = node->node_idx; + reg_errcode_t err = REG_NOERROR; + + switch (node->token.type) + { + case CONCAT: + break; + + case END_OF_RE: + assert (node->next == NULL); + break; + + case OP_DUP_ASTERISK: + case OP_ALT: + { + Idx left, right; + dfa->has_plural_match = 1; + if (node->left != NULL) + left = node->left->first->node_idx; + else + left = node->next->node_idx; + if (node->right != NULL) + right = node->right->first->node_idx; + else + right = node->next->node_idx; + assert (left > -1); + assert (right > -1); + err = re_node_set_init_2 (dfa->edests + idx, left, right); + } + break; + + case ANCHOR: + case OP_OPEN_SUBEXP: + case OP_CLOSE_SUBEXP: + err = re_node_set_init_1 (dfa->edests + idx, node->next->node_idx); + break; + + case OP_BACK_REF: + dfa->nexts[idx] = node->next->node_idx; + if (node->token.type == OP_BACK_REF) + err = re_node_set_init_1 (dfa->edests + idx, dfa->nexts[idx]); + break; + + default: + assert (!IS_EPSILON_NODE (node->token.type)); + dfa->nexts[idx] = node->next->node_idx; + break; + } + + return err; +} + +/* Duplicate the epsilon closure of the node ROOT_NODE. + Note that duplicated nodes have constraint INIT_CONSTRAINT in addition + to their own constraint. */ + +static reg_errcode_t +duplicate_node_closure (re_dfa_t *dfa, Idx top_org_node, Idx top_clone_node, + Idx root_node, unsigned int init_constraint) +{ + Idx org_node, clone_node; + bool ok; + unsigned int constraint = init_constraint; + for (org_node = top_org_node, clone_node = top_clone_node;;) + { + Idx org_dest, clone_dest; + if (dfa->nodes[org_node].type == OP_BACK_REF) + { + /* If the back reference epsilon-transit, its destination must + also have the constraint. Then duplicate the epsilon closure + of the destination of the back reference, and store it in + edests of the back reference. */ + org_dest = dfa->nexts[org_node]; + re_node_set_empty (dfa->edests + clone_node); + clone_dest = duplicate_node (dfa, org_dest, constraint); + if (__glibc_unlikely (clone_dest == -1)) + return REG_ESPACE; + dfa->nexts[clone_node] = dfa->nexts[org_node]; + ok = re_node_set_insert (dfa->edests + clone_node, clone_dest); + if (__glibc_unlikely (! ok)) + return REG_ESPACE; + } + else if (dfa->edests[org_node].nelem == 0) + { + /* In case of the node can't epsilon-transit, don't duplicate the + destination and store the original destination as the + destination of the node. */ + dfa->nexts[clone_node] = dfa->nexts[org_node]; + break; + } + else if (dfa->edests[org_node].nelem == 1) + { + /* In case of the node can epsilon-transit, and it has only one + destination. */ + org_dest = dfa->edests[org_node].elems[0]; + re_node_set_empty (dfa->edests + clone_node); + /* If the node is root_node itself, it means the epsilon closure + has a loop. Then tie it to the destination of the root_node. */ + if (org_node == root_node && clone_node != org_node) + { + ok = re_node_set_insert (dfa->edests + clone_node, org_dest); + if (__glibc_unlikely (! ok)) + return REG_ESPACE; + break; + } + /* In case the node has another constraint, append it. */ + constraint |= dfa->nodes[org_node].constraint; + clone_dest = duplicate_node (dfa, org_dest, constraint); + if (__glibc_unlikely (clone_dest == -1)) + return REG_ESPACE; + ok = re_node_set_insert (dfa->edests + clone_node, clone_dest); + if (__glibc_unlikely (! ok)) + return REG_ESPACE; + } + else /* dfa->edests[org_node].nelem == 2 */ + { + /* In case of the node can epsilon-transit, and it has two + destinations. In the bin_tree_t and DFA, that's '|' and '*'. */ + org_dest = dfa->edests[org_node].elems[0]; + re_node_set_empty (dfa->edests + clone_node); + /* Search for a duplicated node which satisfies the constraint. */ + clone_dest = search_duplicated_node (dfa, org_dest, constraint); + if (clone_dest == -1) + { + /* There is no such duplicated node, create a new one. */ + reg_errcode_t err; + clone_dest = duplicate_node (dfa, org_dest, constraint); + if (__glibc_unlikely (clone_dest == -1)) + return REG_ESPACE; + ok = re_node_set_insert (dfa->edests + clone_node, clone_dest); + if (__glibc_unlikely (! ok)) + return REG_ESPACE; + err = duplicate_node_closure (dfa, org_dest, clone_dest, + root_node, constraint); + if (__glibc_unlikely (err != REG_NOERROR)) + return err; + } + else + { + /* There is a duplicated node which satisfies the constraint, + use it to avoid infinite loop. */ + ok = re_node_set_insert (dfa->edests + clone_node, clone_dest); + if (__glibc_unlikely (! ok)) + return REG_ESPACE; + } + + org_dest = dfa->edests[org_node].elems[1]; + clone_dest = duplicate_node (dfa, org_dest, constraint); + if (__glibc_unlikely (clone_dest == -1)) + return REG_ESPACE; + ok = re_node_set_insert (dfa->edests + clone_node, clone_dest); + if (__glibc_unlikely (! ok)) + return REG_ESPACE; + } + org_node = org_dest; + clone_node = clone_dest; + } + return REG_NOERROR; +} + +/* Search for a node which is duplicated from the node ORG_NODE, and + satisfies the constraint CONSTRAINT. */ + +static Idx +search_duplicated_node (const re_dfa_t *dfa, Idx org_node, + unsigned int constraint) +{ + Idx idx; + for (idx = dfa->nodes_len - 1; dfa->nodes[idx].duplicated && idx > 0; --idx) + { + if (org_node == dfa->org_indices[idx] + && constraint == dfa->nodes[idx].constraint) + return idx; /* Found. */ + } + return -1; /* Not found. */ +} + +/* Duplicate the node whose index is ORG_IDX and set the constraint CONSTRAINT. + Return the index of the new node, or -1 if insufficient storage is + available. */ + +static Idx +duplicate_node (re_dfa_t *dfa, Idx org_idx, unsigned int constraint) +{ + Idx dup_idx = re_dfa_add_node (dfa, dfa->nodes[org_idx]); + if (__glibc_likely (dup_idx != -1)) + { + dfa->nodes[dup_idx].constraint = constraint; + dfa->nodes[dup_idx].constraint |= dfa->nodes[org_idx].constraint; + dfa->nodes[dup_idx].duplicated = 1; + + /* Store the index of the original node. */ + dfa->org_indices[dup_idx] = org_idx; + } + return dup_idx; +} + +static reg_errcode_t +calc_inveclosure (re_dfa_t *dfa) +{ + Idx src, idx; + bool ok; + for (idx = 0; idx < dfa->nodes_len; ++idx) + re_node_set_init_empty (dfa->inveclosures + idx); + + for (src = 0; src < dfa->nodes_len; ++src) + { + Idx *elems = dfa->eclosures[src].elems; + for (idx = 0; idx < dfa->eclosures[src].nelem; ++idx) + { + ok = re_node_set_insert_last (dfa->inveclosures + elems[idx], src); + if (__glibc_unlikely (! ok)) + return REG_ESPACE; + } + } + + return REG_NOERROR; +} + +/* Calculate "eclosure" for all the node in DFA. */ + +static reg_errcode_t +calc_eclosure (re_dfa_t *dfa) +{ + Idx node_idx; + bool incomplete; +#ifdef DEBUG + assert (dfa->nodes_len > 0); +#endif + incomplete = false; + /* For each nodes, calculate epsilon closure. */ + for (node_idx = 0; ; ++node_idx) + { + reg_errcode_t err; + re_node_set eclosure_elem; + if (node_idx == dfa->nodes_len) + { + if (!incomplete) + break; + incomplete = false; + node_idx = 0; + } + +#ifdef DEBUG + assert (dfa->eclosures[node_idx].nelem != -1); +#endif + + /* If we have already calculated, skip it. */ + if (dfa->eclosures[node_idx].nelem != 0) + continue; + /* Calculate epsilon closure of 'node_idx'. */ + err = calc_eclosure_iter (&eclosure_elem, dfa, node_idx, true); + if (__glibc_unlikely (err != REG_NOERROR)) + return err; + + if (dfa->eclosures[node_idx].nelem == 0) + { + incomplete = true; + re_node_set_free (&eclosure_elem); + } + } + return REG_NOERROR; +} + +/* Calculate epsilon closure of NODE. */ + +static reg_errcode_t +calc_eclosure_iter (re_node_set *new_set, re_dfa_t *dfa, Idx node, bool root) +{ + reg_errcode_t err; + Idx i; + re_node_set eclosure; + bool ok; + bool incomplete = false; + err = re_node_set_alloc (&eclosure, dfa->edests[node].nelem + 1); + if (__glibc_unlikely (err != REG_NOERROR)) + return err; + + /* This indicates that we are calculating this node now. + We reference this value to avoid infinite loop. */ + dfa->eclosures[node].nelem = -1; + + /* If the current node has constraints, duplicate all nodes + since they must inherit the constraints. */ + if (dfa->nodes[node].constraint + && dfa->edests[node].nelem + && !dfa->nodes[dfa->edests[node].elems[0]].duplicated) + { + err = duplicate_node_closure (dfa, node, node, node, + dfa->nodes[node].constraint); + if (__glibc_unlikely (err != REG_NOERROR)) + return err; + } + + /* Expand each epsilon destination nodes. */ + if (IS_EPSILON_NODE(dfa->nodes[node].type)) + for (i = 0; i < dfa->edests[node].nelem; ++i) + { + re_node_set eclosure_elem; + Idx edest = dfa->edests[node].elems[i]; + /* If calculating the epsilon closure of 'edest' is in progress, + return intermediate result. */ + if (dfa->eclosures[edest].nelem == -1) + { + incomplete = true; + continue; + } + /* If we haven't calculated the epsilon closure of 'edest' yet, + calculate now. Otherwise use calculated epsilon closure. */ + if (dfa->eclosures[edest].nelem == 0) + { + err = calc_eclosure_iter (&eclosure_elem, dfa, edest, false); + if (__glibc_unlikely (err != REG_NOERROR)) + return err; + } + else + eclosure_elem = dfa->eclosures[edest]; + /* Merge the epsilon closure of 'edest'. */ + err = re_node_set_merge (&eclosure, &eclosure_elem); + if (__glibc_unlikely (err != REG_NOERROR)) + return err; + /* If the epsilon closure of 'edest' is incomplete, + the epsilon closure of this node is also incomplete. */ + if (dfa->eclosures[edest].nelem == 0) + { + incomplete = true; + re_node_set_free (&eclosure_elem); + } + } + + /* An epsilon closure includes itself. */ + ok = re_node_set_insert (&eclosure, node); + if (__glibc_unlikely (! ok)) + return REG_ESPACE; + if (incomplete && !root) + dfa->eclosures[node].nelem = 0; + else + dfa->eclosures[node] = eclosure; + *new_set = eclosure; + return REG_NOERROR; +} + +/* Functions for token which are used in the parser. */ + +/* Fetch a token from INPUT. + We must not use this function inside bracket expressions. */ + +static void +fetch_token (re_token_t *result, re_string_t *input, reg_syntax_t syntax) +{ + re_string_skip_bytes (input, peek_token (result, input, syntax)); +} + +/* Peek a token from INPUT, and return the length of the token. + We must not use this function inside bracket expressions. */ + +static int +peek_token (re_token_t *token, re_string_t *input, reg_syntax_t syntax) +{ + unsigned char c; + + if (re_string_eoi (input)) + { + token->type = END_OF_RE; + return 0; + } + + c = re_string_peek_byte (input, 0); + token->opr.c = c; + + token->word_char = 0; +#ifdef RE_ENABLE_I18N + token->mb_partial = 0; + if (input->mb_cur_max > 1 && + !re_string_first_byte (input, re_string_cur_idx (input))) + { + token->type = CHARACTER; + token->mb_partial = 1; + return 1; + } +#endif + if (c == '\\') + { + unsigned char c2; + if (re_string_cur_idx (input) + 1 >= re_string_length (input)) + { + token->type = BACK_SLASH; + return 1; + } + + c2 = re_string_peek_byte_case (input, 1); + token->opr.c = c2; + token->type = CHARACTER; +#ifdef RE_ENABLE_I18N + if (input->mb_cur_max > 1) + { + wint_t wc = re_string_wchar_at (input, + re_string_cur_idx (input) + 1); + token->word_char = IS_WIDE_WORD_CHAR (wc) != 0; + } + else +#endif + token->word_char = IS_WORD_CHAR (c2) != 0; + + switch (c2) + { + case '|': + if (!(syntax & RE_LIMITED_OPS) && !(syntax & RE_NO_BK_VBAR)) + token->type = OP_ALT; + break; + case '1': case '2': case '3': case '4': case '5': + case '6': case '7': case '8': case '9': + if (!(syntax & RE_NO_BK_REFS)) + { + token->type = OP_BACK_REF; + token->opr.idx = c2 - '1'; + } + break; + case '<': + if (!(syntax & RE_NO_GNU_OPS)) + { + token->type = ANCHOR; + token->opr.ctx_type = WORD_FIRST; + } + break; + case '>': + if (!(syntax & RE_NO_GNU_OPS)) + { + token->type = ANCHOR; + token->opr.ctx_type = WORD_LAST; + } + break; + case 'b': + if (!(syntax & RE_NO_GNU_OPS)) + { + token->type = ANCHOR; + token->opr.ctx_type = WORD_DELIM; + } + break; + case 'B': + if (!(syntax & RE_NO_GNU_OPS)) + { + token->type = ANCHOR; + token->opr.ctx_type = NOT_WORD_DELIM; + } + break; + case 'w': + if (!(syntax & RE_NO_GNU_OPS)) + token->type = OP_WORD; + break; + case 'W': + if (!(syntax & RE_NO_GNU_OPS)) + token->type = OP_NOTWORD; + break; + case 's': + if (!(syntax & RE_NO_GNU_OPS)) + token->type = OP_SPACE; + break; + case 'S': + if (!(syntax & RE_NO_GNU_OPS)) + token->type = OP_NOTSPACE; + break; + case '`': + if (!(syntax & RE_NO_GNU_OPS)) + { + token->type = ANCHOR; + token->opr.ctx_type = BUF_FIRST; + } + break; + case '\'': + if (!(syntax & RE_NO_GNU_OPS)) + { + token->type = ANCHOR; + token->opr.ctx_type = BUF_LAST; + } + break; + case '(': + if (!(syntax & RE_NO_BK_PARENS)) + token->type = OP_OPEN_SUBEXP; + break; + case ')': + if (!(syntax & RE_NO_BK_PARENS)) + token->type = OP_CLOSE_SUBEXP; + break; + case '+': + if (!(syntax & RE_LIMITED_OPS) && (syntax & RE_BK_PLUS_QM)) + token->type = OP_DUP_PLUS; + break; + case '?': + if (!(syntax & RE_LIMITED_OPS) && (syntax & RE_BK_PLUS_QM)) + token->type = OP_DUP_QUESTION; + break; + case '{': + if ((syntax & RE_INTERVALS) && (!(syntax & RE_NO_BK_BRACES))) + token->type = OP_OPEN_DUP_NUM; + break; + case '}': + if ((syntax & RE_INTERVALS) && (!(syntax & RE_NO_BK_BRACES))) + token->type = OP_CLOSE_DUP_NUM; + break; + default: + break; + } + return 2; + } + + token->type = CHARACTER; +#ifdef RE_ENABLE_I18N + if (input->mb_cur_max > 1) + { + wint_t wc = re_string_wchar_at (input, re_string_cur_idx (input)); + token->word_char = IS_WIDE_WORD_CHAR (wc) != 0; + } + else +#endif + token->word_char = IS_WORD_CHAR (token->opr.c); + + switch (c) + { + case '\n': + if (syntax & RE_NEWLINE_ALT) + token->type = OP_ALT; + break; + case '|': + if (!(syntax & RE_LIMITED_OPS) && (syntax & RE_NO_BK_VBAR)) + token->type = OP_ALT; + break; + case '*': + token->type = OP_DUP_ASTERISK; + break; + case '+': + if (!(syntax & RE_LIMITED_OPS) && !(syntax & RE_BK_PLUS_QM)) + token->type = OP_DUP_PLUS; + break; + case '?': + if (!(syntax & RE_LIMITED_OPS) && !(syntax & RE_BK_PLUS_QM)) + token->type = OP_DUP_QUESTION; + break; + case '{': + if ((syntax & RE_INTERVALS) && (syntax & RE_NO_BK_BRACES)) + token->type = OP_OPEN_DUP_NUM; + break; + case '}': + if ((syntax & RE_INTERVALS) && (syntax & RE_NO_BK_BRACES)) + token->type = OP_CLOSE_DUP_NUM; + break; + case '(': + if (syntax & RE_NO_BK_PARENS) + token->type = OP_OPEN_SUBEXP; + break; + case ')': + if (syntax & RE_NO_BK_PARENS) + token->type = OP_CLOSE_SUBEXP; + break; + case '[': + token->type = OP_OPEN_BRACKET; + break; + case '.': + token->type = OP_PERIOD; + break; + case '^': + if (!(syntax & (RE_CONTEXT_INDEP_ANCHORS | RE_CARET_ANCHORS_HERE)) && + re_string_cur_idx (input) != 0) + { + char prev = re_string_peek_byte (input, -1); + if (!(syntax & RE_NEWLINE_ALT) || prev != '\n') + break; + } + token->type = ANCHOR; + token->opr.ctx_type = LINE_FIRST; + break; + case '$': + if (!(syntax & RE_CONTEXT_INDEP_ANCHORS) && + re_string_cur_idx (input) + 1 != re_string_length (input)) + { + re_token_t next; + re_string_skip_bytes (input, 1); + peek_token (&next, input, syntax); + re_string_skip_bytes (input, -1); + if (next.type != OP_ALT && next.type != OP_CLOSE_SUBEXP) + break; + } + token->type = ANCHOR; + token->opr.ctx_type = LINE_LAST; + break; + default: + break; + } + return 1; +} + +/* Peek a token from INPUT, and return the length of the token. + We must not use this function out of bracket expressions. */ + +static int +peek_token_bracket (re_token_t *token, re_string_t *input, reg_syntax_t syntax) +{ + unsigned char c; + if (re_string_eoi (input)) + { + token->type = END_OF_RE; + return 0; + } + c = re_string_peek_byte (input, 0); + token->opr.c = c; + +#ifdef RE_ENABLE_I18N + if (input->mb_cur_max > 1 && + !re_string_first_byte (input, re_string_cur_idx (input))) + { + token->type = CHARACTER; + return 1; + } +#endif /* RE_ENABLE_I18N */ + + if (c == '\\' && (syntax & RE_BACKSLASH_ESCAPE_IN_LISTS) + && re_string_cur_idx (input) + 1 < re_string_length (input)) + { + /* In this case, '\' escape a character. */ + unsigned char c2; + re_string_skip_bytes (input, 1); + c2 = re_string_peek_byte (input, 0); + token->opr.c = c2; + token->type = CHARACTER; + return 1; + } + if (c == '[') /* '[' is a special char in a bracket exps. */ + { + unsigned char c2; + int token_len; + if (re_string_cur_idx (input) + 1 < re_string_length (input)) + c2 = re_string_peek_byte (input, 1); + else + c2 = 0; + token->opr.c = c2; + token_len = 2; + switch (c2) + { + case '.': + token->type = OP_OPEN_COLL_ELEM; + break; + + case '=': + token->type = OP_OPEN_EQUIV_CLASS; + break; + + case ':': + if (syntax & RE_CHAR_CLASSES) + { + token->type = OP_OPEN_CHAR_CLASS; + break; + } + FALLTHROUGH; + default: + token->type = CHARACTER; + token->opr.c = c; + token_len = 1; + break; + } + return token_len; + } + switch (c) + { + case '-': + token->type = OP_CHARSET_RANGE; + break; + case ']': + token->type = OP_CLOSE_BRACKET; + break; + case '^': + token->type = OP_NON_MATCH_LIST; + break; + default: + token->type = CHARACTER; + } + return 1; +} + +/* Functions for parser. */ + +/* Entry point of the parser. + Parse the regular expression REGEXP and return the structure tree. + If an error occurs, ERR is set by error code, and return NULL. + This function build the following tree, from regular expression <reg_exp>: + CAT + / \ + / \ + <reg_exp> EOR + + CAT means concatenation. + EOR means end of regular expression. */ + +static bin_tree_t * +parse (re_string_t *regexp, regex_t *preg, reg_syntax_t syntax, + reg_errcode_t *err) +{ + re_dfa_t *dfa = preg->buffer; + bin_tree_t *tree, *eor, *root; + re_token_t current_token; + dfa->syntax = syntax; + fetch_token (¤t_token, regexp, syntax | RE_CARET_ANCHORS_HERE); + tree = parse_reg_exp (regexp, preg, ¤t_token, syntax, 0, err); + if (__glibc_unlikely (*err != REG_NOERROR && tree == NULL)) + return NULL; + eor = create_tree (dfa, NULL, NULL, END_OF_RE); + if (tree != NULL) + root = create_tree (dfa, tree, eor, CONCAT); + else + root = eor; + if (__glibc_unlikely (eor == NULL || root == NULL)) + { + *err = REG_ESPACE; + return NULL; + } + return root; +} + +/* This function build the following tree, from regular expression + <branch1>|<branch2>: + ALT + / \ + / \ + <branch1> <branch2> + + ALT means alternative, which represents the operator '|'. */ + +static bin_tree_t * +parse_reg_exp (re_string_t *regexp, regex_t *preg, re_token_t *token, + reg_syntax_t syntax, Idx nest, reg_errcode_t *err) +{ + re_dfa_t *dfa = preg->buffer; + bin_tree_t *tree, *branch = NULL; + bitset_word_t initial_bkref_map = dfa->completed_bkref_map; + tree = parse_branch (regexp, preg, token, syntax, nest, err); + if (__glibc_unlikely (*err != REG_NOERROR && tree == NULL)) + return NULL; + + while (token->type == OP_ALT) + { + fetch_token (token, regexp, syntax | RE_CARET_ANCHORS_HERE); + if (token->type != OP_ALT && token->type != END_OF_RE + && (nest == 0 || token->type != OP_CLOSE_SUBEXP)) + { + bitset_word_t accumulated_bkref_map = dfa->completed_bkref_map; + dfa->completed_bkref_map = initial_bkref_map; + branch = parse_branch (regexp, preg, token, syntax, nest, err); + if (__glibc_unlikely (*err != REG_NOERROR && branch == NULL)) + { + if (tree != NULL) + postorder (tree, free_tree, NULL); + return NULL; + } + dfa->completed_bkref_map |= accumulated_bkref_map; + } + else + branch = NULL; + tree = create_tree (dfa, tree, branch, OP_ALT); + if (__glibc_unlikely (tree == NULL)) + { + *err = REG_ESPACE; + return NULL; + } + } + return tree; +} + +/* This function build the following tree, from regular expression + <exp1><exp2>: + CAT + / \ + / \ + <exp1> <exp2> + + CAT means concatenation. */ + +static bin_tree_t * +parse_branch (re_string_t *regexp, regex_t *preg, re_token_t *token, + reg_syntax_t syntax, Idx nest, reg_errcode_t *err) +{ + bin_tree_t *tree, *expr; + re_dfa_t *dfa = preg->buffer; + tree = parse_expression (regexp, preg, token, syntax, nest, err); + if (__glibc_unlikely (*err != REG_NOERROR && tree == NULL)) + return NULL; + + while (token->type != OP_ALT && token->type != END_OF_RE + && (nest == 0 || token->type != OP_CLOSE_SUBEXP)) + { + expr = parse_expression (regexp, preg, token, syntax, nest, err); + if (__glibc_unlikely (*err != REG_NOERROR && expr == NULL)) + { + if (tree != NULL) + postorder (tree, free_tree, NULL); + return NULL; + } + if (tree != NULL && expr != NULL) + { + bin_tree_t *newtree = create_tree (dfa, tree, expr, CONCAT); + if (newtree == NULL) + { + postorder (expr, free_tree, NULL); + postorder (tree, free_tree, NULL); + *err = REG_ESPACE; + return NULL; + } + tree = newtree; + } + else if (tree == NULL) + tree = expr; + /* Otherwise expr == NULL, we don't need to create new tree. */ + } + return tree; +} + +/* This function build the following tree, from regular expression a*: + * + | + a +*/ + +static bin_tree_t * +parse_expression (re_string_t *regexp, regex_t *preg, re_token_t *token, + reg_syntax_t syntax, Idx nest, reg_errcode_t *err) +{ + re_dfa_t *dfa = preg->buffer; + bin_tree_t *tree; + switch (token->type) + { + case CHARACTER: + tree = create_token_tree (dfa, NULL, NULL, token); + if (__glibc_unlikely (tree == NULL)) + { + *err = REG_ESPACE; + return NULL; + } +#ifdef RE_ENABLE_I18N + if (dfa->mb_cur_max > 1) + { + while (!re_string_eoi (regexp) + && !re_string_first_byte (regexp, re_string_cur_idx (regexp))) + { + bin_tree_t *mbc_remain; + fetch_token (token, regexp, syntax); + mbc_remain = create_token_tree (dfa, NULL, NULL, token); + tree = create_tree (dfa, tree, mbc_remain, CONCAT); + if (__glibc_unlikely (mbc_remain == NULL || tree == NULL)) + { + *err = REG_ESPACE; + return NULL; + } + } + } +#endif + break; + + case OP_OPEN_SUBEXP: + tree = parse_sub_exp (regexp, preg, token, syntax, nest + 1, err); + if (__glibc_unlikely (*err != REG_NOERROR && tree == NULL)) + return NULL; + break; + + case OP_OPEN_BRACKET: + tree = parse_bracket_exp (regexp, dfa, token, syntax, err); + if (__glibc_unlikely (*err != REG_NOERROR && tree == NULL)) + return NULL; + break; + + case OP_BACK_REF: + if (!__glibc_likely (dfa->completed_bkref_map & (1 << token->opr.idx))) + { + *err = REG_ESUBREG; + return NULL; + } + dfa->used_bkref_map |= 1 << token->opr.idx; + tree = create_token_tree (dfa, NULL, NULL, token); + if (__glibc_unlikely (tree == NULL)) + { + *err = REG_ESPACE; + return NULL; + } + ++dfa->nbackref; + dfa->has_mb_node = 1; + break; + + case OP_OPEN_DUP_NUM: + if (syntax & RE_CONTEXT_INVALID_DUP) + { + *err = REG_BADRPT; + return NULL; + } + FALLTHROUGH; + case OP_DUP_ASTERISK: + case OP_DUP_PLUS: + case OP_DUP_QUESTION: + if (syntax & RE_CONTEXT_INVALID_OPS) + { + *err = REG_BADRPT; + return NULL; + } + else if (syntax & RE_CONTEXT_INDEP_OPS) + { + fetch_token (token, regexp, syntax); + return parse_expression (regexp, preg, token, syntax, nest, err); + } + FALLTHROUGH; + case OP_CLOSE_SUBEXP: + if ((token->type == OP_CLOSE_SUBEXP) && + !(syntax & RE_UNMATCHED_RIGHT_PAREN_ORD)) + { + *err = REG_ERPAREN; + return NULL; + } + FALLTHROUGH; + case OP_CLOSE_DUP_NUM: + /* We treat it as a normal character. */ + + /* Then we can these characters as normal characters. */ + token->type = CHARACTER; + /* mb_partial and word_char bits should be initialized already + by peek_token. */ + tree = create_token_tree (dfa, NULL, NULL, token); + if (__glibc_unlikely (tree == NULL)) + { + *err = REG_ESPACE; + return NULL; + } + break; + + case ANCHOR: + if ((token->opr.ctx_type + & (WORD_DELIM | NOT_WORD_DELIM | WORD_FIRST | WORD_LAST)) + && dfa->word_ops_used == 0) + init_word_char (dfa); + if (token->opr.ctx_type == WORD_DELIM + || token->opr.ctx_type == NOT_WORD_DELIM) + { + bin_tree_t *tree_first, *tree_last; + if (token->opr.ctx_type == WORD_DELIM) + { + token->opr.ctx_type = WORD_FIRST; + tree_first = create_token_tree (dfa, NULL, NULL, token); + token->opr.ctx_type = WORD_LAST; + } + else + { + token->opr.ctx_type = INSIDE_WORD; + tree_first = create_token_tree (dfa, NULL, NULL, token); + token->opr.ctx_type = INSIDE_NOTWORD; + } + tree_last = create_token_tree (dfa, NULL, NULL, token); + tree = create_tree (dfa, tree_first, tree_last, OP_ALT); + if (__glibc_unlikely (tree_first == NULL || tree_last == NULL + || tree == NULL)) + { + *err = REG_ESPACE; + return NULL; + } + } + else + { + tree = create_token_tree (dfa, NULL, NULL, token); + if (__glibc_unlikely (tree == NULL)) + { + *err = REG_ESPACE; + return NULL; + } + } + /* We must return here, since ANCHORs can't be followed + by repetition operators. + eg. RE"^*" is invalid or "<ANCHOR(^)><CHAR(*)>", + it must not be "<ANCHOR(^)><REPEAT(*)>". */ + fetch_token (token, regexp, syntax); + return tree; + + case OP_PERIOD: + tree = create_token_tree (dfa, NULL, NULL, token); + if (__glibc_unlikely (tree == NULL)) + { + *err = REG_ESPACE; + return NULL; + } + if (dfa->mb_cur_max > 1) + dfa->has_mb_node = 1; + break; + + case OP_WORD: + case OP_NOTWORD: + tree = build_charclass_op (dfa, regexp->trans, + "alnum", + "_", + token->type == OP_NOTWORD, err); + if (__glibc_unlikely (*err != REG_NOERROR && tree == NULL)) + return NULL; + break; + + case OP_SPACE: + case OP_NOTSPACE: + tree = build_charclass_op (dfa, regexp->trans, + "space", + "", + token->type == OP_NOTSPACE, err); + if (__glibc_unlikely (*err != REG_NOERROR && tree == NULL)) + return NULL; + break; + + case OP_ALT: + case END_OF_RE: + return NULL; + + case BACK_SLASH: + *err = REG_EESCAPE; + return NULL; + + default: + /* Must not happen? */ +#ifdef DEBUG + assert (0); +#endif + return NULL; + } + fetch_token (token, regexp, syntax); + + while (token->type == OP_DUP_ASTERISK || token->type == OP_DUP_PLUS + || token->type == OP_DUP_QUESTION || token->type == OP_OPEN_DUP_NUM) + { + bin_tree_t *dup_tree = parse_dup_op (tree, regexp, dfa, token, + syntax, err); + if (__glibc_unlikely (*err != REG_NOERROR && dup_tree == NULL)) + { + if (tree != NULL) + postorder (tree, free_tree, NULL); + return NULL; + } + tree = dup_tree; + /* In BRE consecutive duplications are not allowed. */ + if ((syntax & RE_CONTEXT_INVALID_DUP) + && (token->type == OP_DUP_ASTERISK + || token->type == OP_OPEN_DUP_NUM)) + { + if (tree != NULL) + postorder (tree, free_tree, NULL); + *err = REG_BADRPT; + return NULL; + } + } + + return tree; +} + +/* This function build the following tree, from regular expression + (<reg_exp>): + SUBEXP + | + <reg_exp> +*/ + +static bin_tree_t * +parse_sub_exp (re_string_t *regexp, regex_t *preg, re_token_t *token, + reg_syntax_t syntax, Idx nest, reg_errcode_t *err) +{ + re_dfa_t *dfa = preg->buffer; + bin_tree_t *tree; + size_t cur_nsub; + cur_nsub = preg->re_nsub++; + + fetch_token (token, regexp, syntax | RE_CARET_ANCHORS_HERE); + + /* The subexpression may be a null string. */ + if (token->type == OP_CLOSE_SUBEXP) + tree = NULL; + else + { + tree = parse_reg_exp (regexp, preg, token, syntax, nest, err); + if (__glibc_unlikely (*err == REG_NOERROR + && token->type != OP_CLOSE_SUBEXP)) + { + if (tree != NULL) + postorder (tree, free_tree, NULL); + *err = REG_EPAREN; + } + if (__glibc_unlikely (*err != REG_NOERROR)) + return NULL; + } + + if (cur_nsub <= '9' - '1') + dfa->completed_bkref_map |= 1 << cur_nsub; + + tree = create_tree (dfa, tree, NULL, SUBEXP); + if (__glibc_unlikely (tree == NULL)) + { + *err = REG_ESPACE; + return NULL; + } + tree->token.opr.idx = cur_nsub; + return tree; +} + +/* This function parse repetition operators like "*", "+", "{1,3}" etc. */ + +static bin_tree_t * +parse_dup_op (bin_tree_t *elem, re_string_t *regexp, re_dfa_t *dfa, + re_token_t *token, reg_syntax_t syntax, reg_errcode_t *err) +{ + bin_tree_t *tree = NULL, *old_tree = NULL; + Idx i, start, end, start_idx = re_string_cur_idx (regexp); + re_token_t start_token = *token; + + if (token->type == OP_OPEN_DUP_NUM) + { + end = 0; + start = fetch_number (regexp, token, syntax); + if (start == -1) + { + if (token->type == CHARACTER && token->opr.c == ',') + start = 0; /* We treat "{,m}" as "{0,m}". */ + else + { + *err = REG_BADBR; /* <re>{} is invalid. */ + return NULL; + } + } + if (__glibc_likely (start != -2)) + { + /* We treat "{n}" as "{n,n}". */ + end = ((token->type == OP_CLOSE_DUP_NUM) ? start + : ((token->type == CHARACTER && token->opr.c == ',') + ? fetch_number (regexp, token, syntax) : -2)); + } + if (__glibc_unlikely (start == -2 || end == -2)) + { + /* Invalid sequence. */ + if (__glibc_unlikely (!(syntax & RE_INVALID_INTERVAL_ORD))) + { + if (token->type == END_OF_RE) + *err = REG_EBRACE; + else + *err = REG_BADBR; + + return NULL; + } + + /* If the syntax bit is set, rollback. */ + re_string_set_index (regexp, start_idx); + *token = start_token; + token->type = CHARACTER; + /* mb_partial and word_char bits should be already initialized by + peek_token. */ + return elem; + } + + if (__glibc_unlikely ((end != -1 && start > end) + || token->type != OP_CLOSE_DUP_NUM)) + { + /* First number greater than second. */ + *err = REG_BADBR; + return NULL; + } + + if (__glibc_unlikely (RE_DUP_MAX < (end == -1 ? start : end))) + { + *err = REG_ESIZE; + return NULL; + } + } + else + { + start = (token->type == OP_DUP_PLUS) ? 1 : 0; + end = (token->type == OP_DUP_QUESTION) ? 1 : -1; + } + + fetch_token (token, regexp, syntax); + + if (__glibc_unlikely (elem == NULL)) + return NULL; + if (__glibc_unlikely (start == 0 && end == 0)) + { + postorder (elem, free_tree, NULL); + return NULL; + } + + /* Extract "<re>{n,m}" to "<re><re>...<re><re>{0,<m-n>}". */ + if (__glibc_unlikely (start > 0)) + { + tree = elem; + for (i = 2; i <= start; ++i) + { + elem = duplicate_tree (elem, dfa); + tree = create_tree (dfa, tree, elem, CONCAT); + if (__glibc_unlikely (elem == NULL || tree == NULL)) + goto parse_dup_op_espace; + } + + if (start == end) + return tree; + + /* Duplicate ELEM before it is marked optional. */ + elem = duplicate_tree (elem, dfa); + if (__glibc_unlikely (elem == NULL)) + goto parse_dup_op_espace; + old_tree = tree; + } + else + old_tree = NULL; + + if (elem->token.type == SUBEXP) + { + uintptr_t subidx = elem->token.opr.idx; + postorder (elem, mark_opt_subexp, (void *) subidx); + } + + tree = create_tree (dfa, elem, NULL, + (end == -1 ? OP_DUP_ASTERISK : OP_ALT)); + if (__glibc_unlikely (tree == NULL)) + goto parse_dup_op_espace; + + /* This loop is actually executed only when end != -1, + to rewrite <re>{0,n} as (<re>(<re>...<re>?)?)?... We have + already created the start+1-th copy. */ + if (TYPE_SIGNED (Idx) || end != -1) + for (i = start + 2; i <= end; ++i) + { + elem = duplicate_tree (elem, dfa); + tree = create_tree (dfa, tree, elem, CONCAT); + if (__glibc_unlikely (elem == NULL || tree == NULL)) + goto parse_dup_op_espace; + + tree = create_tree (dfa, tree, NULL, OP_ALT); + if (__glibc_unlikely (tree == NULL)) + goto parse_dup_op_espace; + } + + if (old_tree) + tree = create_tree (dfa, old_tree, tree, CONCAT); + + return tree; + + parse_dup_op_espace: + *err = REG_ESPACE; + return NULL; +} + +/* Size of the names for collating symbol/equivalence_class/character_class. + I'm not sure, but maybe enough. */ +#define BRACKET_NAME_BUF_SIZE 32 + +#ifndef _LIBC + +# ifdef RE_ENABLE_I18N +/* Convert the byte B to the corresponding wide character. In a + unibyte locale, treat B as itself. In a multibyte locale, return + WEOF if B is an encoding error. */ +static wint_t +parse_byte (unsigned char b, re_charset_t *mbcset) +{ + return mbcset == NULL ? b : __btowc (b); +} +# endif + + /* Local function for parse_bracket_exp only used in case of NOT _LIBC. + Build the range expression which starts from START_ELEM, and ends + at END_ELEM. The result are written to MBCSET and SBCSET. + RANGE_ALLOC is the allocated size of mbcset->range_starts, and + mbcset->range_ends, is a pointer argument since we may + update it. */ + +static reg_errcode_t +# ifdef RE_ENABLE_I18N +build_range_exp (const reg_syntax_t syntax, + bitset_t sbcset, + re_charset_t *mbcset, + Idx *range_alloc, + const bracket_elem_t *start_elem, + const bracket_elem_t *end_elem) +# else /* not RE_ENABLE_I18N */ +build_range_exp (const reg_syntax_t syntax, + bitset_t sbcset, + const bracket_elem_t *start_elem, + const bracket_elem_t *end_elem) +# endif /* not RE_ENABLE_I18N */ +{ + unsigned int start_ch, end_ch; + /* Equivalence Classes and Character Classes can't be a range start/end. */ + if (__glibc_unlikely (start_elem->type == EQUIV_CLASS + || start_elem->type == CHAR_CLASS + || end_elem->type == EQUIV_CLASS + || end_elem->type == CHAR_CLASS)) + return REG_ERANGE; + + /* We can handle no multi character collating elements without libc + support. */ + if (__glibc_unlikely ((start_elem->type == COLL_SYM + && strlen ((char *) start_elem->opr.name) > 1) + || (end_elem->type == COLL_SYM + && strlen ((char *) end_elem->opr.name) > 1))) + return REG_ECOLLATE; + +# ifdef RE_ENABLE_I18N + { + wchar_t wc; + wint_t start_wc; + wint_t end_wc; + + start_ch = ((start_elem->type == SB_CHAR) ? start_elem->opr.ch + : ((start_elem->type == COLL_SYM) ? start_elem->opr.name[0] + : 0)); + end_ch = ((end_elem->type == SB_CHAR) ? end_elem->opr.ch + : ((end_elem->type == COLL_SYM) ? end_elem->opr.name[0] + : 0)); + start_wc = ((start_elem->type == SB_CHAR || start_elem->type == COLL_SYM) + ? parse_byte (start_ch, mbcset) : start_elem->opr.wch); + end_wc = ((end_elem->type == SB_CHAR || end_elem->type == COLL_SYM) + ? parse_byte (end_ch, mbcset) : end_elem->opr.wch); + if (start_wc == WEOF || end_wc == WEOF) + return REG_ECOLLATE; + else if (__glibc_unlikely ((syntax & RE_NO_EMPTY_RANGES) + && start_wc > end_wc)) + return REG_ERANGE; + + /* Got valid collation sequence values, add them as a new entry. + However, for !_LIBC we have no collation elements: if the + character set is single byte, the single byte character set + that we build below suffices. parse_bracket_exp passes + no MBCSET if dfa->mb_cur_max == 1. */ + if (mbcset) + { + /* Check the space of the arrays. */ + if (__glibc_unlikely (*range_alloc == mbcset->nranges)) + { + /* There is not enough space, need realloc. */ + wchar_t *new_array_start, *new_array_end; + Idx new_nranges; + + /* +1 in case of mbcset->nranges is 0. */ + new_nranges = 2 * mbcset->nranges + 1; + /* Use realloc since mbcset->range_starts and mbcset->range_ends + are NULL if *range_alloc == 0. */ + new_array_start = re_realloc (mbcset->range_starts, wchar_t, + new_nranges); + new_array_end = re_realloc (mbcset->range_ends, wchar_t, + new_nranges); + + if (__glibc_unlikely (new_array_start == NULL + || new_array_end == NULL)) + { + re_free (new_array_start); + re_free (new_array_end); + return REG_ESPACE; + } + + mbcset->range_starts = new_array_start; + mbcset->range_ends = new_array_end; + *range_alloc = new_nranges; + } + + mbcset->range_starts[mbcset->nranges] = start_wc; + mbcset->range_ends[mbcset->nranges++] = end_wc; + } + + /* Build the table for single byte characters. */ + for (wc = 0; wc < SBC_MAX; ++wc) + { + if (start_wc <= wc && wc <= end_wc) + bitset_set (sbcset, wc); + } + } +# else /* not RE_ENABLE_I18N */ + { + unsigned int ch; + start_ch = ((start_elem->type == SB_CHAR ) ? start_elem->opr.ch + : ((start_elem->type == COLL_SYM) ? start_elem->opr.name[0] + : 0)); + end_ch = ((end_elem->type == SB_CHAR ) ? end_elem->opr.ch + : ((end_elem->type == COLL_SYM) ? end_elem->opr.name[0] + : 0)); + if (start_ch > end_ch) + return REG_ERANGE; + /* Build the table for single byte characters. */ + for (ch = 0; ch < SBC_MAX; ++ch) + if (start_ch <= ch && ch <= end_ch) + bitset_set (sbcset, ch); + } +# endif /* not RE_ENABLE_I18N */ + return REG_NOERROR; +} +#endif /* not _LIBC */ + +#ifndef _LIBC +/* Helper function for parse_bracket_exp only used in case of NOT _LIBC.. + Build the collating element which is represented by NAME. + The result are written to MBCSET and SBCSET. + COLL_SYM_ALLOC is the allocated size of mbcset->coll_sym, is a + pointer argument since we may update it. */ + +static reg_errcode_t +# ifdef RE_ENABLE_I18N +build_collating_symbol (bitset_t sbcset, re_charset_t *mbcset, + Idx *coll_sym_alloc, const unsigned char *name) +# else /* not RE_ENABLE_I18N */ +build_collating_symbol (bitset_t sbcset, const unsigned char *name) +# endif /* not RE_ENABLE_I18N */ +{ + size_t name_len = strlen ((const char *) name); + if (__glibc_unlikely (name_len != 1)) + return REG_ECOLLATE; + else + { + bitset_set (sbcset, name[0]); + return REG_NOERROR; + } +} +#endif /* not _LIBC */ + +/* This function parse bracket expression like "[abc]", "[a-c]", + "[[.a-a.]]" etc. */ + +static bin_tree_t * +parse_bracket_exp (re_string_t *regexp, re_dfa_t *dfa, re_token_t *token, + reg_syntax_t syntax, reg_errcode_t *err) +{ +#ifdef _LIBC + const unsigned char *collseqmb; + const char *collseqwc; + uint32_t nrules; + int32_t table_size; + const int32_t *symb_table; + const unsigned char *extra; + + /* Local function for parse_bracket_exp used in _LIBC environment. + Seek the collating symbol entry corresponding to NAME. + Return the index of the symbol in the SYMB_TABLE, + or -1 if not found. */ + + auto inline int32_t + __attribute__ ((always_inline)) + seek_collating_symbol_entry (const unsigned char *name, size_t name_len) + { + int32_t elem; + + for (elem = 0; elem < table_size; elem++) + if (symb_table[2 * elem] != 0) + { + int32_t idx = symb_table[2 * elem + 1]; + /* Skip the name of collating element name. */ + idx += 1 + extra[idx]; + if (/* Compare the length of the name. */ + name_len == extra[idx] + /* Compare the name. */ + && memcmp (name, &extra[idx + 1], name_len) == 0) + /* Yep, this is the entry. */ + return elem; + } + return -1; + } + + /* Local function for parse_bracket_exp used in _LIBC environment. + Look up the collation sequence value of BR_ELEM. + Return the value if succeeded, UINT_MAX otherwise. */ + + auto inline unsigned int + __attribute__ ((always_inline)) + lookup_collation_sequence_value (bracket_elem_t *br_elem) + { + if (br_elem->type == SB_CHAR) + { + /* + if (MB_CUR_MAX == 1) + */ + if (nrules == 0) + return collseqmb[br_elem->opr.ch]; + else + { + wint_t wc = __btowc (br_elem->opr.ch); + return __collseq_table_lookup (collseqwc, wc); + } + } + else if (br_elem->type == MB_CHAR) + { + if (nrules != 0) + return __collseq_table_lookup (collseqwc, br_elem->opr.wch); + } + else if (br_elem->type == COLL_SYM) + { + size_t sym_name_len = strlen ((char *) br_elem->opr.name); + if (nrules != 0) + { + int32_t elem, idx; + elem = seek_collating_symbol_entry (br_elem->opr.name, + sym_name_len); + if (elem != -1) + { + /* We found the entry. */ + idx = symb_table[2 * elem + 1]; + /* Skip the name of collating element name. */ + idx += 1 + extra[idx]; + /* Skip the byte sequence of the collating element. */ + idx += 1 + extra[idx]; + /* Adjust for the alignment. */ + idx = (idx + 3) & ~3; + /* Skip the multibyte collation sequence value. */ + idx += sizeof (unsigned int); + /* Skip the wide char sequence of the collating element. */ + idx += sizeof (unsigned int) * + (1 + *(unsigned int *) (extra + idx)); + /* Return the collation sequence value. */ + return *(unsigned int *) (extra + idx); + } + else if (sym_name_len == 1) + { + /* No valid character. Match it as a single byte + character. */ + return collseqmb[br_elem->opr.name[0]]; + } + } + else if (sym_name_len == 1) + return collseqmb[br_elem->opr.name[0]]; + } + return UINT_MAX; + } + + /* Local function for parse_bracket_exp used in _LIBC environment. + Build the range expression which starts from START_ELEM, and ends + at END_ELEM. The result are written to MBCSET and SBCSET. + RANGE_ALLOC is the allocated size of mbcset->range_starts, and + mbcset->range_ends, is a pointer argument since we may + update it. */ + + auto inline reg_errcode_t + __attribute__ ((always_inline)) + build_range_exp (bitset_t sbcset, re_charset_t *mbcset, int *range_alloc, + bracket_elem_t *start_elem, bracket_elem_t *end_elem) + { + unsigned int ch; + uint32_t start_collseq; + uint32_t end_collseq; + + /* Equivalence Classes and Character Classes can't be a range + start/end. */ + if (__glibc_unlikely (start_elem->type == EQUIV_CLASS + || start_elem->type == CHAR_CLASS + || end_elem->type == EQUIV_CLASS + || end_elem->type == CHAR_CLASS)) + return REG_ERANGE; + + /* FIXME: Implement rational ranges here, too. */ + start_collseq = lookup_collation_sequence_value (start_elem); + end_collseq = lookup_collation_sequence_value (end_elem); + /* Check start/end collation sequence values. */ + if (__glibc_unlikely (start_collseq == UINT_MAX + || end_collseq == UINT_MAX)) + return REG_ECOLLATE; + if (__glibc_unlikely ((syntax & RE_NO_EMPTY_RANGES) + && start_collseq > end_collseq)) + return REG_ERANGE; + + /* Got valid collation sequence values, add them as a new entry. + However, if we have no collation elements, and the character set + is single byte, the single byte character set that we + build below suffices. */ + if (nrules > 0 || dfa->mb_cur_max > 1) + { + /* Check the space of the arrays. */ + if (__glibc_unlikely (*range_alloc == mbcset->nranges)) + { + /* There is not enough space, need realloc. */ + uint32_t *new_array_start; + uint32_t *new_array_end; + Idx new_nranges; + + /* +1 in case of mbcset->nranges is 0. */ + new_nranges = 2 * mbcset->nranges + 1; + new_array_start = re_realloc (mbcset->range_starts, uint32_t, + new_nranges); + new_array_end = re_realloc (mbcset->range_ends, uint32_t, + new_nranges); + + if (__glibc_unlikely (new_array_start == NULL + || new_array_end == NULL)) + return REG_ESPACE; + + mbcset->range_starts = new_array_start; + mbcset->range_ends = new_array_end; + *range_alloc = new_nranges; + } + + mbcset->range_starts[mbcset->nranges] = start_collseq; + mbcset->range_ends[mbcset->nranges++] = end_collseq; + } + + /* Build the table for single byte characters. */ + for (ch = 0; ch < SBC_MAX; ch++) + { + uint32_t ch_collseq; + /* + if (MB_CUR_MAX == 1) + */ + if (nrules == 0) + ch_collseq = collseqmb[ch]; + else + ch_collseq = __collseq_table_lookup (collseqwc, __btowc (ch)); + if (start_collseq <= ch_collseq && ch_collseq <= end_collseq) + bitset_set (sbcset, ch); + } + return REG_NOERROR; + } + + /* Local function for parse_bracket_exp used in _LIBC environment. + Build the collating element which is represented by NAME. + The result are written to MBCSET and SBCSET. + COLL_SYM_ALLOC is the allocated size of mbcset->coll_sym, is a + pointer argument since we may update it. */ + + auto inline reg_errcode_t + __attribute__ ((always_inline)) + build_collating_symbol (bitset_t sbcset, re_charset_t *mbcset, + Idx *coll_sym_alloc, const unsigned char *name) + { + int32_t elem, idx; + size_t name_len = strlen ((const char *) name); + if (nrules != 0) + { + elem = seek_collating_symbol_entry (name, name_len); + if (elem != -1) + { + /* We found the entry. */ + idx = symb_table[2 * elem + 1]; + /* Skip the name of collating element name. */ + idx += 1 + extra[idx]; + } + else if (name_len == 1) + { + /* No valid character, treat it as a normal + character. */ + bitset_set (sbcset, name[0]); + return REG_NOERROR; + } + else + return REG_ECOLLATE; + + /* Got valid collation sequence, add it as a new entry. */ + /* Check the space of the arrays. */ + if (__glibc_unlikely (*coll_sym_alloc == mbcset->ncoll_syms)) + { + /* Not enough, realloc it. */ + /* +1 in case of mbcset->ncoll_syms is 0. */ + Idx new_coll_sym_alloc = 2 * mbcset->ncoll_syms + 1; + /* Use realloc since mbcset->coll_syms is NULL + if *alloc == 0. */ + int32_t *new_coll_syms = re_realloc (mbcset->coll_syms, int32_t, + new_coll_sym_alloc); + if (__glibc_unlikely (new_coll_syms == NULL)) + return REG_ESPACE; + mbcset->coll_syms = new_coll_syms; + *coll_sym_alloc = new_coll_sym_alloc; + } + mbcset->coll_syms[mbcset->ncoll_syms++] = idx; + return REG_NOERROR; + } + else + { + if (__glibc_unlikely (name_len != 1)) + return REG_ECOLLATE; + else + { + bitset_set (sbcset, name[0]); + return REG_NOERROR; + } + } + } +#endif + + re_token_t br_token; + re_bitset_ptr_t sbcset; +#ifdef RE_ENABLE_I18N + re_charset_t *mbcset; + Idx coll_sym_alloc = 0, range_alloc = 0, mbchar_alloc = 0; + Idx equiv_class_alloc = 0, char_class_alloc = 0; +#endif /* not RE_ENABLE_I18N */ + bool non_match = false; + bin_tree_t *work_tree; + int token_len; + bool first_round = true; +#ifdef _LIBC + collseqmb = (const unsigned char *) + _NL_CURRENT (LC_COLLATE, _NL_COLLATE_COLLSEQMB); + nrules = _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES); + if (nrules) + { + /* + if (MB_CUR_MAX > 1) + */ + collseqwc = _NL_CURRENT (LC_COLLATE, _NL_COLLATE_COLLSEQWC); + table_size = _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_SYMB_HASH_SIZEMB); + symb_table = (const int32_t *) _NL_CURRENT (LC_COLLATE, + _NL_COLLATE_SYMB_TABLEMB); + extra = (const unsigned char *) _NL_CURRENT (LC_COLLATE, + _NL_COLLATE_SYMB_EXTRAMB); + } +#endif + sbcset = (re_bitset_ptr_t) calloc (sizeof (bitset_t), 1); +#ifdef RE_ENABLE_I18N + mbcset = (re_charset_t *) calloc (sizeof (re_charset_t), 1); +#endif /* RE_ENABLE_I18N */ +#ifdef RE_ENABLE_I18N + if (__glibc_unlikely (sbcset == NULL || mbcset == NULL)) +#else + if (__glibc_unlikely (sbcset == NULL)) +#endif /* RE_ENABLE_I18N */ + { + re_free (sbcset); +#ifdef RE_ENABLE_I18N + re_free (mbcset); +#endif + *err = REG_ESPACE; + return NULL; + } + + token_len = peek_token_bracket (token, regexp, syntax); + if (__glibc_unlikely (token->type == END_OF_RE)) + { + *err = REG_BADPAT; + goto parse_bracket_exp_free_return; + } + if (token->type == OP_NON_MATCH_LIST) + { +#ifdef RE_ENABLE_I18N + mbcset->non_match = 1; +#endif /* not RE_ENABLE_I18N */ + non_match = true; + if (syntax & RE_HAT_LISTS_NOT_NEWLINE) + bitset_set (sbcset, '\n'); + re_string_skip_bytes (regexp, token_len); /* Skip a token. */ + token_len = peek_token_bracket (token, regexp, syntax); + if (__glibc_unlikely (token->type == END_OF_RE)) + { + *err = REG_BADPAT; + goto parse_bracket_exp_free_return; + } + } + + /* We treat the first ']' as a normal character. */ + if (token->type == OP_CLOSE_BRACKET) + token->type = CHARACTER; + + while (1) + { + bracket_elem_t start_elem, end_elem; + unsigned char start_name_buf[BRACKET_NAME_BUF_SIZE]; + unsigned char end_name_buf[BRACKET_NAME_BUF_SIZE]; + reg_errcode_t ret; + int token_len2 = 0; + bool is_range_exp = false; + re_token_t token2; + + start_elem.opr.name = start_name_buf; + start_elem.type = COLL_SYM; + ret = parse_bracket_element (&start_elem, regexp, token, token_len, dfa, + syntax, first_round); + if (__glibc_unlikely (ret != REG_NOERROR)) + { + *err = ret; + goto parse_bracket_exp_free_return; + } + first_round = false; + + /* Get information about the next token. We need it in any case. */ + token_len = peek_token_bracket (token, regexp, syntax); + + /* Do not check for ranges if we know they are not allowed. */ + if (start_elem.type != CHAR_CLASS && start_elem.type != EQUIV_CLASS) + { + if (__glibc_unlikely (token->type == END_OF_RE)) + { + *err = REG_EBRACK; + goto parse_bracket_exp_free_return; + } + if (token->type == OP_CHARSET_RANGE) + { + re_string_skip_bytes (regexp, token_len); /* Skip '-'. */ + token_len2 = peek_token_bracket (&token2, regexp, syntax); + if (__glibc_unlikely (token2.type == END_OF_RE)) + { + *err = REG_EBRACK; + goto parse_bracket_exp_free_return; + } + if (token2.type == OP_CLOSE_BRACKET) + { + /* We treat the last '-' as a normal character. */ + re_string_skip_bytes (regexp, -token_len); + token->type = CHARACTER; + } + else + is_range_exp = true; + } + } + + if (is_range_exp == true) + { + end_elem.opr.name = end_name_buf; + end_elem.type = COLL_SYM; + ret = parse_bracket_element (&end_elem, regexp, &token2, token_len2, + dfa, syntax, true); + if (__glibc_unlikely (ret != REG_NOERROR)) + { + *err = ret; + goto parse_bracket_exp_free_return; + } + + token_len = peek_token_bracket (token, regexp, syntax); + +#ifdef _LIBC + *err = build_range_exp (sbcset, mbcset, &range_alloc, + &start_elem, &end_elem); +#else +# ifdef RE_ENABLE_I18N + *err = build_range_exp (syntax, sbcset, + dfa->mb_cur_max > 1 ? mbcset : NULL, + &range_alloc, &start_elem, &end_elem); +# else + *err = build_range_exp (syntax, sbcset, &start_elem, &end_elem); +# endif +#endif /* RE_ENABLE_I18N */ + if (__glibc_unlikely (*err != REG_NOERROR)) + goto parse_bracket_exp_free_return; + } + else + { + switch (start_elem.type) + { + case SB_CHAR: + bitset_set (sbcset, start_elem.opr.ch); + break; +#ifdef RE_ENABLE_I18N + case MB_CHAR: + /* Check whether the array has enough space. */ + if (__glibc_unlikely (mbchar_alloc == mbcset->nmbchars)) + { + wchar_t *new_mbchars; + /* Not enough, realloc it. */ + /* +1 in case of mbcset->nmbchars is 0. */ + mbchar_alloc = 2 * mbcset->nmbchars + 1; + /* Use realloc since array is NULL if *alloc == 0. */ + new_mbchars = re_realloc (mbcset->mbchars, wchar_t, + mbchar_alloc); + if (__glibc_unlikely (new_mbchars == NULL)) + goto parse_bracket_exp_espace; + mbcset->mbchars = new_mbchars; + } + mbcset->mbchars[mbcset->nmbchars++] = start_elem.opr.wch; + break; +#endif /* RE_ENABLE_I18N */ + case EQUIV_CLASS: + *err = build_equiv_class (sbcset, +#ifdef RE_ENABLE_I18N + mbcset, &equiv_class_alloc, +#endif /* RE_ENABLE_I18N */ + start_elem.opr.name); + if (__glibc_unlikely (*err != REG_NOERROR)) + goto parse_bracket_exp_free_return; + break; + case COLL_SYM: + *err = build_collating_symbol (sbcset, +#ifdef RE_ENABLE_I18N + mbcset, &coll_sym_alloc, +#endif /* RE_ENABLE_I18N */ + start_elem.opr.name); + if (__glibc_unlikely (*err != REG_NOERROR)) + goto parse_bracket_exp_free_return; + break; + case CHAR_CLASS: + *err = build_charclass (regexp->trans, sbcset, +#ifdef RE_ENABLE_I18N + mbcset, &char_class_alloc, +#endif /* RE_ENABLE_I18N */ + (const char *) start_elem.opr.name, + syntax); + if (__glibc_unlikely (*err != REG_NOERROR)) + goto parse_bracket_exp_free_return; + break; + default: + assert (0); + break; + } + } + if (__glibc_unlikely (token->type == END_OF_RE)) + { + *err = REG_EBRACK; + goto parse_bracket_exp_free_return; + } + if (token->type == OP_CLOSE_BRACKET) + break; + } + + re_string_skip_bytes (regexp, token_len); /* Skip a token. */ + + /* If it is non-matching list. */ + if (non_match) + bitset_not (sbcset); + +#ifdef RE_ENABLE_I18N + /* Ensure only single byte characters are set. */ + if (dfa->mb_cur_max > 1) + bitset_mask (sbcset, dfa->sb_char); + + if (mbcset->nmbchars || mbcset->ncoll_syms || mbcset->nequiv_classes + || mbcset->nranges || (dfa->mb_cur_max > 1 && (mbcset->nchar_classes + || mbcset->non_match))) + { + bin_tree_t *mbc_tree; + int sbc_idx; + /* Build a tree for complex bracket. */ + dfa->has_mb_node = 1; + br_token.type = COMPLEX_BRACKET; + br_token.opr.mbcset = mbcset; + mbc_tree = create_token_tree (dfa, NULL, NULL, &br_token); + if (__glibc_unlikely (mbc_tree == NULL)) + goto parse_bracket_exp_espace; + for (sbc_idx = 0; sbc_idx < BITSET_WORDS; ++sbc_idx) + if (sbcset[sbc_idx]) + break; + /* If there are no bits set in sbcset, there is no point + of having both SIMPLE_BRACKET and COMPLEX_BRACKET. */ + if (sbc_idx < BITSET_WORDS) + { + /* Build a tree for simple bracket. */ + br_token.type = SIMPLE_BRACKET; + br_token.opr.sbcset = sbcset; + work_tree = create_token_tree (dfa, NULL, NULL, &br_token); + if (__glibc_unlikely (work_tree == NULL)) + goto parse_bracket_exp_espace; + + /* Then join them by ALT node. */ + work_tree = create_tree (dfa, work_tree, mbc_tree, OP_ALT); + if (__glibc_unlikely (work_tree == NULL)) + goto parse_bracket_exp_espace; + } + else + { + re_free (sbcset); + work_tree = mbc_tree; + } + } + else +#endif /* not RE_ENABLE_I18N */ + { +#ifdef RE_ENABLE_I18N + free_charset (mbcset); +#endif + /* Build a tree for simple bracket. */ + br_token.type = SIMPLE_BRACKET; + br_token.opr.sbcset = sbcset; + work_tree = create_token_tree (dfa, NULL, NULL, &br_token); + if (__glibc_unlikely (work_tree == NULL)) + goto parse_bracket_exp_espace; + } + return work_tree; + + parse_bracket_exp_espace: + *err = REG_ESPACE; + parse_bracket_exp_free_return: + re_free (sbcset); +#ifdef RE_ENABLE_I18N + free_charset (mbcset); +#endif /* RE_ENABLE_I18N */ + return NULL; +} + +/* Parse an element in the bracket expression. */ + +static reg_errcode_t +parse_bracket_element (bracket_elem_t *elem, re_string_t *regexp, + re_token_t *token, int token_len, re_dfa_t *dfa, + reg_syntax_t syntax, bool accept_hyphen) +{ +#ifdef RE_ENABLE_I18N + int cur_char_size; + cur_char_size = re_string_char_size_at (regexp, re_string_cur_idx (regexp)); + if (cur_char_size > 1) + { + elem->type = MB_CHAR; + elem->opr.wch = re_string_wchar_at (regexp, re_string_cur_idx (regexp)); + re_string_skip_bytes (regexp, cur_char_size); + return REG_NOERROR; + } +#endif /* RE_ENABLE_I18N */ + re_string_skip_bytes (regexp, token_len); /* Skip a token. */ + if (token->type == OP_OPEN_COLL_ELEM || token->type == OP_OPEN_CHAR_CLASS + || token->type == OP_OPEN_EQUIV_CLASS) + return parse_bracket_symbol (elem, regexp, token); + if (__glibc_unlikely (token->type == OP_CHARSET_RANGE) && !accept_hyphen) + { + /* A '-' must only appear as anything but a range indicator before + the closing bracket. Everything else is an error. */ + re_token_t token2; + (void) peek_token_bracket (&token2, regexp, syntax); + if (token2.type != OP_CLOSE_BRACKET) + /* The actual error value is not standardized since this whole + case is undefined. But ERANGE makes good sense. */ + return REG_ERANGE; + } + elem->type = SB_CHAR; + elem->opr.ch = token->opr.c; + return REG_NOERROR; +} + +/* Parse a bracket symbol in the bracket expression. Bracket symbols are + such as [:<character_class>:], [.<collating_element>.], and + [=<equivalent_class>=]. */ + +static reg_errcode_t +parse_bracket_symbol (bracket_elem_t *elem, re_string_t *regexp, + re_token_t *token) +{ + unsigned char ch, delim = token->opr.c; + int i = 0; + if (re_string_eoi(regexp)) + return REG_EBRACK; + for (;; ++i) + { + if (i >= BRACKET_NAME_BUF_SIZE) + return REG_EBRACK; + if (token->type == OP_OPEN_CHAR_CLASS) + ch = re_string_fetch_byte_case (regexp); + else + ch = re_string_fetch_byte (regexp); + if (re_string_eoi(regexp)) + return REG_EBRACK; + if (ch == delim && re_string_peek_byte (regexp, 0) == ']') + break; + elem->opr.name[i] = ch; + } + re_string_skip_bytes (regexp, 1); + elem->opr.name[i] = '\0'; + switch (token->type) + { + case OP_OPEN_COLL_ELEM: + elem->type = COLL_SYM; + break; + case OP_OPEN_EQUIV_CLASS: + elem->type = EQUIV_CLASS; + break; + case OP_OPEN_CHAR_CLASS: + elem->type = CHAR_CLASS; + break; + default: + break; + } + return REG_NOERROR; +} + + /* Helper function for parse_bracket_exp. + Build the equivalence class which is represented by NAME. + The result are written to MBCSET and SBCSET. + EQUIV_CLASS_ALLOC is the allocated size of mbcset->equiv_classes, + is a pointer argument since we may update it. */ + +static reg_errcode_t +#ifdef RE_ENABLE_I18N +build_equiv_class (bitset_t sbcset, re_charset_t *mbcset, + Idx *equiv_class_alloc, const unsigned char *name) +#else /* not RE_ENABLE_I18N */ +build_equiv_class (bitset_t sbcset, const unsigned char *name) +#endif /* not RE_ENABLE_I18N */ +{ +#ifdef _LIBC + uint32_t nrules = _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES); + if (nrules != 0) + { + const int32_t *table, *indirect; + const unsigned char *weights, *extra, *cp; + unsigned char char_buf[2]; + int32_t idx1, idx2; + unsigned int ch; + size_t len; + /* Calculate the index for equivalence class. */ + cp = name; + table = (const int32_t *) _NL_CURRENT (LC_COLLATE, _NL_COLLATE_TABLEMB); + weights = (const unsigned char *) _NL_CURRENT (LC_COLLATE, + _NL_COLLATE_WEIGHTMB); + extra = (const unsigned char *) _NL_CURRENT (LC_COLLATE, + _NL_COLLATE_EXTRAMB); + indirect = (const int32_t *) _NL_CURRENT (LC_COLLATE, + _NL_COLLATE_INDIRECTMB); + idx1 = findidx (table, indirect, extra, &cp, -1); + if (__glibc_unlikely (idx1 == 0 || *cp != '\0')) + /* This isn't a valid character. */ + return REG_ECOLLATE; + + /* Build single byte matching table for this equivalence class. */ + len = weights[idx1 & 0xffffff]; + for (ch = 0; ch < SBC_MAX; ++ch) + { + char_buf[0] = ch; + cp = char_buf; + idx2 = findidx (table, indirect, extra, &cp, 1); +/* + idx2 = table[ch]; +*/ + if (idx2 == 0) + /* This isn't a valid character. */ + continue; + /* Compare only if the length matches and the collation rule + index is the same. */ + if (len == weights[idx2 & 0xffffff] && (idx1 >> 24) == (idx2 >> 24) + && memcmp (weights + (idx1 & 0xffffff) + 1, + weights + (idx2 & 0xffffff) + 1, len) == 0) + bitset_set (sbcset, ch); + } + /* Check whether the array has enough space. */ + if (__glibc_unlikely (*equiv_class_alloc == mbcset->nequiv_classes)) + { + /* Not enough, realloc it. */ + /* +1 in case of mbcset->nequiv_classes is 0. */ + Idx new_equiv_class_alloc = 2 * mbcset->nequiv_classes + 1; + /* Use realloc since the array is NULL if *alloc == 0. */ + int32_t *new_equiv_classes = re_realloc (mbcset->equiv_classes, + int32_t, + new_equiv_class_alloc); + if (__glibc_unlikely (new_equiv_classes == NULL)) + return REG_ESPACE; + mbcset->equiv_classes = new_equiv_classes; + *equiv_class_alloc = new_equiv_class_alloc; + } + mbcset->equiv_classes[mbcset->nequiv_classes++] = idx1; + } + else +#endif /* _LIBC */ + { + if (__glibc_unlikely (strlen ((const char *) name) != 1)) + return REG_ECOLLATE; + bitset_set (sbcset, *name); + } + return REG_NOERROR; +} + + /* Helper function for parse_bracket_exp. + Build the character class which is represented by NAME. + The result are written to MBCSET and SBCSET. + CHAR_CLASS_ALLOC is the allocated size of mbcset->char_classes, + is a pointer argument since we may update it. */ + +static reg_errcode_t +#ifdef RE_ENABLE_I18N +build_charclass (RE_TRANSLATE_TYPE trans, bitset_t sbcset, + re_charset_t *mbcset, Idx *char_class_alloc, + const char *class_name, reg_syntax_t syntax) +#else /* not RE_ENABLE_I18N */ +build_charclass (RE_TRANSLATE_TYPE trans, bitset_t sbcset, + const char *class_name, reg_syntax_t syntax) +#endif /* not RE_ENABLE_I18N */ +{ + int i; + const char *name = class_name; + + /* In case of REG_ICASE "upper" and "lower" match the both of + upper and lower cases. */ + if ((syntax & RE_ICASE) + && (strcmp (name, "upper") == 0 || strcmp (name, "lower") == 0)) + name = "alpha"; + +#ifdef RE_ENABLE_I18N + /* Check the space of the arrays. */ + if (__glibc_unlikely (*char_class_alloc == mbcset->nchar_classes)) + { + /* Not enough, realloc it. */ + /* +1 in case of mbcset->nchar_classes is 0. */ + Idx new_char_class_alloc = 2 * mbcset->nchar_classes + 1; + /* Use realloc since array is NULL if *alloc == 0. */ + wctype_t *new_char_classes = re_realloc (mbcset->char_classes, wctype_t, + new_char_class_alloc); + if (__glibc_unlikely (new_char_classes == NULL)) + return REG_ESPACE; + mbcset->char_classes = new_char_classes; + *char_class_alloc = new_char_class_alloc; + } + mbcset->char_classes[mbcset->nchar_classes++] = __wctype (name); +#endif /* RE_ENABLE_I18N */ + +#define BUILD_CHARCLASS_LOOP(ctype_func) \ + do { \ + if (__glibc_unlikely (trans != NULL)) \ + { \ + for (i = 0; i < SBC_MAX; ++i) \ + if (ctype_func (i)) \ + bitset_set (sbcset, trans[i]); \ + } \ + else \ + { \ + for (i = 0; i < SBC_MAX; ++i) \ + if (ctype_func (i)) \ + bitset_set (sbcset, i); \ + } \ + } while (0) + + if (strcmp (name, "alnum") == 0) + BUILD_CHARCLASS_LOOP (isalnum); + else if (strcmp (name, "cntrl") == 0) + BUILD_CHARCLASS_LOOP (iscntrl); + else if (strcmp (name, "lower") == 0) + BUILD_CHARCLASS_LOOP (islower); + else if (strcmp (name, "space") == 0) + BUILD_CHARCLASS_LOOP (isspace); + else if (strcmp (name, "alpha") == 0) + BUILD_CHARCLASS_LOOP (isalpha); + else if (strcmp (name, "digit") == 0) + BUILD_CHARCLASS_LOOP (isdigit); + else if (strcmp (name, "print") == 0) + BUILD_CHARCLASS_LOOP (isprint); + else if (strcmp (name, "upper") == 0) + BUILD_CHARCLASS_LOOP (isupper); + else if (strcmp (name, "blank") == 0) + BUILD_CHARCLASS_LOOP (isblank); + else if (strcmp (name, "graph") == 0) + BUILD_CHARCLASS_LOOP (isgraph); + else if (strcmp (name, "punct") == 0) + BUILD_CHARCLASS_LOOP (ispunct); + else if (strcmp (name, "xdigit") == 0) + BUILD_CHARCLASS_LOOP (isxdigit); + else + return REG_ECTYPE; + + return REG_NOERROR; +} + +static bin_tree_t * +build_charclass_op (re_dfa_t *dfa, RE_TRANSLATE_TYPE trans, + const char *class_name, + const char *extra, bool non_match, + reg_errcode_t *err) +{ + re_bitset_ptr_t sbcset; +#ifdef RE_ENABLE_I18N + re_charset_t *mbcset; + Idx alloc = 0; +#endif /* not RE_ENABLE_I18N */ + reg_errcode_t ret; + re_token_t br_token; + bin_tree_t *tree; + + sbcset = (re_bitset_ptr_t) calloc (sizeof (bitset_t), 1); + if (__glibc_unlikely (sbcset == NULL)) + { + *err = REG_ESPACE; + return NULL; + } +#ifdef RE_ENABLE_I18N + mbcset = (re_charset_t *) calloc (sizeof (re_charset_t), 1); + if (__glibc_unlikely (mbcset == NULL)) + { + re_free (sbcset); + *err = REG_ESPACE; + return NULL; + } + mbcset->non_match = non_match; +#endif /* RE_ENABLE_I18N */ + + /* We don't care the syntax in this case. */ + ret = build_charclass (trans, sbcset, +#ifdef RE_ENABLE_I18N + mbcset, &alloc, +#endif /* RE_ENABLE_I18N */ + class_name, 0); + + if (__glibc_unlikely (ret != REG_NOERROR)) + { + re_free (sbcset); +#ifdef RE_ENABLE_I18N + free_charset (mbcset); +#endif /* RE_ENABLE_I18N */ + *err = ret; + return NULL; + } + /* \w match '_' also. */ + for (; *extra; extra++) + bitset_set (sbcset, *extra); + + /* If it is non-matching list. */ + if (non_match) + bitset_not (sbcset); + +#ifdef RE_ENABLE_I18N + /* Ensure only single byte characters are set. */ + if (dfa->mb_cur_max > 1) + bitset_mask (sbcset, dfa->sb_char); +#endif + + /* Build a tree for simple bracket. */ +#if defined GCC_LINT || defined lint + memset (&br_token, 0, sizeof br_token); +#endif + br_token.type = SIMPLE_BRACKET; + br_token.opr.sbcset = sbcset; + tree = create_token_tree (dfa, NULL, NULL, &br_token); + if (__glibc_unlikely (tree == NULL)) + goto build_word_op_espace; + +#ifdef RE_ENABLE_I18N + if (dfa->mb_cur_max > 1) + { + bin_tree_t *mbc_tree; + /* Build a tree for complex bracket. */ + br_token.type = COMPLEX_BRACKET; + br_token.opr.mbcset = mbcset; + dfa->has_mb_node = 1; + mbc_tree = create_token_tree (dfa, NULL, NULL, &br_token); + if (__glibc_unlikely (mbc_tree == NULL)) + goto build_word_op_espace; + /* Then join them by ALT node. */ + tree = create_tree (dfa, tree, mbc_tree, OP_ALT); + if (__glibc_likely (mbc_tree != NULL)) + return tree; + } + else + { + free_charset (mbcset); + return tree; + } +#else /* not RE_ENABLE_I18N */ + return tree; +#endif /* not RE_ENABLE_I18N */ + + build_word_op_espace: + re_free (sbcset); +#ifdef RE_ENABLE_I18N + free_charset (mbcset); +#endif /* RE_ENABLE_I18N */ + *err = REG_ESPACE; + return NULL; +} + +/* This is intended for the expressions like "a{1,3}". + Fetch a number from 'input', and return the number. + Return -1 if the number field is empty like "{,1}". + Return RE_DUP_MAX + 1 if the number field is too large. + Return -2 if an error occurred. */ + +static Idx +fetch_number (re_string_t *input, re_token_t *token, reg_syntax_t syntax) +{ + Idx num = -1; + unsigned char c; + while (1) + { + fetch_token (token, input, syntax); + c = token->opr.c; + if (__glibc_unlikely (token->type == END_OF_RE)) + return -2; + if (token->type == OP_CLOSE_DUP_NUM || c == ',') + break; + num = ((token->type != CHARACTER || c < '0' || '9' < c || num == -2) + ? -2 + : num == -1 + ? c - '0' + : MIN (RE_DUP_MAX + 1, num * 10 + c - '0')); + } + return num; +} + +#ifdef RE_ENABLE_I18N +static void +free_charset (re_charset_t *cset) +{ + re_free (cset->mbchars); +# ifdef _LIBC + re_free (cset->coll_syms); + re_free (cset->equiv_classes); +# endif + re_free (cset->range_starts); + re_free (cset->range_ends); + re_free (cset->char_classes); + re_free (cset); +} +#endif /* RE_ENABLE_I18N */ + +/* Functions for binary tree operation. */ + +/* Create a tree node. */ + +static bin_tree_t * +create_tree (re_dfa_t *dfa, bin_tree_t *left, bin_tree_t *right, + re_token_type_t type) +{ + re_token_t t; +#if defined GCC_LINT || defined lint + memset (&t, 0, sizeof t); +#endif + t.type = type; + return create_token_tree (dfa, left, right, &t); +} + +static bin_tree_t * +create_token_tree (re_dfa_t *dfa, bin_tree_t *left, bin_tree_t *right, + const re_token_t *token) +{ + bin_tree_t *tree; + if (__glibc_unlikely (dfa->str_tree_storage_idx == BIN_TREE_STORAGE_SIZE)) + { + bin_tree_storage_t *storage = re_malloc (bin_tree_storage_t, 1); + + if (storage == NULL) + return NULL; + storage->next = dfa->str_tree_storage; + dfa->str_tree_storage = storage; + dfa->str_tree_storage_idx = 0; + } + tree = &dfa->str_tree_storage->data[dfa->str_tree_storage_idx++]; + + tree->parent = NULL; + tree->left = left; + tree->right = right; + tree->token = *token; + tree->token.duplicated = 0; + tree->token.opt_subexp = 0; + tree->first = NULL; + tree->next = NULL; + tree->node_idx = -1; + + if (left != NULL) + left->parent = tree; + if (right != NULL) + right->parent = tree; + return tree; +} + +/* Mark the tree SRC as an optional subexpression. + To be called from preorder or postorder. */ + +static reg_errcode_t +mark_opt_subexp (void *extra, bin_tree_t *node) +{ + Idx idx = (uintptr_t) extra; + if (node->token.type == SUBEXP && node->token.opr.idx == idx) + node->token.opt_subexp = 1; + + return REG_NOERROR; +} + +/* Free the allocated memory inside NODE. */ + +static void +free_token (re_token_t *node) +{ +#ifdef RE_ENABLE_I18N + if (node->type == COMPLEX_BRACKET && node->duplicated == 0) + free_charset (node->opr.mbcset); + else +#endif /* RE_ENABLE_I18N */ + if (node->type == SIMPLE_BRACKET && node->duplicated == 0) + re_free (node->opr.sbcset); +} + +/* Worker function for tree walking. Free the allocated memory inside NODE + and its children. */ + +static reg_errcode_t +free_tree (void *extra, bin_tree_t *node) +{ + free_token (&node->token); + return REG_NOERROR; +} + + +/* Duplicate the node SRC, and return new node. This is a preorder + visit similar to the one implemented by the generic visitor, but + we need more infrastructure to maintain two parallel trees --- so, + it's easier to duplicate. */ + +static bin_tree_t * +duplicate_tree (const bin_tree_t *root, re_dfa_t *dfa) +{ + const bin_tree_t *node; + bin_tree_t *dup_root; + bin_tree_t **p_new = &dup_root, *dup_node = root->parent; + + for (node = root; ; ) + { + /* Create a new tree and link it back to the current parent. */ + *p_new = create_token_tree (dfa, NULL, NULL, &node->token); + if (*p_new == NULL) + return NULL; + (*p_new)->parent = dup_node; + (*p_new)->token.duplicated = 1; + dup_node = *p_new; + + /* Go to the left node, or up and to the right. */ + if (node->left) + { + node = node->left; + p_new = &dup_node->left; + } + else + { + const bin_tree_t *prev = NULL; + while (node->right == prev || node->right == NULL) + { + prev = node; + node = node->parent; + dup_node = dup_node->parent; + if (!node) + return dup_root; + } + node = node->right; + p_new = &dup_node->right; + } + } +} diff --git a/lib/regex.c b/lib/regex.c new file mode 100644 index 0000000..2a86e10 --- /dev/null +++ b/lib/regex.c @@ -0,0 +1,81 @@ +/* Extended regular expression matching and search library. + Copyright (C) 2002-2018 Free Software Foundation, Inc. + This file is part of the GNU C Library. + Contributed by Isamu Hasegawa <isamu@yamato.ibm.com>. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU General Public + License as published by the Free Software Foundation; either + version 3 of the License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + General Public License for more details. + + You should have received a copy of the GNU General Public + License along with the GNU C Library; if not, see + <https://www.gnu.org/licenses/>. */ + +#ifndef _LIBC +# include <libc-config.h> + +# if (__GNUC__ == 4 && 6 <= __GNUC_MINOR__) || 4 < __GNUC__ +# pragma GCC diagnostic ignored "-Wsuggest-attribute=pure" +# endif +# if (__GNUC__ == 4 && 3 <= __GNUC_MINOR__) || 4 < __GNUC__ +# pragma GCC diagnostic ignored "-Wold-style-definition" +# pragma GCC diagnostic ignored "-Wtype-limits" +# endif +#endif + +/* Make sure no one compiles this code with a C++ compiler. */ +#if defined __cplusplus && defined _LIBC +# error "This is C code, use a C compiler" +#endif + +#ifdef _LIBC +/* We have to keep the namespace clean. */ +# define regfree(preg) __regfree (preg) +# define regexec(pr, st, nm, pm, ef) __regexec (pr, st, nm, pm, ef) +# define regcomp(preg, pattern, cflags) __regcomp (preg, pattern, cflags) +# define regerror(errcode, preg, errbuf, errbuf_size) \ + __regerror(errcode, preg, errbuf, errbuf_size) +# define re_set_registers(bu, re, nu, st, en) \ + __re_set_registers (bu, re, nu, st, en) +# define re_match_2(bufp, string1, size1, string2, size2, pos, regs, stop) \ + __re_match_2 (bufp, string1, size1, string2, size2, pos, regs, stop) +# define re_match(bufp, string, size, pos, regs) \ + __re_match (bufp, string, size, pos, regs) +# define re_search(bufp, string, size, startpos, range, regs) \ + __re_search (bufp, string, size, startpos, range, regs) +# define re_compile_pattern(pattern, length, bufp) \ + __re_compile_pattern (pattern, length, bufp) +# define re_set_syntax(syntax) __re_set_syntax (syntax) +# define re_search_2(bufp, st1, s1, st2, s2, startpos, range, regs, stop) \ + __re_search_2 (bufp, st1, s1, st2, s2, startpos, range, regs, stop) +# define re_compile_fastmap(bufp) __re_compile_fastmap (bufp) + +# include "../locale/localeinfo.h" +#endif + +/* On some systems, limits.h sets RE_DUP_MAX to a lower value than + GNU regex allows. Include it before <regex.h>, which correctly + #undefs RE_DUP_MAX and sets it to the right value. */ +#include <limits.h> + +#include <regex.h> +#include "regex_internal.h" + +#include "regex_internal.c" +#include "regcomp.c" +#include "regexec.c" + +/* Binary backward compatibility. */ +#if _LIBC +# include <shlib-compat.h> +# if SHLIB_COMPAT (libc, GLIBC_2_0, GLIBC_2_3) +link_warning (re_max_failures, "the 're_max_failures' variable is obsolete and will go away.") +int re_max_failures = 2000; +# endif +#endif diff --git a/lib/regex.h b/lib/regex.h new file mode 100644 index 0000000..f2ac950 --- /dev/null +++ b/lib/regex.h @@ -0,0 +1,658 @@ +/* Definitions for data structures and routines for the regular + expression library. + Copyright (C) 1985, 1989-2018 Free Software Foundation, Inc. + This file is part of the GNU C Library. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU General Public + License as published by the Free Software Foundation; either + version 3 of the License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + General Public License for more details. + + You should have received a copy of the GNU General Public + License along with the GNU C Library; if not, see + <https://www.gnu.org/licenses/>. */ + +#ifndef _REGEX_H +#define _REGEX_H 1 + +#include <sys/types.h> + +/* Allow the use in C++ code. */ +#ifdef __cplusplus +extern "C" { +#endif + +/* Define __USE_GNU to declare GNU extensions that violate the + POSIX name space rules. */ +#ifdef _GNU_SOURCE +# define __USE_GNU 1 +#endif + +#ifdef _REGEX_LARGE_OFFSETS + +/* Use types and values that are wide enough to represent signed and + unsigned byte offsets in memory. This currently works only when + the regex code is used outside of the GNU C library; it is not yet + supported within glibc itself, and glibc users should not define + _REGEX_LARGE_OFFSETS. */ + +/* The type of object sizes. */ +typedef size_t __re_size_t; + +/* The type of object sizes, in places where the traditional code + uses unsigned long int. */ +typedef size_t __re_long_size_t; + +#else + +/* The traditional GNU regex implementation mishandles strings longer + than INT_MAX. */ +typedef unsigned int __re_size_t; +typedef unsigned long int __re_long_size_t; + +#endif + +/* The following two types have to be signed and unsigned integer type + wide enough to hold a value of a pointer. For most ANSI compilers + ptrdiff_t and size_t should be likely OK. Still size of these two + types is 2 for Microsoft C. Ugh... */ +typedef long int s_reg_t; +typedef unsigned long int active_reg_t; + +/* The following bits are used to determine the regexp syntax we + recognize. The set/not-set meanings are chosen so that Emacs syntax + remains the value 0. The bits are given in alphabetical order, and + the definitions shifted by one from the previous bit; thus, when we + add or remove a bit, only one other definition need change. */ +typedef unsigned long int reg_syntax_t; + +#ifdef __USE_GNU +/* If this bit is not set, then \ inside a bracket expression is literal. + If set, then such a \ quotes the following character. */ +# define RE_BACKSLASH_ESCAPE_IN_LISTS ((unsigned long int) 1) + +/* If this bit is not set, then + and ? are operators, and \+ and \? are + literals. + If set, then \+ and \? are operators and + and ? are literals. */ +# define RE_BK_PLUS_QM (RE_BACKSLASH_ESCAPE_IN_LISTS << 1) + +/* If this bit is set, then character classes are supported. They are: + [:alpha:], [:upper:], [:lower:], [:digit:], [:alnum:], [:xdigit:], + [:space:], [:print:], [:punct:], [:graph:], and [:cntrl:]. + If not set, then character classes are not supported. */ +# define RE_CHAR_CLASSES (RE_BK_PLUS_QM << 1) + +/* If this bit is set, then ^ and $ are always anchors (outside bracket + expressions, of course). + If this bit is not set, then it depends: + ^ is an anchor if it is at the beginning of a regular + expression or after an open-group or an alternation operator; + $ is an anchor if it is at the end of a regular expression, or + before a close-group or an alternation operator. + + This bit could be (re)combined with RE_CONTEXT_INDEP_OPS, because + POSIX draft 11.2 says that * etc. in leading positions is undefined. + We already implemented a previous draft which made those constructs + invalid, though, so we haven't changed the code back. */ +# define RE_CONTEXT_INDEP_ANCHORS (RE_CHAR_CLASSES << 1) + +/* If this bit is set, then special characters are always special + regardless of where they are in the pattern. + If this bit is not set, then special characters are special only in + some contexts; otherwise they are ordinary. Specifically, + * + ? and intervals are only special when not after the beginning, + open-group, or alternation operator. */ +# define RE_CONTEXT_INDEP_OPS (RE_CONTEXT_INDEP_ANCHORS << 1) + +/* If this bit is set, then *, +, ?, and { cannot be first in an re or + immediately after an alternation or begin-group operator. */ +# define RE_CONTEXT_INVALID_OPS (RE_CONTEXT_INDEP_OPS << 1) + +/* If this bit is set, then . matches newline. + If not set, then it doesn't. */ +# define RE_DOT_NEWLINE (RE_CONTEXT_INVALID_OPS << 1) + +/* If this bit is set, then . doesn't match NUL. + If not set, then it does. */ +# define RE_DOT_NOT_NULL (RE_DOT_NEWLINE << 1) + +/* If this bit is set, nonmatching lists [^...] do not match newline. + If not set, they do. */ +# define RE_HAT_LISTS_NOT_NEWLINE (RE_DOT_NOT_NULL << 1) + +/* If this bit is set, either \{...\} or {...} defines an + interval, depending on RE_NO_BK_BRACES. + If not set, \{, \}, {, and } are literals. */ +# define RE_INTERVALS (RE_HAT_LISTS_NOT_NEWLINE << 1) + +/* If this bit is set, +, ? and | aren't recognized as operators. + If not set, they are. */ +# define RE_LIMITED_OPS (RE_INTERVALS << 1) + +/* If this bit is set, newline is an alternation operator. + If not set, newline is literal. */ +# define RE_NEWLINE_ALT (RE_LIMITED_OPS << 1) + +/* If this bit is set, then '{...}' defines an interval, and \{ and \} + are literals. + If not set, then '\{...\}' defines an interval. */ +# define RE_NO_BK_BRACES (RE_NEWLINE_ALT << 1) + +/* If this bit is set, (...) defines a group, and \( and \) are literals. + If not set, \(...\) defines a group, and ( and ) are literals. */ +# define RE_NO_BK_PARENS (RE_NO_BK_BRACES << 1) + +/* If this bit is set, then \<digit> matches <digit>. + If not set, then \<digit> is a back-reference. */ +# define RE_NO_BK_REFS (RE_NO_BK_PARENS << 1) + +/* If this bit is set, then | is an alternation operator, and \| is literal. + If not set, then \| is an alternation operator, and | is literal. */ +# define RE_NO_BK_VBAR (RE_NO_BK_REFS << 1) + +/* If this bit is set, then an ending range point collating higher + than the starting range point, as in [z-a], is invalid. + If not set, then when ending range point collates higher than the + starting range point, the range is ignored. */ +# define RE_NO_EMPTY_RANGES (RE_NO_BK_VBAR << 1) + +/* If this bit is set, then an unmatched ) is ordinary. + If not set, then an unmatched ) is invalid. */ +# define RE_UNMATCHED_RIGHT_PAREN_ORD (RE_NO_EMPTY_RANGES << 1) + +/* If this bit is set, succeed as soon as we match the whole pattern, + without further backtracking. */ +# define RE_NO_POSIX_BACKTRACKING (RE_UNMATCHED_RIGHT_PAREN_ORD << 1) + +/* If this bit is set, do not process the GNU regex operators. + If not set, then the GNU regex operators are recognized. */ +# define RE_NO_GNU_OPS (RE_NO_POSIX_BACKTRACKING << 1) + +/* If this bit is set, turn on internal regex debugging. + If not set, and debugging was on, turn it off. + This only works if regex.c is compiled -DDEBUG. + We define this bit always, so that all that's needed to turn on + debugging is to recompile regex.c; the calling code can always have + this bit set, and it won't affect anything in the normal case. */ +# define RE_DEBUG (RE_NO_GNU_OPS << 1) + +/* If this bit is set, a syntactically invalid interval is treated as + a string of ordinary characters. For example, the ERE 'a{1' is + treated as 'a\{1'. */ +# define RE_INVALID_INTERVAL_ORD (RE_DEBUG << 1) + +/* If this bit is set, then ignore case when matching. + If not set, then case is significant. */ +# define RE_ICASE (RE_INVALID_INTERVAL_ORD << 1) + +/* This bit is used internally like RE_CONTEXT_INDEP_ANCHORS but only + for ^, because it is difficult to scan the regex backwards to find + whether ^ should be special. */ +# define RE_CARET_ANCHORS_HERE (RE_ICASE << 1) + +/* If this bit is set, then \{ cannot be first in a regex or + immediately after an alternation, open-group or \} operator. */ +# define RE_CONTEXT_INVALID_DUP (RE_CARET_ANCHORS_HERE << 1) + +/* If this bit is set, then no_sub will be set to 1 during + re_compile_pattern. */ +# define RE_NO_SUB (RE_CONTEXT_INVALID_DUP << 1) +#endif + +/* This global variable defines the particular regexp syntax to use (for + some interfaces). When a regexp is compiled, the syntax used is + stored in the pattern buffer, so changing this does not affect + already-compiled regexps. */ +extern reg_syntax_t re_syntax_options; + +#ifdef __USE_GNU +/* Define combinations of the above bits for the standard possibilities. + (The [[[ comments delimit what gets put into the Texinfo file, so + don't delete them!) */ +/* [[[begin syntaxes]]] */ +# define RE_SYNTAX_EMACS 0 + +# define RE_SYNTAX_AWK \ + (RE_BACKSLASH_ESCAPE_IN_LISTS | RE_DOT_NOT_NULL \ + | RE_NO_BK_PARENS | RE_NO_BK_REFS \ + | RE_NO_BK_VBAR | RE_NO_EMPTY_RANGES \ + | RE_DOT_NEWLINE | RE_CONTEXT_INDEP_ANCHORS \ + | RE_CHAR_CLASSES \ + | RE_UNMATCHED_RIGHT_PAREN_ORD | RE_NO_GNU_OPS) + +# define RE_SYNTAX_GNU_AWK \ + ((RE_SYNTAX_POSIX_EXTENDED | RE_BACKSLASH_ESCAPE_IN_LISTS \ + | RE_INVALID_INTERVAL_ORD) \ + & ~(RE_DOT_NOT_NULL | RE_CONTEXT_INDEP_OPS \ + | RE_CONTEXT_INVALID_OPS )) + +# define RE_SYNTAX_POSIX_AWK \ + (RE_SYNTAX_POSIX_EXTENDED | RE_BACKSLASH_ESCAPE_IN_LISTS \ + | RE_INTERVALS | RE_NO_GNU_OPS \ + | RE_INVALID_INTERVAL_ORD) + +# define RE_SYNTAX_GREP \ + ((RE_SYNTAX_POSIX_BASIC | RE_NEWLINE_ALT) \ + & ~(RE_CONTEXT_INVALID_DUP | RE_DOT_NOT_NULL)) + +# define RE_SYNTAX_EGREP \ + ((RE_SYNTAX_POSIX_EXTENDED | RE_INVALID_INTERVAL_ORD | RE_NEWLINE_ALT) \ + & ~(RE_CONTEXT_INVALID_OPS | RE_DOT_NOT_NULL)) + +/* POSIX grep -E behavior is no longer incompatible with GNU. */ +# define RE_SYNTAX_POSIX_EGREP \ + RE_SYNTAX_EGREP + +/* P1003.2/D11.2, section 4.20.7.1, lines 5078ff. */ +# define RE_SYNTAX_ED RE_SYNTAX_POSIX_BASIC + +# define RE_SYNTAX_SED RE_SYNTAX_POSIX_BASIC + +/* Syntax bits common to both basic and extended POSIX regex syntax. */ +# define _RE_SYNTAX_POSIX_COMMON \ + (RE_CHAR_CLASSES | RE_DOT_NEWLINE | RE_DOT_NOT_NULL \ + | RE_INTERVALS | RE_NO_EMPTY_RANGES) + +# define RE_SYNTAX_POSIX_BASIC \ + (_RE_SYNTAX_POSIX_COMMON | RE_BK_PLUS_QM | RE_CONTEXT_INVALID_DUP) + +/* Differs from ..._POSIX_BASIC only in that RE_BK_PLUS_QM becomes + RE_LIMITED_OPS, i.e., \? \+ \| are not recognized. Actually, this + isn't minimal, since other operators, such as \`, aren't disabled. */ +# define RE_SYNTAX_POSIX_MINIMAL_BASIC \ + (_RE_SYNTAX_POSIX_COMMON | RE_LIMITED_OPS) + +# define RE_SYNTAX_POSIX_EXTENDED \ + (_RE_SYNTAX_POSIX_COMMON | RE_CONTEXT_INDEP_ANCHORS \ + | RE_CONTEXT_INDEP_OPS | RE_NO_BK_BRACES \ + | RE_NO_BK_PARENS | RE_NO_BK_VBAR \ + | RE_CONTEXT_INVALID_OPS | RE_UNMATCHED_RIGHT_PAREN_ORD) + +/* Differs from ..._POSIX_EXTENDED in that RE_CONTEXT_INDEP_OPS is + removed and RE_NO_BK_REFS is added. */ +# define RE_SYNTAX_POSIX_MINIMAL_EXTENDED \ + (_RE_SYNTAX_POSIX_COMMON | RE_CONTEXT_INDEP_ANCHORS \ + | RE_CONTEXT_INVALID_OPS | RE_NO_BK_BRACES \ + | RE_NO_BK_PARENS | RE_NO_BK_REFS \ + | RE_NO_BK_VBAR | RE_UNMATCHED_RIGHT_PAREN_ORD) +/* [[[end syntaxes]]] */ + +/* Maximum number of duplicates an interval can allow. POSIX-conforming + systems might define this in <limits.h>, but we want our + value, so remove any previous define. */ +# ifdef _REGEX_INCLUDE_LIMITS_H +# include <limits.h> +# endif +# ifdef RE_DUP_MAX +# undef RE_DUP_MAX +# endif + +/* RE_DUP_MAX is 2**15 - 1 because an earlier implementation stored + the counter as a 2-byte signed integer. This is no longer true, so + RE_DUP_MAX could be increased to (INT_MAX / 10 - 1), or to + ((SIZE_MAX - 9) / 10) if _REGEX_LARGE_OFFSETS is defined. + However, there would be a huge performance problem if someone + actually used a pattern like a\{214748363\}, so RE_DUP_MAX retains + its historical value. */ +# define RE_DUP_MAX (0x7fff) +#endif + + +/* POSIX 'cflags' bits (i.e., information for 'regcomp'). */ + +/* If this bit is set, then use extended regular expression syntax. + If not set, then use basic regular expression syntax. */ +#define REG_EXTENDED 1 + +/* If this bit is set, then ignore case when matching. + If not set, then case is significant. */ +#define REG_ICASE (1 << 1) + +/* If this bit is set, then anchors do not match at newline + characters in the string. + If not set, then anchors do match at newlines. */ +#define REG_NEWLINE (1 << 2) + +/* If this bit is set, then report only success or fail in regexec. + If not set, then returns differ between not matching and errors. */ +#define REG_NOSUB (1 << 3) + + +/* POSIX 'eflags' bits (i.e., information for regexec). */ + +/* If this bit is set, then the beginning-of-line operator doesn't match + the beginning of the string (presumably because it's not the + beginning of a line). + If not set, then the beginning-of-line operator does match the + beginning of the string. */ +#define REG_NOTBOL 1 + +/* Like REG_NOTBOL, except for the end-of-line. */ +#define REG_NOTEOL (1 << 1) + +/* Use PMATCH[0] to delimit the start and end of the search in the + buffer. */ +#define REG_STARTEND (1 << 2) + + +/* If any error codes are removed, changed, or added, update the + '__re_error_msgid' table in regcomp.c. */ + +typedef enum +{ + _REG_ENOSYS = -1, /* This will never happen for this implementation. */ + _REG_NOERROR = 0, /* Success. */ + _REG_NOMATCH, /* Didn't find a match (for regexec). */ + + /* POSIX regcomp return error codes. (In the order listed in the + standard.) */ + _REG_BADPAT, /* Invalid pattern. */ + _REG_ECOLLATE, /* Invalid collating element. */ + _REG_ECTYPE, /* Invalid character class name. */ + _REG_EESCAPE, /* Trailing backslash. */ + _REG_ESUBREG, /* Invalid back reference. */ + _REG_EBRACK, /* Unmatched left bracket. */ + _REG_EPAREN, /* Parenthesis imbalance. */ + _REG_EBRACE, /* Unmatched \{. */ + _REG_BADBR, /* Invalid contents of \{\}. */ + _REG_ERANGE, /* Invalid range end. */ + _REG_ESPACE, /* Ran out of memory. */ + _REG_BADRPT, /* No preceding re for repetition op. */ + + /* Error codes we've added. */ + _REG_EEND, /* Premature end. */ + _REG_ESIZE, /* Too large (e.g., repeat count too large). */ + _REG_ERPAREN /* Unmatched ) or \); not returned from regcomp. */ +} reg_errcode_t; + +#if defined _XOPEN_SOURCE || defined __USE_XOPEN2K +# define REG_ENOSYS _REG_ENOSYS +#endif +#define REG_NOERROR _REG_NOERROR +#define REG_NOMATCH _REG_NOMATCH +#define REG_BADPAT _REG_BADPAT +#define REG_ECOLLATE _REG_ECOLLATE +#define REG_ECTYPE _REG_ECTYPE +#define REG_EESCAPE _REG_EESCAPE +#define REG_ESUBREG _REG_ESUBREG +#define REG_EBRACK _REG_EBRACK +#define REG_EPAREN _REG_EPAREN +#define REG_EBRACE _REG_EBRACE +#define REG_BADBR _REG_BADBR +#define REG_ERANGE _REG_ERANGE +#define REG_ESPACE _REG_ESPACE +#define REG_BADRPT _REG_BADRPT +#define REG_EEND _REG_EEND +#define REG_ESIZE _REG_ESIZE +#define REG_ERPAREN _REG_ERPAREN + +/* This data structure represents a compiled pattern. Before calling + the pattern compiler, the fields 'buffer', 'allocated', 'fastmap', + and 'translate' can be set. After the pattern has been compiled, + the fields 're_nsub', 'not_bol' and 'not_eol' are available. All + other fields are private to the regex routines. */ + +#ifndef RE_TRANSLATE_TYPE +# define __RE_TRANSLATE_TYPE unsigned char * +# ifdef __USE_GNU +# define RE_TRANSLATE_TYPE __RE_TRANSLATE_TYPE +# endif +#endif + +#ifdef __USE_GNU +# define __REPB_PREFIX(name) name +#else +# define __REPB_PREFIX(name) __##name +#endif + +struct re_pattern_buffer +{ + /* Space that holds the compiled pattern. The type + 'struct re_dfa_t' is private and is not declared here. */ + struct re_dfa_t *__REPB_PREFIX(buffer); + + /* Number of bytes to which 'buffer' points. */ + __re_long_size_t __REPB_PREFIX(allocated); + + /* Number of bytes actually used in 'buffer'. */ + __re_long_size_t __REPB_PREFIX(used); + + /* Syntax setting with which the pattern was compiled. */ + reg_syntax_t __REPB_PREFIX(syntax); + + /* Pointer to a fastmap, if any, otherwise zero. re_search uses the + fastmap, if there is one, to skip over impossible starting points + for matches. */ + char *__REPB_PREFIX(fastmap); + + /* Either a translate table to apply to all characters before + comparing them, or zero for no translation. The translation is + applied to a pattern when it is compiled and to a string when it + is matched. */ + __RE_TRANSLATE_TYPE __REPB_PREFIX(translate); + + /* Number of subexpressions found by the compiler. */ + size_t re_nsub; + + /* Zero if this pattern cannot match the empty string, one else. + Well, in truth it's used only in 're_search_2', to see whether or + not we should use the fastmap, so we don't set this absolutely + perfectly; see 're_compile_fastmap' (the "duplicate" case). */ + unsigned __REPB_PREFIX(can_be_null) : 1; + + /* If REGS_UNALLOCATED, allocate space in the 'regs' structure + for 'max (RE_NREGS, re_nsub + 1)' groups. + If REGS_REALLOCATE, reallocate space if necessary. + If REGS_FIXED, use what's there. */ +#ifdef __USE_GNU +# define REGS_UNALLOCATED 0 +# define REGS_REALLOCATE 1 +# define REGS_FIXED 2 +#endif + unsigned __REPB_PREFIX(regs_allocated) : 2; + + /* Set to zero when 're_compile_pattern' compiles a pattern; set to + one by 're_compile_fastmap' if it updates the fastmap. */ + unsigned __REPB_PREFIX(fastmap_accurate) : 1; + + /* If set, 're_match_2' does not return information about + subexpressions. */ + unsigned __REPB_PREFIX(no_sub) : 1; + + /* If set, a beginning-of-line anchor doesn't match at the beginning + of the string. */ + unsigned __REPB_PREFIX(not_bol) : 1; + + /* Similarly for an end-of-line anchor. */ + unsigned __REPB_PREFIX(not_eol) : 1; + + /* If true, an anchor at a newline matches. */ + unsigned __REPB_PREFIX(newline_anchor) : 1; +}; + +typedef struct re_pattern_buffer regex_t; + +/* Type for byte offsets within the string. POSIX mandates this. */ +#ifdef _REGEX_LARGE_OFFSETS +/* POSIX 1003.1-2008 requires that regoff_t be at least as wide as + ptrdiff_t and ssize_t. We don't know of any hosts where ptrdiff_t + is wider than ssize_t, so ssize_t is safe. ptrdiff_t is not + visible here, so use ssize_t. */ +typedef ssize_t regoff_t; +#else +/* The traditional GNU regex implementation mishandles strings longer + than INT_MAX. */ +typedef int regoff_t; +#endif + + +#ifdef __USE_GNU +/* This is the structure we store register match data in. See + regex.texinfo for a full description of what registers match. */ +struct re_registers +{ + __re_size_t num_regs; + regoff_t *start; + regoff_t *end; +}; + + +/* If 'regs_allocated' is REGS_UNALLOCATED in the pattern buffer, + 're_match_2' returns information about at least this many registers + the first time a 'regs' structure is passed. */ +# ifndef RE_NREGS +# define RE_NREGS 30 +# endif +#endif + + +/* POSIX specification for registers. Aside from the different names than + 're_registers', POSIX uses an array of structures, instead of a + structure of arrays. */ +typedef struct +{ + regoff_t rm_so; /* Byte offset from string's start to substring's start. */ + regoff_t rm_eo; /* Byte offset from string's start to substring's end. */ +} regmatch_t; + +/* Declarations for routines. */ + +#ifdef __USE_GNU +/* Sets the current default syntax to SYNTAX, and return the old syntax. + You can also simply assign to the 're_syntax_options' variable. */ +extern reg_syntax_t re_set_syntax (reg_syntax_t __syntax); + +/* Compile the regular expression PATTERN, with length LENGTH + and syntax given by the global 're_syntax_options', into the buffer + BUFFER. Return NULL if successful, and an error string if not. + + To free the allocated storage, you must call 'regfree' on BUFFER. + Note that the translate table must either have been initialized by + 'regcomp', with a malloc'ed value, or set to NULL before calling + 'regfree'. */ +extern const char *re_compile_pattern (const char *__pattern, size_t __length, + struct re_pattern_buffer *__buffer); + + +/* Compile a fastmap for the compiled pattern in BUFFER; used to + accelerate searches. Return 0 if successful and -2 if was an + internal error. */ +extern int re_compile_fastmap (struct re_pattern_buffer *__buffer); + + +/* Search in the string STRING (with length LENGTH) for the pattern + compiled into BUFFER. Start searching at position START, for RANGE + characters. Return the starting position of the match, -1 for no + match, or -2 for an internal error. Also return register + information in REGS (if REGS and BUFFER->no_sub are nonzero). */ +extern regoff_t re_search (struct re_pattern_buffer *__buffer, + const char *__String, regoff_t __length, + regoff_t __start, regoff_t __range, + struct re_registers *__regs); + + +/* Like 're_search', but search in the concatenation of STRING1 and + STRING2. Also, stop searching at index START + STOP. */ +extern regoff_t re_search_2 (struct re_pattern_buffer *__buffer, + const char *__string1, regoff_t __length1, + const char *__string2, regoff_t __length2, + regoff_t __start, regoff_t __range, + struct re_registers *__regs, + regoff_t __stop); + + +/* Like 're_search', but return how many characters in STRING the regexp + in BUFFER matched, starting at position START. */ +extern regoff_t re_match (struct re_pattern_buffer *__buffer, + const char *__String, regoff_t __length, + regoff_t __start, struct re_registers *__regs); + + +/* Relates to 're_match' as 're_search_2' relates to 're_search'. */ +extern regoff_t re_match_2 (struct re_pattern_buffer *__buffer, + const char *__string1, regoff_t __length1, + const char *__string2, regoff_t __length2, + regoff_t __start, struct re_registers *__regs, + regoff_t __stop); + + +/* Set REGS to hold NUM_REGS registers, storing them in STARTS and + ENDS. Subsequent matches using BUFFER and REGS will use this memory + for recording register information. STARTS and ENDS must be + allocated with malloc, and must each be at least 'NUM_REGS * sizeof + (regoff_t)' bytes long. + + If NUM_REGS == 0, then subsequent matches should allocate their own + register data. + + Unless this function is called, the first search or match using + BUFFER will allocate its own register data, without + freeing the old data. */ +extern void re_set_registers (struct re_pattern_buffer *__buffer, + struct re_registers *__regs, + __re_size_t __num_regs, + regoff_t *__starts, regoff_t *__ends); +#endif /* Use GNU */ + +#if defined _REGEX_RE_COMP || (defined _LIBC && defined __USE_MISC) +# ifndef _CRAY +/* 4.2 bsd compatibility. */ +extern char *re_comp (const char *); +extern int re_exec (const char *); +# endif +#endif + +/* For plain 'restrict', use glibc's __restrict if defined. + Otherwise, GCC 2.95 and later have "__restrict"; C99 compilers have + "restrict", and "configure" may have defined "restrict". + Other compilers use __restrict, __restrict__, and _Restrict, and + 'configure' might #define 'restrict' to those words, so pick a + different name. */ +#ifndef _Restrict_ +# if defined __restrict || 2 < __GNUC__ + (95 <= __GNUC_MINOR__) +# define _Restrict_ __restrict +# elif 199901L <= __STDC_VERSION__ || defined restrict +# define _Restrict_ restrict +# else +# define _Restrict_ +# endif +#endif +/* For [restrict], use glibc's __restrict_arr if available. + Otherwise, GCC 3.1 (not in C++ mode) and C99 support [restrict]. */ +#ifndef _Restrict_arr_ +# ifdef __restrict_arr +# define _Restrict_arr_ __restrict_arr +# elif ((199901L <= __STDC_VERSION__ || 3 < __GNUC__ + (1 <= __GNUC_MINOR__)) \ + && !defined __GNUG__) +# define _Restrict_arr_ _Restrict_ +# else +# define _Restrict_arr_ +# endif +#endif + +/* POSIX compatibility. */ +extern int regcomp (regex_t *_Restrict_ __preg, + const char *_Restrict_ __pattern, + int __cflags); + +extern int regexec (const regex_t *_Restrict_ __preg, + const char *_Restrict_ __String, size_t __nmatch, + regmatch_t __pmatch[_Restrict_arr_], + int __eflags); + +extern size_t regerror (int __errcode, const regex_t *_Restrict_ __preg, + char *_Restrict_ __errbuf, size_t __errbuf_size); + +extern void regfree (regex_t *__preg); + + +#ifdef __cplusplus +} +#endif /* C++ */ + +#endif /* regex.h */ diff --git a/lib/regex_internal.c b/lib/regex_internal.c new file mode 100644 index 0000000..f13def3 --- /dev/null +++ b/lib/regex_internal.c @@ -0,0 +1,1746 @@ +/* Extended regular expression matching and search library. + Copyright (C) 2002-2018 Free Software Foundation, Inc. + This file is part of the GNU C Library. + Contributed by Isamu Hasegawa <isamu@yamato.ibm.com>. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU General Public + License as published by the Free Software Foundation; either + version 3 of the License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + General Public License for more details. + + You should have received a copy of the GNU General Public + License along with the GNU C Library; if not, see + <https://www.gnu.org/licenses/>. */ + +static void re_string_construct_common (const char *str, Idx len, + re_string_t *pstr, + RE_TRANSLATE_TYPE trans, bool icase, + const re_dfa_t *dfa); +static re_dfastate_t *create_ci_newstate (const re_dfa_t *dfa, + const re_node_set *nodes, + re_hashval_t hash); +static re_dfastate_t *create_cd_newstate (const re_dfa_t *dfa, + const re_node_set *nodes, + unsigned int context, + re_hashval_t hash); +static reg_errcode_t re_string_realloc_buffers (re_string_t *pstr, + Idx new_buf_len); +#ifdef RE_ENABLE_I18N +static void build_wcs_buffer (re_string_t *pstr); +static reg_errcode_t build_wcs_upper_buffer (re_string_t *pstr); +#endif /* RE_ENABLE_I18N */ +static void build_upper_buffer (re_string_t *pstr); +static void re_string_translate_buffer (re_string_t *pstr); +static unsigned int re_string_context_at (const re_string_t *input, Idx idx, + int eflags) __attribute__ ((pure)); + +/* Functions for string operation. */ + +/* This function allocate the buffers. It is necessary to call + re_string_reconstruct before using the object. */ + +static reg_errcode_t +__attribute_warn_unused_result__ +re_string_allocate (re_string_t *pstr, const char *str, Idx len, Idx init_len, + RE_TRANSLATE_TYPE trans, bool icase, const re_dfa_t *dfa) +{ + reg_errcode_t ret; + Idx init_buf_len; + + /* Ensure at least one character fits into the buffers. */ + if (init_len < dfa->mb_cur_max) + init_len = dfa->mb_cur_max; + init_buf_len = (len + 1 < init_len) ? len + 1: init_len; + re_string_construct_common (str, len, pstr, trans, icase, dfa); + + ret = re_string_realloc_buffers (pstr, init_buf_len); + if (__glibc_unlikely (ret != REG_NOERROR)) + return ret; + + pstr->word_char = dfa->word_char; + pstr->word_ops_used = dfa->word_ops_used; + pstr->mbs = pstr->mbs_allocated ? pstr->mbs : (unsigned char *) str; + pstr->valid_len = (pstr->mbs_allocated || dfa->mb_cur_max > 1) ? 0 : len; + pstr->valid_raw_len = pstr->valid_len; + return REG_NOERROR; +} + +/* This function allocate the buffers, and initialize them. */ + +static reg_errcode_t +__attribute_warn_unused_result__ +re_string_construct (re_string_t *pstr, const char *str, Idx len, + RE_TRANSLATE_TYPE trans, bool icase, const re_dfa_t *dfa) +{ + reg_errcode_t ret; + memset (pstr, '\0', sizeof (re_string_t)); + re_string_construct_common (str, len, pstr, trans, icase, dfa); + + if (len > 0) + { + ret = re_string_realloc_buffers (pstr, len + 1); + if (__glibc_unlikely (ret != REG_NOERROR)) + return ret; + } + pstr->mbs = pstr->mbs_allocated ? pstr->mbs : (unsigned char *) str; + + if (icase) + { +#ifdef RE_ENABLE_I18N + if (dfa->mb_cur_max > 1) + { + while (1) + { + ret = build_wcs_upper_buffer (pstr); + if (__glibc_unlikely (ret != REG_NOERROR)) + return ret; + if (pstr->valid_raw_len >= len) + break; + if (pstr->bufs_len > pstr->valid_len + dfa->mb_cur_max) + break; + ret = re_string_realloc_buffers (pstr, pstr->bufs_len * 2); + if (__glibc_unlikely (ret != REG_NOERROR)) + return ret; + } + } + else +#endif /* RE_ENABLE_I18N */ + build_upper_buffer (pstr); + } + else + { +#ifdef RE_ENABLE_I18N + if (dfa->mb_cur_max > 1) + build_wcs_buffer (pstr); + else +#endif /* RE_ENABLE_I18N */ + { + if (trans != NULL) + re_string_translate_buffer (pstr); + else + { + pstr->valid_len = pstr->bufs_len; + pstr->valid_raw_len = pstr->bufs_len; + } + } + } + + return REG_NOERROR; +} + +/* Helper functions for re_string_allocate, and re_string_construct. */ + +static reg_errcode_t +__attribute_warn_unused_result__ +re_string_realloc_buffers (re_string_t *pstr, Idx new_buf_len) +{ +#ifdef RE_ENABLE_I18N + if (pstr->mb_cur_max > 1) + { + wint_t *new_wcs; + + /* Avoid overflow in realloc. */ + const size_t max_object_size = MAX (sizeof (wint_t), sizeof (Idx)); + if (__glibc_unlikely (MIN (IDX_MAX, SIZE_MAX / max_object_size) + < new_buf_len)) + return REG_ESPACE; + + new_wcs = re_realloc (pstr->wcs, wint_t, new_buf_len); + if (__glibc_unlikely (new_wcs == NULL)) + return REG_ESPACE; + pstr->wcs = new_wcs; + if (pstr->offsets != NULL) + { + Idx *new_offsets = re_realloc (pstr->offsets, Idx, new_buf_len); + if (__glibc_unlikely (new_offsets == NULL)) + return REG_ESPACE; + pstr->offsets = new_offsets; + } + } +#endif /* RE_ENABLE_I18N */ + if (pstr->mbs_allocated) + { + unsigned char *new_mbs = re_realloc (pstr->mbs, unsigned char, + new_buf_len); + if (__glibc_unlikely (new_mbs == NULL)) + return REG_ESPACE; + pstr->mbs = new_mbs; + } + pstr->bufs_len = new_buf_len; + return REG_NOERROR; +} + + +static void +re_string_construct_common (const char *str, Idx len, re_string_t *pstr, + RE_TRANSLATE_TYPE trans, bool icase, + const re_dfa_t *dfa) +{ + pstr->raw_mbs = (const unsigned char *) str; + pstr->len = len; + pstr->raw_len = len; + pstr->trans = trans; + pstr->icase = icase; + pstr->mbs_allocated = (trans != NULL || icase); + pstr->mb_cur_max = dfa->mb_cur_max; + pstr->is_utf8 = dfa->is_utf8; + pstr->map_notascii = dfa->map_notascii; + pstr->stop = pstr->len; + pstr->raw_stop = pstr->stop; +} + +#ifdef RE_ENABLE_I18N + +/* Build wide character buffer PSTR->WCS. + If the byte sequence of the string are: + <mb1>(0), <mb1>(1), <mb2>(0), <mb2>(1), <sb3> + Then wide character buffer will be: + <wc1> , WEOF , <wc2> , WEOF , <wc3> + We use WEOF for padding, they indicate that the position isn't + a first byte of a multibyte character. + + Note that this function assumes PSTR->VALID_LEN elements are already + built and starts from PSTR->VALID_LEN. */ + +static void +build_wcs_buffer (re_string_t *pstr) +{ +#ifdef _LIBC + unsigned char buf[MB_LEN_MAX]; + assert (MB_LEN_MAX >= pstr->mb_cur_max); +#else + unsigned char buf[64]; +#endif + mbstate_t prev_st; + Idx byte_idx, end_idx, remain_len; + size_t mbclen; + + /* Build the buffers from pstr->valid_len to either pstr->len or + pstr->bufs_len. */ + end_idx = (pstr->bufs_len > pstr->len) ? pstr->len : pstr->bufs_len; + for (byte_idx = pstr->valid_len; byte_idx < end_idx;) + { + wchar_t wc; + const char *p; + + remain_len = end_idx - byte_idx; + prev_st = pstr->cur_state; + /* Apply the translation if we need. */ + if (__glibc_unlikely (pstr->trans != NULL)) + { + int i, ch; + + for (i = 0; i < pstr->mb_cur_max && i < remain_len; ++i) + { + ch = pstr->raw_mbs [pstr->raw_mbs_idx + byte_idx + i]; + buf[i] = pstr->mbs[byte_idx + i] = pstr->trans[ch]; + } + p = (const char *) buf; + } + else + p = (const char *) pstr->raw_mbs + pstr->raw_mbs_idx + byte_idx; + mbclen = __mbrtowc (&wc, p, remain_len, &pstr->cur_state); + if (__glibc_unlikely (mbclen == (size_t) -1 || mbclen == 0 + || (mbclen == (size_t) -2 + && pstr->bufs_len >= pstr->len))) + { + /* We treat these cases as a singlebyte character. */ + mbclen = 1; + wc = (wchar_t) pstr->raw_mbs[pstr->raw_mbs_idx + byte_idx]; + if (__glibc_unlikely (pstr->trans != NULL)) + wc = pstr->trans[wc]; + pstr->cur_state = prev_st; + } + else if (__glibc_unlikely (mbclen == (size_t) -2)) + { + /* The buffer doesn't have enough space, finish to build. */ + pstr->cur_state = prev_st; + break; + } + + /* Write wide character and padding. */ + pstr->wcs[byte_idx++] = wc; + /* Write paddings. */ + for (remain_len = byte_idx + mbclen - 1; byte_idx < remain_len ;) + pstr->wcs[byte_idx++] = WEOF; + } + pstr->valid_len = byte_idx; + pstr->valid_raw_len = byte_idx; +} + +/* Build wide character buffer PSTR->WCS like build_wcs_buffer, + but for REG_ICASE. */ + +static reg_errcode_t +__attribute_warn_unused_result__ +build_wcs_upper_buffer (re_string_t *pstr) +{ + mbstate_t prev_st; + Idx src_idx, byte_idx, end_idx, remain_len; + size_t mbclen; +#ifdef _LIBC + char buf[MB_LEN_MAX]; + assert (MB_LEN_MAX >= pstr->mb_cur_max); +#else + char buf[64]; +#endif + + byte_idx = pstr->valid_len; + end_idx = (pstr->bufs_len > pstr->len) ? pstr->len : pstr->bufs_len; + + /* The following optimization assumes that ASCII characters can be + mapped to wide characters with a simple cast. */ + if (! pstr->map_notascii && pstr->trans == NULL && !pstr->offsets_needed) + { + while (byte_idx < end_idx) + { + wchar_t wc; + + if (isascii (pstr->raw_mbs[pstr->raw_mbs_idx + byte_idx]) + && mbsinit (&pstr->cur_state)) + { + /* In case of a singlebyte character. */ + pstr->mbs[byte_idx] + = toupper (pstr->raw_mbs[pstr->raw_mbs_idx + byte_idx]); + /* The next step uses the assumption that wchar_t is encoded + ASCII-safe: all ASCII values can be converted like this. */ + pstr->wcs[byte_idx] = (wchar_t) pstr->mbs[byte_idx]; + ++byte_idx; + continue; + } + + remain_len = end_idx - byte_idx; + prev_st = pstr->cur_state; + mbclen = __mbrtowc (&wc, + ((const char *) pstr->raw_mbs + pstr->raw_mbs_idx + + byte_idx), remain_len, &pstr->cur_state); + if (__glibc_likely (0 < mbclen && mbclen < (size_t) -2)) + { + wchar_t wcu = __towupper (wc); + if (wcu != wc) + { + size_t mbcdlen; + + mbcdlen = __wcrtomb (buf, wcu, &prev_st); + if (__glibc_likely (mbclen == mbcdlen)) + memcpy (pstr->mbs + byte_idx, buf, mbclen); + else + { + src_idx = byte_idx; + goto offsets_needed; + } + } + else + memcpy (pstr->mbs + byte_idx, + pstr->raw_mbs + pstr->raw_mbs_idx + byte_idx, mbclen); + pstr->wcs[byte_idx++] = wcu; + /* Write paddings. */ + for (remain_len = byte_idx + mbclen - 1; byte_idx < remain_len ;) + pstr->wcs[byte_idx++] = WEOF; + } + else if (mbclen == (size_t) -1 || mbclen == 0 + || (mbclen == (size_t) -2 && pstr->bufs_len >= pstr->len)) + { + /* It is an invalid character, an incomplete character + at the end of the string, or '\0'. Just use the byte. */ + int ch = pstr->raw_mbs[pstr->raw_mbs_idx + byte_idx]; + pstr->mbs[byte_idx] = ch; + /* And also cast it to wide char. */ + pstr->wcs[byte_idx++] = (wchar_t) ch; + if (__glibc_unlikely (mbclen == (size_t) -1)) + pstr->cur_state = prev_st; + } + else + { + /* The buffer doesn't have enough space, finish to build. */ + pstr->cur_state = prev_st; + break; + } + } + pstr->valid_len = byte_idx; + pstr->valid_raw_len = byte_idx; + return REG_NOERROR; + } + else + for (src_idx = pstr->valid_raw_len; byte_idx < end_idx;) + { + wchar_t wc; + const char *p; + offsets_needed: + remain_len = end_idx - byte_idx; + prev_st = pstr->cur_state; + if (__glibc_unlikely (pstr->trans != NULL)) + { + int i, ch; + + for (i = 0; i < pstr->mb_cur_max && i < remain_len; ++i) + { + ch = pstr->raw_mbs [pstr->raw_mbs_idx + src_idx + i]; + buf[i] = pstr->trans[ch]; + } + p = (const char *) buf; + } + else + p = (const char *) pstr->raw_mbs + pstr->raw_mbs_idx + src_idx; + mbclen = __mbrtowc (&wc, p, remain_len, &pstr->cur_state); + if (__glibc_likely (0 < mbclen && mbclen < (size_t) -2)) + { + wchar_t wcu = __towupper (wc); + if (wcu != wc) + { + size_t mbcdlen; + + mbcdlen = __wcrtomb ((char *) buf, wcu, &prev_st); + if (__glibc_likely (mbclen == mbcdlen)) + memcpy (pstr->mbs + byte_idx, buf, mbclen); + else if (mbcdlen != (size_t) -1) + { + size_t i; + + if (byte_idx + mbcdlen > pstr->bufs_len) + { + pstr->cur_state = prev_st; + break; + } + + if (pstr->offsets == NULL) + { + pstr->offsets = re_malloc (Idx, pstr->bufs_len); + + if (pstr->offsets == NULL) + return REG_ESPACE; + } + if (!pstr->offsets_needed) + { + for (i = 0; i < (size_t) byte_idx; ++i) + pstr->offsets[i] = i; + pstr->offsets_needed = 1; + } + + memcpy (pstr->mbs + byte_idx, buf, mbcdlen); + pstr->wcs[byte_idx] = wcu; + pstr->offsets[byte_idx] = src_idx; + for (i = 1; i < mbcdlen; ++i) + { + pstr->offsets[byte_idx + i] + = src_idx + (i < mbclen ? i : mbclen - 1); + pstr->wcs[byte_idx + i] = WEOF; + } + pstr->len += mbcdlen - mbclen; + if (pstr->raw_stop > src_idx) + pstr->stop += mbcdlen - mbclen; + end_idx = (pstr->bufs_len > pstr->len) + ? pstr->len : pstr->bufs_len; + byte_idx += mbcdlen; + src_idx += mbclen; + continue; + } + else + memcpy (pstr->mbs + byte_idx, p, mbclen); + } + else + memcpy (pstr->mbs + byte_idx, p, mbclen); + + if (__glibc_unlikely (pstr->offsets_needed != 0)) + { + size_t i; + for (i = 0; i < mbclen; ++i) + pstr->offsets[byte_idx + i] = src_idx + i; + } + src_idx += mbclen; + + pstr->wcs[byte_idx++] = wcu; + /* Write paddings. */ + for (remain_len = byte_idx + mbclen - 1; byte_idx < remain_len ;) + pstr->wcs[byte_idx++] = WEOF; + } + else if (mbclen == (size_t) -1 || mbclen == 0 + || (mbclen == (size_t) -2 && pstr->bufs_len >= pstr->len)) + { + /* It is an invalid character or '\0'. Just use the byte. */ + int ch = pstr->raw_mbs[pstr->raw_mbs_idx + src_idx]; + + if (__glibc_unlikely (pstr->trans != NULL)) + ch = pstr->trans [ch]; + pstr->mbs[byte_idx] = ch; + + if (__glibc_unlikely (pstr->offsets_needed != 0)) + pstr->offsets[byte_idx] = src_idx; + ++src_idx; + + /* And also cast it to wide char. */ + pstr->wcs[byte_idx++] = (wchar_t) ch; + if (__glibc_unlikely (mbclen == (size_t) -1)) + pstr->cur_state = prev_st; + } + else + { + /* The buffer doesn't have enough space, finish to build. */ + pstr->cur_state = prev_st; + break; + } + } + pstr->valid_len = byte_idx; + pstr->valid_raw_len = src_idx; + return REG_NOERROR; +} + +/* Skip characters until the index becomes greater than NEW_RAW_IDX. + Return the index. */ + +static Idx +re_string_skip_chars (re_string_t *pstr, Idx new_raw_idx, wint_t *last_wc) +{ + mbstate_t prev_st; + Idx rawbuf_idx; + size_t mbclen; + wint_t wc = WEOF; + + /* Skip the characters which are not necessary to check. */ + for (rawbuf_idx = pstr->raw_mbs_idx + pstr->valid_raw_len; + rawbuf_idx < new_raw_idx;) + { + wchar_t wc2; + Idx remain_len = pstr->raw_len - rawbuf_idx; + prev_st = pstr->cur_state; + mbclen = __mbrtowc (&wc2, (const char *) pstr->raw_mbs + rawbuf_idx, + remain_len, &pstr->cur_state); + if (__glibc_unlikely (mbclen == (size_t) -2 || mbclen == (size_t) -1 + || mbclen == 0)) + { + /* We treat these cases as a single byte character. */ + if (mbclen == 0 || remain_len == 0) + wc = L'\0'; + else + wc = *(unsigned char *) (pstr->raw_mbs + rawbuf_idx); + mbclen = 1; + pstr->cur_state = prev_st; + } + else + wc = wc2; + /* Then proceed the next character. */ + rawbuf_idx += mbclen; + } + *last_wc = wc; + return rawbuf_idx; +} +#endif /* RE_ENABLE_I18N */ + +/* Build the buffer PSTR->MBS, and apply the translation if we need. + This function is used in case of REG_ICASE. */ + +static void +build_upper_buffer (re_string_t *pstr) +{ + Idx char_idx, end_idx; + end_idx = (pstr->bufs_len > pstr->len) ? pstr->len : pstr->bufs_len; + + for (char_idx = pstr->valid_len; char_idx < end_idx; ++char_idx) + { + int ch = pstr->raw_mbs[pstr->raw_mbs_idx + char_idx]; + if (__glibc_unlikely (pstr->trans != NULL)) + ch = pstr->trans[ch]; + pstr->mbs[char_idx] = toupper (ch); + } + pstr->valid_len = char_idx; + pstr->valid_raw_len = char_idx; +} + +/* Apply TRANS to the buffer in PSTR. */ + +static void +re_string_translate_buffer (re_string_t *pstr) +{ + Idx buf_idx, end_idx; + end_idx = (pstr->bufs_len > pstr->len) ? pstr->len : pstr->bufs_len; + + for (buf_idx = pstr->valid_len; buf_idx < end_idx; ++buf_idx) + { + int ch = pstr->raw_mbs[pstr->raw_mbs_idx + buf_idx]; + pstr->mbs[buf_idx] = pstr->trans[ch]; + } + + pstr->valid_len = buf_idx; + pstr->valid_raw_len = buf_idx; +} + +/* This function re-construct the buffers. + Concretely, convert to wide character in case of pstr->mb_cur_max > 1, + convert to upper case in case of REG_ICASE, apply translation. */ + +static reg_errcode_t +__attribute_warn_unused_result__ +re_string_reconstruct (re_string_t *pstr, Idx idx, int eflags) +{ + Idx offset; + + if (__glibc_unlikely (pstr->raw_mbs_idx <= idx)) + offset = idx - pstr->raw_mbs_idx; + else + { + /* Reset buffer. */ +#ifdef RE_ENABLE_I18N + if (pstr->mb_cur_max > 1) + memset (&pstr->cur_state, '\0', sizeof (mbstate_t)); +#endif /* RE_ENABLE_I18N */ + pstr->len = pstr->raw_len; + pstr->stop = pstr->raw_stop; + pstr->valid_len = 0; + pstr->raw_mbs_idx = 0; + pstr->valid_raw_len = 0; + pstr->offsets_needed = 0; + pstr->tip_context = ((eflags & REG_NOTBOL) ? CONTEXT_BEGBUF + : CONTEXT_NEWLINE | CONTEXT_BEGBUF); + if (!pstr->mbs_allocated) + pstr->mbs = (unsigned char *) pstr->raw_mbs; + offset = idx; + } + + if (__glibc_likely (offset != 0)) + { + /* Should the already checked characters be kept? */ + if (__glibc_likely (offset < pstr->valid_raw_len)) + { + /* Yes, move them to the front of the buffer. */ +#ifdef RE_ENABLE_I18N + if (__glibc_unlikely (pstr->offsets_needed)) + { + Idx low = 0, high = pstr->valid_len, mid; + do + { + mid = (high + low) / 2; + if (pstr->offsets[mid] > offset) + high = mid; + else if (pstr->offsets[mid] < offset) + low = mid + 1; + else + break; + } + while (low < high); + if (pstr->offsets[mid] < offset) + ++mid; + pstr->tip_context = re_string_context_at (pstr, mid - 1, + eflags); + /* This can be quite complicated, so handle specially + only the common and easy case where the character with + different length representation of lower and upper + case is present at or after offset. */ + if (pstr->valid_len > offset + && mid == offset && pstr->offsets[mid] == offset) + { + memmove (pstr->wcs, pstr->wcs + offset, + (pstr->valid_len - offset) * sizeof (wint_t)); + memmove (pstr->mbs, pstr->mbs + offset, pstr->valid_len - offset); + pstr->valid_len -= offset; + pstr->valid_raw_len -= offset; + for (low = 0; low < pstr->valid_len; low++) + pstr->offsets[low] = pstr->offsets[low + offset] - offset; + } + else + { + /* Otherwise, just find out how long the partial multibyte + character at offset is and fill it with WEOF/255. */ + pstr->len = pstr->raw_len - idx + offset; + pstr->stop = pstr->raw_stop - idx + offset; + pstr->offsets_needed = 0; + while (mid > 0 && pstr->offsets[mid - 1] == offset) + --mid; + while (mid < pstr->valid_len) + if (pstr->wcs[mid] != WEOF) + break; + else + ++mid; + if (mid == pstr->valid_len) + pstr->valid_len = 0; + else + { + pstr->valid_len = pstr->offsets[mid] - offset; + if (pstr->valid_len) + { + for (low = 0; low < pstr->valid_len; ++low) + pstr->wcs[low] = WEOF; + memset (pstr->mbs, 255, pstr->valid_len); + } + } + pstr->valid_raw_len = pstr->valid_len; + } + } + else +#endif + { + pstr->tip_context = re_string_context_at (pstr, offset - 1, + eflags); +#ifdef RE_ENABLE_I18N + if (pstr->mb_cur_max > 1) + memmove (pstr->wcs, pstr->wcs + offset, + (pstr->valid_len - offset) * sizeof (wint_t)); +#endif /* RE_ENABLE_I18N */ + if (__glibc_unlikely (pstr->mbs_allocated)) + memmove (pstr->mbs, pstr->mbs + offset, + pstr->valid_len - offset); + pstr->valid_len -= offset; + pstr->valid_raw_len -= offset; +#if defined DEBUG && DEBUG + assert (pstr->valid_len > 0); +#endif + } + } + else + { +#ifdef RE_ENABLE_I18N + /* No, skip all characters until IDX. */ + Idx prev_valid_len = pstr->valid_len; + + if (__glibc_unlikely (pstr->offsets_needed)) + { + pstr->len = pstr->raw_len - idx + offset; + pstr->stop = pstr->raw_stop - idx + offset; + pstr->offsets_needed = 0; + } +#endif + pstr->valid_len = 0; +#ifdef RE_ENABLE_I18N + if (pstr->mb_cur_max > 1) + { + Idx wcs_idx; + wint_t wc = WEOF; + + if (pstr->is_utf8) + { + const unsigned char *raw, *p, *end; + + /* Special case UTF-8. Multi-byte chars start with any + byte other than 0x80 - 0xbf. */ + raw = pstr->raw_mbs + pstr->raw_mbs_idx; + end = raw + (offset - pstr->mb_cur_max); + if (end < pstr->raw_mbs) + end = pstr->raw_mbs; + p = raw + offset - 1; +#ifdef _LIBC + /* We know the wchar_t encoding is UCS4, so for the simple + case, ASCII characters, skip the conversion step. */ + if (isascii (*p) && __glibc_likely (pstr->trans == NULL)) + { + memset (&pstr->cur_state, '\0', sizeof (mbstate_t)); + /* pstr->valid_len = 0; */ + wc = (wchar_t) *p; + } + else +#endif + for (; p >= end; --p) + if ((*p & 0xc0) != 0x80) + { + mbstate_t cur_state; + wchar_t wc2; + Idx mlen = raw + pstr->len - p; + unsigned char buf[6]; + size_t mbclen; + + const unsigned char *pp = p; + if (__glibc_unlikely (pstr->trans != NULL)) + { + int i = mlen < 6 ? mlen : 6; + while (--i >= 0) + buf[i] = pstr->trans[p[i]]; + pp = buf; + } + /* XXX Don't use mbrtowc, we know which conversion + to use (UTF-8 -> UCS4). */ + memset (&cur_state, 0, sizeof (cur_state)); + mbclen = __mbrtowc (&wc2, (const char *) pp, mlen, + &cur_state); + if (raw + offset - p <= mbclen + && mbclen < (size_t) -2) + { + memset (&pstr->cur_state, '\0', + sizeof (mbstate_t)); + pstr->valid_len = mbclen - (raw + offset - p); + wc = wc2; + } + break; + } + } + + if (wc == WEOF) + pstr->valid_len = re_string_skip_chars (pstr, idx, &wc) - idx; + if (wc == WEOF) + pstr->tip_context + = re_string_context_at (pstr, prev_valid_len - 1, eflags); + else + pstr->tip_context = ((__glibc_unlikely (pstr->word_ops_used != 0) + && IS_WIDE_WORD_CHAR (wc)) + ? CONTEXT_WORD + : ((IS_WIDE_NEWLINE (wc) + && pstr->newline_anchor) + ? CONTEXT_NEWLINE : 0)); + if (__glibc_unlikely (pstr->valid_len)) + { + for (wcs_idx = 0; wcs_idx < pstr->valid_len; ++wcs_idx) + pstr->wcs[wcs_idx] = WEOF; + if (pstr->mbs_allocated) + memset (pstr->mbs, 255, pstr->valid_len); + } + pstr->valid_raw_len = pstr->valid_len; + } + else +#endif /* RE_ENABLE_I18N */ + { + int c = pstr->raw_mbs[pstr->raw_mbs_idx + offset - 1]; + pstr->valid_raw_len = 0; + if (pstr->trans) + c = pstr->trans[c]; + pstr->tip_context = (bitset_contain (pstr->word_char, c) + ? CONTEXT_WORD + : ((IS_NEWLINE (c) && pstr->newline_anchor) + ? CONTEXT_NEWLINE : 0)); + } + } + if (!__glibc_unlikely (pstr->mbs_allocated)) + pstr->mbs += offset; + } + pstr->raw_mbs_idx = idx; + pstr->len -= offset; + pstr->stop -= offset; + + /* Then build the buffers. */ +#ifdef RE_ENABLE_I18N + if (pstr->mb_cur_max > 1) + { + if (pstr->icase) + { + reg_errcode_t ret = build_wcs_upper_buffer (pstr); + if (__glibc_unlikely (ret != REG_NOERROR)) + return ret; + } + else + build_wcs_buffer (pstr); + } + else +#endif /* RE_ENABLE_I18N */ + if (__glibc_unlikely (pstr->mbs_allocated)) + { + if (pstr->icase) + build_upper_buffer (pstr); + else if (pstr->trans != NULL) + re_string_translate_buffer (pstr); + } + else + pstr->valid_len = pstr->len; + + pstr->cur_idx = 0; + return REG_NOERROR; +} + +static unsigned char +__attribute__ ((pure)) +re_string_peek_byte_case (const re_string_t *pstr, Idx idx) +{ + int ch; + Idx off; + + /* Handle the common (easiest) cases first. */ + if (__glibc_likely (!pstr->mbs_allocated)) + return re_string_peek_byte (pstr, idx); + +#ifdef RE_ENABLE_I18N + if (pstr->mb_cur_max > 1 + && ! re_string_is_single_byte_char (pstr, pstr->cur_idx + idx)) + return re_string_peek_byte (pstr, idx); +#endif + + off = pstr->cur_idx + idx; +#ifdef RE_ENABLE_I18N + if (pstr->offsets_needed) + off = pstr->offsets[off]; +#endif + + ch = pstr->raw_mbs[pstr->raw_mbs_idx + off]; + +#ifdef RE_ENABLE_I18N + /* Ensure that e.g. for tr_TR.UTF-8 BACKSLASH DOTLESS SMALL LETTER I + this function returns CAPITAL LETTER I instead of first byte of + DOTLESS SMALL LETTER I. The latter would confuse the parser, + since peek_byte_case doesn't advance cur_idx in any way. */ + if (pstr->offsets_needed && !isascii (ch)) + return re_string_peek_byte (pstr, idx); +#endif + + return ch; +} + +static unsigned char +re_string_fetch_byte_case (re_string_t *pstr) +{ + if (__glibc_likely (!pstr->mbs_allocated)) + return re_string_fetch_byte (pstr); + +#ifdef RE_ENABLE_I18N + if (pstr->offsets_needed) + { + Idx off; + int ch; + + /* For tr_TR.UTF-8 [[:islower:]] there is + [[: CAPITAL LETTER I WITH DOT lower:]] in mbs. Skip + in that case the whole multi-byte character and return + the original letter. On the other side, with + [[: DOTLESS SMALL LETTER I return [[:I, as doing + anything else would complicate things too much. */ + + if (!re_string_first_byte (pstr, pstr->cur_idx)) + return re_string_fetch_byte (pstr); + + off = pstr->offsets[pstr->cur_idx]; + ch = pstr->raw_mbs[pstr->raw_mbs_idx + off]; + + if (! isascii (ch)) + return re_string_fetch_byte (pstr); + + re_string_skip_bytes (pstr, + re_string_char_size_at (pstr, pstr->cur_idx)); + return ch; + } +#endif + + return pstr->raw_mbs[pstr->raw_mbs_idx + pstr->cur_idx++]; +} + +static void +re_string_destruct (re_string_t *pstr) +{ +#ifdef RE_ENABLE_I18N + re_free (pstr->wcs); + re_free (pstr->offsets); +#endif /* RE_ENABLE_I18N */ + if (pstr->mbs_allocated) + re_free (pstr->mbs); +} + +/* Return the context at IDX in INPUT. */ + +static unsigned int +re_string_context_at (const re_string_t *input, Idx idx, int eflags) +{ + int c; + if (__glibc_unlikely (idx < 0)) + /* In this case, we use the value stored in input->tip_context, + since we can't know the character in input->mbs[-1] here. */ + return input->tip_context; + if (__glibc_unlikely (idx == input->len)) + return ((eflags & REG_NOTEOL) ? CONTEXT_ENDBUF + : CONTEXT_NEWLINE | CONTEXT_ENDBUF); +#ifdef RE_ENABLE_I18N + if (input->mb_cur_max > 1) + { + wint_t wc; + Idx wc_idx = idx; + while(input->wcs[wc_idx] == WEOF) + { +#if defined DEBUG && DEBUG + /* It must not happen. */ + assert (wc_idx >= 0); +#endif + --wc_idx; + if (wc_idx < 0) + return input->tip_context; + } + wc = input->wcs[wc_idx]; + if (__glibc_unlikely (input->word_ops_used != 0) + && IS_WIDE_WORD_CHAR (wc)) + return CONTEXT_WORD; + return (IS_WIDE_NEWLINE (wc) && input->newline_anchor + ? CONTEXT_NEWLINE : 0); + } + else +#endif + { + c = re_string_byte_at (input, idx); + if (bitset_contain (input->word_char, c)) + return CONTEXT_WORD; + return IS_NEWLINE (c) && input->newline_anchor ? CONTEXT_NEWLINE : 0; + } +} + +/* Functions for set operation. */ + +static reg_errcode_t +__attribute_warn_unused_result__ +re_node_set_alloc (re_node_set *set, Idx size) +{ + set->alloc = size; + set->nelem = 0; + set->elems = re_malloc (Idx, size); + if (__glibc_unlikely (set->elems == NULL) + && (MALLOC_0_IS_NONNULL || size != 0)) + return REG_ESPACE; + return REG_NOERROR; +} + +static reg_errcode_t +__attribute_warn_unused_result__ +re_node_set_init_1 (re_node_set *set, Idx elem) +{ + set->alloc = 1; + set->nelem = 1; + set->elems = re_malloc (Idx, 1); + if (__glibc_unlikely (set->elems == NULL)) + { + set->alloc = set->nelem = 0; + return REG_ESPACE; + } + set->elems[0] = elem; + return REG_NOERROR; +} + +static reg_errcode_t +__attribute_warn_unused_result__ +re_node_set_init_2 (re_node_set *set, Idx elem1, Idx elem2) +{ + set->alloc = 2; + set->elems = re_malloc (Idx, 2); + if (__glibc_unlikely (set->elems == NULL)) + return REG_ESPACE; + if (elem1 == elem2) + { + set->nelem = 1; + set->elems[0] = elem1; + } + else + { + set->nelem = 2; + if (elem1 < elem2) + { + set->elems[0] = elem1; + set->elems[1] = elem2; + } + else + { + set->elems[0] = elem2; + set->elems[1] = elem1; + } + } + return REG_NOERROR; +} + +static reg_errcode_t +__attribute_warn_unused_result__ +re_node_set_init_copy (re_node_set *dest, const re_node_set *src) +{ + dest->nelem = src->nelem; + if (src->nelem > 0) + { + dest->alloc = dest->nelem; + dest->elems = re_malloc (Idx, dest->alloc); + if (__glibc_unlikely (dest->elems == NULL)) + { + dest->alloc = dest->nelem = 0; + return REG_ESPACE; + } + memcpy (dest->elems, src->elems, src->nelem * sizeof (Idx)); + } + else + re_node_set_init_empty (dest); + return REG_NOERROR; +} + +/* Calculate the intersection of the sets SRC1 and SRC2. And merge it to + DEST. Return value indicate the error code or REG_NOERROR if succeeded. + Note: We assume dest->elems is NULL, when dest->alloc is 0. */ + +static reg_errcode_t +__attribute_warn_unused_result__ +re_node_set_add_intersect (re_node_set *dest, const re_node_set *src1, + const re_node_set *src2) +{ + Idx i1, i2, is, id, delta, sbase; + if (src1->nelem == 0 || src2->nelem == 0) + return REG_NOERROR; + + /* We need dest->nelem + 2 * elems_in_intersection; this is a + conservative estimate. */ + if (src1->nelem + src2->nelem + dest->nelem > dest->alloc) + { + Idx new_alloc = src1->nelem + src2->nelem + dest->alloc; + Idx *new_elems = re_realloc (dest->elems, Idx, new_alloc); + if (__glibc_unlikely (new_elems == NULL)) + return REG_ESPACE; + dest->elems = new_elems; + dest->alloc = new_alloc; + } + + /* Find the items in the intersection of SRC1 and SRC2, and copy + into the top of DEST those that are not already in DEST itself. */ + sbase = dest->nelem + src1->nelem + src2->nelem; + i1 = src1->nelem - 1; + i2 = src2->nelem - 1; + id = dest->nelem - 1; + for (;;) + { + if (src1->elems[i1] == src2->elems[i2]) + { + /* Try to find the item in DEST. Maybe we could binary search? */ + while (id >= 0 && dest->elems[id] > src1->elems[i1]) + --id; + + if (id < 0 || dest->elems[id] != src1->elems[i1]) + dest->elems[--sbase] = src1->elems[i1]; + + if (--i1 < 0 || --i2 < 0) + break; + } + + /* Lower the highest of the two items. */ + else if (src1->elems[i1] < src2->elems[i2]) + { + if (--i2 < 0) + break; + } + else + { + if (--i1 < 0) + break; + } + } + + id = dest->nelem - 1; + is = dest->nelem + src1->nelem + src2->nelem - 1; + delta = is - sbase + 1; + + /* Now copy. When DELTA becomes zero, the remaining + DEST elements are already in place; this is more or + less the same loop that is in re_node_set_merge. */ + dest->nelem += delta; + if (delta > 0 && id >= 0) + for (;;) + { + if (dest->elems[is] > dest->elems[id]) + { + /* Copy from the top. */ + dest->elems[id + delta--] = dest->elems[is--]; + if (delta == 0) + break; + } + else + { + /* Slide from the bottom. */ + dest->elems[id + delta] = dest->elems[id]; + if (--id < 0) + break; + } + } + + /* Copy remaining SRC elements. */ + memcpy (dest->elems, dest->elems + sbase, delta * sizeof (Idx)); + + return REG_NOERROR; +} + +/* Calculate the union set of the sets SRC1 and SRC2. And store it to + DEST. Return value indicate the error code or REG_NOERROR if succeeded. */ + +static reg_errcode_t +__attribute_warn_unused_result__ +re_node_set_init_union (re_node_set *dest, const re_node_set *src1, + const re_node_set *src2) +{ + Idx i1, i2, id; + if (src1 != NULL && src1->nelem > 0 && src2 != NULL && src2->nelem > 0) + { + dest->alloc = src1->nelem + src2->nelem; + dest->elems = re_malloc (Idx, dest->alloc); + if (__glibc_unlikely (dest->elems == NULL)) + return REG_ESPACE; + } + else + { + if (src1 != NULL && src1->nelem > 0) + return re_node_set_init_copy (dest, src1); + else if (src2 != NULL && src2->nelem > 0) + return re_node_set_init_copy (dest, src2); + else + re_node_set_init_empty (dest); + return REG_NOERROR; + } + for (i1 = i2 = id = 0 ; i1 < src1->nelem && i2 < src2->nelem ;) + { + if (src1->elems[i1] > src2->elems[i2]) + { + dest->elems[id++] = src2->elems[i2++]; + continue; + } + if (src1->elems[i1] == src2->elems[i2]) + ++i2; + dest->elems[id++] = src1->elems[i1++]; + } + if (i1 < src1->nelem) + { + memcpy (dest->elems + id, src1->elems + i1, + (src1->nelem - i1) * sizeof (Idx)); + id += src1->nelem - i1; + } + else if (i2 < src2->nelem) + { + memcpy (dest->elems + id, src2->elems + i2, + (src2->nelem - i2) * sizeof (Idx)); + id += src2->nelem - i2; + } + dest->nelem = id; + return REG_NOERROR; +} + +/* Calculate the union set of the sets DEST and SRC. And store it to + DEST. Return value indicate the error code or REG_NOERROR if succeeded. */ + +static reg_errcode_t +__attribute_warn_unused_result__ +re_node_set_merge (re_node_set *dest, const re_node_set *src) +{ + Idx is, id, sbase, delta; + if (src == NULL || src->nelem == 0) + return REG_NOERROR; + if (dest->alloc < 2 * src->nelem + dest->nelem) + { + Idx new_alloc = 2 * (src->nelem + dest->alloc); + Idx *new_buffer = re_realloc (dest->elems, Idx, new_alloc); + if (__glibc_unlikely (new_buffer == NULL)) + return REG_ESPACE; + dest->elems = new_buffer; + dest->alloc = new_alloc; + } + + if (__glibc_unlikely (dest->nelem == 0)) + { + dest->nelem = src->nelem; + memcpy (dest->elems, src->elems, src->nelem * sizeof (Idx)); + return REG_NOERROR; + } + + /* Copy into the top of DEST the items of SRC that are not + found in DEST. Maybe we could binary search in DEST? */ + for (sbase = dest->nelem + 2 * src->nelem, + is = src->nelem - 1, id = dest->nelem - 1; is >= 0 && id >= 0; ) + { + if (dest->elems[id] == src->elems[is]) + is--, id--; + else if (dest->elems[id] < src->elems[is]) + dest->elems[--sbase] = src->elems[is--]; + else /* if (dest->elems[id] > src->elems[is]) */ + --id; + } + + if (is >= 0) + { + /* If DEST is exhausted, the remaining items of SRC must be unique. */ + sbase -= is + 1; + memcpy (dest->elems + sbase, src->elems, (is + 1) * sizeof (Idx)); + } + + id = dest->nelem - 1; + is = dest->nelem + 2 * src->nelem - 1; + delta = is - sbase + 1; + if (delta == 0) + return REG_NOERROR; + + /* Now copy. When DELTA becomes zero, the remaining + DEST elements are already in place. */ + dest->nelem += delta; + for (;;) + { + if (dest->elems[is] > dest->elems[id]) + { + /* Copy from the top. */ + dest->elems[id + delta--] = dest->elems[is--]; + if (delta == 0) + break; + } + else + { + /* Slide from the bottom. */ + dest->elems[id + delta] = dest->elems[id]; + if (--id < 0) + { + /* Copy remaining SRC elements. */ + memcpy (dest->elems, dest->elems + sbase, + delta * sizeof (Idx)); + break; + } + } + } + + return REG_NOERROR; +} + +/* Insert the new element ELEM to the re_node_set* SET. + SET should not already have ELEM. + Return true if successful. */ + +static bool +__attribute_warn_unused_result__ +re_node_set_insert (re_node_set *set, Idx elem) +{ + Idx idx; + /* In case the set is empty. */ + if (set->alloc == 0) + return __glibc_likely (re_node_set_init_1 (set, elem) == REG_NOERROR); + + if (__glibc_unlikely (set->nelem) == 0) + { + /* We already guaranteed above that set->alloc != 0. */ + set->elems[0] = elem; + ++set->nelem; + return true; + } + + /* Realloc if we need. */ + if (set->alloc == set->nelem) + { + Idx *new_elems; + set->alloc = set->alloc * 2; + new_elems = re_realloc (set->elems, Idx, set->alloc); + if (__glibc_unlikely (new_elems == NULL)) + return false; + set->elems = new_elems; + } + + /* Move the elements which follows the new element. Test the + first element separately to skip a check in the inner loop. */ + if (elem < set->elems[0]) + { + idx = 0; + for (idx = set->nelem; idx > 0; idx--) + set->elems[idx] = set->elems[idx - 1]; + } + else + { + for (idx = set->nelem; set->elems[idx - 1] > elem; idx--) + set->elems[idx] = set->elems[idx - 1]; + } + + /* Insert the new element. */ + set->elems[idx] = elem; + ++set->nelem; + return true; +} + +/* Insert the new element ELEM to the re_node_set* SET. + SET should not already have any element greater than or equal to ELEM. + Return true if successful. */ + +static bool +__attribute_warn_unused_result__ +re_node_set_insert_last (re_node_set *set, Idx elem) +{ + /* Realloc if we need. */ + if (set->alloc == set->nelem) + { + Idx *new_elems; + set->alloc = (set->alloc + 1) * 2; + new_elems = re_realloc (set->elems, Idx, set->alloc); + if (__glibc_unlikely (new_elems == NULL)) + return false; + set->elems = new_elems; + } + + /* Insert the new element. */ + set->elems[set->nelem++] = elem; + return true; +} + +/* Compare two node sets SET1 and SET2. + Return true if SET1 and SET2 are equivalent. */ + +static bool +__attribute__ ((pure)) +re_node_set_compare (const re_node_set *set1, const re_node_set *set2) +{ + Idx i; + if (set1 == NULL || set2 == NULL || set1->nelem != set2->nelem) + return false; + for (i = set1->nelem ; --i >= 0 ; ) + if (set1->elems[i] != set2->elems[i]) + return false; + return true; +} + +/* Return (idx + 1) if SET contains the element ELEM, return 0 otherwise. */ + +static Idx +__attribute__ ((pure)) +re_node_set_contains (const re_node_set *set, Idx elem) +{ + __re_size_t idx, right, mid; + if (set->nelem <= 0) + return 0; + + /* Binary search the element. */ + idx = 0; + right = set->nelem - 1; + while (idx < right) + { + mid = (idx + right) / 2; + if (set->elems[mid] < elem) + idx = mid + 1; + else + right = mid; + } + return set->elems[idx] == elem ? idx + 1 : 0; +} + +static void +re_node_set_remove_at (re_node_set *set, Idx idx) +{ + if (idx < 0 || idx >= set->nelem) + return; + --set->nelem; + for (; idx < set->nelem; idx++) + set->elems[idx] = set->elems[idx + 1]; +} + + +/* Add the token TOKEN to dfa->nodes, and return the index of the token. + Or return -1 if an error occurred. */ + +static Idx +re_dfa_add_node (re_dfa_t *dfa, re_token_t token) +{ + if (__glibc_unlikely (dfa->nodes_len >= dfa->nodes_alloc)) + { + size_t new_nodes_alloc = dfa->nodes_alloc * 2; + Idx *new_nexts, *new_indices; + re_node_set *new_edests, *new_eclosures; + re_token_t *new_nodes; + + /* Avoid overflows in realloc. */ + const size_t max_object_size = MAX (sizeof (re_token_t), + MAX (sizeof (re_node_set), + sizeof (Idx))); + if (__glibc_unlikely (MIN (IDX_MAX, SIZE_MAX / max_object_size) + < new_nodes_alloc)) + return -1; + + new_nodes = re_realloc (dfa->nodes, re_token_t, new_nodes_alloc); + if (__glibc_unlikely (new_nodes == NULL)) + return -1; + dfa->nodes = new_nodes; + new_nexts = re_realloc (dfa->nexts, Idx, new_nodes_alloc); + new_indices = re_realloc (dfa->org_indices, Idx, new_nodes_alloc); + new_edests = re_realloc (dfa->edests, re_node_set, new_nodes_alloc); + new_eclosures = re_realloc (dfa->eclosures, re_node_set, new_nodes_alloc); + if (__glibc_unlikely (new_nexts == NULL || new_indices == NULL + || new_edests == NULL || new_eclosures == NULL)) + { + re_free (new_nexts); + re_free (new_indices); + re_free (new_edests); + re_free (new_eclosures); + return -1; + } + dfa->nexts = new_nexts; + dfa->org_indices = new_indices; + dfa->edests = new_edests; + dfa->eclosures = new_eclosures; + dfa->nodes_alloc = new_nodes_alloc; + } + dfa->nodes[dfa->nodes_len] = token; + dfa->nodes[dfa->nodes_len].constraint = 0; +#ifdef RE_ENABLE_I18N + dfa->nodes[dfa->nodes_len].accept_mb = + ((token.type == OP_PERIOD && dfa->mb_cur_max > 1) + || token.type == COMPLEX_BRACKET); +#endif + dfa->nexts[dfa->nodes_len] = -1; + re_node_set_init_empty (dfa->edests + dfa->nodes_len); + re_node_set_init_empty (dfa->eclosures + dfa->nodes_len); + return dfa->nodes_len++; +} + +static re_hashval_t +calc_state_hash (const re_node_set *nodes, unsigned int context) +{ + re_hashval_t hash = nodes->nelem + context; + Idx i; + for (i = 0 ; i < nodes->nelem ; i++) + hash += nodes->elems[i]; + return hash; +} + +/* Search for the state whose node_set is equivalent to NODES. + Return the pointer to the state, if we found it in the DFA. + Otherwise create the new one and return it. In case of an error + return NULL and set the error code in ERR. + Note: - We assume NULL as the invalid state, then it is possible that + return value is NULL and ERR is REG_NOERROR. + - We never return non-NULL value in case of any errors, it is for + optimization. */ + +static re_dfastate_t * +__attribute_warn_unused_result__ +re_acquire_state (reg_errcode_t *err, const re_dfa_t *dfa, + const re_node_set *nodes) +{ + re_hashval_t hash; + re_dfastate_t *new_state; + struct re_state_table_entry *spot; + Idx i; +#if defined GCC_LINT || defined lint + /* Suppress bogus uninitialized-variable warnings. */ + *err = REG_NOERROR; +#endif + if (__glibc_unlikely (nodes->nelem == 0)) + { + *err = REG_NOERROR; + return NULL; + } + hash = calc_state_hash (nodes, 0); + spot = dfa->state_table + (hash & dfa->state_hash_mask); + + for (i = 0 ; i < spot->num ; i++) + { + re_dfastate_t *state = spot->array[i]; + if (hash != state->hash) + continue; + if (re_node_set_compare (&state->nodes, nodes)) + return state; + } + + /* There are no appropriate state in the dfa, create the new one. */ + new_state = create_ci_newstate (dfa, nodes, hash); + if (__glibc_unlikely (new_state == NULL)) + *err = REG_ESPACE; + + return new_state; +} + +/* Search for the state whose node_set is equivalent to NODES and + whose context is equivalent to CONTEXT. + Return the pointer to the state, if we found it in the DFA. + Otherwise create the new one and return it. In case of an error + return NULL and set the error code in ERR. + Note: - We assume NULL as the invalid state, then it is possible that + return value is NULL and ERR is REG_NOERROR. + - We never return non-NULL value in case of any errors, it is for + optimization. */ + +static re_dfastate_t * +__attribute_warn_unused_result__ +re_acquire_state_context (reg_errcode_t *err, const re_dfa_t *dfa, + const re_node_set *nodes, unsigned int context) +{ + re_hashval_t hash; + re_dfastate_t *new_state; + struct re_state_table_entry *spot; + Idx i; +#if defined GCC_LINT || defined lint + /* Suppress bogus uninitialized-variable warnings. */ + *err = REG_NOERROR; +#endif + if (nodes->nelem == 0) + { + *err = REG_NOERROR; + return NULL; + } + hash = calc_state_hash (nodes, context); + spot = dfa->state_table + (hash & dfa->state_hash_mask); + + for (i = 0 ; i < spot->num ; i++) + { + re_dfastate_t *state = spot->array[i]; + if (state->hash == hash + && state->context == context + && re_node_set_compare (state->entrance_nodes, nodes)) + return state; + } + /* There are no appropriate state in 'dfa', create the new one. */ + new_state = create_cd_newstate (dfa, nodes, context, hash); + if (__glibc_unlikely (new_state == NULL)) + *err = REG_ESPACE; + + return new_state; +} + +/* Finish initialization of the new state NEWSTATE, and using its hash value + HASH put in the appropriate bucket of DFA's state table. Return value + indicates the error code if failed. */ + +static reg_errcode_t +__attribute_warn_unused_result__ +register_state (const re_dfa_t *dfa, re_dfastate_t *newstate, + re_hashval_t hash) +{ + struct re_state_table_entry *spot; + reg_errcode_t err; + Idx i; + + newstate->hash = hash; + err = re_node_set_alloc (&newstate->non_eps_nodes, newstate->nodes.nelem); + if (__glibc_unlikely (err != REG_NOERROR)) + return REG_ESPACE; + for (i = 0; i < newstate->nodes.nelem; i++) + { + Idx elem = newstate->nodes.elems[i]; + if (!IS_EPSILON_NODE (dfa->nodes[elem].type)) + if (! re_node_set_insert_last (&newstate->non_eps_nodes, elem)) + return REG_ESPACE; + } + + spot = dfa->state_table + (hash & dfa->state_hash_mask); + if (__glibc_unlikely (spot->alloc <= spot->num)) + { + Idx new_alloc = 2 * spot->num + 2; + re_dfastate_t **new_array = re_realloc (spot->array, re_dfastate_t *, + new_alloc); + if (__glibc_unlikely (new_array == NULL)) + return REG_ESPACE; + spot->array = new_array; + spot->alloc = new_alloc; + } + spot->array[spot->num++] = newstate; + return REG_NOERROR; +} + +static void +free_state (re_dfastate_t *state) +{ + re_node_set_free (&state->non_eps_nodes); + re_node_set_free (&state->inveclosure); + if (state->entrance_nodes != &state->nodes) + { + re_node_set_free (state->entrance_nodes); + re_free (state->entrance_nodes); + } + re_node_set_free (&state->nodes); + re_free (state->word_trtable); + re_free (state->trtable); + re_free (state); +} + +/* Create the new state which is independent of contexts. + Return the new state if succeeded, otherwise return NULL. */ + +static re_dfastate_t * +__attribute_warn_unused_result__ +create_ci_newstate (const re_dfa_t *dfa, const re_node_set *nodes, + re_hashval_t hash) +{ + Idx i; + reg_errcode_t err; + re_dfastate_t *newstate; + + newstate = (re_dfastate_t *) calloc (sizeof (re_dfastate_t), 1); + if (__glibc_unlikely (newstate == NULL)) + return NULL; + err = re_node_set_init_copy (&newstate->nodes, nodes); + if (__glibc_unlikely (err != REG_NOERROR)) + { + re_free (newstate); + return NULL; + } + + newstate->entrance_nodes = &newstate->nodes; + for (i = 0 ; i < nodes->nelem ; i++) + { + re_token_t *node = dfa->nodes + nodes->elems[i]; + re_token_type_t type = node->type; + if (type == CHARACTER && !node->constraint) + continue; +#ifdef RE_ENABLE_I18N + newstate->accept_mb |= node->accept_mb; +#endif /* RE_ENABLE_I18N */ + + /* If the state has the halt node, the state is a halt state. */ + if (type == END_OF_RE) + newstate->halt = 1; + else if (type == OP_BACK_REF) + newstate->has_backref = 1; + else if (type == ANCHOR || node->constraint) + newstate->has_constraint = 1; + } + err = register_state (dfa, newstate, hash); + if (__glibc_unlikely (err != REG_NOERROR)) + { + free_state (newstate); + newstate = NULL; + } + return newstate; +} + +/* Create the new state which is depend on the context CONTEXT. + Return the new state if succeeded, otherwise return NULL. */ + +static re_dfastate_t * +__attribute_warn_unused_result__ +create_cd_newstate (const re_dfa_t *dfa, const re_node_set *nodes, + unsigned int context, re_hashval_t hash) +{ + Idx i, nctx_nodes = 0; + reg_errcode_t err; + re_dfastate_t *newstate; + + newstate = (re_dfastate_t *) calloc (sizeof (re_dfastate_t), 1); + if (__glibc_unlikely (newstate == NULL)) + return NULL; + err = re_node_set_init_copy (&newstate->nodes, nodes); + if (__glibc_unlikely (err != REG_NOERROR)) + { + re_free (newstate); + return NULL; + } + + newstate->context = context; + newstate->entrance_nodes = &newstate->nodes; + + for (i = 0 ; i < nodes->nelem ; i++) + { + re_token_t *node = dfa->nodes + nodes->elems[i]; + re_token_type_t type = node->type; + unsigned int constraint = node->constraint; + + if (type == CHARACTER && !constraint) + continue; +#ifdef RE_ENABLE_I18N + newstate->accept_mb |= node->accept_mb; +#endif /* RE_ENABLE_I18N */ + + /* If the state has the halt node, the state is a halt state. */ + if (type == END_OF_RE) + newstate->halt = 1; + else if (type == OP_BACK_REF) + newstate->has_backref = 1; + + if (constraint) + { + if (newstate->entrance_nodes == &newstate->nodes) + { + newstate->entrance_nodes = re_malloc (re_node_set, 1); + if (__glibc_unlikely (newstate->entrance_nodes == NULL)) + { + free_state (newstate); + return NULL; + } + if (re_node_set_init_copy (newstate->entrance_nodes, nodes) + != REG_NOERROR) + return NULL; + nctx_nodes = 0; + newstate->has_constraint = 1; + } + + if (NOT_SATISFY_PREV_CONSTRAINT (constraint,context)) + { + re_node_set_remove_at (&newstate->nodes, i - nctx_nodes); + ++nctx_nodes; + } + } + } + err = register_state (dfa, newstate, hash); + if (__glibc_unlikely (err != REG_NOERROR)) + { + free_state (newstate); + newstate = NULL; + } + return newstate; +} diff --git a/lib/regex_internal.h b/lib/regex_internal.h new file mode 100644 index 0000000..b0e49cd --- /dev/null +++ b/lib/regex_internal.h @@ -0,0 +1,896 @@ +/* Extended regular expression matching and search library. + Copyright (C) 2002-2018 Free Software Foundation, Inc. + This file is part of the GNU C Library. + Contributed by Isamu Hasegawa <isamu@yamato.ibm.com>. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU General Public + License as published by the Free Software Foundation; either + version 3 of the License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + General Public License for more details. + + You should have received a copy of the GNU General Public + License along with the GNU C Library; if not, see + <https://www.gnu.org/licenses/>. */ + +#ifndef _REGEX_INTERNAL_H +#define _REGEX_INTERNAL_H 1 + +#include <assert.h> +#include <ctype.h> +#include <stdio.h> +#include <stdlib.h> +#include <string.h> + +#include <langinfo.h> +#include <locale.h> +#include <wchar.h> +#include <wctype.h> +#include <stdbool.h> +#include <stdint.h> + +#include <intprops.h> + +#ifdef _LIBC +# include <libc-lock.h> +# define lock_define(name) __libc_lock_define (, name) +# define lock_init(lock) (__libc_lock_init (lock), 0) +# define lock_fini(lock) ((void) 0) +# define lock_lock(lock) __libc_lock_lock (lock) +# define lock_unlock(lock) __libc_lock_unlock (lock) +#elif defined GNULIB_LOCK && !defined USE_UNLOCKED_IO +# include "glthread/lock.h" + /* Use gl_lock_define if empty macro arguments are known to work. + Otherwise, fall back on less-portable substitutes. */ +# if ((defined __GNUC__ && !defined __STRICT_ANSI__) \ + || (defined __STDC_VERSION__ && 199901L <= __STDC_VERSION__)) +# define lock_define(name) gl_lock_define (, name) +# elif USE_POSIX_THREADS +# define lock_define(name) pthread_mutex_t name; +# elif USE_PTH_THREADS +# define lock_define(name) pth_mutex_t name; +# elif USE_SOLARIS_THREADS +# define lock_define(name) mutex_t name; +# elif USE_WINDOWS_THREADS +# define lock_define(name) gl_lock_t name; +# else +# define lock_define(name) +# endif +# define lock_init(lock) glthread_lock_init (&(lock)) +# define lock_fini(lock) glthread_lock_destroy (&(lock)) +# define lock_lock(lock) glthread_lock_lock (&(lock)) +# define lock_unlock(lock) glthread_lock_unlock (&(lock)) +#elif defined GNULIB_PTHREAD && !defined USE_UNLOCKED_IO +# include <pthread.h> +# define lock_define(name) pthread_mutex_t name; +# define lock_init(lock) pthread_mutex_init (&(lock), 0) +# define lock_fini(lock) pthread_mutex_destroy (&(lock)) +# define lock_lock(lock) pthread_mutex_lock (&(lock)) +# define lock_unlock(lock) pthread_mutex_unlock (&(lock)) +#else +# define lock_define(name) +# define lock_init(lock) 0 +# define lock_fini(lock) ((void) 0) + /* The 'dfa' avoids an "unused variable 'dfa'" warning from GCC. */ +# define lock_lock(lock) ((void) dfa) +# define lock_unlock(lock) ((void) 0) +#endif + +/* In case that the system doesn't have isblank(). */ +#if !defined _LIBC && ! (defined isblank || (HAVE_ISBLANK && HAVE_DECL_ISBLANK)) +# define isblank(ch) ((ch) == ' ' || (ch) == '\t') +#endif + +#ifdef _LIBC +# ifndef _RE_DEFINE_LOCALE_FUNCTIONS +# define _RE_DEFINE_LOCALE_FUNCTIONS 1 +# include <locale/localeinfo.h> +# include <locale/coll-lookup.h> +# endif +#endif + +/* This is for other GNU distributions with internationalized messages. */ +#if (HAVE_LIBINTL_H && ENABLE_NLS) || defined _LIBC +# include <libintl.h> +# ifdef _LIBC +# undef gettext +# define gettext(msgid) \ + __dcgettext (_libc_intl_domainname, msgid, LC_MESSAGES) +# endif +#else +# undef gettext +# define gettext(msgid) (msgid) +#endif + +#ifndef gettext_noop +/* This define is so xgettext can find the internationalizable + strings. */ +# define gettext_noop(String) String +#endif + +#if (defined MB_CUR_MAX && HAVE_WCTYPE_H && HAVE_ISWCTYPE) || _LIBC +# define RE_ENABLE_I18N +#endif + +/* Number of ASCII characters. */ +#define ASCII_CHARS 0x80 + +/* Number of single byte characters. */ +#define SBC_MAX (UCHAR_MAX + 1) + +#define COLL_ELEM_LEN_MAX 8 + +/* The character which represents newline. */ +#define NEWLINE_CHAR '\n' +#define WIDE_NEWLINE_CHAR L'\n' + +/* Rename to standard API for using out of glibc. */ +#ifndef _LIBC +# undef __wctype +# undef __iswalnum +# undef __iswctype +# undef __towlower +# undef __towupper +# define __wctype wctype +# define __iswalnum iswalnum +# define __iswctype iswctype +# define __towlower towlower +# define __towupper towupper +# define __btowc btowc +# define __mbrtowc mbrtowc +# define __wcrtomb wcrtomb +# define __regfree regfree +# define attribute_hidden +#endif /* not _LIBC */ + +#if __GNUC__ < 3 + (__GNUC_MINOR__ < 1) +# define __attribute__(arg) +#endif + +#ifndef SSIZE_MAX +# define SSIZE_MAX ((ssize_t) (SIZE_MAX / 2)) +#endif + +/* The type of indexes into strings. This is signed, not size_t, + since the API requires indexes to fit in regoff_t anyway, and using + signed integers makes the code a bit smaller and presumably faster. + The traditional GNU regex implementation uses int for indexes. + The POSIX-compatible implementation uses a possibly-wider type. + The name 'Idx' is three letters to minimize the hassle of + reindenting a lot of regex code that formerly used 'int'. */ +typedef regoff_t Idx; +#ifdef _REGEX_LARGE_OFFSETS +# define IDX_MAX SSIZE_MAX +#else +# define IDX_MAX INT_MAX +#endif + +/* A hash value, suitable for computing hash tables. */ +typedef __re_size_t re_hashval_t; + +/* An integer used to represent a set of bits. It must be unsigned, + and must be at least as wide as unsigned int. */ +typedef unsigned long int bitset_word_t; +/* All bits set in a bitset_word_t. */ +#define BITSET_WORD_MAX ULONG_MAX + +/* Number of bits in a bitset_word_t. For portability to hosts with + padding bits, do not use '(sizeof (bitset_word_t) * CHAR_BIT)'; + instead, deduce it directly from BITSET_WORD_MAX. Avoid + greater-than-32-bit integers and unconditional shifts by more than + 31 bits, as they're not portable. */ +#if BITSET_WORD_MAX == 0xffffffffUL +# define BITSET_WORD_BITS 32 +#elif BITSET_WORD_MAX >> 31 >> 4 == 1 +# define BITSET_WORD_BITS 36 +#elif BITSET_WORD_MAX >> 31 >> 16 == 1 +# define BITSET_WORD_BITS 48 +#elif BITSET_WORD_MAX >> 31 >> 28 == 1 +# define BITSET_WORD_BITS 60 +#elif BITSET_WORD_MAX >> 31 >> 31 >> 1 == 1 +# define BITSET_WORD_BITS 64 +#elif BITSET_WORD_MAX >> 31 >> 31 >> 9 == 1 +# define BITSET_WORD_BITS 72 +#elif BITSET_WORD_MAX >> 31 >> 31 >> 31 >> 31 >> 3 == 1 +# define BITSET_WORD_BITS 128 +#elif BITSET_WORD_MAX >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 7 == 1 +# define BITSET_WORD_BITS 256 +#elif BITSET_WORD_MAX >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 7 > 1 +# define BITSET_WORD_BITS 257 /* any value > SBC_MAX will do here */ +# if BITSET_WORD_BITS <= SBC_MAX +# error "Invalid SBC_MAX" +# endif +#else +# error "Add case for new bitset_word_t size" +#endif + +/* Number of bitset_word_t values in a bitset_t. */ +#define BITSET_WORDS ((SBC_MAX + BITSET_WORD_BITS - 1) / BITSET_WORD_BITS) + +typedef bitset_word_t bitset_t[BITSET_WORDS]; +typedef bitset_word_t *re_bitset_ptr_t; +typedef const bitset_word_t *re_const_bitset_ptr_t; + +#define PREV_WORD_CONSTRAINT 0x0001 +#define PREV_NOTWORD_CONSTRAINT 0x0002 +#define NEXT_WORD_CONSTRAINT 0x0004 +#define NEXT_NOTWORD_CONSTRAINT 0x0008 +#define PREV_NEWLINE_CONSTRAINT 0x0010 +#define NEXT_NEWLINE_CONSTRAINT 0x0020 +#define PREV_BEGBUF_CONSTRAINT 0x0040 +#define NEXT_ENDBUF_CONSTRAINT 0x0080 +#define WORD_DELIM_CONSTRAINT 0x0100 +#define NOT_WORD_DELIM_CONSTRAINT 0x0200 + +typedef enum +{ + INSIDE_WORD = PREV_WORD_CONSTRAINT | NEXT_WORD_CONSTRAINT, + WORD_FIRST = PREV_NOTWORD_CONSTRAINT | NEXT_WORD_CONSTRAINT, + WORD_LAST = PREV_WORD_CONSTRAINT | NEXT_NOTWORD_CONSTRAINT, + INSIDE_NOTWORD = PREV_NOTWORD_CONSTRAINT | NEXT_NOTWORD_CONSTRAINT, + LINE_FIRST = PREV_NEWLINE_CONSTRAINT, + LINE_LAST = NEXT_NEWLINE_CONSTRAINT, + BUF_FIRST = PREV_BEGBUF_CONSTRAINT, + BUF_LAST = NEXT_ENDBUF_CONSTRAINT, + WORD_DELIM = WORD_DELIM_CONSTRAINT, + NOT_WORD_DELIM = NOT_WORD_DELIM_CONSTRAINT +} re_context_type; + +typedef struct +{ + Idx alloc; + Idx nelem; + Idx *elems; +} re_node_set; + +typedef enum +{ + NON_TYPE = 0, + + /* Node type, These are used by token, node, tree. */ + CHARACTER = 1, + END_OF_RE = 2, + SIMPLE_BRACKET = 3, + OP_BACK_REF = 4, + OP_PERIOD = 5, +#ifdef RE_ENABLE_I18N + COMPLEX_BRACKET = 6, + OP_UTF8_PERIOD = 7, +#endif /* RE_ENABLE_I18N */ + + /* We define EPSILON_BIT as a macro so that OP_OPEN_SUBEXP is used + when the debugger shows values of this enum type. */ +#define EPSILON_BIT 8 + OP_OPEN_SUBEXP = EPSILON_BIT | 0, + OP_CLOSE_SUBEXP = EPSILON_BIT | 1, + OP_ALT = EPSILON_BIT | 2, + OP_DUP_ASTERISK = EPSILON_BIT | 3, + ANCHOR = EPSILON_BIT | 4, + + /* Tree type, these are used only by tree. */ + CONCAT = 16, + SUBEXP = 17, + + /* Token type, these are used only by token. */ + OP_DUP_PLUS = 18, + OP_DUP_QUESTION, + OP_OPEN_BRACKET, + OP_CLOSE_BRACKET, + OP_CHARSET_RANGE, + OP_OPEN_DUP_NUM, + OP_CLOSE_DUP_NUM, + OP_NON_MATCH_LIST, + OP_OPEN_COLL_ELEM, + OP_CLOSE_COLL_ELEM, + OP_OPEN_EQUIV_CLASS, + OP_CLOSE_EQUIV_CLASS, + OP_OPEN_CHAR_CLASS, + OP_CLOSE_CHAR_CLASS, + OP_WORD, + OP_NOTWORD, + OP_SPACE, + OP_NOTSPACE, + BACK_SLASH + +} re_token_type_t; + +#ifdef RE_ENABLE_I18N +typedef struct +{ + /* Multibyte characters. */ + wchar_t *mbchars; + + /* Collating symbols. */ +# ifdef _LIBC + int32_t *coll_syms; +# endif + + /* Equivalence classes. */ +# ifdef _LIBC + int32_t *equiv_classes; +# endif + + /* Range expressions. */ +# ifdef _LIBC + uint32_t *range_starts; + uint32_t *range_ends; +# else /* not _LIBC */ + wchar_t *range_starts; + wchar_t *range_ends; +# endif /* not _LIBC */ + + /* Character classes. */ + wctype_t *char_classes; + + /* If this character set is the non-matching list. */ + unsigned int non_match : 1; + + /* # of multibyte characters. */ + Idx nmbchars; + + /* # of collating symbols. */ + Idx ncoll_syms; + + /* # of equivalence classes. */ + Idx nequiv_classes; + + /* # of range expressions. */ + Idx nranges; + + /* # of character classes. */ + Idx nchar_classes; +} re_charset_t; +#endif /* RE_ENABLE_I18N */ + +typedef struct +{ + union + { + unsigned char c; /* for CHARACTER */ + re_bitset_ptr_t sbcset; /* for SIMPLE_BRACKET */ +#ifdef RE_ENABLE_I18N + re_charset_t *mbcset; /* for COMPLEX_BRACKET */ +#endif /* RE_ENABLE_I18N */ + Idx idx; /* for BACK_REF */ + re_context_type ctx_type; /* for ANCHOR */ + } opr; +#if __GNUC__ >= 2 && !defined __STRICT_ANSI__ + re_token_type_t type : 8; +#else + re_token_type_t type; +#endif + unsigned int constraint : 10; /* context constraint */ + unsigned int duplicated : 1; + unsigned int opt_subexp : 1; +#ifdef RE_ENABLE_I18N + unsigned int accept_mb : 1; + /* These 2 bits can be moved into the union if needed (e.g. if running out + of bits; move opr.c to opr.c.c and move the flags to opr.c.flags). */ + unsigned int mb_partial : 1; +#endif + unsigned int word_char : 1; +} re_token_t; + +#define IS_EPSILON_NODE(type) ((type) & EPSILON_BIT) + +struct re_string_t +{ + /* Indicate the raw buffer which is the original string passed as an + argument of regexec(), re_search(), etc.. */ + const unsigned char *raw_mbs; + /* Store the multibyte string. In case of "case insensitive mode" like + REG_ICASE, upper cases of the string are stored, otherwise MBS points + the same address that RAW_MBS points. */ + unsigned char *mbs; +#ifdef RE_ENABLE_I18N + /* Store the wide character string which is corresponding to MBS. */ + wint_t *wcs; + Idx *offsets; + mbstate_t cur_state; +#endif + /* Index in RAW_MBS. Each character mbs[i] corresponds to + raw_mbs[raw_mbs_idx + i]. */ + Idx raw_mbs_idx; + /* The length of the valid characters in the buffers. */ + Idx valid_len; + /* The corresponding number of bytes in raw_mbs array. */ + Idx valid_raw_len; + /* The length of the buffers MBS and WCS. */ + Idx bufs_len; + /* The index in MBS, which is updated by re_string_fetch_byte. */ + Idx cur_idx; + /* length of RAW_MBS array. */ + Idx raw_len; + /* This is RAW_LEN - RAW_MBS_IDX + VALID_LEN - VALID_RAW_LEN. */ + Idx len; + /* End of the buffer may be shorter than its length in the cases such + as re_match_2, re_search_2. Then, we use STOP for end of the buffer + instead of LEN. */ + Idx raw_stop; + /* This is RAW_STOP - RAW_MBS_IDX adjusted through OFFSETS. */ + Idx stop; + + /* The context of mbs[0]. We store the context independently, since + the context of mbs[0] may be different from raw_mbs[0], which is + the beginning of the input string. */ + unsigned int tip_context; + /* The translation passed as a part of an argument of re_compile_pattern. */ + RE_TRANSLATE_TYPE trans; + /* Copy of re_dfa_t's word_char. */ + re_const_bitset_ptr_t word_char; + /* true if REG_ICASE. */ + unsigned char icase; + unsigned char is_utf8; + unsigned char map_notascii; + unsigned char mbs_allocated; + unsigned char offsets_needed; + unsigned char newline_anchor; + unsigned char word_ops_used; + int mb_cur_max; +}; +typedef struct re_string_t re_string_t; + + +struct re_dfa_t; +typedef struct re_dfa_t re_dfa_t; + +#ifndef _LIBC +# define IS_IN(libc) false +#endif + +#define re_string_peek_byte(pstr, offset) \ + ((pstr)->mbs[(pstr)->cur_idx + offset]) +#define re_string_fetch_byte(pstr) \ + ((pstr)->mbs[(pstr)->cur_idx++]) +#define re_string_first_byte(pstr, idx) \ + ((idx) == (pstr)->valid_len || (pstr)->wcs[idx] != WEOF) +#define re_string_is_single_byte_char(pstr, idx) \ + ((pstr)->wcs[idx] != WEOF && ((pstr)->valid_len == (idx) + 1 \ + || (pstr)->wcs[(idx) + 1] != WEOF)) +#define re_string_eoi(pstr) ((pstr)->stop <= (pstr)->cur_idx) +#define re_string_cur_idx(pstr) ((pstr)->cur_idx) +#define re_string_get_buffer(pstr) ((pstr)->mbs) +#define re_string_length(pstr) ((pstr)->len) +#define re_string_byte_at(pstr,idx) ((pstr)->mbs[idx]) +#define re_string_skip_bytes(pstr,idx) ((pstr)->cur_idx += (idx)) +#define re_string_set_index(pstr,idx) ((pstr)->cur_idx = (idx)) + +#if defined _LIBC || HAVE_ALLOCA +# include <alloca.h> +#endif + +#ifndef _LIBC +# if HAVE_ALLOCA +/* The OS usually guarantees only one guard page at the bottom of the stack, + and a page size can be as small as 4096 bytes. So we cannot safely + allocate anything larger than 4096 bytes. Also care for the possibility + of a few compiler-allocated temporary stack slots. */ +# define __libc_use_alloca(n) ((n) < 4032) +# else +/* alloca is implemented with malloc, so just use malloc. */ +# define __libc_use_alloca(n) 0 +# undef alloca +# define alloca(n) malloc (n) +# endif +#endif + +#ifdef _LIBC +# define MALLOC_0_IS_NONNULL 1 +#elif !defined MALLOC_0_IS_NONNULL +# define MALLOC_0_IS_NONNULL 0 +#endif + +#ifndef MAX +# define MAX(a,b) ((a) < (b) ? (b) : (a)) +#endif +#ifndef MIN +# define MIN(a,b) ((a) < (b) ? (a) : (b)) +#endif + +#define re_malloc(t,n) ((t *) malloc ((n) * sizeof (t))) +#define re_realloc(p,t,n) ((t *) realloc (p, (n) * sizeof (t))) +#define re_free(p) free (p) + +struct bin_tree_t +{ + struct bin_tree_t *parent; + struct bin_tree_t *left; + struct bin_tree_t *right; + struct bin_tree_t *first; + struct bin_tree_t *next; + + re_token_t token; + + /* 'node_idx' is the index in dfa->nodes, if 'type' == 0. + Otherwise 'type' indicate the type of this node. */ + Idx node_idx; +}; +typedef struct bin_tree_t bin_tree_t; + +#define BIN_TREE_STORAGE_SIZE \ + ((1024 - sizeof (void *)) / sizeof (bin_tree_t)) + +struct bin_tree_storage_t +{ + struct bin_tree_storage_t *next; + bin_tree_t data[BIN_TREE_STORAGE_SIZE]; +}; +typedef struct bin_tree_storage_t bin_tree_storage_t; + +#define CONTEXT_WORD 1 +#define CONTEXT_NEWLINE (CONTEXT_WORD << 1) +#define CONTEXT_BEGBUF (CONTEXT_NEWLINE << 1) +#define CONTEXT_ENDBUF (CONTEXT_BEGBUF << 1) + +#define IS_WORD_CONTEXT(c) ((c) & CONTEXT_WORD) +#define IS_NEWLINE_CONTEXT(c) ((c) & CONTEXT_NEWLINE) +#define IS_BEGBUF_CONTEXT(c) ((c) & CONTEXT_BEGBUF) +#define IS_ENDBUF_CONTEXT(c) ((c) & CONTEXT_ENDBUF) +#define IS_ORDINARY_CONTEXT(c) ((c) == 0) + +#define IS_WORD_CHAR(ch) (isalnum (ch) || (ch) == '_') +#define IS_NEWLINE(ch) ((ch) == NEWLINE_CHAR) +#define IS_WIDE_WORD_CHAR(ch) (__iswalnum (ch) || (ch) == L'_') +#define IS_WIDE_NEWLINE(ch) ((ch) == WIDE_NEWLINE_CHAR) + +#define NOT_SATISFY_PREV_CONSTRAINT(constraint,context) \ + ((((constraint) & PREV_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \ + || ((constraint & PREV_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \ + || ((constraint & PREV_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context))\ + || ((constraint & PREV_BEGBUF_CONSTRAINT) && !IS_BEGBUF_CONTEXT (context))) + +#define NOT_SATISFY_NEXT_CONSTRAINT(constraint,context) \ + ((((constraint) & NEXT_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \ + || (((constraint) & NEXT_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \ + || (((constraint) & NEXT_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context)) \ + || (((constraint) & NEXT_ENDBUF_CONSTRAINT) && !IS_ENDBUF_CONTEXT (context))) + +struct re_dfastate_t +{ + re_hashval_t hash; + re_node_set nodes; + re_node_set non_eps_nodes; + re_node_set inveclosure; + re_node_set *entrance_nodes; + struct re_dfastate_t **trtable, **word_trtable; + unsigned int context : 4; + unsigned int halt : 1; + /* If this state can accept "multi byte". + Note that we refer to multibyte characters, and multi character + collating elements as "multi byte". */ + unsigned int accept_mb : 1; + /* If this state has backreference node(s). */ + unsigned int has_backref : 1; + unsigned int has_constraint : 1; +}; +typedef struct re_dfastate_t re_dfastate_t; + +struct re_state_table_entry +{ + Idx num; + Idx alloc; + re_dfastate_t **array; +}; + +/* Array type used in re_sub_match_last_t and re_sub_match_top_t. */ + +typedef struct +{ + Idx next_idx; + Idx alloc; + re_dfastate_t **array; +} state_array_t; + +/* Store information about the node NODE whose type is OP_CLOSE_SUBEXP. */ + +typedef struct +{ + Idx node; + Idx str_idx; /* The position NODE match at. */ + state_array_t path; +} re_sub_match_last_t; + +/* Store information about the node NODE whose type is OP_OPEN_SUBEXP. + And information about the node, whose type is OP_CLOSE_SUBEXP, + corresponding to NODE is stored in LASTS. */ + +typedef struct +{ + Idx str_idx; + Idx node; + state_array_t *path; + Idx alasts; /* Allocation size of LASTS. */ + Idx nlasts; /* The number of LASTS. */ + re_sub_match_last_t **lasts; +} re_sub_match_top_t; + +struct re_backref_cache_entry +{ + Idx node; + Idx str_idx; + Idx subexp_from; + Idx subexp_to; + char more; + char unused; + unsigned short int eps_reachable_subexps_map; +}; + +typedef struct +{ + /* The string object corresponding to the input string. */ + re_string_t input; +#if defined _LIBC || (defined __STDC_VERSION__ && __STDC_VERSION__ >= 199901L) + const re_dfa_t *const dfa; +#else + const re_dfa_t *dfa; +#endif + /* EFLAGS of the argument of regexec. */ + int eflags; + /* Where the matching ends. */ + Idx match_last; + Idx last_node; + /* The state log used by the matcher. */ + re_dfastate_t **state_log; + Idx state_log_top; + /* Back reference cache. */ + Idx nbkref_ents; + Idx abkref_ents; + struct re_backref_cache_entry *bkref_ents; + int max_mb_elem_len; + Idx nsub_tops; + Idx asub_tops; + re_sub_match_top_t **sub_tops; +} re_match_context_t; + +typedef struct +{ + re_dfastate_t **sifted_states; + re_dfastate_t **limited_states; + Idx last_node; + Idx last_str_idx; + re_node_set limits; +} re_sift_context_t; + +struct re_fail_stack_ent_t +{ + Idx idx; + Idx node; + regmatch_t *regs; + re_node_set eps_via_nodes; +}; + +struct re_fail_stack_t +{ + Idx num; + Idx alloc; + struct re_fail_stack_ent_t *stack; +}; + +struct re_dfa_t +{ + re_token_t *nodes; + size_t nodes_alloc; + size_t nodes_len; + Idx *nexts; + Idx *org_indices; + re_node_set *edests; + re_node_set *eclosures; + re_node_set *inveclosures; + struct re_state_table_entry *state_table; + re_dfastate_t *init_state; + re_dfastate_t *init_state_word; + re_dfastate_t *init_state_nl; + re_dfastate_t *init_state_begbuf; + bin_tree_t *str_tree; + bin_tree_storage_t *str_tree_storage; + re_bitset_ptr_t sb_char; + int str_tree_storage_idx; + + /* number of subexpressions 're_nsub' is in regex_t. */ + re_hashval_t state_hash_mask; + Idx init_node; + Idx nbackref; /* The number of backreference in this dfa. */ + + /* Bitmap expressing which backreference is used. */ + bitset_word_t used_bkref_map; + bitset_word_t completed_bkref_map; + + unsigned int has_plural_match : 1; + /* If this dfa has "multibyte node", which is a backreference or + a node which can accept multibyte character or multi character + collating element. */ + unsigned int has_mb_node : 1; + unsigned int is_utf8 : 1; + unsigned int map_notascii : 1; + unsigned int word_ops_used : 1; + int mb_cur_max; + bitset_t word_char; + reg_syntax_t syntax; + Idx *subexp_map; +#ifdef DEBUG + char* re_str; +#endif + lock_define (lock) +}; + +#define re_node_set_init_empty(set) memset (set, '\0', sizeof (re_node_set)) +#define re_node_set_remove(set,id) \ + (re_node_set_remove_at (set, re_node_set_contains (set, id) - 1)) +#define re_node_set_empty(p) ((p)->nelem = 0) +#define re_node_set_free(set) re_free ((set)->elems) + + +typedef enum +{ + SB_CHAR, + MB_CHAR, + EQUIV_CLASS, + COLL_SYM, + CHAR_CLASS +} bracket_elem_type; + +typedef struct +{ + bracket_elem_type type; + union + { + unsigned char ch; + unsigned char *name; + wchar_t wch; + } opr; +} bracket_elem_t; + + +/* Functions for bitset_t operation. */ + +static inline void +bitset_set (bitset_t set, Idx i) +{ + set[i / BITSET_WORD_BITS] |= (bitset_word_t) 1 << i % BITSET_WORD_BITS; +} + +static inline void +bitset_clear (bitset_t set, Idx i) +{ + set[i / BITSET_WORD_BITS] &= ~ ((bitset_word_t) 1 << i % BITSET_WORD_BITS); +} + +static inline bool +bitset_contain (const bitset_t set, Idx i) +{ + return (set[i / BITSET_WORD_BITS] >> i % BITSET_WORD_BITS) & 1; +} + +static inline void +bitset_empty (bitset_t set) +{ + memset (set, '\0', sizeof (bitset_t)); +} + +static inline void +bitset_set_all (bitset_t set) +{ + memset (set, -1, sizeof (bitset_word_t) * (SBC_MAX / BITSET_WORD_BITS)); + if (SBC_MAX % BITSET_WORD_BITS != 0) + set[BITSET_WORDS - 1] = + ((bitset_word_t) 1 << SBC_MAX % BITSET_WORD_BITS) - 1; +} + +static inline void +bitset_copy (bitset_t dest, const bitset_t src) +{ + memcpy (dest, src, sizeof (bitset_t)); +} + +static inline void +bitset_not (bitset_t set) +{ + int bitset_i; + for (bitset_i = 0; bitset_i < SBC_MAX / BITSET_WORD_BITS; ++bitset_i) + set[bitset_i] = ~set[bitset_i]; + if (SBC_MAX % BITSET_WORD_BITS != 0) + set[BITSET_WORDS - 1] = + ((((bitset_word_t) 1 << SBC_MAX % BITSET_WORD_BITS) - 1) + & ~set[BITSET_WORDS - 1]); +} + +static inline void +bitset_merge (bitset_t dest, const bitset_t src) +{ + int bitset_i; + for (bitset_i = 0; bitset_i < BITSET_WORDS; ++bitset_i) + dest[bitset_i] |= src[bitset_i]; +} + +static inline void +bitset_mask (bitset_t dest, const bitset_t src) +{ + int bitset_i; + for (bitset_i = 0; bitset_i < BITSET_WORDS; ++bitset_i) + dest[bitset_i] &= src[bitset_i]; +} + +#ifdef RE_ENABLE_I18N +/* Functions for re_string. */ +static int +__attribute__ ((pure, unused)) +re_string_char_size_at (const re_string_t *pstr, Idx idx) +{ + int byte_idx; + if (pstr->mb_cur_max == 1) + return 1; + for (byte_idx = 1; idx + byte_idx < pstr->valid_len; ++byte_idx) + if (pstr->wcs[idx + byte_idx] != WEOF) + break; + return byte_idx; +} + +static wint_t +__attribute__ ((pure, unused)) +re_string_wchar_at (const re_string_t *pstr, Idx idx) +{ + if (pstr->mb_cur_max == 1) + return (wint_t) pstr->mbs[idx]; + return (wint_t) pstr->wcs[idx]; +} + +# ifdef _LIBC +# include <locale/weight.h> +# endif + +static int +__attribute__ ((pure, unused)) +re_string_elem_size_at (const re_string_t *pstr, Idx idx) +{ +# ifdef _LIBC + const unsigned char *p, *extra; + const int32_t *table, *indirect; + uint_fast32_t nrules = _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES); + + if (nrules != 0) + { + table = (const int32_t *) _NL_CURRENT (LC_COLLATE, _NL_COLLATE_TABLEMB); + extra = (const unsigned char *) + _NL_CURRENT (LC_COLLATE, _NL_COLLATE_EXTRAMB); + indirect = (const int32_t *) _NL_CURRENT (LC_COLLATE, + _NL_COLLATE_INDIRECTMB); + p = pstr->mbs + idx; + findidx (table, indirect, extra, &p, pstr->len - idx); + return p - pstr->mbs - idx; + } + else +# endif /* _LIBC */ + return 1; +} +#endif /* RE_ENABLE_I18N */ + +#ifndef __GNUC_PREREQ +# if defined __GNUC__ && defined __GNUC_MINOR__ +# define __GNUC_PREREQ(maj, min) \ + ((__GNUC__ << 16) + __GNUC_MINOR__ >= ((maj) << 16) + (min)) +# else +# define __GNUC_PREREQ(maj, min) 0 +# endif +#endif + +#if __GNUC_PREREQ (3,4) +# undef __attribute_warn_unused_result__ +# define __attribute_warn_unused_result__ \ + __attribute__ ((__warn_unused_result__)) +#else +# define __attribute_warn_unused_result__ /* empty */ +#endif + +#ifndef FALLTHROUGH +# if __GNUC__ < 7 +# define FALLTHROUGH ((void) 0) +# else +# define FALLTHROUGH __attribute__ ((__fallthrough__)) +# endif +#endif + +#endif /* _REGEX_INTERNAL_H */ diff --git a/lib/regexec.c b/lib/regexec.c new file mode 100644 index 0000000..c7d099c --- /dev/null +++ b/lib/regexec.c @@ -0,0 +1,4334 @@ +/* Extended regular expression matching and search library. + Copyright (C) 2002-2018 Free Software Foundation, Inc. + This file is part of the GNU C Library. + Contributed by Isamu Hasegawa <isamu@yamato.ibm.com>. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU General Public + License as published by the Free Software Foundation; either + version 3 of the License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + General Public License for more details. + + You should have received a copy of the GNU General Public + License along with the GNU C Library; if not, see + <https://www.gnu.org/licenses/>. */ + +static reg_errcode_t match_ctx_init (re_match_context_t *cache, int eflags, + Idx n); +static void match_ctx_clean (re_match_context_t *mctx); +static void match_ctx_free (re_match_context_t *cache); +static reg_errcode_t match_ctx_add_entry (re_match_context_t *cache, Idx node, + Idx str_idx, Idx from, Idx to); +static Idx search_cur_bkref_entry (const re_match_context_t *mctx, Idx str_idx); +static reg_errcode_t match_ctx_add_subtop (re_match_context_t *mctx, Idx node, + Idx str_idx); +static re_sub_match_last_t * match_ctx_add_sublast (re_sub_match_top_t *subtop, + Idx node, Idx str_idx); +static void sift_ctx_init (re_sift_context_t *sctx, re_dfastate_t **sifted_sts, + re_dfastate_t **limited_sts, Idx last_node, + Idx last_str_idx); +static reg_errcode_t re_search_internal (const regex_t *preg, + const char *string, Idx length, + Idx start, Idx last_start, Idx stop, + size_t nmatch, regmatch_t pmatch[], + int eflags); +static regoff_t re_search_2_stub (struct re_pattern_buffer *bufp, + const char *string1, Idx length1, + const char *string2, Idx length2, + Idx start, regoff_t range, + struct re_registers *regs, + Idx stop, bool ret_len); +static regoff_t re_search_stub (struct re_pattern_buffer *bufp, + const char *string, Idx length, Idx start, + regoff_t range, Idx stop, + struct re_registers *regs, + bool ret_len); +static unsigned re_copy_regs (struct re_registers *regs, regmatch_t *pmatch, + Idx nregs, int regs_allocated); +static reg_errcode_t prune_impossible_nodes (re_match_context_t *mctx); +static Idx check_matching (re_match_context_t *mctx, bool fl_longest_match, + Idx *p_match_first); +static Idx check_halt_state_context (const re_match_context_t *mctx, + const re_dfastate_t *state, Idx idx); +static void update_regs (const re_dfa_t *dfa, regmatch_t *pmatch, + regmatch_t *prev_idx_match, Idx cur_node, + Idx cur_idx, Idx nmatch); +static reg_errcode_t push_fail_stack (struct re_fail_stack_t *fs, + Idx str_idx, Idx dest_node, Idx nregs, + regmatch_t *regs, + re_node_set *eps_via_nodes); +static reg_errcode_t set_regs (const regex_t *preg, + const re_match_context_t *mctx, + size_t nmatch, regmatch_t *pmatch, + bool fl_backtrack); +static reg_errcode_t free_fail_stack_return (struct re_fail_stack_t *fs); + +#ifdef RE_ENABLE_I18N +static int sift_states_iter_mb (const re_match_context_t *mctx, + re_sift_context_t *sctx, + Idx node_idx, Idx str_idx, Idx max_str_idx); +#endif /* RE_ENABLE_I18N */ +static reg_errcode_t sift_states_backward (const re_match_context_t *mctx, + re_sift_context_t *sctx); +static reg_errcode_t build_sifted_states (const re_match_context_t *mctx, + re_sift_context_t *sctx, Idx str_idx, + re_node_set *cur_dest); +static reg_errcode_t update_cur_sifted_state (const re_match_context_t *mctx, + re_sift_context_t *sctx, + Idx str_idx, + re_node_set *dest_nodes); +static reg_errcode_t add_epsilon_src_nodes (const re_dfa_t *dfa, + re_node_set *dest_nodes, + const re_node_set *candidates); +static bool check_dst_limits (const re_match_context_t *mctx, + const re_node_set *limits, + Idx dst_node, Idx dst_idx, Idx src_node, + Idx src_idx); +static int check_dst_limits_calc_pos_1 (const re_match_context_t *mctx, + int boundaries, Idx subexp_idx, + Idx from_node, Idx bkref_idx); +static int check_dst_limits_calc_pos (const re_match_context_t *mctx, + Idx limit, Idx subexp_idx, + Idx node, Idx str_idx, + Idx bkref_idx); +static reg_errcode_t check_subexp_limits (const re_dfa_t *dfa, + re_node_set *dest_nodes, + const re_node_set *candidates, + re_node_set *limits, + struct re_backref_cache_entry *bkref_ents, + Idx str_idx); +static reg_errcode_t sift_states_bkref (const re_match_context_t *mctx, + re_sift_context_t *sctx, + Idx str_idx, const re_node_set *candidates); +static reg_errcode_t merge_state_array (const re_dfa_t *dfa, + re_dfastate_t **dst, + re_dfastate_t **src, Idx num); +static re_dfastate_t *find_recover_state (reg_errcode_t *err, + re_match_context_t *mctx); +static re_dfastate_t *transit_state (reg_errcode_t *err, + re_match_context_t *mctx, + re_dfastate_t *state); +static re_dfastate_t *merge_state_with_log (reg_errcode_t *err, + re_match_context_t *mctx, + re_dfastate_t *next_state); +static reg_errcode_t check_subexp_matching_top (re_match_context_t *mctx, + re_node_set *cur_nodes, + Idx str_idx); +#if 0 +static re_dfastate_t *transit_state_sb (reg_errcode_t *err, + re_match_context_t *mctx, + re_dfastate_t *pstate); +#endif +#ifdef RE_ENABLE_I18N +static reg_errcode_t transit_state_mb (re_match_context_t *mctx, + re_dfastate_t *pstate); +#endif /* RE_ENABLE_I18N */ +static reg_errcode_t transit_state_bkref (re_match_context_t *mctx, + const re_node_set *nodes); +static reg_errcode_t get_subexp (re_match_context_t *mctx, + Idx bkref_node, Idx bkref_str_idx); +static reg_errcode_t get_subexp_sub (re_match_context_t *mctx, + const re_sub_match_top_t *sub_top, + re_sub_match_last_t *sub_last, + Idx bkref_node, Idx bkref_str); +static Idx find_subexp_node (const re_dfa_t *dfa, const re_node_set *nodes, + Idx subexp_idx, int type); +static reg_errcode_t check_arrival (re_match_context_t *mctx, + state_array_t *path, Idx top_node, + Idx top_str, Idx last_node, Idx last_str, + int type); +static reg_errcode_t check_arrival_add_next_nodes (re_match_context_t *mctx, + Idx str_idx, + re_node_set *cur_nodes, + re_node_set *next_nodes); +static reg_errcode_t check_arrival_expand_ecl (const re_dfa_t *dfa, + re_node_set *cur_nodes, + Idx ex_subexp, int type); +static reg_errcode_t check_arrival_expand_ecl_sub (const re_dfa_t *dfa, + re_node_set *dst_nodes, + Idx target, Idx ex_subexp, + int type); +static reg_errcode_t expand_bkref_cache (re_match_context_t *mctx, + re_node_set *cur_nodes, Idx cur_str, + Idx subexp_num, int type); +static bool build_trtable (const re_dfa_t *dfa, re_dfastate_t *state); +#ifdef RE_ENABLE_I18N +static int check_node_accept_bytes (const re_dfa_t *dfa, Idx node_idx, + const re_string_t *input, Idx idx); +# ifdef _LIBC +static unsigned int find_collation_sequence_value (const unsigned char *mbs, + size_t name_len); +# endif /* _LIBC */ +#endif /* RE_ENABLE_I18N */ +static Idx group_nodes_into_DFAstates (const re_dfa_t *dfa, + const re_dfastate_t *state, + re_node_set *states_node, + bitset_t *states_ch); +static bool check_node_accept (const re_match_context_t *mctx, + const re_token_t *node, Idx idx); +static reg_errcode_t extend_buffers (re_match_context_t *mctx, int min_len); + +/* Entry point for POSIX code. */ + +/* regexec searches for a given pattern, specified by PREG, in the + string STRING. + + If NMATCH is zero or REG_NOSUB was set in the cflags argument to + 'regcomp', we ignore PMATCH. Otherwise, we assume PMATCH has at + least NMATCH elements, and we set them to the offsets of the + corresponding matched substrings. + + EFLAGS specifies "execution flags" which affect matching: if + REG_NOTBOL is set, then ^ does not match at the beginning of the + string; if REG_NOTEOL is set, then $ does not match at the end. + + We return 0 if we find a match and REG_NOMATCH if not. */ + +int +regexec (const regex_t *_Restrict_ preg, const char *_Restrict_ string, + size_t nmatch, regmatch_t pmatch[], int eflags) +{ + reg_errcode_t err; + Idx start, length; + re_dfa_t *dfa = preg->buffer; + + if (eflags & ~(REG_NOTBOL | REG_NOTEOL | REG_STARTEND)) + return REG_BADPAT; + + if (eflags & REG_STARTEND) + { + start = pmatch[0].rm_so; + length = pmatch[0].rm_eo; + } + else + { + start = 0; + length = strlen (string); + } + + lock_lock (dfa->lock); + if (preg->no_sub) + err = re_search_internal (preg, string, length, start, length, + length, 0, NULL, eflags); + else + err = re_search_internal (preg, string, length, start, length, + length, nmatch, pmatch, eflags); + lock_unlock (dfa->lock); + return err != REG_NOERROR; +} + +#ifdef _LIBC +libc_hidden_def (__regexec) + +# include <shlib-compat.h> +versioned_symbol (libc, __regexec, regexec, GLIBC_2_3_4); + +# if SHLIB_COMPAT (libc, GLIBC_2_0, GLIBC_2_3_4) +__typeof__ (__regexec) __compat_regexec; + +int +attribute_compat_text_section +__compat_regexec (const regex_t *_Restrict_ preg, + const char *_Restrict_ string, size_t nmatch, + regmatch_t pmatch[], int eflags) +{ + return regexec (preg, string, nmatch, pmatch, + eflags & (REG_NOTBOL | REG_NOTEOL)); +} +compat_symbol (libc, __compat_regexec, regexec, GLIBC_2_0); +# endif +#endif + +/* Entry points for GNU code. */ + +/* re_match, re_search, re_match_2, re_search_2 + + The former two functions operate on STRING with length LENGTH, + while the later two operate on concatenation of STRING1 and STRING2 + with lengths LENGTH1 and LENGTH2, respectively. + + re_match() matches the compiled pattern in BUFP against the string, + starting at index START. + + re_search() first tries matching at index START, then it tries to match + starting from index START + 1, and so on. The last start position tried + is START + RANGE. (Thus RANGE = 0 forces re_search to operate the same + way as re_match().) + + The parameter STOP of re_{match,search}_2 specifies that no match exceeding + the first STOP characters of the concatenation of the strings should be + concerned. + + If REGS is not NULL, and BUFP->no_sub is not set, the offsets of the match + and all groups is stored in REGS. (For the "_2" variants, the offsets are + computed relative to the concatenation, not relative to the individual + strings.) + + On success, re_match* functions return the length of the match, re_search* + return the position of the start of the match. Return value -1 means no + match was found and -2 indicates an internal error. */ + +regoff_t +re_match (struct re_pattern_buffer *bufp, const char *string, Idx length, + Idx start, struct re_registers *regs) +{ + return re_search_stub (bufp, string, length, start, 0, length, regs, true); +} +#ifdef _LIBC +weak_alias (__re_match, re_match) +#endif + +regoff_t +re_search (struct re_pattern_buffer *bufp, const char *string, Idx length, + Idx start, regoff_t range, struct re_registers *regs) +{ + return re_search_stub (bufp, string, length, start, range, length, regs, + false); +} +#ifdef _LIBC +weak_alias (__re_search, re_search) +#endif + +regoff_t +re_match_2 (struct re_pattern_buffer *bufp, const char *string1, Idx length1, + const char *string2, Idx length2, Idx start, + struct re_registers *regs, Idx stop) +{ + return re_search_2_stub (bufp, string1, length1, string2, length2, + start, 0, regs, stop, true); +} +#ifdef _LIBC +weak_alias (__re_match_2, re_match_2) +#endif + +regoff_t +re_search_2 (struct re_pattern_buffer *bufp, const char *string1, Idx length1, + const char *string2, Idx length2, Idx start, regoff_t range, + struct re_registers *regs, Idx stop) +{ + return re_search_2_stub (bufp, string1, length1, string2, length2, + start, range, regs, stop, false); +} +#ifdef _LIBC +weak_alias (__re_search_2, re_search_2) +#endif + +static regoff_t +re_search_2_stub (struct re_pattern_buffer *bufp, const char *string1, + Idx length1, const char *string2, Idx length2, Idx start, + regoff_t range, struct re_registers *regs, + Idx stop, bool ret_len) +{ + const char *str; + regoff_t rval; + Idx len; + char *s = NULL; + + if (__glibc_unlikely ((length1 < 0 || length2 < 0 || stop < 0 + || INT_ADD_WRAPV (length1, length2, &len)))) + return -2; + + /* Concatenate the strings. */ + if (length2 > 0) + if (length1 > 0) + { + s = re_malloc (char, len); + + if (__glibc_unlikely (s == NULL)) + return -2; +#ifdef _LIBC + memcpy (__mempcpy (s, string1, length1), string2, length2); +#else + memcpy (s, string1, length1); + memcpy (s + length1, string2, length2); +#endif + str = s; + } + else + str = string2; + else + str = string1; + + rval = re_search_stub (bufp, str, len, start, range, stop, regs, + ret_len); + re_free (s); + return rval; +} + +/* The parameters have the same meaning as those of re_search. + Additional parameters: + If RET_LEN is true the length of the match is returned (re_match style); + otherwise the position of the match is returned. */ + +static regoff_t +re_search_stub (struct re_pattern_buffer *bufp, const char *string, Idx length, + Idx start, regoff_t range, Idx stop, struct re_registers *regs, + bool ret_len) +{ + reg_errcode_t result; + regmatch_t *pmatch; + Idx nregs; + regoff_t rval; + int eflags = 0; + re_dfa_t *dfa = bufp->buffer; + Idx last_start = start + range; + + /* Check for out-of-range. */ + if (__glibc_unlikely (start < 0 || start > length)) + return -1; + if (__glibc_unlikely (length < last_start + || (0 <= range && last_start < start))) + last_start = length; + else if (__glibc_unlikely (last_start < 0 + || (range < 0 && start <= last_start))) + last_start = 0; + + lock_lock (dfa->lock); + + eflags |= (bufp->not_bol) ? REG_NOTBOL : 0; + eflags |= (bufp->not_eol) ? REG_NOTEOL : 0; + + /* Compile fastmap if we haven't yet. */ + if (start < last_start && bufp->fastmap != NULL && !bufp->fastmap_accurate) + re_compile_fastmap (bufp); + + if (__glibc_unlikely (bufp->no_sub)) + regs = NULL; + + /* We need at least 1 register. */ + if (regs == NULL) + nregs = 1; + else if (__glibc_unlikely (bufp->regs_allocated == REGS_FIXED + && regs->num_regs <= bufp->re_nsub)) + { + nregs = regs->num_regs; + if (__glibc_unlikely (nregs < 1)) + { + /* Nothing can be copied to regs. */ + regs = NULL; + nregs = 1; + } + } + else + nregs = bufp->re_nsub + 1; + pmatch = re_malloc (regmatch_t, nregs); + if (__glibc_unlikely (pmatch == NULL)) + { + rval = -2; + goto out; + } + + result = re_search_internal (bufp, string, length, start, last_start, stop, + nregs, pmatch, eflags); + + rval = 0; + + /* I hope we needn't fill their regs with -1's when no match was found. */ + if (result != REG_NOERROR) + rval = result == REG_NOMATCH ? -1 : -2; + else if (regs != NULL) + { + /* If caller wants register contents data back, copy them. */ + bufp->regs_allocated = re_copy_regs (regs, pmatch, nregs, + bufp->regs_allocated); + if (__glibc_unlikely (bufp->regs_allocated == REGS_UNALLOCATED)) + rval = -2; + } + + if (__glibc_likely (rval == 0)) + { + if (ret_len) + { + assert (pmatch[0].rm_so == start); + rval = pmatch[0].rm_eo - start; + } + else + rval = pmatch[0].rm_so; + } + re_free (pmatch); + out: + lock_unlock (dfa->lock); + return rval; +} + +static unsigned +re_copy_regs (struct re_registers *regs, regmatch_t *pmatch, Idx nregs, + int regs_allocated) +{ + int rval = REGS_REALLOCATE; + Idx i; + Idx need_regs = nregs + 1; + /* We need one extra element beyond 'num_regs' for the '-1' marker GNU code + uses. */ + + /* Have the register data arrays been allocated? */ + if (regs_allocated == REGS_UNALLOCATED) + { /* No. So allocate them with malloc. */ + regs->start = re_malloc (regoff_t, need_regs); + if (__glibc_unlikely (regs->start == NULL)) + return REGS_UNALLOCATED; + regs->end = re_malloc (regoff_t, need_regs); + if (__glibc_unlikely (regs->end == NULL)) + { + re_free (regs->start); + return REGS_UNALLOCATED; + } + regs->num_regs = need_regs; + } + else if (regs_allocated == REGS_REALLOCATE) + { /* Yes. If we need more elements than were already + allocated, reallocate them. If we need fewer, just + leave it alone. */ + if (__glibc_unlikely (need_regs > regs->num_regs)) + { + regoff_t *new_start = re_realloc (regs->start, regoff_t, need_regs); + regoff_t *new_end; + if (__glibc_unlikely (new_start == NULL)) + return REGS_UNALLOCATED; + new_end = re_realloc (regs->end, regoff_t, need_regs); + if (__glibc_unlikely (new_end == NULL)) + { + re_free (new_start); + return REGS_UNALLOCATED; + } + regs->start = new_start; + regs->end = new_end; + regs->num_regs = need_regs; + } + } + else + { + assert (regs_allocated == REGS_FIXED); + /* This function may not be called with REGS_FIXED and nregs too big. */ + assert (regs->num_regs >= nregs); + rval = REGS_FIXED; + } + + /* Copy the regs. */ + for (i = 0; i < nregs; ++i) + { + regs->start[i] = pmatch[i].rm_so; + regs->end[i] = pmatch[i].rm_eo; + } + for ( ; i < regs->num_regs; ++i) + regs->start[i] = regs->end[i] = -1; + + return rval; +} + +/* Set REGS to hold NUM_REGS registers, storing them in STARTS and + ENDS. Subsequent matches using PATTERN_BUFFER and REGS will use + this memory for recording register information. STARTS and ENDS + must be allocated using the malloc library routine, and must each + be at least NUM_REGS * sizeof (regoff_t) bytes long. + + If NUM_REGS == 0, then subsequent matches should allocate their own + register data. + + Unless this function is called, the first search or match using + PATTERN_BUFFER will allocate its own register data, without + freeing the old data. */ + +void +re_set_registers (struct re_pattern_buffer *bufp, struct re_registers *regs, + __re_size_t num_regs, regoff_t *starts, regoff_t *ends) +{ + if (num_regs) + { + bufp->regs_allocated = REGS_REALLOCATE; + regs->num_regs = num_regs; + regs->start = starts; + regs->end = ends; + } + else + { + bufp->regs_allocated = REGS_UNALLOCATED; + regs->num_regs = 0; + regs->start = regs->end = NULL; + } +} +#ifdef _LIBC +weak_alias (__re_set_registers, re_set_registers) +#endif + +/* Entry points compatible with 4.2 BSD regex library. We don't define + them unless specifically requested. */ + +#if defined _REGEX_RE_COMP || defined _LIBC +int +# ifdef _LIBC +weak_function +# endif +re_exec (const char *s) +{ + return 0 == regexec (&re_comp_buf, s, 0, NULL, 0); +} +#endif /* _REGEX_RE_COMP */ + +/* Internal entry point. */ + +/* Searches for a compiled pattern PREG in the string STRING, whose + length is LENGTH. NMATCH, PMATCH, and EFLAGS have the same + meaning as with regexec. LAST_START is START + RANGE, where + START and RANGE have the same meaning as with re_search. + Return REG_NOERROR if we find a match, and REG_NOMATCH if not, + otherwise return the error code. + Note: We assume front end functions already check ranges. + (0 <= LAST_START && LAST_START <= LENGTH) */ + +static reg_errcode_t +__attribute_warn_unused_result__ +re_search_internal (const regex_t *preg, const char *string, Idx length, + Idx start, Idx last_start, Idx stop, size_t nmatch, + regmatch_t pmatch[], int eflags) +{ + reg_errcode_t err; + const re_dfa_t *dfa = preg->buffer; + Idx left_lim, right_lim; + int incr; + bool fl_longest_match; + int match_kind; + Idx match_first; + Idx match_last = -1; + Idx extra_nmatch; + bool sb; + int ch; +#if defined _LIBC || (defined __STDC_VERSION__ && __STDC_VERSION__ >= 199901L) + re_match_context_t mctx = { .dfa = dfa }; +#else + re_match_context_t mctx; +#endif + char *fastmap = ((preg->fastmap != NULL && preg->fastmap_accurate + && start != last_start && !preg->can_be_null) + ? preg->fastmap : NULL); + RE_TRANSLATE_TYPE t = preg->translate; + +#if !(defined _LIBC || (defined __STDC_VERSION__ && __STDC_VERSION__ >= 199901L)) + memset (&mctx, '\0', sizeof (re_match_context_t)); + mctx.dfa = dfa; +#endif + + extra_nmatch = (nmatch > preg->re_nsub) ? nmatch - (preg->re_nsub + 1) : 0; + nmatch -= extra_nmatch; + + /* Check if the DFA haven't been compiled. */ + if (__glibc_unlikely (preg->used == 0 || dfa->init_state == NULL + || dfa->init_state_word == NULL + || dfa->init_state_nl == NULL + || dfa->init_state_begbuf == NULL)) + return REG_NOMATCH; + +#ifdef DEBUG + /* We assume front-end functions already check them. */ + assert (0 <= last_start && last_start <= length); +#endif + + /* If initial states with non-begbuf contexts have no elements, + the regex must be anchored. If preg->newline_anchor is set, + we'll never use init_state_nl, so do not check it. */ + if (dfa->init_state->nodes.nelem == 0 + && dfa->init_state_word->nodes.nelem == 0 + && (dfa->init_state_nl->nodes.nelem == 0 + || !preg->newline_anchor)) + { + if (start != 0 && last_start != 0) + return REG_NOMATCH; + start = last_start = 0; + } + + /* We must check the longest matching, if nmatch > 0. */ + fl_longest_match = (nmatch != 0 || dfa->nbackref); + + err = re_string_allocate (&mctx.input, string, length, dfa->nodes_len + 1, + preg->translate, (preg->syntax & RE_ICASE) != 0, + dfa); + if (__glibc_unlikely (err != REG_NOERROR)) + goto free_return; + mctx.input.stop = stop; + mctx.input.raw_stop = stop; + mctx.input.newline_anchor = preg->newline_anchor; + + err = match_ctx_init (&mctx, eflags, dfa->nbackref * 2); + if (__glibc_unlikely (err != REG_NOERROR)) + goto free_return; + + /* We will log all the DFA states through which the dfa pass, + if nmatch > 1, or this dfa has "multibyte node", which is a + back-reference or a node which can accept multibyte character or + multi character collating element. */ + if (nmatch > 1 || dfa->has_mb_node) + { + /* Avoid overflow. */ + if (__glibc_unlikely ((MIN (IDX_MAX, SIZE_MAX / sizeof (re_dfastate_t *)) + <= mctx.input.bufs_len))) + { + err = REG_ESPACE; + goto free_return; + } + + mctx.state_log = re_malloc (re_dfastate_t *, mctx.input.bufs_len + 1); + if (__glibc_unlikely (mctx.state_log == NULL)) + { + err = REG_ESPACE; + goto free_return; + } + } + else + mctx.state_log = NULL; + + match_first = start; + mctx.input.tip_context = (eflags & REG_NOTBOL) ? CONTEXT_BEGBUF + : CONTEXT_NEWLINE | CONTEXT_BEGBUF; + + /* Check incrementally whether the input string matches. */ + incr = (last_start < start) ? -1 : 1; + left_lim = (last_start < start) ? last_start : start; + right_lim = (last_start < start) ? start : last_start; + sb = dfa->mb_cur_max == 1; + match_kind = + (fastmap + ? ((sb || !(preg->syntax & RE_ICASE || t) ? 4 : 0) + | (start <= last_start ? 2 : 0) + | (t != NULL ? 1 : 0)) + : 8); + + for (;; match_first += incr) + { + err = REG_NOMATCH; + if (match_first < left_lim || right_lim < match_first) + goto free_return; + + /* Advance as rapidly as possible through the string, until we + find a plausible place to start matching. This may be done + with varying efficiency, so there are various possibilities: + only the most common of them are specialized, in order to + save on code size. We use a switch statement for speed. */ + switch (match_kind) + { + case 8: + /* No fastmap. */ + break; + + case 7: + /* Fastmap with single-byte translation, match forward. */ + while (__glibc_likely (match_first < right_lim) + && !fastmap[t[(unsigned char) string[match_first]]]) + ++match_first; + goto forward_match_found_start_or_reached_end; + + case 6: + /* Fastmap without translation, match forward. */ + while (__glibc_likely (match_first < right_lim) + && !fastmap[(unsigned char) string[match_first]]) + ++match_first; + + forward_match_found_start_or_reached_end: + if (__glibc_unlikely (match_first == right_lim)) + { + ch = match_first >= length + ? 0 : (unsigned char) string[match_first]; + if (!fastmap[t ? t[ch] : ch]) + goto free_return; + } + break; + + case 4: + case 5: + /* Fastmap without multi-byte translation, match backwards. */ + while (match_first >= left_lim) + { + ch = match_first >= length + ? 0 : (unsigned char) string[match_first]; + if (fastmap[t ? t[ch] : ch]) + break; + --match_first; + } + if (match_first < left_lim) + goto free_return; + break; + + default: + /* In this case, we can't determine easily the current byte, + since it might be a component byte of a multibyte + character. Then we use the constructed buffer instead. */ + for (;;) + { + /* If MATCH_FIRST is out of the valid range, reconstruct the + buffers. */ + __re_size_t offset = match_first - mctx.input.raw_mbs_idx; + if (__glibc_unlikely (offset + >= (__re_size_t) mctx.input.valid_raw_len)) + { + err = re_string_reconstruct (&mctx.input, match_first, + eflags); + if (__glibc_unlikely (err != REG_NOERROR)) + goto free_return; + + offset = match_first - mctx.input.raw_mbs_idx; + } + /* If MATCH_FIRST is out of the buffer, leave it as '\0'. + Note that MATCH_FIRST must not be smaller than 0. */ + ch = (match_first >= length + ? 0 : re_string_byte_at (&mctx.input, offset)); + if (fastmap[ch]) + break; + match_first += incr; + if (match_first < left_lim || match_first > right_lim) + { + err = REG_NOMATCH; + goto free_return; + } + } + break; + } + + /* Reconstruct the buffers so that the matcher can assume that + the matching starts from the beginning of the buffer. */ + err = re_string_reconstruct (&mctx.input, match_first, eflags); + if (__glibc_unlikely (err != REG_NOERROR)) + goto free_return; + +#ifdef RE_ENABLE_I18N + /* Don't consider this char as a possible match start if it part, + yet isn't the head, of a multibyte character. */ + if (!sb && !re_string_first_byte (&mctx.input, 0)) + continue; +#endif + + /* It seems to be appropriate one, then use the matcher. */ + /* We assume that the matching starts from 0. */ + mctx.state_log_top = mctx.nbkref_ents = mctx.max_mb_elem_len = 0; + match_last = check_matching (&mctx, fl_longest_match, + start <= last_start ? &match_first : NULL); + if (match_last != -1) + { + if (__glibc_unlikely (match_last == -2)) + { + err = REG_ESPACE; + goto free_return; + } + else + { + mctx.match_last = match_last; + if ((!preg->no_sub && nmatch > 1) || dfa->nbackref) + { + re_dfastate_t *pstate = mctx.state_log[match_last]; + mctx.last_node = check_halt_state_context (&mctx, pstate, + match_last); + } + if ((!preg->no_sub && nmatch > 1 && dfa->has_plural_match) + || dfa->nbackref) + { + err = prune_impossible_nodes (&mctx); + if (err == REG_NOERROR) + break; + if (__glibc_unlikely (err != REG_NOMATCH)) + goto free_return; + match_last = -1; + } + else + break; /* We found a match. */ + } + } + + match_ctx_clean (&mctx); + } + +#ifdef DEBUG + assert (match_last != -1); + assert (err == REG_NOERROR); +#endif + + /* Set pmatch[] if we need. */ + if (nmatch > 0) + { + Idx reg_idx; + + /* Initialize registers. */ + for (reg_idx = 1; reg_idx < nmatch; ++reg_idx) + pmatch[reg_idx].rm_so = pmatch[reg_idx].rm_eo = -1; + + /* Set the points where matching start/end. */ + pmatch[0].rm_so = 0; + pmatch[0].rm_eo = mctx.match_last; + /* FIXME: This function should fail if mctx.match_last exceeds + the maximum possible regoff_t value. We need a new error + code REG_OVERFLOW. */ + + if (!preg->no_sub && nmatch > 1) + { + err = set_regs (preg, &mctx, nmatch, pmatch, + dfa->has_plural_match && dfa->nbackref > 0); + if (__glibc_unlikely (err != REG_NOERROR)) + goto free_return; + } + + /* At last, add the offset to each register, since we slid + the buffers so that we could assume that the matching starts + from 0. */ + for (reg_idx = 0; reg_idx < nmatch; ++reg_idx) + if (pmatch[reg_idx].rm_so != -1) + { +#ifdef RE_ENABLE_I18N + if (__glibc_unlikely (mctx.input.offsets_needed != 0)) + { + pmatch[reg_idx].rm_so = + (pmatch[reg_idx].rm_so == mctx.input.valid_len + ? mctx.input.valid_raw_len + : mctx.input.offsets[pmatch[reg_idx].rm_so]); + pmatch[reg_idx].rm_eo = + (pmatch[reg_idx].rm_eo == mctx.input.valid_len + ? mctx.input.valid_raw_len + : mctx.input.offsets[pmatch[reg_idx].rm_eo]); + } +#else + assert (mctx.input.offsets_needed == 0); +#endif + pmatch[reg_idx].rm_so += match_first; + pmatch[reg_idx].rm_eo += match_first; + } + for (reg_idx = 0; reg_idx < extra_nmatch; ++reg_idx) + { + pmatch[nmatch + reg_idx].rm_so = -1; + pmatch[nmatch + reg_idx].rm_eo = -1; + } + + if (dfa->subexp_map) + for (reg_idx = 0; reg_idx + 1 < nmatch; reg_idx++) + if (dfa->subexp_map[reg_idx] != reg_idx) + { + pmatch[reg_idx + 1].rm_so + = pmatch[dfa->subexp_map[reg_idx] + 1].rm_so; + pmatch[reg_idx + 1].rm_eo + = pmatch[dfa->subexp_map[reg_idx] + 1].rm_eo; + } + } + + free_return: + re_free (mctx.state_log); + if (dfa->nbackref) + match_ctx_free (&mctx); + re_string_destruct (&mctx.input); + return err; +} + +static reg_errcode_t +__attribute_warn_unused_result__ +prune_impossible_nodes (re_match_context_t *mctx) +{ + const re_dfa_t *const dfa = mctx->dfa; + Idx halt_node, match_last; + reg_errcode_t ret; + re_dfastate_t **sifted_states; + re_dfastate_t **lim_states = NULL; + re_sift_context_t sctx; +#ifdef DEBUG + assert (mctx->state_log != NULL); +#endif + match_last = mctx->match_last; + halt_node = mctx->last_node; + + /* Avoid overflow. */ + if (__glibc_unlikely (MIN (IDX_MAX, SIZE_MAX / sizeof (re_dfastate_t *)) + <= match_last)) + return REG_ESPACE; + + sifted_states = re_malloc (re_dfastate_t *, match_last + 1); + if (__glibc_unlikely (sifted_states == NULL)) + { + ret = REG_ESPACE; + goto free_return; + } + if (dfa->nbackref) + { + lim_states = re_malloc (re_dfastate_t *, match_last + 1); + if (__glibc_unlikely (lim_states == NULL)) + { + ret = REG_ESPACE; + goto free_return; + } + while (1) + { + memset (lim_states, '\0', + sizeof (re_dfastate_t *) * (match_last + 1)); + sift_ctx_init (&sctx, sifted_states, lim_states, halt_node, + match_last); + ret = sift_states_backward (mctx, &sctx); + re_node_set_free (&sctx.limits); + if (__glibc_unlikely (ret != REG_NOERROR)) + goto free_return; + if (sifted_states[0] != NULL || lim_states[0] != NULL) + break; + do + { + --match_last; + if (match_last < 0) + { + ret = REG_NOMATCH; + goto free_return; + } + } while (mctx->state_log[match_last] == NULL + || !mctx->state_log[match_last]->halt); + halt_node = check_halt_state_context (mctx, + mctx->state_log[match_last], + match_last); + } + ret = merge_state_array (dfa, sifted_states, lim_states, + match_last + 1); + re_free (lim_states); + lim_states = NULL; + if (__glibc_unlikely (ret != REG_NOERROR)) + goto free_return; + } + else + { + sift_ctx_init (&sctx, sifted_states, lim_states, halt_node, match_last); + ret = sift_states_backward (mctx, &sctx); + re_node_set_free (&sctx.limits); + if (__glibc_unlikely (ret != REG_NOERROR)) + goto free_return; + if (sifted_states[0] == NULL) + { + ret = REG_NOMATCH; + goto free_return; + } + } + re_free (mctx->state_log); + mctx->state_log = sifted_states; + sifted_states = NULL; + mctx->last_node = halt_node; + mctx->match_last = match_last; + ret = REG_NOERROR; + free_return: + re_free (sifted_states); + re_free (lim_states); + return ret; +} + +/* Acquire an initial state and return it. + We must select appropriate initial state depending on the context, + since initial states may have constraints like "\<", "^", etc.. */ + +static inline re_dfastate_t * +__attribute__ ((always_inline)) +acquire_init_state_context (reg_errcode_t *err, const re_match_context_t *mctx, + Idx idx) +{ + const re_dfa_t *const dfa = mctx->dfa; + if (dfa->init_state->has_constraint) + { + unsigned int context; + context = re_string_context_at (&mctx->input, idx - 1, mctx->eflags); + if (IS_WORD_CONTEXT (context)) + return dfa->init_state_word; + else if (IS_ORDINARY_CONTEXT (context)) + return dfa->init_state; + else if (IS_BEGBUF_CONTEXT (context) && IS_NEWLINE_CONTEXT (context)) + return dfa->init_state_begbuf; + else if (IS_NEWLINE_CONTEXT (context)) + return dfa->init_state_nl; + else if (IS_BEGBUF_CONTEXT (context)) + { + /* It is relatively rare case, then calculate on demand. */ + return re_acquire_state_context (err, dfa, + dfa->init_state->entrance_nodes, + context); + } + else + /* Must not happen? */ + return dfa->init_state; + } + else + return dfa->init_state; +} + +/* Check whether the regular expression match input string INPUT or not, + and return the index where the matching end. Return -1 if + there is no match, and return -2 in case of an error. + FL_LONGEST_MATCH means we want the POSIX longest matching. + If P_MATCH_FIRST is not NULL, and the match fails, it is set to the + next place where we may want to try matching. + Note that the matcher assumes that the matching starts from the current + index of the buffer. */ + +static Idx +__attribute_warn_unused_result__ +check_matching (re_match_context_t *mctx, bool fl_longest_match, + Idx *p_match_first) +{ + const re_dfa_t *const dfa = mctx->dfa; + reg_errcode_t err; + Idx match = 0; + Idx match_last = -1; + Idx cur_str_idx = re_string_cur_idx (&mctx->input); + re_dfastate_t *cur_state; + bool at_init_state = p_match_first != NULL; + Idx next_start_idx = cur_str_idx; + + err = REG_NOERROR; + cur_state = acquire_init_state_context (&err, mctx, cur_str_idx); + /* An initial state must not be NULL (invalid). */ + if (__glibc_unlikely (cur_state == NULL)) + { + assert (err == REG_ESPACE); + return -2; + } + + if (mctx->state_log != NULL) + { + mctx->state_log[cur_str_idx] = cur_state; + + /* Check OP_OPEN_SUBEXP in the initial state in case that we use them + later. E.g. Processing back references. */ + if (__glibc_unlikely (dfa->nbackref)) + { + at_init_state = false; + err = check_subexp_matching_top (mctx, &cur_state->nodes, 0); + if (__glibc_unlikely (err != REG_NOERROR)) + return err; + + if (cur_state->has_backref) + { + err = transit_state_bkref (mctx, &cur_state->nodes); + if (__glibc_unlikely (err != REG_NOERROR)) + return err; + } + } + } + + /* If the RE accepts NULL string. */ + if (__glibc_unlikely (cur_state->halt)) + { + if (!cur_state->has_constraint + || check_halt_state_context (mctx, cur_state, cur_str_idx)) + { + if (!fl_longest_match) + return cur_str_idx; + else + { + match_last = cur_str_idx; + match = 1; + } + } + } + + while (!re_string_eoi (&mctx->input)) + { + re_dfastate_t *old_state = cur_state; + Idx next_char_idx = re_string_cur_idx (&mctx->input) + 1; + + if ((__glibc_unlikely (next_char_idx >= mctx->input.bufs_len) + && mctx->input.bufs_len < mctx->input.len) + || (__glibc_unlikely (next_char_idx >= mctx->input.valid_len) + && mctx->input.valid_len < mctx->input.len)) + { + err = extend_buffers (mctx, next_char_idx + 1); + if (__glibc_unlikely (err != REG_NOERROR)) + { + assert (err == REG_ESPACE); + return -2; + } + } + + cur_state = transit_state (&err, mctx, cur_state); + if (mctx->state_log != NULL) + cur_state = merge_state_with_log (&err, mctx, cur_state); + + if (cur_state == NULL) + { + /* Reached the invalid state or an error. Try to recover a valid + state using the state log, if available and if we have not + already found a valid (even if not the longest) match. */ + if (__glibc_unlikely (err != REG_NOERROR)) + return -2; + + if (mctx->state_log == NULL + || (match && !fl_longest_match) + || (cur_state = find_recover_state (&err, mctx)) == NULL) + break; + } + + if (__glibc_unlikely (at_init_state)) + { + if (old_state == cur_state) + next_start_idx = next_char_idx; + else + at_init_state = false; + } + + if (cur_state->halt) + { + /* Reached a halt state. + Check the halt state can satisfy the current context. */ + if (!cur_state->has_constraint + || check_halt_state_context (mctx, cur_state, + re_string_cur_idx (&mctx->input))) + { + /* We found an appropriate halt state. */ + match_last = re_string_cur_idx (&mctx->input); + match = 1; + + /* We found a match, do not modify match_first below. */ + p_match_first = NULL; + if (!fl_longest_match) + break; + } + } + } + + if (p_match_first) + *p_match_first += next_start_idx; + + return match_last; +} + +/* Check NODE match the current context. */ + +static bool +check_halt_node_context (const re_dfa_t *dfa, Idx node, unsigned int context) +{ + re_token_type_t type = dfa->nodes[node].type; + unsigned int constraint = dfa->nodes[node].constraint; + if (type != END_OF_RE) + return false; + if (!constraint) + return true; + if (NOT_SATISFY_NEXT_CONSTRAINT (constraint, context)) + return false; + return true; +} + +/* Check the halt state STATE match the current context. + Return 0 if not match, if the node, STATE has, is a halt node and + match the context, return the node. */ + +static Idx +check_halt_state_context (const re_match_context_t *mctx, + const re_dfastate_t *state, Idx idx) +{ + Idx i; + unsigned int context; +#ifdef DEBUG + assert (state->halt); +#endif + context = re_string_context_at (&mctx->input, idx, mctx->eflags); + for (i = 0; i < state->nodes.nelem; ++i) + if (check_halt_node_context (mctx->dfa, state->nodes.elems[i], context)) + return state->nodes.elems[i]; + return 0; +} + +/* Compute the next node to which "NFA" transit from NODE("NFA" is a NFA + corresponding to the DFA). + Return the destination node, and update EPS_VIA_NODES; + return -1 in case of errors. */ + +static Idx +proceed_next_node (const re_match_context_t *mctx, Idx nregs, regmatch_t *regs, + Idx *pidx, Idx node, re_node_set *eps_via_nodes, + struct re_fail_stack_t *fs) +{ + const re_dfa_t *const dfa = mctx->dfa; + Idx i; + bool ok; + if (IS_EPSILON_NODE (dfa->nodes[node].type)) + { + re_node_set *cur_nodes = &mctx->state_log[*pidx]->nodes; + re_node_set *edests = &dfa->edests[node]; + Idx dest_node; + ok = re_node_set_insert (eps_via_nodes, node); + if (__glibc_unlikely (! ok)) + return -2; + /* Pick up a valid destination, or return -1 if none + is found. */ + for (dest_node = -1, i = 0; i < edests->nelem; ++i) + { + Idx candidate = edests->elems[i]; + if (!re_node_set_contains (cur_nodes, candidate)) + continue; + if (dest_node == -1) + dest_node = candidate; + + else + { + /* In order to avoid infinite loop like "(a*)*", return the second + epsilon-transition if the first was already considered. */ + if (re_node_set_contains (eps_via_nodes, dest_node)) + return candidate; + + /* Otherwise, push the second epsilon-transition on the fail stack. */ + else if (fs != NULL + && push_fail_stack (fs, *pidx, candidate, nregs, regs, + eps_via_nodes)) + return -2; + + /* We know we are going to exit. */ + break; + } + } + return dest_node; + } + else + { + Idx naccepted = 0; + re_token_type_t type = dfa->nodes[node].type; + +#ifdef RE_ENABLE_I18N + if (dfa->nodes[node].accept_mb) + naccepted = check_node_accept_bytes (dfa, node, &mctx->input, *pidx); + else +#endif /* RE_ENABLE_I18N */ + if (type == OP_BACK_REF) + { + Idx subexp_idx = dfa->nodes[node].opr.idx + 1; + naccepted = regs[subexp_idx].rm_eo - regs[subexp_idx].rm_so; + if (fs != NULL) + { + if (regs[subexp_idx].rm_so == -1 || regs[subexp_idx].rm_eo == -1) + return -1; + else if (naccepted) + { + char *buf = (char *) re_string_get_buffer (&mctx->input); + if (memcmp (buf + regs[subexp_idx].rm_so, buf + *pidx, + naccepted) != 0) + return -1; + } + } + + if (naccepted == 0) + { + Idx dest_node; + ok = re_node_set_insert (eps_via_nodes, node); + if (__glibc_unlikely (! ok)) + return -2; + dest_node = dfa->edests[node].elems[0]; + if (re_node_set_contains (&mctx->state_log[*pidx]->nodes, + dest_node)) + return dest_node; + } + } + + if (naccepted != 0 + || check_node_accept (mctx, dfa->nodes + node, *pidx)) + { + Idx dest_node = dfa->nexts[node]; + *pidx = (naccepted == 0) ? *pidx + 1 : *pidx + naccepted; + if (fs && (*pidx > mctx->match_last || mctx->state_log[*pidx] == NULL + || !re_node_set_contains (&mctx->state_log[*pidx]->nodes, + dest_node))) + return -1; + re_node_set_empty (eps_via_nodes); + return dest_node; + } + } + return -1; +} + +static reg_errcode_t +__attribute_warn_unused_result__ +push_fail_stack (struct re_fail_stack_t *fs, Idx str_idx, Idx dest_node, + Idx nregs, regmatch_t *regs, re_node_set *eps_via_nodes) +{ + reg_errcode_t err; + Idx num = fs->num++; + if (fs->num == fs->alloc) + { + struct re_fail_stack_ent_t *new_array; + new_array = re_realloc (fs->stack, struct re_fail_stack_ent_t, + fs->alloc * 2); + if (new_array == NULL) + return REG_ESPACE; + fs->alloc *= 2; + fs->stack = new_array; + } + fs->stack[num].idx = str_idx; + fs->stack[num].node = dest_node; + fs->stack[num].regs = re_malloc (regmatch_t, nregs); + if (fs->stack[num].regs == NULL) + return REG_ESPACE; + memcpy (fs->stack[num].regs, regs, sizeof (regmatch_t) * nregs); + err = re_node_set_init_copy (&fs->stack[num].eps_via_nodes, eps_via_nodes); + return err; +} + +static Idx +pop_fail_stack (struct re_fail_stack_t *fs, Idx *pidx, Idx nregs, + regmatch_t *regs, re_node_set *eps_via_nodes) +{ + Idx num = --fs->num; + assert (num >= 0); + *pidx = fs->stack[num].idx; + memcpy (regs, fs->stack[num].regs, sizeof (regmatch_t) * nregs); + re_node_set_free (eps_via_nodes); + re_free (fs->stack[num].regs); + *eps_via_nodes = fs->stack[num].eps_via_nodes; + return fs->stack[num].node; +} + +/* Set the positions where the subexpressions are starts/ends to registers + PMATCH. + Note: We assume that pmatch[0] is already set, and + pmatch[i].rm_so == pmatch[i].rm_eo == -1 for 0 < i < nmatch. */ + +static reg_errcode_t +__attribute_warn_unused_result__ +set_regs (const regex_t *preg, const re_match_context_t *mctx, size_t nmatch, + regmatch_t *pmatch, bool fl_backtrack) +{ + const re_dfa_t *dfa = preg->buffer; + Idx idx, cur_node; + re_node_set eps_via_nodes; + struct re_fail_stack_t *fs; + struct re_fail_stack_t fs_body = { 0, 2, NULL }; + regmatch_t *prev_idx_match; + bool prev_idx_match_malloced = false; + +#ifdef DEBUG + assert (nmatch > 1); + assert (mctx->state_log != NULL); +#endif + if (fl_backtrack) + { + fs = &fs_body; + fs->stack = re_malloc (struct re_fail_stack_ent_t, fs->alloc); + if (fs->stack == NULL) + return REG_ESPACE; + } + else + fs = NULL; + + cur_node = dfa->init_node; + re_node_set_init_empty (&eps_via_nodes); + + if (__libc_use_alloca (nmatch * sizeof (regmatch_t))) + prev_idx_match = (regmatch_t *) alloca (nmatch * sizeof (regmatch_t)); + else + { + prev_idx_match = re_malloc (regmatch_t, nmatch); + if (prev_idx_match == NULL) + { + free_fail_stack_return (fs); + return REG_ESPACE; + } + prev_idx_match_malloced = true; + } + memcpy (prev_idx_match, pmatch, sizeof (regmatch_t) * nmatch); + + for (idx = pmatch[0].rm_so; idx <= pmatch[0].rm_eo ;) + { + update_regs (dfa, pmatch, prev_idx_match, cur_node, idx, nmatch); + + if (idx == pmatch[0].rm_eo && cur_node == mctx->last_node) + { + Idx reg_idx; + if (fs) + { + for (reg_idx = 0; reg_idx < nmatch; ++reg_idx) + if (pmatch[reg_idx].rm_so > -1 && pmatch[reg_idx].rm_eo == -1) + break; + if (reg_idx == nmatch) + { + re_node_set_free (&eps_via_nodes); + if (prev_idx_match_malloced) + re_free (prev_idx_match); + return free_fail_stack_return (fs); + } + cur_node = pop_fail_stack (fs, &idx, nmatch, pmatch, + &eps_via_nodes); + } + else + { + re_node_set_free (&eps_via_nodes); + if (prev_idx_match_malloced) + re_free (prev_idx_match); + return REG_NOERROR; + } + } + + /* Proceed to next node. */ + cur_node = proceed_next_node (mctx, nmatch, pmatch, &idx, cur_node, + &eps_via_nodes, fs); + + if (__glibc_unlikely (cur_node < 0)) + { + if (__glibc_unlikely (cur_node == -2)) + { + re_node_set_free (&eps_via_nodes); + if (prev_idx_match_malloced) + re_free (prev_idx_match); + free_fail_stack_return (fs); + return REG_ESPACE; + } + if (fs) + cur_node = pop_fail_stack (fs, &idx, nmatch, pmatch, + &eps_via_nodes); + else + { + re_node_set_free (&eps_via_nodes); + if (prev_idx_match_malloced) + re_free (prev_idx_match); + return REG_NOMATCH; + } + } + } + re_node_set_free (&eps_via_nodes); + if (prev_idx_match_malloced) + re_free (prev_idx_match); + return free_fail_stack_return (fs); +} + +static reg_errcode_t +free_fail_stack_return (struct re_fail_stack_t *fs) +{ + if (fs) + { + Idx fs_idx; + for (fs_idx = 0; fs_idx < fs->num; ++fs_idx) + { + re_node_set_free (&fs->stack[fs_idx].eps_via_nodes); + re_free (fs->stack[fs_idx].regs); + } + re_free (fs->stack); + } + return REG_NOERROR; +} + +static void +update_regs (const re_dfa_t *dfa, regmatch_t *pmatch, + regmatch_t *prev_idx_match, Idx cur_node, Idx cur_idx, Idx nmatch) +{ + int type = dfa->nodes[cur_node].type; + if (type == OP_OPEN_SUBEXP) + { + Idx reg_num = dfa->nodes[cur_node].opr.idx + 1; + + /* We are at the first node of this sub expression. */ + if (reg_num < nmatch) + { + pmatch[reg_num].rm_so = cur_idx; + pmatch[reg_num].rm_eo = -1; + } + } + else if (type == OP_CLOSE_SUBEXP) + { + Idx reg_num = dfa->nodes[cur_node].opr.idx + 1; + if (reg_num < nmatch) + { + /* We are at the last node of this sub expression. */ + if (pmatch[reg_num].rm_so < cur_idx) + { + pmatch[reg_num].rm_eo = cur_idx; + /* This is a non-empty match or we are not inside an optional + subexpression. Accept this right away. */ + memcpy (prev_idx_match, pmatch, sizeof (regmatch_t) * nmatch); + } + else + { + if (dfa->nodes[cur_node].opt_subexp + && prev_idx_match[reg_num].rm_so != -1) + /* We transited through an empty match for an optional + subexpression, like (a?)*, and this is not the subexp's + first match. Copy back the old content of the registers + so that matches of an inner subexpression are undone as + well, like in ((a?))*. */ + memcpy (pmatch, prev_idx_match, sizeof (regmatch_t) * nmatch); + else + /* We completed a subexpression, but it may be part of + an optional one, so do not update PREV_IDX_MATCH. */ + pmatch[reg_num].rm_eo = cur_idx; + } + } + } +} + +/* This function checks the STATE_LOG from the SCTX->last_str_idx to 0 + and sift the nodes in each states according to the following rules. + Updated state_log will be wrote to STATE_LOG. + + Rules: We throw away the Node 'a' in the STATE_LOG[STR_IDX] if... + 1. When STR_IDX == MATCH_LAST(the last index in the state_log): + If 'a' isn't the LAST_NODE and 'a' can't epsilon transit to + the LAST_NODE, we throw away the node 'a'. + 2. When 0 <= STR_IDX < MATCH_LAST and 'a' accepts + string 's' and transit to 'b': + i. If 'b' isn't in the STATE_LOG[STR_IDX+strlen('s')], we throw + away the node 'a'. + ii. If 'b' is in the STATE_LOG[STR_IDX+strlen('s')] but 'b' is + thrown away, we throw away the node 'a'. + 3. When 0 <= STR_IDX < MATCH_LAST and 'a' epsilon transit to 'b': + i. If 'b' isn't in the STATE_LOG[STR_IDX], we throw away the + node 'a'. + ii. If 'b' is in the STATE_LOG[STR_IDX] but 'b' is thrown away, + we throw away the node 'a'. */ + +#define STATE_NODE_CONTAINS(state,node) \ + ((state) != NULL && re_node_set_contains (&(state)->nodes, node)) + +static reg_errcode_t +sift_states_backward (const re_match_context_t *mctx, re_sift_context_t *sctx) +{ + reg_errcode_t err; + int null_cnt = 0; + Idx str_idx = sctx->last_str_idx; + re_node_set cur_dest; + +#ifdef DEBUG + assert (mctx->state_log != NULL && mctx->state_log[str_idx] != NULL); +#endif + + /* Build sifted state_log[str_idx]. It has the nodes which can epsilon + transit to the last_node and the last_node itself. */ + err = re_node_set_init_1 (&cur_dest, sctx->last_node); + if (__glibc_unlikely (err != REG_NOERROR)) + return err; + err = update_cur_sifted_state (mctx, sctx, str_idx, &cur_dest); + if (__glibc_unlikely (err != REG_NOERROR)) + goto free_return; + + /* Then check each states in the state_log. */ + while (str_idx > 0) + { + /* Update counters. */ + null_cnt = (sctx->sifted_states[str_idx] == NULL) ? null_cnt + 1 : 0; + if (null_cnt > mctx->max_mb_elem_len) + { + memset (sctx->sifted_states, '\0', + sizeof (re_dfastate_t *) * str_idx); + re_node_set_free (&cur_dest); + return REG_NOERROR; + } + re_node_set_empty (&cur_dest); + --str_idx; + + if (mctx->state_log[str_idx]) + { + err = build_sifted_states (mctx, sctx, str_idx, &cur_dest); + if (__glibc_unlikely (err != REG_NOERROR)) + goto free_return; + } + + /* Add all the nodes which satisfy the following conditions: + - It can epsilon transit to a node in CUR_DEST. + - It is in CUR_SRC. + And update state_log. */ + err = update_cur_sifted_state (mctx, sctx, str_idx, &cur_dest); + if (__glibc_unlikely (err != REG_NOERROR)) + goto free_return; + } + err = REG_NOERROR; + free_return: + re_node_set_free (&cur_dest); + return err; +} + +static reg_errcode_t +__attribute_warn_unused_result__ +build_sifted_states (const re_match_context_t *mctx, re_sift_context_t *sctx, + Idx str_idx, re_node_set *cur_dest) +{ + const re_dfa_t *const dfa = mctx->dfa; + const re_node_set *cur_src = &mctx->state_log[str_idx]->non_eps_nodes; + Idx i; + + /* Then build the next sifted state. + We build the next sifted state on 'cur_dest', and update + 'sifted_states[str_idx]' with 'cur_dest'. + Note: + 'cur_dest' is the sifted state from 'state_log[str_idx + 1]'. + 'cur_src' points the node_set of the old 'state_log[str_idx]' + (with the epsilon nodes pre-filtered out). */ + for (i = 0; i < cur_src->nelem; i++) + { + Idx prev_node = cur_src->elems[i]; + int naccepted = 0; + bool ok; + +#ifdef DEBUG + re_token_type_t type = dfa->nodes[prev_node].type; + assert (!IS_EPSILON_NODE (type)); +#endif +#ifdef RE_ENABLE_I18N + /* If the node may accept "multi byte". */ + if (dfa->nodes[prev_node].accept_mb) + naccepted = sift_states_iter_mb (mctx, sctx, prev_node, + str_idx, sctx->last_str_idx); +#endif /* RE_ENABLE_I18N */ + + /* We don't check backreferences here. + See update_cur_sifted_state(). */ + if (!naccepted + && check_node_accept (mctx, dfa->nodes + prev_node, str_idx) + && STATE_NODE_CONTAINS (sctx->sifted_states[str_idx + 1], + dfa->nexts[prev_node])) + naccepted = 1; + + if (naccepted == 0) + continue; + + if (sctx->limits.nelem) + { + Idx to_idx = str_idx + naccepted; + if (check_dst_limits (mctx, &sctx->limits, + dfa->nexts[prev_node], to_idx, + prev_node, str_idx)) + continue; + } + ok = re_node_set_insert (cur_dest, prev_node); + if (__glibc_unlikely (! ok)) + return REG_ESPACE; + } + + return REG_NOERROR; +} + +/* Helper functions. */ + +static reg_errcode_t +clean_state_log_if_needed (re_match_context_t *mctx, Idx next_state_log_idx) +{ + Idx top = mctx->state_log_top; + + if ((next_state_log_idx >= mctx->input.bufs_len + && mctx->input.bufs_len < mctx->input.len) + || (next_state_log_idx >= mctx->input.valid_len + && mctx->input.valid_len < mctx->input.len)) + { + reg_errcode_t err; + err = extend_buffers (mctx, next_state_log_idx + 1); + if (__glibc_unlikely (err != REG_NOERROR)) + return err; + } + + if (top < next_state_log_idx) + { + memset (mctx->state_log + top + 1, '\0', + sizeof (re_dfastate_t *) * (next_state_log_idx - top)); + mctx->state_log_top = next_state_log_idx; + } + return REG_NOERROR; +} + +static reg_errcode_t +merge_state_array (const re_dfa_t *dfa, re_dfastate_t **dst, + re_dfastate_t **src, Idx num) +{ + Idx st_idx; + reg_errcode_t err; + for (st_idx = 0; st_idx < num; ++st_idx) + { + if (dst[st_idx] == NULL) + dst[st_idx] = src[st_idx]; + else if (src[st_idx] != NULL) + { + re_node_set merged_set; + err = re_node_set_init_union (&merged_set, &dst[st_idx]->nodes, + &src[st_idx]->nodes); + if (__glibc_unlikely (err != REG_NOERROR)) + return err; + dst[st_idx] = re_acquire_state (&err, dfa, &merged_set); + re_node_set_free (&merged_set); + if (__glibc_unlikely (err != REG_NOERROR)) + return err; + } + } + return REG_NOERROR; +} + +static reg_errcode_t +update_cur_sifted_state (const re_match_context_t *mctx, + re_sift_context_t *sctx, Idx str_idx, + re_node_set *dest_nodes) +{ + const re_dfa_t *const dfa = mctx->dfa; + reg_errcode_t err = REG_NOERROR; + const re_node_set *candidates; + candidates = ((mctx->state_log[str_idx] == NULL) ? NULL + : &mctx->state_log[str_idx]->nodes); + + if (dest_nodes->nelem == 0) + sctx->sifted_states[str_idx] = NULL; + else + { + if (candidates) + { + /* At first, add the nodes which can epsilon transit to a node in + DEST_NODE. */ + err = add_epsilon_src_nodes (dfa, dest_nodes, candidates); + if (__glibc_unlikely (err != REG_NOERROR)) + return err; + + /* Then, check the limitations in the current sift_context. */ + if (sctx->limits.nelem) + { + err = check_subexp_limits (dfa, dest_nodes, candidates, &sctx->limits, + mctx->bkref_ents, str_idx); + if (__glibc_unlikely (err != REG_NOERROR)) + return err; + } + } + + sctx->sifted_states[str_idx] = re_acquire_state (&err, dfa, dest_nodes); + if (__glibc_unlikely (err != REG_NOERROR)) + return err; + } + + if (candidates && mctx->state_log[str_idx]->has_backref) + { + err = sift_states_bkref (mctx, sctx, str_idx, candidates); + if (__glibc_unlikely (err != REG_NOERROR)) + return err; + } + return REG_NOERROR; +} + +static reg_errcode_t +__attribute_warn_unused_result__ +add_epsilon_src_nodes (const re_dfa_t *dfa, re_node_set *dest_nodes, + const re_node_set *candidates) +{ + reg_errcode_t err = REG_NOERROR; + Idx i; + + re_dfastate_t *state = re_acquire_state (&err, dfa, dest_nodes); + if (__glibc_unlikely (err != REG_NOERROR)) + return err; + + if (!state->inveclosure.alloc) + { + err = re_node_set_alloc (&state->inveclosure, dest_nodes->nelem); + if (__glibc_unlikely (err != REG_NOERROR)) + return REG_ESPACE; + for (i = 0; i < dest_nodes->nelem; i++) + { + err = re_node_set_merge (&state->inveclosure, + dfa->inveclosures + dest_nodes->elems[i]); + if (__glibc_unlikely (err != REG_NOERROR)) + return REG_ESPACE; + } + } + return re_node_set_add_intersect (dest_nodes, candidates, + &state->inveclosure); +} + +static reg_errcode_t +sub_epsilon_src_nodes (const re_dfa_t *dfa, Idx node, re_node_set *dest_nodes, + const re_node_set *candidates) +{ + Idx ecl_idx; + reg_errcode_t err; + re_node_set *inv_eclosure = dfa->inveclosures + node; + re_node_set except_nodes; + re_node_set_init_empty (&except_nodes); + for (ecl_idx = 0; ecl_idx < inv_eclosure->nelem; ++ecl_idx) + { + Idx cur_node = inv_eclosure->elems[ecl_idx]; + if (cur_node == node) + continue; + if (IS_EPSILON_NODE (dfa->nodes[cur_node].type)) + { + Idx edst1 = dfa->edests[cur_node].elems[0]; + Idx edst2 = ((dfa->edests[cur_node].nelem > 1) + ? dfa->edests[cur_node].elems[1] : -1); + if ((!re_node_set_contains (inv_eclosure, edst1) + && re_node_set_contains (dest_nodes, edst1)) + || (edst2 > 0 + && !re_node_set_contains (inv_eclosure, edst2) + && re_node_set_contains (dest_nodes, edst2))) + { + err = re_node_set_add_intersect (&except_nodes, candidates, + dfa->inveclosures + cur_node); + if (__glibc_unlikely (err != REG_NOERROR)) + { + re_node_set_free (&except_nodes); + return err; + } + } + } + } + for (ecl_idx = 0; ecl_idx < inv_eclosure->nelem; ++ecl_idx) + { + Idx cur_node = inv_eclosure->elems[ecl_idx]; + if (!re_node_set_contains (&except_nodes, cur_node)) + { + Idx idx = re_node_set_contains (dest_nodes, cur_node) - 1; + re_node_set_remove_at (dest_nodes, idx); + } + } + re_node_set_free (&except_nodes); + return REG_NOERROR; +} + +static bool +check_dst_limits (const re_match_context_t *mctx, const re_node_set *limits, + Idx dst_node, Idx dst_idx, Idx src_node, Idx src_idx) +{ + const re_dfa_t *const dfa = mctx->dfa; + Idx lim_idx, src_pos, dst_pos; + + Idx dst_bkref_idx = search_cur_bkref_entry (mctx, dst_idx); + Idx src_bkref_idx = search_cur_bkref_entry (mctx, src_idx); + for (lim_idx = 0; lim_idx < limits->nelem; ++lim_idx) + { + Idx subexp_idx; + struct re_backref_cache_entry *ent; + ent = mctx->bkref_ents + limits->elems[lim_idx]; + subexp_idx = dfa->nodes[ent->node].opr.idx; + + dst_pos = check_dst_limits_calc_pos (mctx, limits->elems[lim_idx], + subexp_idx, dst_node, dst_idx, + dst_bkref_idx); + src_pos = check_dst_limits_calc_pos (mctx, limits->elems[lim_idx], + subexp_idx, src_node, src_idx, + src_bkref_idx); + + /* In case of: + <src> <dst> ( <subexp> ) + ( <subexp> ) <src> <dst> + ( <subexp1> <src> <subexp2> <dst> <subexp3> ) */ + if (src_pos == dst_pos) + continue; /* This is unrelated limitation. */ + else + return true; + } + return false; +} + +static int +check_dst_limits_calc_pos_1 (const re_match_context_t *mctx, int boundaries, + Idx subexp_idx, Idx from_node, Idx bkref_idx) +{ + const re_dfa_t *const dfa = mctx->dfa; + const re_node_set *eclosures = dfa->eclosures + from_node; + Idx node_idx; + + /* Else, we are on the boundary: examine the nodes on the epsilon + closure. */ + for (node_idx = 0; node_idx < eclosures->nelem; ++node_idx) + { + Idx node = eclosures->elems[node_idx]; + switch (dfa->nodes[node].type) + { + case OP_BACK_REF: + if (bkref_idx != -1) + { + struct re_backref_cache_entry *ent = mctx->bkref_ents + bkref_idx; + do + { + Idx dst; + int cpos; + + if (ent->node != node) + continue; + + if (subexp_idx < BITSET_WORD_BITS + && !(ent->eps_reachable_subexps_map + & ((bitset_word_t) 1 << subexp_idx))) + continue; + + /* Recurse trying to reach the OP_OPEN_SUBEXP and + OP_CLOSE_SUBEXP cases below. But, if the + destination node is the same node as the source + node, don't recurse because it would cause an + infinite loop: a regex that exhibits this behavior + is ()\1*\1* */ + dst = dfa->edests[node].elems[0]; + if (dst == from_node) + { + if (boundaries & 1) + return -1; + else /* if (boundaries & 2) */ + return 0; + } + + cpos = + check_dst_limits_calc_pos_1 (mctx, boundaries, subexp_idx, + dst, bkref_idx); + if (cpos == -1 /* && (boundaries & 1) */) + return -1; + if (cpos == 0 && (boundaries & 2)) + return 0; + + if (subexp_idx < BITSET_WORD_BITS) + ent->eps_reachable_subexps_map + &= ~((bitset_word_t) 1 << subexp_idx); + } + while (ent++->more); + } + break; + + case OP_OPEN_SUBEXP: + if ((boundaries & 1) && subexp_idx == dfa->nodes[node].opr.idx) + return -1; + break; + + case OP_CLOSE_SUBEXP: + if ((boundaries & 2) && subexp_idx == dfa->nodes[node].opr.idx) + return 0; + break; + + default: + break; + } + } + + return (boundaries & 2) ? 1 : 0; +} + +static int +check_dst_limits_calc_pos (const re_match_context_t *mctx, Idx limit, + Idx subexp_idx, Idx from_node, Idx str_idx, + Idx bkref_idx) +{ + struct re_backref_cache_entry *lim = mctx->bkref_ents + limit; + int boundaries; + + /* If we are outside the range of the subexpression, return -1 or 1. */ + if (str_idx < lim->subexp_from) + return -1; + + if (lim->subexp_to < str_idx) + return 1; + + /* If we are within the subexpression, return 0. */ + boundaries = (str_idx == lim->subexp_from); + boundaries |= (str_idx == lim->subexp_to) << 1; + if (boundaries == 0) + return 0; + + /* Else, examine epsilon closure. */ + return check_dst_limits_calc_pos_1 (mctx, boundaries, subexp_idx, + from_node, bkref_idx); +} + +/* Check the limitations of sub expressions LIMITS, and remove the nodes + which are against limitations from DEST_NODES. */ + +static reg_errcode_t +check_subexp_limits (const re_dfa_t *dfa, re_node_set *dest_nodes, + const re_node_set *candidates, re_node_set *limits, + struct re_backref_cache_entry *bkref_ents, Idx str_idx) +{ + reg_errcode_t err; + Idx node_idx, lim_idx; + + for (lim_idx = 0; lim_idx < limits->nelem; ++lim_idx) + { + Idx subexp_idx; + struct re_backref_cache_entry *ent; + ent = bkref_ents + limits->elems[lim_idx]; + + if (str_idx <= ent->subexp_from || ent->str_idx < str_idx) + continue; /* This is unrelated limitation. */ + + subexp_idx = dfa->nodes[ent->node].opr.idx; + if (ent->subexp_to == str_idx) + { + Idx ops_node = -1; + Idx cls_node = -1; + for (node_idx = 0; node_idx < dest_nodes->nelem; ++node_idx) + { + Idx node = dest_nodes->elems[node_idx]; + re_token_type_t type = dfa->nodes[node].type; + if (type == OP_OPEN_SUBEXP + && subexp_idx == dfa->nodes[node].opr.idx) + ops_node = node; + else if (type == OP_CLOSE_SUBEXP + && subexp_idx == dfa->nodes[node].opr.idx) + cls_node = node; + } + + /* Check the limitation of the open subexpression. */ + /* Note that (ent->subexp_to = str_idx != ent->subexp_from). */ + if (ops_node >= 0) + { + err = sub_epsilon_src_nodes (dfa, ops_node, dest_nodes, + candidates); + if (__glibc_unlikely (err != REG_NOERROR)) + return err; + } + + /* Check the limitation of the close subexpression. */ + if (cls_node >= 0) + for (node_idx = 0; node_idx < dest_nodes->nelem; ++node_idx) + { + Idx node = dest_nodes->elems[node_idx]; + if (!re_node_set_contains (dfa->inveclosures + node, + cls_node) + && !re_node_set_contains (dfa->eclosures + node, + cls_node)) + { + /* It is against this limitation. + Remove it form the current sifted state. */ + err = sub_epsilon_src_nodes (dfa, node, dest_nodes, + candidates); + if (__glibc_unlikely (err != REG_NOERROR)) + return err; + --node_idx; + } + } + } + else /* (ent->subexp_to != str_idx) */ + { + for (node_idx = 0; node_idx < dest_nodes->nelem; ++node_idx) + { + Idx node = dest_nodes->elems[node_idx]; + re_token_type_t type = dfa->nodes[node].type; + if (type == OP_CLOSE_SUBEXP || type == OP_OPEN_SUBEXP) + { + if (subexp_idx != dfa->nodes[node].opr.idx) + continue; + /* It is against this limitation. + Remove it form the current sifted state. */ + err = sub_epsilon_src_nodes (dfa, node, dest_nodes, + candidates); + if (__glibc_unlikely (err != REG_NOERROR)) + return err; + } + } + } + } + return REG_NOERROR; +} + +static reg_errcode_t +__attribute_warn_unused_result__ +sift_states_bkref (const re_match_context_t *mctx, re_sift_context_t *sctx, + Idx str_idx, const re_node_set *candidates) +{ + const re_dfa_t *const dfa = mctx->dfa; + reg_errcode_t err; + Idx node_idx, node; + re_sift_context_t local_sctx; + Idx first_idx = search_cur_bkref_entry (mctx, str_idx); + + if (first_idx == -1) + return REG_NOERROR; + + local_sctx.sifted_states = NULL; /* Mark that it hasn't been initialized. */ + + for (node_idx = 0; node_idx < candidates->nelem; ++node_idx) + { + Idx enabled_idx; + re_token_type_t type; + struct re_backref_cache_entry *entry; + node = candidates->elems[node_idx]; + type = dfa->nodes[node].type; + /* Avoid infinite loop for the REs like "()\1+". */ + if (node == sctx->last_node && str_idx == sctx->last_str_idx) + continue; + if (type != OP_BACK_REF) + continue; + + entry = mctx->bkref_ents + first_idx; + enabled_idx = first_idx; + do + { + Idx subexp_len; + Idx to_idx; + Idx dst_node; + bool ok; + re_dfastate_t *cur_state; + + if (entry->node != node) + continue; + subexp_len = entry->subexp_to - entry->subexp_from; + to_idx = str_idx + subexp_len; + dst_node = (subexp_len ? dfa->nexts[node] + : dfa->edests[node].elems[0]); + + if (to_idx > sctx->last_str_idx + || sctx->sifted_states[to_idx] == NULL + || !STATE_NODE_CONTAINS (sctx->sifted_states[to_idx], dst_node) + || check_dst_limits (mctx, &sctx->limits, node, + str_idx, dst_node, to_idx)) + continue; + + if (local_sctx.sifted_states == NULL) + { + local_sctx = *sctx; + err = re_node_set_init_copy (&local_sctx.limits, &sctx->limits); + if (__glibc_unlikely (err != REG_NOERROR)) + goto free_return; + } + local_sctx.last_node = node; + local_sctx.last_str_idx = str_idx; + ok = re_node_set_insert (&local_sctx.limits, enabled_idx); + if (__glibc_unlikely (! ok)) + { + err = REG_ESPACE; + goto free_return; + } + cur_state = local_sctx.sifted_states[str_idx]; + err = sift_states_backward (mctx, &local_sctx); + if (__glibc_unlikely (err != REG_NOERROR)) + goto free_return; + if (sctx->limited_states != NULL) + { + err = merge_state_array (dfa, sctx->limited_states, + local_sctx.sifted_states, + str_idx + 1); + if (__glibc_unlikely (err != REG_NOERROR)) + goto free_return; + } + local_sctx.sifted_states[str_idx] = cur_state; + re_node_set_remove (&local_sctx.limits, enabled_idx); + + /* mctx->bkref_ents may have changed, reload the pointer. */ + entry = mctx->bkref_ents + enabled_idx; + } + while (enabled_idx++, entry++->more); + } + err = REG_NOERROR; + free_return: + if (local_sctx.sifted_states != NULL) + { + re_node_set_free (&local_sctx.limits); + } + + return err; +} + + +#ifdef RE_ENABLE_I18N +static int +sift_states_iter_mb (const re_match_context_t *mctx, re_sift_context_t *sctx, + Idx node_idx, Idx str_idx, Idx max_str_idx) +{ + const re_dfa_t *const dfa = mctx->dfa; + int naccepted; + /* Check the node can accept "multi byte". */ + naccepted = check_node_accept_bytes (dfa, node_idx, &mctx->input, str_idx); + if (naccepted > 0 && str_idx + naccepted <= max_str_idx && + !STATE_NODE_CONTAINS (sctx->sifted_states[str_idx + naccepted], + dfa->nexts[node_idx])) + /* The node can't accept the "multi byte", or the + destination was already thrown away, then the node + could't accept the current input "multi byte". */ + naccepted = 0; + /* Otherwise, it is sure that the node could accept + 'naccepted' bytes input. */ + return naccepted; +} +#endif /* RE_ENABLE_I18N */ + + +/* Functions for state transition. */ + +/* Return the next state to which the current state STATE will transit by + accepting the current input byte, and update STATE_LOG if necessary. + If STATE can accept a multibyte char/collating element/back reference + update the destination of STATE_LOG. */ + +static re_dfastate_t * +__attribute_warn_unused_result__ +transit_state (reg_errcode_t *err, re_match_context_t *mctx, + re_dfastate_t *state) +{ + re_dfastate_t **trtable; + unsigned char ch; + +#ifdef RE_ENABLE_I18N + /* If the current state can accept multibyte. */ + if (__glibc_unlikely (state->accept_mb)) + { + *err = transit_state_mb (mctx, state); + if (__glibc_unlikely (*err != REG_NOERROR)) + return NULL; + } +#endif /* RE_ENABLE_I18N */ + + /* Then decide the next state with the single byte. */ +#if 0 + if (0) + /* don't use transition table */ + return transit_state_sb (err, mctx, state); +#endif + + /* Use transition table */ + ch = re_string_fetch_byte (&mctx->input); + for (;;) + { + trtable = state->trtable; + if (__glibc_likely (trtable != NULL)) + return trtable[ch]; + + trtable = state->word_trtable; + if (__glibc_likely (trtable != NULL)) + { + unsigned int context; + context + = re_string_context_at (&mctx->input, + re_string_cur_idx (&mctx->input) - 1, + mctx->eflags); + if (IS_WORD_CONTEXT (context)) + return trtable[ch + SBC_MAX]; + else + return trtable[ch]; + } + + if (!build_trtable (mctx->dfa, state)) + { + *err = REG_ESPACE; + return NULL; + } + + /* Retry, we now have a transition table. */ + } +} + +/* Update the state_log if we need */ +static re_dfastate_t * +merge_state_with_log (reg_errcode_t *err, re_match_context_t *mctx, + re_dfastate_t *next_state) +{ + const re_dfa_t *const dfa = mctx->dfa; + Idx cur_idx = re_string_cur_idx (&mctx->input); + + if (cur_idx > mctx->state_log_top) + { + mctx->state_log[cur_idx] = next_state; + mctx->state_log_top = cur_idx; + } + else if (mctx->state_log[cur_idx] == 0) + { + mctx->state_log[cur_idx] = next_state; + } + else + { + re_dfastate_t *pstate; + unsigned int context; + re_node_set next_nodes, *log_nodes, *table_nodes = NULL; + /* If (state_log[cur_idx] != 0), it implies that cur_idx is + the destination of a multibyte char/collating element/ + back reference. Then the next state is the union set of + these destinations and the results of the transition table. */ + pstate = mctx->state_log[cur_idx]; + log_nodes = pstate->entrance_nodes; + if (next_state != NULL) + { + table_nodes = next_state->entrance_nodes; + *err = re_node_set_init_union (&next_nodes, table_nodes, + log_nodes); + if (__glibc_unlikely (*err != REG_NOERROR)) + return NULL; + } + else + next_nodes = *log_nodes; + /* Note: We already add the nodes of the initial state, + then we don't need to add them here. */ + + context = re_string_context_at (&mctx->input, + re_string_cur_idx (&mctx->input) - 1, + mctx->eflags); + next_state = mctx->state_log[cur_idx] + = re_acquire_state_context (err, dfa, &next_nodes, context); + /* We don't need to check errors here, since the return value of + this function is next_state and ERR is already set. */ + + if (table_nodes != NULL) + re_node_set_free (&next_nodes); + } + + if (__glibc_unlikely (dfa->nbackref) && next_state != NULL) + { + /* Check OP_OPEN_SUBEXP in the current state in case that we use them + later. We must check them here, since the back references in the + next state might use them. */ + *err = check_subexp_matching_top (mctx, &next_state->nodes, + cur_idx); + if (__glibc_unlikely (*err != REG_NOERROR)) + return NULL; + + /* If the next state has back references. */ + if (next_state->has_backref) + { + *err = transit_state_bkref (mctx, &next_state->nodes); + if (__glibc_unlikely (*err != REG_NOERROR)) + return NULL; + next_state = mctx->state_log[cur_idx]; + } + } + + return next_state; +} + +/* Skip bytes in the input that correspond to part of a + multi-byte match, then look in the log for a state + from which to restart matching. */ +static re_dfastate_t * +find_recover_state (reg_errcode_t *err, re_match_context_t *mctx) +{ + re_dfastate_t *cur_state; + do + { + Idx max = mctx->state_log_top; + Idx cur_str_idx = re_string_cur_idx (&mctx->input); + + do + { + if (++cur_str_idx > max) + return NULL; + re_string_skip_bytes (&mctx->input, 1); + } + while (mctx->state_log[cur_str_idx] == NULL); + + cur_state = merge_state_with_log (err, mctx, NULL); + } + while (*err == REG_NOERROR && cur_state == NULL); + return cur_state; +} + +/* Helper functions for transit_state. */ + +/* From the node set CUR_NODES, pick up the nodes whose types are + OP_OPEN_SUBEXP and which have corresponding back references in the regular + expression. And register them to use them later for evaluating the + corresponding back references. */ + +static reg_errcode_t +check_subexp_matching_top (re_match_context_t *mctx, re_node_set *cur_nodes, + Idx str_idx) +{ + const re_dfa_t *const dfa = mctx->dfa; + Idx node_idx; + reg_errcode_t err; + + /* TODO: This isn't efficient. + Because there might be more than one nodes whose types are + OP_OPEN_SUBEXP and whose index is SUBEXP_IDX, we must check all + nodes. + E.g. RE: (a){2} */ + for (node_idx = 0; node_idx < cur_nodes->nelem; ++node_idx) + { + Idx node = cur_nodes->elems[node_idx]; + if (dfa->nodes[node].type == OP_OPEN_SUBEXP + && dfa->nodes[node].opr.idx < BITSET_WORD_BITS + && (dfa->used_bkref_map + & ((bitset_word_t) 1 << dfa->nodes[node].opr.idx))) + { + err = match_ctx_add_subtop (mctx, node, str_idx); + if (__glibc_unlikely (err != REG_NOERROR)) + return err; + } + } + return REG_NOERROR; +} + +#if 0 +/* Return the next state to which the current state STATE will transit by + accepting the current input byte. */ + +static re_dfastate_t * +transit_state_sb (reg_errcode_t *err, re_match_context_t *mctx, + re_dfastate_t *state) +{ + const re_dfa_t *const dfa = mctx->dfa; + re_node_set next_nodes; + re_dfastate_t *next_state; + Idx node_cnt, cur_str_idx = re_string_cur_idx (&mctx->input); + unsigned int context; + + *err = re_node_set_alloc (&next_nodes, state->nodes.nelem + 1); + if (__glibc_unlikely (*err != REG_NOERROR)) + return NULL; + for (node_cnt = 0; node_cnt < state->nodes.nelem; ++node_cnt) + { + Idx cur_node = state->nodes.elems[node_cnt]; + if (check_node_accept (mctx, dfa->nodes + cur_node, cur_str_idx)) + { + *err = re_node_set_merge (&next_nodes, + dfa->eclosures + dfa->nexts[cur_node]); + if (__glibc_unlikely (*err != REG_NOERROR)) + { + re_node_set_free (&next_nodes); + return NULL; + } + } + } + context = re_string_context_at (&mctx->input, cur_str_idx, mctx->eflags); + next_state = re_acquire_state_context (err, dfa, &next_nodes, context); + /* We don't need to check errors here, since the return value of + this function is next_state and ERR is already set. */ + + re_node_set_free (&next_nodes); + re_string_skip_bytes (&mctx->input, 1); + return next_state; +} +#endif + +#ifdef RE_ENABLE_I18N +static reg_errcode_t +transit_state_mb (re_match_context_t *mctx, re_dfastate_t *pstate) +{ + const re_dfa_t *const dfa = mctx->dfa; + reg_errcode_t err; + Idx i; + + for (i = 0; i < pstate->nodes.nelem; ++i) + { + re_node_set dest_nodes, *new_nodes; + Idx cur_node_idx = pstate->nodes.elems[i]; + int naccepted; + Idx dest_idx; + unsigned int context; + re_dfastate_t *dest_state; + + if (!dfa->nodes[cur_node_idx].accept_mb) + continue; + + if (dfa->nodes[cur_node_idx].constraint) + { + context = re_string_context_at (&mctx->input, + re_string_cur_idx (&mctx->input), + mctx->eflags); + if (NOT_SATISFY_NEXT_CONSTRAINT (dfa->nodes[cur_node_idx].constraint, + context)) + continue; + } + + /* How many bytes the node can accept? */ + naccepted = check_node_accept_bytes (dfa, cur_node_idx, &mctx->input, + re_string_cur_idx (&mctx->input)); + if (naccepted == 0) + continue; + + /* The node can accepts 'naccepted' bytes. */ + dest_idx = re_string_cur_idx (&mctx->input) + naccepted; + mctx->max_mb_elem_len = ((mctx->max_mb_elem_len < naccepted) ? naccepted + : mctx->max_mb_elem_len); + err = clean_state_log_if_needed (mctx, dest_idx); + if (__glibc_unlikely (err != REG_NOERROR)) + return err; +#ifdef DEBUG + assert (dfa->nexts[cur_node_idx] != -1); +#endif + new_nodes = dfa->eclosures + dfa->nexts[cur_node_idx]; + + dest_state = mctx->state_log[dest_idx]; + if (dest_state == NULL) + dest_nodes = *new_nodes; + else + { + err = re_node_set_init_union (&dest_nodes, + dest_state->entrance_nodes, new_nodes); + if (__glibc_unlikely (err != REG_NOERROR)) + return err; + } + context = re_string_context_at (&mctx->input, dest_idx - 1, + mctx->eflags); + mctx->state_log[dest_idx] + = re_acquire_state_context (&err, dfa, &dest_nodes, context); + if (dest_state != NULL) + re_node_set_free (&dest_nodes); + if (__glibc_unlikely (mctx->state_log[dest_idx] == NULL + && err != REG_NOERROR)) + return err; + } + return REG_NOERROR; +} +#endif /* RE_ENABLE_I18N */ + +static reg_errcode_t +transit_state_bkref (re_match_context_t *mctx, const re_node_set *nodes) +{ + const re_dfa_t *const dfa = mctx->dfa; + reg_errcode_t err; + Idx i; + Idx cur_str_idx = re_string_cur_idx (&mctx->input); + + for (i = 0; i < nodes->nelem; ++i) + { + Idx dest_str_idx, prev_nelem, bkc_idx; + Idx node_idx = nodes->elems[i]; + unsigned int context; + const re_token_t *node = dfa->nodes + node_idx; + re_node_set *new_dest_nodes; + + /* Check whether 'node' is a backreference or not. */ + if (node->type != OP_BACK_REF) + continue; + + if (node->constraint) + { + context = re_string_context_at (&mctx->input, cur_str_idx, + mctx->eflags); + if (NOT_SATISFY_NEXT_CONSTRAINT (node->constraint, context)) + continue; + } + + /* 'node' is a backreference. + Check the substring which the substring matched. */ + bkc_idx = mctx->nbkref_ents; + err = get_subexp (mctx, node_idx, cur_str_idx); + if (__glibc_unlikely (err != REG_NOERROR)) + goto free_return; + + /* And add the epsilon closures (which is 'new_dest_nodes') of + the backreference to appropriate state_log. */ +#ifdef DEBUG + assert (dfa->nexts[node_idx] != -1); +#endif + for (; bkc_idx < mctx->nbkref_ents; ++bkc_idx) + { + Idx subexp_len; + re_dfastate_t *dest_state; + struct re_backref_cache_entry *bkref_ent; + bkref_ent = mctx->bkref_ents + bkc_idx; + if (bkref_ent->node != node_idx || bkref_ent->str_idx != cur_str_idx) + continue; + subexp_len = bkref_ent->subexp_to - bkref_ent->subexp_from; + new_dest_nodes = (subexp_len == 0 + ? dfa->eclosures + dfa->edests[node_idx].elems[0] + : dfa->eclosures + dfa->nexts[node_idx]); + dest_str_idx = (cur_str_idx + bkref_ent->subexp_to + - bkref_ent->subexp_from); + context = re_string_context_at (&mctx->input, dest_str_idx - 1, + mctx->eflags); + dest_state = mctx->state_log[dest_str_idx]; + prev_nelem = ((mctx->state_log[cur_str_idx] == NULL) ? 0 + : mctx->state_log[cur_str_idx]->nodes.nelem); + /* Add 'new_dest_node' to state_log. */ + if (dest_state == NULL) + { + mctx->state_log[dest_str_idx] + = re_acquire_state_context (&err, dfa, new_dest_nodes, + context); + if (__glibc_unlikely (mctx->state_log[dest_str_idx] == NULL + && err != REG_NOERROR)) + goto free_return; + } + else + { + re_node_set dest_nodes; + err = re_node_set_init_union (&dest_nodes, + dest_state->entrance_nodes, + new_dest_nodes); + if (__glibc_unlikely (err != REG_NOERROR)) + { + re_node_set_free (&dest_nodes); + goto free_return; + } + mctx->state_log[dest_str_idx] + = re_acquire_state_context (&err, dfa, &dest_nodes, context); + re_node_set_free (&dest_nodes); + if (__glibc_unlikely (mctx->state_log[dest_str_idx] == NULL + && err != REG_NOERROR)) + goto free_return; + } + /* We need to check recursively if the backreference can epsilon + transit. */ + if (subexp_len == 0 + && mctx->state_log[cur_str_idx]->nodes.nelem > prev_nelem) + { + err = check_subexp_matching_top (mctx, new_dest_nodes, + cur_str_idx); + if (__glibc_unlikely (err != REG_NOERROR)) + goto free_return; + err = transit_state_bkref (mctx, new_dest_nodes); + if (__glibc_unlikely (err != REG_NOERROR)) + goto free_return; + } + } + } + err = REG_NOERROR; + free_return: + return err; +} + +/* Enumerate all the candidates which the backreference BKREF_NODE can match + at BKREF_STR_IDX, and register them by match_ctx_add_entry(). + Note that we might collect inappropriate candidates here. + However, the cost of checking them strictly here is too high, then we + delay these checking for prune_impossible_nodes(). */ + +static reg_errcode_t +__attribute_warn_unused_result__ +get_subexp (re_match_context_t *mctx, Idx bkref_node, Idx bkref_str_idx) +{ + const re_dfa_t *const dfa = mctx->dfa; + Idx subexp_num, sub_top_idx; + const char *buf = (const char *) re_string_get_buffer (&mctx->input); + /* Return if we have already checked BKREF_NODE at BKREF_STR_IDX. */ + Idx cache_idx = search_cur_bkref_entry (mctx, bkref_str_idx); + if (cache_idx != -1) + { + const struct re_backref_cache_entry *entry + = mctx->bkref_ents + cache_idx; + do + if (entry->node == bkref_node) + return REG_NOERROR; /* We already checked it. */ + while (entry++->more); + } + + subexp_num = dfa->nodes[bkref_node].opr.idx; + + /* For each sub expression */ + for (sub_top_idx = 0; sub_top_idx < mctx->nsub_tops; ++sub_top_idx) + { + reg_errcode_t err; + re_sub_match_top_t *sub_top = mctx->sub_tops[sub_top_idx]; + re_sub_match_last_t *sub_last; + Idx sub_last_idx, sl_str, bkref_str_off; + + if (dfa->nodes[sub_top->node].opr.idx != subexp_num) + continue; /* It isn't related. */ + + sl_str = sub_top->str_idx; + bkref_str_off = bkref_str_idx; + /* At first, check the last node of sub expressions we already + evaluated. */ + for (sub_last_idx = 0; sub_last_idx < sub_top->nlasts; ++sub_last_idx) + { + regoff_t sl_str_diff; + sub_last = sub_top->lasts[sub_last_idx]; + sl_str_diff = sub_last->str_idx - sl_str; + /* The matched string by the sub expression match with the substring + at the back reference? */ + if (sl_str_diff > 0) + { + if (__glibc_unlikely (bkref_str_off + sl_str_diff + > mctx->input.valid_len)) + { + /* Not enough chars for a successful match. */ + if (bkref_str_off + sl_str_diff > mctx->input.len) + break; + + err = clean_state_log_if_needed (mctx, + bkref_str_off + + sl_str_diff); + if (__glibc_unlikely (err != REG_NOERROR)) + return err; + buf = (const char *) re_string_get_buffer (&mctx->input); + } + if (memcmp (buf + bkref_str_off, buf + sl_str, sl_str_diff) != 0) + /* We don't need to search this sub expression any more. */ + break; + } + bkref_str_off += sl_str_diff; + sl_str += sl_str_diff; + err = get_subexp_sub (mctx, sub_top, sub_last, bkref_node, + bkref_str_idx); + + /* Reload buf, since the preceding call might have reallocated + the buffer. */ + buf = (const char *) re_string_get_buffer (&mctx->input); + + if (err == REG_NOMATCH) + continue; + if (__glibc_unlikely (err != REG_NOERROR)) + return err; + } + + if (sub_last_idx < sub_top->nlasts) + continue; + if (sub_last_idx > 0) + ++sl_str; + /* Then, search for the other last nodes of the sub expression. */ + for (; sl_str <= bkref_str_idx; ++sl_str) + { + Idx cls_node; + regoff_t sl_str_off; + const re_node_set *nodes; + sl_str_off = sl_str - sub_top->str_idx; + /* The matched string by the sub expression match with the substring + at the back reference? */ + if (sl_str_off > 0) + { + if (__glibc_unlikely (bkref_str_off >= mctx->input.valid_len)) + { + /* If we are at the end of the input, we cannot match. */ + if (bkref_str_off >= mctx->input.len) + break; + + err = extend_buffers (mctx, bkref_str_off + 1); + if (__glibc_unlikely (err != REG_NOERROR)) + return err; + + buf = (const char *) re_string_get_buffer (&mctx->input); + } + if (buf [bkref_str_off++] != buf[sl_str - 1]) + break; /* We don't need to search this sub expression + any more. */ + } + if (mctx->state_log[sl_str] == NULL) + continue; + /* Does this state have a ')' of the sub expression? */ + nodes = &mctx->state_log[sl_str]->nodes; + cls_node = find_subexp_node (dfa, nodes, subexp_num, + OP_CLOSE_SUBEXP); + if (cls_node == -1) + continue; /* No. */ + if (sub_top->path == NULL) + { + sub_top->path = calloc (sizeof (state_array_t), + sl_str - sub_top->str_idx + 1); + if (sub_top->path == NULL) + return REG_ESPACE; + } + /* Can the OP_OPEN_SUBEXP node arrive the OP_CLOSE_SUBEXP node + in the current context? */ + err = check_arrival (mctx, sub_top->path, sub_top->node, + sub_top->str_idx, cls_node, sl_str, + OP_CLOSE_SUBEXP); + if (err == REG_NOMATCH) + continue; + if (__glibc_unlikely (err != REG_NOERROR)) + return err; + sub_last = match_ctx_add_sublast (sub_top, cls_node, sl_str); + if (__glibc_unlikely (sub_last == NULL)) + return REG_ESPACE; + err = get_subexp_sub (mctx, sub_top, sub_last, bkref_node, + bkref_str_idx); + buf = (const char *) re_string_get_buffer (&mctx->input); + if (err == REG_NOMATCH) + continue; + if (__glibc_unlikely (err != REG_NOERROR)) + return err; + } + } + return REG_NOERROR; +} + +/* Helper functions for get_subexp(). */ + +/* Check SUB_LAST can arrive to the back reference BKREF_NODE at BKREF_STR. + If it can arrive, register the sub expression expressed with SUB_TOP + and SUB_LAST. */ + +static reg_errcode_t +get_subexp_sub (re_match_context_t *mctx, const re_sub_match_top_t *sub_top, + re_sub_match_last_t *sub_last, Idx bkref_node, Idx bkref_str) +{ + reg_errcode_t err; + Idx to_idx; + /* Can the subexpression arrive the back reference? */ + err = check_arrival (mctx, &sub_last->path, sub_last->node, + sub_last->str_idx, bkref_node, bkref_str, + OP_OPEN_SUBEXP); + if (err != REG_NOERROR) + return err; + err = match_ctx_add_entry (mctx, bkref_node, bkref_str, sub_top->str_idx, + sub_last->str_idx); + if (__glibc_unlikely (err != REG_NOERROR)) + return err; + to_idx = bkref_str + sub_last->str_idx - sub_top->str_idx; + return clean_state_log_if_needed (mctx, to_idx); +} + +/* Find the first node which is '(' or ')' and whose index is SUBEXP_IDX. + Search '(' if FL_OPEN, or search ')' otherwise. + TODO: This function isn't efficient... + Because there might be more than one nodes whose types are + OP_OPEN_SUBEXP and whose index is SUBEXP_IDX, we must check all + nodes. + E.g. RE: (a){2} */ + +static Idx +find_subexp_node (const re_dfa_t *dfa, const re_node_set *nodes, + Idx subexp_idx, int type) +{ + Idx cls_idx; + for (cls_idx = 0; cls_idx < nodes->nelem; ++cls_idx) + { + Idx cls_node = nodes->elems[cls_idx]; + const re_token_t *node = dfa->nodes + cls_node; + if (node->type == type + && node->opr.idx == subexp_idx) + return cls_node; + } + return -1; +} + +/* Check whether the node TOP_NODE at TOP_STR can arrive to the node + LAST_NODE at LAST_STR. We record the path onto PATH since it will be + heavily reused. + Return REG_NOERROR if it can arrive, or REG_NOMATCH otherwise. */ + +static reg_errcode_t +__attribute_warn_unused_result__ +check_arrival (re_match_context_t *mctx, state_array_t *path, Idx top_node, + Idx top_str, Idx last_node, Idx last_str, int type) +{ + const re_dfa_t *const dfa = mctx->dfa; + reg_errcode_t err = REG_NOERROR; + Idx subexp_num, backup_cur_idx, str_idx, null_cnt; + re_dfastate_t *cur_state = NULL; + re_node_set *cur_nodes, next_nodes; + re_dfastate_t **backup_state_log; + unsigned int context; + + subexp_num = dfa->nodes[top_node].opr.idx; + /* Extend the buffer if we need. */ + if (__glibc_unlikely (path->alloc < last_str + mctx->max_mb_elem_len + 1)) + { + re_dfastate_t **new_array; + Idx old_alloc = path->alloc; + Idx incr_alloc = last_str + mctx->max_mb_elem_len + 1; + Idx new_alloc; + if (__glibc_unlikely (IDX_MAX - old_alloc < incr_alloc)) + return REG_ESPACE; + new_alloc = old_alloc + incr_alloc; + if (__glibc_unlikely (SIZE_MAX / sizeof (re_dfastate_t *) < new_alloc)) + return REG_ESPACE; + new_array = re_realloc (path->array, re_dfastate_t *, new_alloc); + if (__glibc_unlikely (new_array == NULL)) + return REG_ESPACE; + path->array = new_array; + path->alloc = new_alloc; + memset (new_array + old_alloc, '\0', + sizeof (re_dfastate_t *) * (path->alloc - old_alloc)); + } + + str_idx = path->next_idx ? path->next_idx : top_str; + + /* Temporary modify MCTX. */ + backup_state_log = mctx->state_log; + backup_cur_idx = mctx->input.cur_idx; + mctx->state_log = path->array; + mctx->input.cur_idx = str_idx; + + /* Setup initial node set. */ + context = re_string_context_at (&mctx->input, str_idx - 1, mctx->eflags); + if (str_idx == top_str) + { + err = re_node_set_init_1 (&next_nodes, top_node); + if (__glibc_unlikely (err != REG_NOERROR)) + return err; + err = check_arrival_expand_ecl (dfa, &next_nodes, subexp_num, type); + if (__glibc_unlikely (err != REG_NOERROR)) + { + re_node_set_free (&next_nodes); + return err; + } + } + else + { + cur_state = mctx->state_log[str_idx]; + if (cur_state && cur_state->has_backref) + { + err = re_node_set_init_copy (&next_nodes, &cur_state->nodes); + if (__glibc_unlikely (err != REG_NOERROR)) + return err; + } + else + re_node_set_init_empty (&next_nodes); + } + if (str_idx == top_str || (cur_state && cur_state->has_backref)) + { + if (next_nodes.nelem) + { + err = expand_bkref_cache (mctx, &next_nodes, str_idx, + subexp_num, type); + if (__glibc_unlikely (err != REG_NOERROR)) + { + re_node_set_free (&next_nodes); + return err; + } + } + cur_state = re_acquire_state_context (&err, dfa, &next_nodes, context); + if (__glibc_unlikely (cur_state == NULL && err != REG_NOERROR)) + { + re_node_set_free (&next_nodes); + return err; + } + mctx->state_log[str_idx] = cur_state; + } + + for (null_cnt = 0; str_idx < last_str && null_cnt <= mctx->max_mb_elem_len;) + { + re_node_set_empty (&next_nodes); + if (mctx->state_log[str_idx + 1]) + { + err = re_node_set_merge (&next_nodes, + &mctx->state_log[str_idx + 1]->nodes); + if (__glibc_unlikely (err != REG_NOERROR)) + { + re_node_set_free (&next_nodes); + return err; + } + } + if (cur_state) + { + err = check_arrival_add_next_nodes (mctx, str_idx, + &cur_state->non_eps_nodes, + &next_nodes); + if (__glibc_unlikely (err != REG_NOERROR)) + { + re_node_set_free (&next_nodes); + return err; + } + } + ++str_idx; + if (next_nodes.nelem) + { + err = check_arrival_expand_ecl (dfa, &next_nodes, subexp_num, type); + if (__glibc_unlikely (err != REG_NOERROR)) + { + re_node_set_free (&next_nodes); + return err; + } + err = expand_bkref_cache (mctx, &next_nodes, str_idx, + subexp_num, type); + if (__glibc_unlikely (err != REG_NOERROR)) + { + re_node_set_free (&next_nodes); + return err; + } + } + context = re_string_context_at (&mctx->input, str_idx - 1, mctx->eflags); + cur_state = re_acquire_state_context (&err, dfa, &next_nodes, context); + if (__glibc_unlikely (cur_state == NULL && err != REG_NOERROR)) + { + re_node_set_free (&next_nodes); + return err; + } + mctx->state_log[str_idx] = cur_state; + null_cnt = cur_state == NULL ? null_cnt + 1 : 0; + } + re_node_set_free (&next_nodes); + cur_nodes = (mctx->state_log[last_str] == NULL ? NULL + : &mctx->state_log[last_str]->nodes); + path->next_idx = str_idx; + + /* Fix MCTX. */ + mctx->state_log = backup_state_log; + mctx->input.cur_idx = backup_cur_idx; + + /* Then check the current node set has the node LAST_NODE. */ + if (cur_nodes != NULL && re_node_set_contains (cur_nodes, last_node)) + return REG_NOERROR; + + return REG_NOMATCH; +} + +/* Helper functions for check_arrival. */ + +/* Calculate the destination nodes of CUR_NODES at STR_IDX, and append them + to NEXT_NODES. + TODO: This function is similar to the functions transit_state*(), + however this function has many additional works. + Can't we unify them? */ + +static reg_errcode_t +__attribute_warn_unused_result__ +check_arrival_add_next_nodes (re_match_context_t *mctx, Idx str_idx, + re_node_set *cur_nodes, re_node_set *next_nodes) +{ + const re_dfa_t *const dfa = mctx->dfa; + bool ok; + Idx cur_idx; +#ifdef RE_ENABLE_I18N + reg_errcode_t err = REG_NOERROR; +#endif + re_node_set union_set; + re_node_set_init_empty (&union_set); + for (cur_idx = 0; cur_idx < cur_nodes->nelem; ++cur_idx) + { + int naccepted = 0; + Idx cur_node = cur_nodes->elems[cur_idx]; +#ifdef DEBUG + re_token_type_t type = dfa->nodes[cur_node].type; + assert (!IS_EPSILON_NODE (type)); +#endif +#ifdef RE_ENABLE_I18N + /* If the node may accept "multi byte". */ + if (dfa->nodes[cur_node].accept_mb) + { + naccepted = check_node_accept_bytes (dfa, cur_node, &mctx->input, + str_idx); + if (naccepted > 1) + { + re_dfastate_t *dest_state; + Idx next_node = dfa->nexts[cur_node]; + Idx next_idx = str_idx + naccepted; + dest_state = mctx->state_log[next_idx]; + re_node_set_empty (&union_set); + if (dest_state) + { + err = re_node_set_merge (&union_set, &dest_state->nodes); + if (__glibc_unlikely (err != REG_NOERROR)) + { + re_node_set_free (&union_set); + return err; + } + } + ok = re_node_set_insert (&union_set, next_node); + if (__glibc_unlikely (! ok)) + { + re_node_set_free (&union_set); + return REG_ESPACE; + } + mctx->state_log[next_idx] = re_acquire_state (&err, dfa, + &union_set); + if (__glibc_unlikely (mctx->state_log[next_idx] == NULL + && err != REG_NOERROR)) + { + re_node_set_free (&union_set); + return err; + } + } + } +#endif /* RE_ENABLE_I18N */ + if (naccepted + || check_node_accept (mctx, dfa->nodes + cur_node, str_idx)) + { + ok = re_node_set_insert (next_nodes, dfa->nexts[cur_node]); + if (__glibc_unlikely (! ok)) + { + re_node_set_free (&union_set); + return REG_ESPACE; + } + } + } + re_node_set_free (&union_set); + return REG_NOERROR; +} + +/* For all the nodes in CUR_NODES, add the epsilon closures of them to + CUR_NODES, however exclude the nodes which are: + - inside the sub expression whose number is EX_SUBEXP, if FL_OPEN. + - out of the sub expression whose number is EX_SUBEXP, if !FL_OPEN. +*/ + +static reg_errcode_t +check_arrival_expand_ecl (const re_dfa_t *dfa, re_node_set *cur_nodes, + Idx ex_subexp, int type) +{ + reg_errcode_t err; + Idx idx, outside_node; + re_node_set new_nodes; +#ifdef DEBUG + assert (cur_nodes->nelem); +#endif + err = re_node_set_alloc (&new_nodes, cur_nodes->nelem); + if (__glibc_unlikely (err != REG_NOERROR)) + return err; + /* Create a new node set NEW_NODES with the nodes which are epsilon + closures of the node in CUR_NODES. */ + + for (idx = 0; idx < cur_nodes->nelem; ++idx) + { + Idx cur_node = cur_nodes->elems[idx]; + const re_node_set *eclosure = dfa->eclosures + cur_node; + outside_node = find_subexp_node (dfa, eclosure, ex_subexp, type); + if (outside_node == -1) + { + /* There are no problematic nodes, just merge them. */ + err = re_node_set_merge (&new_nodes, eclosure); + if (__glibc_unlikely (err != REG_NOERROR)) + { + re_node_set_free (&new_nodes); + return err; + } + } + else + { + /* There are problematic nodes, re-calculate incrementally. */ + err = check_arrival_expand_ecl_sub (dfa, &new_nodes, cur_node, + ex_subexp, type); + if (__glibc_unlikely (err != REG_NOERROR)) + { + re_node_set_free (&new_nodes); + return err; + } + } + } + re_node_set_free (cur_nodes); + *cur_nodes = new_nodes; + return REG_NOERROR; +} + +/* Helper function for check_arrival_expand_ecl. + Check incrementally the epsilon closure of TARGET, and if it isn't + problematic append it to DST_NODES. */ + +static reg_errcode_t +__attribute_warn_unused_result__ +check_arrival_expand_ecl_sub (const re_dfa_t *dfa, re_node_set *dst_nodes, + Idx target, Idx ex_subexp, int type) +{ + Idx cur_node; + for (cur_node = target; !re_node_set_contains (dst_nodes, cur_node);) + { + bool ok; + + if (dfa->nodes[cur_node].type == type + && dfa->nodes[cur_node].opr.idx == ex_subexp) + { + if (type == OP_CLOSE_SUBEXP) + { + ok = re_node_set_insert (dst_nodes, cur_node); + if (__glibc_unlikely (! ok)) + return REG_ESPACE; + } + break; + } + ok = re_node_set_insert (dst_nodes, cur_node); + if (__glibc_unlikely (! ok)) + return REG_ESPACE; + if (dfa->edests[cur_node].nelem == 0) + break; + if (dfa->edests[cur_node].nelem == 2) + { + reg_errcode_t err; + err = check_arrival_expand_ecl_sub (dfa, dst_nodes, + dfa->edests[cur_node].elems[1], + ex_subexp, type); + if (__glibc_unlikely (err != REG_NOERROR)) + return err; + } + cur_node = dfa->edests[cur_node].elems[0]; + } + return REG_NOERROR; +} + + +/* For all the back references in the current state, calculate the + destination of the back references by the appropriate entry + in MCTX->BKREF_ENTS. */ + +static reg_errcode_t +__attribute_warn_unused_result__ +expand_bkref_cache (re_match_context_t *mctx, re_node_set *cur_nodes, + Idx cur_str, Idx subexp_num, int type) +{ + const re_dfa_t *const dfa = mctx->dfa; + reg_errcode_t err; + Idx cache_idx_start = search_cur_bkref_entry (mctx, cur_str); + struct re_backref_cache_entry *ent; + + if (cache_idx_start == -1) + return REG_NOERROR; + + restart: + ent = mctx->bkref_ents + cache_idx_start; + do + { + Idx to_idx, next_node; + + /* Is this entry ENT is appropriate? */ + if (!re_node_set_contains (cur_nodes, ent->node)) + continue; /* No. */ + + to_idx = cur_str + ent->subexp_to - ent->subexp_from; + /* Calculate the destination of the back reference, and append it + to MCTX->STATE_LOG. */ + if (to_idx == cur_str) + { + /* The backreference did epsilon transit, we must re-check all the + node in the current state. */ + re_node_set new_dests; + reg_errcode_t err2, err3; + next_node = dfa->edests[ent->node].elems[0]; + if (re_node_set_contains (cur_nodes, next_node)) + continue; + err = re_node_set_init_1 (&new_dests, next_node); + err2 = check_arrival_expand_ecl (dfa, &new_dests, subexp_num, type); + err3 = re_node_set_merge (cur_nodes, &new_dests); + re_node_set_free (&new_dests); + if (__glibc_unlikely (err != REG_NOERROR || err2 != REG_NOERROR + || err3 != REG_NOERROR)) + { + err = (err != REG_NOERROR ? err + : (err2 != REG_NOERROR ? err2 : err3)); + return err; + } + /* TODO: It is still inefficient... */ + goto restart; + } + else + { + re_node_set union_set; + next_node = dfa->nexts[ent->node]; + if (mctx->state_log[to_idx]) + { + bool ok; + if (re_node_set_contains (&mctx->state_log[to_idx]->nodes, + next_node)) + continue; + err = re_node_set_init_copy (&union_set, + &mctx->state_log[to_idx]->nodes); + ok = re_node_set_insert (&union_set, next_node); + if (__glibc_unlikely (err != REG_NOERROR || ! ok)) + { + re_node_set_free (&union_set); + err = err != REG_NOERROR ? err : REG_ESPACE; + return err; + } + } + else + { + err = re_node_set_init_1 (&union_set, next_node); + if (__glibc_unlikely (err != REG_NOERROR)) + return err; + } + mctx->state_log[to_idx] = re_acquire_state (&err, dfa, &union_set); + re_node_set_free (&union_set); + if (__glibc_unlikely (mctx->state_log[to_idx] == NULL + && err != REG_NOERROR)) + return err; + } + } + while (ent++->more); + return REG_NOERROR; +} + +/* Build transition table for the state. + Return true if successful. */ + +static bool +build_trtable (const re_dfa_t *dfa, re_dfastate_t *state) +{ + reg_errcode_t err; + Idx i, j; + int ch; + bool need_word_trtable = false; + bitset_word_t elem, mask; + bool dests_node_malloced = false; + bool dest_states_malloced = false; + Idx ndests; /* Number of the destination states from 'state'. */ + re_dfastate_t **trtable; + re_dfastate_t **dest_states = NULL, **dest_states_word, **dest_states_nl; + re_node_set follows, *dests_node; + bitset_t *dests_ch; + bitset_t acceptable; + + struct dests_alloc + { + re_node_set dests_node[SBC_MAX]; + bitset_t dests_ch[SBC_MAX]; + } *dests_alloc; + + /* We build DFA states which corresponds to the destination nodes + from 'state'. 'dests_node[i]' represents the nodes which i-th + destination state contains, and 'dests_ch[i]' represents the + characters which i-th destination state accepts. */ + if (__libc_use_alloca (sizeof (struct dests_alloc))) + dests_alloc = (struct dests_alloc *) alloca (sizeof (struct dests_alloc)); + else + { + dests_alloc = re_malloc (struct dests_alloc, 1); + if (__glibc_unlikely (dests_alloc == NULL)) + return false; + dests_node_malloced = true; + } + dests_node = dests_alloc->dests_node; + dests_ch = dests_alloc->dests_ch; + + /* Initialize transition table. */ + state->word_trtable = state->trtable = NULL; + + /* At first, group all nodes belonging to 'state' into several + destinations. */ + ndests = group_nodes_into_DFAstates (dfa, state, dests_node, dests_ch); + if (__glibc_unlikely (ndests <= 0)) + { + if (dests_node_malloced) + re_free (dests_alloc); + /* Return false in case of an error, true otherwise. */ + if (ndests == 0) + { + state->trtable = (re_dfastate_t **) + calloc (sizeof (re_dfastate_t *), SBC_MAX); + if (__glibc_unlikely (state->trtable == NULL)) + return false; + return true; + } + return false; + } + + err = re_node_set_alloc (&follows, ndests + 1); + if (__glibc_unlikely (err != REG_NOERROR)) + goto out_free; + + /* Avoid arithmetic overflow in size calculation. */ + size_t ndests_max + = ((SIZE_MAX - (sizeof (re_node_set) + sizeof (bitset_t)) * SBC_MAX) + / (3 * sizeof (re_dfastate_t *))); + if (__glibc_unlikely (ndests_max < ndests)) + goto out_free; + + if (__libc_use_alloca ((sizeof (re_node_set) + sizeof (bitset_t)) * SBC_MAX + + ndests * 3 * sizeof (re_dfastate_t *))) + dest_states = (re_dfastate_t **) + alloca (ndests * 3 * sizeof (re_dfastate_t *)); + else + { + dest_states = re_malloc (re_dfastate_t *, ndests * 3); + if (__glibc_unlikely (dest_states == NULL)) + { +out_free: + if (dest_states_malloced) + re_free (dest_states); + re_node_set_free (&follows); + for (i = 0; i < ndests; ++i) + re_node_set_free (dests_node + i); + if (dests_node_malloced) + re_free (dests_alloc); + return false; + } + dest_states_malloced = true; + } + dest_states_word = dest_states + ndests; + dest_states_nl = dest_states_word + ndests; + bitset_empty (acceptable); + + /* Then build the states for all destinations. */ + for (i = 0; i < ndests; ++i) + { + Idx next_node; + re_node_set_empty (&follows); + /* Merge the follows of this destination states. */ + for (j = 0; j < dests_node[i].nelem; ++j) + { + next_node = dfa->nexts[dests_node[i].elems[j]]; + if (next_node != -1) + { + err = re_node_set_merge (&follows, dfa->eclosures + next_node); + if (__glibc_unlikely (err != REG_NOERROR)) + goto out_free; + } + } + dest_states[i] = re_acquire_state_context (&err, dfa, &follows, 0); + if (__glibc_unlikely (dest_states[i] == NULL && err != REG_NOERROR)) + goto out_free; + /* If the new state has context constraint, + build appropriate states for these contexts. */ + if (dest_states[i]->has_constraint) + { + dest_states_word[i] = re_acquire_state_context (&err, dfa, &follows, + CONTEXT_WORD); + if (__glibc_unlikely (dest_states_word[i] == NULL + && err != REG_NOERROR)) + goto out_free; + + if (dest_states[i] != dest_states_word[i] && dfa->mb_cur_max > 1) + need_word_trtable = true; + + dest_states_nl[i] = re_acquire_state_context (&err, dfa, &follows, + CONTEXT_NEWLINE); + if (__glibc_unlikely (dest_states_nl[i] == NULL && err != REG_NOERROR)) + goto out_free; + } + else + { + dest_states_word[i] = dest_states[i]; + dest_states_nl[i] = dest_states[i]; + } + bitset_merge (acceptable, dests_ch[i]); + } + + if (!__glibc_unlikely (need_word_trtable)) + { + /* We don't care about whether the following character is a word + character, or we are in a single-byte character set so we can + discern by looking at the character code: allocate a + 256-entry transition table. */ + trtable = state->trtable = + (re_dfastate_t **) calloc (sizeof (re_dfastate_t *), SBC_MAX); + if (__glibc_unlikely (trtable == NULL)) + goto out_free; + + /* For all characters ch...: */ + for (i = 0; i < BITSET_WORDS; ++i) + for (ch = i * BITSET_WORD_BITS, elem = acceptable[i], mask = 1; + elem; + mask <<= 1, elem >>= 1, ++ch) + if (__glibc_unlikely (elem & 1)) + { + /* There must be exactly one destination which accepts + character ch. See group_nodes_into_DFAstates. */ + for (j = 0; (dests_ch[j][i] & mask) == 0; ++j) + ; + + /* j-th destination accepts the word character ch. */ + if (dfa->word_char[i] & mask) + trtable[ch] = dest_states_word[j]; + else + trtable[ch] = dest_states[j]; + } + } + else + { + /* We care about whether the following character is a word + character, and we are in a multi-byte character set: discern + by looking at the character code: build two 256-entry + transition tables, one starting at trtable[0] and one + starting at trtable[SBC_MAX]. */ + trtable = state->word_trtable = + (re_dfastate_t **) calloc (sizeof (re_dfastate_t *), 2 * SBC_MAX); + if (__glibc_unlikely (trtable == NULL)) + goto out_free; + + /* For all characters ch...: */ + for (i = 0; i < BITSET_WORDS; ++i) + for (ch = i * BITSET_WORD_BITS, elem = acceptable[i], mask = 1; + elem; + mask <<= 1, elem >>= 1, ++ch) + if (__glibc_unlikely (elem & 1)) + { + /* There must be exactly one destination which accepts + character ch. See group_nodes_into_DFAstates. */ + for (j = 0; (dests_ch[j][i] & mask) == 0; ++j) + ; + + /* j-th destination accepts the word character ch. */ + trtable[ch] = dest_states[j]; + trtable[ch + SBC_MAX] = dest_states_word[j]; + } + } + + /* new line */ + if (bitset_contain (acceptable, NEWLINE_CHAR)) + { + /* The current state accepts newline character. */ + for (j = 0; j < ndests; ++j) + if (bitset_contain (dests_ch[j], NEWLINE_CHAR)) + { + /* k-th destination accepts newline character. */ + trtable[NEWLINE_CHAR] = dest_states_nl[j]; + if (need_word_trtable) + trtable[NEWLINE_CHAR + SBC_MAX] = dest_states_nl[j]; + /* There must be only one destination which accepts + newline. See group_nodes_into_DFAstates. */ + break; + } + } + + if (dest_states_malloced) + re_free (dest_states); + + re_node_set_free (&follows); + for (i = 0; i < ndests; ++i) + re_node_set_free (dests_node + i); + + if (dests_node_malloced) + re_free (dests_alloc); + + return true; +} + +/* Group all nodes belonging to STATE into several destinations. + Then for all destinations, set the nodes belonging to the destination + to DESTS_NODE[i] and set the characters accepted by the destination + to DEST_CH[i]. This function return the number of destinations. */ + +static Idx +group_nodes_into_DFAstates (const re_dfa_t *dfa, const re_dfastate_t *state, + re_node_set *dests_node, bitset_t *dests_ch) +{ + reg_errcode_t err; + bool ok; + Idx i, j, k; + Idx ndests; /* Number of the destinations from 'state'. */ + bitset_t accepts; /* Characters a node can accept. */ + const re_node_set *cur_nodes = &state->nodes; + bitset_empty (accepts); + ndests = 0; + + /* For all the nodes belonging to 'state', */ + for (i = 0; i < cur_nodes->nelem; ++i) + { + re_token_t *node = &dfa->nodes[cur_nodes->elems[i]]; + re_token_type_t type = node->type; + unsigned int constraint = node->constraint; + + /* Enumerate all single byte character this node can accept. */ + if (type == CHARACTER) + bitset_set (accepts, node->opr.c); + else if (type == SIMPLE_BRACKET) + { + bitset_merge (accepts, node->opr.sbcset); + } + else if (type == OP_PERIOD) + { +#ifdef RE_ENABLE_I18N + if (dfa->mb_cur_max > 1) + bitset_merge (accepts, dfa->sb_char); + else +#endif + bitset_set_all (accepts); + if (!(dfa->syntax & RE_DOT_NEWLINE)) + bitset_clear (accepts, '\n'); + if (dfa->syntax & RE_DOT_NOT_NULL) + bitset_clear (accepts, '\0'); + } +#ifdef RE_ENABLE_I18N + else if (type == OP_UTF8_PERIOD) + { + if (ASCII_CHARS % BITSET_WORD_BITS == 0) + memset (accepts, -1, ASCII_CHARS / CHAR_BIT); + else + bitset_merge (accepts, utf8_sb_map); + if (!(dfa->syntax & RE_DOT_NEWLINE)) + bitset_clear (accepts, '\n'); + if (dfa->syntax & RE_DOT_NOT_NULL) + bitset_clear (accepts, '\0'); + } +#endif + else + continue; + + /* Check the 'accepts' and sift the characters which are not + match it the context. */ + if (constraint) + { + if (constraint & NEXT_NEWLINE_CONSTRAINT) + { + bool accepts_newline = bitset_contain (accepts, NEWLINE_CHAR); + bitset_empty (accepts); + if (accepts_newline) + bitset_set (accepts, NEWLINE_CHAR); + else + continue; + } + if (constraint & NEXT_ENDBUF_CONSTRAINT) + { + bitset_empty (accepts); + continue; + } + + if (constraint & NEXT_WORD_CONSTRAINT) + { + bitset_word_t any_set = 0; + if (type == CHARACTER && !node->word_char) + { + bitset_empty (accepts); + continue; + } +#ifdef RE_ENABLE_I18N + if (dfa->mb_cur_max > 1) + for (j = 0; j < BITSET_WORDS; ++j) + any_set |= (accepts[j] &= (dfa->word_char[j] | ~dfa->sb_char[j])); + else +#endif + for (j = 0; j < BITSET_WORDS; ++j) + any_set |= (accepts[j] &= dfa->word_char[j]); + if (!any_set) + continue; + } + if (constraint & NEXT_NOTWORD_CONSTRAINT) + { + bitset_word_t any_set = 0; + if (type == CHARACTER && node->word_char) + { + bitset_empty (accepts); + continue; + } +#ifdef RE_ENABLE_I18N + if (dfa->mb_cur_max > 1) + for (j = 0; j < BITSET_WORDS; ++j) + any_set |= (accepts[j] &= ~(dfa->word_char[j] & dfa->sb_char[j])); + else +#endif + for (j = 0; j < BITSET_WORDS; ++j) + any_set |= (accepts[j] &= ~dfa->word_char[j]); + if (!any_set) + continue; + } + } + + /* Then divide 'accepts' into DFA states, or create a new + state. Above, we make sure that accepts is not empty. */ + for (j = 0; j < ndests; ++j) + { + bitset_t intersec; /* Intersection sets, see below. */ + bitset_t remains; + /* Flags, see below. */ + bitset_word_t has_intersec, not_subset, not_consumed; + + /* Optimization, skip if this state doesn't accept the character. */ + if (type == CHARACTER && !bitset_contain (dests_ch[j], node->opr.c)) + continue; + + /* Enumerate the intersection set of this state and 'accepts'. */ + has_intersec = 0; + for (k = 0; k < BITSET_WORDS; ++k) + has_intersec |= intersec[k] = accepts[k] & dests_ch[j][k]; + /* And skip if the intersection set is empty. */ + if (!has_intersec) + continue; + + /* Then check if this state is a subset of 'accepts'. */ + not_subset = not_consumed = 0; + for (k = 0; k < BITSET_WORDS; ++k) + { + not_subset |= remains[k] = ~accepts[k] & dests_ch[j][k]; + not_consumed |= accepts[k] = accepts[k] & ~dests_ch[j][k]; + } + + /* If this state isn't a subset of 'accepts', create a + new group state, which has the 'remains'. */ + if (not_subset) + { + bitset_copy (dests_ch[ndests], remains); + bitset_copy (dests_ch[j], intersec); + err = re_node_set_init_copy (dests_node + ndests, &dests_node[j]); + if (__glibc_unlikely (err != REG_NOERROR)) + goto error_return; + ++ndests; + } + + /* Put the position in the current group. */ + ok = re_node_set_insert (&dests_node[j], cur_nodes->elems[i]); + if (__glibc_unlikely (! ok)) + goto error_return; + + /* If all characters are consumed, go to next node. */ + if (!not_consumed) + break; + } + /* Some characters remain, create a new group. */ + if (j == ndests) + { + bitset_copy (dests_ch[ndests], accepts); + err = re_node_set_init_1 (dests_node + ndests, cur_nodes->elems[i]); + if (__glibc_unlikely (err != REG_NOERROR)) + goto error_return; + ++ndests; + bitset_empty (accepts); + } + } + return ndests; + error_return: + for (j = 0; j < ndests; ++j) + re_node_set_free (dests_node + j); + return -1; +} + +#ifdef RE_ENABLE_I18N +/* Check how many bytes the node 'dfa->nodes[node_idx]' accepts. + Return the number of the bytes the node accepts. + STR_IDX is the current index of the input string. + + This function handles the nodes which can accept one character, or + one collating element like '.', '[a-z]', opposite to the other nodes + can only accept one byte. */ + +# ifdef _LIBC +# include <locale/weight.h> +# endif + +static int +check_node_accept_bytes (const re_dfa_t *dfa, Idx node_idx, + const re_string_t *input, Idx str_idx) +{ + const re_token_t *node = dfa->nodes + node_idx; + int char_len, elem_len; + Idx i; + + if (__glibc_unlikely (node->type == OP_UTF8_PERIOD)) + { + unsigned char c = re_string_byte_at (input, str_idx), d; + if (__glibc_likely (c < 0xc2)) + return 0; + + if (str_idx + 2 > input->len) + return 0; + + d = re_string_byte_at (input, str_idx + 1); + if (c < 0xe0) + return (d < 0x80 || d > 0xbf) ? 0 : 2; + else if (c < 0xf0) + { + char_len = 3; + if (c == 0xe0 && d < 0xa0) + return 0; + } + else if (c < 0xf8) + { + char_len = 4; + if (c == 0xf0 && d < 0x90) + return 0; + } + else if (c < 0xfc) + { + char_len = 5; + if (c == 0xf8 && d < 0x88) + return 0; + } + else if (c < 0xfe) + { + char_len = 6; + if (c == 0xfc && d < 0x84) + return 0; + } + else + return 0; + + if (str_idx + char_len > input->len) + return 0; + + for (i = 1; i < char_len; ++i) + { + d = re_string_byte_at (input, str_idx + i); + if (d < 0x80 || d > 0xbf) + return 0; + } + return char_len; + } + + char_len = re_string_char_size_at (input, str_idx); + if (node->type == OP_PERIOD) + { + if (char_len <= 1) + return 0; + /* FIXME: I don't think this if is needed, as both '\n' + and '\0' are char_len == 1. */ + /* '.' accepts any one character except the following two cases. */ + if ((!(dfa->syntax & RE_DOT_NEWLINE) && + re_string_byte_at (input, str_idx) == '\n') || + ((dfa->syntax & RE_DOT_NOT_NULL) && + re_string_byte_at (input, str_idx) == '\0')) + return 0; + return char_len; + } + + elem_len = re_string_elem_size_at (input, str_idx); + if ((elem_len <= 1 && char_len <= 1) || char_len == 0) + return 0; + + if (node->type == COMPLEX_BRACKET) + { + const re_charset_t *cset = node->opr.mbcset; +# ifdef _LIBC + const unsigned char *pin + = ((const unsigned char *) re_string_get_buffer (input) + str_idx); + Idx j; + uint32_t nrules; +# endif /* _LIBC */ + int match_len = 0; + wchar_t wc = ((cset->nranges || cset->nchar_classes || cset->nmbchars) + ? re_string_wchar_at (input, str_idx) : 0); + + /* match with multibyte character? */ + for (i = 0; i < cset->nmbchars; ++i) + if (wc == cset->mbchars[i]) + { + match_len = char_len; + goto check_node_accept_bytes_match; + } + /* match with character_class? */ + for (i = 0; i < cset->nchar_classes; ++i) + { + wctype_t wt = cset->char_classes[i]; + if (__iswctype (wc, wt)) + { + match_len = char_len; + goto check_node_accept_bytes_match; + } + } + +# ifdef _LIBC + nrules = _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES); + if (nrules != 0) + { + unsigned int in_collseq = 0; + const int32_t *table, *indirect; + const unsigned char *weights, *extra; + const char *collseqwc; + + /* match with collating_symbol? */ + if (cset->ncoll_syms) + extra = (const unsigned char *) + _NL_CURRENT (LC_COLLATE, _NL_COLLATE_SYMB_EXTRAMB); + for (i = 0; i < cset->ncoll_syms; ++i) + { + const unsigned char *coll_sym = extra + cset->coll_syms[i]; + /* Compare the length of input collating element and + the length of current collating element. */ + if (*coll_sym != elem_len) + continue; + /* Compare each bytes. */ + for (j = 0; j < *coll_sym; j++) + if (pin[j] != coll_sym[1 + j]) + break; + if (j == *coll_sym) + { + /* Match if every bytes is equal. */ + match_len = j; + goto check_node_accept_bytes_match; + } + } + + if (cset->nranges) + { + if (elem_len <= char_len) + { + collseqwc = _NL_CURRENT (LC_COLLATE, _NL_COLLATE_COLLSEQWC); + in_collseq = __collseq_table_lookup (collseqwc, wc); + } + else + in_collseq = find_collation_sequence_value (pin, elem_len); + } + /* match with range expression? */ + /* FIXME: Implement rational ranges here, too. */ + for (i = 0; i < cset->nranges; ++i) + if (cset->range_starts[i] <= in_collseq + && in_collseq <= cset->range_ends[i]) + { + match_len = elem_len; + goto check_node_accept_bytes_match; + } + + /* match with equivalence_class? */ + if (cset->nequiv_classes) + { + const unsigned char *cp = pin; + table = (const int32_t *) + _NL_CURRENT (LC_COLLATE, _NL_COLLATE_TABLEMB); + weights = (const unsigned char *) + _NL_CURRENT (LC_COLLATE, _NL_COLLATE_WEIGHTMB); + extra = (const unsigned char *) + _NL_CURRENT (LC_COLLATE, _NL_COLLATE_EXTRAMB); + indirect = (const int32_t *) + _NL_CURRENT (LC_COLLATE, _NL_COLLATE_INDIRECTMB); + int32_t idx = findidx (table, indirect, extra, &cp, elem_len); + int32_t rule = idx >> 24; + idx &= 0xffffff; + if (idx > 0) + { + size_t weight_len = weights[idx]; + for (i = 0; i < cset->nequiv_classes; ++i) + { + int32_t equiv_class_idx = cset->equiv_classes[i]; + int32_t equiv_class_rule = equiv_class_idx >> 24; + equiv_class_idx &= 0xffffff; + if (weights[equiv_class_idx] == weight_len + && equiv_class_rule == rule + && memcmp (weights + idx + 1, + weights + equiv_class_idx + 1, + weight_len) == 0) + { + match_len = elem_len; + goto check_node_accept_bytes_match; + } + } + } + } + } + else +# endif /* _LIBC */ + { + /* match with range expression? */ + for (i = 0; i < cset->nranges; ++i) + { + if (cset->range_starts[i] <= wc && wc <= cset->range_ends[i]) + { + match_len = char_len; + goto check_node_accept_bytes_match; + } + } + } + check_node_accept_bytes_match: + if (!cset->non_match) + return match_len; + else + { + if (match_len > 0) + return 0; + else + return (elem_len > char_len) ? elem_len : char_len; + } + } + return 0; +} + +# ifdef _LIBC +static unsigned int +find_collation_sequence_value (const unsigned char *mbs, size_t mbs_len) +{ + uint32_t nrules = _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES); + if (nrules == 0) + { + if (mbs_len == 1) + { + /* No valid character. Match it as a single byte character. */ + const unsigned char *collseq = (const unsigned char *) + _NL_CURRENT (LC_COLLATE, _NL_COLLATE_COLLSEQMB); + return collseq[mbs[0]]; + } + return UINT_MAX; + } + else + { + int32_t idx; + const unsigned char *extra = (const unsigned char *) + _NL_CURRENT (LC_COLLATE, _NL_COLLATE_SYMB_EXTRAMB); + int32_t extrasize = (const unsigned char *) + _NL_CURRENT (LC_COLLATE, _NL_COLLATE_SYMB_EXTRAMB + 1) - extra; + + for (idx = 0; idx < extrasize;) + { + int mbs_cnt; + bool found = false; + int32_t elem_mbs_len; + /* Skip the name of collating element name. */ + idx = idx + extra[idx] + 1; + elem_mbs_len = extra[idx++]; + if (mbs_len == elem_mbs_len) + { + for (mbs_cnt = 0; mbs_cnt < elem_mbs_len; ++mbs_cnt) + if (extra[idx + mbs_cnt] != mbs[mbs_cnt]) + break; + if (mbs_cnt == elem_mbs_len) + /* Found the entry. */ + found = true; + } + /* Skip the byte sequence of the collating element. */ + idx += elem_mbs_len; + /* Adjust for the alignment. */ + idx = (idx + 3) & ~3; + /* Skip the collation sequence value. */ + idx += sizeof (uint32_t); + /* Skip the wide char sequence of the collating element. */ + idx = idx + sizeof (uint32_t) * (*(int32_t *) (extra + idx) + 1); + /* If we found the entry, return the sequence value. */ + if (found) + return *(uint32_t *) (extra + idx); + /* Skip the collation sequence value. */ + idx += sizeof (uint32_t); + } + return UINT_MAX; + } +} +# endif /* _LIBC */ +#endif /* RE_ENABLE_I18N */ + +/* Check whether the node accepts the byte which is IDX-th + byte of the INPUT. */ + +static bool +check_node_accept (const re_match_context_t *mctx, const re_token_t *node, + Idx idx) +{ + unsigned char ch; + ch = re_string_byte_at (&mctx->input, idx); + switch (node->type) + { + case CHARACTER: + if (node->opr.c != ch) + return false; + break; + + case SIMPLE_BRACKET: + if (!bitset_contain (node->opr.sbcset, ch)) + return false; + break; + +#ifdef RE_ENABLE_I18N + case OP_UTF8_PERIOD: + if (ch >= ASCII_CHARS) + return false; + FALLTHROUGH; +#endif + case OP_PERIOD: + if ((ch == '\n' && !(mctx->dfa->syntax & RE_DOT_NEWLINE)) + || (ch == '\0' && (mctx->dfa->syntax & RE_DOT_NOT_NULL))) + return false; + break; + + default: + return false; + } + + if (node->constraint) + { + /* The node has constraints. Check whether the current context + satisfies the constraints. */ + unsigned int context = re_string_context_at (&mctx->input, idx, + mctx->eflags); + if (NOT_SATISFY_NEXT_CONSTRAINT (node->constraint, context)) + return false; + } + + return true; +} + +/* Extend the buffers, if the buffers have run out. */ + +static reg_errcode_t +__attribute_warn_unused_result__ +extend_buffers (re_match_context_t *mctx, int min_len) +{ + reg_errcode_t ret; + re_string_t *pstr = &mctx->input; + + /* Avoid overflow. */ + if (__glibc_unlikely (MIN (IDX_MAX, SIZE_MAX / sizeof (re_dfastate_t *)) / 2 + <= pstr->bufs_len)) + return REG_ESPACE; + + /* Double the lengths of the buffers, but allocate at least MIN_LEN. */ + ret = re_string_realloc_buffers (pstr, + MAX (min_len, + MIN (pstr->len, pstr->bufs_len * 2))); + if (__glibc_unlikely (ret != REG_NOERROR)) + return ret; + + if (mctx->state_log != NULL) + { + /* And double the length of state_log. */ + /* XXX We have no indication of the size of this buffer. If this + allocation fail we have no indication that the state_log array + does not have the right size. */ + re_dfastate_t **new_array = re_realloc (mctx->state_log, re_dfastate_t *, + pstr->bufs_len + 1); + if (__glibc_unlikely (new_array == NULL)) + return REG_ESPACE; + mctx->state_log = new_array; + } + + /* Then reconstruct the buffers. */ + if (pstr->icase) + { +#ifdef RE_ENABLE_I18N + if (pstr->mb_cur_max > 1) + { + ret = build_wcs_upper_buffer (pstr); + if (__glibc_unlikely (ret != REG_NOERROR)) + return ret; + } + else +#endif /* RE_ENABLE_I18N */ + build_upper_buffer (pstr); + } + else + { +#ifdef RE_ENABLE_I18N + if (pstr->mb_cur_max > 1) + build_wcs_buffer (pstr); + else +#endif /* RE_ENABLE_I18N */ + { + if (pstr->trans != NULL) + re_string_translate_buffer (pstr); + } + } + return REG_NOERROR; +} + + +/* Functions for matching context. */ + +/* Initialize MCTX. */ + +static reg_errcode_t +__attribute_warn_unused_result__ +match_ctx_init (re_match_context_t *mctx, int eflags, Idx n) +{ + mctx->eflags = eflags; + mctx->match_last = -1; + if (n > 0) + { + /* Avoid overflow. */ + size_t max_object_size = + MAX (sizeof (struct re_backref_cache_entry), + sizeof (re_sub_match_top_t *)); + if (__glibc_unlikely (MIN (IDX_MAX, SIZE_MAX / max_object_size) < n)) + return REG_ESPACE; + + mctx->bkref_ents = re_malloc (struct re_backref_cache_entry, n); + mctx->sub_tops = re_malloc (re_sub_match_top_t *, n); + if (__glibc_unlikely (mctx->bkref_ents == NULL || mctx->sub_tops == NULL)) + return REG_ESPACE; + } + /* Already zero-ed by the caller. + else + mctx->bkref_ents = NULL; + mctx->nbkref_ents = 0; + mctx->nsub_tops = 0; */ + mctx->abkref_ents = n; + mctx->max_mb_elem_len = 1; + mctx->asub_tops = n; + return REG_NOERROR; +} + +/* Clean the entries which depend on the current input in MCTX. + This function must be invoked when the matcher changes the start index + of the input, or changes the input string. */ + +static void +match_ctx_clean (re_match_context_t *mctx) +{ + Idx st_idx; + for (st_idx = 0; st_idx < mctx->nsub_tops; ++st_idx) + { + Idx sl_idx; + re_sub_match_top_t *top = mctx->sub_tops[st_idx]; + for (sl_idx = 0; sl_idx < top->nlasts; ++sl_idx) + { + re_sub_match_last_t *last = top->lasts[sl_idx]; + re_free (last->path.array); + re_free (last); + } + re_free (top->lasts); + if (top->path) + { + re_free (top->path->array); + re_free (top->path); + } + re_free (top); + } + + mctx->nsub_tops = 0; + mctx->nbkref_ents = 0; +} + +/* Free all the memory associated with MCTX. */ + +static void +match_ctx_free (re_match_context_t *mctx) +{ + /* First, free all the memory associated with MCTX->SUB_TOPS. */ + match_ctx_clean (mctx); + re_free (mctx->sub_tops); + re_free (mctx->bkref_ents); +} + +/* Add a new backreference entry to MCTX. + Note that we assume that caller never call this function with duplicate + entry, and call with STR_IDX which isn't smaller than any existing entry. +*/ + +static reg_errcode_t +__attribute_warn_unused_result__ +match_ctx_add_entry (re_match_context_t *mctx, Idx node, Idx str_idx, Idx from, + Idx to) +{ + if (mctx->nbkref_ents >= mctx->abkref_ents) + { + struct re_backref_cache_entry* new_entry; + new_entry = re_realloc (mctx->bkref_ents, struct re_backref_cache_entry, + mctx->abkref_ents * 2); + if (__glibc_unlikely (new_entry == NULL)) + { + re_free (mctx->bkref_ents); + return REG_ESPACE; + } + mctx->bkref_ents = new_entry; + memset (mctx->bkref_ents + mctx->nbkref_ents, '\0', + sizeof (struct re_backref_cache_entry) * mctx->abkref_ents); + mctx->abkref_ents *= 2; + } + if (mctx->nbkref_ents > 0 + && mctx->bkref_ents[mctx->nbkref_ents - 1].str_idx == str_idx) + mctx->bkref_ents[mctx->nbkref_ents - 1].more = 1; + + mctx->bkref_ents[mctx->nbkref_ents].node = node; + mctx->bkref_ents[mctx->nbkref_ents].str_idx = str_idx; + mctx->bkref_ents[mctx->nbkref_ents].subexp_from = from; + mctx->bkref_ents[mctx->nbkref_ents].subexp_to = to; + + /* This is a cache that saves negative results of check_dst_limits_calc_pos. + If bit N is clear, means that this entry won't epsilon-transition to + an OP_OPEN_SUBEXP or OP_CLOSE_SUBEXP for the N+1-th subexpression. If + it is set, check_dst_limits_calc_pos_1 will recurse and try to find one + such node. + + A backreference does not epsilon-transition unless it is empty, so set + to all zeros if FROM != TO. */ + mctx->bkref_ents[mctx->nbkref_ents].eps_reachable_subexps_map + = (from == to ? -1 : 0); + + mctx->bkref_ents[mctx->nbkref_ents++].more = 0; + if (mctx->max_mb_elem_len < to - from) + mctx->max_mb_elem_len = to - from; + return REG_NOERROR; +} + +/* Return the first entry with the same str_idx, or -1 if none is + found. Note that MCTX->BKREF_ENTS is already sorted by MCTX->STR_IDX. */ + +static Idx +search_cur_bkref_entry (const re_match_context_t *mctx, Idx str_idx) +{ + Idx left, right, mid, last; + last = right = mctx->nbkref_ents; + for (left = 0; left < right;) + { + mid = (left + right) / 2; + if (mctx->bkref_ents[mid].str_idx < str_idx) + left = mid + 1; + else + right = mid; + } + if (left < last && mctx->bkref_ents[left].str_idx == str_idx) + return left; + else + return -1; +} + +/* Register the node NODE, whose type is OP_OPEN_SUBEXP, and which matches + at STR_IDX. */ + +static reg_errcode_t +__attribute_warn_unused_result__ +match_ctx_add_subtop (re_match_context_t *mctx, Idx node, Idx str_idx) +{ +#ifdef DEBUG + assert (mctx->sub_tops != NULL); + assert (mctx->asub_tops > 0); +#endif + if (__glibc_unlikely (mctx->nsub_tops == mctx->asub_tops)) + { + Idx new_asub_tops = mctx->asub_tops * 2; + re_sub_match_top_t **new_array = re_realloc (mctx->sub_tops, + re_sub_match_top_t *, + new_asub_tops); + if (__glibc_unlikely (new_array == NULL)) + return REG_ESPACE; + mctx->sub_tops = new_array; + mctx->asub_tops = new_asub_tops; + } + mctx->sub_tops[mctx->nsub_tops] = calloc (1, sizeof (re_sub_match_top_t)); + if (__glibc_unlikely (mctx->sub_tops[mctx->nsub_tops] == NULL)) + return REG_ESPACE; + mctx->sub_tops[mctx->nsub_tops]->node = node; + mctx->sub_tops[mctx->nsub_tops++]->str_idx = str_idx; + return REG_NOERROR; +} + +/* Register the node NODE, whose type is OP_CLOSE_SUBEXP, and which matches + at STR_IDX, whose corresponding OP_OPEN_SUBEXP is SUB_TOP. */ + +static re_sub_match_last_t * +match_ctx_add_sublast (re_sub_match_top_t *subtop, Idx node, Idx str_idx) +{ + re_sub_match_last_t *new_entry; + if (__glibc_unlikely (subtop->nlasts == subtop->alasts)) + { + Idx new_alasts = 2 * subtop->alasts + 1; + re_sub_match_last_t **new_array = re_realloc (subtop->lasts, + re_sub_match_last_t *, + new_alasts); + if (__glibc_unlikely (new_array == NULL)) + return NULL; + subtop->lasts = new_array; + subtop->alasts = new_alasts; + } + new_entry = calloc (1, sizeof (re_sub_match_last_t)); + if (__glibc_likely (new_entry != NULL)) + { + subtop->lasts[subtop->nlasts] = new_entry; + new_entry->node = node; + new_entry->str_idx = str_idx; + ++subtop->nlasts; + } + return new_entry; +} + +static void +sift_ctx_init (re_sift_context_t *sctx, re_dfastate_t **sifted_sts, + re_dfastate_t **limited_sts, Idx last_node, Idx last_str_idx) +{ + sctx->sifted_states = sifted_sts; + sctx->limited_states = limited_sts; + sctx->last_node = last_node; + sctx->last_str_idx = last_str_idx; + re_node_set_init_empty (&sctx->limits); +} diff --git a/lib/rename.c b/lib/rename.c new file mode 100644 index 0000000..0b457aa --- /dev/null +++ b/lib/rename.c @@ -0,0 +1,475 @@ +/* Work around rename bugs in some systems. + + Copyright (C) 2001-2003, 2005-2006, 2009-2018 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +/* Written by Volker Borchert, Eric Blake. */ + +#include <config.h> + +#include <stdio.h> + +#undef rename + +#if defined _WIN32 && ! defined __CYGWIN__ +/* The mingw rename has problems with trailing slashes; it also + requires use of native Windows calls to allow atomic renames over + existing files. */ + +# include <errno.h> +# include <stdbool.h> +# include <stdlib.h> +# include <sys/stat.h> +# include <unistd.h> + +# define WIN32_LEAN_AND_MEAN +# include <windows.h> + +# include "dirname.h" + +/* Rename the file SRC to DST. This replacement is necessary on + Windows, on which the system rename function will not replace + an existing DST. */ +int +rpl_rename (char const *src, char const *dst) +{ + int error; + size_t src_len = strlen (src); + size_t dst_len = strlen (dst); + char *src_base = last_component (src); + char *dst_base = last_component (dst); + bool src_slash; + bool dst_slash; + bool dst_exists; + struct stat src_st; + struct stat dst_st; + + /* Filter out dot as last component. */ + if (!src_len || !dst_len) + { + errno = ENOENT; + return -1; + } + if (*src_base == '.') + { + size_t len = base_len (src_base); + if (len == 1 || (len == 2 && src_base[1] == '.')) + { + errno = EINVAL; + return -1; + } + } + if (*dst_base == '.') + { + size_t len = base_len (dst_base); + if (len == 1 || (len == 2 && dst_base[1] == '.')) + { + errno = EINVAL; + return -1; + } + } + + /* Presence of a trailing slash requires directory semantics. If + the source does not exist, or if the destination cannot be turned + into a directory, give up now. Otherwise, strip trailing slashes + before calling rename. There are no symlinks on mingw, so stat + works instead of lstat. */ + src_slash = ISSLASH (src[src_len - 1]); + dst_slash = ISSLASH (dst[dst_len - 1]); + if (stat (src, &src_st)) + return -1; + if (stat (dst, &dst_st)) + { + if (errno != ENOENT || (!S_ISDIR (src_st.st_mode) && dst_slash)) + return -1; + dst_exists = false; + } + else + { + if (S_ISDIR (dst_st.st_mode) != S_ISDIR (src_st.st_mode)) + { + errno = S_ISDIR (dst_st.st_mode) ? EISDIR : ENOTDIR; + return -1; + } + dst_exists = true; + } + + /* There are no symlinks, so if a file existed with a trailing + slash, it must be a directory, and we don't have to worry about + stripping strip trailing slash. However, mingw refuses to + replace an existing empty directory, so we have to help it out. + And canonicalize_file_name is not yet ported to mingw; however, + for directories, getcwd works as a viable alternative. Ensure + that we can get back to where we started before using it; later + attempts to return are fatal. Note that we can end up losing a + directory if rename then fails, but it was empty, so not much + damage was done. */ + if (dst_exists && S_ISDIR (dst_st.st_mode)) + { + char *cwd = getcwd (NULL, 0); + char *src_temp; + char *dst_temp; + if (!cwd || chdir (cwd)) + return -1; + if (IS_ABSOLUTE_FILE_NAME (src)) + { + dst_temp = chdir (dst) ? NULL : getcwd (NULL, 0); + src_temp = chdir (src) ? NULL : getcwd (NULL, 0); + } + else + { + src_temp = chdir (src) ? NULL : getcwd (NULL, 0); + if (!IS_ABSOLUTE_FILE_NAME (dst) && chdir (cwd)) + abort (); + dst_temp = chdir (dst) ? NULL : getcwd (NULL, 0); + } + if (chdir (cwd)) + abort (); + free (cwd); + if (!src_temp || !dst_temp) + { + free (src_temp); + free (dst_temp); + errno = ENOMEM; + return -1; + } + src_len = strlen (src_temp); + if (strncmp (src_temp, dst_temp, src_len) == 0 + && (ISSLASH (dst_temp[src_len]) || dst_temp[src_len] == '\0')) + { + error = dst_temp[src_len]; + free (src_temp); + free (dst_temp); + if (error) + { + errno = EINVAL; + return -1; + } + return 0; + } + if (rmdir (dst)) + { + error = errno; + free (src_temp); + free (dst_temp); + errno = error; + return -1; + } + free (src_temp); + free (dst_temp); + } + + /* MoveFileEx works if SRC is a directory without any flags, but + fails with MOVEFILE_REPLACE_EXISTING, so try without flags first. + Thankfully, MoveFileEx handles hard links correctly, even though + rename() does not. */ + if (MoveFileEx (src, dst, 0)) + return 0; + + /* Retry with MOVEFILE_REPLACE_EXISTING if the move failed + due to the destination already existing. */ + error = GetLastError (); + if (error == ERROR_FILE_EXISTS || error == ERROR_ALREADY_EXISTS) + { + if (MoveFileEx (src, dst, MOVEFILE_REPLACE_EXISTING)) + return 0; + + error = GetLastError (); + } + + switch (error) + { + case ERROR_FILE_NOT_FOUND: + case ERROR_PATH_NOT_FOUND: + case ERROR_BAD_PATHNAME: + case ERROR_DIRECTORY: + errno = ENOENT; + break; + + case ERROR_ACCESS_DENIED: + case ERROR_SHARING_VIOLATION: + errno = EACCES; + break; + + case ERROR_OUTOFMEMORY: + errno = ENOMEM; + break; + + case ERROR_CURRENT_DIRECTORY: + errno = EBUSY; + break; + + case ERROR_NOT_SAME_DEVICE: + errno = EXDEV; + break; + + case ERROR_WRITE_PROTECT: + errno = EROFS; + break; + + case ERROR_WRITE_FAULT: + case ERROR_READ_FAULT: + case ERROR_GEN_FAILURE: + errno = EIO; + break; + + case ERROR_HANDLE_DISK_FULL: + case ERROR_DISK_FULL: + case ERROR_DISK_TOO_FRAGMENTED: + errno = ENOSPC; + break; + + case ERROR_FILE_EXISTS: + case ERROR_ALREADY_EXISTS: + errno = EEXIST; + break; + + case ERROR_BUFFER_OVERFLOW: + case ERROR_FILENAME_EXCED_RANGE: + errno = ENAMETOOLONG; + break; + + case ERROR_INVALID_NAME: + case ERROR_DELETE_PENDING: + errno = EPERM; /* ? */ + break; + +# ifndef ERROR_FILE_TOO_LARGE +/* This value is documented but not defined in all versions of windows.h. */ +# define ERROR_FILE_TOO_LARGE 223 +# endif + case ERROR_FILE_TOO_LARGE: + errno = EFBIG; + break; + + default: + errno = EINVAL; + break; + } + + return -1; +} + +#else /* ! W32 platform */ + +# include <errno.h> +# include <stdio.h> +# include <stdlib.h> +# include <string.h> +# include <sys/stat.h> +# include <unistd.h> + +# include "dirname.h" +# include "same-inode.h" + +/* Rename the file SRC to DST, fixing any trailing slash bugs. */ + +int +rpl_rename (char const *src, char const *dst) +{ + size_t src_len = strlen (src); + size_t dst_len = strlen (dst); + char *src_temp = (char *) src; + char *dst_temp = (char *) dst; + bool src_slash; + bool dst_slash; + bool dst_exists _GL_UNUSED; + int ret_val = -1; + int rename_errno = ENOTDIR; + struct stat src_st; + struct stat dst_st; + + if (!src_len || !dst_len) + return rename (src, dst); /* Let strace see the ENOENT failure. */ + +# if RENAME_DEST_EXISTS_BUG + { + char *src_base = last_component (src); + char *dst_base = last_component (dst); + if (*src_base == '.') + { + size_t len = base_len (src_base); + if (len == 1 || (len == 2 && src_base[1] == '.')) + { + errno = EINVAL; + return -1; + } + } + if (*dst_base == '.') + { + size_t len = base_len (dst_base); + if (len == 1 || (len == 2 && dst_base[1] == '.')) + { + errno = EINVAL; + return -1; + } + } + } +# endif /* RENAME_DEST_EXISTS_BUG */ + + src_slash = src[src_len - 1] == '/'; + dst_slash = dst[dst_len - 1] == '/'; + +# if !RENAME_HARD_LINK_BUG && !RENAME_DEST_EXISTS_BUG + /* If there are no trailing slashes, then trust the native + implementation unless we also suspect issues with hard link + detection or file/directory conflicts. */ + if (!src_slash && !dst_slash) + return rename (src, dst); +# endif /* !RENAME_HARD_LINK_BUG && !RENAME_DEST_EXISTS_BUG */ + + /* Presence of a trailing slash requires directory semantics. If + the source does not exist, or if the destination cannot be turned + into a directory, give up now. Otherwise, strip trailing slashes + before calling rename. */ + if (lstat (src, &src_st)) + return -1; + if (lstat (dst, &dst_st)) + { + if (errno != ENOENT || (!S_ISDIR (src_st.st_mode) && dst_slash)) + return -1; + dst_exists = false; + } + else + { + if (S_ISDIR (dst_st.st_mode) != S_ISDIR (src_st.st_mode)) + { + errno = S_ISDIR (dst_st.st_mode) ? EISDIR : ENOTDIR; + return -1; + } +# if RENAME_HARD_LINK_BUG + if (SAME_INODE (src_st, dst_st)) + return 0; +# endif /* RENAME_HARD_LINK_BUG */ + dst_exists = true; + } + +# if (RENAME_TRAILING_SLASH_SOURCE_BUG || RENAME_DEST_EXISTS_BUG \ + || RENAME_HARD_LINK_BUG) + /* If the only bug was that a trailing slash was allowed on a + non-existing file destination, as in Solaris 10, then we've + already covered that situation. But if there is any problem with + a trailing slash on an existing source or destination, as in + Solaris 9, or if a directory can overwrite a symlink, as on + Cygwin 1.5, or if directories cannot be created with trailing + slash, as on NetBSD 1.6, then we must strip the offending slash + and check that we have not encountered a symlink instead of a + directory. + + Stripping a trailing slash interferes with POSIX semantics, where + rename behavior on a symlink with a trailing slash operates on + the corresponding target directory. We prefer the GNU semantics + of rejecting any use of a symlink with trailing slash, but do not + enforce them, since Solaris 10 is able to obey POSIX semantics + and there might be clients expecting it, as counter-intuitive as + those semantics are. + + Technically, we could also follow the POSIX behavior by chasing a + readlink trail, but that is harder to implement. */ + if (src_slash) + { + src_temp = strdup (src); + if (!src_temp) + { + /* Rather than rely on strdup-posix, we set errno ourselves. */ + rename_errno = ENOMEM; + goto out; + } + strip_trailing_slashes (src_temp); + if (lstat (src_temp, &src_st)) + { + rename_errno = errno; + goto out; + } + if (S_ISLNK (src_st.st_mode)) + goto out; + } + if (dst_slash) + { + dst_temp = strdup (dst); + if (!dst_temp) + { + rename_errno = ENOMEM; + goto out; + } + strip_trailing_slashes (dst_temp); + if (lstat (dst_temp, &dst_st)) + { + if (errno != ENOENT) + { + rename_errno = errno; + goto out; + } + } + else if (S_ISLNK (dst_st.st_mode)) + goto out; + } +# endif /* RENAME_TRAILING_SLASH_SOURCE_BUG || RENAME_DEST_EXISTS_BUG + || RENAME_HARD_LINK_BUG */ + +# if RENAME_DEST_EXISTS_BUG + /* Cygwin 1.5 sometimes behaves oddly when moving a non-empty + directory on top of an empty one (the old directory name can + reappear if the new directory tree is removed). Work around this + by removing the target first, but don't remove the target if it + is a subdirectory of the source. Note that we can end up losing + a directory if rename then fails, but it was empty, so not much + damage was done. */ + if (dst_exists && S_ISDIR (dst_st.st_mode)) + { + if (src_st.st_dev != dst_st.st_dev) + { + rename_errno = EXDEV; + goto out; + } + if (src_temp != src) + free (src_temp); + src_temp = canonicalize_file_name (src); + if (dst_temp != dst) + free (dst_temp); + dst_temp = canonicalize_file_name (dst); + if (!src_temp || !dst_temp) + { + rename_errno = ENOMEM; + goto out; + } + src_len = strlen (src_temp); + if (strncmp (src_temp, dst_temp, src_len) == 0 + && dst_temp[src_len] == '/') + { + rename_errno = EINVAL; + goto out; + } + if (rmdir (dst)) + { + rename_errno = errno; + goto out; + } + } +# endif /* RENAME_DEST_EXISTS_BUG */ + + ret_val = rename (src_temp, dst_temp); + rename_errno = errno; + + out: _GL_UNUSED_LABEL; + + if (src_temp != src) + free (src_temp); + if (dst_temp != dst) + free (dst_temp); + errno = rename_errno; + return ret_val; +} +#endif /* ! W32 platform */ diff --git a/lib/rmdir.c b/lib/rmdir.c new file mode 100644 index 0000000..7dc7aec --- /dev/null +++ b/lib/rmdir.c @@ -0,0 +1,53 @@ +/* Work around rmdir bugs. + + Copyright (C) 1988, 1990, 1999, 2003-2006, 2009-2018 Free Software + Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +#include <config.h> + +#include <unistd.h> + +#include <errno.h> +#include <string.h> + +#include "dosname.h" + +#undef rmdir + +/* Remove directory DIR. + Return 0 if successful, -1 if not. */ + +int +rpl_rmdir (char const *dir) +{ + /* Work around cygwin 1.5.x bug where rmdir("dir/./") succeeds. */ + size_t len = strlen (dir); + int result; + while (len && ISSLASH (dir[len - 1])) + len--; + if (len && dir[len - 1] == '.' && (1 == len || ISSLASH (dir[len - 2]))) + { + errno = EINVAL; + return -1; + } + result = rmdir (dir); + /* Work around mingw bug, where rmdir("file/") fails with EINVAL + instead of ENOTDIR. We've already filtered out trailing ., the + only reason allowed by POSIX for EINVAL. */ + if (result == -1 && errno == EINVAL) + errno = ENOTDIR; + return result; +} diff --git a/lib/same-inode.h b/lib/same-inode.h new file mode 100644 index 0000000..6837795 --- /dev/null +++ b/lib/same-inode.h @@ -0,0 +1,47 @@ +/* Determine whether two stat buffers are known to refer to the same file. + + Copyright (C) 2006, 2009-2018 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +#ifndef SAME_INODE_H +# define SAME_INODE_H 1 + +# include <sys/types.h> + +# if defined __VMS && __CRTL_VER < 80200000 +# define SAME_INODE(a, b) \ + ((a).st_ino[0] == (b).st_ino[0] \ + && (a).st_ino[1] == (b).st_ino[1] \ + && (a).st_ino[2] == (b).st_ino[2] \ + && (a).st_dev == (b).st_dev) +# elif defined _WIN32 && ! defined __CYGWIN__ + /* Native Windows. */ +# if _GL_WINDOWS_STAT_INODES + /* stat() and fstat() set st_dev and st_ino to 0 if information about + the inode is not available. */ +# define SAME_INODE(a, b) \ + (!((a).st_ino == 0 && (a).st_dev == 0) \ + && (a).st_ino == (b).st_ino && (a).st_dev == (b).st_dev) +# else + /* stat() and fstat() set st_ino to 0 always. */ +# define SAME_INODE(a, b) 0 +# endif +# else +# define SAME_INODE(a, b) \ + ((a).st_ino == (b).st_ino \ + && (a).st_dev == (b).st_dev) +# endif + +#endif diff --git a/lib/se-context.c b/lib/se-context.c new file mode 100644 index 0000000..2e5aa13 --- /dev/null +++ b/lib/se-context.c @@ -0,0 +1,3 @@ +#include <config.h> +#define SE_CONTEXT_INLINE _GL_EXTERN_INLINE +#include <selinux/context.h> diff --git a/lib/se-context.in.h b/lib/se-context.in.h new file mode 100644 index 0000000..c3890a2 --- /dev/null +++ b/lib/se-context.in.h @@ -0,0 +1,46 @@ +#ifndef SELINUX_CONTEXT_H +# define SELINUX_CONTEXT_H + +# include <errno.h> + +#ifndef _GL_INLINE_HEADER_BEGIN + #error "Please include config.h first." +#endif +_GL_INLINE_HEADER_BEGIN +#ifndef SE_CONTEXT_INLINE +# define SE_CONTEXT_INLINE _GL_INLINE +#endif + +/* The definition of _GL_UNUSED_PARAMETER is copied here. */ + +typedef int context_t; +SE_CONTEXT_INLINE context_t context_new (char const *s _GL_UNUSED_PARAMETER) + { errno = ENOTSUP; return 0; } +SE_CONTEXT_INLINE char *context_str (context_t con _GL_UNUSED_PARAMETER) + { errno = ENOTSUP; return (void *) 0; } +SE_CONTEXT_INLINE void context_free (context_t c _GL_UNUSED_PARAMETER) {} + +SE_CONTEXT_INLINE int context_user_set (context_t sc _GL_UNUSED_PARAMETER, + char const *s _GL_UNUSED_PARAMETER) + { errno = ENOTSUP; return -1; } +SE_CONTEXT_INLINE int context_role_set (context_t sc _GL_UNUSED_PARAMETER, + char const *s _GL_UNUSED_PARAMETER) + { errno = ENOTSUP; return -1; } +SE_CONTEXT_INLINE int context_range_set (context_t sc _GL_UNUSED_PARAMETER, + char const *s _GL_UNUSED_PARAMETER) + { errno = ENOTSUP; return -1; } +SE_CONTEXT_INLINE int context_type_set (context_t sc _GL_UNUSED_PARAMETER, + char const *s _GL_UNUSED_PARAMETER) + { errno = ENOTSUP; return -1; } +SE_CONTEXT_INLINE char *context_type_get (context_t sc _GL_UNUSED_PARAMETER) + { errno = ENOTSUP; return (void *) 0; } +SE_CONTEXT_INLINE char *context_range_get (context_t sc _GL_UNUSED_PARAMETER) + { errno = ENOTSUP; return (void *) 0; } +SE_CONTEXT_INLINE char *context_role_get (context_t sc _GL_UNUSED_PARAMETER) + { errno = ENOTSUP; return (void *) 0; } +SE_CONTEXT_INLINE char *context_user_get (context_t sc _GL_UNUSED_PARAMETER) + { errno = ENOTSUP; return (void *) 0; } + +_GL_INLINE_HEADER_END + +#endif diff --git a/lib/se-selinux.c b/lib/se-selinux.c new file mode 100644 index 0000000..ee4e57f --- /dev/null +++ b/lib/se-selinux.c @@ -0,0 +1,3 @@ +#include <config.h> +#define SE_SELINUX_INLINE _GL_EXTERN_INLINE +#include <selinux/selinux.h> diff --git a/lib/se-selinux.in.h b/lib/se-selinux.in.h new file mode 100644 index 0000000..f7610b3 --- /dev/null +++ b/lib/se-selinux.in.h @@ -0,0 +1,119 @@ +/* Replacement <selinux/selinux.h> for platforms that lack it. + Copyright (C) 2008-2018 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +#if __GNUC__ >= 3 +@PRAGMA_SYSTEM_HEADER@ +#endif +@PRAGMA_COLUMNS@ + +#if HAVE_SELINUX_SELINUX_H + +#@INCLUDE_NEXT@ @NEXT_SELINUX_SELINUX_H@ + +#else +# if !defined _@GUARD_PREFIX@_SELINUX_SELINUX_H +# define _@GUARD_PREFIX@_SELINUX_SELINUX_H + +# include <sys/types.h> +# include <errno.h> + +# ifndef _GL_INLINE_HEADER_BEGIN + #error "Please include config.h first." +# endif +_GL_INLINE_HEADER_BEGIN +# ifndef SE_SELINUX_INLINE +# define SE_SELINUX_INLINE _GL_INLINE +# endif + +/* The definition of _GL_UNUSED_PARAMETER is copied here. */ + +# if !GNULIB_defined_security_types + +typedef unsigned short security_class_t; +typedef char *security_context_t; +# define is_selinux_enabled() 0 + +SE_SELINUX_INLINE int +getcon (security_context_t *con _GL_UNUSED_PARAMETER) + { errno = ENOTSUP; return -1; } +SE_SELINUX_INLINE void +freecon (security_context_t con _GL_UNUSED_PARAMETER) {} + +SE_SELINUX_INLINE int +getfscreatecon (security_context_t *con _GL_UNUSED_PARAMETER) + { errno = ENOTSUP; return -1; } +SE_SELINUX_INLINE int +setfscreatecon (security_context_t con _GL_UNUSED_PARAMETER) + { errno = ENOTSUP; return -1; } +SE_SELINUX_INLINE int +matchpathcon (char const *file _GL_UNUSED_PARAMETER, + mode_t m _GL_UNUSED_PARAMETER, + security_context_t *con _GL_UNUSED_PARAMETER) + { errno = ENOTSUP; return -1; } +SE_SELINUX_INLINE int +getfilecon (char const *file _GL_UNUSED_PARAMETER, + security_context_t *con _GL_UNUSED_PARAMETER) + { errno = ENOTSUP; return -1; } +SE_SELINUX_INLINE int +lgetfilecon (char const *file _GL_UNUSED_PARAMETER, + security_context_t *con _GL_UNUSED_PARAMETER) + { errno = ENOTSUP; return -1; } +SE_SELINUX_INLINE int +fgetfilecon (int fd, security_context_t *con _GL_UNUSED_PARAMETER) + { errno = ENOTSUP; return -1; } +SE_SELINUX_INLINE int +setfilecon (char const *file _GL_UNUSED_PARAMETER, + security_context_t con _GL_UNUSED_PARAMETER) + { errno = ENOTSUP; return -1; } +SE_SELINUX_INLINE int +lsetfilecon (char const *file _GL_UNUSED_PARAMETER, + security_context_t con _GL_UNUSED_PARAMETER) + { errno = ENOTSUP; return -1; } +SE_SELINUX_INLINE int +fsetfilecon (int fd _GL_UNUSED_PARAMETER, + security_context_t con _GL_UNUSED_PARAMETER) + { errno = ENOTSUP; return -1; } + +SE_SELINUX_INLINE int +security_check_context (security_context_t con _GL_UNUSED_PARAMETER) + { errno = ENOTSUP; return -1; } +SE_SELINUX_INLINE int +security_check_context_raw (security_context_t con _GL_UNUSED_PARAMETER) + { errno = ENOTSUP; return -1; } +SE_SELINUX_INLINE int +setexeccon (security_context_t con _GL_UNUSED_PARAMETER) + { errno = ENOTSUP; return -1; } +SE_SELINUX_INLINE int +security_compute_create (security_context_t scon _GL_UNUSED_PARAMETER, + security_context_t tcon _GL_UNUSED_PARAMETER, + security_class_t tclass _GL_UNUSED_PARAMETER, + security_context_t *newcon _GL_UNUSED_PARAMETER) + { errno = ENOTSUP; return -1; } +SE_SELINUX_INLINE security_class_t +string_to_security_class (char const *name) + { errno = ENOTSUP; return 0; } +SE_SELINUX_INLINE int +matchpathcon_init_prefix (char const *path _GL_UNUSED_PARAMETER, + char const *prefix _GL_UNUSED_PARAMETER) + { errno = ENOTSUP; return -1; } + +# define GNULIB_defined_security_types 1 +# endif + +_GL_INLINE_HEADER_END + +# endif +#endif diff --git a/lib/set-acl.c b/lib/set-acl.c new file mode 100644 index 0000000..830beb1 --- /dev/null +++ b/lib/set-acl.c @@ -0,0 +1,48 @@ +/* set-acl.c - set access control list equivalent to a mode + + Copyright (C) 2002-2003, 2005-2018 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. + + Written by Paul Eggert and Andreas Gruenbacher, and Bruno Haible. */ + +#include <config.h> + +#include "acl.h" + +#include <errno.h> + +#include "quote.h" +#include "error.h" +#include "gettext.h" +#define _(msgid) gettext (msgid) + + +/* Set the access control lists of a file. If DESC is a valid file + descriptor, use file descriptor operations where available, else use + filename based operations on NAME. If access control lists are not + available, fchmod the target file to MODE. Also sets the + non-permission bits of the destination file (S_ISUID, S_ISGID, S_ISVTX) + to those from MODE if any are set. + Return 0 if successful. On failure, output a diagnostic, set errno and + return -1. */ + +int +set_acl (char const *name, int desc, mode_t mode) +{ + int ret = qset_acl (name, desc, mode); + if (ret != 0) + error (0, errno, _("setting permissions for %s"), quote (name)); + return ret; +} diff --git a/lib/set-permissions.c b/lib/set-permissions.c new file mode 100644 index 0000000..a415e13 --- /dev/null +++ b/lib/set-permissions.c @@ -0,0 +1,847 @@ +/* Set permissions of a file. -*- coding: utf-8 -*- + + Copyright (C) 2002-2003, 2005-2018 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. + + Written by Paul Eggert, Andreas Grünbacher, and Bruno Haible. */ + +#include <config.h> + +#include "acl.h" + +#include "acl-internal.h" + +#if USE_ACL +# if ! defined HAVE_ACL_FROM_MODE && defined HAVE_ACL_FROM_TEXT /* FreeBSD, IRIX, Tru64, Cygwin >= 2.5 */ +# if HAVE_ACL_GET_FILE && !HAVE_ACL_TYPE_EXTENDED + +static acl_t +acl_from_mode (mode_t mode) +{ +# if HAVE_ACL_FREE_TEXT /* Tru64 */ + char acl_text[] = "u::---,g::---,o::---,"; +# else /* FreeBSD, IRIX, Cygwin >= 2.5 */ + char acl_text[] = "u::---,g::---,o::---"; +# endif + + if (mode & S_IRUSR) acl_text[ 3] = 'r'; + if (mode & S_IWUSR) acl_text[ 4] = 'w'; + if (mode & S_IXUSR) acl_text[ 5] = 'x'; + if (mode & S_IRGRP) acl_text[10] = 'r'; + if (mode & S_IWGRP) acl_text[11] = 'w'; + if (mode & S_IXGRP) acl_text[12] = 'x'; + if (mode & S_IROTH) acl_text[17] = 'r'; + if (mode & S_IWOTH) acl_text[18] = 'w'; + if (mode & S_IXOTH) acl_text[19] = 'x'; + + return acl_from_text (acl_text); +} +# endif +# endif + +# if HAVE_FACL && defined GETACL /* Solaris, Cygwin < 2.5, not HP-UX */ +static int +set_acls_from_mode (const char *name, int desc, mode_t mode, bool *must_chmod) +{ +# ifdef ACE_GETACL + /* Solaris also has a different variant of ACLs, used in ZFS and NFSv4 + file systems (whereas the other ones are used in UFS file systems). */ + + /* The flags in the ace_t structure changed in a binary incompatible way + when ACL_NO_TRIVIAL etc. were introduced in <sys/acl.h> version 1.15. + How to distinguish the two conventions at runtime? + We fetch the existing ACL. In the old convention, usually three ACEs have + a_flags = ACE_OWNER / ACE_GROUP / ACE_OTHER, in the range 0x0100..0x0400. + In the new convention, these values are not used. */ + int convention; + + { + /* Initially, try to read the entries into a stack-allocated buffer. + Use malloc if it does not fit. */ + enum + { + alloc_init = 4000 / sizeof (ace_t), /* >= 3 */ + alloc_max = MIN (INT_MAX, SIZE_MAX / sizeof (ace_t)) + }; + ace_t buf[alloc_init]; + size_t alloc = alloc_init; + ace_t *entries = buf; + ace_t *malloced = NULL; + int count; + + for (;;) + { + count = (desc != -1 + ? facl (desc, ACE_GETACL, alloc, entries) + : acl (name, ACE_GETACL, alloc, entries)); + if (count < 0 && errno == ENOSPC) + { + /* Increase the size of the buffer. */ + free (malloced); + if (alloc > alloc_max / 2) + { + errno = ENOMEM; + return -1; + } + alloc = 2 * alloc; /* <= alloc_max */ + entries = malloced = (ace_t *) malloc (alloc * sizeof (ace_t)); + if (entries == NULL) + { + errno = ENOMEM; + return -1; + } + continue; + } + break; + } + + if (count <= 0) + convention = -1; + else + { + int i; + + convention = 0; + for (i = 0; i < count; i++) + if (entries[i].a_flags & (OLD_ACE_OWNER | OLD_ACE_GROUP | OLD_ACE_OTHER)) + { + convention = 1; + break; + } + } + free (malloced); + } + + if (convention >= 0) + { + ace_t entries[6]; + int count; + int ret; + + if (convention) + { + /* Running on Solaris 10. */ + entries[0].a_type = OLD_ALLOW; + entries[0].a_flags = OLD_ACE_OWNER; + entries[0].a_who = 0; /* irrelevant */ + entries[0].a_access_mask = (mode >> 6) & 7; + entries[1].a_type = OLD_ALLOW; + entries[1].a_flags = OLD_ACE_GROUP; + entries[1].a_who = 0; /* irrelevant */ + entries[1].a_access_mask = (mode >> 3) & 7; + entries[2].a_type = OLD_ALLOW; + entries[2].a_flags = OLD_ACE_OTHER; + entries[2].a_who = 0; + entries[2].a_access_mask = mode & 7; + count = 3; + } + else + { + /* Running on Solaris 10 (newer version) or Solaris 11. + The details here were found through "/bin/ls -lvd somefiles". */ + entries[0].a_type = NEW_ACE_ACCESS_DENIED_ACE_TYPE; + entries[0].a_flags = NEW_ACE_OWNER; + entries[0].a_who = 0; /* irrelevant */ + entries[0].a_access_mask = 0; + entries[1].a_type = NEW_ACE_ACCESS_ALLOWED_ACE_TYPE; + entries[1].a_flags = NEW_ACE_OWNER; + entries[1].a_who = 0; /* irrelevant */ + entries[1].a_access_mask = NEW_ACE_WRITE_NAMED_ATTRS + | NEW_ACE_WRITE_ATTRIBUTES + | NEW_ACE_WRITE_ACL + | NEW_ACE_WRITE_OWNER; + if (mode & 0400) + entries[1].a_access_mask |= NEW_ACE_READ_DATA; + else + entries[0].a_access_mask |= NEW_ACE_READ_DATA; + if (mode & 0200) + entries[1].a_access_mask |= NEW_ACE_WRITE_DATA | NEW_ACE_APPEND_DATA; + else + entries[0].a_access_mask |= NEW_ACE_WRITE_DATA | NEW_ACE_APPEND_DATA; + if (mode & 0100) + entries[1].a_access_mask |= NEW_ACE_EXECUTE; + else + entries[0].a_access_mask |= NEW_ACE_EXECUTE; + entries[2].a_type = NEW_ACE_ACCESS_DENIED_ACE_TYPE; + entries[2].a_flags = NEW_ACE_GROUP | NEW_ACE_IDENTIFIER_GROUP; + entries[2].a_who = 0; /* irrelevant */ + entries[2].a_access_mask = 0; + entries[3].a_type = NEW_ACE_ACCESS_ALLOWED_ACE_TYPE; + entries[3].a_flags = NEW_ACE_GROUP | NEW_ACE_IDENTIFIER_GROUP; + entries[3].a_who = 0; /* irrelevant */ + entries[3].a_access_mask = 0; + if (mode & 0040) + entries[3].a_access_mask |= NEW_ACE_READ_DATA; + else + entries[2].a_access_mask |= NEW_ACE_READ_DATA; + if (mode & 0020) + entries[3].a_access_mask |= NEW_ACE_WRITE_DATA | NEW_ACE_APPEND_DATA; + else + entries[2].a_access_mask |= NEW_ACE_WRITE_DATA | NEW_ACE_APPEND_DATA; + if (mode & 0010) + entries[3].a_access_mask |= NEW_ACE_EXECUTE; + else + entries[2].a_access_mask |= NEW_ACE_EXECUTE; + entries[4].a_type = NEW_ACE_ACCESS_DENIED_ACE_TYPE; + entries[4].a_flags = NEW_ACE_EVERYONE; + entries[4].a_who = 0; + entries[4].a_access_mask = NEW_ACE_WRITE_NAMED_ATTRS + | NEW_ACE_WRITE_ATTRIBUTES + | NEW_ACE_WRITE_ACL + | NEW_ACE_WRITE_OWNER; + entries[5].a_type = NEW_ACE_ACCESS_ALLOWED_ACE_TYPE; + entries[5].a_flags = NEW_ACE_EVERYONE; + entries[5].a_who = 0; + entries[5].a_access_mask = NEW_ACE_READ_NAMED_ATTRS + | NEW_ACE_READ_ATTRIBUTES + | NEW_ACE_READ_ACL + | NEW_ACE_SYNCHRONIZE; + if (mode & 0004) + entries[5].a_access_mask |= NEW_ACE_READ_DATA; + else + entries[4].a_access_mask |= NEW_ACE_READ_DATA; + if (mode & 0002) + entries[5].a_access_mask |= NEW_ACE_WRITE_DATA | NEW_ACE_APPEND_DATA; + else + entries[4].a_access_mask |= NEW_ACE_WRITE_DATA | NEW_ACE_APPEND_DATA; + if (mode & 0001) + entries[5].a_access_mask |= NEW_ACE_EXECUTE; + else + entries[4].a_access_mask |= NEW_ACE_EXECUTE; + count = 6; + } + if (desc != -1) + ret = facl (desc, ACE_SETACL, count, entries); + else + ret = acl (name, ACE_SETACL, count, entries); + if (ret < 0 && errno != EINVAL && errno != ENOTSUP) + { + if (errno == ENOSYS) + { + *must_chmod = true; + return 0; + } + return -1; + } + if (ret == 0) + return 0; + } +# endif + + { + aclent_t entries[3]; + int ret; + + entries[0].a_type = USER_OBJ; + entries[0].a_id = 0; /* irrelevant */ + entries[0].a_perm = (mode >> 6) & 7; + entries[1].a_type = GROUP_OBJ; + entries[1].a_id = 0; /* irrelevant */ + entries[1].a_perm = (mode >> 3) & 7; + entries[2].a_type = OTHER_OBJ; + entries[2].a_id = 0; + entries[2].a_perm = mode & 7; + + if (desc != -1) + ret = facl (desc, SETACL, + sizeof (entries) / sizeof (aclent_t), entries); + else + ret = acl (name, SETACL, + sizeof (entries) / sizeof (aclent_t), entries); + if (ret < 0) + { + if (errno == ENOSYS || errno == EOPNOTSUPP) + { + *must_chmod = true; + return 0; + } + return -1; + } + return 0; + } +} + +# elif HAVE_GETACL /* HP-UX */ +static int +context_acl_from_mode (struct permission_context *ctx, const char *name, int desc) +{ + struct stat statbuf; + int ret; + + if (desc != -1) + ret = fstat (desc, &statbuf); + else + ret = stat (name, &statbuf); + if (ret < 0) + return -1; + + ctx->entries[0].uid = statbuf.st_uid; + ctx->entries[0].gid = ACL_NSGROUP; + ctx->entries[0].mode = (ctx->mode >> 6) & 7; + ctx->entries[1].uid = ACL_NSUSER; + ctx->entries[1].gid = statbuf.st_gid; + ctx->entries[1].mode = (ctx->mode >> 3) & 7; + ctx->entries[2].uid = ACL_NSUSER; + ctx->entries[2].gid = ACL_NSGROUP; + ctx->entries[2].mode = ctx->mode & 7; + ctx->count = 3; + return 0; +} + +# if HAVE_ACLV_H /* HP-UX >= 11.11 */ +static int +context_aclv_from_mode (struct permission_context *ctx) +{ + int ret; + + ctx->aclv_entries[0].a_type = USER_OBJ; + ctx->aclv_entries[0].a_id = 0; /* irrelevant */ + ctx->aclv_entries[0].a_perm = (ctx->mode >> 6) & 7; + ctx->aclv_entries[1].a_type = GROUP_OBJ; + ctx->aclv_entries[1].a_id = 0; /* irrelevant */ + ctx->aclv_entries[1].a_perm = (ctx->mode >> 3) & 7; + ctx->aclv_entries[2].a_type = CLASS_OBJ; + ctx->aclv_entries[2].a_id = 0; + ctx->aclv_entries[2].a_perm = (ctx->mode >> 3) & 7; + ctx->aclv_entries[3].a_type = OTHER_OBJ; + ctx->aclv_entries[3].a_id = 0; + ctx->aclv_entries[3].a_perm = ctx->mode & 7; + ctx->aclv_count = 4; + + ret = aclsort (ctx->aclv_count, 1, ctx->aclv_entries); + if (ret > 0) + abort (); + return ret; +} +# endif + +# elif HAVE_ACLX_GET && defined ACL_AIX_WIP /* AIX */ +static int +set_acls_from_mode (const char *name, int desc, mode_t mode, bool *must_chmod) +{ + acl_type_list_t types; + size_t types_size = sizeof (types); + acl_type_t type; + + if (aclx_gettypes (name, &types, &types_size) < 0 + || types.num_entries == 0) + { + *must_chmod = true; + return 0; + } + + /* XXX Do we need to clear all types of ACLs for the given file, or is it + sufficient to clear the first one? */ + type = types.entries[0]; + if (type.u64 == ACL_AIXC) + { + union { struct acl a; char room[128]; } u; + int ret; + + u.a.acl_len = (char *) &u.a.acl_ext[0] - (char *) &u.a; /* no entries */ + u.a.acl_mode = mode & ~(S_IXACL | 0777); + u.a.u_access = (mode >> 6) & 7; + u.a.g_access = (mode >> 3) & 7; + u.a.o_access = mode & 7; + + if (desc != -1) + ret = aclx_fput (desc, SET_ACL | SET_MODE_S_BITS, + type, &u.a, u.a.acl_len, mode); + else + ret = aclx_put (name, SET_ACL | SET_MODE_S_BITS, + type, &u.a, u.a.acl_len, mode); + if (!(ret < 0 && errno == ENOSYS)) + return ret; + } + else if (type.u64 == ACL_NFS4) + { + union { nfs4_acl_int_t a; char room[128]; } u; + nfs4_ace_int_t *ace; + int ret; + + u.a.aclVersion = NFS4_ACL_INT_STRUCT_VERSION; + u.a.aclEntryN = 0; + ace = &u.a.aclEntry[0]; + { + ace->flags = ACE4_ID_SPECIAL; + ace->aceWho.special_whoid = ACE4_WHO_OWNER; + ace->aceType = ACE4_ACCESS_ALLOWED_ACE_TYPE; + ace->aceFlags = 0; + ace->aceMask = + (mode & 0400 ? ACE4_READ_DATA | ACE4_LIST_DIRECTORY : 0) + | (mode & 0200 + ? ACE4_WRITE_DATA | ACE4_ADD_FILE | ACE4_APPEND_DATA + | ACE4_ADD_SUBDIRECTORY + : 0) + | (mode & 0100 ? ACE4_EXECUTE : 0); + ace->aceWhoString[0] = '\0'; + ace->entryLen = (char *) &ace->aceWhoString[4] - (char *) ace; + ace = (nfs4_ace_int_t *) (char *) &ace->aceWhoString[4]; + u.a.aclEntryN++; + } + { + ace->flags = ACE4_ID_SPECIAL; + ace->aceWho.special_whoid = ACE4_WHO_GROUP; + ace->aceType = ACE4_ACCESS_ALLOWED_ACE_TYPE; + ace->aceFlags = 0; + ace->aceMask = + (mode & 0040 ? ACE4_READ_DATA | ACE4_LIST_DIRECTORY : 0) + | (mode & 0020 + ? ACE4_WRITE_DATA | ACE4_ADD_FILE | ACE4_APPEND_DATA + | ACE4_ADD_SUBDIRECTORY + : 0) + | (mode & 0010 ? ACE4_EXECUTE : 0); + ace->aceWhoString[0] = '\0'; + ace->entryLen = (char *) &ace->aceWhoString[4] - (char *) ace; + ace = (nfs4_ace_int_t *) (char *) &ace->aceWhoString[4]; + u.a.aclEntryN++; + } + { + ace->flags = ACE4_ID_SPECIAL; + ace->aceWho.special_whoid = ACE4_WHO_EVERYONE; + ace->aceType = ACE4_ACCESS_ALLOWED_ACE_TYPE; + ace->aceFlags = 0; + ace->aceMask = + (mode & 0004 ? ACE4_READ_DATA | ACE4_LIST_DIRECTORY : 0) + | (mode & 0002 + ? ACE4_WRITE_DATA | ACE4_ADD_FILE | ACE4_APPEND_DATA + | ACE4_ADD_SUBDIRECTORY + : 0) + | (mode & 0001 ? ACE4_EXECUTE : 0); + ace->aceWhoString[0] = '\0'; + ace->entryLen = (char *) &ace->aceWhoString[4] - (char *) ace; + ace = (nfs4_ace_int_t *) (char *) &ace->aceWhoString[4]; + u.a.aclEntryN++; + } + u.a.aclLength = (char *) ace - (char *) &u.a; + + if (desc != -1) + ret = aclx_fput (desc, SET_ACL | SET_MODE_S_BITS, + type, &u.a, u.a.aclLength, mode); + else + ret = aclx_put (name, SET_ACL | SET_MODE_S_BITS, + type, &u.a, u.a.aclLength, mode); + if (!(ret < 0 && errno == ENOSYS)) + return ret; + } + + *must_chmod = true; + return 0; +} + +# elif HAVE_STATACL /* older AIX */ +static int +context_acl_from_mode (struct permission_context *ctx) +{ + ctx->u.a.acl_len = (char *) &ctx->u.a.acl_ext[0] - (char *) &ctx->u.a; /* no entries */ + ctx->u.a.acl_mode = ctx->mode & ~(S_IXACL | 0777); + ctx->u.a.u_access = (ctx->mode >> 6) & 7; + ctx->u.a.g_access = (ctx->mode >> 3) & 7; + ctx->u.a.o_access = ctx->mode & 7; + ctx->have_u = true; + return 0; +} + +# elif HAVE_ACLSORT /* NonStop Kernel */ +static int +context_acl_from_mode (struct permission_context *ctx) +{ + int ret; + + ctx->entries[0].a_type = USER_OBJ; + ctx->entries[0].a_id = 0; /* irrelevant */ + ctx->entries[0].a_perm = (ctx->mode >> 6) & 7; + ctx->entries[1].a_type = GROUP_OBJ; + ctx->entries[1].a_id = 0; /* irrelevant */ + ctx->entries[1].a_perm = (ctx->mode >> 3) & 7; + ctx->entries[2].a_type = CLASS_OBJ; + ctx->entries[2].a_id = 0; + ctx->entries[2].a_perm = (ctx->mode >> 3) & 7; + ctx->entries[3].a_type = OTHER_OBJ; + ctx->entries[3].a_id = 0; + ctx->entries[3].a_perm = ctx->mode & 7; + ctx->count = 4; + + ret = aclsort (ctx->count, 1, entries); + if (ret > 0) + abort (); + return ret; +} +# endif + +static int +set_acls (struct permission_context *ctx, const char *name, int desc, + int from_mode, bool *must_chmod, bool *acls_set) +{ + int ret = 0; + +# if HAVE_ACL_GET_FILE + /* POSIX 1003.1e (draft 17 -- abandoned) specific version. */ + /* Linux, FreeBSD, Mac OS X, IRIX, Tru64, Cygwin >= 2.5 */ +# if !HAVE_ACL_TYPE_EXTENDED + /* Linux, FreeBSD, IRIX, Tru64, Cygwin >= 2.5 */ + +# ifndef HAVE_ACL_FROM_TEXT +# error Must have acl_from_text (see POSIX 1003.1e draft 17). +# endif +# ifndef HAVE_ACL_DELETE_DEF_FILE +# error Must have acl_delete_def_file (see POSIX 1003.1e draft 17). +# endif + + if (! ctx->acls_not_supported) + { + if (ret == 0 && from_mode) + { + if (ctx->acl) + acl_free (ctx->acl); + ctx->acl = acl_from_mode (ctx->mode); + if (ctx->acl == NULL) + ret = -1; + } + + if (ret == 0 && ctx->acl) + { + if (HAVE_ACL_SET_FD && desc != -1) + ret = acl_set_fd (desc, ctx->acl); + else + ret = acl_set_file (name, ACL_TYPE_ACCESS, ctx->acl); + if (ret != 0) + { + if (! acl_errno_valid (errno)) + { + ctx->acls_not_supported = true; + if (from_mode || acl_access_nontrivial (ctx->acl) == 0) + ret = 0; + } + } + else + { + *acls_set = true; + if (S_ISDIR(ctx->mode)) + { + if (! from_mode && ctx->default_acl && + acl_default_nontrivial (ctx->default_acl)) + ret = acl_set_file (name, ACL_TYPE_DEFAULT, + ctx->default_acl); + else + ret = acl_delete_def_file (name); + } + } + } + } + +# if HAVE_ACL_TYPE_NFS4 /* FreeBSD */ + + /* File systems either support POSIX ACLs (for example, ufs) or NFS4 ACLs + (for example, zfs). */ + + /* TODO: Implement setting ACLs once get_permissions() reads them. */ + +# endif + +# else /* HAVE_ACL_TYPE_EXTENDED */ + /* Mac OS X */ + + /* On Mac OS X, acl_get_file (name, ACL_TYPE_ACCESS) + and acl_get_file (name, ACL_TYPE_DEFAULT) + always return NULL / EINVAL. You have to use + acl_get_file (name, ACL_TYPE_EXTENDED) + or acl_get_fd (open (name, ...)) + to retrieve an ACL. + On the other hand, + acl_set_file (name, ACL_TYPE_ACCESS, acl) + and acl_set_file (name, ACL_TYPE_DEFAULT, acl) + have the same effect as + acl_set_file (name, ACL_TYPE_EXTENDED, acl): + Each of these calls sets the file's ACL. */ + + if (ctx->acl == NULL) + { + acl_t acl; + + /* Remove ACLs if the file has ACLs. */ + if (HAVE_ACL_GET_FD && desc != -1) + acl = acl_get_fd (desc); + else + acl = acl_get_file (name, ACL_TYPE_EXTENDED); + if (acl) + { + acl_free (acl); + + acl = acl_init (0); + if (acl) + { + if (HAVE_ACL_SET_FD && desc != -1) + ret = acl_set_fd (desc, acl); + else + ret = acl_set_file (name, ACL_TYPE_EXTENDED, acl); + acl_free (acl); + } + else + ret = -1; + } + } + else + { + if (HAVE_ACL_SET_FD && desc != -1) + ret = acl_set_fd (desc, ctx->acl); + else + ret = acl_set_file (name, ACL_TYPE_EXTENDED, ctx->acl); + if (ret != 0) + { + if (! acl_errno_valid (errno) + && ! acl_extended_nontrivial (ctx->acl)) + ret = 0; + } + } + *acls_set = true; + +# endif + +# elif defined GETACL /* Solaris, Cygwin, not HP-UX */ + + /* Solaris 2.5 through Solaris 10, Cygwin, and contemporaneous versions + of Unixware. The acl() call returns the access and default ACL both + at once. */ + + /* If both ace_entries and entries are available, try SETACL before + ACE_SETACL, because SETACL cannot fail with ENOTSUP whereas ACE_SETACL + can. */ + + if (from_mode) + return set_acls_from_mode (name, desc, ctx->mode, must_chmod); + + if (ret == 0 && ctx->count) + { + if (desc != -1) + ret = facl (desc, SETACL, ctx->count, ctx->entries); + else + ret = acl (name, SETACL, ctx->count, ctx->entries); + if (ret < 0) + { + if ((errno == ENOSYS || errno == EOPNOTSUPP || errno == EINVAL) + && acl_nontrivial (ctx->count, ctx->entries) == 0) + ret = 0; + } + else + *acls_set = true; + } + +# ifdef ACE_GETACL + if (ret == 0 && ctx->ace_count) + { + if (desc != -1) + ret = facl (desc, ACE_SETACL, ctx->ace_count, ctx->ace_entries); + else + ret = acl (name, ACE_SETACL, ctx->ace_count, ctx->ace_entries); + if (ret < 0) + { + if ((errno == ENOSYS || errno == EINVAL || errno == ENOTSUP) + && acl_ace_nontrivial (ctx->ace_count, ctx->ace_entries) == 0) + ret = 0; + } + else + *acls_set = true; + } +# endif + +# elif HAVE_GETACL /* HP-UX */ + + if (from_mode) + ret = context_acl_from_mode (ctx, name, desc); + + if (ret == 0 && ctx->count > 0) + { + if (desc != -1) + ret = fsetacl (desc, ctx->count, ctx->entries); + else + ret = setacl (name, ctx->count, ctx->entries); + if (ret < 0) + { + if ((errno == ENOSYS || errno == EOPNOTSUPP || errno == ENOTSUP) + && (from_mode || !acl_nontrivial (ctx->count, ctx->entries))) + ret = 0; + } + else + *acls_set = true; + } + +# if HAVE_ACLV_H + if (from_mode) + ret = context_aclv_from_mode (ctx); + + if (ret == 0 && ctx->aclv_count > 0) + { + ret = acl ((char *) name, ACL_SET, ctx->aclv_count, ctx->aclv_entries); + if (ret < 0) + { + if ((errno == ENOSYS || errno == EOPNOTSUPP || errno == EINVAL) + && (from_mode || !aclv_nontrivial (ctx->aclv_count, ctx->aclv_entries))) + ret = 0; + } + else + *acls_set = true; + } +# endif + +# elif HAVE_ACLX_GET && ACL_AIX_WIP /* AIX */ + + /* TODO: Implement setting ACLs once get_permissions() reads them. */ + + if (from_mode) + ret = set_acls_from_mode (name, desc, mode, must_chmod); + +# elif HAVE_STATACL /* older AIX */ + + if (from_mode) + ret = context_acl_from_mode (ctx); + + if (ret == 0 && ctx->have_u) + { + if (desc != -1) + ret = fchacl (desc, &ctx->u.a, ctx->u.a.acl_len); + else + ret = chacl ((char *) name, &ctx->u.a, ctx->u.a.acl_len); + if (ret < 0) + { + if (errno == ENOSYS && from_mode) + ret = 0; + } + else + *acls_set = true; + } + +# elif HAVE_ACLSORT /* NonStop Kernel */ + + if (from_mode) + ret = context_acl_from_mode (ctx); + + if (ret == 0 && ctx->count) + { + ret = acl ((char *) name, ACL_SET, ctx->count, ctx->entries); + if (ret != 0) + { + if (!acl_nontrivial (ctx->count, ctx->entries)) + ret = 0; + } + else + *acls_set = true; + } + +# else /* No ACLs */ + + /* Nothing to do. */ + +# endif + + return ret; +} +#endif + +/* If DESC is a valid file descriptor use fchmod to change the + file's mode to MODE on systems that have fchmod. On systems + that don't have fchmod and if DESC is invalid, use chmod on + NAME instead. + Return 0 if successful. Return -1 and set errno upon failure. */ + +int +chmod_or_fchmod (const char *name, int desc, mode_t mode) +{ + if (HAVE_FCHMOD && desc != -1) + return fchmod (desc, mode); + else + return chmod (name, mode); +} + +/* Set the permissions in CTX on a file. If DESC is a valid file descriptor, + use file descriptor operations, else use filename based operations on NAME. + If access control lists are not available, fchmod the target file to the + mode in CTX. Also sets the non-permission bits of the destination file + (S_ISUID, S_ISGID, S_ISVTX) to those from the mode in CTX if any are set. + Return 0 if successful. Return -1 and set errno upon failure. */ + +int +set_permissions (struct permission_context *ctx, const char *name, int desc) +{ + bool acls_set _GL_UNUSED = false; + bool early_chmod; + bool must_chmod = false; + int ret = 0; + +#if USE_ACL +# if HAVE_STATACL + /* older AIX */ + /* There is no need to call chmod_or_fchmod, since the mode + bits S_ISUID, S_ISGID, S_ISVTX are also stored in the ACL. */ + + early_chmod = false; +# else + /* All other platforms */ + /* On Cygwin, it is necessary to call chmod before acl, because + chmod can change the contents of the ACL (in ways that don't + change the allowed accesses, but still visible). */ + + early_chmod = (! MODE_INSIDE_ACL || (ctx->mode & (S_ISUID | S_ISGID | S_ISVTX))); +# endif +#else + /* No ACLs */ + + early_chmod = true; +#endif + + if (early_chmod) + { + ret = chmod_or_fchmod (name, desc, ctx->mode); + if (ret != 0) + return -1; + } + +#if USE_ACL + ret = set_acls (ctx, name, desc, false, &must_chmod, &acls_set); + if (! acls_set) + { + int saved_errno = ret ? errno : 0; + + /* If we can't set an acl which we expect to be able to set, try setting + the permissions to ctx->mode. Due to possible inherited permissions, + we cannot simply chmod. */ + + ret = set_acls (ctx, name, desc, true, &must_chmod, &acls_set); + if (! acls_set) + must_chmod = true; + + if (saved_errno) + { + errno = saved_errno; + ret = -1; + } + } +#endif + + if (must_chmod && ! early_chmod) + { + int saved_errno = ret ? errno : 0; + + ret = chmod_or_fchmod (name, desc, ctx->mode); + + if (saved_errno) + { + errno = saved_errno; + ret = -1; + } + } + + return ret; +} diff --git a/lib/stat-macros.h b/lib/stat-macros.h new file mode 100644 index 0000000..690216c --- /dev/null +++ b/lib/stat-macros.h @@ -0,0 +1,3 @@ +/* All the mode bits that can be affected by chmod. */ +#define CHMOD_MODE_BITS \ + (S_ISUID | S_ISGID | S_ISVTX | S_IRWXU | S_IRWXG | S_IRWXO) diff --git a/lib/stat-time.c b/lib/stat-time.c new file mode 100644 index 0000000..81b83dd --- /dev/null +++ b/lib/stat-time.c @@ -0,0 +1,3 @@ +#include <config.h> +#define _GL_STAT_TIME_INLINE _GL_EXTERN_INLINE +#include "stat-time.h" diff --git a/lib/stat-time.h b/lib/stat-time.h new file mode 100644 index 0000000..69ebe85 --- /dev/null +++ b/lib/stat-time.h @@ -0,0 +1,252 @@ +/* stat-related time functions. + + Copyright (C) 2005, 2007, 2009-2018 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +/* Written by Paul Eggert. */ + +#ifndef STAT_TIME_H +#define STAT_TIME_H 1 + +#include "intprops.h" + +#include <errno.h> +#include <stddef.h> +#include <sys/stat.h> +#include <time.h> + +#ifndef _GL_INLINE_HEADER_BEGIN + #error "Please include config.h first." +#endif +_GL_INLINE_HEADER_BEGIN +#ifndef _GL_STAT_TIME_INLINE +# define _GL_STAT_TIME_INLINE _GL_INLINE +#endif + +#ifdef __cplusplus +extern "C" { +#endif + +/* STAT_TIMESPEC (ST, ST_XTIM) is the ST_XTIM member for *ST of type + struct timespec, if available. If not, then STAT_TIMESPEC_NS (ST, + ST_XTIM) is the nanosecond component of the ST_XTIM member for *ST, + if available. ST_XTIM can be st_atim, st_ctim, st_mtim, or st_birthtim + for access, status change, data modification, or birth (creation) + time respectively. + + These macros are private to stat-time.h. */ +#if _GL_WINDOWS_STAT_TIMESPEC || defined HAVE_STRUCT_STAT_ST_ATIM_TV_NSEC +# if _GL_WINDOWS_STAT_TIMESPEC || defined TYPEOF_STRUCT_STAT_ST_ATIM_IS_STRUCT_TIMESPEC +# define STAT_TIMESPEC(st, st_xtim) ((st)->st_xtim) +# else +# define STAT_TIMESPEC_NS(st, st_xtim) ((st)->st_xtim.tv_nsec) +# endif +#elif defined HAVE_STRUCT_STAT_ST_ATIMESPEC_TV_NSEC +# define STAT_TIMESPEC(st, st_xtim) ((st)->st_xtim##espec) +#elif defined HAVE_STRUCT_STAT_ST_ATIMENSEC +# define STAT_TIMESPEC_NS(st, st_xtim) ((st)->st_xtim##ensec) +#elif defined HAVE_STRUCT_STAT_ST_ATIM_ST__TIM_TV_NSEC +# define STAT_TIMESPEC_NS(st, st_xtim) ((st)->st_xtim.st__tim.tv_nsec) +#endif + +/* Return the nanosecond component of *ST's access time. */ +_GL_STAT_TIME_INLINE long int _GL_ATTRIBUTE_PURE +get_stat_atime_ns (struct stat const *st) +{ +# if defined STAT_TIMESPEC + return STAT_TIMESPEC (st, st_atim).tv_nsec; +# elif defined STAT_TIMESPEC_NS + return STAT_TIMESPEC_NS (st, st_atim); +# else + return 0; +# endif +} + +/* Return the nanosecond component of *ST's status change time. */ +_GL_STAT_TIME_INLINE long int _GL_ATTRIBUTE_PURE +get_stat_ctime_ns (struct stat const *st) +{ +# if defined STAT_TIMESPEC + return STAT_TIMESPEC (st, st_ctim).tv_nsec; +# elif defined STAT_TIMESPEC_NS + return STAT_TIMESPEC_NS (st, st_ctim); +# else + return 0; +# endif +} + +/* Return the nanosecond component of *ST's data modification time. */ +_GL_STAT_TIME_INLINE long int _GL_ATTRIBUTE_PURE +get_stat_mtime_ns (struct stat const *st) +{ +# if defined STAT_TIMESPEC + return STAT_TIMESPEC (st, st_mtim).tv_nsec; +# elif defined STAT_TIMESPEC_NS + return STAT_TIMESPEC_NS (st, st_mtim); +# else + return 0; +# endif +} + +/* Return the nanosecond component of *ST's birth time. */ +_GL_STAT_TIME_INLINE long int _GL_ATTRIBUTE_PURE +get_stat_birthtime_ns (struct stat const *st _GL_UNUSED) +{ +# if defined HAVE_STRUCT_STAT_ST_BIRTHTIMESPEC_TV_NSEC + return STAT_TIMESPEC (st, st_birthtim).tv_nsec; +# elif defined HAVE_STRUCT_STAT_ST_BIRTHTIMENSEC + return STAT_TIMESPEC_NS (st, st_birthtim); +# else + return 0; +# endif +} + +/* Return *ST's access time. */ +_GL_STAT_TIME_INLINE struct timespec _GL_ATTRIBUTE_PURE +get_stat_atime (struct stat const *st) +{ +#ifdef STAT_TIMESPEC + return STAT_TIMESPEC (st, st_atim); +#else + struct timespec t; + t.tv_sec = st->st_atime; + t.tv_nsec = get_stat_atime_ns (st); + return t; +#endif +} + +/* Return *ST's status change time. */ +_GL_STAT_TIME_INLINE struct timespec _GL_ATTRIBUTE_PURE +get_stat_ctime (struct stat const *st) +{ +#ifdef STAT_TIMESPEC + return STAT_TIMESPEC (st, st_ctim); +#else + struct timespec t; + t.tv_sec = st->st_ctime; + t.tv_nsec = get_stat_ctime_ns (st); + return t; +#endif +} + +/* Return *ST's data modification time. */ +_GL_STAT_TIME_INLINE struct timespec _GL_ATTRIBUTE_PURE +get_stat_mtime (struct stat const *st) +{ +#ifdef STAT_TIMESPEC + return STAT_TIMESPEC (st, st_mtim); +#else + struct timespec t; + t.tv_sec = st->st_mtime; + t.tv_nsec = get_stat_mtime_ns (st); + return t; +#endif +} + +/* Return *ST's birth time, if available; otherwise return a value + with tv_sec and tv_nsec both equal to -1. */ +_GL_STAT_TIME_INLINE struct timespec _GL_ATTRIBUTE_PURE +get_stat_birthtime (struct stat const *st _GL_UNUSED) +{ + struct timespec t; + +#if (defined HAVE_STRUCT_STAT_ST_BIRTHTIMESPEC_TV_NSEC \ + || defined HAVE_STRUCT_STAT_ST_BIRTHTIM_TV_NSEC) + t = STAT_TIMESPEC (st, st_birthtim); +#elif defined HAVE_STRUCT_STAT_ST_BIRTHTIMENSEC + t.tv_sec = st->st_birthtime; + t.tv_nsec = st->st_birthtimensec; +#elif defined _WIN32 && ! defined __CYGWIN__ + /* Native Windows platforms (but not Cygwin) put the "file creation + time" in st_ctime (!). See + <https://msdn.microsoft.com/en-us/library/14h5k7ff(VS.80).aspx>. */ +# if _GL_WINDOWS_STAT_TIMESPEC + t = st->st_ctim; +# else + t.tv_sec = st->st_ctime; + t.tv_nsec = 0; +# endif +#else + /* Birth time is not supported. */ + t.tv_sec = -1; + t.tv_nsec = -1; +#endif + +#if (defined HAVE_STRUCT_STAT_ST_BIRTHTIMESPEC_TV_NSEC \ + || defined HAVE_STRUCT_STAT_ST_BIRTHTIM_TV_NSEC \ + || defined HAVE_STRUCT_STAT_ST_BIRTHTIMENSEC) + /* FreeBSD and NetBSD sometimes signal the absence of knowledge by + using zero. Attempt to work around this problem. Alas, this can + report failure even for valid timestamps. Also, NetBSD + sometimes returns junk in the birth time fields; work around this + bug if it is detected. */ + if (! (t.tv_sec && 0 <= t.tv_nsec && t.tv_nsec < 1000000000)) + { + t.tv_sec = -1; + t.tv_nsec = -1; + } +#endif + + return t; +} + +/* If a stat-like function returned RESULT, normalize the timestamps + in *ST, in case this platform suffers from the Solaris 11 bug where + tv_nsec might be negative. Return the adjusted RESULT, setting + errno to EOVERFLOW if normalization overflowed. This function + is intended to be private to this .h file. */ +_GL_STAT_TIME_INLINE int +stat_time_normalize (int result, struct stat *st _GL_UNUSED) +{ +#if defined __sun && defined STAT_TIMESPEC + if (result == 0) + { + long int timespec_hz = 1000000000; + short int const ts_off[] = { offsetof (struct stat, st_atim), + offsetof (struct stat, st_mtim), + offsetof (struct stat, st_ctim) }; + int i; + for (i = 0; i < sizeof ts_off / sizeof *ts_off; i++) + { + struct timespec *ts = (struct timespec *) ((char *) st + ts_off[i]); + long int q = ts->tv_nsec / timespec_hz; + long int r = ts->tv_nsec % timespec_hz; + if (r < 0) + { + r += timespec_hz; + q--; + } + ts->tv_nsec = r; + /* Overflow is possible, as Solaris 11 stat can yield + tv_sec == TYPE_MINIMUM (time_t) && tv_nsec == -1000000000. + INT_ADD_WRAPV is OK, since time_t is signed on Solaris. */ + if (INT_ADD_WRAPV (q, ts->tv_sec, &ts->tv_sec)) + { + errno = EOVERFLOW; + return -1; + } + } + } +#endif + return result; +} + +#ifdef __cplusplus +} +#endif + +_GL_INLINE_HEADER_END + +#endif diff --git a/lib/stat-w32.c b/lib/stat-w32.c new file mode 100644 index 0000000..c7a53f9 --- /dev/null +++ b/lib/stat-w32.c @@ -0,0 +1,425 @@ +/* Core of implementation of fstat and stat for native Windows. + Copyright (C) 2017-2018 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +/* Written by Bruno Haible. */ + +#include <config.h> + +#if defined _WIN32 && ! defined __CYGWIN__ + +/* Ensure that <windows.h> defines FILE_ID_INFO. */ +#undef _WIN32_WINNT +#define _WIN32_WINNT _WIN32_WINNT_WIN8 + +#include <sys/types.h> +#include <sys/stat.h> +#include <errno.h> +#include <limits.h> +#include <string.h> +#include <unistd.h> +#include <windows.h> + +/* Specification. */ +#include "stat-w32.h" + +#include "pathmax.h" +#include "verify.h" + +/* Avoid warnings from gcc -Wcast-function-type. */ +#define GetProcAddress \ + (void *) GetProcAddress + +#if _GL_WINDOWS_STAT_INODES == 2 +/* GetFileInformationByHandleEx was introduced only in Windows Vista. */ +typedef DWORD (WINAPI * GetFileInformationByHandleExFuncType) (HANDLE hFile, + FILE_INFO_BY_HANDLE_CLASS fiClass, + LPVOID lpBuffer, + DWORD dwBufferSize); +static GetFileInformationByHandleExFuncType GetFileInformationByHandleExFunc = NULL; +#endif +/* GetFinalPathNameByHandle was introduced only in Windows Vista. */ +typedef DWORD (WINAPI * GetFinalPathNameByHandleFuncType) (HANDLE hFile, + LPTSTR lpFilePath, + DWORD lenFilePath, + DWORD dwFlags); +static GetFinalPathNameByHandleFuncType GetFinalPathNameByHandleFunc = NULL; +static BOOL initialized = FALSE; + +static void +initialize (void) +{ + HMODULE kernel32 = LoadLibrary ("kernel32.dll"); + if (kernel32 != NULL) + { +#if _GL_WINDOWS_STAT_INODES == 2 + GetFileInformationByHandleExFunc = + (GetFileInformationByHandleExFuncType) GetProcAddress (kernel32, "GetFileInformationByHandleEx"); +#endif + GetFinalPathNameByHandleFunc = + (GetFinalPathNameByHandleFuncType) GetProcAddress (kernel32, "GetFinalPathNameByHandleA"); + } + initialized = TRUE; +} + +/* Converts a FILETIME to GMT time since 1970-01-01 00:00:00. */ +#if _GL_WINDOWS_STAT_TIMESPEC +struct timespec +_gl_convert_FILETIME_to_timespec (const FILETIME *ft) +{ + struct timespec result; + /* FILETIME: <https://msdn.microsoft.com/en-us/library/ms724284.aspx> */ + unsigned long long since_1601 = + ((unsigned long long) ft->dwHighDateTime << 32) + | (unsigned long long) ft->dwLowDateTime; + if (since_1601 == 0) + { + result.tv_sec = 0; + result.tv_nsec = 0; + } + else + { + /* Between 1601-01-01 and 1970-01-01 there were 280 normal years and 89 + leap years, in total 134774 days. */ + unsigned long long since_1970 = + since_1601 - (unsigned long long) 134774 * (unsigned long long) 86400 * (unsigned long long) 10000000; + result.tv_sec = since_1970 / (unsigned long long) 10000000; + result.tv_nsec = (unsigned long) (since_1970 % (unsigned long long) 10000000) * 100; + } + return result; +} +#else +time_t +_gl_convert_FILETIME_to_POSIX (const FILETIME *ft) +{ + /* FILETIME: <https://msdn.microsoft.com/en-us/library/ms724284.aspx> */ + unsigned long long since_1601 = + ((unsigned long long) ft->dwHighDateTime << 32) + | (unsigned long long) ft->dwLowDateTime; + if (since_1601 == 0) + return 0; + else + { + /* Between 1601-01-01 and 1970-01-01 there were 280 normal years and 89 + leap years, in total 134774 days. */ + unsigned long long since_1970 = + since_1601 - (unsigned long long) 134774 * (unsigned long long) 86400 * (unsigned long long) 10000000; + return since_1970 / (unsigned long long) 10000000; + } +} +#endif + +/* Fill *BUF with information about the file designated by H. + PATH is the file name, if known, otherwise NULL. + Return 0 if successful, or -1 with errno set upon failure. */ +int +_gl_fstat_by_handle (HANDLE h, const char *path, struct stat *buf) +{ + /* GetFileType + <https://msdn.microsoft.com/en-us/library/aa364960.aspx> */ + DWORD type = GetFileType (h); + if (type == FILE_TYPE_DISK) + { + if (!initialized) + initialize (); + + /* st_mode can be determined through + GetFileAttributesEx + <https://msdn.microsoft.com/en-us/library/aa364946.aspx> + <https://msdn.microsoft.com/en-us/library/aa365739.aspx> + or through + GetFileInformationByHandle + <https://msdn.microsoft.com/en-us/library/aa364952.aspx> + <https://msdn.microsoft.com/en-us/library/aa363788.aspx> + or through + GetFileInformationByHandleEx with argument FileBasicInfo + <https://msdn.microsoft.com/en-us/library/aa364953.aspx> + <https://msdn.microsoft.com/en-us/library/aa364217.aspx> + The latter requires -D_WIN32_WINNT=_WIN32_WINNT_VISTA or higher. */ + BY_HANDLE_FILE_INFORMATION info; + if (! GetFileInformationByHandle (h, &info)) + goto failed; + + /* Test for error conditions before starting to fill *buf. */ + if (sizeof (buf->st_size) <= 4 && info.nFileSizeHigh > 0) + { + errno = EOVERFLOW; + return -1; + } + +#if _GL_WINDOWS_STAT_INODES + /* st_ino can be determined through + GetFileInformationByHandle + <https://msdn.microsoft.com/en-us/library/aa364952.aspx> + <https://msdn.microsoft.com/en-us/library/aa363788.aspx> + as 64 bits, or through + GetFileInformationByHandleEx with argument FileIdInfo + <https://msdn.microsoft.com/en-us/library/aa364953.aspx> + <https://msdn.microsoft.com/en-us/library/hh802691.aspx> + as 128 bits. + The latter requires -D_WIN32_WINNT=_WIN32_WINNT_WIN8 or higher. */ + /* Experiments show that GetFileInformationByHandleEx does not provide + much more information than GetFileInformationByHandle: + * The dwVolumeSerialNumber from GetFileInformationByHandle is equal + to the low 32 bits of the 64-bit VolumeSerialNumber from + GetFileInformationByHandleEx, and is apparently sufficient for + identifying the device. + * The nFileIndex from GetFileInformationByHandle is equal to the low + 64 bits of the 128-bit FileId from GetFileInformationByHandleEx, + and the high 64 bits of this 128-bit FileId are zero. + * On a FAT file system, GetFileInformationByHandleEx fails with error + ERROR_INVALID_PARAMETER, whereas GetFileInformationByHandle + succeeds. + * On a CIFS/SMB file system, GetFileInformationByHandleEx fails with + error ERROR_INVALID_LEVEL, whereas GetFileInformationByHandle + succeeds. */ +# if _GL_WINDOWS_STAT_INODES == 2 + if (GetFileInformationByHandleExFunc != NULL) + { + FILE_ID_INFO id; + if (GetFileInformationByHandleExFunc (h, FileIdInfo, &id, sizeof (id))) + { + buf->st_dev = id.VolumeSerialNumber; + verify (sizeof (ino_t) == sizeof (id.FileId)); + memcpy (&buf->st_ino, &id.FileId, sizeof (ino_t)); + goto ino_done; + } + else + { + switch (GetLastError ()) + { + case ERROR_INVALID_PARAMETER: /* older Windows version, or FAT */ + case ERROR_INVALID_LEVEL: /* CIFS/SMB file system */ + goto fallback; + default: + goto failed; + } + } + } + fallback: ; + /* Fallback for older Windows versions. */ + buf->st_dev = info.dwVolumeSerialNumber; + buf->st_ino._gl_ino[0] = ((ULONGLONG) info.nFileIndexHigh << 32) | (ULONGLONG) info.nFileIndexLow; + buf->st_ino._gl_ino[1] = 0; + ino_done: ; +# else /* _GL_WINDOWS_STAT_INODES == 1 */ + buf->st_dev = info.dwVolumeSerialNumber; + buf->st_ino = ((ULONGLONG) info.nFileIndexHigh << 32) | (ULONGLONG) info.nFileIndexLow; +# endif +#else + /* st_ino is not wide enough for identifying a file on a device. + Without st_ino, st_dev is pointless. */ + buf->st_dev = 0; + buf->st_ino = 0; +#endif + + /* st_mode. */ + unsigned int mode = + /* XXX How to handle FILE_ATTRIBUTE_REPARSE_POINT ? */ + ((info.dwFileAttributes & FILE_ATTRIBUTE_DIRECTORY) ? _S_IFDIR | S_IEXEC_UGO : _S_IFREG) + | S_IREAD_UGO + | ((info.dwFileAttributes & FILE_ATTRIBUTE_READONLY) ? 0 : S_IWRITE_UGO); + if (!(info.dwFileAttributes & FILE_ATTRIBUTE_DIRECTORY)) + { + /* Determine whether the file is executable by looking at the file + name suffix. + If the file name is already known, use it. Otherwise, for + non-empty files, it can be determined through + GetFinalPathNameByHandle + <https://msdn.microsoft.com/en-us/library/aa364962.aspx> + or through + GetFileInformationByHandleEx with argument FileNameInfo + <https://msdn.microsoft.com/en-us/library/aa364953.aspx> + <https://msdn.microsoft.com/en-us/library/aa364388.aspx> + Both require -D_WIN32_WINNT=_WIN32_WINNT_VISTA or higher. */ + if (info.nFileSizeHigh > 0 || info.nFileSizeLow > 0) + { + char fpath[PATH_MAX]; + if (path != NULL + || (GetFinalPathNameByHandleFunc != NULL + && GetFinalPathNameByHandleFunc (h, fpath, sizeof (fpath), VOLUME_NAME_NONE) + < sizeof (fpath) + && (path = fpath, 1))) + { + const char *last_dot = NULL; + const char *p; + for (p = path; *p != '\0'; p++) + if (*p == '.') + last_dot = p; + if (last_dot != NULL) + { + const char *suffix = last_dot + 1; + if (_stricmp (suffix, "exe") == 0 + || _stricmp (suffix, "bat") == 0 + || _stricmp (suffix, "cmd") == 0 + || _stricmp (suffix, "com") == 0) + mode |= S_IEXEC_UGO; + } + } + else + /* Cannot determine file name. Pretend that it is executable. */ + mode |= S_IEXEC_UGO; + } + } + buf->st_mode = mode; + + /* st_nlink can be determined through + GetFileInformationByHandle + <https://msdn.microsoft.com/en-us/library/aa364952.aspx> + <https://msdn.microsoft.com/en-us/library/aa363788.aspx> + or through + GetFileInformationByHandleEx with argument FileStandardInfo + <https://msdn.microsoft.com/en-us/library/aa364953.aspx> + <https://msdn.microsoft.com/en-us/library/aa364401.aspx> + The latter requires -D_WIN32_WINNT=_WIN32_WINNT_VISTA or higher. */ + buf->st_nlink = (info.nNumberOfLinks > SHRT_MAX ? SHRT_MAX : info.nNumberOfLinks); + + /* There's no easy way to map the Windows SID concept to an integer. */ + buf->st_uid = 0; + buf->st_gid = 0; + + /* st_rdev is irrelevant for normal files and directories. */ + buf->st_rdev = 0; + + /* st_size can be determined through + GetFileSizeEx + <https://msdn.microsoft.com/en-us/library/aa364957.aspx> + or through + GetFileAttributesEx + <https://msdn.microsoft.com/en-us/library/aa364946.aspx> + <https://msdn.microsoft.com/en-us/library/aa365739.aspx> + or through + GetFileInformationByHandle + <https://msdn.microsoft.com/en-us/library/aa364952.aspx> + <https://msdn.microsoft.com/en-us/library/aa363788.aspx> + or through + GetFileInformationByHandleEx with argument FileStandardInfo + <https://msdn.microsoft.com/en-us/library/aa364953.aspx> + <https://msdn.microsoft.com/en-us/library/aa364401.aspx> + The latter requires -D_WIN32_WINNT=_WIN32_WINNT_VISTA or higher. */ + if (sizeof (buf->st_size) <= 4) + /* Range check already done above. */ + buf->st_size = info.nFileSizeLow; + else + buf->st_size = ((long long) info.nFileSizeHigh << 32) | (long long) info.nFileSizeLow; + + /* st_atime, st_mtime, st_ctime can be determined through + GetFileTime + <https://msdn.microsoft.com/en-us/library/ms724320.aspx> + or through + GetFileAttributesEx + <https://msdn.microsoft.com/en-us/library/aa364946.aspx> + <https://msdn.microsoft.com/en-us/library/aa365739.aspx> + or through + GetFileInformationByHandle + <https://msdn.microsoft.com/en-us/library/aa364952.aspx> + <https://msdn.microsoft.com/en-us/library/aa363788.aspx> + or through + GetFileInformationByHandleEx with argument FileBasicInfo + <https://msdn.microsoft.com/en-us/library/aa364953.aspx> + <https://msdn.microsoft.com/en-us/library/aa364217.aspx> + The latter requires -D_WIN32_WINNT=_WIN32_WINNT_VISTA or higher. */ +#if _GL_WINDOWS_STAT_TIMESPEC + buf->st_atim = _gl_convert_FILETIME_to_timespec (&info.ftLastAccessTime); + buf->st_mtim = _gl_convert_FILETIME_to_timespec (&info.ftLastWriteTime); + buf->st_ctim = _gl_convert_FILETIME_to_timespec (&info.ftCreationTime); +#else + buf->st_atime = _gl_convert_FILETIME_to_POSIX (&info.ftLastAccessTime); + buf->st_mtime = _gl_convert_FILETIME_to_POSIX (&info.ftLastWriteTime); + buf->st_ctime = _gl_convert_FILETIME_to_POSIX (&info.ftCreationTime); +#endif + + return 0; + } + else if (type == FILE_TYPE_CHAR || type == FILE_TYPE_PIPE) + { + buf->st_dev = 0; +#if _GL_WINDOWS_STAT_INODES == 2 + buf->st_ino._gl_ino[0] = buf->st_ino._gl_ino[1] = 0; +#else + buf->st_ino = 0; +#endif + buf->st_mode = (type == FILE_TYPE_PIPE ? _S_IFIFO : _S_IFCHR); + buf->st_nlink = 1; + buf->st_uid = 0; + buf->st_gid = 0; + buf->st_rdev = 0; + if (type == FILE_TYPE_PIPE) + { + /* PeekNamedPipe + <https://msdn.microsoft.com/en-us/library/aa365779.aspx> */ + DWORD bytes_available; + if (PeekNamedPipe (h, NULL, 0, NULL, &bytes_available, NULL)) + buf->st_size = bytes_available; + else + buf->st_size = 0; + } + else + buf->st_size = 0; +#if _GL_WINDOWS_STAT_TIMESPEC + buf->st_atim.tv_sec = 0; buf->st_atim.tv_nsec = 0; + buf->st_mtim.tv_sec = 0; buf->st_mtim.tv_nsec = 0; + buf->st_ctim.tv_sec = 0; buf->st_ctim.tv_nsec = 0; +#else + buf->st_atime = 0; + buf->st_mtime = 0; + buf->st_ctime = 0; +#endif + return 0; + } + else + { + errno = ENOENT; + return -1; + } + + failed: + { + DWORD error = GetLastError (); + #if 0 + fprintf (stderr, "_gl_fstat_by_handle error 0x%x\n", (unsigned int) error); + #endif + switch (error) + { + case ERROR_ACCESS_DENIED: + case ERROR_SHARING_VIOLATION: + errno = EACCES; + break; + + case ERROR_OUTOFMEMORY: + errno = ENOMEM; + break; + + case ERROR_WRITE_FAULT: + case ERROR_READ_FAULT: + case ERROR_GEN_FAILURE: + errno = EIO; + break; + + default: + errno = EINVAL; + break; + } + return -1; + } +} + +#else + +/* This declaration is solely to ensure that after preprocessing + this file is never empty. */ +typedef int dummy; + +#endif diff --git a/lib/stat-w32.h b/lib/stat-w32.h new file mode 100644 index 0000000..9a3ecc1 --- /dev/null +++ b/lib/stat-w32.h @@ -0,0 +1,37 @@ +/* Core of implementation of fstat and stat for native Windows. + Copyright (C) 2017-2018 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +#ifndef _STAT_W32_H +#define _STAT_W32_H 1 + +/* Converts a FILETIME to GMT time since 1970-01-01 00:00:00. */ +#if _GL_WINDOWS_STAT_TIMESPEC +extern struct timespec _gl_convert_FILETIME_to_timespec (const FILETIME *ft); +#else +extern time_t _gl_convert_FILETIME_to_POSIX (const FILETIME *ft); +#endif + +/* Fill *BUF with information about the file designated by H. + PATH is the file name, if known, otherwise NULL. + Return 0 if successful, or -1 with errno set upon failure. */ +extern int _gl_fstat_by_handle (HANDLE h, const char *path, struct stat *buf); + +/* Bitmasks for st_mode. */ +#define S_IREAD_UGO (_S_IREAD | (_S_IREAD >> 3) | (_S_IREAD >> 6)) +#define S_IWRITE_UGO (_S_IWRITE | (_S_IWRITE >> 3) | (_S_IWRITE >> 6)) +#define S_IEXEC_UGO (_S_IEXEC | (_S_IEXEC >> 3) | (_S_IEXEC >> 6)) + +#endif /* _STAT_W32_H */ diff --git a/lib/stat.c b/lib/stat.c new file mode 100644 index 0000000..18c0caa --- /dev/null +++ b/lib/stat.c @@ -0,0 +1,429 @@ +/* Work around platform bugs in stat. + Copyright (C) 2009-2018 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +/* Written by Eric Blake and Bruno Haible. */ + +/* If the user's config.h happens to include <sys/stat.h>, let it include only + the system's <sys/stat.h> here, so that orig_stat doesn't recurse to + rpl_stat. */ +#define __need_system_sys_stat_h +#include <config.h> + +/* Get the original definition of stat. It might be defined as a macro. */ +#include <sys/types.h> +#include <sys/stat.h> +#undef __need_system_sys_stat_h + +#if defined _WIN32 && ! defined __CYGWIN__ +# define WINDOWS_NATIVE +#endif + +#if !defined WINDOWS_NATIVE + +static int +orig_stat (const char *filename, struct stat *buf) +{ + return stat (filename, buf); +} + +#endif + +/* Specification. */ +/* Write "sys/stat.h" here, not <sys/stat.h>, otherwise OSF/1 5.1 DTK cc + eliminates this include because of the preliminary #include <sys/stat.h> + above. */ +#include "sys/stat.h" + +#include "stat-time.h" + +#include <errno.h> +#include <limits.h> +#include <stdbool.h> +#include <string.h> +#include "filename.h" +#include "malloca.h" +#include "verify.h" + +#ifdef WINDOWS_NATIVE +# define WIN32_LEAN_AND_MEAN +# include <windows.h> +# include "stat-w32.h" +#endif + +#ifdef WINDOWS_NATIVE +/* Return TRUE if the given file name denotes an UNC root. */ +static BOOL +is_unc_root (const char *rname) +{ + /* Test whether it has the syntax '\\server\share'. */ + if (ISSLASH (rname[0]) && ISSLASH (rname[1])) + { + /* It starts with two slashes. Find the next slash. */ + const char *p = rname + 2; + const char *q = p; + while (*q != '\0' && !ISSLASH (*q)) + q++; + if (q > p && *q != '\0') + { + /* Found the next slash at q. */ + q++; + const char *r = q; + while (*r != '\0' && !ISSLASH (*r)) + r++; + if (r > q && *r == '\0') + return TRUE; + } + } + return FALSE; +} +#endif + +/* Store information about NAME into ST. Work around bugs with + trailing slashes. Mingw has other bugs (such as st_ino always + being 0 on success) which this wrapper does not work around. But + at least this implementation provides the ability to emulate fchdir + correctly. */ + +int +rpl_stat (char const *name, struct stat *buf) +{ +#ifdef WINDOWS_NATIVE + /* Fill the fields ourselves, because the original stat function returns + values for st_atime, st_mtime, st_ctime that depend on the current time + zone. See + <https://lists.gnu.org/r/bug-gnulib/2017-04/msg00134.html> */ + /* XXX Should we convert to wchar_t* and prepend '\\?\', in order to work + around length limitations + <https://msdn.microsoft.com/en-us/library/aa365247.aspx> ? */ + + /* POSIX <http://pubs.opengroup.org/onlinepubs/9699919799/basedefs/V1_chap04.html#tag_04_13> + specifies: "More than two leading <slash> characters shall be treated as + a single <slash> character." */ + if (ISSLASH (name[0]) && ISSLASH (name[1]) && ISSLASH (name[2])) + { + name += 2; + while (ISSLASH (name[1])) + name++; + } + + size_t len = strlen (name); + size_t drive_prefix_len = (HAS_DEVICE (name) ? 2 : 0); + + /* Remove trailing slashes (except the very first one, at position + drive_prefix_len), but remember their presence. */ + size_t rlen; + bool check_dir = false; + + rlen = len; + while (rlen > drive_prefix_len && ISSLASH (name[rlen-1])) + { + check_dir = true; + if (rlen == drive_prefix_len + 1) + break; + rlen--; + } + + /* Handle '' and 'C:'. */ + if (!check_dir && rlen == drive_prefix_len) + { + errno = ENOENT; + return -1; + } + + /* Handle '\\'. */ + if (rlen == 1 && ISSLASH (name[0]) && len >= 2) + { + errno = ENOENT; + return -1; + } + + const char *rname; + char *malloca_rname; + if (rlen == len) + { + rname = name; + malloca_rname = NULL; + } + else + { + malloca_rname = malloca (rlen + 1); + if (malloca_rname == NULL) + { + errno = ENOMEM; + return -1; + } + memcpy (malloca_rname, name, rlen); + malloca_rname[rlen] = '\0'; + rname = malloca_rname; + } + + /* There are two ways to get at the requested information: + - by scanning the parent directory and examining the relevant + directory entry, + - by opening the file directly. + The first approach fails for root directories (e.g. 'C:\') and + UNC root directories (e.g. '\\server\share'). + The second approach fails for some system files (e.g. 'C:\pagefile.sys' + and 'C:\hiberfil.sys'): ERROR_SHARING_VIOLATION. + The second approach gives more information (in particular, correct + st_dev, st_ino, st_nlink fields). + So we use the second approach and, as a fallback except for root and + UNC root directories, also the first approach. */ + { + int ret; + + { + /* Approach based on the file. */ + + /* Open a handle to the file. + CreateFile + <https://msdn.microsoft.com/en-us/library/aa363858.aspx> + <https://msdn.microsoft.com/en-us/library/aa363874.aspx> */ + HANDLE h = + CreateFile (rname, + FILE_READ_ATTRIBUTES, + FILE_SHARE_READ | FILE_SHARE_WRITE | FILE_SHARE_DELETE, + NULL, + OPEN_EXISTING, + /* FILE_FLAG_POSIX_SEMANTICS (treat file names that differ only + in case as different) makes sense only when applied to *all* + filesystem operations. */ + FILE_FLAG_BACKUP_SEMANTICS /* | FILE_FLAG_POSIX_SEMANTICS */, + NULL); + if (h != INVALID_HANDLE_VALUE) + { + ret = _gl_fstat_by_handle (h, rname, buf); + CloseHandle (h); + goto done; + } + } + + /* Test for root and UNC root directories. */ + if ((rlen == drive_prefix_len + 1 && ISSLASH (rname[drive_prefix_len])) + || is_unc_root (rname)) + goto failed; + + /* Fallback. */ + { + /* Approach based on the directory entry. */ + + if (strchr (rname, '?') != NULL || strchr (rname, '*') != NULL) + { + /* Other Windows API functions would fail with error + ERROR_INVALID_NAME. */ + if (malloca_rname != NULL) + freea (malloca_rname); + errno = ENOENT; + return -1; + } + + /* Get the details about the directory entry. This can be done through + FindFirstFile + <https://msdn.microsoft.com/en-us/library/aa364418.aspx> + <https://msdn.microsoft.com/en-us/library/aa365740.aspx> + or through + FindFirstFileEx with argument FindExInfoBasic + <https://msdn.microsoft.com/en-us/library/aa364419.aspx> + <https://msdn.microsoft.com/en-us/library/aa364415.aspx> + <https://msdn.microsoft.com/en-us/library/aa365740.aspx> */ + WIN32_FIND_DATA info; + HANDLE h = FindFirstFile (rname, &info); + if (h == INVALID_HANDLE_VALUE) + goto failed; + + /* Test for error conditions before starting to fill *buf. */ + if (sizeof (buf->st_size) <= 4 && info.nFileSizeHigh > 0) + { + FindClose (h); + if (malloca_rname != NULL) + freea (malloca_rname); + errno = EOVERFLOW; + return -1; + } + +# if _GL_WINDOWS_STAT_INODES + buf->st_dev = 0; +# if _GL_WINDOWS_STAT_INODES == 2 + buf->st_ino._gl_ino[0] = buf->st_ino._gl_ino[1] = 0; +# else /* _GL_WINDOWS_STAT_INODES == 1 */ + buf->st_ino = 0; +# endif +# else + /* st_ino is not wide enough for identifying a file on a device. + Without st_ino, st_dev is pointless. */ + buf->st_dev = 0; + buf->st_ino = 0; +# endif + + /* st_mode. */ + unsigned int mode = + /* XXX How to handle FILE_ATTRIBUTE_REPARSE_POINT ? */ + ((info.dwFileAttributes & FILE_ATTRIBUTE_DIRECTORY) ? _S_IFDIR | S_IEXEC_UGO : _S_IFREG) + | S_IREAD_UGO + | ((info.dwFileAttributes & FILE_ATTRIBUTE_READONLY) ? 0 : S_IWRITE_UGO); + if (!(info.dwFileAttributes & FILE_ATTRIBUTE_DIRECTORY)) + { + /* Determine whether the file is executable by looking at the file + name suffix. */ + if (info.nFileSizeHigh > 0 || info.nFileSizeLow > 0) + { + const char *last_dot = NULL; + const char *p; + for (p = info.cFileName; *p != '\0'; p++) + if (*p == '.') + last_dot = p; + if (last_dot != NULL) + { + const char *suffix = last_dot + 1; + if (_stricmp (suffix, "exe") == 0 + || _stricmp (suffix, "bat") == 0 + || _stricmp (suffix, "cmd") == 0 + || _stricmp (suffix, "com") == 0) + mode |= S_IEXEC_UGO; + } + } + } + buf->st_mode = mode; + + /* st_nlink. Ignore hard links here. */ + buf->st_nlink = 1; + + /* There's no easy way to map the Windows SID concept to an integer. */ + buf->st_uid = 0; + buf->st_gid = 0; + + /* st_rdev is irrelevant for normal files and directories. */ + buf->st_rdev = 0; + + /* st_size. */ + if (sizeof (buf->st_size) <= 4) + /* Range check already done above. */ + buf->st_size = info.nFileSizeLow; + else + buf->st_size = ((long long) info.nFileSizeHigh << 32) | (long long) info.nFileSizeLow; + + /* st_atime, st_mtime, st_ctime. */ +# if _GL_WINDOWS_STAT_TIMESPEC + buf->st_atim = _gl_convert_FILETIME_to_timespec (&info.ftLastAccessTime); + buf->st_mtim = _gl_convert_FILETIME_to_timespec (&info.ftLastWriteTime); + buf->st_ctim = _gl_convert_FILETIME_to_timespec (&info.ftCreationTime); +# else + buf->st_atime = _gl_convert_FILETIME_to_POSIX (&info.ftLastAccessTime); + buf->st_mtime = _gl_convert_FILETIME_to_POSIX (&info.ftLastWriteTime); + buf->st_ctime = _gl_convert_FILETIME_to_POSIX (&info.ftCreationTime); +# endif + + FindClose (h); + + ret = 0; + } + + done: + if (ret >= 0 && check_dir && !S_ISDIR (buf->st_mode)) + { + errno = ENOTDIR; + ret = -1; + } + if (malloca_rname != NULL) + { + int saved_errno = errno; + freea (malloca_rname); + errno = saved_errno; + } + return ret; + } + + failed: + { + DWORD error = GetLastError (); + #if 0 + fprintf (stderr, "rpl_stat error 0x%x\n", (unsigned int) error); + #endif + + if (malloca_rname != NULL) + freea (malloca_rname); + + switch (error) + { + /* Some of these errors probably cannot happen with the specific flags + that we pass to CreateFile. But who knows... */ + case ERROR_FILE_NOT_FOUND: /* The last component of rname does not exist. */ + case ERROR_PATH_NOT_FOUND: /* Some directory component in rname does not exist. */ + case ERROR_BAD_PATHNAME: /* rname is such as '\\server'. */ + case ERROR_BAD_NET_NAME: /* rname is such as '\\server\nonexistentshare'. */ + case ERROR_INVALID_NAME: /* rname contains wildcards, misplaced colon, etc. */ + case ERROR_DIRECTORY: + errno = ENOENT; + break; + + case ERROR_ACCESS_DENIED: /* rname is such as 'C:\System Volume Information\foo'. */ + case ERROR_SHARING_VIOLATION: /* rname is such as 'C:\pagefile.sys' (second approach only). */ + /* XXX map to EACCESS or EPERM? */ + errno = EACCES; + break; + + case ERROR_OUTOFMEMORY: + errno = ENOMEM; + break; + + case ERROR_WRITE_PROTECT: + errno = EROFS; + break; + + case ERROR_WRITE_FAULT: + case ERROR_READ_FAULT: + case ERROR_GEN_FAILURE: + errno = EIO; + break; + + case ERROR_BUFFER_OVERFLOW: + case ERROR_FILENAME_EXCED_RANGE: + errno = ENAMETOOLONG; + break; + + case ERROR_DELETE_PENDING: /* XXX map to EACCESS or EPERM? */ + errno = EPERM; + break; + + default: + errno = EINVAL; + break; + } + + return -1; + } +#else + int result = orig_stat (name, buf); + if (result == 0) + { +# if REPLACE_FUNC_STAT_FILE + /* Solaris 9 mistakenly succeeds when given a non-directory with a + trailing slash. */ + if (!S_ISDIR (buf->st_mode)) + { + size_t len = strlen (name); + if (ISSLASH (name[len - 1])) + { + errno = ENOTDIR; + return -1; + } + } +# endif /* REPLACE_FUNC_STAT_FILE */ + result = stat_time_normalize (result, buf); + } + return result; +#endif +} diff --git a/lib/stdalign.in.h b/lib/stdalign.in.h new file mode 100644 index 0000000..bbfa9ac --- /dev/null +++ b/lib/stdalign.in.h @@ -0,0 +1,121 @@ +/* A substitute for ISO C11 <stdalign.h>. + + Copyright 2011-2018 Free Software Foundation, Inc. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3, or (at your option) + any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, see <https://www.gnu.org/licenses/>. */ + +/* Written by Paul Eggert and Bruno Haible. */ + +#ifndef _GL_STDALIGN_H +#define _GL_STDALIGN_H + +/* ISO C11 <stdalign.h> for platforms that lack it. + + References: + ISO C11 (latest free draft + <http://www.open-std.org/jtc1/sc22/wg14/www/docs/n1570.pdf>) + sections 6.5.3.4, 6.7.5, 7.15. + C++11 (latest free draft + <http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2011/n3242.pdf>) + section 18.10. */ + +/* alignof (TYPE), also known as _Alignof (TYPE), yields the alignment + requirement of a structure member (i.e., slot or field) that is of + type TYPE, as an integer constant expression. + + This differs from GCC's __alignof__ operator, which can yield a + better-performing alignment for an object of that type. For + example, on x86 with GCC, __alignof__ (double) and __alignof__ + (long long) are 8, whereas alignof (double) and alignof (long long) + are 4 unless the option '-malign-double' is used. + + The result cannot be used as a value for an 'enum' constant, if you + want to be portable to HP-UX 10.20 cc and AIX 3.2.5 xlc. + + Include <stddef.h> for offsetof. */ +#include <stddef.h> + +/* FreeBSD 9.1 <sys/cdefs.h>, included by <stddef.h> and lots of other + standard headers, defines conflicting implementations of _Alignas + and _Alignof that are no better than ours; override them. */ +#undef _Alignas +#undef _Alignof + +/* GCC releases before GCC 4.9 had a bug in _Alignof. See GCC bug 52023 + <https://gcc.gnu.org/bugzilla/show_bug.cgi?id=52023>. */ +#if (!defined __STDC_VERSION__ || __STDC_VERSION__ < 201112 \ + || (defined __GNUC__ && __GNUC__ < 4 + (__GNUC_MINOR__ < 9))) +# ifdef __cplusplus +# if 201103 <= __cplusplus +# define _Alignof(type) alignof (type) +# else + template <class __t> struct __alignof_helper { char __a; __t __b; }; +# define _Alignof(type) offsetof (__alignof_helper<type>, __b) +# endif +# else +# define _Alignof(type) offsetof (struct { char __a; type __b; }, __b) +# endif +#endif +#if ! (defined __cplusplus && 201103 <= __cplusplus) +# define alignof _Alignof +#endif +#define __alignof_is_defined 1 + +/* alignas (A), also known as _Alignas (A), aligns a variable or type + to the alignment A, where A is an integer constant expression. For + example: + + int alignas (8) foo; + struct s { int a; int alignas (8) bar; }; + + aligns the address of FOO and the offset of BAR to be multiples of 8. + + A should be a power of two that is at least the type's alignment + and at most the implementation's alignment limit. This limit is + 2**28 on typical GNUish hosts, and 2**13 on MSVC. To be portable + to MSVC through at least version 10.0, A should be an integer + constant, as MSVC does not support expressions such as 1 << 3. + To be portable to Sun C 5.11, do not align auto variables to + anything stricter than their default alignment. + + The following C11 requirements are not supported here: + + - If A is zero, alignas has no effect. + - alignas can be used multiple times; the strictest one wins. + - alignas (TYPE) is equivalent to alignas (alignof (TYPE)). + + */ + +#if !defined __STDC_VERSION__ || __STDC_VERSION__ < 201112 +# if defined __cplusplus && 201103 <= __cplusplus +# define _Alignas(a) alignas (a) +# elif ((defined __APPLE__ && defined __MACH__ \ + ? 4 < __GNUC__ + (1 <= __GNUC_MINOR__) \ + : __GNUC__) \ + || (__ia64 && (61200 <= __HP_cc || 61200 <= __HP_aCC)) \ + || __ICC || 0x590 <= __SUNPRO_C || 0x0600 <= __xlC__) +# define _Alignas(a) __attribute__ ((__aligned__ (a))) +# elif 1300 <= _MSC_VER +# define _Alignas(a) __declspec (align (a)) +# endif +#endif +#if ((defined _Alignas && ! (defined __cplusplus && 201103 <= __cplusplus)) \ + || (defined __STDC_VERSION__ && 201112 <= __STDC_VERSION__)) +# define alignas _Alignas +#endif +#if defined alignas || (defined __cplusplus && 201103 <= __cplusplus) +# define __alignas_is_defined 1 +#endif + +#endif /* _GL_STDALIGN_H */ diff --git a/lib/stdarg.in.h b/lib/stdarg.in.h new file mode 100644 index 0000000..46af0c2 --- /dev/null +++ b/lib/stdarg.in.h @@ -0,0 +1,35 @@ +/* Substitute for and wrapper around <stdarg.h>. + Copyright (C) 2008-2018 Free Software Foundation, Inc. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3, or (at your option) + any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, see <https://www.gnu.org/licenses/>. */ + +#ifndef _@GUARD_PREFIX@_STDARG_H + +#if __GNUC__ >= 3 +@PRAGMA_SYSTEM_HEADER@ +#endif +@PRAGMA_COLUMNS@ + +/* The include_next requires a split double-inclusion guard. */ +#@INCLUDE_NEXT@ @NEXT_STDARG_H@ + +#ifndef _@GUARD_PREFIX@_STDARG_H +#define _@GUARD_PREFIX@_STDARG_H + +#ifndef va_copy +# define va_copy(a,b) ((a) = (b)) +#endif + +#endif /* _@GUARD_PREFIX@_STDARG_H */ +#endif /* _@GUARD_PREFIX@_STDARG_H */ diff --git a/lib/stdbool.in.h b/lib/stdbool.in.h new file mode 100644 index 0000000..ebf1a2f --- /dev/null +++ b/lib/stdbool.in.h @@ -0,0 +1,132 @@ +/* Copyright (C) 2001-2003, 2006-2018 Free Software Foundation, Inc. + Written by Bruno Haible <haible@clisp.cons.org>, 2001. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3, or (at your option) + any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, see <https://www.gnu.org/licenses/>. */ + +#ifndef _GL_STDBOOL_H +#define _GL_STDBOOL_H + +/* ISO C 99 <stdbool.h> for platforms that lack it. */ + +/* Usage suggestions: + + Programs that use <stdbool.h> should be aware of some limitations + and standards compliance issues. + + Standards compliance: + + - <stdbool.h> must be #included before 'bool', 'false', 'true' + can be used. + + - You cannot assume that sizeof (bool) == 1. + + - Programs should not undefine the macros bool, true, and false, + as C99 lists that as an "obsolescent feature". + + Limitations of this substitute, when used in a C89 environment: + + - <stdbool.h> must be #included before the '_Bool' type can be used. + + - You cannot assume that _Bool is a typedef; it might be a macro. + + - Bit-fields of type 'bool' are not supported. Portable code + should use 'unsigned int foo : 1;' rather than 'bool foo : 1;'. + + - In C99, casts and automatic conversions to '_Bool' or 'bool' are + performed in such a way that every nonzero value gets converted + to 'true', and zero gets converted to 'false'. This doesn't work + with this substitute. With this substitute, only the values 0 and 1 + give the expected result when converted to _Bool' or 'bool'. + + - C99 allows the use of (_Bool)0.0 in constant expressions, but + this substitute cannot always provide this property. + + Also, it is suggested that programs use 'bool' rather than '_Bool'; + this isn't required, but 'bool' is more common. */ + + +/* 7.16. Boolean type and values */ + +/* BeOS <sys/socket.h> already #defines false 0, true 1. We use the same + definitions below, but temporarily we have to #undef them. */ +#if defined __BEOS__ && !defined __HAIKU__ +# include <OS.h> /* defines bool but not _Bool */ +# undef false +# undef true +#endif + +#ifdef __cplusplus +# define _Bool bool +# define bool bool +#else +# if defined __BEOS__ && !defined __HAIKU__ + /* A compiler known to have 'bool'. */ + /* If the compiler already has both 'bool' and '_Bool', we can assume they + are the same types. */ +# if !@HAVE__BOOL@ +typedef bool _Bool; +# endif +# else +# if !defined __GNUC__ + /* If @HAVE__BOOL@: + Some HP-UX cc and AIX IBM C compiler versions have compiler bugs when + the built-in _Bool type is used. See + https://gcc.gnu.org/ml/gcc-patches/2003-12/msg02303.html + https://lists.gnu.org/r/bug-coreutils/2005-11/msg00161.html + https://lists.gnu.org/r/bug-coreutils/2005-10/msg00086.html + Similar bugs are likely with other compilers as well; this file + wouldn't be used if <stdbool.h> was working. + So we override the _Bool type. + If !@HAVE__BOOL@: + Need to define _Bool ourselves. As 'signed char' or as an enum type? + Use of a typedef, with SunPRO C, leads to a stupid + "warning: _Bool is a keyword in ISO C99". + Use of an enum type, with IRIX cc, leads to a stupid + "warning(1185): enumerated type mixed with another type". + Even the existence of an enum type, without a typedef, + "Invalid enumerator. (badenum)" with HP-UX cc on Tru64. + The only benefit of the enum, debuggability, is not important + with these compilers. So use 'signed char' and no enum. */ +# define _Bool signed char +# else + /* With this compiler, trust the _Bool type if the compiler has it. */ +# if !@HAVE__BOOL@ + /* For the sake of symbolic names in gdb, define true and false as + enum constants, not only as macros. + It is tempting to write + typedef enum { false = 0, true = 1 } _Bool; + so that gdb prints values of type 'bool' symbolically. But then + values of type '_Bool' might promote to 'int' or 'unsigned int' + (see ISO C 99 6.7.2.2.(4)); however, '_Bool' must promote to 'int' + (see ISO C 99 6.3.1.1.(2)). So add a negative value to the + enum; this ensures that '_Bool' promotes to 'int'. */ +typedef enum { _Bool_must_promote_to_int = -1, false = 0, true = 1 } _Bool; +# endif +# endif +# endif +# define bool _Bool +#endif + +/* The other macros must be usable in preprocessor directives. */ +#ifdef __cplusplus +# define false false +# define true true +#else +# define false 0 +# define true 1 +#endif + +#define __bool_true_false_are_defined 1 + +#endif /* _GL_STDBOOL_H */ diff --git a/lib/stddef.in.h b/lib/stddef.in.h new file mode 100644 index 0000000..8329a9b --- /dev/null +++ b/lib/stddef.in.h @@ -0,0 +1,114 @@ +/* A substitute for POSIX 2008 <stddef.h>, for platforms that have issues. + + Copyright (C) 2009-2018 Free Software Foundation, Inc. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3, or (at your option) + any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, see <https://www.gnu.org/licenses/>. */ + +/* Written by Eric Blake. */ + +/* + * POSIX 2008 <stddef.h> for platforms that have issues. + * <http://www.opengroup.org/susv3xbd/stddef.h.html> + */ + +#if __GNUC__ >= 3 +@PRAGMA_SYSTEM_HEADER@ +#endif +@PRAGMA_COLUMNS@ + +#if defined __need_wchar_t || defined __need_size_t \ + || defined __need_ptrdiff_t || defined __need_NULL \ + || defined __need_wint_t +/* Special invocation convention inside gcc header files. In + particular, gcc provides a version of <stddef.h> that blindly + redefines NULL even when __need_wint_t was defined, even though + wint_t is not normally provided by <stddef.h>. Hence, we must + remember if special invocation has ever been used to obtain wint_t, + in which case we need to clean up NULL yet again. */ + +# if !(defined _@GUARD_PREFIX@_STDDEF_H && defined _GL_STDDEF_WINT_T) +# ifdef __need_wint_t +# define _GL_STDDEF_WINT_T +# endif +# @INCLUDE_NEXT@ @NEXT_STDDEF_H@ +# endif + +#else +/* Normal invocation convention. */ + +# ifndef _@GUARD_PREFIX@_STDDEF_H + +/* The include_next requires a split double-inclusion guard. */ + +# @INCLUDE_NEXT@ @NEXT_STDDEF_H@ + +/* On NetBSD 5.0, the definition of NULL lacks proper parentheses. */ +# if (@REPLACE_NULL@ \ + && (!defined _@GUARD_PREFIX@_STDDEF_H || defined _GL_STDDEF_WINT_T)) +# undef NULL +# ifdef __cplusplus + /* ISO C++ says that the macro NULL must expand to an integer constant + expression, hence '((void *) 0)' is not allowed in C++. */ +# if __GNUG__ >= 3 + /* GNU C++ has a __null macro that behaves like an integer ('int' or + 'long') but has the same size as a pointer. Use that, to avoid + warnings. */ +# define NULL __null +# else +# define NULL 0L +# endif +# else +# define NULL ((void *) 0) +# endif +# endif + +# ifndef _@GUARD_PREFIX@_STDDEF_H +# define _@GUARD_PREFIX@_STDDEF_H + +/* Some platforms lack wchar_t. */ +#if !@HAVE_WCHAR_T@ +# define wchar_t int +#endif + +/* Some platforms lack max_align_t. The check for _GCC_MAX_ALIGN_T is + a hack in case the configure-time test was done with g++ even though + we are currently compiling with gcc. */ +#if ! (@HAVE_MAX_ALIGN_T@ || defined _GCC_MAX_ALIGN_T) +# if !GNULIB_defined_max_align_t +/* On the x86, the maximum storage alignment of double, long, etc. is 4, + but GCC's C11 ABI for x86 says that max_align_t has an alignment of 8, + and the C11 standard allows this. Work around this problem by + using __alignof__ (which returns 8 for double) rather than _Alignof + (which returns 4), and align each union member accordingly. */ +# ifdef __GNUC__ +# define _GL_STDDEF_ALIGNAS(type) \ + __attribute__ ((__aligned__ (__alignof__ (type)))) +# else +# define _GL_STDDEF_ALIGNAS(type) /* */ +# endif +typedef union +{ + char *__p _GL_STDDEF_ALIGNAS (char *); + double __d _GL_STDDEF_ALIGNAS (double); + long double __ld _GL_STDDEF_ALIGNAS (long double); + long int __i _GL_STDDEF_ALIGNAS (long int); +} rpl_max_align_t; +# define max_align_t rpl_max_align_t +# define GNULIB_defined_max_align_t 1 +# endif +#endif + +# endif /* _@GUARD_PREFIX@_STDDEF_H */ +# endif /* _@GUARD_PREFIX@_STDDEF_H */ +#endif /* __need_XXX */ diff --git a/lib/stdint.in.h b/lib/stdint.in.h new file mode 100644 index 0000000..94e7c81 --- /dev/null +++ b/lib/stdint.in.h @@ -0,0 +1,726 @@ +/* Copyright (C) 2001-2002, 2004-2018 Free Software Foundation, Inc. + Written by Paul Eggert, Bruno Haible, Sam Steingold, Peter Burwood. + This file is part of gnulib. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3, or (at your option) + any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, see <https://www.gnu.org/licenses/>. */ + +/* + * ISO C 99 <stdint.h> for platforms that lack it. + * <http://www.opengroup.org/susv3xbd/stdint.h.html> + */ + +#ifndef _@GUARD_PREFIX@_STDINT_H + +#if __GNUC__ >= 3 +@PRAGMA_SYSTEM_HEADER@ +#endif +@PRAGMA_COLUMNS@ + +/* When including a system file that in turn includes <inttypes.h>, + use the system <inttypes.h>, not our substitute. This avoids + problems with (for example) VMS, whose <sys/bitypes.h> includes + <inttypes.h>. */ +#define _GL_JUST_INCLUDE_SYSTEM_INTTYPES_H + +/* On Android (Bionic libc), <sys/types.h> includes this file before + having defined 'time_t'. Therefore in this case avoid including + other system header files; just include the system's <stdint.h>. + Ideally we should test __BIONIC__ here, but it is only defined after + <sys/cdefs.h> has been included; hence test __ANDROID__ instead. */ +#if defined __ANDROID__ && defined _GL_INCLUDING_SYS_TYPES_H +# @INCLUDE_NEXT@ @NEXT_STDINT_H@ +#else + +/* Get those types that are already defined in other system include + files, so that we can "#define int8_t signed char" below without + worrying about a later system include file containing a "typedef + signed char int8_t;" that will get messed up by our macro. Our + macros should all be consistent with the system versions, except + for the "fast" types and macros, which we recommend against using + in public interfaces due to compiler differences. */ + +#if @HAVE_STDINT_H@ +# if defined __sgi && ! defined __c99 + /* Bypass IRIX's <stdint.h> if in C89 mode, since it merely annoys users + with "This header file is to be used only for c99 mode compilations" + diagnostics. */ +# define __STDINT_H__ +# endif + + /* Some pre-C++11 <stdint.h> implementations need this. */ +# ifdef __cplusplus +# ifndef __STDC_CONSTANT_MACROS +# define __STDC_CONSTANT_MACROS 1 +# endif +# ifndef __STDC_LIMIT_MACROS +# define __STDC_LIMIT_MACROS 1 +# endif +# endif + + /* Other systems may have an incomplete or buggy <stdint.h>. + Include it before <inttypes.h>, since any "#include <stdint.h>" + in <inttypes.h> would reinclude us, skipping our contents because + _@GUARD_PREFIX@_STDINT_H is defined. + The include_next requires a split double-inclusion guard. */ +# @INCLUDE_NEXT@ @NEXT_STDINT_H@ +#endif + +#if ! defined _@GUARD_PREFIX@_STDINT_H && ! defined _GL_JUST_INCLUDE_SYSTEM_STDINT_H +#define _@GUARD_PREFIX@_STDINT_H + +/* Get SCHAR_MIN, SCHAR_MAX, UCHAR_MAX, INT_MIN, INT_MAX, + LONG_MIN, LONG_MAX, ULONG_MAX, _GL_INTEGER_WIDTH. */ +#include <limits.h> + +/* Override WINT_MIN and WINT_MAX if gnulib's <wchar.h> or <wctype.h> overrides + wint_t. */ +#if @GNULIB_OVERRIDES_WINT_T@ +# undef WINT_MIN +# undef WINT_MAX +# define WINT_MIN 0x0U +# define WINT_MAX 0xffffffffU +#endif + +#if ! @HAVE_C99_STDINT_H@ + +/* <sys/types.h> defines some of the stdint.h types as well, on glibc, + IRIX 6.5, and OpenBSD 3.8 (via <machine/types.h>). + AIX 5.2 <sys/types.h> isn't needed and causes troubles. + Mac OS X 10.4.6 <sys/types.h> includes <stdint.h> (which is us), but + relies on the system <stdint.h> definitions, so include + <sys/types.h> after @NEXT_STDINT_H@. */ +# if @HAVE_SYS_TYPES_H@ && ! defined _AIX +# include <sys/types.h> +# endif + +# if @HAVE_INTTYPES_H@ + /* In OpenBSD 3.8, <inttypes.h> includes <machine/types.h>, which defines + int{8,16,32,64}_t, uint{8,16,32,64}_t and __BIT_TYPES_DEFINED__. + <inttypes.h> also defines intptr_t and uintptr_t. */ +# include <inttypes.h> +# elif @HAVE_SYS_INTTYPES_H@ + /* Solaris 7 <sys/inttypes.h> has the types except the *_fast*_t types, and + the macros except for *_FAST*_*, INTPTR_MIN, PTRDIFF_MIN, PTRDIFF_MAX. */ +# include <sys/inttypes.h> +# endif + +# if @HAVE_SYS_BITYPES_H@ && ! defined __BIT_TYPES_DEFINED__ + /* Linux libc4 >= 4.6.7 and libc5 have a <sys/bitypes.h> that defines + int{8,16,32,64}_t and __BIT_TYPES_DEFINED__. In libc5 >= 5.2.2 it is + included by <sys/types.h>. */ +# include <sys/bitypes.h> +# endif + +# undef _GL_JUST_INCLUDE_SYSTEM_INTTYPES_H + +/* Minimum and maximum values for an integer type under the usual assumption. + Return an unspecified value if BITS == 0, adding a check to pacify + picky compilers. */ + +/* These are separate macros, because if you try to merge these macros into + a single one, HP-UX cc rejects the resulting expression in constant + expressions. */ +# define _STDINT_UNSIGNED_MIN(bits, zero) \ + (zero) +# define _STDINT_SIGNED_MIN(bits, zero) \ + (~ _STDINT_MAX (1, bits, zero)) + +# define _STDINT_MAX(signed, bits, zero) \ + (((((zero) + 1) << ((bits) ? (bits) - 1 - (signed) : 0)) - 1) * 2 + 1) + +#if !GNULIB_defined_stdint_types + +/* 7.18.1.1. Exact-width integer types */ + +/* Here we assume a standard architecture where the hardware integer + types have 8, 16, 32, optionally 64 bits. */ + +# undef int8_t +# undef uint8_t +typedef signed char gl_int8_t; +typedef unsigned char gl_uint8_t; +# define int8_t gl_int8_t +# define uint8_t gl_uint8_t + +# undef int16_t +# undef uint16_t +typedef short int gl_int16_t; +typedef unsigned short int gl_uint16_t; +# define int16_t gl_int16_t +# define uint16_t gl_uint16_t + +# undef int32_t +# undef uint32_t +typedef int gl_int32_t; +typedef unsigned int gl_uint32_t; +# define int32_t gl_int32_t +# define uint32_t gl_uint32_t + +/* If the system defines INT64_MAX, assume int64_t works. That way, + if the underlying platform defines int64_t to be a 64-bit long long + int, the code below won't mistakenly define it to be a 64-bit long + int, which would mess up C++ name mangling. We must use #ifdef + rather than #if, to avoid an error with HP-UX 10.20 cc. */ + +# ifdef INT64_MAX +# define GL_INT64_T +# else +/* Do not undefine int64_t if gnulib is not being used with 64-bit + types, since otherwise it breaks platforms like Tandem/NSK. */ +# if LONG_MAX >> 31 >> 31 == 1 +# undef int64_t +typedef long int gl_int64_t; +# define int64_t gl_int64_t +# define GL_INT64_T +# elif defined _MSC_VER +# undef int64_t +typedef __int64 gl_int64_t; +# define int64_t gl_int64_t +# define GL_INT64_T +# elif @HAVE_LONG_LONG_INT@ +# undef int64_t +typedef long long int gl_int64_t; +# define int64_t gl_int64_t +# define GL_INT64_T +# endif +# endif + +# ifdef UINT64_MAX +# define GL_UINT64_T +# else +# if ULONG_MAX >> 31 >> 31 >> 1 == 1 +# undef uint64_t +typedef unsigned long int gl_uint64_t; +# define uint64_t gl_uint64_t +# define GL_UINT64_T +# elif defined _MSC_VER +# undef uint64_t +typedef unsigned __int64 gl_uint64_t; +# define uint64_t gl_uint64_t +# define GL_UINT64_T +# elif @HAVE_UNSIGNED_LONG_LONG_INT@ +# undef uint64_t +typedef unsigned long long int gl_uint64_t; +# define uint64_t gl_uint64_t +# define GL_UINT64_T +# endif +# endif + +/* Avoid collision with Solaris 2.5.1 <pthread.h> etc. */ +# define _UINT8_T +# define _UINT32_T +# define _UINT64_T + + +/* 7.18.1.2. Minimum-width integer types */ + +/* Here we assume a standard architecture where the hardware integer + types have 8, 16, 32, optionally 64 bits. Therefore the leastN_t types + are the same as the corresponding N_t types. */ + +# undef int_least8_t +# undef uint_least8_t +# undef int_least16_t +# undef uint_least16_t +# undef int_least32_t +# undef uint_least32_t +# undef int_least64_t +# undef uint_least64_t +# define int_least8_t int8_t +# define uint_least8_t uint8_t +# define int_least16_t int16_t +# define uint_least16_t uint16_t +# define int_least32_t int32_t +# define uint_least32_t uint32_t +# ifdef GL_INT64_T +# define int_least64_t int64_t +# endif +# ifdef GL_UINT64_T +# define uint_least64_t uint64_t +# endif + +/* 7.18.1.3. Fastest minimum-width integer types */ + +/* Note: Other <stdint.h> substitutes may define these types differently. + It is not recommended to use these types in public header files. */ + +/* Here we assume a standard architecture where the hardware integer + types have 8, 16, 32, optionally 64 bits. Therefore the fastN_t types + are taken from the same list of types. The following code normally + uses types consistent with glibc, as that lessens the chance of + incompatibility with older GNU hosts. */ + +# undef int_fast8_t +# undef uint_fast8_t +# undef int_fast16_t +# undef uint_fast16_t +# undef int_fast32_t +# undef uint_fast32_t +# undef int_fast64_t +# undef uint_fast64_t +typedef signed char gl_int_fast8_t; +typedef unsigned char gl_uint_fast8_t; + +# ifdef __sun +/* Define types compatible with SunOS 5.10, so that code compiled under + earlier SunOS versions works with code compiled under SunOS 5.10. */ +typedef int gl_int_fast32_t; +typedef unsigned int gl_uint_fast32_t; +# else +typedef long int gl_int_fast32_t; +typedef unsigned long int gl_uint_fast32_t; +# endif +typedef gl_int_fast32_t gl_int_fast16_t; +typedef gl_uint_fast32_t gl_uint_fast16_t; + +# define int_fast8_t gl_int_fast8_t +# define uint_fast8_t gl_uint_fast8_t +# define int_fast16_t gl_int_fast16_t +# define uint_fast16_t gl_uint_fast16_t +# define int_fast32_t gl_int_fast32_t +# define uint_fast32_t gl_uint_fast32_t +# ifdef GL_INT64_T +# define int_fast64_t int64_t +# endif +# ifdef GL_UINT64_T +# define uint_fast64_t uint64_t +# endif + +/* 7.18.1.4. Integer types capable of holding object pointers */ + +/* kLIBC's stdint.h defines _INTPTR_T_DECLARED and needs its own + definitions of intptr_t and uintptr_t (which use int and unsigned) + to avoid clashes with declarations of system functions like sbrk. */ +# ifndef _INTPTR_T_DECLARED +# undef intptr_t +# undef uintptr_t +typedef long int gl_intptr_t; +typedef unsigned long int gl_uintptr_t; +# define intptr_t gl_intptr_t +# define uintptr_t gl_uintptr_t +# endif + +/* 7.18.1.5. Greatest-width integer types */ + +/* Note: These types are compiler dependent. It may be unwise to use them in + public header files. */ + +/* If the system defines INTMAX_MAX, assume that intmax_t works, and + similarly for UINTMAX_MAX and uintmax_t. This avoids problems with + assuming one type where another is used by the system. */ + +# ifndef INTMAX_MAX +# undef INTMAX_C +# undef intmax_t +# if @HAVE_LONG_LONG_INT@ && LONG_MAX >> 30 == 1 +typedef long long int gl_intmax_t; +# define intmax_t gl_intmax_t +# elif defined GL_INT64_T +# define intmax_t int64_t +# else +typedef long int gl_intmax_t; +# define intmax_t gl_intmax_t +# endif +# endif + +# ifndef UINTMAX_MAX +# undef UINTMAX_C +# undef uintmax_t +# if @HAVE_UNSIGNED_LONG_LONG_INT@ && ULONG_MAX >> 31 == 1 +typedef unsigned long long int gl_uintmax_t; +# define uintmax_t gl_uintmax_t +# elif defined GL_UINT64_T +# define uintmax_t uint64_t +# else +typedef unsigned long int gl_uintmax_t; +# define uintmax_t gl_uintmax_t +# endif +# endif + +/* Verify that intmax_t and uintmax_t have the same size. Too much code + breaks if this is not the case. If this check fails, the reason is likely + to be found in the autoconf macros. */ +typedef int _verify_intmax_size[sizeof (intmax_t) == sizeof (uintmax_t) + ? 1 : -1]; + +# define GNULIB_defined_stdint_types 1 +# endif /* !GNULIB_defined_stdint_types */ + +/* 7.18.2. Limits of specified-width integer types */ + +/* 7.18.2.1. Limits of exact-width integer types */ + +/* Here we assume a standard architecture where the hardware integer + types have 8, 16, 32, optionally 64 bits. */ + +# undef INT8_MIN +# undef INT8_MAX +# undef UINT8_MAX +# define INT8_MIN (~ INT8_MAX) +# define INT8_MAX 127 +# define UINT8_MAX 255 + +# undef INT16_MIN +# undef INT16_MAX +# undef UINT16_MAX +# define INT16_MIN (~ INT16_MAX) +# define INT16_MAX 32767 +# define UINT16_MAX 65535 + +# undef INT32_MIN +# undef INT32_MAX +# undef UINT32_MAX +# define INT32_MIN (~ INT32_MAX) +# define INT32_MAX 2147483647 +# define UINT32_MAX 4294967295U + +# if defined GL_INT64_T && ! defined INT64_MAX +/* Prefer (- INTMAX_C (1) << 63) over (~ INT64_MAX) because SunPRO C 5.0 + evaluates the latter incorrectly in preprocessor expressions. */ +# define INT64_MIN (- INTMAX_C (1) << 63) +# define INT64_MAX INTMAX_C (9223372036854775807) +# endif + +# if defined GL_UINT64_T && ! defined UINT64_MAX +# define UINT64_MAX UINTMAX_C (18446744073709551615) +# endif + +/* 7.18.2.2. Limits of minimum-width integer types */ + +/* Here we assume a standard architecture where the hardware integer + types have 8, 16, 32, optionally 64 bits. Therefore the leastN_t types + are the same as the corresponding N_t types. */ + +# undef INT_LEAST8_MIN +# undef INT_LEAST8_MAX +# undef UINT_LEAST8_MAX +# define INT_LEAST8_MIN INT8_MIN +# define INT_LEAST8_MAX INT8_MAX +# define UINT_LEAST8_MAX UINT8_MAX + +# undef INT_LEAST16_MIN +# undef INT_LEAST16_MAX +# undef UINT_LEAST16_MAX +# define INT_LEAST16_MIN INT16_MIN +# define INT_LEAST16_MAX INT16_MAX +# define UINT_LEAST16_MAX UINT16_MAX + +# undef INT_LEAST32_MIN +# undef INT_LEAST32_MAX +# undef UINT_LEAST32_MAX +# define INT_LEAST32_MIN INT32_MIN +# define INT_LEAST32_MAX INT32_MAX +# define UINT_LEAST32_MAX UINT32_MAX + +# undef INT_LEAST64_MIN +# undef INT_LEAST64_MAX +# ifdef GL_INT64_T +# define INT_LEAST64_MIN INT64_MIN +# define INT_LEAST64_MAX INT64_MAX +# endif + +# undef UINT_LEAST64_MAX +# ifdef GL_UINT64_T +# define UINT_LEAST64_MAX UINT64_MAX +# endif + +/* 7.18.2.3. Limits of fastest minimum-width integer types */ + +/* Here we assume a standard architecture where the hardware integer + types have 8, 16, 32, optionally 64 bits. Therefore the fastN_t types + are taken from the same list of types. */ + +# undef INT_FAST8_MIN +# undef INT_FAST8_MAX +# undef UINT_FAST8_MAX +# define INT_FAST8_MIN SCHAR_MIN +# define INT_FAST8_MAX SCHAR_MAX +# define UINT_FAST8_MAX UCHAR_MAX + +# undef INT_FAST16_MIN +# undef INT_FAST16_MAX +# undef UINT_FAST16_MAX +# define INT_FAST16_MIN INT_FAST32_MIN +# define INT_FAST16_MAX INT_FAST32_MAX +# define UINT_FAST16_MAX UINT_FAST32_MAX + +# undef INT_FAST32_MIN +# undef INT_FAST32_MAX +# undef UINT_FAST32_MAX +# ifdef __sun +# define INT_FAST32_MIN INT_MIN +# define INT_FAST32_MAX INT_MAX +# define UINT_FAST32_MAX UINT_MAX +# else +# define INT_FAST32_MIN LONG_MIN +# define INT_FAST32_MAX LONG_MAX +# define UINT_FAST32_MAX ULONG_MAX +# endif + +# undef INT_FAST64_MIN +# undef INT_FAST64_MAX +# ifdef GL_INT64_T +# define INT_FAST64_MIN INT64_MIN +# define INT_FAST64_MAX INT64_MAX +# endif + +# undef UINT_FAST64_MAX +# ifdef GL_UINT64_T +# define UINT_FAST64_MAX UINT64_MAX +# endif + +/* 7.18.2.4. Limits of integer types capable of holding object pointers */ + +# undef INTPTR_MIN +# undef INTPTR_MAX +# undef UINTPTR_MAX +# define INTPTR_MIN LONG_MIN +# define INTPTR_MAX LONG_MAX +# define UINTPTR_MAX ULONG_MAX + +/* 7.18.2.5. Limits of greatest-width integer types */ + +# ifndef INTMAX_MAX +# undef INTMAX_MIN +# ifdef INT64_MAX +# define INTMAX_MIN INT64_MIN +# define INTMAX_MAX INT64_MAX +# else +# define INTMAX_MIN INT32_MIN +# define INTMAX_MAX INT32_MAX +# endif +# endif + +# ifndef UINTMAX_MAX +# ifdef UINT64_MAX +# define UINTMAX_MAX UINT64_MAX +# else +# define UINTMAX_MAX UINT32_MAX +# endif +# endif + +/* 7.18.3. Limits of other integer types */ + +/* ptrdiff_t limits */ +# undef PTRDIFF_MIN +# undef PTRDIFF_MAX +# if @APPLE_UNIVERSAL_BUILD@ +# ifdef _LP64 +# define PTRDIFF_MIN _STDINT_SIGNED_MIN (64, 0l) +# define PTRDIFF_MAX _STDINT_MAX (1, 64, 0l) +# else +# define PTRDIFF_MIN _STDINT_SIGNED_MIN (32, 0) +# define PTRDIFF_MAX _STDINT_MAX (1, 32, 0) +# endif +# else +# define PTRDIFF_MIN \ + _STDINT_SIGNED_MIN (@BITSIZEOF_PTRDIFF_T@, 0@PTRDIFF_T_SUFFIX@) +# define PTRDIFF_MAX \ + _STDINT_MAX (1, @BITSIZEOF_PTRDIFF_T@, 0@PTRDIFF_T_SUFFIX@) +# endif + +/* sig_atomic_t limits */ +# undef SIG_ATOMIC_MIN +# undef SIG_ATOMIC_MAX +# if @HAVE_SIGNED_SIG_ATOMIC_T@ +# define SIG_ATOMIC_MIN \ + _STDINT_SIGNED_MIN (@BITSIZEOF_SIG_ATOMIC_T@, 0@SIG_ATOMIC_T_SUFFIX@) +# else +# define SIG_ATOMIC_MIN \ + _STDINT_UNSIGNED_MIN (@BITSIZEOF_SIG_ATOMIC_T@, 0@SIG_ATOMIC_T_SUFFIX@) +# endif +# define SIG_ATOMIC_MAX \ + _STDINT_MAX (@HAVE_SIGNED_SIG_ATOMIC_T@, @BITSIZEOF_SIG_ATOMIC_T@, \ + 0@SIG_ATOMIC_T_SUFFIX@) + + +/* size_t limit */ +# undef SIZE_MAX +# if @APPLE_UNIVERSAL_BUILD@ +# ifdef _LP64 +# define SIZE_MAX _STDINT_MAX (0, 64, 0ul) +# else +# define SIZE_MAX _STDINT_MAX (0, 32, 0ul) +# endif +# else +# define SIZE_MAX _STDINT_MAX (0, @BITSIZEOF_SIZE_T@, 0@SIZE_T_SUFFIX@) +# endif + +/* wchar_t limits */ +/* Get WCHAR_MIN, WCHAR_MAX. + This include is not on the top, above, because on OSF/1 4.0 we have a + sequence of nested includes + <wchar.h> -> <stdio.h> -> <getopt.h> -> <stdlib.h>, and the latter includes + <stdint.h> and assumes its types are already defined. */ +# if @HAVE_WCHAR_H@ && ! (defined WCHAR_MIN && defined WCHAR_MAX) + /* BSD/OS 4.0.1 has a bug: <stddef.h>, <stdio.h> and <time.h> must be + included before <wchar.h>. */ +# include <stddef.h> +# include <stdio.h> +# include <time.h> +# define _GL_JUST_INCLUDE_SYSTEM_WCHAR_H +# include <wchar.h> +# undef _GL_JUST_INCLUDE_SYSTEM_WCHAR_H +# endif +# undef WCHAR_MIN +# undef WCHAR_MAX +# if @HAVE_SIGNED_WCHAR_T@ +# define WCHAR_MIN \ + _STDINT_SIGNED_MIN (@BITSIZEOF_WCHAR_T@, 0@WCHAR_T_SUFFIX@) +# else +# define WCHAR_MIN \ + _STDINT_UNSIGNED_MIN (@BITSIZEOF_WCHAR_T@, 0@WCHAR_T_SUFFIX@) +# endif +# define WCHAR_MAX \ + _STDINT_MAX (@HAVE_SIGNED_WCHAR_T@, @BITSIZEOF_WCHAR_T@, 0@WCHAR_T_SUFFIX@) + +/* wint_t limits */ +# undef WINT_MIN +# undef WINT_MAX +# if @HAVE_SIGNED_WINT_T@ +# define WINT_MIN \ + _STDINT_SIGNED_MIN (@BITSIZEOF_WINT_T@, 0@WINT_T_SUFFIX@) +# else +# define WINT_MIN \ + _STDINT_UNSIGNED_MIN (@BITSIZEOF_WINT_T@, 0@WINT_T_SUFFIX@) +# endif +# define WINT_MAX \ + _STDINT_MAX (@HAVE_SIGNED_WINT_T@, @BITSIZEOF_WINT_T@, 0@WINT_T_SUFFIX@) + +/* 7.18.4. Macros for integer constants */ + +/* 7.18.4.1. Macros for minimum-width integer constants */ +/* According to ISO C 99 Technical Corrigendum 1 */ + +/* Here we assume a standard architecture where the hardware integer + types have 8, 16, 32, optionally 64 bits, and int is 32 bits. */ + +# undef INT8_C +# undef UINT8_C +# define INT8_C(x) x +# define UINT8_C(x) x + +# undef INT16_C +# undef UINT16_C +# define INT16_C(x) x +# define UINT16_C(x) x + +# undef INT32_C +# undef UINT32_C +# define INT32_C(x) x +# define UINT32_C(x) x ## U + +# undef INT64_C +# undef UINT64_C +# if LONG_MAX >> 31 >> 31 == 1 +# define INT64_C(x) x##L +# elif defined _MSC_VER +# define INT64_C(x) x##i64 +# elif @HAVE_LONG_LONG_INT@ +# define INT64_C(x) x##LL +# endif +# if ULONG_MAX >> 31 >> 31 >> 1 == 1 +# define UINT64_C(x) x##UL +# elif defined _MSC_VER +# define UINT64_C(x) x##ui64 +# elif @HAVE_UNSIGNED_LONG_LONG_INT@ +# define UINT64_C(x) x##ULL +# endif + +/* 7.18.4.2. Macros for greatest-width integer constants */ + +# ifndef INTMAX_C +# if @HAVE_LONG_LONG_INT@ && LONG_MAX >> 30 == 1 +# define INTMAX_C(x) x##LL +# elif defined GL_INT64_T +# define INTMAX_C(x) INT64_C(x) +# else +# define INTMAX_C(x) x##L +# endif +# endif + +# ifndef UINTMAX_C +# if @HAVE_UNSIGNED_LONG_LONG_INT@ && ULONG_MAX >> 31 == 1 +# define UINTMAX_C(x) x##ULL +# elif defined GL_UINT64_T +# define UINTMAX_C(x) UINT64_C(x) +# else +# define UINTMAX_C(x) x##UL +# endif +# endif + +#endif /* !@HAVE_C99_STDINT_H@ */ + +/* Macros specified by ISO/IEC TS 18661-1:2014. */ + +#if (!defined UINTMAX_WIDTH \ + && (defined _GNU_SOURCE || defined __STDC_WANT_IEC_60559_BFP_EXT__)) +# ifdef INT8_MAX +# define INT8_WIDTH _GL_INTEGER_WIDTH (INT8_MIN, INT8_MAX) +# endif +# ifdef UINT8_MAX +# define UINT8_WIDTH _GL_INTEGER_WIDTH (0, UINT8_MAX) +# endif +# ifdef INT16_MAX +# define INT16_WIDTH _GL_INTEGER_WIDTH (INT16_MIN, INT16_MAX) +# endif +# ifdef UINT16_MAX +# define UINT16_WIDTH _GL_INTEGER_WIDTH (0, UINT16_MAX) +# endif +# ifdef INT32_MAX +# define INT32_WIDTH _GL_INTEGER_WIDTH (INT32_MIN, INT32_MAX) +# endif +# ifdef UINT32_MAX +# define UINT32_WIDTH _GL_INTEGER_WIDTH (0, UINT32_MAX) +# endif +# ifdef INT64_MAX +# define INT64_WIDTH _GL_INTEGER_WIDTH (INT64_MIN, INT64_MAX) +# endif +# ifdef UINT64_MAX +# define UINT64_WIDTH _GL_INTEGER_WIDTH (0, UINT64_MAX) +# endif +# define INT_LEAST8_WIDTH _GL_INTEGER_WIDTH (INT_LEAST8_MIN, INT_LEAST8_MAX) +# define UINT_LEAST8_WIDTH _GL_INTEGER_WIDTH (0, UINT_LEAST8_MAX) +# define INT_LEAST16_WIDTH _GL_INTEGER_WIDTH (INT_LEAST16_MIN, INT_LEAST16_MAX) +# define UINT_LEAST16_WIDTH _GL_INTEGER_WIDTH (0, UINT_LEAST16_MAX) +# define INT_LEAST32_WIDTH _GL_INTEGER_WIDTH (INT_LEAST32_MIN, INT_LEAST32_MAX) +# define UINT_LEAST32_WIDTH _GL_INTEGER_WIDTH (0, UINT_LEAST32_MAX) +# define INT_LEAST64_WIDTH _GL_INTEGER_WIDTH (INT_LEAST64_MIN, INT_LEAST64_MAX) +# define UINT_LEAST64_WIDTH _GL_INTEGER_WIDTH (0, UINT_LEAST64_MAX) +# define INT_FAST8_WIDTH _GL_INTEGER_WIDTH (INT_FAST8_MIN, INT_FAST8_MAX) +# define UINT_FAST8_WIDTH _GL_INTEGER_WIDTH (0, UINT_FAST8_MAX) +# define INT_FAST16_WIDTH _GL_INTEGER_WIDTH (INT_FAST16_MIN, INT_FAST16_MAX) +# define UINT_FAST16_WIDTH _GL_INTEGER_WIDTH (0, UINT_FAST16_MAX) +# define INT_FAST32_WIDTH _GL_INTEGER_WIDTH (INT_FAST32_MIN, INT_FAST32_MAX) +# define UINT_FAST32_WIDTH _GL_INTEGER_WIDTH (0, UINT_FAST32_MAX) +# define INT_FAST64_WIDTH _GL_INTEGER_WIDTH (INT_FAST64_MIN, INT_FAST64_MAX) +# define UINT_FAST64_WIDTH _GL_INTEGER_WIDTH (0, UINT_FAST64_MAX) +# define INTPTR_WIDTH _GL_INTEGER_WIDTH (INTPTR_MIN, INTPTR_MAX) +# define UINTPTR_WIDTH _GL_INTEGER_WIDTH (0, UINTPTR_MAX) +# define INTMAX_WIDTH _GL_INTEGER_WIDTH (INTMAX_MIN, INTMAX_MAX) +# define UINTMAX_WIDTH _GL_INTEGER_WIDTH (0, UINTMAX_MAX) +# define PTRDIFF_WIDTH _GL_INTEGER_WIDTH (PTRDIFF_MIN, PTRDIFF_MAX) +# define SIZE_WIDTH _GL_INTEGER_WIDTH (0, SIZE_MAX) +# define WCHAR_WIDTH _GL_INTEGER_WIDTH (WCHAR_MIN, WCHAR_MAX) +# ifdef WINT_MAX +# define WINT_WIDTH _GL_INTEGER_WIDTH (WINT_MIN, WINT_MAX) +# endif +# ifdef SIG_ATOMIC_MAX +# define SIG_ATOMIC_WIDTH _GL_INTEGER_WIDTH (SIG_ATOMIC_MIN, SIG_ATOMIC_MAX) +# endif +#endif /* !WINT_WIDTH && (_GNU_SOURCE || __STDC_WANT_IEC_60559_BFP_EXT__) */ + +#endif /* _@GUARD_PREFIX@_STDINT_H */ +#endif /* !(defined __ANDROID__ && ...) */ +#endif /* !defined _@GUARD_PREFIX@_STDINT_H && !defined _GL_JUST_INCLUDE_SYSTEM_STDINT_H */ diff --git a/lib/stdio-impl.h b/lib/stdio-impl.h new file mode 100644 index 0000000..393ef0c --- /dev/null +++ b/lib/stdio-impl.h @@ -0,0 +1,202 @@ +/* Implementation details of FILE streams. + Copyright (C) 2007-2008, 2010-2018 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +/* Many stdio implementations have the same logic and therefore can share + the same implementation of stdio extension API, except that some fields + have different naming conventions, or their access requires some casts. */ + +/* Glibc 2.28 made _IO_IN_BACKUP private. For now, work around this + problem by defining it ourselves. FIXME: Do not rely on glibc + internals. */ +#if !defined _IO_IN_BACKUP && defined _IO_EOF_SEEN +# define _IO_IN_BACKUP 0x100 +#endif + +/* BSD stdio derived implementations. */ + +#if defined __NetBSD__ /* NetBSD */ +/* Get __NetBSD_Version__. */ +# include <sys/param.h> +#endif + +#include <errno.h> /* For detecting Plan9. */ + +#if defined __sferror || defined __DragonFly__ || defined __ANDROID__ + /* FreeBSD, NetBSD, OpenBSD, DragonFly, Mac OS X, Cygwin, Minix 3, Android */ + +# if defined __DragonFly__ /* DragonFly */ + /* See <https://gitweb.dragonflybsd.org/dragonfly.git/blob_plain/HEAD:/lib/libc/stdio/priv_stdio.h>. */ +# define fp_ ((struct { struct __FILE_public pub; \ + struct { unsigned char *_base; int _size; } _bf; \ + void *cookie; \ + void *_close; \ + void *_read; \ + void *_seek; \ + void *_write; \ + struct { unsigned char *_base; int _size; } _ub; \ + int _ur; \ + unsigned char _ubuf[3]; \ + unsigned char _nbuf[1]; \ + struct { unsigned char *_base; int _size; } _lb; \ + int _blksize; \ + fpos_t _offset; \ + /* More fields, not relevant here. */ \ + } *) fp) + /* See <https://gitweb.dragonflybsd.org/dragonfly.git/blob_plain/HEAD:/include/stdio.h>. */ +# define _p pub._p +# define _flags pub._flags +# define _r pub._r +# define _w pub._w +# elif defined __ANDROID__ /* Android */ + /* Up to this commit from 2015-10-12 + <https://android.googlesource.com/platform/bionic.git/+/f0141dfab10a4b332769d52fa76631a64741297a> + the innards of FILE were public, and fp_ub could be defined like for OpenBSD, + see <https://android.googlesource.com/platform/bionic.git/+/e78392637d5086384a5631ddfdfa8d7ec8326ee3/libc/stdio/fileext.h> + and <https://android.googlesource.com/platform/bionic.git/+/e78392637d5086384a5631ddfdfa8d7ec8326ee3/libc/stdio/local.h>. + After this commit, the innards of FILE are hidden. */ +# define fp_ ((struct { unsigned char *_p; \ + int _r; \ + int _w; \ + int _flags; \ + int _file; \ + struct { unsigned char *_base; size_t _size; } _bf; \ + int _lbfsize; \ + void *_cookie; \ + void *_close; \ + void *_read; \ + void *_seek; \ + void *_write; \ + struct { unsigned char *_base; size_t _size; } _ext; \ + unsigned char *_up; \ + int _ur; \ + unsigned char _ubuf[3]; \ + unsigned char _nbuf[1]; \ + struct { unsigned char *_base; size_t _size; } _lb; \ + int _blksize; \ + fpos_t _offset; \ + /* More fields, not relevant here. */ \ + } *) fp) +# else +# define fp_ fp +# endif + +# if (defined __NetBSD__ && __NetBSD_Version__ >= 105270000) || defined __OpenBSD__ || defined __minix /* NetBSD >= 1.5ZA, OpenBSD, Minix 3 */ + /* See <http://cvsweb.netbsd.org/bsdweb.cgi/src/lib/libc/stdio/fileext.h?rev=HEAD&content-type=text/x-cvsweb-markup> + and <https://cvsweb.openbsd.org/cgi-bin/cvsweb/src/lib/libc/stdio/fileext.h?rev=HEAD&content-type=text/x-cvsweb-markup> + and <https://github.com/Stichting-MINIX-Research-Foundation/minix/blob/master/lib/libc/stdio/fileext.h> */ + struct __sfileext + { + struct __sbuf _ub; /* ungetc buffer */ + /* More fields, not relevant here. */ + }; +# define fp_ub ((struct __sfileext *) fp->_ext._base)->_ub +# elif defined __ANDROID__ /* Android */ + struct __sfileext + { + struct { unsigned char *_base; size_t _size; } _ub; /* ungetc buffer */ + /* More fields, not relevant here. */ + }; +# define fp_ub ((struct __sfileext *) fp_->_ext._base)->_ub +# else /* FreeBSD, NetBSD <= 1.5Z, DragonFly, Mac OS X, Cygwin */ +# define fp_ub fp_->_ub +# endif + +# define HASUB(fp) (fp_ub._base != NULL) + +# if defined __ANDROID__ /* Android */ + /* Needed after this commit from 2016-01-25 + <https://android.googlesource.com/platform/bionic.git/+/e70e0e9267d069bf56a5078c99307e08a7280de7> */ +# ifndef __SEOF +# define __SLBF 1 +# define __SNBF 2 +# define __SRD 4 +# define __SWR 8 +# define __SRW 0x10 +# define __SEOF 0x20 +# define __SERR 0x40 +# endif +# ifndef __SOFF +# define __SOFF 0x1000 +# endif +# endif + +#endif + + +/* SystemV derived implementations. */ + +#ifdef __TANDEM /* NonStop Kernel */ +# ifndef _IOERR +/* These values were determined by the program 'stdioext-flags' at + <https://lists.gnu.org/r/bug-gnulib/2010-12/msg00165.html>. */ +# define _IOERR 0x40 +# define _IOREAD 0x80 +# define _IOWRT 0x4 +# define _IORW 0x100 +# endif +#endif + +#if defined _IOERR + +# if defined __sun && defined _LP64 /* Solaris/{SPARC,AMD64} 64-bit */ +# define fp_ ((struct { unsigned char *_ptr; \ + unsigned char *_base; \ + unsigned char *_end; \ + long _cnt; \ + int _file; \ + unsigned int _flag; \ + } *) fp) +# elif defined __VMS /* OpenVMS */ +# define fp_ ((struct _iobuf *) fp) +# else +# define fp_ fp +# endif + +# if defined _SCO_DS /* OpenServer */ +# define _cnt __cnt +# define _ptr __ptr +# define _base __base +# define _flag __flag +# endif + +#elif defined _WIN32 && ! defined __CYGWIN__ /* newer Windows with MSVC */ + +/* <stdio.h> does not define the innards of FILE any more. */ +# define WINDOWS_OPAQUE_FILE + +struct _gl_real_FILE +{ + /* Note: Compared to older Windows and to mingw, it has the fields + _base and _cnt swapped. */ + unsigned char *_ptr; + unsigned char *_base; + int _cnt; + int _flag; + int _file; + int _charbuf; + int _bufsiz; +}; +# define fp_ ((struct _gl_real_FILE *) fp) + +/* These values were determined by a program similar to the one at + <https://lists.gnu.org/r/bug-gnulib/2010-12/msg00165.html>. */ +# define _IOREAD 0x1 +# define _IOWRT 0x2 +# define _IORW 0x4 +# define _IOEOF 0x8 +# define _IOERR 0x10 + +#endif diff --git a/lib/stdio.in.h b/lib/stdio.in.h new file mode 100644 index 0000000..ff7c9c8 --- /dev/null +++ b/lib/stdio.in.h @@ -0,0 +1,1377 @@ +/* A GNU-like <stdio.h>. + + Copyright (C) 2004, 2007-2018 Free Software Foundation, Inc. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3, or (at your option) + any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, see <https://www.gnu.org/licenses/>. */ + +#if __GNUC__ >= 3 +@PRAGMA_SYSTEM_HEADER@ +#endif +@PRAGMA_COLUMNS@ + +#if defined __need_FILE || defined __need___FILE || defined _GL_ALREADY_INCLUDING_STDIO_H +/* Special invocation convention: + - Inside glibc header files. + - On OSF/1 5.1 we have a sequence of nested includes + <stdio.h> -> <getopt.h> -> <ctype.h> -> <sys/localedef.h> -> + <sys/lc_core.h> -> <nl_types.h> -> <mesg.h> -> <stdio.h>. + In this situation, the functions are not yet declared, therefore we cannot + provide the C++ aliases. */ + +#@INCLUDE_NEXT@ @NEXT_STDIO_H@ + +#else +/* Normal invocation convention. */ + +#ifndef _@GUARD_PREFIX@_STDIO_H + +#define _GL_ALREADY_INCLUDING_STDIO_H + +/* The include_next requires a split double-inclusion guard. */ +#@INCLUDE_NEXT@ @NEXT_STDIO_H@ + +#undef _GL_ALREADY_INCLUDING_STDIO_H + +#ifndef _@GUARD_PREFIX@_STDIO_H +#define _@GUARD_PREFIX@_STDIO_H + +/* Get va_list. Needed on many systems, including glibc 2.8. */ +#include <stdarg.h> + +#include <stddef.h> + +/* Get off_t and ssize_t. Needed on many systems, including glibc 2.8 + and eglibc 2.11.2. + May also define off_t to a 64-bit type on native Windows. */ +#include <sys/types.h> + +/* The __attribute__ feature is available in gcc versions 2.5 and later. + The __-protected variants of the attributes 'format' and 'printf' are + accepted by gcc versions 2.6.4 (effectively 2.7) and later. + We enable _GL_ATTRIBUTE_FORMAT only if these are supported too, because + gnulib and libintl do '#define printf __printf__' when they override + the 'printf' function. */ +#if __GNUC__ > 2 || (__GNUC__ == 2 && __GNUC_MINOR__ >= 7) +# define _GL_ATTRIBUTE_FORMAT(spec) __attribute__ ((__format__ spec)) +#else +# define _GL_ATTRIBUTE_FORMAT(spec) /* empty */ +#endif + +/* _GL_ATTRIBUTE_FORMAT_PRINTF + indicates to GCC that the function takes a format string and arguments, + where the format string directives are the ones standardized by ISO C99 + and POSIX. */ +#if __GNUC__ > 4 || (__GNUC__ == 4 && __GNUC_MINOR__ >= 4) +# define _GL_ATTRIBUTE_FORMAT_PRINTF(formatstring_parameter, first_argument) \ + _GL_ATTRIBUTE_FORMAT ((__gnu_printf__, formatstring_parameter, first_argument)) +#else +# define _GL_ATTRIBUTE_FORMAT_PRINTF(formatstring_parameter, first_argument) \ + _GL_ATTRIBUTE_FORMAT ((__printf__, formatstring_parameter, first_argument)) +#endif + +/* _GL_ATTRIBUTE_FORMAT_PRINTF_SYSTEM is like _GL_ATTRIBUTE_FORMAT_PRINTF, + except that it indicates to GCC that the supported format string directives + are the ones of the system printf(), rather than the ones standardized by + ISO C99 and POSIX. */ +#if GNULIB_PRINTF_ATTRIBUTE_FLAVOR_GNU +# define _GL_ATTRIBUTE_FORMAT_PRINTF_SYSTEM(formatstring_parameter, first_argument) \ + _GL_ATTRIBUTE_FORMAT_PRINTF (formatstring_parameter, first_argument) +#else +# define _GL_ATTRIBUTE_FORMAT_PRINTF_SYSTEM(formatstring_parameter, first_argument) \ + _GL_ATTRIBUTE_FORMAT ((__printf__, formatstring_parameter, first_argument)) +#endif + +/* _GL_ATTRIBUTE_FORMAT_SCANF + indicates to GCC that the function takes a format string and arguments, + where the format string directives are the ones standardized by ISO C99 + and POSIX. */ +#if __GNUC__ > 4 || (__GNUC__ == 4 && __GNUC_MINOR__ >= 4) +# define _GL_ATTRIBUTE_FORMAT_SCANF(formatstring_parameter, first_argument) \ + _GL_ATTRIBUTE_FORMAT ((__gnu_scanf__, formatstring_parameter, first_argument)) +#else +# define _GL_ATTRIBUTE_FORMAT_SCANF(formatstring_parameter, first_argument) \ + _GL_ATTRIBUTE_FORMAT ((__scanf__, formatstring_parameter, first_argument)) +#endif + +/* _GL_ATTRIBUTE_FORMAT_SCANF_SYSTEM is like _GL_ATTRIBUTE_FORMAT_SCANF, + except that it indicates to GCC that the supported format string directives + are the ones of the system scanf(), rather than the ones standardized by + ISO C99 and POSIX. */ +#define _GL_ATTRIBUTE_FORMAT_SCANF_SYSTEM(formatstring_parameter, first_argument) \ + _GL_ATTRIBUTE_FORMAT ((__scanf__, formatstring_parameter, first_argument)) + +/* Solaris 10 and NetBSD 7.0 declare renameat in <unistd.h>, not in <stdio.h>. */ +/* But in any case avoid namespace pollution on glibc systems. */ +#if (@GNULIB_RENAMEAT@ || defined GNULIB_POSIXCHECK) && (defined __sun || defined __NetBSD__) \ + && ! defined __GLIBC__ +# include <unistd.h> +#endif + +/* MSVC declares 'perror' in <stdlib.h>, not in <stdio.h>. We must include + it before we #define perror rpl_perror. */ +/* But in any case avoid namespace pollution on glibc systems. */ +#if (@GNULIB_PERROR@ || defined GNULIB_POSIXCHECK) \ + && (defined _WIN32 && ! defined __CYGWIN__) \ + && ! defined __GLIBC__ +# include <stdlib.h> +#endif + +/* MSVC declares 'remove' in <io.h>, not in <stdio.h>. We must include + it before we #define remove rpl_remove. */ +/* MSVC declares 'rename' in <io.h>, not in <stdio.h>. We must include + it before we #define rename rpl_rename. */ +/* But in any case avoid namespace pollution on glibc systems. */ +#if (@GNULIB_REMOVE@ || @GNULIB_RENAME@ || defined GNULIB_POSIXCHECK) \ + && (defined _WIN32 && ! defined __CYGWIN__) \ + && ! defined __GLIBC__ +# include <io.h> +#endif + + +/* The definitions of _GL_FUNCDECL_RPL etc. are copied here. */ + +/* The definition of _GL_ARG_NONNULL is copied here. */ + +/* The definition of _GL_WARN_ON_USE is copied here. */ + +/* Macros for stringification. */ +#define _GL_STDIO_STRINGIZE(token) #token +#define _GL_STDIO_MACROEXPAND_AND_STRINGIZE(token) _GL_STDIO_STRINGIZE(token) + +/* When also using extern inline, suppress the use of static inline in + standard headers of problematic Apple configurations, as Libc at + least through Libc-825.26 (2013-04-09) mishandles it; see, e.g., + <https://lists.gnu.org/r/bug-gnulib/2012-12/msg00023.html>. + Perhaps Apple will fix this some day. */ +#if (defined _GL_EXTERN_INLINE_IN_USE && defined __APPLE__ \ + && defined __GNUC__ && defined __STDC__) +# undef putc_unlocked +#endif + +#if @GNULIB_DPRINTF@ +# if @REPLACE_DPRINTF@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# define dprintf rpl_dprintf +# endif +_GL_FUNCDECL_RPL (dprintf, int, (int fd, const char *format, ...) + _GL_ATTRIBUTE_FORMAT_PRINTF (2, 3) + _GL_ARG_NONNULL ((2))); +_GL_CXXALIAS_RPL (dprintf, int, (int fd, const char *format, ...)); +# else +# if !@HAVE_DPRINTF@ +_GL_FUNCDECL_SYS (dprintf, int, (int fd, const char *format, ...) + _GL_ATTRIBUTE_FORMAT_PRINTF (2, 3) + _GL_ARG_NONNULL ((2))); +# endif +_GL_CXXALIAS_SYS (dprintf, int, (int fd, const char *format, ...)); +# endif +_GL_CXXALIASWARN (dprintf); +#elif defined GNULIB_POSIXCHECK +# undef dprintf +# if HAVE_RAW_DECL_DPRINTF +_GL_WARN_ON_USE (dprintf, "dprintf is unportable - " + "use gnulib module dprintf for portability"); +# endif +#endif + +#if @GNULIB_FCLOSE@ +/* Close STREAM and its underlying file descriptor. */ +# if @REPLACE_FCLOSE@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# define fclose rpl_fclose +# endif +_GL_FUNCDECL_RPL (fclose, int, (FILE *stream) _GL_ARG_NONNULL ((1))); +_GL_CXXALIAS_RPL (fclose, int, (FILE *stream)); +# else +_GL_CXXALIAS_SYS (fclose, int, (FILE *stream)); +# endif +_GL_CXXALIASWARN (fclose); +#elif defined GNULIB_POSIXCHECK +# undef fclose +/* Assume fclose is always declared. */ +_GL_WARN_ON_USE (fclose, "fclose is not always POSIX compliant - " + "use gnulib module fclose for portable POSIX compliance"); +#endif + +#if @GNULIB_FDOPEN@ +# if @REPLACE_FDOPEN@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef fdopen +# define fdopen rpl_fdopen +# endif +_GL_FUNCDECL_RPL (fdopen, FILE *, (int fd, const char *mode) + _GL_ARG_NONNULL ((2))); +_GL_CXXALIAS_RPL (fdopen, FILE *, (int fd, const char *mode)); +# else +_GL_CXXALIAS_SYS (fdopen, FILE *, (int fd, const char *mode)); +# endif +_GL_CXXALIASWARN (fdopen); +#elif defined GNULIB_POSIXCHECK +# undef fdopen +/* Assume fdopen is always declared. */ +_GL_WARN_ON_USE (fdopen, "fdopen on native Windows platforms is not POSIX compliant - " + "use gnulib module fdopen for portability"); +#endif + +#if @GNULIB_FFLUSH@ +/* Flush all pending data on STREAM according to POSIX rules. Both + output and seekable input streams are supported. + Note! LOSS OF DATA can occur if fflush is applied on an input stream + that is _not_seekable_ or on an update stream that is _not_seekable_ + and in which the most recent operation was input. Seekability can + be tested with lseek(fileno(fp),0,SEEK_CUR). */ +# if @REPLACE_FFLUSH@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# define fflush rpl_fflush +# endif +_GL_FUNCDECL_RPL (fflush, int, (FILE *gl_stream)); +_GL_CXXALIAS_RPL (fflush, int, (FILE *gl_stream)); +# else +_GL_CXXALIAS_SYS (fflush, int, (FILE *gl_stream)); +# endif +_GL_CXXALIASWARN (fflush); +#elif defined GNULIB_POSIXCHECK +# undef fflush +/* Assume fflush is always declared. */ +_GL_WARN_ON_USE (fflush, "fflush is not always POSIX compliant - " + "use gnulib module fflush for portable POSIX compliance"); +#endif + +#if @GNULIB_FGETC@ +# if @REPLACE_STDIO_READ_FUNCS@ && @GNULIB_STDIO_H_NONBLOCKING@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef fgetc +# define fgetc rpl_fgetc +# endif +_GL_FUNCDECL_RPL (fgetc, int, (FILE *stream) _GL_ARG_NONNULL ((1))); +_GL_CXXALIAS_RPL (fgetc, int, (FILE *stream)); +# else +_GL_CXXALIAS_SYS (fgetc, int, (FILE *stream)); +# endif +_GL_CXXALIASWARN (fgetc); +#endif + +#if @GNULIB_FGETS@ +# if @REPLACE_STDIO_READ_FUNCS@ && @GNULIB_STDIO_H_NONBLOCKING@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef fgets +# define fgets rpl_fgets +# endif +_GL_FUNCDECL_RPL (fgets, char *, (char *s, int n, FILE *stream) + _GL_ARG_NONNULL ((1, 3))); +_GL_CXXALIAS_RPL (fgets, char *, (char *s, int n, FILE *stream)); +# else +_GL_CXXALIAS_SYS (fgets, char *, (char *s, int n, FILE *stream)); +# endif +_GL_CXXALIASWARN (fgets); +#endif + +#if @GNULIB_FOPEN@ +# if @REPLACE_FOPEN@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef fopen +# define fopen rpl_fopen +# endif +_GL_FUNCDECL_RPL (fopen, FILE *, (const char *filename, const char *mode) + _GL_ARG_NONNULL ((1, 2))); +_GL_CXXALIAS_RPL (fopen, FILE *, (const char *filename, const char *mode)); +# else +_GL_CXXALIAS_SYS (fopen, FILE *, (const char *filename, const char *mode)); +# endif +_GL_CXXALIASWARN (fopen); +#elif defined GNULIB_POSIXCHECK +# undef fopen +/* Assume fopen is always declared. */ +_GL_WARN_ON_USE (fopen, "fopen on native Windows platforms is not POSIX compliant - " + "use gnulib module fopen for portability"); +#endif + +#if @GNULIB_FPRINTF_POSIX@ || @GNULIB_FPRINTF@ +# if (@GNULIB_FPRINTF_POSIX@ && @REPLACE_FPRINTF@) \ + || (@GNULIB_FPRINTF@ && @REPLACE_STDIO_WRITE_FUNCS@ && (@GNULIB_STDIO_H_NONBLOCKING@ || @GNULIB_STDIO_H_SIGPIPE@)) +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# define fprintf rpl_fprintf +# endif +# define GNULIB_overrides_fprintf 1 +# if @GNULIB_FPRINTF_POSIX@ || @GNULIB_VFPRINTF_POSIX@ +_GL_FUNCDECL_RPL (fprintf, int, (FILE *fp, const char *format, ...) + _GL_ATTRIBUTE_FORMAT_PRINTF (2, 3) + _GL_ARG_NONNULL ((1, 2))); +# else +_GL_FUNCDECL_RPL (fprintf, int, (FILE *fp, const char *format, ...) + _GL_ATTRIBUTE_FORMAT_PRINTF_SYSTEM (2, 3) + _GL_ARG_NONNULL ((1, 2))); +# endif +_GL_CXXALIAS_RPL (fprintf, int, (FILE *fp, const char *format, ...)); +# else +_GL_CXXALIAS_SYS (fprintf, int, (FILE *fp, const char *format, ...)); +# endif +_GL_CXXALIASWARN (fprintf); +#endif +#if !@GNULIB_FPRINTF_POSIX@ && defined GNULIB_POSIXCHECK +# if !GNULIB_overrides_fprintf +# undef fprintf +# endif +/* Assume fprintf is always declared. */ +_GL_WARN_ON_USE (fprintf, "fprintf is not always POSIX compliant - " + "use gnulib module fprintf-posix for portable " + "POSIX compliance"); +#endif + +#if @GNULIB_FPURGE@ +/* Discard all pending buffered I/O data on STREAM. + STREAM must not be wide-character oriented. + When discarding pending output, the file position is set back to where it + was before the write calls. When discarding pending input, the file + position is advanced to match the end of the previously read input. + Return 0 if successful. Upon error, return -1 and set errno. */ +# if @REPLACE_FPURGE@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# define fpurge rpl_fpurge +# endif +_GL_FUNCDECL_RPL (fpurge, int, (FILE *gl_stream) _GL_ARG_NONNULL ((1))); +_GL_CXXALIAS_RPL (fpurge, int, (FILE *gl_stream)); +# else +# if !@HAVE_DECL_FPURGE@ +_GL_FUNCDECL_SYS (fpurge, int, (FILE *gl_stream) _GL_ARG_NONNULL ((1))); +# endif +_GL_CXXALIAS_SYS (fpurge, int, (FILE *gl_stream)); +# endif +_GL_CXXALIASWARN (fpurge); +#elif defined GNULIB_POSIXCHECK +# undef fpurge +# if HAVE_RAW_DECL_FPURGE +_GL_WARN_ON_USE (fpurge, "fpurge is not always present - " + "use gnulib module fpurge for portability"); +# endif +#endif + +#if @GNULIB_FPUTC@ +# if @REPLACE_STDIO_WRITE_FUNCS@ && (@GNULIB_STDIO_H_NONBLOCKING@ || @GNULIB_STDIO_H_SIGPIPE@) +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef fputc +# define fputc rpl_fputc +# endif +_GL_FUNCDECL_RPL (fputc, int, (int c, FILE *stream) _GL_ARG_NONNULL ((2))); +_GL_CXXALIAS_RPL (fputc, int, (int c, FILE *stream)); +# else +_GL_CXXALIAS_SYS (fputc, int, (int c, FILE *stream)); +# endif +_GL_CXXALIASWARN (fputc); +#endif + +#if @GNULIB_FPUTS@ +# if @REPLACE_STDIO_WRITE_FUNCS@ && (@GNULIB_STDIO_H_NONBLOCKING@ || @GNULIB_STDIO_H_SIGPIPE@) +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef fputs +# define fputs rpl_fputs +# endif +_GL_FUNCDECL_RPL (fputs, int, (const char *string, FILE *stream) + _GL_ARG_NONNULL ((1, 2))); +_GL_CXXALIAS_RPL (fputs, int, (const char *string, FILE *stream)); +# else +_GL_CXXALIAS_SYS (fputs, int, (const char *string, FILE *stream)); +# endif +_GL_CXXALIASWARN (fputs); +#endif + +#if @GNULIB_FREAD@ +# if @REPLACE_STDIO_READ_FUNCS@ && @GNULIB_STDIO_H_NONBLOCKING@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef fread +# define fread rpl_fread +# endif +_GL_FUNCDECL_RPL (fread, size_t, (void *ptr, size_t s, size_t n, FILE *stream) + _GL_ARG_NONNULL ((4))); +_GL_CXXALIAS_RPL (fread, size_t, (void *ptr, size_t s, size_t n, FILE *stream)); +# else +_GL_CXXALIAS_SYS (fread, size_t, (void *ptr, size_t s, size_t n, FILE *stream)); +# endif +_GL_CXXALIASWARN (fread); +#endif + +#if @GNULIB_FREOPEN@ +# if @REPLACE_FREOPEN@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef freopen +# define freopen rpl_freopen +# endif +_GL_FUNCDECL_RPL (freopen, FILE *, + (const char *filename, const char *mode, FILE *stream) + _GL_ARG_NONNULL ((2, 3))); +_GL_CXXALIAS_RPL (freopen, FILE *, + (const char *filename, const char *mode, FILE *stream)); +# else +_GL_CXXALIAS_SYS (freopen, FILE *, + (const char *filename, const char *mode, FILE *stream)); +# endif +_GL_CXXALIASWARN (freopen); +#elif defined GNULIB_POSIXCHECK +# undef freopen +/* Assume freopen is always declared. */ +_GL_WARN_ON_USE (freopen, + "freopen on native Windows platforms is not POSIX compliant - " + "use gnulib module freopen for portability"); +#endif + +#if @GNULIB_FSCANF@ +# if @REPLACE_STDIO_READ_FUNCS@ && @GNULIB_STDIO_H_NONBLOCKING@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef fscanf +# define fscanf rpl_fscanf +# endif +_GL_FUNCDECL_RPL (fscanf, int, (FILE *stream, const char *format, ...) + _GL_ATTRIBUTE_FORMAT_SCANF_SYSTEM (2, 3) + _GL_ARG_NONNULL ((1, 2))); +_GL_CXXALIAS_RPL (fscanf, int, (FILE *stream, const char *format, ...)); +# else +_GL_CXXALIAS_SYS (fscanf, int, (FILE *stream, const char *format, ...)); +# endif +_GL_CXXALIASWARN (fscanf); +#endif + + +/* Set up the following warnings, based on which modules are in use. + GNU Coding Standards discourage the use of fseek, since it imposes + an arbitrary limitation on some 32-bit hosts. Remember that the + fseek module depends on the fseeko module, so we only have three + cases to consider: + + 1. The developer is not using either module. Issue a warning under + GNULIB_POSIXCHECK for both functions, to remind them that both + functions have bugs on some systems. _GL_NO_LARGE_FILES has no + impact on this warning. + + 2. The developer is using both modules. They may be unaware of the + arbitrary limitations of fseek, so issue a warning under + GNULIB_POSIXCHECK. On the other hand, they may be using both + modules intentionally, so the developer can define + _GL_NO_LARGE_FILES in the compilation units where the use of fseek + is safe, to silence the warning. + + 3. The developer is using the fseeko module, but not fseek. Gnulib + guarantees that fseek will still work around platform bugs in that + case, but we presume that the developer is aware of the pitfalls of + fseek and was trying to avoid it, so issue a warning even when + GNULIB_POSIXCHECK is undefined. Again, _GL_NO_LARGE_FILES can be + defined to silence the warning in particular compilation units. + In C++ compilations with GNULIB_NAMESPACE, in order to avoid that + fseek gets defined as a macro, it is recommended that the developer + uses the fseek module, even if he is not calling the fseek function. + + Most gnulib clients that perform stream operations should fall into + category 3. */ + +#if @GNULIB_FSEEK@ +# if defined GNULIB_POSIXCHECK && !defined _GL_NO_LARGE_FILES +# define _GL_FSEEK_WARN /* Category 2, above. */ +# undef fseek +# endif +# if @REPLACE_FSEEK@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef fseek +# define fseek rpl_fseek +# endif +_GL_FUNCDECL_RPL (fseek, int, (FILE *fp, long offset, int whence) + _GL_ARG_NONNULL ((1))); +_GL_CXXALIAS_RPL (fseek, int, (FILE *fp, long offset, int whence)); +# else +_GL_CXXALIAS_SYS (fseek, int, (FILE *fp, long offset, int whence)); +# endif +_GL_CXXALIASWARN (fseek); +#endif + +#if @GNULIB_FSEEKO@ +# if !@GNULIB_FSEEK@ && !defined _GL_NO_LARGE_FILES +# define _GL_FSEEK_WARN /* Category 3, above. */ +# undef fseek +# endif +# if @REPLACE_FSEEKO@ +/* Provide an fseeko function that is aware of a preceding fflush(), and which + detects pipes. */ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef fseeko +# define fseeko rpl_fseeko +# endif +_GL_FUNCDECL_RPL (fseeko, int, (FILE *fp, off_t offset, int whence) + _GL_ARG_NONNULL ((1))); +_GL_CXXALIAS_RPL (fseeko, int, (FILE *fp, off_t offset, int whence)); +# else +# if ! @HAVE_DECL_FSEEKO@ +_GL_FUNCDECL_SYS (fseeko, int, (FILE *fp, off_t offset, int whence) + _GL_ARG_NONNULL ((1))); +# endif +_GL_CXXALIAS_SYS (fseeko, int, (FILE *fp, off_t offset, int whence)); +# endif +_GL_CXXALIASWARN (fseeko); +#elif defined GNULIB_POSIXCHECK +# define _GL_FSEEK_WARN /* Category 1, above. */ +# undef fseek +# undef fseeko +# if HAVE_RAW_DECL_FSEEKO +_GL_WARN_ON_USE (fseeko, "fseeko is unportable - " + "use gnulib module fseeko for portability"); +# endif +#endif + +#ifdef _GL_FSEEK_WARN +# undef _GL_FSEEK_WARN +/* Here, either fseek is undefined (but C89 guarantees that it is + declared), or it is defined as rpl_fseek (declared above). */ +_GL_WARN_ON_USE (fseek, "fseek cannot handle files larger than 4 GB " + "on 32-bit platforms - " + "use fseeko function for handling of large files"); +#endif + + +/* ftell, ftello. See the comments on fseek/fseeko. */ + +#if @GNULIB_FTELL@ +# if defined GNULIB_POSIXCHECK && !defined _GL_NO_LARGE_FILES +# define _GL_FTELL_WARN /* Category 2, above. */ +# undef ftell +# endif +# if @REPLACE_FTELL@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef ftell +# define ftell rpl_ftell +# endif +_GL_FUNCDECL_RPL (ftell, long, (FILE *fp) _GL_ARG_NONNULL ((1))); +_GL_CXXALIAS_RPL (ftell, long, (FILE *fp)); +# else +_GL_CXXALIAS_SYS (ftell, long, (FILE *fp)); +# endif +_GL_CXXALIASWARN (ftell); +#endif + +#if @GNULIB_FTELLO@ +# if !@GNULIB_FTELL@ && !defined _GL_NO_LARGE_FILES +# define _GL_FTELL_WARN /* Category 3, above. */ +# undef ftell +# endif +# if @REPLACE_FTELLO@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef ftello +# define ftello rpl_ftello +# endif +_GL_FUNCDECL_RPL (ftello, off_t, (FILE *fp) _GL_ARG_NONNULL ((1))); +_GL_CXXALIAS_RPL (ftello, off_t, (FILE *fp)); +# else +# if ! @HAVE_DECL_FTELLO@ +_GL_FUNCDECL_SYS (ftello, off_t, (FILE *fp) _GL_ARG_NONNULL ((1))); +# endif +_GL_CXXALIAS_SYS (ftello, off_t, (FILE *fp)); +# endif +_GL_CXXALIASWARN (ftello); +#elif defined GNULIB_POSIXCHECK +# define _GL_FTELL_WARN /* Category 1, above. */ +# undef ftell +# undef ftello +# if HAVE_RAW_DECL_FTELLO +_GL_WARN_ON_USE (ftello, "ftello is unportable - " + "use gnulib module ftello for portability"); +# endif +#endif + +#ifdef _GL_FTELL_WARN +# undef _GL_FTELL_WARN +/* Here, either ftell is undefined (but C89 guarantees that it is + declared), or it is defined as rpl_ftell (declared above). */ +_GL_WARN_ON_USE (ftell, "ftell cannot handle files larger than 4 GB " + "on 32-bit platforms - " + "use ftello function for handling of large files"); +#endif + + +#if @GNULIB_FWRITE@ +# if @REPLACE_STDIO_WRITE_FUNCS@ && (@GNULIB_STDIO_H_NONBLOCKING@ || @GNULIB_STDIO_H_SIGPIPE@) +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef fwrite +# define fwrite rpl_fwrite +# endif +_GL_FUNCDECL_RPL (fwrite, size_t, + (const void *ptr, size_t s, size_t n, FILE *stream) + _GL_ARG_NONNULL ((1, 4))); +_GL_CXXALIAS_RPL (fwrite, size_t, + (const void *ptr, size_t s, size_t n, FILE *stream)); +# else +_GL_CXXALIAS_SYS (fwrite, size_t, + (const void *ptr, size_t s, size_t n, FILE *stream)); + +/* Work around bug 11959 when fortifying glibc 2.4 through 2.15 + <https://sourceware.org/bugzilla/show_bug.cgi?id=11959>, + which sometimes causes an unwanted diagnostic for fwrite calls. + This affects only function declaration attributes under certain + versions of gcc and clang, and is not needed for C++. */ +# if (0 < __USE_FORTIFY_LEVEL \ + && __GLIBC__ == 2 && 4 <= __GLIBC_MINOR__ && __GLIBC_MINOR__ <= 15 \ + && 3 < __GNUC__ + (4 <= __GNUC_MINOR__) \ + && !defined __cplusplus) +# undef fwrite +# undef fwrite_unlocked +extern size_t __REDIRECT (rpl_fwrite, + (const void *__restrict, size_t, size_t, + FILE *__restrict), + fwrite); +extern size_t __REDIRECT (rpl_fwrite_unlocked, + (const void *__restrict, size_t, size_t, + FILE *__restrict), + fwrite_unlocked); +# define fwrite rpl_fwrite +# define fwrite_unlocked rpl_fwrite_unlocked +# endif +# endif +_GL_CXXALIASWARN (fwrite); +#endif + +#if @GNULIB_GETC@ +# if @REPLACE_STDIO_READ_FUNCS@ && @GNULIB_STDIO_H_NONBLOCKING@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef getc +# define getc rpl_fgetc +# endif +_GL_FUNCDECL_RPL (fgetc, int, (FILE *stream) _GL_ARG_NONNULL ((1))); +_GL_CXXALIAS_RPL_1 (getc, rpl_fgetc, int, (FILE *stream)); +# else +_GL_CXXALIAS_SYS (getc, int, (FILE *stream)); +# endif +_GL_CXXALIASWARN (getc); +#endif + +#if @GNULIB_GETCHAR@ +# if @REPLACE_STDIO_READ_FUNCS@ && @GNULIB_STDIO_H_NONBLOCKING@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef getchar +# define getchar rpl_getchar +# endif +_GL_FUNCDECL_RPL (getchar, int, (void)); +_GL_CXXALIAS_RPL (getchar, int, (void)); +# else +_GL_CXXALIAS_SYS (getchar, int, (void)); +# endif +_GL_CXXALIASWARN (getchar); +#endif + +#if @GNULIB_GETDELIM@ +/* Read input, up to (and including) the next occurrence of DELIMITER, from + STREAM, store it in *LINEPTR (and NUL-terminate it). + *LINEPTR is a pointer returned from malloc (or NULL), pointing to *LINESIZE + bytes of space. It is realloc'd as necessary. + Return the number of bytes read and stored at *LINEPTR (not including the + NUL terminator), or -1 on error or EOF. */ +# if @REPLACE_GETDELIM@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef getdelim +# define getdelim rpl_getdelim +# endif +_GL_FUNCDECL_RPL (getdelim, ssize_t, + (char **lineptr, size_t *linesize, int delimiter, + FILE *stream) + _GL_ARG_NONNULL ((1, 2, 4))); +_GL_CXXALIAS_RPL (getdelim, ssize_t, + (char **lineptr, size_t *linesize, int delimiter, + FILE *stream)); +# else +# if !@HAVE_DECL_GETDELIM@ +_GL_FUNCDECL_SYS (getdelim, ssize_t, + (char **lineptr, size_t *linesize, int delimiter, + FILE *stream) + _GL_ARG_NONNULL ((1, 2, 4))); +# endif +_GL_CXXALIAS_SYS (getdelim, ssize_t, + (char **lineptr, size_t *linesize, int delimiter, + FILE *stream)); +# endif +_GL_CXXALIASWARN (getdelim); +#elif defined GNULIB_POSIXCHECK +# undef getdelim +# if HAVE_RAW_DECL_GETDELIM +_GL_WARN_ON_USE (getdelim, "getdelim is unportable - " + "use gnulib module getdelim for portability"); +# endif +#endif + +#if @GNULIB_GETLINE@ +/* Read a line, up to (and including) the next newline, from STREAM, store it + in *LINEPTR (and NUL-terminate it). + *LINEPTR is a pointer returned from malloc (or NULL), pointing to *LINESIZE + bytes of space. It is realloc'd as necessary. + Return the number of bytes read and stored at *LINEPTR (not including the + NUL terminator), or -1 on error or EOF. */ +# if @REPLACE_GETLINE@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef getline +# define getline rpl_getline +# endif +_GL_FUNCDECL_RPL (getline, ssize_t, + (char **lineptr, size_t *linesize, FILE *stream) + _GL_ARG_NONNULL ((1, 2, 3))); +_GL_CXXALIAS_RPL (getline, ssize_t, + (char **lineptr, size_t *linesize, FILE *stream)); +# else +# if !@HAVE_DECL_GETLINE@ +_GL_FUNCDECL_SYS (getline, ssize_t, + (char **lineptr, size_t *linesize, FILE *stream) + _GL_ARG_NONNULL ((1, 2, 3))); +# endif +_GL_CXXALIAS_SYS (getline, ssize_t, + (char **lineptr, size_t *linesize, FILE *stream)); +# endif +# if @HAVE_DECL_GETLINE@ +_GL_CXXALIASWARN (getline); +# endif +#elif defined GNULIB_POSIXCHECK +# undef getline +# if HAVE_RAW_DECL_GETLINE +_GL_WARN_ON_USE (getline, "getline is unportable - " + "use gnulib module getline for portability"); +# endif +#endif + +/* It is very rare that the developer ever has full control of stdin, + so any use of gets warrants an unconditional warning; besides, C11 + removed it. */ +#undef gets +#if HAVE_RAW_DECL_GETS && !defined __cplusplus +_GL_WARN_ON_USE (gets, "gets is a security hole - use fgets instead"); +#endif + +#if @GNULIB_OBSTACK_PRINTF@ || @GNULIB_OBSTACK_PRINTF_POSIX@ +struct obstack; +/* Grow an obstack with formatted output. Return the number of + bytes added to OBS. No trailing nul byte is added, and the + object should be closed with obstack_finish before use. Upon + memory allocation error, call obstack_alloc_failed_handler. Upon + other error, return -1. */ +# if @REPLACE_OBSTACK_PRINTF@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# define obstack_printf rpl_obstack_printf +# endif +_GL_FUNCDECL_RPL (obstack_printf, int, + (struct obstack *obs, const char *format, ...) + _GL_ATTRIBUTE_FORMAT_PRINTF (2, 3) + _GL_ARG_NONNULL ((1, 2))); +_GL_CXXALIAS_RPL (obstack_printf, int, + (struct obstack *obs, const char *format, ...)); +# else +# if !@HAVE_DECL_OBSTACK_PRINTF@ +_GL_FUNCDECL_SYS (obstack_printf, int, + (struct obstack *obs, const char *format, ...) + _GL_ATTRIBUTE_FORMAT_PRINTF (2, 3) + _GL_ARG_NONNULL ((1, 2))); +# endif +_GL_CXXALIAS_SYS (obstack_printf, int, + (struct obstack *obs, const char *format, ...)); +# endif +_GL_CXXALIASWARN (obstack_printf); +# if @REPLACE_OBSTACK_PRINTF@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# define obstack_vprintf rpl_obstack_vprintf +# endif +_GL_FUNCDECL_RPL (obstack_vprintf, int, + (struct obstack *obs, const char *format, va_list args) + _GL_ATTRIBUTE_FORMAT_PRINTF (2, 0) + _GL_ARG_NONNULL ((1, 2))); +_GL_CXXALIAS_RPL (obstack_vprintf, int, + (struct obstack *obs, const char *format, va_list args)); +# else +# if !@HAVE_DECL_OBSTACK_PRINTF@ +_GL_FUNCDECL_SYS (obstack_vprintf, int, + (struct obstack *obs, const char *format, va_list args) + _GL_ATTRIBUTE_FORMAT_PRINTF (2, 0) + _GL_ARG_NONNULL ((1, 2))); +# endif +_GL_CXXALIAS_SYS (obstack_vprintf, int, + (struct obstack *obs, const char *format, va_list args)); +# endif +_GL_CXXALIASWARN (obstack_vprintf); +#endif + +#if @GNULIB_PCLOSE@ +# if !@HAVE_PCLOSE@ +_GL_FUNCDECL_SYS (pclose, int, (FILE *stream) _GL_ARG_NONNULL ((1))); +# endif +_GL_CXXALIAS_SYS (pclose, int, (FILE *stream)); +_GL_CXXALIASWARN (pclose); +#elif defined GNULIB_POSIXCHECK +# undef pclose +# if HAVE_RAW_DECL_PCLOSE +_GL_WARN_ON_USE (pclose, "pclose is unportable - " + "use gnulib module pclose for more portability"); +# endif +#endif + +#if @GNULIB_PERROR@ +/* Print a message to standard error, describing the value of ERRNO, + (if STRING is not NULL and not empty) prefixed with STRING and ": ", + and terminated with a newline. */ +# if @REPLACE_PERROR@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# define perror rpl_perror +# endif +_GL_FUNCDECL_RPL (perror, void, (const char *string)); +_GL_CXXALIAS_RPL (perror, void, (const char *string)); +# else +_GL_CXXALIAS_SYS (perror, void, (const char *string)); +# endif +_GL_CXXALIASWARN (perror); +#elif defined GNULIB_POSIXCHECK +# undef perror +/* Assume perror is always declared. */ +_GL_WARN_ON_USE (perror, "perror is not always POSIX compliant - " + "use gnulib module perror for portability"); +#endif + +#if @GNULIB_POPEN@ +# if @REPLACE_POPEN@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef popen +# define popen rpl_popen +# endif +_GL_FUNCDECL_RPL (popen, FILE *, (const char *cmd, const char *mode) + _GL_ARG_NONNULL ((1, 2))); +_GL_CXXALIAS_RPL (popen, FILE *, (const char *cmd, const char *mode)); +# else +# if !@HAVE_POPEN@ +_GL_FUNCDECL_SYS (popen, FILE *, (const char *cmd, const char *mode) + _GL_ARG_NONNULL ((1, 2))); +# endif +_GL_CXXALIAS_SYS (popen, FILE *, (const char *cmd, const char *mode)); +# endif +_GL_CXXALIASWARN (popen); +#elif defined GNULIB_POSIXCHECK +# undef popen +# if HAVE_RAW_DECL_POPEN +_GL_WARN_ON_USE (popen, "popen is buggy on some platforms - " + "use gnulib module popen or pipe for more portability"); +# endif +#endif + +#if @GNULIB_PRINTF_POSIX@ || @GNULIB_PRINTF@ +# if (@GNULIB_PRINTF_POSIX@ && @REPLACE_PRINTF@) \ + || (@GNULIB_PRINTF@ && @REPLACE_STDIO_WRITE_FUNCS@ && (@GNULIB_STDIO_H_NONBLOCKING@ || @GNULIB_STDIO_H_SIGPIPE@)) +# if defined __GNUC__ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +/* Don't break __attribute__((format(printf,M,N))). */ +# define printf __printf__ +# endif +# if @GNULIB_PRINTF_POSIX@ || @GNULIB_VFPRINTF_POSIX@ +_GL_FUNCDECL_RPL_1 (__printf__, int, + (const char *format, ...) + __asm__ (@ASM_SYMBOL_PREFIX@ + _GL_STDIO_MACROEXPAND_AND_STRINGIZE(rpl_printf)) + _GL_ATTRIBUTE_FORMAT_PRINTF (1, 2) + _GL_ARG_NONNULL ((1))); +# else +_GL_FUNCDECL_RPL_1 (__printf__, int, + (const char *format, ...) + __asm__ (@ASM_SYMBOL_PREFIX@ + _GL_STDIO_MACROEXPAND_AND_STRINGIZE(rpl_printf)) + _GL_ATTRIBUTE_FORMAT_PRINTF_SYSTEM (1, 2) + _GL_ARG_NONNULL ((1))); +# endif +_GL_CXXALIAS_RPL_1 (printf, __printf__, int, (const char *format, ...)); +# else +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# define printf rpl_printf +# endif +_GL_FUNCDECL_RPL (printf, int, + (const char *format, ...) + _GL_ATTRIBUTE_FORMAT_PRINTF (1, 2) + _GL_ARG_NONNULL ((1))); +_GL_CXXALIAS_RPL (printf, int, (const char *format, ...)); +# endif +# define GNULIB_overrides_printf 1 +# else +_GL_CXXALIAS_SYS (printf, int, (const char *format, ...)); +# endif +_GL_CXXALIASWARN (printf); +#endif +#if !@GNULIB_PRINTF_POSIX@ && defined GNULIB_POSIXCHECK +# if !GNULIB_overrides_printf +# undef printf +# endif +/* Assume printf is always declared. */ +_GL_WARN_ON_USE (printf, "printf is not always POSIX compliant - " + "use gnulib module printf-posix for portable " + "POSIX compliance"); +#endif + +#if @GNULIB_PUTC@ +# if @REPLACE_STDIO_WRITE_FUNCS@ && (@GNULIB_STDIO_H_NONBLOCKING@ || @GNULIB_STDIO_H_SIGPIPE@) +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef putc +# define putc rpl_fputc +# endif +_GL_FUNCDECL_RPL (fputc, int, (int c, FILE *stream) _GL_ARG_NONNULL ((2))); +_GL_CXXALIAS_RPL_1 (putc, rpl_fputc, int, (int c, FILE *stream)); +# else +_GL_CXXALIAS_SYS (putc, int, (int c, FILE *stream)); +# endif +_GL_CXXALIASWARN (putc); +#endif + +#if @GNULIB_PUTCHAR@ +# if @REPLACE_STDIO_WRITE_FUNCS@ && (@GNULIB_STDIO_H_NONBLOCKING@ || @GNULIB_STDIO_H_SIGPIPE@) +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef putchar +# define putchar rpl_putchar +# endif +_GL_FUNCDECL_RPL (putchar, int, (int c)); +_GL_CXXALIAS_RPL (putchar, int, (int c)); +# else +_GL_CXXALIAS_SYS (putchar, int, (int c)); +# endif +_GL_CXXALIASWARN (putchar); +#endif + +#if @GNULIB_PUTS@ +# if @REPLACE_STDIO_WRITE_FUNCS@ && (@GNULIB_STDIO_H_NONBLOCKING@ || @GNULIB_STDIO_H_SIGPIPE@) +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef puts +# define puts rpl_puts +# endif +_GL_FUNCDECL_RPL (puts, int, (const char *string) _GL_ARG_NONNULL ((1))); +_GL_CXXALIAS_RPL (puts, int, (const char *string)); +# else +_GL_CXXALIAS_SYS (puts, int, (const char *string)); +# endif +_GL_CXXALIASWARN (puts); +#endif + +#if @GNULIB_REMOVE@ +# if @REPLACE_REMOVE@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef remove +# define remove rpl_remove +# endif +_GL_FUNCDECL_RPL (remove, int, (const char *name) _GL_ARG_NONNULL ((1))); +_GL_CXXALIAS_RPL (remove, int, (const char *name)); +# else +_GL_CXXALIAS_SYS (remove, int, (const char *name)); +# endif +_GL_CXXALIASWARN (remove); +#elif defined GNULIB_POSIXCHECK +# undef remove +/* Assume remove is always declared. */ +_GL_WARN_ON_USE (remove, "remove cannot handle directories on some platforms - " + "use gnulib module remove for more portability"); +#endif + +#if @GNULIB_RENAME@ +# if @REPLACE_RENAME@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef rename +# define rename rpl_rename +# endif +_GL_FUNCDECL_RPL (rename, int, + (const char *old_filename, const char *new_filename) + _GL_ARG_NONNULL ((1, 2))); +_GL_CXXALIAS_RPL (rename, int, + (const char *old_filename, const char *new_filename)); +# else +_GL_CXXALIAS_SYS (rename, int, + (const char *old_filename, const char *new_filename)); +# endif +_GL_CXXALIASWARN (rename); +#elif defined GNULIB_POSIXCHECK +# undef rename +/* Assume rename is always declared. */ +_GL_WARN_ON_USE (rename, "rename is buggy on some platforms - " + "use gnulib module rename for more portability"); +#endif + +#if @GNULIB_RENAMEAT@ +# if @REPLACE_RENAMEAT@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef renameat +# define renameat rpl_renameat +# endif +_GL_FUNCDECL_RPL (renameat, int, + (int fd1, char const *file1, int fd2, char const *file2) + _GL_ARG_NONNULL ((2, 4))); +_GL_CXXALIAS_RPL (renameat, int, + (int fd1, char const *file1, int fd2, char const *file2)); +# else +# if !@HAVE_RENAMEAT@ +_GL_FUNCDECL_SYS (renameat, int, + (int fd1, char const *file1, int fd2, char const *file2) + _GL_ARG_NONNULL ((2, 4))); +# endif +_GL_CXXALIAS_SYS (renameat, int, + (int fd1, char const *file1, int fd2, char const *file2)); +# endif +_GL_CXXALIASWARN (renameat); +#elif defined GNULIB_POSIXCHECK +# undef renameat +# if HAVE_RAW_DECL_RENAMEAT +_GL_WARN_ON_USE (renameat, "renameat is not portable - " + "use gnulib module renameat for portability"); +# endif +#endif + +#if @GNULIB_SCANF@ +# if @REPLACE_STDIO_READ_FUNCS@ && @GNULIB_STDIO_H_NONBLOCKING@ +# if defined __GNUC__ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef scanf +/* Don't break __attribute__((format(scanf,M,N))). */ +# define scanf __scanf__ +# endif +_GL_FUNCDECL_RPL_1 (__scanf__, int, + (const char *format, ...) + __asm__ (@ASM_SYMBOL_PREFIX@ + _GL_STDIO_MACROEXPAND_AND_STRINGIZE(rpl_scanf)) + _GL_ATTRIBUTE_FORMAT_SCANF_SYSTEM (1, 2) + _GL_ARG_NONNULL ((1))); +_GL_CXXALIAS_RPL_1 (scanf, __scanf__, int, (const char *format, ...)); +# else +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef scanf +# define scanf rpl_scanf +# endif +_GL_FUNCDECL_RPL (scanf, int, (const char *format, ...) + _GL_ATTRIBUTE_FORMAT_SCANF_SYSTEM (1, 2) + _GL_ARG_NONNULL ((1))); +_GL_CXXALIAS_RPL (scanf, int, (const char *format, ...)); +# endif +# else +_GL_CXXALIAS_SYS (scanf, int, (const char *format, ...)); +# endif +_GL_CXXALIASWARN (scanf); +#endif + +#if @GNULIB_SNPRINTF@ +# if @REPLACE_SNPRINTF@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# define snprintf rpl_snprintf +# endif +_GL_FUNCDECL_RPL (snprintf, int, + (char *str, size_t size, const char *format, ...) + _GL_ATTRIBUTE_FORMAT_PRINTF (3, 4) + _GL_ARG_NONNULL ((3))); +_GL_CXXALIAS_RPL (snprintf, int, + (char *str, size_t size, const char *format, ...)); +# else +# if !@HAVE_DECL_SNPRINTF@ +_GL_FUNCDECL_SYS (snprintf, int, + (char *str, size_t size, const char *format, ...) + _GL_ATTRIBUTE_FORMAT_PRINTF (3, 4) + _GL_ARG_NONNULL ((3))); +# endif +_GL_CXXALIAS_SYS (snprintf, int, + (char *str, size_t size, const char *format, ...)); +# endif +_GL_CXXALIASWARN (snprintf); +#elif defined GNULIB_POSIXCHECK +# undef snprintf +# if HAVE_RAW_DECL_SNPRINTF +_GL_WARN_ON_USE (snprintf, "snprintf is unportable - " + "use gnulib module snprintf for portability"); +# endif +#endif + +/* Some people would argue that all sprintf uses should be warned about + (for example, OpenBSD issues a link warning for it), + since it can cause security holes due to buffer overruns. + However, we believe that sprintf can be used safely, and is more + efficient than snprintf in those safe cases; and as proof of our + belief, we use sprintf in several gnulib modules. So this header + intentionally avoids adding a warning to sprintf except when + GNULIB_POSIXCHECK is defined. */ + +#if @GNULIB_SPRINTF_POSIX@ +# if @REPLACE_SPRINTF@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# define sprintf rpl_sprintf +# endif +_GL_FUNCDECL_RPL (sprintf, int, (char *str, const char *format, ...) + _GL_ATTRIBUTE_FORMAT_PRINTF (2, 3) + _GL_ARG_NONNULL ((1, 2))); +_GL_CXXALIAS_RPL (sprintf, int, (char *str, const char *format, ...)); +# else +_GL_CXXALIAS_SYS (sprintf, int, (char *str, const char *format, ...)); +# endif +_GL_CXXALIASWARN (sprintf); +#elif defined GNULIB_POSIXCHECK +# undef sprintf +/* Assume sprintf is always declared. */ +_GL_WARN_ON_USE (sprintf, "sprintf is not always POSIX compliant - " + "use gnulib module sprintf-posix for portable " + "POSIX compliance"); +#endif + +#if @GNULIB_TMPFILE@ +# if @REPLACE_TMPFILE@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# define tmpfile rpl_tmpfile +# endif +_GL_FUNCDECL_RPL (tmpfile, FILE *, (void)); +_GL_CXXALIAS_RPL (tmpfile, FILE *, (void)); +# else +_GL_CXXALIAS_SYS (tmpfile, FILE *, (void)); +# endif +_GL_CXXALIASWARN (tmpfile); +#elif defined GNULIB_POSIXCHECK +# undef tmpfile +# if HAVE_RAW_DECL_TMPFILE +_GL_WARN_ON_USE (tmpfile, "tmpfile is not usable on mingw - " + "use gnulib module tmpfile for portability"); +# endif +#endif + +#if @GNULIB_VASPRINTF@ +/* Write formatted output to a string dynamically allocated with malloc(). + If the memory allocation succeeds, store the address of the string in + *RESULT and return the number of resulting bytes, excluding the trailing + NUL. Upon memory allocation error, or some other error, return -1. */ +# if @REPLACE_VASPRINTF@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# define asprintf rpl_asprintf +# endif +_GL_FUNCDECL_RPL (asprintf, int, + (char **result, const char *format, ...) + _GL_ATTRIBUTE_FORMAT_PRINTF (2, 3) + _GL_ARG_NONNULL ((1, 2))); +_GL_CXXALIAS_RPL (asprintf, int, + (char **result, const char *format, ...)); +# else +# if !@HAVE_VASPRINTF@ +_GL_FUNCDECL_SYS (asprintf, int, + (char **result, const char *format, ...) + _GL_ATTRIBUTE_FORMAT_PRINTF (2, 3) + _GL_ARG_NONNULL ((1, 2))); +# endif +_GL_CXXALIAS_SYS (asprintf, int, + (char **result, const char *format, ...)); +# endif +_GL_CXXALIASWARN (asprintf); +# if @REPLACE_VASPRINTF@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# define vasprintf rpl_vasprintf +# endif +_GL_FUNCDECL_RPL (vasprintf, int, + (char **result, const char *format, va_list args) + _GL_ATTRIBUTE_FORMAT_PRINTF (2, 0) + _GL_ARG_NONNULL ((1, 2))); +_GL_CXXALIAS_RPL (vasprintf, int, + (char **result, const char *format, va_list args)); +# else +# if !@HAVE_VASPRINTF@ +_GL_FUNCDECL_SYS (vasprintf, int, + (char **result, const char *format, va_list args) + _GL_ATTRIBUTE_FORMAT_PRINTF (2, 0) + _GL_ARG_NONNULL ((1, 2))); +# endif +_GL_CXXALIAS_SYS (vasprintf, int, + (char **result, const char *format, va_list args)); +# endif +_GL_CXXALIASWARN (vasprintf); +#endif + +#if @GNULIB_VDPRINTF@ +# if @REPLACE_VDPRINTF@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# define vdprintf rpl_vdprintf +# endif +_GL_FUNCDECL_RPL (vdprintf, int, (int fd, const char *format, va_list args) + _GL_ATTRIBUTE_FORMAT_PRINTF (2, 0) + _GL_ARG_NONNULL ((2))); +_GL_CXXALIAS_RPL (vdprintf, int, (int fd, const char *format, va_list args)); +# else +# if !@HAVE_VDPRINTF@ +_GL_FUNCDECL_SYS (vdprintf, int, (int fd, const char *format, va_list args) + _GL_ATTRIBUTE_FORMAT_PRINTF (2, 0) + _GL_ARG_NONNULL ((2))); +# endif +/* Need to cast, because on Solaris, the third parameter will likely be + __va_list args. */ +_GL_CXXALIAS_SYS_CAST (vdprintf, int, + (int fd, const char *format, va_list args)); +# endif +_GL_CXXALIASWARN (vdprintf); +#elif defined GNULIB_POSIXCHECK +# undef vdprintf +# if HAVE_RAW_DECL_VDPRINTF +_GL_WARN_ON_USE (vdprintf, "vdprintf is unportable - " + "use gnulib module vdprintf for portability"); +# endif +#endif + +#if @GNULIB_VFPRINTF_POSIX@ || @GNULIB_VFPRINTF@ +# if (@GNULIB_VFPRINTF_POSIX@ && @REPLACE_VFPRINTF@) \ + || (@GNULIB_VFPRINTF@ && @REPLACE_STDIO_WRITE_FUNCS@ && (@GNULIB_STDIO_H_NONBLOCKING@ || @GNULIB_STDIO_H_SIGPIPE@)) +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# define vfprintf rpl_vfprintf +# endif +# define GNULIB_overrides_vfprintf 1 +# if @GNULIB_VFPRINTF_POSIX@ +_GL_FUNCDECL_RPL (vfprintf, int, (FILE *fp, const char *format, va_list args) + _GL_ATTRIBUTE_FORMAT_PRINTF (2, 0) + _GL_ARG_NONNULL ((1, 2))); +# else +_GL_FUNCDECL_RPL (vfprintf, int, (FILE *fp, const char *format, va_list args) + _GL_ATTRIBUTE_FORMAT_PRINTF_SYSTEM (2, 0) + _GL_ARG_NONNULL ((1, 2))); +# endif +_GL_CXXALIAS_RPL (vfprintf, int, (FILE *fp, const char *format, va_list args)); +# else +/* Need to cast, because on Solaris, the third parameter is + __va_list args + and GCC's fixincludes did not change this to __gnuc_va_list. */ +_GL_CXXALIAS_SYS_CAST (vfprintf, int, + (FILE *fp, const char *format, va_list args)); +# endif +_GL_CXXALIASWARN (vfprintf); +#endif +#if !@GNULIB_VFPRINTF_POSIX@ && defined GNULIB_POSIXCHECK +# if !GNULIB_overrides_vfprintf +# undef vfprintf +# endif +/* Assume vfprintf is always declared. */ +_GL_WARN_ON_USE (vfprintf, "vfprintf is not always POSIX compliant - " + "use gnulib module vfprintf-posix for portable " + "POSIX compliance"); +#endif + +#if @GNULIB_VFSCANF@ +# if @REPLACE_STDIO_READ_FUNCS@ && @GNULIB_STDIO_H_NONBLOCKING@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef vfscanf +# define vfscanf rpl_vfscanf +# endif +_GL_FUNCDECL_RPL (vfscanf, int, + (FILE *stream, const char *format, va_list args) + _GL_ATTRIBUTE_FORMAT_SCANF_SYSTEM (2, 0) + _GL_ARG_NONNULL ((1, 2))); +_GL_CXXALIAS_RPL (vfscanf, int, + (FILE *stream, const char *format, va_list args)); +# else +_GL_CXXALIAS_SYS (vfscanf, int, + (FILE *stream, const char *format, va_list args)); +# endif +_GL_CXXALIASWARN (vfscanf); +#endif + +#if @GNULIB_VPRINTF_POSIX@ || @GNULIB_VPRINTF@ +# if (@GNULIB_VPRINTF_POSIX@ && @REPLACE_VPRINTF@) \ + || (@GNULIB_VPRINTF@ && @REPLACE_STDIO_WRITE_FUNCS@ && (@GNULIB_STDIO_H_NONBLOCKING@ || @GNULIB_STDIO_H_SIGPIPE@)) +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# define vprintf rpl_vprintf +# endif +# define GNULIB_overrides_vprintf 1 +# if @GNULIB_VPRINTF_POSIX@ || @GNULIB_VFPRINTF_POSIX@ +_GL_FUNCDECL_RPL (vprintf, int, (const char *format, va_list args) + _GL_ATTRIBUTE_FORMAT_PRINTF (1, 0) + _GL_ARG_NONNULL ((1))); +# else +_GL_FUNCDECL_RPL (vprintf, int, (const char *format, va_list args) + _GL_ATTRIBUTE_FORMAT_PRINTF_SYSTEM (1, 0) + _GL_ARG_NONNULL ((1))); +# endif +_GL_CXXALIAS_RPL (vprintf, int, (const char *format, va_list args)); +# else +/* Need to cast, because on Solaris, the second parameter is + __va_list args + and GCC's fixincludes did not change this to __gnuc_va_list. */ +_GL_CXXALIAS_SYS_CAST (vprintf, int, (const char *format, va_list args)); +# endif +_GL_CXXALIASWARN (vprintf); +#endif +#if !@GNULIB_VPRINTF_POSIX@ && defined GNULIB_POSIXCHECK +# if !GNULIB_overrides_vprintf +# undef vprintf +# endif +/* Assume vprintf is always declared. */ +_GL_WARN_ON_USE (vprintf, "vprintf is not always POSIX compliant - " + "use gnulib module vprintf-posix for portable " + "POSIX compliance"); +#endif + +#if @GNULIB_VSCANF@ +# if @REPLACE_STDIO_READ_FUNCS@ && @GNULIB_STDIO_H_NONBLOCKING@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef vscanf +# define vscanf rpl_vscanf +# endif +_GL_FUNCDECL_RPL (vscanf, int, (const char *format, va_list args) + _GL_ATTRIBUTE_FORMAT_SCANF_SYSTEM (1, 0) + _GL_ARG_NONNULL ((1))); +_GL_CXXALIAS_RPL (vscanf, int, (const char *format, va_list args)); +# else +_GL_CXXALIAS_SYS (vscanf, int, (const char *format, va_list args)); +# endif +_GL_CXXALIASWARN (vscanf); +#endif + +#if @GNULIB_VSNPRINTF@ +# if @REPLACE_VSNPRINTF@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# define vsnprintf rpl_vsnprintf +# endif +_GL_FUNCDECL_RPL (vsnprintf, int, + (char *str, size_t size, const char *format, va_list args) + _GL_ATTRIBUTE_FORMAT_PRINTF (3, 0) + _GL_ARG_NONNULL ((3))); +_GL_CXXALIAS_RPL (vsnprintf, int, + (char *str, size_t size, const char *format, va_list args)); +# else +# if !@HAVE_DECL_VSNPRINTF@ +_GL_FUNCDECL_SYS (vsnprintf, int, + (char *str, size_t size, const char *format, va_list args) + _GL_ATTRIBUTE_FORMAT_PRINTF (3, 0) + _GL_ARG_NONNULL ((3))); +# endif +_GL_CXXALIAS_SYS (vsnprintf, int, + (char *str, size_t size, const char *format, va_list args)); +# endif +_GL_CXXALIASWARN (vsnprintf); +#elif defined GNULIB_POSIXCHECK +# undef vsnprintf +# if HAVE_RAW_DECL_VSNPRINTF +_GL_WARN_ON_USE (vsnprintf, "vsnprintf is unportable - " + "use gnulib module vsnprintf for portability"); +# endif +#endif + +#if @GNULIB_VSPRINTF_POSIX@ +# if @REPLACE_VSPRINTF@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# define vsprintf rpl_vsprintf +# endif +_GL_FUNCDECL_RPL (vsprintf, int, + (char *str, const char *format, va_list args) + _GL_ATTRIBUTE_FORMAT_PRINTF (2, 0) + _GL_ARG_NONNULL ((1, 2))); +_GL_CXXALIAS_RPL (vsprintf, int, + (char *str, const char *format, va_list args)); +# else +/* Need to cast, because on Solaris, the third parameter is + __va_list args + and GCC's fixincludes did not change this to __gnuc_va_list. */ +_GL_CXXALIAS_SYS_CAST (vsprintf, int, + (char *str, const char *format, va_list args)); +# endif +_GL_CXXALIASWARN (vsprintf); +#elif defined GNULIB_POSIXCHECK +# undef vsprintf +/* Assume vsprintf is always declared. */ +_GL_WARN_ON_USE (vsprintf, "vsprintf is not always POSIX compliant - " + "use gnulib module vsprintf-posix for portable " + "POSIX compliance"); +#endif + +#endif /* _@GUARD_PREFIX@_STDIO_H */ +#endif /* _@GUARD_PREFIX@_STDIO_H */ +#endif diff --git a/lib/stdlib.in.h b/lib/stdlib.in.h new file mode 100644 index 0000000..441c018 --- /dev/null +++ b/lib/stdlib.in.h @@ -0,0 +1,1013 @@ +/* A GNU-like <stdlib.h>. + + Copyright (C) 1995, 2001-2004, 2006-2018 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +#if __GNUC__ >= 3 +@PRAGMA_SYSTEM_HEADER@ +#endif +@PRAGMA_COLUMNS@ + +#if defined __need_system_stdlib_h || defined __need_malloc_and_calloc +/* Special invocation conventions inside some gnulib header files, + and inside some glibc header files, respectively. */ + +#@INCLUDE_NEXT@ @NEXT_STDLIB_H@ + +#else +/* Normal invocation convention. */ + +#ifndef _@GUARD_PREFIX@_STDLIB_H + +/* The include_next requires a split double-inclusion guard. */ +#@INCLUDE_NEXT@ @NEXT_STDLIB_H@ + +#ifndef _@GUARD_PREFIX@_STDLIB_H +#define _@GUARD_PREFIX@_STDLIB_H + +/* NetBSD 5.0 mis-defines NULL. */ +#include <stddef.h> + +/* MirBSD 10 defines WEXITSTATUS in <sys/wait.h>, not in <stdlib.h>. */ +#if @GNULIB_SYSTEM_POSIX@ && !defined WEXITSTATUS +# include <sys/wait.h> +#endif + +/* Solaris declares getloadavg() in <sys/loadavg.h>. */ +#if (@GNULIB_GETLOADAVG@ || defined GNULIB_POSIXCHECK) && @HAVE_SYS_LOADAVG_H@ +/* OpenIndiana has a bug: <sys/time.h> must be included before + <sys/loadavg.h>. */ +# include <sys/time.h> +# include <sys/loadavg.h> +#endif + +/* Native Windows platforms declare mktemp() in <io.h>. */ +#if 0 && (defined _WIN32 && ! defined __CYGWIN__) +# include <io.h> +#endif + +#if @GNULIB_RANDOM_R@ + +/* OSF/1 5.1 declares 'struct random_data' in <random.h>, which is included + from <stdlib.h> if _REENTRANT is defined. Include it whenever we need + 'struct random_data'. */ +# if @HAVE_RANDOM_H@ +# include <random.h> +# endif + +# if !@HAVE_STRUCT_RANDOM_DATA@ || @REPLACE_RANDOM_R@ || !@HAVE_RANDOM_R@ +# include <stdint.h> +# endif + +# if !@HAVE_STRUCT_RANDOM_DATA@ +/* Define 'struct random_data'. + But allow multiple gnulib generated <stdlib.h> replacements to coexist. */ +# if !GNULIB_defined_struct_random_data +struct random_data +{ + int32_t *fptr; /* Front pointer. */ + int32_t *rptr; /* Rear pointer. */ + int32_t *state; /* Array of state values. */ + int rand_type; /* Type of random number generator. */ + int rand_deg; /* Degree of random number generator. */ + int rand_sep; /* Distance between front and rear. */ + int32_t *end_ptr; /* Pointer behind state table. */ +}; +# define GNULIB_defined_struct_random_data 1 +# endif +# endif +#endif + +#if (@GNULIB_MKSTEMP@ || @GNULIB_MKSTEMPS@ || @GNULIB_MKOSTEMP@ || @GNULIB_MKOSTEMPS@ || @GNULIB_GETSUBOPT@ || defined GNULIB_POSIXCHECK) && ! defined __GLIBC__ && !(defined _WIN32 && ! defined __CYGWIN__) +/* On Mac OS X 10.3, only <unistd.h> declares mkstemp. */ +/* On Mac OS X 10.5, only <unistd.h> declares mkstemps. */ +/* On Mac OS X 10.13, only <unistd.h> declares mkostemp and mkostemps. */ +/* On Cygwin 1.7.1, only <unistd.h> declares getsubopt. */ +/* But avoid namespace pollution on glibc systems and native Windows. */ +# include <unistd.h> +#endif + +/* The __attribute__ feature is available in gcc versions 2.5 and later. + The attribute __pure__ was added in gcc 2.96. */ +#if __GNUC__ > 2 || (__GNUC__ == 2 && __GNUC_MINOR__ >= 96) +# define _GL_ATTRIBUTE_PURE __attribute__ ((__pure__)) +#else +# define _GL_ATTRIBUTE_PURE /* empty */ +#endif + +/* The definition of _Noreturn is copied here. */ + +/* The definitions of _GL_FUNCDECL_RPL etc. are copied here. */ + +/* The definition of _GL_ARG_NONNULL is copied here. */ + +/* The definition of _GL_WARN_ON_USE is copied here. */ + + +/* Some systems do not define EXIT_*, despite otherwise supporting C89. */ +#ifndef EXIT_SUCCESS +# define EXIT_SUCCESS 0 +#endif +/* Tandem/NSK and other platforms that define EXIT_FAILURE as -1 interfere + with proper operation of xargs. */ +#ifndef EXIT_FAILURE +# define EXIT_FAILURE 1 +#elif EXIT_FAILURE != 1 +# undef EXIT_FAILURE +# define EXIT_FAILURE 1 +#endif + + +#if @GNULIB__EXIT@ +/* Terminate the current process with the given return code, without running + the 'atexit' handlers. */ +# if !@HAVE__EXIT@ +_GL_FUNCDECL_SYS (_Exit, _Noreturn void, (int status)); +# endif +_GL_CXXALIAS_SYS (_Exit, void, (int status)); +_GL_CXXALIASWARN (_Exit); +#elif defined GNULIB_POSIXCHECK +# undef _Exit +# if HAVE_RAW_DECL__EXIT +_GL_WARN_ON_USE (_Exit, "_Exit is unportable - " + "use gnulib module _Exit for portability"); +# endif +#endif + + +#if @GNULIB_ATOLL@ +/* Parse a signed decimal integer. + Returns the value of the integer. Errors are not detected. */ +# if !@HAVE_ATOLL@ +_GL_FUNCDECL_SYS (atoll, long long, (const char *string) + _GL_ATTRIBUTE_PURE + _GL_ARG_NONNULL ((1))); +# endif +_GL_CXXALIAS_SYS (atoll, long long, (const char *string)); +_GL_CXXALIASWARN (atoll); +#elif defined GNULIB_POSIXCHECK +# undef atoll +# if HAVE_RAW_DECL_ATOLL +_GL_WARN_ON_USE (atoll, "atoll is unportable - " + "use gnulib module atoll for portability"); +# endif +#endif + +#if @GNULIB_CALLOC_POSIX@ +# if @REPLACE_CALLOC@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef calloc +# define calloc rpl_calloc +# endif +_GL_FUNCDECL_RPL (calloc, void *, (size_t nmemb, size_t size)); +_GL_CXXALIAS_RPL (calloc, void *, (size_t nmemb, size_t size)); +# else +_GL_CXXALIAS_SYS (calloc, void *, (size_t nmemb, size_t size)); +# endif +_GL_CXXALIASWARN (calloc); +#elif defined GNULIB_POSIXCHECK +# undef calloc +/* Assume calloc is always declared. */ +_GL_WARN_ON_USE (calloc, "calloc is not POSIX compliant everywhere - " + "use gnulib module calloc-posix for portability"); +#endif + +#if @GNULIB_CANONICALIZE_FILE_NAME@ +# if @REPLACE_CANONICALIZE_FILE_NAME@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# define canonicalize_file_name rpl_canonicalize_file_name +# endif +_GL_FUNCDECL_RPL (canonicalize_file_name, char *, (const char *name) + _GL_ARG_NONNULL ((1))); +_GL_CXXALIAS_RPL (canonicalize_file_name, char *, (const char *name)); +# else +# if !@HAVE_CANONICALIZE_FILE_NAME@ +_GL_FUNCDECL_SYS (canonicalize_file_name, char *, (const char *name) + _GL_ARG_NONNULL ((1))); +# endif +_GL_CXXALIAS_SYS (canonicalize_file_name, char *, (const char *name)); +# endif +_GL_CXXALIASWARN (canonicalize_file_name); +#elif defined GNULIB_POSIXCHECK +# undef canonicalize_file_name +# if HAVE_RAW_DECL_CANONICALIZE_FILE_NAME +_GL_WARN_ON_USE (canonicalize_file_name, + "canonicalize_file_name is unportable - " + "use gnulib module canonicalize-lgpl for portability"); +# endif +#endif + +#if @GNULIB_GETLOADAVG@ +/* Store max(NELEM,3) load average numbers in LOADAVG[]. + The three numbers are the load average of the last 1 minute, the last 5 + minutes, and the last 15 minutes, respectively. + LOADAVG is an array of NELEM numbers. */ +# if !@HAVE_DECL_GETLOADAVG@ +_GL_FUNCDECL_SYS (getloadavg, int, (double loadavg[], int nelem) + _GL_ARG_NONNULL ((1))); +# endif +_GL_CXXALIAS_SYS (getloadavg, int, (double loadavg[], int nelem)); +_GL_CXXALIASWARN (getloadavg); +#elif defined GNULIB_POSIXCHECK +# undef getloadavg +# if HAVE_RAW_DECL_GETLOADAVG +_GL_WARN_ON_USE (getloadavg, "getloadavg is not portable - " + "use gnulib module getloadavg for portability"); +# endif +#endif + +#if @GNULIB_GETSUBOPT@ +/* Assuming *OPTIONP is a comma separated list of elements of the form + "token" or "token=value", getsubopt parses the first of these elements. + If the first element refers to a "token" that is member of the given + NULL-terminated array of tokens: + - It replaces the comma with a NUL byte, updates *OPTIONP to point past + the first option and the comma, sets *VALUEP to the value of the + element (or NULL if it doesn't contain an "=" sign), + - It returns the index of the "token" in the given array of tokens. + Otherwise it returns -1, and *OPTIONP and *VALUEP are undefined. + For more details see the POSIX:2001 specification. + http://www.opengroup.org/susv3xsh/getsubopt.html */ +# if !@HAVE_GETSUBOPT@ +_GL_FUNCDECL_SYS (getsubopt, int, + (char **optionp, char *const *tokens, char **valuep) + _GL_ARG_NONNULL ((1, 2, 3))); +# endif +_GL_CXXALIAS_SYS (getsubopt, int, + (char **optionp, char *const *tokens, char **valuep)); +_GL_CXXALIASWARN (getsubopt); +#elif defined GNULIB_POSIXCHECK +# undef getsubopt +# if HAVE_RAW_DECL_GETSUBOPT +_GL_WARN_ON_USE (getsubopt, "getsubopt is unportable - " + "use gnulib module getsubopt for portability"); +# endif +#endif + +#if @GNULIB_GRANTPT@ +/* Change the ownership and access permission of the slave side of the + pseudo-terminal whose master side is specified by FD. */ +# if !@HAVE_GRANTPT@ +_GL_FUNCDECL_SYS (grantpt, int, (int fd)); +# endif +_GL_CXXALIAS_SYS (grantpt, int, (int fd)); +_GL_CXXALIASWARN (grantpt); +#elif defined GNULIB_POSIXCHECK +# undef grantpt +# if HAVE_RAW_DECL_GRANTPT +_GL_WARN_ON_USE (grantpt, "grantpt is not portable - " + "use gnulib module grantpt for portability"); +# endif +#endif + +/* If _GL_USE_STDLIB_ALLOC is nonzero, the including module does not + rely on GNU or POSIX semantics for malloc and realloc (for example, + by never specifying a zero size), so it does not need malloc or + realloc to be redefined. */ +#if @GNULIB_MALLOC_POSIX@ +# if @REPLACE_MALLOC@ +# if !((defined __cplusplus && defined GNULIB_NAMESPACE) \ + || _GL_USE_STDLIB_ALLOC) +# undef malloc +# define malloc rpl_malloc +# endif +_GL_FUNCDECL_RPL (malloc, void *, (size_t size)); +_GL_CXXALIAS_RPL (malloc, void *, (size_t size)); +# else +_GL_CXXALIAS_SYS (malloc, void *, (size_t size)); +# endif +_GL_CXXALIASWARN (malloc); +#elif defined GNULIB_POSIXCHECK && !_GL_USE_STDLIB_ALLOC +# undef malloc +/* Assume malloc is always declared. */ +_GL_WARN_ON_USE (malloc, "malloc is not POSIX compliant everywhere - " + "use gnulib module malloc-posix for portability"); +#endif + +/* Convert a multibyte character to a wide character. */ +#if @GNULIB_MBTOWC@ +# if @REPLACE_MBTOWC@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef mbtowc +# define mbtowc rpl_mbtowc +# endif +_GL_FUNCDECL_RPL (mbtowc, int, (wchar_t *pwc, const char *s, size_t n)); +_GL_CXXALIAS_RPL (mbtowc, int, (wchar_t *pwc, const char *s, size_t n)); +# else +_GL_CXXALIAS_SYS (mbtowc, int, (wchar_t *pwc, const char *s, size_t n)); +# endif +_GL_CXXALIASWARN (mbtowc); +#endif + +#if @GNULIB_MKDTEMP@ +/* Create a unique temporary directory from TEMPLATE. + The last six characters of TEMPLATE must be "XXXXXX"; + they are replaced with a string that makes the directory name unique. + Returns TEMPLATE, or a null pointer if it cannot get a unique name. + The directory is created mode 700. */ +# if !@HAVE_MKDTEMP@ +_GL_FUNCDECL_SYS (mkdtemp, char *, (char * /*template*/) _GL_ARG_NONNULL ((1))); +# endif +_GL_CXXALIAS_SYS (mkdtemp, char *, (char * /*template*/)); +_GL_CXXALIASWARN (mkdtemp); +#elif defined GNULIB_POSIXCHECK +# undef mkdtemp +# if HAVE_RAW_DECL_MKDTEMP +_GL_WARN_ON_USE (mkdtemp, "mkdtemp is unportable - " + "use gnulib module mkdtemp for portability"); +# endif +#endif + +#if @GNULIB_MKOSTEMP@ +/* Create a unique temporary file from TEMPLATE. + The last six characters of TEMPLATE must be "XXXXXX"; + they are replaced with a string that makes the file name unique. + The flags are a bitmask, possibly including O_CLOEXEC (defined in <fcntl.h>) + and O_TEXT, O_BINARY (defined in "binary-io.h"). + The file is then created, with the specified flags, ensuring it didn't exist + before. + The file is created read-write (mask at least 0600 & ~umask), but it may be + world-readable and world-writable (mask 0666 & ~umask), depending on the + implementation. + Returns the open file descriptor if successful, otherwise -1 and errno + set. */ +# if !@HAVE_MKOSTEMP@ +_GL_FUNCDECL_SYS (mkostemp, int, (char * /*template*/, int /*flags*/) + _GL_ARG_NONNULL ((1))); +# endif +_GL_CXXALIAS_SYS (mkostemp, int, (char * /*template*/, int /*flags*/)); +_GL_CXXALIASWARN (mkostemp); +#elif defined GNULIB_POSIXCHECK +# undef mkostemp +# if HAVE_RAW_DECL_MKOSTEMP +_GL_WARN_ON_USE (mkostemp, "mkostemp is unportable - " + "use gnulib module mkostemp for portability"); +# endif +#endif + +#if @GNULIB_MKOSTEMPS@ +/* Create a unique temporary file from TEMPLATE. + The last six characters of TEMPLATE before a suffix of length + SUFFIXLEN must be "XXXXXX"; + they are replaced with a string that makes the file name unique. + The flags are a bitmask, possibly including O_CLOEXEC (defined in <fcntl.h>) + and O_TEXT, O_BINARY (defined in "binary-io.h"). + The file is then created, with the specified flags, ensuring it didn't exist + before. + The file is created read-write (mask at least 0600 & ~umask), but it may be + world-readable and world-writable (mask 0666 & ~umask), depending on the + implementation. + Returns the open file descriptor if successful, otherwise -1 and errno + set. */ +# if !@HAVE_MKOSTEMPS@ +_GL_FUNCDECL_SYS (mkostemps, int, + (char * /*template*/, int /*suffixlen*/, int /*flags*/) + _GL_ARG_NONNULL ((1))); +# endif +_GL_CXXALIAS_SYS (mkostemps, int, + (char * /*template*/, int /*suffixlen*/, int /*flags*/)); +_GL_CXXALIASWARN (mkostemps); +#elif defined GNULIB_POSIXCHECK +# undef mkostemps +# if HAVE_RAW_DECL_MKOSTEMPS +_GL_WARN_ON_USE (mkostemps, "mkostemps is unportable - " + "use gnulib module mkostemps for portability"); +# endif +#endif + +#if @GNULIB_MKSTEMP@ +/* Create a unique temporary file from TEMPLATE. + The last six characters of TEMPLATE must be "XXXXXX"; + they are replaced with a string that makes the file name unique. + The file is then created, ensuring it didn't exist before. + The file is created read-write (mask at least 0600 & ~umask), but it may be + world-readable and world-writable (mask 0666 & ~umask), depending on the + implementation. + Returns the open file descriptor if successful, otherwise -1 and errno + set. */ +# if @REPLACE_MKSTEMP@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# define mkstemp rpl_mkstemp +# endif +_GL_FUNCDECL_RPL (mkstemp, int, (char * /*template*/) _GL_ARG_NONNULL ((1))); +_GL_CXXALIAS_RPL (mkstemp, int, (char * /*template*/)); +# else +# if ! @HAVE_MKSTEMP@ +_GL_FUNCDECL_SYS (mkstemp, int, (char * /*template*/) _GL_ARG_NONNULL ((1))); +# endif +_GL_CXXALIAS_SYS (mkstemp, int, (char * /*template*/)); +# endif +_GL_CXXALIASWARN (mkstemp); +#elif defined GNULIB_POSIXCHECK +# undef mkstemp +# if HAVE_RAW_DECL_MKSTEMP +_GL_WARN_ON_USE (mkstemp, "mkstemp is unportable - " + "use gnulib module mkstemp for portability"); +# endif +#endif + +#if @GNULIB_MKSTEMPS@ +/* Create a unique temporary file from TEMPLATE. + The last six characters of TEMPLATE prior to a suffix of length + SUFFIXLEN must be "XXXXXX"; + they are replaced with a string that makes the file name unique. + The file is then created, ensuring it didn't exist before. + The file is created read-write (mask at least 0600 & ~umask), but it may be + world-readable and world-writable (mask 0666 & ~umask), depending on the + implementation. + Returns the open file descriptor if successful, otherwise -1 and errno + set. */ +# if !@HAVE_MKSTEMPS@ +_GL_FUNCDECL_SYS (mkstemps, int, (char * /*template*/, int /*suffixlen*/) + _GL_ARG_NONNULL ((1))); +# endif +_GL_CXXALIAS_SYS (mkstemps, int, (char * /*template*/, int /*suffixlen*/)); +_GL_CXXALIASWARN (mkstemps); +#elif defined GNULIB_POSIXCHECK +# undef mkstemps +# if HAVE_RAW_DECL_MKSTEMPS +_GL_WARN_ON_USE (mkstemps, "mkstemps is unportable - " + "use gnulib module mkstemps for portability"); +# endif +#endif + +#if @GNULIB_POSIX_OPENPT@ +/* Return an FD open to the master side of a pseudo-terminal. Flags should + include O_RDWR, and may also include O_NOCTTY. */ +# if !@HAVE_POSIX_OPENPT@ +_GL_FUNCDECL_SYS (posix_openpt, int, (int flags)); +# endif +_GL_CXXALIAS_SYS (posix_openpt, int, (int flags)); +_GL_CXXALIASWARN (posix_openpt); +#elif defined GNULIB_POSIXCHECK +# undef posix_openpt +# if HAVE_RAW_DECL_POSIX_OPENPT +_GL_WARN_ON_USE (posix_openpt, "posix_openpt is not portable - " + "use gnulib module posix_openpt for portability"); +# endif +#endif + +#if @GNULIB_PTSNAME@ +/* Return the pathname of the pseudo-terminal slave associated with + the master FD is open on, or NULL on errors. */ +# if @REPLACE_PTSNAME@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef ptsname +# define ptsname rpl_ptsname +# endif +_GL_FUNCDECL_RPL (ptsname, char *, (int fd)); +_GL_CXXALIAS_RPL (ptsname, char *, (int fd)); +# else +# if !@HAVE_PTSNAME@ +_GL_FUNCDECL_SYS (ptsname, char *, (int fd)); +# endif +_GL_CXXALIAS_SYS (ptsname, char *, (int fd)); +# endif +_GL_CXXALIASWARN (ptsname); +#elif defined GNULIB_POSIXCHECK +# undef ptsname +# if HAVE_RAW_DECL_PTSNAME +_GL_WARN_ON_USE (ptsname, "ptsname is not portable - " + "use gnulib module ptsname for portability"); +# endif +#endif + +#if @GNULIB_PTSNAME_R@ +/* Set the pathname of the pseudo-terminal slave associated with + the master FD is open on and return 0, or set errno and return + non-zero on errors. */ +# if @REPLACE_PTSNAME_R@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef ptsname_r +# define ptsname_r rpl_ptsname_r +# endif +_GL_FUNCDECL_RPL (ptsname_r, int, (int fd, char *buf, size_t len)); +_GL_CXXALIAS_RPL (ptsname_r, int, (int fd, char *buf, size_t len)); +# else +# if !@HAVE_PTSNAME_R@ +_GL_FUNCDECL_SYS (ptsname_r, int, (int fd, char *buf, size_t len)); +# endif +_GL_CXXALIAS_SYS (ptsname_r, int, (int fd, char *buf, size_t len)); +# endif +_GL_CXXALIASWARN (ptsname_r); +#elif defined GNULIB_POSIXCHECK +# undef ptsname_r +# if HAVE_RAW_DECL_PTSNAME_R +_GL_WARN_ON_USE (ptsname_r, "ptsname_r is not portable - " + "use gnulib module ptsname_r for portability"); +# endif +#endif + +#if @GNULIB_PUTENV@ +# if @REPLACE_PUTENV@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef putenv +# define putenv rpl_putenv +# endif +_GL_FUNCDECL_RPL (putenv, int, (char *string) _GL_ARG_NONNULL ((1))); +_GL_CXXALIAS_RPL (putenv, int, (char *string)); +# else +_GL_CXXALIAS_SYS (putenv, int, (char *string)); +# endif +_GL_CXXALIASWARN (putenv); +#endif + +#if @GNULIB_QSORT_R@ +/* Sort an array of NMEMB elements, starting at address BASE, each element + occupying SIZE bytes, in ascending order according to the comparison + function COMPARE. */ +# if @REPLACE_QSORT_R@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef qsort_r +# define qsort_r rpl_qsort_r +# endif +_GL_FUNCDECL_RPL (qsort_r, void, (void *base, size_t nmemb, size_t size, + int (*compare) (void const *, void const *, + void *), + void *arg) _GL_ARG_NONNULL ((1, 4))); +_GL_CXXALIAS_RPL (qsort_r, void, (void *base, size_t nmemb, size_t size, + int (*compare) (void const *, void const *, + void *), + void *arg)); +# else +# if !@HAVE_QSORT_R@ +_GL_FUNCDECL_SYS (qsort_r, void, (void *base, size_t nmemb, size_t size, + int (*compare) (void const *, void const *, + void *), + void *arg) _GL_ARG_NONNULL ((1, 4))); +# endif +_GL_CXXALIAS_SYS (qsort_r, void, (void *base, size_t nmemb, size_t size, + int (*compare) (void const *, void const *, + void *), + void *arg)); +# endif +_GL_CXXALIASWARN (qsort_r); +#elif defined GNULIB_POSIXCHECK +# undef qsort_r +# if HAVE_RAW_DECL_QSORT_R +_GL_WARN_ON_USE (qsort_r, "qsort_r is not portable - " + "use gnulib module qsort_r for portability"); +# endif +#endif + + +#if @GNULIB_RANDOM_R@ +# if !@HAVE_RANDOM_R@ +# ifndef RAND_MAX +# define RAND_MAX 2147483647 +# endif +# endif +#endif + + +#if @GNULIB_RANDOM@ +# if !@HAVE_RANDOM@ +_GL_FUNCDECL_SYS (random, long, (void)); +# endif +_GL_CXXALIAS_SYS (random, long, (void)); +_GL_CXXALIASWARN (random); +#elif defined GNULIB_POSIXCHECK +# undef random +# if HAVE_RAW_DECL_RANDOM +_GL_WARN_ON_USE (random, "random is unportable - " + "use gnulib module random for portability"); +# endif +#endif + +#if @GNULIB_RANDOM@ +# if !@HAVE_RANDOM@ +_GL_FUNCDECL_SYS (srandom, void, (unsigned int seed)); +# endif +_GL_CXXALIAS_SYS (srandom, void, (unsigned int seed)); +_GL_CXXALIASWARN (srandom); +#elif defined GNULIB_POSIXCHECK +# undef srandom +# if HAVE_RAW_DECL_SRANDOM +_GL_WARN_ON_USE (srandom, "srandom is unportable - " + "use gnulib module random for portability"); +# endif +#endif + +#if @GNULIB_RANDOM@ +# if !@HAVE_RANDOM@ || !@HAVE_DECL_INITSTATE@ +_GL_FUNCDECL_SYS (initstate, char *, + (unsigned int seed, char *buf, size_t buf_size) + _GL_ARG_NONNULL ((2))); +# endif +_GL_CXXALIAS_SYS (initstate, char *, + (unsigned int seed, char *buf, size_t buf_size)); +_GL_CXXALIASWARN (initstate); +#elif defined GNULIB_POSIXCHECK +# undef initstate +# if HAVE_RAW_DECL_INITSTATE_R +_GL_WARN_ON_USE (initstate, "initstate is unportable - " + "use gnulib module random for portability"); +# endif +#endif + +#if @GNULIB_RANDOM@ +# if !@HAVE_RANDOM@ || !@HAVE_DECL_SETSTATE@ +_GL_FUNCDECL_SYS (setstate, char *, (char *arg_state) _GL_ARG_NONNULL ((1))); +# endif +_GL_CXXALIAS_SYS (setstate, char *, (char *arg_state)); +_GL_CXXALIASWARN (setstate); +#elif defined GNULIB_POSIXCHECK +# undef setstate +# if HAVE_RAW_DECL_SETSTATE_R +_GL_WARN_ON_USE (setstate, "setstate is unportable - " + "use gnulib module random for portability"); +# endif +#endif + + +#if @GNULIB_RANDOM_R@ +# if @REPLACE_RANDOM_R@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef random_r +# define random_r rpl_random_r +# endif +_GL_FUNCDECL_RPL (random_r, int, (struct random_data *buf, int32_t *result) + _GL_ARG_NONNULL ((1, 2))); +_GL_CXXALIAS_RPL (random_r, int, (struct random_data *buf, int32_t *result)); +# else +# if !@HAVE_RANDOM_R@ +_GL_FUNCDECL_SYS (random_r, int, (struct random_data *buf, int32_t *result) + _GL_ARG_NONNULL ((1, 2))); +# endif +_GL_CXXALIAS_SYS (random_r, int, (struct random_data *buf, int32_t *result)); +# endif +_GL_CXXALIASWARN (random_r); +#elif defined GNULIB_POSIXCHECK +# undef random_r +# if HAVE_RAW_DECL_RANDOM_R +_GL_WARN_ON_USE (random_r, "random_r is unportable - " + "use gnulib module random_r for portability"); +# endif +#endif + +#if @GNULIB_RANDOM_R@ +# if @REPLACE_RANDOM_R@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef srandom_r +# define srandom_r rpl_srandom_r +# endif +_GL_FUNCDECL_RPL (srandom_r, int, + (unsigned int seed, struct random_data *rand_state) + _GL_ARG_NONNULL ((2))); +_GL_CXXALIAS_RPL (srandom_r, int, + (unsigned int seed, struct random_data *rand_state)); +# else +# if !@HAVE_RANDOM_R@ +_GL_FUNCDECL_SYS (srandom_r, int, + (unsigned int seed, struct random_data *rand_state) + _GL_ARG_NONNULL ((2))); +# endif +_GL_CXXALIAS_SYS (srandom_r, int, + (unsigned int seed, struct random_data *rand_state)); +# endif +_GL_CXXALIASWARN (srandom_r); +#elif defined GNULIB_POSIXCHECK +# undef srandom_r +# if HAVE_RAW_DECL_SRANDOM_R +_GL_WARN_ON_USE (srandom_r, "srandom_r is unportable - " + "use gnulib module random_r for portability"); +# endif +#endif + +#if @GNULIB_RANDOM_R@ +# if @REPLACE_RANDOM_R@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef initstate_r +# define initstate_r rpl_initstate_r +# endif +_GL_FUNCDECL_RPL (initstate_r, int, + (unsigned int seed, char *buf, size_t buf_size, + struct random_data *rand_state) + _GL_ARG_NONNULL ((2, 4))); +_GL_CXXALIAS_RPL (initstate_r, int, + (unsigned int seed, char *buf, size_t buf_size, + struct random_data *rand_state)); +# else +# if !@HAVE_RANDOM_R@ +_GL_FUNCDECL_SYS (initstate_r, int, + (unsigned int seed, char *buf, size_t buf_size, + struct random_data *rand_state) + _GL_ARG_NONNULL ((2, 4))); +# endif +_GL_CXXALIAS_SYS (initstate_r, int, + (unsigned int seed, char *buf, size_t buf_size, + struct random_data *rand_state)); +# endif +_GL_CXXALIASWARN (initstate_r); +#elif defined GNULIB_POSIXCHECK +# undef initstate_r +# if HAVE_RAW_DECL_INITSTATE_R +_GL_WARN_ON_USE (initstate_r, "initstate_r is unportable - " + "use gnulib module random_r for portability"); +# endif +#endif + +#if @GNULIB_RANDOM_R@ +# if @REPLACE_RANDOM_R@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef setstate_r +# define setstate_r rpl_setstate_r +# endif +_GL_FUNCDECL_RPL (setstate_r, int, + (char *arg_state, struct random_data *rand_state) + _GL_ARG_NONNULL ((1, 2))); +_GL_CXXALIAS_RPL (setstate_r, int, + (char *arg_state, struct random_data *rand_state)); +# else +# if !@HAVE_RANDOM_R@ +_GL_FUNCDECL_SYS (setstate_r, int, + (char *arg_state, struct random_data *rand_state) + _GL_ARG_NONNULL ((1, 2))); +# endif +_GL_CXXALIAS_SYS (setstate_r, int, + (char *arg_state, struct random_data *rand_state)); +# endif +_GL_CXXALIASWARN (setstate_r); +#elif defined GNULIB_POSIXCHECK +# undef setstate_r +# if HAVE_RAW_DECL_SETSTATE_R +_GL_WARN_ON_USE (setstate_r, "setstate_r is unportable - " + "use gnulib module random_r for portability"); +# endif +#endif + + +#if @GNULIB_REALLOC_POSIX@ +# if @REPLACE_REALLOC@ +# if !((defined __cplusplus && defined GNULIB_NAMESPACE) \ + || _GL_USE_STDLIB_ALLOC) +# undef realloc +# define realloc rpl_realloc +# endif +_GL_FUNCDECL_RPL (realloc, void *, (void *ptr, size_t size)); +_GL_CXXALIAS_RPL (realloc, void *, (void *ptr, size_t size)); +# else +_GL_CXXALIAS_SYS (realloc, void *, (void *ptr, size_t size)); +# endif +_GL_CXXALIASWARN (realloc); +#elif defined GNULIB_POSIXCHECK && !_GL_USE_STDLIB_ALLOC +# undef realloc +/* Assume realloc is always declared. */ +_GL_WARN_ON_USE (realloc, "realloc is not POSIX compliant everywhere - " + "use gnulib module realloc-posix for portability"); +#endif + + +#if @GNULIB_REALLOCARRAY@ +# if ! @HAVE_REALLOCARRAY@ +_GL_FUNCDECL_SYS (reallocarray, void *, + (void *ptr, size_t nmemb, size_t size)); +# endif +_GL_CXXALIAS_SYS (reallocarray, void *, + (void *ptr, size_t nmemb, size_t size)); +_GL_CXXALIASWARN (reallocarray); +#elif defined GNULIB_POSIXCHECK +# undef reallocarray +# if HAVE_RAW_DECL_REALLOCARRAY +_GL_WARN_ON_USE (reallocarray, "reallocarray is not portable - " + "use gnulib module reallocarray for portability"); +# endif +#endif + +#if @GNULIB_REALPATH@ +# if @REPLACE_REALPATH@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# define realpath rpl_realpath +# endif +_GL_FUNCDECL_RPL (realpath, char *, (const char *name, char *resolved) + _GL_ARG_NONNULL ((1))); +_GL_CXXALIAS_RPL (realpath, char *, (const char *name, char *resolved)); +# else +# if !@HAVE_REALPATH@ +_GL_FUNCDECL_SYS (realpath, char *, (const char *name, char *resolved) + _GL_ARG_NONNULL ((1))); +# endif +_GL_CXXALIAS_SYS (realpath, char *, (const char *name, char *resolved)); +# endif +_GL_CXXALIASWARN (realpath); +#elif defined GNULIB_POSIXCHECK +# undef realpath +# if HAVE_RAW_DECL_REALPATH +_GL_WARN_ON_USE (realpath, "realpath is unportable - use gnulib module " + "canonicalize or canonicalize-lgpl for portability"); +# endif +#endif + +#if @GNULIB_RPMATCH@ +/* Test a user response to a question. + Return 1 if it is affirmative, 0 if it is negative, or -1 if not clear. */ +# if !@HAVE_RPMATCH@ +_GL_FUNCDECL_SYS (rpmatch, int, (const char *response) _GL_ARG_NONNULL ((1))); +# endif +_GL_CXXALIAS_SYS (rpmatch, int, (const char *response)); +_GL_CXXALIASWARN (rpmatch); +#elif defined GNULIB_POSIXCHECK +# undef rpmatch +# if HAVE_RAW_DECL_RPMATCH +_GL_WARN_ON_USE (rpmatch, "rpmatch is unportable - " + "use gnulib module rpmatch for portability"); +# endif +#endif + +#if @GNULIB_SECURE_GETENV@ +/* Look up NAME in the environment, returning 0 in insecure situations. */ +# if !@HAVE_SECURE_GETENV@ +_GL_FUNCDECL_SYS (secure_getenv, char *, + (char const *name) _GL_ARG_NONNULL ((1))); +# endif +_GL_CXXALIAS_SYS (secure_getenv, char *, (char const *name)); +_GL_CXXALIASWARN (secure_getenv); +#elif defined GNULIB_POSIXCHECK +# undef secure_getenv +# if HAVE_RAW_DECL_SECURE_GETENV +_GL_WARN_ON_USE (secure_getenv, "secure_getenv is unportable - " + "use gnulib module secure_getenv for portability"); +# endif +#endif + +#if @GNULIB_SETENV@ +/* Set NAME to VALUE in the environment. + If REPLACE is nonzero, overwrite an existing value. */ +# if @REPLACE_SETENV@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef setenv +# define setenv rpl_setenv +# endif +_GL_FUNCDECL_RPL (setenv, int, + (const char *name, const char *value, int replace) + _GL_ARG_NONNULL ((1))); +_GL_CXXALIAS_RPL (setenv, int, + (const char *name, const char *value, int replace)); +# else +# if !@HAVE_DECL_SETENV@ +_GL_FUNCDECL_SYS (setenv, int, + (const char *name, const char *value, int replace) + _GL_ARG_NONNULL ((1))); +# endif +_GL_CXXALIAS_SYS (setenv, int, + (const char *name, const char *value, int replace)); +# endif +# if !(@REPLACE_SETENV@ && !@HAVE_DECL_SETENV@) +_GL_CXXALIASWARN (setenv); +# endif +#elif defined GNULIB_POSIXCHECK +# undef setenv +# if HAVE_RAW_DECL_SETENV +_GL_WARN_ON_USE (setenv, "setenv is unportable - " + "use gnulib module setenv for portability"); +# endif +#endif + +#if @GNULIB_STRTOD@ + /* Parse a double from STRING, updating ENDP if appropriate. */ +# if @REPLACE_STRTOD@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# define strtod rpl_strtod +# endif +_GL_FUNCDECL_RPL (strtod, double, (const char *str, char **endp) + _GL_ARG_NONNULL ((1))); +_GL_CXXALIAS_RPL (strtod, double, (const char *str, char **endp)); +# else +# if !@HAVE_STRTOD@ +_GL_FUNCDECL_SYS (strtod, double, (const char *str, char **endp) + _GL_ARG_NONNULL ((1))); +# endif +_GL_CXXALIAS_SYS (strtod, double, (const char *str, char **endp)); +# endif +_GL_CXXALIASWARN (strtod); +#elif defined GNULIB_POSIXCHECK +# undef strtod +# if HAVE_RAW_DECL_STRTOD +_GL_WARN_ON_USE (strtod, "strtod is unportable - " + "use gnulib module strtod for portability"); +# endif +#endif + +#if @GNULIB_STRTOLL@ +/* Parse a signed integer whose textual representation starts at STRING. + The integer is expected to be in base BASE (2 <= BASE <= 36); if BASE == 0, + it may be decimal or octal (with prefix "0") or hexadecimal (with prefix + "0x"). + If ENDPTR is not NULL, the address of the first byte after the integer is + stored in *ENDPTR. + Upon overflow, the return value is LLONG_MAX or LLONG_MIN, and errno is set + to ERANGE. */ +# if !@HAVE_STRTOLL@ +_GL_FUNCDECL_SYS (strtoll, long long, + (const char *string, char **endptr, int base) + _GL_ARG_NONNULL ((1))); +# endif +_GL_CXXALIAS_SYS (strtoll, long long, + (const char *string, char **endptr, int base)); +_GL_CXXALIASWARN (strtoll); +#elif defined GNULIB_POSIXCHECK +# undef strtoll +# if HAVE_RAW_DECL_STRTOLL +_GL_WARN_ON_USE (strtoll, "strtoll is unportable - " + "use gnulib module strtoll for portability"); +# endif +#endif + +#if @GNULIB_STRTOULL@ +/* Parse an unsigned integer whose textual representation starts at STRING. + The integer is expected to be in base BASE (2 <= BASE <= 36); if BASE == 0, + it may be decimal or octal (with prefix "0") or hexadecimal (with prefix + "0x"). + If ENDPTR is not NULL, the address of the first byte after the integer is + stored in *ENDPTR. + Upon overflow, the return value is ULLONG_MAX, and errno is set to + ERANGE. */ +# if !@HAVE_STRTOULL@ +_GL_FUNCDECL_SYS (strtoull, unsigned long long, + (const char *string, char **endptr, int base) + _GL_ARG_NONNULL ((1))); +# endif +_GL_CXXALIAS_SYS (strtoull, unsigned long long, + (const char *string, char **endptr, int base)); +_GL_CXXALIASWARN (strtoull); +#elif defined GNULIB_POSIXCHECK +# undef strtoull +# if HAVE_RAW_DECL_STRTOULL +_GL_WARN_ON_USE (strtoull, "strtoull is unportable - " + "use gnulib module strtoull for portability"); +# endif +#endif + +#if @GNULIB_UNLOCKPT@ +/* Unlock the slave side of the pseudo-terminal whose master side is specified + by FD, so that it can be opened. */ +# if !@HAVE_UNLOCKPT@ +_GL_FUNCDECL_SYS (unlockpt, int, (int fd)); +# endif +_GL_CXXALIAS_SYS (unlockpt, int, (int fd)); +_GL_CXXALIASWARN (unlockpt); +#elif defined GNULIB_POSIXCHECK +# undef unlockpt +# if HAVE_RAW_DECL_UNLOCKPT +_GL_WARN_ON_USE (unlockpt, "unlockpt is not portable - " + "use gnulib module unlockpt for portability"); +# endif +#endif + +#if @GNULIB_UNSETENV@ +/* Remove the variable NAME from the environment. */ +# if @REPLACE_UNSETENV@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef unsetenv +# define unsetenv rpl_unsetenv +# endif +_GL_FUNCDECL_RPL (unsetenv, int, (const char *name) _GL_ARG_NONNULL ((1))); +_GL_CXXALIAS_RPL (unsetenv, int, (const char *name)); +# else +# if !@HAVE_DECL_UNSETENV@ +_GL_FUNCDECL_SYS (unsetenv, int, (const char *name) _GL_ARG_NONNULL ((1))); +# endif +_GL_CXXALIAS_SYS (unsetenv, int, (const char *name)); +# endif +# if !(@REPLACE_UNSETENV@ && !@HAVE_DECL_UNSETENV@) +_GL_CXXALIASWARN (unsetenv); +# endif +#elif defined GNULIB_POSIXCHECK +# undef unsetenv +# if HAVE_RAW_DECL_UNSETENV +_GL_WARN_ON_USE (unsetenv, "unsetenv is unportable - " + "use gnulib module unsetenv for portability"); +# endif +#endif + +/* Convert a wide character to a multibyte character. */ +#if @GNULIB_WCTOMB@ +# if @REPLACE_WCTOMB@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef wctomb +# define wctomb rpl_wctomb +# endif +_GL_FUNCDECL_RPL (wctomb, int, (char *s, wchar_t wc)); +_GL_CXXALIAS_RPL (wctomb, int, (char *s, wchar_t wc)); +# else +_GL_CXXALIAS_SYS (wctomb, int, (char *s, wchar_t wc)); +# endif +_GL_CXXALIASWARN (wctomb); +#endif + + +#endif /* _@GUARD_PREFIX@_STDLIB_H */ +#endif /* _@GUARD_PREFIX@_STDLIB_H */ +#endif diff --git a/lib/streq.h b/lib/streq.h new file mode 100644 index 0000000..bde1b95 --- /dev/null +++ b/lib/streq.h @@ -0,0 +1,176 @@ +/* Optimized string comparison. + Copyright (C) 2001-2002, 2007, 2009-2018 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify it + under the terms of the GNU General Public License as published + by the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +/* Written by Bruno Haible <bruno@clisp.org>. */ + +#ifndef _GL_STREQ_H +#define _GL_STREQ_H + +#include <string.h> + +/* STREQ_OPT allows to optimize string comparison with a small literal string. + STREQ_OPT (s, "EUC-KR", 'E', 'U', 'C', '-', 'K', 'R', 0, 0, 0) + is semantically equivalent to + strcmp (s, "EUC-KR") == 0 + just faster. */ + +/* Help GCC to generate good code for string comparisons with + immediate strings. */ +#if defined (__GNUC__) && defined (__OPTIMIZE__) + +static inline int +streq9 (const char *s1, const char *s2) +{ + return strcmp (s1 + 9, s2 + 9) == 0; +} + +static inline int +streq8 (const char *s1, const char *s2, char s28) +{ + if (s1[8] == s28) + { + if (s28 == 0) + return 1; + else + return streq9 (s1, s2); + } + else + return 0; +} + +static inline int +streq7 (const char *s1, const char *s2, char s27, char s28) +{ + if (s1[7] == s27) + { + if (s27 == 0) + return 1; + else + return streq8 (s1, s2, s28); + } + else + return 0; +} + +static inline int +streq6 (const char *s1, const char *s2, char s26, char s27, char s28) +{ + if (s1[6] == s26) + { + if (s26 == 0) + return 1; + else + return streq7 (s1, s2, s27, s28); + } + else + return 0; +} + +static inline int +streq5 (const char *s1, const char *s2, char s25, char s26, char s27, char s28) +{ + if (s1[5] == s25) + { + if (s25 == 0) + return 1; + else + return streq6 (s1, s2, s26, s27, s28); + } + else + return 0; +} + +static inline int +streq4 (const char *s1, const char *s2, char s24, char s25, char s26, char s27, char s28) +{ + if (s1[4] == s24) + { + if (s24 == 0) + return 1; + else + return streq5 (s1, s2, s25, s26, s27, s28); + } + else + return 0; +} + +static inline int +streq3 (const char *s1, const char *s2, char s23, char s24, char s25, char s26, char s27, char s28) +{ + if (s1[3] == s23) + { + if (s23 == 0) + return 1; + else + return streq4 (s1, s2, s24, s25, s26, s27, s28); + } + else + return 0; +} + +static inline int +streq2 (const char *s1, const char *s2, char s22, char s23, char s24, char s25, char s26, char s27, char s28) +{ + if (s1[2] == s22) + { + if (s22 == 0) + return 1; + else + return streq3 (s1, s2, s23, s24, s25, s26, s27, s28); + } + else + return 0; +} + +static inline int +streq1 (const char *s1, const char *s2, char s21, char s22, char s23, char s24, char s25, char s26, char s27, char s28) +{ + if (s1[1] == s21) + { + if (s21 == 0) + return 1; + else + return streq2 (s1, s2, s22, s23, s24, s25, s26, s27, s28); + } + else + return 0; +} + +static inline int +streq0 (const char *s1, const char *s2, char s20, char s21, char s22, char s23, char s24, char s25, char s26, char s27, char s28) +{ + if (s1[0] == s20) + { + if (s20 == 0) + return 1; + else + return streq1 (s1, s2, s21, s22, s23, s24, s25, s26, s27, s28); + } + else + return 0; +} + +#define STREQ_OPT(s1,s2,s20,s21,s22,s23,s24,s25,s26,s27,s28) \ + streq0 (s1, s2, s20, s21, s22, s23, s24, s25, s26, s27, s28) + +#else + +#define STREQ_OPT(s1,s2,s20,s21,s22,s23,s24,s25,s26,s27,s28) \ + (strcmp (s1, s2) == 0) + +#endif + +#endif /* _GL_STREQ_H */ diff --git a/lib/strerror-override.c b/lib/strerror-override.c new file mode 100644 index 0000000..d3ad21e --- /dev/null +++ b/lib/strerror-override.c @@ -0,0 +1,302 @@ +/* strerror-override.c --- POSIX compatible system error routine + + Copyright (C) 2010-2018 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +/* Written by Bruno Haible <bruno@clisp.org>, 2010. */ + +#include <config.h> + +#include "strerror-override.h" + +#include <errno.h> + +#if GNULIB_defined_EWINSOCK /* native Windows platforms */ +# if HAVE_WINSOCK2_H +# include <winsock2.h> +# endif +#endif + +/* If ERRNUM maps to an errno value defined by gnulib, return a string + describing the error. Otherwise return NULL. */ +const char * +strerror_override (int errnum) +{ + /* These error messages are taken from glibc/sysdeps/gnu/errlist.c. */ + switch (errnum) + { +#if REPLACE_STRERROR_0 + case 0: + return "Success"; +#endif + +#if GNULIB_defined_ESOCK /* native Windows platforms with older <errno.h> */ + case EINPROGRESS: + return "Operation now in progress"; + case EALREADY: + return "Operation already in progress"; + case ENOTSOCK: + return "Socket operation on non-socket"; + case EDESTADDRREQ: + return "Destination address required"; + case EMSGSIZE: + return "Message too long"; + case EPROTOTYPE: + return "Protocol wrong type for socket"; + case ENOPROTOOPT: + return "Protocol not available"; + case EPROTONOSUPPORT: + return "Protocol not supported"; + case EOPNOTSUPP: + return "Operation not supported"; + case EAFNOSUPPORT: + return "Address family not supported by protocol"; + case EADDRINUSE: + return "Address already in use"; + case EADDRNOTAVAIL: + return "Cannot assign requested address"; + case ENETDOWN: + return "Network is down"; + case ENETUNREACH: + return "Network is unreachable"; + case ECONNRESET: + return "Connection reset by peer"; + case ENOBUFS: + return "No buffer space available"; + case EISCONN: + return "Transport endpoint is already connected"; + case ENOTCONN: + return "Transport endpoint is not connected"; + case ETIMEDOUT: + return "Connection timed out"; + case ECONNREFUSED: + return "Connection refused"; + case ELOOP: + return "Too many levels of symbolic links"; + case EHOSTUNREACH: + return "No route to host"; + case EWOULDBLOCK: + return "Operation would block"; +#endif +#if GNULIB_defined_ESTREAMS /* native Windows platforms with older <errno.h> */ + case ETXTBSY: + return "Text file busy"; + case ENODATA: + return "No data available"; + case ENOSR: + return "Out of streams resources"; + case ENOSTR: + return "Device not a stream"; + case ETIME: + return "Timer expired"; + case EOTHER: + return "Other error"; +#endif +#if GNULIB_defined_EWINSOCK /* native Windows platforms */ + case ESOCKTNOSUPPORT: + return "Socket type not supported"; + case EPFNOSUPPORT: + return "Protocol family not supported"; + case ESHUTDOWN: + return "Cannot send after transport endpoint shutdown"; + case ETOOMANYREFS: + return "Too many references: cannot splice"; + case EHOSTDOWN: + return "Host is down"; + case EPROCLIM: + return "Too many processes"; + case EUSERS: + return "Too many users"; + case EDQUOT: + return "Disk quota exceeded"; + case ESTALE: + return "Stale NFS file handle"; + case EREMOTE: + return "Object is remote"; +# if HAVE_WINSOCK2_H + /* WSA_INVALID_HANDLE maps to EBADF */ + /* WSA_NOT_ENOUGH_MEMORY maps to ENOMEM */ + /* WSA_INVALID_PARAMETER maps to EINVAL */ + case WSA_OPERATION_ABORTED: + return "Overlapped operation aborted"; + case WSA_IO_INCOMPLETE: + return "Overlapped I/O event object not in signaled state"; + case WSA_IO_PENDING: + return "Overlapped operations will complete later"; + /* WSAEINTR maps to EINTR */ + /* WSAEBADF maps to EBADF */ + /* WSAEACCES maps to EACCES */ + /* WSAEFAULT maps to EFAULT */ + /* WSAEINVAL maps to EINVAL */ + /* WSAEMFILE maps to EMFILE */ + /* WSAEWOULDBLOCK maps to EWOULDBLOCK */ + /* WSAEINPROGRESS maps to EINPROGRESS */ + /* WSAEALREADY maps to EALREADY */ + /* WSAENOTSOCK maps to ENOTSOCK */ + /* WSAEDESTADDRREQ maps to EDESTADDRREQ */ + /* WSAEMSGSIZE maps to EMSGSIZE */ + /* WSAEPROTOTYPE maps to EPROTOTYPE */ + /* WSAENOPROTOOPT maps to ENOPROTOOPT */ + /* WSAEPROTONOSUPPORT maps to EPROTONOSUPPORT */ + /* WSAESOCKTNOSUPPORT is ESOCKTNOSUPPORT */ + /* WSAEOPNOTSUPP maps to EOPNOTSUPP */ + /* WSAEPFNOSUPPORT is EPFNOSUPPORT */ + /* WSAEAFNOSUPPORT maps to EAFNOSUPPORT */ + /* WSAEADDRINUSE maps to EADDRINUSE */ + /* WSAEADDRNOTAVAIL maps to EADDRNOTAVAIL */ + /* WSAENETDOWN maps to ENETDOWN */ + /* WSAENETUNREACH maps to ENETUNREACH */ + /* WSAENETRESET maps to ENETRESET */ + /* WSAECONNABORTED maps to ECONNABORTED */ + /* WSAECONNRESET maps to ECONNRESET */ + /* WSAENOBUFS maps to ENOBUFS */ + /* WSAEISCONN maps to EISCONN */ + /* WSAENOTCONN maps to ENOTCONN */ + /* WSAESHUTDOWN is ESHUTDOWN */ + /* WSAETOOMANYREFS is ETOOMANYREFS */ + /* WSAETIMEDOUT maps to ETIMEDOUT */ + /* WSAECONNREFUSED maps to ECONNREFUSED */ + /* WSAELOOP maps to ELOOP */ + /* WSAENAMETOOLONG maps to ENAMETOOLONG */ + /* WSAEHOSTDOWN is EHOSTDOWN */ + /* WSAEHOSTUNREACH maps to EHOSTUNREACH */ + /* WSAENOTEMPTY maps to ENOTEMPTY */ + /* WSAEPROCLIM is EPROCLIM */ + /* WSAEUSERS is EUSERS */ + /* WSAEDQUOT is EDQUOT */ + /* WSAESTALE is ESTALE */ + /* WSAEREMOTE is EREMOTE */ + case WSASYSNOTREADY: + return "Network subsystem is unavailable"; + case WSAVERNOTSUPPORTED: + return "Winsock.dll version out of range"; + case WSANOTINITIALISED: + return "Successful WSAStartup not yet performed"; + case WSAEDISCON: + return "Graceful shutdown in progress"; + case WSAENOMORE: case WSA_E_NO_MORE: + return "No more results"; + case WSAECANCELLED: case WSA_E_CANCELLED: + return "Call was canceled"; + case WSAEINVALIDPROCTABLE: + return "Procedure call table is invalid"; + case WSAEINVALIDPROVIDER: + return "Service provider is invalid"; + case WSAEPROVIDERFAILEDINIT: + return "Service provider failed to initialize"; + case WSASYSCALLFAILURE: + return "System call failure"; + case WSASERVICE_NOT_FOUND: + return "Service not found"; + case WSATYPE_NOT_FOUND: + return "Class type not found"; + case WSAEREFUSED: + return "Database query was refused"; + case WSAHOST_NOT_FOUND: + return "Host not found"; + case WSATRY_AGAIN: + return "Nonauthoritative host not found"; + case WSANO_RECOVERY: + return "Nonrecoverable error"; + case WSANO_DATA: + return "Valid name, no data record of requested type"; + /* WSA_QOS_* omitted */ +# endif +#endif + +#if GNULIB_defined_ENOMSG + case ENOMSG: + return "No message of desired type"; +#endif + +#if GNULIB_defined_EIDRM + case EIDRM: + return "Identifier removed"; +#endif + +#if GNULIB_defined_ENOLINK + case ENOLINK: + return "Link has been severed"; +#endif + +#if GNULIB_defined_EPROTO + case EPROTO: + return "Protocol error"; +#endif + +#if GNULIB_defined_EMULTIHOP + case EMULTIHOP: + return "Multihop attempted"; +#endif + +#if GNULIB_defined_EBADMSG + case EBADMSG: + return "Bad message"; +#endif + +#if GNULIB_defined_EOVERFLOW + case EOVERFLOW: + return "Value too large for defined data type"; +#endif + +#if GNULIB_defined_ENOTSUP + case ENOTSUP: + return "Not supported"; +#endif + +#if GNULIB_defined_ENETRESET + case ENETRESET: + return "Network dropped connection on reset"; +#endif + +#if GNULIB_defined_ECONNABORTED + case ECONNABORTED: + return "Software caused connection abort"; +#endif + +#if GNULIB_defined_ESTALE + case ESTALE: + return "Stale NFS file handle"; +#endif + +#if GNULIB_defined_EDQUOT + case EDQUOT: + return "Disk quota exceeded"; +#endif + +#if GNULIB_defined_ECANCELED + case ECANCELED: + return "Operation canceled"; +#endif + +#if GNULIB_defined_EOWNERDEAD + case EOWNERDEAD: + return "Owner died"; +#endif + +#if GNULIB_defined_ENOTRECOVERABLE + case ENOTRECOVERABLE: + return "State not recoverable"; +#endif + +#if GNULIB_defined_EILSEQ + case EILSEQ: + return "Invalid or incomplete multibyte or wide character"; +#endif + + default: + return NULL; + } +} diff --git a/lib/strerror-override.h b/lib/strerror-override.h new file mode 100644 index 0000000..a96fd08 --- /dev/null +++ b/lib/strerror-override.h @@ -0,0 +1,56 @@ +/* strerror-override.h --- POSIX compatible system error routine + + Copyright (C) 2010-2018 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +#ifndef _GL_STRERROR_OVERRIDE_H +# define _GL_STRERROR_OVERRIDE_H + +# include <errno.h> +# include <stddef.h> + +/* Reasonable buffer size that should never trigger ERANGE; if this + proves too small, we intentionally abort(), to remind us to fix + this value. */ +# define STACKBUF_LEN 256 + +/* If ERRNUM maps to an errno value defined by gnulib, return a string + describing the error. Otherwise return NULL. */ +# if REPLACE_STRERROR_0 \ + || GNULIB_defined_ESOCK \ + || GNULIB_defined_ESTREAMS \ + || GNULIB_defined_EWINSOCK \ + || GNULIB_defined_ENOMSG \ + || GNULIB_defined_EIDRM \ + || GNULIB_defined_ENOLINK \ + || GNULIB_defined_EPROTO \ + || GNULIB_defined_EMULTIHOP \ + || GNULIB_defined_EBADMSG \ + || GNULIB_defined_EOVERFLOW \ + || GNULIB_defined_ENOTSUP \ + || GNULIB_defined_ENETRESET \ + || GNULIB_defined_ECONNABORTED \ + || GNULIB_defined_ESTALE \ + || GNULIB_defined_EDQUOT \ + || GNULIB_defined_ECANCELED \ + || GNULIB_defined_EOWNERDEAD \ + || GNULIB_defined_ENOTRECOVERABLE \ + || GNULIB_defined_EILSEQ +extern const char *strerror_override (int errnum) _GL_ATTRIBUTE_CONST; +# else +# define strerror_override(ignored) NULL +# endif + +#endif /* _GL_STRERROR_OVERRIDE_H */ diff --git a/lib/strerror.c b/lib/strerror.c new file mode 100644 index 0000000..36397bb --- /dev/null +++ b/lib/strerror.c @@ -0,0 +1,71 @@ +/* strerror.c --- POSIX compatible system error routine + + Copyright (C) 2007-2018 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +#include <config.h> + +/* Specification. */ +#include <string.h> + +#include <errno.h> +#include <stdio.h> +#include <stdlib.h> +#include <string.h> + +#include "intprops.h" +#include "strerror-override.h" +#include "verify.h" + +/* Use the system functions, not the gnulib overrides in this file. */ +#undef sprintf + +char * +strerror (int n) +#undef strerror +{ + static char buf[STACKBUF_LEN]; + size_t len; + + /* Cast away const, due to the historical signature of strerror; + callers should not be modifying the string. */ + const char *msg = strerror_override (n); + if (msg) + return (char *) msg; + + msg = strerror (n); + + /* Our strerror_r implementation might use the system's strerror + buffer, so all other clients of strerror have to see the error + copied into a buffer that we manage. This is not thread-safe, + even if the system strerror is, but portable programs shouldn't + be using strerror if they care about thread-safety. */ + if (!msg || !*msg) + { + static char const fmt[] = "Unknown error %d"; + verify (sizeof buf >= sizeof (fmt) + INT_STRLEN_BOUND (n)); + sprintf (buf, fmt, n); + errno = EINVAL; + return buf; + } + + /* Fix STACKBUF_LEN if this ever aborts. */ + len = strlen (msg); + if (sizeof buf <= len) + abort (); + + memcpy (buf, msg, len + 1); + return buf; +} diff --git a/lib/string.in.h b/lib/string.in.h new file mode 100644 index 0000000..29f4ba6 --- /dev/null +++ b/lib/string.in.h @@ -0,0 +1,1063 @@ +/* A GNU-like <string.h>. + + Copyright (C) 1995-1996, 2001-2018 Free Software Foundation, Inc. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3, or (at your option) + any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, see <https://www.gnu.org/licenses/>. */ + +#if __GNUC__ >= 3 +@PRAGMA_SYSTEM_HEADER@ +#endif +@PRAGMA_COLUMNS@ + +#if defined _GL_ALREADY_INCLUDING_STRING_H +/* Special invocation convention: + - On OS X/NetBSD we have a sequence of nested includes + <string.h> -> <strings.h> -> "string.h" + In this situation system _chk variants due to -D_FORTIFY_SOURCE + might be used after any replacements defined here. */ + +#@INCLUDE_NEXT@ @NEXT_STRING_H@ + +#else +/* Normal invocation convention. */ + +#ifndef _@GUARD_PREFIX@_STRING_H + +#define _GL_ALREADY_INCLUDING_STRING_H + +/* The include_next requires a split double-inclusion guard. */ +#@INCLUDE_NEXT@ @NEXT_STRING_H@ + +#undef _GL_ALREADY_INCLUDING_STRING_H + +#ifndef _@GUARD_PREFIX@_STRING_H +#define _@GUARD_PREFIX@_STRING_H + +/* NetBSD 5.0 mis-defines NULL. */ +#include <stddef.h> + +/* MirBSD defines mbslen as a macro. */ +#if @GNULIB_MBSLEN@ && defined __MirBSD__ +# include <wchar.h> +#endif + +/* The __attribute__ feature is available in gcc versions 2.5 and later. + The attribute __pure__ was added in gcc 2.96. */ +#if __GNUC__ > 2 || (__GNUC__ == 2 && __GNUC_MINOR__ >= 96) +# define _GL_ATTRIBUTE_PURE __attribute__ ((__pure__)) +#else +# define _GL_ATTRIBUTE_PURE /* empty */ +#endif + +/* NetBSD 5.0 declares strsignal in <unistd.h>, not in <string.h>. */ +/* But in any case avoid namespace pollution on glibc systems. */ +#if (@GNULIB_STRSIGNAL@ || defined GNULIB_POSIXCHECK) && defined __NetBSD__ \ + && ! defined __GLIBC__ +# include <unistd.h> +#endif + +/* The definitions of _GL_FUNCDECL_RPL etc. are copied here. */ + +/* The definition of _GL_ARG_NONNULL is copied here. */ + +/* The definition of _GL_WARN_ON_USE is copied here. */ + + +/* Clear a block of memory. The compiler will not delete a call to + this function, even if the block is dead after the call. */ +#if @GNULIB_EXPLICIT_BZERO@ +# if ! @HAVE_EXPLICIT_BZERO@ +_GL_FUNCDECL_SYS (explicit_bzero, void, + (void *__dest, size_t __n) _GL_ARG_NONNULL ((1))); +# endif +_GL_CXXALIAS_SYS (explicit_bzero, void, (void *__dest, size_t __n)); +_GL_CXXALIASWARN (explicit_bzero); +#elif defined GNULIB_POSIXCHECK +# undef explicit_bzero +# if HAVE_RAW_DECL_EXPLICIT_BZERO +_GL_WARN_ON_USE (explicit_bzero, "explicit_bzero is unportable - " + "use gnulib module explicit_bzero for portability"); +# endif +#endif + +/* Find the index of the least-significant set bit. */ +#if @GNULIB_FFSL@ +# if !@HAVE_FFSL@ +_GL_FUNCDECL_SYS (ffsl, int, (long int i)); +# endif +_GL_CXXALIAS_SYS (ffsl, int, (long int i)); +_GL_CXXALIASWARN (ffsl); +#elif defined GNULIB_POSIXCHECK +# undef ffsl +# if HAVE_RAW_DECL_FFSL +_GL_WARN_ON_USE (ffsl, "ffsl is not portable - use the ffsl module"); +# endif +#endif + + +/* Find the index of the least-significant set bit. */ +#if @GNULIB_FFSLL@ +# if !@HAVE_FFSLL@ +_GL_FUNCDECL_SYS (ffsll, int, (long long int i)); +# endif +_GL_CXXALIAS_SYS (ffsll, int, (long long int i)); +_GL_CXXALIASWARN (ffsll); +#elif defined GNULIB_POSIXCHECK +# undef ffsll +# if HAVE_RAW_DECL_FFSLL +_GL_WARN_ON_USE (ffsll, "ffsll is not portable - use the ffsll module"); +# endif +#endif + + +/* Return the first instance of C within N bytes of S, or NULL. */ +#if @GNULIB_MEMCHR@ +# if @REPLACE_MEMCHR@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# define memchr rpl_memchr +# endif +_GL_FUNCDECL_RPL (memchr, void *, (void const *__s, int __c, size_t __n) + _GL_ATTRIBUTE_PURE + _GL_ARG_NONNULL ((1))); +_GL_CXXALIAS_RPL (memchr, void *, (void const *__s, int __c, size_t __n)); +# else +# if ! @HAVE_MEMCHR@ +_GL_FUNCDECL_SYS (memchr, void *, (void const *__s, int __c, size_t __n) + _GL_ATTRIBUTE_PURE + _GL_ARG_NONNULL ((1))); +# endif + /* On some systems, this function is defined as an overloaded function: + extern "C" { const void * std::memchr (const void *, int, size_t); } + extern "C++" { void * std::memchr (void *, int, size_t); } */ +_GL_CXXALIAS_SYS_CAST2 (memchr, + void *, (void const *__s, int __c, size_t __n), + void const *, (void const *__s, int __c, size_t __n)); +# endif +# if ((__GLIBC__ == 2 && __GLIBC_MINOR__ >= 10) && !defined __UCLIBC__) \ + && (__GNUC__ > 4 || (__GNUC__ == 4 && __GNUC_MINOR__ >= 4)) +_GL_CXXALIASWARN1 (memchr, void *, (void *__s, int __c, size_t __n)); +_GL_CXXALIASWARN1 (memchr, void const *, + (void const *__s, int __c, size_t __n)); +# else +_GL_CXXALIASWARN (memchr); +# endif +#elif defined GNULIB_POSIXCHECK +# undef memchr +/* Assume memchr is always declared. */ +_GL_WARN_ON_USE (memchr, "memchr has platform-specific bugs - " + "use gnulib module memchr for portability" ); +#endif + +/* Return the first occurrence of NEEDLE in HAYSTACK. */ +#if @GNULIB_MEMMEM@ +# if @REPLACE_MEMMEM@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# define memmem rpl_memmem +# endif +_GL_FUNCDECL_RPL (memmem, void *, + (void const *__haystack, size_t __haystack_len, + void const *__needle, size_t __needle_len) + _GL_ATTRIBUTE_PURE + _GL_ARG_NONNULL ((1, 3))); +_GL_CXXALIAS_RPL (memmem, void *, + (void const *__haystack, size_t __haystack_len, + void const *__needle, size_t __needle_len)); +# else +# if ! @HAVE_DECL_MEMMEM@ +_GL_FUNCDECL_SYS (memmem, void *, + (void const *__haystack, size_t __haystack_len, + void const *__needle, size_t __needle_len) + _GL_ATTRIBUTE_PURE + _GL_ARG_NONNULL ((1, 3))); +# endif +_GL_CXXALIAS_SYS (memmem, void *, + (void const *__haystack, size_t __haystack_len, + void const *__needle, size_t __needle_len)); +# endif +_GL_CXXALIASWARN (memmem); +#elif defined GNULIB_POSIXCHECK +# undef memmem +# if HAVE_RAW_DECL_MEMMEM +_GL_WARN_ON_USE (memmem, "memmem is unportable and often quadratic - " + "use gnulib module memmem-simple for portability, " + "and module memmem for speed" ); +# endif +#endif + +/* Copy N bytes of SRC to DEST, return pointer to bytes after the + last written byte. */ +#if @GNULIB_MEMPCPY@ +# if ! @HAVE_MEMPCPY@ +_GL_FUNCDECL_SYS (mempcpy, void *, + (void *restrict __dest, void const *restrict __src, + size_t __n) + _GL_ARG_NONNULL ((1, 2))); +# endif +_GL_CXXALIAS_SYS (mempcpy, void *, + (void *restrict __dest, void const *restrict __src, + size_t __n)); +_GL_CXXALIASWARN (mempcpy); +#elif defined GNULIB_POSIXCHECK +# undef mempcpy +# if HAVE_RAW_DECL_MEMPCPY +_GL_WARN_ON_USE (mempcpy, "mempcpy is unportable - " + "use gnulib module mempcpy for portability"); +# endif +#endif + +/* Search backwards through a block for a byte (specified as an int). */ +#if @GNULIB_MEMRCHR@ +# if ! @HAVE_DECL_MEMRCHR@ +_GL_FUNCDECL_SYS (memrchr, void *, (void const *, int, size_t) + _GL_ATTRIBUTE_PURE + _GL_ARG_NONNULL ((1))); +# endif + /* On some systems, this function is defined as an overloaded function: + extern "C++" { const void * std::memrchr (const void *, int, size_t); } + extern "C++" { void * std::memrchr (void *, int, size_t); } */ +_GL_CXXALIAS_SYS_CAST2 (memrchr, + void *, (void const *, int, size_t), + void const *, (void const *, int, size_t)); +# if ((__GLIBC__ == 2 && __GLIBC_MINOR__ >= 10) && !defined __UCLIBC__) \ + && (__GNUC__ > 4 || (__GNUC__ == 4 && __GNUC_MINOR__ >= 4)) +_GL_CXXALIASWARN1 (memrchr, void *, (void *, int, size_t)); +_GL_CXXALIASWARN1 (memrchr, void const *, (void const *, int, size_t)); +# else +_GL_CXXALIASWARN (memrchr); +# endif +#elif defined GNULIB_POSIXCHECK +# undef memrchr +# if HAVE_RAW_DECL_MEMRCHR +_GL_WARN_ON_USE (memrchr, "memrchr is unportable - " + "use gnulib module memrchr for portability"); +# endif +#endif + +/* Find the first occurrence of C in S. More efficient than + memchr(S,C,N), at the expense of undefined behavior if C does not + occur within N bytes. */ +#if @GNULIB_RAWMEMCHR@ +# if ! @HAVE_RAWMEMCHR@ +_GL_FUNCDECL_SYS (rawmemchr, void *, (void const *__s, int __c_in) + _GL_ATTRIBUTE_PURE + _GL_ARG_NONNULL ((1))); +# endif + /* On some systems, this function is defined as an overloaded function: + extern "C++" { const void * std::rawmemchr (const void *, int); } + extern "C++" { void * std::rawmemchr (void *, int); } */ +_GL_CXXALIAS_SYS_CAST2 (rawmemchr, + void *, (void const *__s, int __c_in), + void const *, (void const *__s, int __c_in)); +# if ((__GLIBC__ == 2 && __GLIBC_MINOR__ >= 10) && !defined __UCLIBC__) \ + && (__GNUC__ > 4 || (__GNUC__ == 4 && __GNUC_MINOR__ >= 4)) +_GL_CXXALIASWARN1 (rawmemchr, void *, (void *__s, int __c_in)); +_GL_CXXALIASWARN1 (rawmemchr, void const *, (void const *__s, int __c_in)); +# else +_GL_CXXALIASWARN (rawmemchr); +# endif +#elif defined GNULIB_POSIXCHECK +# undef rawmemchr +# if HAVE_RAW_DECL_RAWMEMCHR +_GL_WARN_ON_USE (rawmemchr, "rawmemchr is unportable - " + "use gnulib module rawmemchr for portability"); +# endif +#endif + +/* Copy SRC to DST, returning the address of the terminating '\0' in DST. */ +#if @GNULIB_STPCPY@ +# if ! @HAVE_STPCPY@ +_GL_FUNCDECL_SYS (stpcpy, char *, + (char *restrict __dst, char const *restrict __src) + _GL_ARG_NONNULL ((1, 2))); +# endif +_GL_CXXALIAS_SYS (stpcpy, char *, + (char *restrict __dst, char const *restrict __src)); +_GL_CXXALIASWARN (stpcpy); +#elif defined GNULIB_POSIXCHECK +# undef stpcpy +# if HAVE_RAW_DECL_STPCPY +_GL_WARN_ON_USE (stpcpy, "stpcpy is unportable - " + "use gnulib module stpcpy for portability"); +# endif +#endif + +/* Copy no more than N bytes of SRC to DST, returning a pointer past the + last non-NUL byte written into DST. */ +#if @GNULIB_STPNCPY@ +# if @REPLACE_STPNCPY@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef stpncpy +# define stpncpy rpl_stpncpy +# endif +_GL_FUNCDECL_RPL (stpncpy, char *, + (char *restrict __dst, char const *restrict __src, + size_t __n) + _GL_ARG_NONNULL ((1, 2))); +_GL_CXXALIAS_RPL (stpncpy, char *, + (char *restrict __dst, char const *restrict __src, + size_t __n)); +# else +# if ! @HAVE_STPNCPY@ +_GL_FUNCDECL_SYS (stpncpy, char *, + (char *restrict __dst, char const *restrict __src, + size_t __n) + _GL_ARG_NONNULL ((1, 2))); +# endif +_GL_CXXALIAS_SYS (stpncpy, char *, + (char *restrict __dst, char const *restrict __src, + size_t __n)); +# endif +_GL_CXXALIASWARN (stpncpy); +#elif defined GNULIB_POSIXCHECK +# undef stpncpy +# if HAVE_RAW_DECL_STPNCPY +_GL_WARN_ON_USE (stpncpy, "stpncpy is unportable - " + "use gnulib module stpncpy for portability"); +# endif +#endif + +#if defined GNULIB_POSIXCHECK +/* strchr() does not work with multibyte strings if the locale encoding is + GB18030 and the character to be searched is a digit. */ +# undef strchr +/* Assume strchr is always declared. */ +_GL_WARN_ON_USE (strchr, "strchr cannot work correctly on character strings " + "in some multibyte locales - " + "use mbschr if you care about internationalization"); +#endif + +/* Find the first occurrence of C in S or the final NUL byte. */ +#if @GNULIB_STRCHRNUL@ +# if @REPLACE_STRCHRNUL@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# define strchrnul rpl_strchrnul +# endif +_GL_FUNCDECL_RPL (strchrnul, char *, (const char *__s, int __c_in) + _GL_ATTRIBUTE_PURE + _GL_ARG_NONNULL ((1))); +_GL_CXXALIAS_RPL (strchrnul, char *, + (const char *str, int ch)); +# else +# if ! @HAVE_STRCHRNUL@ +_GL_FUNCDECL_SYS (strchrnul, char *, (char const *__s, int __c_in) + _GL_ATTRIBUTE_PURE + _GL_ARG_NONNULL ((1))); +# endif + /* On some systems, this function is defined as an overloaded function: + extern "C++" { const char * std::strchrnul (const char *, int); } + extern "C++" { char * std::strchrnul (char *, int); } */ +_GL_CXXALIAS_SYS_CAST2 (strchrnul, + char *, (char const *__s, int __c_in), + char const *, (char const *__s, int __c_in)); +# endif +# if ((__GLIBC__ == 2 && __GLIBC_MINOR__ >= 10) && !defined __UCLIBC__) \ + && (__GNUC__ > 4 || (__GNUC__ == 4 && __GNUC_MINOR__ >= 4)) +_GL_CXXALIASWARN1 (strchrnul, char *, (char *__s, int __c_in)); +_GL_CXXALIASWARN1 (strchrnul, char const *, (char const *__s, int __c_in)); +# else +_GL_CXXALIASWARN (strchrnul); +# endif +#elif defined GNULIB_POSIXCHECK +# undef strchrnul +# if HAVE_RAW_DECL_STRCHRNUL +_GL_WARN_ON_USE (strchrnul, "strchrnul is unportable - " + "use gnulib module strchrnul for portability"); +# endif +#endif + +/* Duplicate S, returning an identical malloc'd string. */ +#if @GNULIB_STRDUP@ +# if @REPLACE_STRDUP@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef strdup +# define strdup rpl_strdup +# endif +_GL_FUNCDECL_RPL (strdup, char *, (char const *__s) _GL_ARG_NONNULL ((1))); +_GL_CXXALIAS_RPL (strdup, char *, (char const *__s)); +# else +# if defined __cplusplus && defined GNULIB_NAMESPACE && defined strdup + /* strdup exists as a function and as a macro. Get rid of the macro. */ +# undef strdup +# endif +# if !(@HAVE_DECL_STRDUP@ || defined strdup) +_GL_FUNCDECL_SYS (strdup, char *, (char const *__s) _GL_ARG_NONNULL ((1))); +# endif +_GL_CXXALIAS_SYS (strdup, char *, (char const *__s)); +# endif +_GL_CXXALIASWARN (strdup); +#elif defined GNULIB_POSIXCHECK +# undef strdup +# if HAVE_RAW_DECL_STRDUP +_GL_WARN_ON_USE (strdup, "strdup is unportable - " + "use gnulib module strdup for portability"); +# endif +#endif + +/* Append no more than N characters from SRC onto DEST. */ +#if @GNULIB_STRNCAT@ +# if @REPLACE_STRNCAT@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef strncat +# define strncat rpl_strncat +# endif +_GL_FUNCDECL_RPL (strncat, char *, (char *dest, const char *src, size_t n) + _GL_ARG_NONNULL ((1, 2))); +_GL_CXXALIAS_RPL (strncat, char *, (char *dest, const char *src, size_t n)); +# else +_GL_CXXALIAS_SYS (strncat, char *, (char *dest, const char *src, size_t n)); +# endif +_GL_CXXALIASWARN (strncat); +#elif defined GNULIB_POSIXCHECK +# undef strncat +# if HAVE_RAW_DECL_STRNCAT +_GL_WARN_ON_USE (strncat, "strncat is unportable - " + "use gnulib module strncat for portability"); +# endif +#endif + +/* Return a newly allocated copy of at most N bytes of STRING. */ +#if @GNULIB_STRNDUP@ +# if @REPLACE_STRNDUP@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef strndup +# define strndup rpl_strndup +# endif +_GL_FUNCDECL_RPL (strndup, char *, (char const *__s, size_t __n) + _GL_ARG_NONNULL ((1))); +_GL_CXXALIAS_RPL (strndup, char *, (char const *__s, size_t __n)); +# else +# if ! @HAVE_DECL_STRNDUP@ +_GL_FUNCDECL_SYS (strndup, char *, (char const *__s, size_t __n) + _GL_ARG_NONNULL ((1))); +# endif +_GL_CXXALIAS_SYS (strndup, char *, (char const *__s, size_t __n)); +# endif +_GL_CXXALIASWARN (strndup); +#elif defined GNULIB_POSIXCHECK +# undef strndup +# if HAVE_RAW_DECL_STRNDUP +_GL_WARN_ON_USE (strndup, "strndup is unportable - " + "use gnulib module strndup for portability"); +# endif +#endif + +/* Find the length (number of bytes) of STRING, but scan at most + MAXLEN bytes. If no '\0' terminator is found in that many bytes, + return MAXLEN. */ +#if @GNULIB_STRNLEN@ +# if @REPLACE_STRNLEN@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef strnlen +# define strnlen rpl_strnlen +# endif +_GL_FUNCDECL_RPL (strnlen, size_t, (char const *__s, size_t __maxlen) + _GL_ATTRIBUTE_PURE + _GL_ARG_NONNULL ((1))); +_GL_CXXALIAS_RPL (strnlen, size_t, (char const *__s, size_t __maxlen)); +# else +# if ! @HAVE_DECL_STRNLEN@ +_GL_FUNCDECL_SYS (strnlen, size_t, (char const *__s, size_t __maxlen) + _GL_ATTRIBUTE_PURE + _GL_ARG_NONNULL ((1))); +# endif +_GL_CXXALIAS_SYS (strnlen, size_t, (char const *__s, size_t __maxlen)); +# endif +_GL_CXXALIASWARN (strnlen); +#elif defined GNULIB_POSIXCHECK +# undef strnlen +# if HAVE_RAW_DECL_STRNLEN +_GL_WARN_ON_USE (strnlen, "strnlen is unportable - " + "use gnulib module strnlen for portability"); +# endif +#endif + +#if defined GNULIB_POSIXCHECK +/* strcspn() assumes the second argument is a list of single-byte characters. + Even in this simple case, it does not work with multibyte strings if the + locale encoding is GB18030 and one of the characters to be searched is a + digit. */ +# undef strcspn +/* Assume strcspn is always declared. */ +_GL_WARN_ON_USE (strcspn, "strcspn cannot work correctly on character strings " + "in multibyte locales - " + "use mbscspn if you care about internationalization"); +#endif + +/* Find the first occurrence in S of any character in ACCEPT. */ +#if @GNULIB_STRPBRK@ +# if ! @HAVE_STRPBRK@ +_GL_FUNCDECL_SYS (strpbrk, char *, (char const *__s, char const *__accept) + _GL_ATTRIBUTE_PURE + _GL_ARG_NONNULL ((1, 2))); +# endif + /* On some systems, this function is defined as an overloaded function: + extern "C" { const char * strpbrk (const char *, const char *); } + extern "C++" { char * strpbrk (char *, const char *); } */ +_GL_CXXALIAS_SYS_CAST2 (strpbrk, + char *, (char const *__s, char const *__accept), + const char *, (char const *__s, char const *__accept)); +# if ((__GLIBC__ == 2 && __GLIBC_MINOR__ >= 10) && !defined __UCLIBC__) \ + && (__GNUC__ > 4 || (__GNUC__ == 4 && __GNUC_MINOR__ >= 4)) +_GL_CXXALIASWARN1 (strpbrk, char *, (char *__s, char const *__accept)); +_GL_CXXALIASWARN1 (strpbrk, char const *, + (char const *__s, char const *__accept)); +# else +_GL_CXXALIASWARN (strpbrk); +# endif +# if defined GNULIB_POSIXCHECK +/* strpbrk() assumes the second argument is a list of single-byte characters. + Even in this simple case, it does not work with multibyte strings if the + locale encoding is GB18030 and one of the characters to be searched is a + digit. */ +# undef strpbrk +_GL_WARN_ON_USE (strpbrk, "strpbrk cannot work correctly on character strings " + "in multibyte locales - " + "use mbspbrk if you care about internationalization"); +# endif +#elif defined GNULIB_POSIXCHECK +# undef strpbrk +# if HAVE_RAW_DECL_STRPBRK +_GL_WARN_ON_USE (strpbrk, "strpbrk is unportable - " + "use gnulib module strpbrk for portability"); +# endif +#endif + +#if defined GNULIB_POSIXCHECK +/* strspn() assumes the second argument is a list of single-byte characters. + Even in this simple case, it cannot work with multibyte strings. */ +# undef strspn +/* Assume strspn is always declared. */ +_GL_WARN_ON_USE (strspn, "strspn cannot work correctly on character strings " + "in multibyte locales - " + "use mbsspn if you care about internationalization"); +#endif + +#if defined GNULIB_POSIXCHECK +/* strrchr() does not work with multibyte strings if the locale encoding is + GB18030 and the character to be searched is a digit. */ +# undef strrchr +/* Assume strrchr is always declared. */ +_GL_WARN_ON_USE (strrchr, "strrchr cannot work correctly on character strings " + "in some multibyte locales - " + "use mbsrchr if you care about internationalization"); +#endif + +/* Search the next delimiter (char listed in DELIM) starting at *STRINGP. + If one is found, overwrite it with a NUL, and advance *STRINGP + to point to the next char after it. Otherwise, set *STRINGP to NULL. + If *STRINGP was already NULL, nothing happens. + Return the old value of *STRINGP. + + This is a variant of strtok() that is multithread-safe and supports + empty fields. + + Caveat: It modifies the original string. + Caveat: These functions cannot be used on constant strings. + Caveat: The identity of the delimiting character is lost. + Caveat: It doesn't work with multibyte strings unless all of the delimiter + characters are ASCII characters < 0x30. + + See also strtok_r(). */ +#if @GNULIB_STRSEP@ +# if ! @HAVE_STRSEP@ +_GL_FUNCDECL_SYS (strsep, char *, + (char **restrict __stringp, char const *restrict __delim) + _GL_ARG_NONNULL ((1, 2))); +# endif +_GL_CXXALIAS_SYS (strsep, char *, + (char **restrict __stringp, char const *restrict __delim)); +_GL_CXXALIASWARN (strsep); +# if defined GNULIB_POSIXCHECK +# undef strsep +_GL_WARN_ON_USE (strsep, "strsep cannot work correctly on character strings " + "in multibyte locales - " + "use mbssep if you care about internationalization"); +# endif +#elif defined GNULIB_POSIXCHECK +# undef strsep +# if HAVE_RAW_DECL_STRSEP +_GL_WARN_ON_USE (strsep, "strsep is unportable - " + "use gnulib module strsep for portability"); +# endif +#endif + +#if @GNULIB_STRSTR@ +# if @REPLACE_STRSTR@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# define strstr rpl_strstr +# endif +_GL_FUNCDECL_RPL (strstr, char *, (const char *haystack, const char *needle) + _GL_ATTRIBUTE_PURE + _GL_ARG_NONNULL ((1, 2))); +_GL_CXXALIAS_RPL (strstr, char *, (const char *haystack, const char *needle)); +# else + /* On some systems, this function is defined as an overloaded function: + extern "C++" { const char * strstr (const char *, const char *); } + extern "C++" { char * strstr (char *, const char *); } */ +_GL_CXXALIAS_SYS_CAST2 (strstr, + char *, (const char *haystack, const char *needle), + const char *, (const char *haystack, const char *needle)); +# endif +# if ((__GLIBC__ == 2 && __GLIBC_MINOR__ >= 10) && !defined __UCLIBC__) \ + && (__GNUC__ > 4 || (__GNUC__ == 4 && __GNUC_MINOR__ >= 4)) +_GL_CXXALIASWARN1 (strstr, char *, (char *haystack, const char *needle)); +_GL_CXXALIASWARN1 (strstr, const char *, + (const char *haystack, const char *needle)); +# else +_GL_CXXALIASWARN (strstr); +# endif +#elif defined GNULIB_POSIXCHECK +/* strstr() does not work with multibyte strings if the locale encoding is + different from UTF-8: + POSIX says that it operates on "strings", and "string" in POSIX is defined + as a sequence of bytes, not of characters. */ +# undef strstr +/* Assume strstr is always declared. */ +_GL_WARN_ON_USE (strstr, "strstr is quadratic on many systems, and cannot " + "work correctly on character strings in most " + "multibyte locales - " + "use mbsstr if you care about internationalization, " + "or use strstr if you care about speed"); +#endif + +/* Find the first occurrence of NEEDLE in HAYSTACK, using case-insensitive + comparison. */ +#if @GNULIB_STRCASESTR@ +# if @REPLACE_STRCASESTR@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# define strcasestr rpl_strcasestr +# endif +_GL_FUNCDECL_RPL (strcasestr, char *, + (const char *haystack, const char *needle) + _GL_ATTRIBUTE_PURE + _GL_ARG_NONNULL ((1, 2))); +_GL_CXXALIAS_RPL (strcasestr, char *, + (const char *haystack, const char *needle)); +# else +# if ! @HAVE_STRCASESTR@ +_GL_FUNCDECL_SYS (strcasestr, char *, + (const char *haystack, const char *needle) + _GL_ATTRIBUTE_PURE + _GL_ARG_NONNULL ((1, 2))); +# endif + /* On some systems, this function is defined as an overloaded function: + extern "C++" { const char * strcasestr (const char *, const char *); } + extern "C++" { char * strcasestr (char *, const char *); } */ +_GL_CXXALIAS_SYS_CAST2 (strcasestr, + char *, (const char *haystack, const char *needle), + const char *, (const char *haystack, const char *needle)); +# endif +# if ((__GLIBC__ == 2 && __GLIBC_MINOR__ >= 10) && !defined __UCLIBC__) \ + && (__GNUC__ > 4 || (__GNUC__ == 4 && __GNUC_MINOR__ >= 4)) +_GL_CXXALIASWARN1 (strcasestr, char *, (char *haystack, const char *needle)); +_GL_CXXALIASWARN1 (strcasestr, const char *, + (const char *haystack, const char *needle)); +# else +_GL_CXXALIASWARN (strcasestr); +# endif +#elif defined GNULIB_POSIXCHECK +/* strcasestr() does not work with multibyte strings: + It is a glibc extension, and glibc implements it only for unibyte + locales. */ +# undef strcasestr +# if HAVE_RAW_DECL_STRCASESTR +_GL_WARN_ON_USE (strcasestr, "strcasestr does work correctly on character " + "strings in multibyte locales - " + "use mbscasestr if you care about " + "internationalization, or use c-strcasestr if you want " + "a locale independent function"); +# endif +#endif + +/* Parse S into tokens separated by characters in DELIM. + If S is NULL, the saved pointer in SAVE_PTR is used as + the next starting point. For example: + char s[] = "-abc-=-def"; + char *sp; + x = strtok_r(s, "-", &sp); // x = "abc", sp = "=-def" + x = strtok_r(NULL, "-=", &sp); // x = "def", sp = NULL + x = strtok_r(NULL, "=", &sp); // x = NULL + // s = "abc\0-def\0" + + This is a variant of strtok() that is multithread-safe. + + For the POSIX documentation for this function, see: + http://www.opengroup.org/susv3xsh/strtok.html + + Caveat: It modifies the original string. + Caveat: These functions cannot be used on constant strings. + Caveat: The identity of the delimiting character is lost. + Caveat: It doesn't work with multibyte strings unless all of the delimiter + characters are ASCII characters < 0x30. + + See also strsep(). */ +#if @GNULIB_STRTOK_R@ +# if @REPLACE_STRTOK_R@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef strtok_r +# define strtok_r rpl_strtok_r +# endif +_GL_FUNCDECL_RPL (strtok_r, char *, + (char *restrict s, char const *restrict delim, + char **restrict save_ptr) + _GL_ARG_NONNULL ((2, 3))); +_GL_CXXALIAS_RPL (strtok_r, char *, + (char *restrict s, char const *restrict delim, + char **restrict save_ptr)); +# else +# if @UNDEFINE_STRTOK_R@ || defined GNULIB_POSIXCHECK +# undef strtok_r +# endif +# if ! @HAVE_DECL_STRTOK_R@ +_GL_FUNCDECL_SYS (strtok_r, char *, + (char *restrict s, char const *restrict delim, + char **restrict save_ptr) + _GL_ARG_NONNULL ((2, 3))); +# endif +_GL_CXXALIAS_SYS (strtok_r, char *, + (char *restrict s, char const *restrict delim, + char **restrict save_ptr)); +# endif +_GL_CXXALIASWARN (strtok_r); +# if defined GNULIB_POSIXCHECK +_GL_WARN_ON_USE (strtok_r, "strtok_r cannot work correctly on character " + "strings in multibyte locales - " + "use mbstok_r if you care about internationalization"); +# endif +#elif defined GNULIB_POSIXCHECK +# undef strtok_r +# if HAVE_RAW_DECL_STRTOK_R +_GL_WARN_ON_USE (strtok_r, "strtok_r is unportable - " + "use gnulib module strtok_r for portability"); +# endif +#endif + + +/* The following functions are not specified by POSIX. They are gnulib + extensions. */ + +#if @GNULIB_MBSLEN@ +/* Return the number of multibyte characters in the character string STRING. + This considers multibyte characters, unlike strlen, which counts bytes. */ +# ifdef __MirBSD__ /* MirBSD defines mbslen as a macro. Override it. */ +# undef mbslen +# endif +# if @HAVE_MBSLEN@ /* AIX, OSF/1, MirBSD define mbslen already in libc. */ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# define mbslen rpl_mbslen +# endif +_GL_FUNCDECL_RPL (mbslen, size_t, (const char *string) + _GL_ATTRIBUTE_PURE + _GL_ARG_NONNULL ((1))); +_GL_CXXALIAS_RPL (mbslen, size_t, (const char *string)); +# else +_GL_FUNCDECL_SYS (mbslen, size_t, (const char *string) + _GL_ATTRIBUTE_PURE + _GL_ARG_NONNULL ((1))); +_GL_CXXALIAS_SYS (mbslen, size_t, (const char *string)); +# endif +_GL_CXXALIASWARN (mbslen); +#endif + +#if @GNULIB_MBSNLEN@ +/* Return the number of multibyte characters in the character string starting + at STRING and ending at STRING + LEN. */ +_GL_EXTERN_C size_t mbsnlen (const char *string, size_t len) + _GL_ATTRIBUTE_PURE + _GL_ARG_NONNULL ((1)); +#endif + +#if @GNULIB_MBSCHR@ +/* Locate the first single-byte character C in the character string STRING, + and return a pointer to it. Return NULL if C is not found in STRING. + Unlike strchr(), this function works correctly in multibyte locales with + encodings such as GB18030. */ +# if defined __hpux +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# define mbschr rpl_mbschr /* avoid collision with HP-UX function */ +# endif +_GL_FUNCDECL_RPL (mbschr, char *, (const char *string, int c) + _GL_ATTRIBUTE_PURE + _GL_ARG_NONNULL ((1))); +_GL_CXXALIAS_RPL (mbschr, char *, (const char *string, int c)); +# else +_GL_FUNCDECL_SYS (mbschr, char *, (const char *string, int c) + _GL_ATTRIBUTE_PURE + _GL_ARG_NONNULL ((1))); +_GL_CXXALIAS_SYS (mbschr, char *, (const char *string, int c)); +# endif +_GL_CXXALIASWARN (mbschr); +#endif + +#if @GNULIB_MBSRCHR@ +/* Locate the last single-byte character C in the character string STRING, + and return a pointer to it. Return NULL if C is not found in STRING. + Unlike strrchr(), this function works correctly in multibyte locales with + encodings such as GB18030. */ +# if defined __hpux || defined __INTERIX +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# define mbsrchr rpl_mbsrchr /* avoid collision with system function */ +# endif +_GL_FUNCDECL_RPL (mbsrchr, char *, (const char *string, int c) + _GL_ATTRIBUTE_PURE + _GL_ARG_NONNULL ((1))); +_GL_CXXALIAS_RPL (mbsrchr, char *, (const char *string, int c)); +# else +_GL_FUNCDECL_SYS (mbsrchr, char *, (const char *string, int c) + _GL_ATTRIBUTE_PURE + _GL_ARG_NONNULL ((1))); +_GL_CXXALIAS_SYS (mbsrchr, char *, (const char *string, int c)); +# endif +_GL_CXXALIASWARN (mbsrchr); +#endif + +#if @GNULIB_MBSSTR@ +/* Find the first occurrence of the character string NEEDLE in the character + string HAYSTACK. Return NULL if NEEDLE is not found in HAYSTACK. + Unlike strstr(), this function works correctly in multibyte locales with + encodings different from UTF-8. */ +_GL_EXTERN_C char * mbsstr (const char *haystack, const char *needle) + _GL_ATTRIBUTE_PURE + _GL_ARG_NONNULL ((1, 2)); +#endif + +#if @GNULIB_MBSCASECMP@ +/* Compare the character strings S1 and S2, ignoring case, returning less than, + equal to or greater than zero if S1 is lexicographically less than, equal to + or greater than S2. + Note: This function may, in multibyte locales, return 0 for strings of + different lengths! + Unlike strcasecmp(), this function works correctly in multibyte locales. */ +_GL_EXTERN_C int mbscasecmp (const char *s1, const char *s2) + _GL_ATTRIBUTE_PURE + _GL_ARG_NONNULL ((1, 2)); +#endif + +#if @GNULIB_MBSNCASECMP@ +/* Compare the initial segment of the character string S1 consisting of at most + N characters with the initial segment of the character string S2 consisting + of at most N characters, ignoring case, returning less than, equal to or + greater than zero if the initial segment of S1 is lexicographically less + than, equal to or greater than the initial segment of S2. + Note: This function may, in multibyte locales, return 0 for initial segments + of different lengths! + Unlike strncasecmp(), this function works correctly in multibyte locales. + But beware that N is not a byte count but a character count! */ +_GL_EXTERN_C int mbsncasecmp (const char *s1, const char *s2, size_t n) + _GL_ATTRIBUTE_PURE + _GL_ARG_NONNULL ((1, 2)); +#endif + +#if @GNULIB_MBSPCASECMP@ +/* Compare the initial segment of the character string STRING consisting of + at most mbslen (PREFIX) characters with the character string PREFIX, + ignoring case. If the two match, return a pointer to the first byte + after this prefix in STRING. Otherwise, return NULL. + Note: This function may, in multibyte locales, return non-NULL if STRING + is of smaller length than PREFIX! + Unlike strncasecmp(), this function works correctly in multibyte + locales. */ +_GL_EXTERN_C char * mbspcasecmp (const char *string, const char *prefix) + _GL_ATTRIBUTE_PURE + _GL_ARG_NONNULL ((1, 2)); +#endif + +#if @GNULIB_MBSCASESTR@ +/* Find the first occurrence of the character string NEEDLE in the character + string HAYSTACK, using case-insensitive comparison. + Note: This function may, in multibyte locales, return success even if + strlen (haystack) < strlen (needle) ! + Unlike strcasestr(), this function works correctly in multibyte locales. */ +_GL_EXTERN_C char * mbscasestr (const char *haystack, const char *needle) + _GL_ATTRIBUTE_PURE + _GL_ARG_NONNULL ((1, 2)); +#endif + +#if @GNULIB_MBSCSPN@ +/* Find the first occurrence in the character string STRING of any character + in the character string ACCEPT. Return the number of bytes from the + beginning of the string to this occurrence, or to the end of the string + if none exists. + Unlike strcspn(), this function works correctly in multibyte locales. */ +_GL_EXTERN_C size_t mbscspn (const char *string, const char *accept) + _GL_ATTRIBUTE_PURE + _GL_ARG_NONNULL ((1, 2)); +#endif + +#if @GNULIB_MBSPBRK@ +/* Find the first occurrence in the character string STRING of any character + in the character string ACCEPT. Return the pointer to it, or NULL if none + exists. + Unlike strpbrk(), this function works correctly in multibyte locales. */ +# if defined __hpux +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# define mbspbrk rpl_mbspbrk /* avoid collision with HP-UX function */ +# endif +_GL_FUNCDECL_RPL (mbspbrk, char *, (const char *string, const char *accept) + _GL_ATTRIBUTE_PURE + _GL_ARG_NONNULL ((1, 2))); +_GL_CXXALIAS_RPL (mbspbrk, char *, (const char *string, const char *accept)); +# else +_GL_FUNCDECL_SYS (mbspbrk, char *, (const char *string, const char *accept) + _GL_ATTRIBUTE_PURE + _GL_ARG_NONNULL ((1, 2))); +_GL_CXXALIAS_SYS (mbspbrk, char *, (const char *string, const char *accept)); +# endif +_GL_CXXALIASWARN (mbspbrk); +#endif + +#if @GNULIB_MBSSPN@ +/* Find the first occurrence in the character string STRING of any character + not in the character string REJECT. Return the number of bytes from the + beginning of the string to this occurrence, or to the end of the string + if none exists. + Unlike strspn(), this function works correctly in multibyte locales. */ +_GL_EXTERN_C size_t mbsspn (const char *string, const char *reject) + _GL_ATTRIBUTE_PURE + _GL_ARG_NONNULL ((1, 2)); +#endif + +#if @GNULIB_MBSSEP@ +/* Search the next delimiter (multibyte character listed in the character + string DELIM) starting at the character string *STRINGP. + If one is found, overwrite it with a NUL, and advance *STRINGP to point + to the next multibyte character after it. Otherwise, set *STRINGP to NULL. + If *STRINGP was already NULL, nothing happens. + Return the old value of *STRINGP. + + This is a variant of mbstok_r() that supports empty fields. + + Caveat: It modifies the original string. + Caveat: These functions cannot be used on constant strings. + Caveat: The identity of the delimiting character is lost. + + See also mbstok_r(). */ +_GL_EXTERN_C char * mbssep (char **stringp, const char *delim) + _GL_ARG_NONNULL ((1, 2)); +#endif + +#if @GNULIB_MBSTOK_R@ +/* Parse the character string STRING into tokens separated by characters in + the character string DELIM. + If STRING is NULL, the saved pointer in SAVE_PTR is used as + the next starting point. For example: + char s[] = "-abc-=-def"; + char *sp; + x = mbstok_r(s, "-", &sp); // x = "abc", sp = "=-def" + x = mbstok_r(NULL, "-=", &sp); // x = "def", sp = NULL + x = mbstok_r(NULL, "=", &sp); // x = NULL + // s = "abc\0-def\0" + + Caveat: It modifies the original string. + Caveat: These functions cannot be used on constant strings. + Caveat: The identity of the delimiting character is lost. + + See also mbssep(). */ +_GL_EXTERN_C char * mbstok_r (char *string, const char *delim, char **save_ptr) + _GL_ARG_NONNULL ((2, 3)); +#endif + +/* Map any int, typically from errno, into an error message. */ +#if @GNULIB_STRERROR@ +# if @REPLACE_STRERROR@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef strerror +# define strerror rpl_strerror +# endif +_GL_FUNCDECL_RPL (strerror, char *, (int)); +_GL_CXXALIAS_RPL (strerror, char *, (int)); +# else +_GL_CXXALIAS_SYS (strerror, char *, (int)); +# endif +_GL_CXXALIASWARN (strerror); +#elif defined GNULIB_POSIXCHECK +# undef strerror +/* Assume strerror is always declared. */ +_GL_WARN_ON_USE (strerror, "strerror is unportable - " + "use gnulib module strerror to guarantee non-NULL result"); +#endif + +/* Map any int, typically from errno, into an error message. Multithread-safe. + Uses the POSIX declaration, not the glibc declaration. */ +#if @GNULIB_STRERROR_R@ +# if @REPLACE_STRERROR_R@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef strerror_r +# define strerror_r rpl_strerror_r +# endif +_GL_FUNCDECL_RPL (strerror_r, int, (int errnum, char *buf, size_t buflen) + _GL_ARG_NONNULL ((2))); +_GL_CXXALIAS_RPL (strerror_r, int, (int errnum, char *buf, size_t buflen)); +# else +# if !@HAVE_DECL_STRERROR_R@ +_GL_FUNCDECL_SYS (strerror_r, int, (int errnum, char *buf, size_t buflen) + _GL_ARG_NONNULL ((2))); +# endif +_GL_CXXALIAS_SYS (strerror_r, int, (int errnum, char *buf, size_t buflen)); +# endif +# if @HAVE_DECL_STRERROR_R@ +_GL_CXXALIASWARN (strerror_r); +# endif +#elif defined GNULIB_POSIXCHECK +# undef strerror_r +# if HAVE_RAW_DECL_STRERROR_R +_GL_WARN_ON_USE (strerror_r, "strerror_r is unportable - " + "use gnulib module strerror_r-posix for portability"); +# endif +#endif + +#if @GNULIB_STRSIGNAL@ +# if @REPLACE_STRSIGNAL@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# define strsignal rpl_strsignal +# endif +_GL_FUNCDECL_RPL (strsignal, char *, (int __sig)); +_GL_CXXALIAS_RPL (strsignal, char *, (int __sig)); +# else +# if ! @HAVE_DECL_STRSIGNAL@ +_GL_FUNCDECL_SYS (strsignal, char *, (int __sig)); +# endif +/* Need to cast, because on Cygwin 1.5.x systems, the return type is + 'const char *'. */ +_GL_CXXALIAS_SYS_CAST (strsignal, char *, (int __sig)); +# endif +_GL_CXXALIASWARN (strsignal); +#elif defined GNULIB_POSIXCHECK +# undef strsignal +# if HAVE_RAW_DECL_STRSIGNAL +_GL_WARN_ON_USE (strsignal, "strsignal is unportable - " + "use gnulib module strsignal for portability"); +# endif +#endif + +#if @GNULIB_STRVERSCMP@ +# if !@HAVE_STRVERSCMP@ +_GL_FUNCDECL_SYS (strverscmp, int, (const char *, const char *) + _GL_ATTRIBUTE_PURE + _GL_ARG_NONNULL ((1, 2))); +# endif +_GL_CXXALIAS_SYS (strverscmp, int, (const char *, const char *)); +_GL_CXXALIASWARN (strverscmp); +#elif defined GNULIB_POSIXCHECK +# undef strverscmp +# if HAVE_RAW_DECL_STRVERSCMP +_GL_WARN_ON_USE (strverscmp, "strverscmp is unportable - " + "use gnulib module strverscmp for portability"); +# endif +#endif + + +#endif /* _@GUARD_PREFIX@_STRING_H */ +#endif /* _@GUARD_PREFIX@_STRING_H */ +#endif diff --git a/lib/stripslash.c b/lib/stripslash.c new file mode 100644 index 0000000..60c6781 --- /dev/null +++ b/lib/stripslash.c @@ -0,0 +1,45 @@ +/* stripslash.c -- remove redundant trailing slashes from a file name + + Copyright (C) 1990, 2001, 2003-2006, 2009-2018 Free Software Foundation, + Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +#include <config.h> + +#include "dirname.h" + +/* Remove trailing slashes from FILE. Return true if a trailing slash + was removed. This is useful when using file name completion from a + shell that adds a "/" after directory names (such as tcsh and + bash), because on symlinks to directories, several system calls + have different semantics according to whether a trailing slash is + present. */ + +bool +strip_trailing_slashes (char *file) +{ + char *base = last_component (file); + char *base_lim; + bool had_slash; + + /* last_component returns "" for file system roots, but we need to turn + "///" into "/". */ + if (! *base) + base = file; + base_lim = base + base_len (base); + had_slash = (*base_lim != '\0'); + *base_lim = '\0'; + return had_slash; +} diff --git a/lib/strverscmp.c b/lib/strverscmp.c new file mode 100644 index 0000000..4eac791 --- /dev/null +++ b/lib/strverscmp.c @@ -0,0 +1,117 @@ +/* Compare strings while treating digits characters numerically. + Copyright (C) 1997-2018 Free Software Foundation, Inc. + This file is part of the GNU C Library. + Contributed by Jean-François Bignolles <bignolle@ecoledoc.ibp.fr>, 1997. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU General Public + License as published by the Free Software Foundation; either + version 3 of the License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + General Public License for more details. + + You should have received a copy of the GNU General Public + License along with the GNU C Library; if not, see + <https://www.gnu.org/licenses/>. */ + +#if !_LIBC +# include <config.h> +#endif + +#include <string.h> +#include <ctype.h> + +/* states: S_N: normal, S_I: comparing integral part, S_F: comparing + fractional parts, S_Z: idem but with leading Zeroes only */ +#define S_N 0x0 +#define S_I 0x3 +#define S_F 0x6 +#define S_Z 0x9 + +/* result_type: CMP: return diff; LEN: compare using len_diff/diff */ +#define CMP 2 +#define LEN 3 + +#ifndef weak_alias +# define __strverscmp strverscmp +#endif + +/* Compare S1 and S2 as strings holding indices/version numbers, + returning less than, equal to or greater than zero if S1 is less than, + equal to or greater than S2 (for more info, see the texinfo doc). +*/ + +int +__strverscmp (const char *s1, const char *s2) +{ + const unsigned char *p1 = (const unsigned char *) s1; + const unsigned char *p2 = (const unsigned char *) s2; + unsigned char c1, c2; + int state; + int diff; + + /* Symbol(s) 0 [1-9] others + Transition (10) 0 (01) d (00) x */ + static const unsigned char next_state[] = + { + /* state x d 0 */ + /* S_N */ S_N, S_I, S_Z, + /* S_I */ S_N, S_I, S_I, + /* S_F */ S_N, S_F, S_F, + /* S_Z */ S_N, S_F, S_Z + }; + + static const signed char result_type[] = + { + /* state x/x x/d x/0 d/x d/d d/0 0/x 0/d 0/0 */ + + /* S_N */ CMP, CMP, CMP, CMP, LEN, CMP, CMP, CMP, CMP, + /* S_I */ CMP, -1, -1, +1, LEN, LEN, +1, LEN, LEN, + /* S_F */ CMP, CMP, CMP, CMP, CMP, CMP, CMP, CMP, CMP, + /* S_Z */ CMP, +1, +1, -1, CMP, CMP, -1, CMP, CMP + }; + + if (p1 == p2) + return 0; + + c1 = *p1++; + c2 = *p2++; + /* Hint: '0' is a digit too. */ + state = S_N + ((c1 == '0') + (isdigit (c1) != 0)); + + while ((diff = c1 - c2) == 0) + { + if (c1 == '\0') + return diff; + + state = next_state[state]; + c1 = *p1++; + c2 = *p2++; + state += (c1 == '0') + (isdigit (c1) != 0); + } + + state = result_type[state * 3 + (((c2 == '0') + (isdigit (c2) != 0)))]; + + switch (state) + { + case CMP: + return diff; + + case LEN: + while (isdigit (*p1++)) + if (!isdigit (*p2++)) + return 1; + + return isdigit (*p2) ? -1 : diff; + + default: + return state; + } +} +#ifdef weak_alias +libc_hidden_def (__strverscmp) +weak_alias (__strverscmp, strverscmp) +#endif diff --git a/lib/sys_stat.in.h b/lib/sys_stat.in.h new file mode 100644 index 0000000..6ae6ac5 --- /dev/null +++ b/lib/sys_stat.in.h @@ -0,0 +1,816 @@ +/* Provide a more complete sys/stat.h header file. + Copyright (C) 2005-2018 Free Software Foundation, Inc. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3, or (at your option) + any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, see <https://www.gnu.org/licenses/>. */ + +/* Written by Eric Blake, Paul Eggert, and Jim Meyering. */ + +/* This file is supposed to be used on platforms where <sys/stat.h> is + incomplete. It is intended to provide definitions and prototypes + needed by an application. Start with what the system provides. */ + +#if __GNUC__ >= 3 +@PRAGMA_SYSTEM_HEADER@ +#endif +@PRAGMA_COLUMNS@ + +#if defined __need_system_sys_stat_h +/* Special invocation convention. */ + +#@INCLUDE_NEXT@ @NEXT_SYS_STAT_H@ + +#else +/* Normal invocation convention. */ + +#ifndef _@GUARD_PREFIX@_SYS_STAT_H + +/* Get nlink_t. + May also define off_t to a 64-bit type on native Windows. */ +#include <sys/types.h> + +/* Get struct timespec. */ +#include <time.h> + +/* The include_next requires a split double-inclusion guard. */ +#@INCLUDE_NEXT@ @NEXT_SYS_STAT_H@ + +#ifndef _@GUARD_PREFIX@_SYS_STAT_H +#define _@GUARD_PREFIX@_SYS_STAT_H + +/* The definitions of _GL_FUNCDECL_RPL etc. are copied here. */ + +/* The definition of _GL_ARG_NONNULL is copied here. */ + +/* The definition of _GL_WARN_ON_USE is copied here. */ + +/* Before doing "#define mkdir rpl_mkdir" below, we need to include all + headers that may declare mkdir(). Native Windows platforms declare mkdir + in <io.h> and/or <direct.h>, not in <unistd.h>. */ +#if defined _WIN32 && ! defined __CYGWIN__ +# include <io.h> /* mingw32, mingw64 */ +# include <direct.h> /* mingw64, MSVC 9 */ +#endif + +/* Native Windows platforms declare umask() in <io.h>. */ +#if 0 && (defined _WIN32 && ! defined __CYGWIN__) +# include <io.h> +#endif + +/* Large File Support on native Windows. */ +#if @WINDOWS_64_BIT_ST_SIZE@ +# define stat _stati64 +#endif + +/* Optionally, override 'struct stat' on native Windows. */ +#if @GNULIB_OVERRIDES_STRUCT_STAT@ + +# undef stat +# if @GNULIB_STAT@ +# define stat rpl_stat +# else + /* Provoke a clear link error if stat() is used as a function and + module 'stat' is not in use. */ +# define stat stat_used_without_requesting_gnulib_module_stat +# endif + +# if !GNULIB_defined_struct_stat +struct stat +{ + dev_t st_dev; + ino_t st_ino; + mode_t st_mode; + nlink_t st_nlink; +# if 0 + uid_t st_uid; +# else /* uid_t is not defined by default on native Windows. */ + short st_uid; +# endif +# if 0 + gid_t st_gid; +# else /* gid_t is not defined by default on native Windows. */ + short st_gid; +# endif + dev_t st_rdev; + off_t st_size; +# if 0 + blksize_t st_blksize; + blkcnt_t st_blocks; +# endif + +# if @WINDOWS_STAT_TIMESPEC@ + struct timespec st_atim; + struct timespec st_mtim; + struct timespec st_ctim; +# else + time_t st_atime; + time_t st_mtime; + time_t st_ctime; +# endif +}; +# if @WINDOWS_STAT_TIMESPEC@ +# define st_atime st_atim.tv_sec +# define st_mtime st_mtim.tv_sec +# define st_ctime st_ctim.tv_sec + /* Indicator, for gnulib internal purposes. */ +# define _GL_WINDOWS_STAT_TIMESPEC 1 +# endif +# define GNULIB_defined_struct_stat 1 +# endif + +/* Other possible values of st_mode. */ +# if 0 +# define _S_IFBLK 0x6000 +# endif +# if 0 +# define _S_IFLNK 0xA000 +# endif +# if 0 +# define _S_IFSOCK 0xC000 +# endif + +#endif + +#ifndef S_IFIFO +# ifdef _S_IFIFO +# define S_IFIFO _S_IFIFO +# endif +#endif + +#ifndef S_IFMT +# define S_IFMT 0170000 +#endif + +#if STAT_MACROS_BROKEN +# undef S_ISBLK +# undef S_ISCHR +# undef S_ISDIR +# undef S_ISFIFO +# undef S_ISLNK +# undef S_ISNAM +# undef S_ISMPB +# undef S_ISMPC +# undef S_ISNWK +# undef S_ISREG +# undef S_ISSOCK +#endif + +#ifndef S_ISBLK +# ifdef S_IFBLK +# define S_ISBLK(m) (((m) & S_IFMT) == S_IFBLK) +# else +# define S_ISBLK(m) 0 +# endif +#endif + +#ifndef S_ISCHR +# ifdef S_IFCHR +# define S_ISCHR(m) (((m) & S_IFMT) == S_IFCHR) +# else +# define S_ISCHR(m) 0 +# endif +#endif + +#ifndef S_ISDIR +# ifdef S_IFDIR +# define S_ISDIR(m) (((m) & S_IFMT) == S_IFDIR) +# else +# define S_ISDIR(m) 0 +# endif +#endif + +#ifndef S_ISDOOR /* Solaris 2.5 and up */ +# define S_ISDOOR(m) 0 +#endif + +#ifndef S_ISFIFO +# ifdef S_IFIFO +# define S_ISFIFO(m) (((m) & S_IFMT) == S_IFIFO) +# else +# define S_ISFIFO(m) 0 +# endif +#endif + +#ifndef S_ISLNK +# ifdef S_IFLNK +# define S_ISLNK(m) (((m) & S_IFMT) == S_IFLNK) +# else +# define S_ISLNK(m) 0 +# endif +#endif + +#ifndef S_ISMPB /* V7 */ +# ifdef S_IFMPB +# define S_ISMPB(m) (((m) & S_IFMT) == S_IFMPB) +# define S_ISMPC(m) (((m) & S_IFMT) == S_IFMPC) +# else +# define S_ISMPB(m) 0 +# define S_ISMPC(m) 0 +# endif +#endif + +#ifndef S_ISMPX /* AIX */ +# define S_ISMPX(m) 0 +#endif + +#ifndef S_ISNAM /* Xenix */ +# ifdef S_IFNAM +# define S_ISNAM(m) (((m) & S_IFMT) == S_IFNAM) +# else +# define S_ISNAM(m) 0 +# endif +#endif + +#ifndef S_ISNWK /* HP/UX */ +# ifdef S_IFNWK +# define S_ISNWK(m) (((m) & S_IFMT) == S_IFNWK) +# else +# define S_ISNWK(m) 0 +# endif +#endif + +#ifndef S_ISPORT /* Solaris 10 and up */ +# define S_ISPORT(m) 0 +#endif + +#ifndef S_ISREG +# ifdef S_IFREG +# define S_ISREG(m) (((m) & S_IFMT) == S_IFREG) +# else +# define S_ISREG(m) 0 +# endif +#endif + +#ifndef S_ISSOCK +# ifdef S_IFSOCK +# define S_ISSOCK(m) (((m) & S_IFMT) == S_IFSOCK) +# else +# define S_ISSOCK(m) 0 +# endif +#endif + + +#ifndef S_TYPEISMQ +# define S_TYPEISMQ(p) 0 +#endif + +#ifndef S_TYPEISTMO +# define S_TYPEISTMO(p) 0 +#endif + + +#ifndef S_TYPEISSEM +# ifdef S_INSEM +# define S_TYPEISSEM(p) (S_ISNAM ((p)->st_mode) && (p)->st_rdev == S_INSEM) +# else +# define S_TYPEISSEM(p) 0 +# endif +#endif + +#ifndef S_TYPEISSHM +# ifdef S_INSHD +# define S_TYPEISSHM(p) (S_ISNAM ((p)->st_mode) && (p)->st_rdev == S_INSHD) +# else +# define S_TYPEISSHM(p) 0 +# endif +#endif + +/* high performance ("contiguous data") */ +#ifndef S_ISCTG +# define S_ISCTG(p) 0 +#endif + +/* Cray DMF (data migration facility): off line, with data */ +#ifndef S_ISOFD +# define S_ISOFD(p) 0 +#endif + +/* Cray DMF (data migration facility): off line, with no data */ +#ifndef S_ISOFL +# define S_ISOFL(p) 0 +#endif + +/* 4.4BSD whiteout */ +#ifndef S_ISWHT +# define S_ISWHT(m) 0 +#endif + +/* If any of the following are undefined, + define them to their de facto standard values. */ +#if !S_ISUID +# define S_ISUID 04000 +#endif +#if !S_ISGID +# define S_ISGID 02000 +#endif + +/* S_ISVTX is a common extension to POSIX. */ +#ifndef S_ISVTX +# define S_ISVTX 01000 +#endif + +#if !S_IRUSR && S_IREAD +# define S_IRUSR S_IREAD +#endif +#if !S_IRUSR +# define S_IRUSR 00400 +#endif +#if !S_IRGRP +# define S_IRGRP (S_IRUSR >> 3) +#endif +#if !S_IROTH +# define S_IROTH (S_IRUSR >> 6) +#endif + +#if !S_IWUSR && S_IWRITE +# define S_IWUSR S_IWRITE +#endif +#if !S_IWUSR +# define S_IWUSR 00200 +#endif +#if !S_IWGRP +# define S_IWGRP (S_IWUSR >> 3) +#endif +#if !S_IWOTH +# define S_IWOTH (S_IWUSR >> 6) +#endif + +#if !S_IXUSR && S_IEXEC +# define S_IXUSR S_IEXEC +#endif +#if !S_IXUSR +# define S_IXUSR 00100 +#endif +#if !S_IXGRP +# define S_IXGRP (S_IXUSR >> 3) +#endif +#if !S_IXOTH +# define S_IXOTH (S_IXUSR >> 6) +#endif + +#if !S_IRWXU +# define S_IRWXU (S_IRUSR | S_IWUSR | S_IXUSR) +#endif +#if !S_IRWXG +# define S_IRWXG (S_IRGRP | S_IWGRP | S_IXGRP) +#endif +#if !S_IRWXO +# define S_IRWXO (S_IROTH | S_IWOTH | S_IXOTH) +#endif + +/* S_IXUGO is a common extension to POSIX. */ +#if !S_IXUGO +# define S_IXUGO (S_IXUSR | S_IXGRP | S_IXOTH) +#endif + +#ifndef S_IRWXUGO +# define S_IRWXUGO (S_IRWXU | S_IRWXG | S_IRWXO) +#endif + +/* Macros for futimens and utimensat. */ +#ifndef UTIME_NOW +# define UTIME_NOW (-1) +# define UTIME_OMIT (-2) +#endif + + +#if @GNULIB_FCHMODAT@ +# if !@HAVE_FCHMODAT@ +_GL_FUNCDECL_SYS (fchmodat, int, + (int fd, char const *file, mode_t mode, int flag) + _GL_ARG_NONNULL ((2))); +# endif +_GL_CXXALIAS_SYS (fchmodat, int, + (int fd, char const *file, mode_t mode, int flag)); +_GL_CXXALIASWARN (fchmodat); +#elif defined GNULIB_POSIXCHECK +# undef fchmodat +# if HAVE_RAW_DECL_FCHMODAT +_GL_WARN_ON_USE (fchmodat, "fchmodat is not portable - " + "use gnulib module openat for portability"); +# endif +#endif + + +#if @GNULIB_FSTAT@ +# if @REPLACE_FSTAT@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef fstat +# define fstat rpl_fstat +# endif +_GL_FUNCDECL_RPL (fstat, int, (int fd, struct stat *buf) _GL_ARG_NONNULL ((2))); +_GL_CXXALIAS_RPL (fstat, int, (int fd, struct stat *buf)); +# else +_GL_CXXALIAS_SYS (fstat, int, (int fd, struct stat *buf)); +# endif +_GL_CXXALIASWARN (fstat); +#elif @GNULIB_OVERRIDES_STRUCT_STAT@ +# undef fstat +# define fstat fstat_used_without_requesting_gnulib_module_fstat +#elif @WINDOWS_64_BIT_ST_SIZE@ +/* Above, we define stat to _stati64. */ +# define fstat _fstati64 +#elif defined GNULIB_POSIXCHECK +# undef fstat +# if HAVE_RAW_DECL_FSTAT +_GL_WARN_ON_USE (fstat, "fstat has portability problems - " + "use gnulib module fstat for portability"); +# endif +#endif + + +#if @GNULIB_FSTATAT@ +# if @REPLACE_FSTATAT@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef fstatat +# define fstatat rpl_fstatat +# endif +_GL_FUNCDECL_RPL (fstatat, int, + (int fd, char const *name, struct stat *st, int flags) + _GL_ARG_NONNULL ((2, 3))); +_GL_CXXALIAS_RPL (fstatat, int, + (int fd, char const *name, struct stat *st, int flags)); +# else +# if !@HAVE_FSTATAT@ +_GL_FUNCDECL_SYS (fstatat, int, + (int fd, char const *name, struct stat *st, int flags) + _GL_ARG_NONNULL ((2, 3))); +# endif +_GL_CXXALIAS_SYS (fstatat, int, + (int fd, char const *name, struct stat *st, int flags)); +# endif +_GL_CXXALIASWARN (fstatat); +#elif @GNULIB_OVERRIDES_STRUCT_STAT@ +# undef fstatat +# define fstatat fstatat_used_without_requesting_gnulib_module_fstatat +#elif defined GNULIB_POSIXCHECK +# undef fstatat +# if HAVE_RAW_DECL_FSTATAT +_GL_WARN_ON_USE (fstatat, "fstatat is not portable - " + "use gnulib module openat for portability"); +# endif +#endif + + +#if @GNULIB_FUTIMENS@ +/* Use the rpl_ prefix also on Solaris <= 9, because on Solaris 9 our futimens + implementation relies on futimesat, which on Solaris 10 makes an invocation + to futimens that is meant to invoke the libc's futimens(), not gnulib's + futimens(). */ +# if @REPLACE_FUTIMENS@ || (!@HAVE_FUTIMENS@ && defined __sun) +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef futimens +# define futimens rpl_futimens +# endif +_GL_FUNCDECL_RPL (futimens, int, (int fd, struct timespec const times[2])); +_GL_CXXALIAS_RPL (futimens, int, (int fd, struct timespec const times[2])); +# else +# if !@HAVE_FUTIMENS@ +_GL_FUNCDECL_SYS (futimens, int, (int fd, struct timespec const times[2])); +# endif +_GL_CXXALIAS_SYS (futimens, int, (int fd, struct timespec const times[2])); +# endif +# if @HAVE_FUTIMENS@ +_GL_CXXALIASWARN (futimens); +# endif +#elif defined GNULIB_POSIXCHECK +# undef futimens +# if HAVE_RAW_DECL_FUTIMENS +_GL_WARN_ON_USE (futimens, "futimens is not portable - " + "use gnulib module futimens for portability"); +# endif +#endif + + +#if @GNULIB_LCHMOD@ +/* Change the mode of FILENAME to MODE, without dereferencing it if FILENAME + denotes a symbolic link. */ +# if !@HAVE_LCHMOD@ +/* The lchmod replacement follows symbolic links. Callers should take + this into account; lchmod should be applied only to arguments that + are known to not be symbolic links. On hosts that lack lchmod, + this can lead to race conditions between the check and the + invocation of lchmod, but we know of no workarounds that are + reliable in general. You might try requesting support for lchmod + from your operating system supplier. */ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# define lchmod chmod +# endif +/* Need to cast, because on mingw, the second parameter of chmod is + int mode. */ +_GL_CXXALIAS_RPL_CAST_1 (lchmod, chmod, int, + (const char *filename, mode_t mode)); +# else +# if 0 /* assume already declared */ +_GL_FUNCDECL_SYS (lchmod, int, (const char *filename, mode_t mode) + _GL_ARG_NONNULL ((1))); +# endif +_GL_CXXALIAS_SYS (lchmod, int, (const char *filename, mode_t mode)); +# endif +# if @HAVE_LCHMOD@ +_GL_CXXALIASWARN (lchmod); +# endif +#elif defined GNULIB_POSIXCHECK +# undef lchmod +# if HAVE_RAW_DECL_LCHMOD +_GL_WARN_ON_USE (lchmod, "lchmod is unportable - " + "use gnulib module lchmod for portability"); +# endif +#endif + + +#if @GNULIB_LSTAT@ +# if ! @HAVE_LSTAT@ +/* mingw does not support symlinks, therefore it does not have lstat. But + without links, stat does just fine. */ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# define lstat stat +# endif +_GL_CXXALIAS_RPL_1 (lstat, stat, int, (const char *name, struct stat *buf)); +# elif @REPLACE_LSTAT@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef lstat +# define lstat rpl_lstat +# endif +_GL_FUNCDECL_RPL (lstat, int, (const char *name, struct stat *buf) + _GL_ARG_NONNULL ((1, 2))); +_GL_CXXALIAS_RPL (lstat, int, (const char *name, struct stat *buf)); +# else +_GL_CXXALIAS_SYS (lstat, int, (const char *name, struct stat *buf)); +# endif +# if @HAVE_LSTAT@ +_GL_CXXALIASWARN (lstat); +# endif +#elif @GNULIB_OVERRIDES_STRUCT_STAT@ +# undef lstat +# define lstat lstat_used_without_requesting_gnulib_module_lstat +#elif defined GNULIB_POSIXCHECK +# undef lstat +# if HAVE_RAW_DECL_LSTAT +_GL_WARN_ON_USE (lstat, "lstat is unportable - " + "use gnulib module lstat for portability"); +# endif +#endif + + +#if @REPLACE_MKDIR@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef mkdir +# define mkdir rpl_mkdir +# endif +_GL_FUNCDECL_RPL (mkdir, int, (char const *name, mode_t mode) + _GL_ARG_NONNULL ((1))); +_GL_CXXALIAS_RPL (mkdir, int, (char const *name, mode_t mode)); +#else +/* mingw's _mkdir() function has 1 argument, but we pass 2 arguments. + Additionally, it declares _mkdir (and depending on compile flags, an + alias mkdir), only in the nonstandard includes <direct.h> and <io.h>, + which are included above. */ +# if defined _WIN32 && ! defined __CYGWIN__ + +# if !GNULIB_defined_rpl_mkdir +static int +rpl_mkdir (char const *name, mode_t mode) +{ + return _mkdir (name); +} +# define GNULIB_defined_rpl_mkdir 1 +# endif + +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# define mkdir rpl_mkdir +# endif +_GL_CXXALIAS_RPL (mkdir, int, (char const *name, mode_t mode)); +# else +_GL_CXXALIAS_SYS (mkdir, int, (char const *name, mode_t mode)); +# endif +#endif +_GL_CXXALIASWARN (mkdir); + + +#if @GNULIB_MKDIRAT@ +# if !@HAVE_MKDIRAT@ +_GL_FUNCDECL_SYS (mkdirat, int, (int fd, char const *file, mode_t mode) + _GL_ARG_NONNULL ((2))); +# endif +_GL_CXXALIAS_SYS (mkdirat, int, (int fd, char const *file, mode_t mode)); +_GL_CXXALIASWARN (mkdirat); +#elif defined GNULIB_POSIXCHECK +# undef mkdirat +# if HAVE_RAW_DECL_MKDIRAT +_GL_WARN_ON_USE (mkdirat, "mkdirat is not portable - " + "use gnulib module openat for portability"); +# endif +#endif + + +#if @GNULIB_MKFIFO@ +# if @REPLACE_MKFIFO@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef mkfifo +# define mkfifo rpl_mkfifo +# endif +_GL_FUNCDECL_RPL (mkfifo, int, (char const *file, mode_t mode) + _GL_ARG_NONNULL ((1))); +_GL_CXXALIAS_RPL (mkfifo, int, (char const *file, mode_t mode)); +# else +# if !@HAVE_MKFIFO@ +_GL_FUNCDECL_SYS (mkfifo, int, (char const *file, mode_t mode) + _GL_ARG_NONNULL ((1))); +# endif +_GL_CXXALIAS_SYS (mkfifo, int, (char const *file, mode_t mode)); +# endif +_GL_CXXALIASWARN (mkfifo); +#elif defined GNULIB_POSIXCHECK +# undef mkfifo +# if HAVE_RAW_DECL_MKFIFO +_GL_WARN_ON_USE (mkfifo, "mkfifo is not portable - " + "use gnulib module mkfifo for portability"); +# endif +#endif + + +#if @GNULIB_MKFIFOAT@ +# if !@HAVE_MKFIFOAT@ +_GL_FUNCDECL_SYS (mkfifoat, int, (int fd, char const *file, mode_t mode) + _GL_ARG_NONNULL ((2))); +# endif +_GL_CXXALIAS_SYS (mkfifoat, int, (int fd, char const *file, mode_t mode)); +_GL_CXXALIASWARN (mkfifoat); +#elif defined GNULIB_POSIXCHECK +# undef mkfifoat +# if HAVE_RAW_DECL_MKFIFOAT +_GL_WARN_ON_USE (mkfifoat, "mkfifoat is not portable - " + "use gnulib module mkfifoat for portability"); +# endif +#endif + + +#if @GNULIB_MKNOD@ +# if @REPLACE_MKNOD@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef mknod +# define mknod rpl_mknod +# endif +_GL_FUNCDECL_RPL (mknod, int, (char const *file, mode_t mode, dev_t dev) + _GL_ARG_NONNULL ((1))); +_GL_CXXALIAS_RPL (mknod, int, (char const *file, mode_t mode, dev_t dev)); +# else +# if !@HAVE_MKNOD@ +_GL_FUNCDECL_SYS (mknod, int, (char const *file, mode_t mode, dev_t dev) + _GL_ARG_NONNULL ((1))); +# endif +/* Need to cast, because on OSF/1 5.1, the third parameter is '...'. */ +_GL_CXXALIAS_SYS_CAST (mknod, int, (char const *file, mode_t mode, dev_t dev)); +# endif +_GL_CXXALIASWARN (mknod); +#elif defined GNULIB_POSIXCHECK +# undef mknod +# if HAVE_RAW_DECL_MKNOD +_GL_WARN_ON_USE (mknod, "mknod is not portable - " + "use gnulib module mknod for portability"); +# endif +#endif + + +#if @GNULIB_MKNODAT@ +# if !@HAVE_MKNODAT@ +_GL_FUNCDECL_SYS (mknodat, int, + (int fd, char const *file, mode_t mode, dev_t dev) + _GL_ARG_NONNULL ((2))); +# endif +_GL_CXXALIAS_SYS (mknodat, int, + (int fd, char const *file, mode_t mode, dev_t dev)); +_GL_CXXALIASWARN (mknodat); +#elif defined GNULIB_POSIXCHECK +# undef mknodat +# if HAVE_RAW_DECL_MKNODAT +_GL_WARN_ON_USE (mknodat, "mknodat is not portable - " + "use gnulib module mkfifoat for portability"); +# endif +#endif + + +#if @GNULIB_STAT@ +# if @REPLACE_STAT@ +# if !@GNULIB_OVERRIDES_STRUCT_STAT@ + /* We can't use the object-like #define stat rpl_stat, because of + struct stat. This means that rpl_stat will not be used if the user + does (stat)(a,b). Oh well. */ +# if defined _AIX && defined stat && defined _LARGE_FILES + /* With _LARGE_FILES defined, AIX (only) defines stat to stat64, + so we have to replace stat64() instead of stat(). */ +# undef stat64 +# define stat64(name, st) rpl_stat (name, st) +# elif @WINDOWS_64_BIT_ST_SIZE@ + /* Above, we define stat to _stati64. */ +# if defined __MINGW32__ && defined _stati64 +# ifndef _USE_32BIT_TIME_T + /* The system headers define _stati64 to _stat64. */ +# undef _stat64 +# define _stat64(name, st) rpl_stat (name, st) +# endif +# elif defined _MSC_VER && defined _stati64 +# ifdef _USE_32BIT_TIME_T + /* The system headers define _stati64 to _stat32i64. */ +# undef _stat32i64 +# define _stat32i64(name, st) rpl_stat (name, st) +# else + /* The system headers define _stati64 to _stat64. */ +# undef _stat64 +# define _stat64(name, st) rpl_stat (name, st) +# endif +# else +# undef _stati64 +# define _stati64(name, st) rpl_stat (name, st) +# endif +# elif defined __MINGW32__ && defined stat +# ifdef _USE_32BIT_TIME_T + /* The system headers define stat to _stat32i64. */ +# undef _stat32i64 +# define _stat32i64(name, st) rpl_stat (name, st) +# else + /* The system headers define stat to _stat64. */ +# undef _stat64 +# define _stat64(name, st) rpl_stat (name, st) +# endif +# elif defined _MSC_VER && defined stat +# ifdef _USE_32BIT_TIME_T + /* The system headers define stat to _stat32. */ +# undef _stat32 +# define _stat32(name, st) rpl_stat (name, st) +# else + /* The system headers define stat to _stat64i32. */ +# undef _stat64i32 +# define _stat64i32(name, st) rpl_stat (name, st) +# endif +# else /* !(_AIX || __MINGW32__ || _MSC_VER) */ +# undef stat +# define stat(name, st) rpl_stat (name, st) +# endif /* !_LARGE_FILES */ +# endif /* !@GNULIB_OVERRIDES_STRUCT_STAT@ */ +_GL_EXTERN_C int stat (const char *name, struct stat *buf) + _GL_ARG_NONNULL ((1, 2)); +# endif +#elif @GNULIB_OVERRIDES_STRUCT_STAT@ +/* see above: + #define stat stat_used_without_requesting_gnulib_module_stat + */ +#elif defined GNULIB_POSIXCHECK +# undef stat +# if HAVE_RAW_DECL_STAT +_GL_WARN_ON_USE (stat, "stat is unportable - " + "use gnulib module stat for portability"); +# endif +#endif + + +#if @GNULIB_UTIMENSAT@ +/* Use the rpl_ prefix also on Solaris <= 9, because on Solaris 9 our utimensat + implementation relies on futimesat, which on Solaris 10 makes an invocation + to utimensat that is meant to invoke the libc's utimensat(), not gnulib's + utimensat(). */ +# if @REPLACE_UTIMENSAT@ || (!@HAVE_UTIMENSAT@ && defined __sun) +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef utimensat +# define utimensat rpl_utimensat +# endif +_GL_FUNCDECL_RPL (utimensat, int, (int fd, char const *name, + struct timespec const times[2], int flag) + _GL_ARG_NONNULL ((2))); +_GL_CXXALIAS_RPL (utimensat, int, (int fd, char const *name, + struct timespec const times[2], int flag)); +# else +# if !@HAVE_UTIMENSAT@ +_GL_FUNCDECL_SYS (utimensat, int, (int fd, char const *name, + struct timespec const times[2], int flag) + _GL_ARG_NONNULL ((2))); +# endif +_GL_CXXALIAS_SYS (utimensat, int, (int fd, char const *name, + struct timespec const times[2], int flag)); +# endif +# if @HAVE_UTIMENSAT@ +_GL_CXXALIASWARN (utimensat); +# endif +#elif defined GNULIB_POSIXCHECK +# undef utimensat +# if HAVE_RAW_DECL_UTIMENSAT +_GL_WARN_ON_USE (utimensat, "utimensat is not portable - " + "use gnulib module utimensat for portability"); +# endif +#endif + + +#endif /* _@GUARD_PREFIX@_SYS_STAT_H */ +#endif /* _@GUARD_PREFIX@_SYS_STAT_H */ +#endif diff --git a/lib/sys_time.in.h b/lib/sys_time.in.h new file mode 100644 index 0000000..b4a0e49 --- /dev/null +++ b/lib/sys_time.in.h @@ -0,0 +1,220 @@ +/* Provide a more complete sys/time.h. + + Copyright (C) 2007-2018 Free Software Foundation, Inc. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3, or (at your option) + any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, see <https://www.gnu.org/licenses/>. */ + +/* Written by Paul Eggert. */ + +#ifndef _@GUARD_PREFIX@_SYS_TIME_H + +#if __GNUC__ >= 3 +@PRAGMA_SYSTEM_HEADER@ +#endif +@PRAGMA_COLUMNS@ + +/* On Cygwin and on many BSDish systems, <sys/time.h> includes itself + recursively via <sys/select.h>. + Simply delegate to the system's header in this case; it is a no-op. + Without this extra ifdef, the C++ gettimeofday declaration below + would be a forward declaration in gnulib's nested <sys/time.h>. */ +#if defined _CYGWIN_SYS_TIME_H || defined _SYS_TIME_H || defined _SYS_TIME_H_ +# @INCLUDE_NEXT@ @NEXT_SYS_TIME_H@ +#else + +/* The include_next requires a split double-inclusion guard. */ +#if @HAVE_SYS_TIME_H@ +# @INCLUDE_NEXT@ @NEXT_SYS_TIME_H@ +#endif + +#ifndef _@GUARD_PREFIX@_SYS_TIME_H +#define _@GUARD_PREFIX@_SYS_TIME_H + +#if ! @HAVE_SYS_TIME_H@ +# include <time.h> +#endif + +/* On native Windows with MSVC, get the 'struct timeval' type. + Also, on native Windows with a 64-bit time_t, where we are overriding the + 'struct timeval' type, get all declarations of system functions whose + signature contains 'struct timeval'. */ +#if (defined _MSC_VER || @REPLACE_STRUCT_TIMEVAL@) && @HAVE_WINSOCK2_H@ && !defined _GL_INCLUDING_WINSOCK2_H +# define _GL_INCLUDING_WINSOCK2_H +# include <winsock2.h> +# undef _GL_INCLUDING_WINSOCK2_H +#endif + +/* The definitions of _GL_FUNCDECL_RPL etc. are copied here. */ + +/* The definition of _GL_ARG_NONNULL is copied here. */ + +/* The definition of _GL_WARN_ON_USE is copied here. */ + +#ifdef __cplusplus +extern "C" { +#endif + +#if !@HAVE_STRUCT_TIMEVAL@ || @REPLACE_STRUCT_TIMEVAL@ + +# if @REPLACE_STRUCT_TIMEVAL@ +# define timeval rpl_timeval +# endif + +# if !GNULIB_defined_struct_timeval +struct timeval +{ + time_t tv_sec; + long int tv_usec; +}; +# define GNULIB_defined_struct_timeval 1 +# endif + +#endif + +#ifdef __cplusplus +} +#endif + +#if @GNULIB_GETTIMEOFDAY@ +# if @REPLACE_GETTIMEOFDAY@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef gettimeofday +# define gettimeofday rpl_gettimeofday +# endif +_GL_FUNCDECL_RPL (gettimeofday, int, + (struct timeval *restrict, void *restrict) + _GL_ARG_NONNULL ((1))); +_GL_CXXALIAS_RPL (gettimeofday, int, + (struct timeval *restrict, void *restrict)); +# else +# if !@HAVE_GETTIMEOFDAY@ +_GL_FUNCDECL_SYS (gettimeofday, int, + (struct timeval *restrict, void *restrict) + _GL_ARG_NONNULL ((1))); +# endif +/* Need to cast, because on glibc systems, by default, the second argument is + struct timezone *. */ +_GL_CXXALIAS_SYS_CAST (gettimeofday, int, + (struct timeval *restrict, void *restrict)); +# endif +_GL_CXXALIASWARN (gettimeofday); +# if defined __cplusplus && defined GNULIB_NAMESPACE +namespace GNULIB_NAMESPACE { + typedef ::timeval +#undef timeval + timeval; +} +# endif +#elif defined GNULIB_POSIXCHECK +# undef gettimeofday +# if HAVE_RAW_DECL_GETTIMEOFDAY +_GL_WARN_ON_USE (gettimeofday, "gettimeofday is unportable - " + "use gnulib module gettimeofday for portability"); +# endif +#endif + +/* Hide some function declarations from <winsock2.h>. */ + +#if defined _MSC_VER && @HAVE_WINSOCK2_H@ +# if !defined _@GUARD_PREFIX@_UNISTD_H +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef close +# define close close_used_without_including_unistd_h +# else + _GL_WARN_ON_USE (close, + "close() used without including <unistd.h>"); +# endif +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef gethostname +# define gethostname gethostname_used_without_including_unistd_h +# else + _GL_WARN_ON_USE (gethostname, + "gethostname() used without including <unistd.h>"); +# endif +# endif +# if !defined _@GUARD_PREFIX@_SYS_SOCKET_H +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef socket +# define socket socket_used_without_including_sys_socket_h +# undef connect +# define connect connect_used_without_including_sys_socket_h +# undef accept +# define accept accept_used_without_including_sys_socket_h +# undef bind +# define bind bind_used_without_including_sys_socket_h +# undef getpeername +# define getpeername getpeername_used_without_including_sys_socket_h +# undef getsockname +# define getsockname getsockname_used_without_including_sys_socket_h +# undef getsockopt +# define getsockopt getsockopt_used_without_including_sys_socket_h +# undef listen +# define listen listen_used_without_including_sys_socket_h +# undef recv +# define recv recv_used_without_including_sys_socket_h +# undef send +# define send send_used_without_including_sys_socket_h +# undef recvfrom +# define recvfrom recvfrom_used_without_including_sys_socket_h +# undef sendto +# define sendto sendto_used_without_including_sys_socket_h +# undef setsockopt +# define setsockopt setsockopt_used_without_including_sys_socket_h +# undef shutdown +# define shutdown shutdown_used_without_including_sys_socket_h +# else + _GL_WARN_ON_USE (socket, + "socket() used without including <sys/socket.h>"); + _GL_WARN_ON_USE (connect, + "connect() used without including <sys/socket.h>"); + _GL_WARN_ON_USE (accept, + "accept() used without including <sys/socket.h>"); + _GL_WARN_ON_USE (bind, + "bind() used without including <sys/socket.h>"); + _GL_WARN_ON_USE (getpeername, + "getpeername() used without including <sys/socket.h>"); + _GL_WARN_ON_USE (getsockname, + "getsockname() used without including <sys/socket.h>"); + _GL_WARN_ON_USE (getsockopt, + "getsockopt() used without including <sys/socket.h>"); + _GL_WARN_ON_USE (listen, + "listen() used without including <sys/socket.h>"); + _GL_WARN_ON_USE (recv, + "recv() used without including <sys/socket.h>"); + _GL_WARN_ON_USE (send, + "send() used without including <sys/socket.h>"); + _GL_WARN_ON_USE (recvfrom, + "recvfrom() used without including <sys/socket.h>"); + _GL_WARN_ON_USE (sendto, + "sendto() used without including <sys/socket.h>"); + _GL_WARN_ON_USE (setsockopt, + "setsockopt() used without including <sys/socket.h>"); + _GL_WARN_ON_USE (shutdown, + "shutdown() used without including <sys/socket.h>"); +# endif +# endif +# if !defined _@GUARD_PREFIX@_SYS_SELECT_H +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef select +# define select select_used_without_including_sys_select_h +# else + _GL_WARN_ON_USE (select, + "select() used without including <sys/select.h>"); +# endif +# endif +#endif + +#endif /* _@GUARD_PREFIX@_SYS_TIME_H */ +#endif /* _CYGWIN_SYS_TIME_H */ +#endif /* _@GUARD_PREFIX@_SYS_TIME_H */ diff --git a/lib/sys_types.in.h b/lib/sys_types.in.h new file mode 100644 index 0000000..7f8c1c4 --- /dev/null +++ b/lib/sys_types.in.h @@ -0,0 +1,106 @@ +/* Provide a more complete sys/types.h. + + Copyright (C) 2011-2018 Free Software Foundation, Inc. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3, or (at your option) + any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, see <https://www.gnu.org/licenses/>. */ + +#if __GNUC__ >= 3 +@PRAGMA_SYSTEM_HEADER@ +#endif +@PRAGMA_COLUMNS@ + +#if defined _WIN32 && !defined __CYGWIN__ \ + && (defined __need_off_t || defined __need___off64_t \ + || defined __need_ssize_t || defined __need_time_t) + +/* Special invocation convention inside mingw header files. */ + +#@INCLUDE_NEXT@ @NEXT_SYS_TYPES_H@ + +#else +/* Normal invocation convention. */ + +#ifndef _@GUARD_PREFIX@_SYS_TYPES_H + +/* The include_next requires a split double-inclusion guard. */ +# define _GL_INCLUDING_SYS_TYPES_H +#@INCLUDE_NEXT@ @NEXT_SYS_TYPES_H@ +# undef _GL_INCLUDING_SYS_TYPES_H + +#ifndef _@GUARD_PREFIX@_SYS_TYPES_H +#define _@GUARD_PREFIX@_SYS_TYPES_H + +/* Override off_t if Large File Support is requested on native Windows. */ +#if @WINDOWS_64_BIT_OFF_T@ +/* Same as int64_t in <stdint.h>. */ +# if defined _MSC_VER +# define off_t __int64 +# else +# define off_t long long int +# endif +/* Indicator, for gnulib internal purposes. */ +# define _GL_WINDOWS_64_BIT_OFF_T 1 +#endif + +/* Override dev_t and ino_t if distinguishable inodes support is requested + on native Windows. */ +#if @WINDOWS_STAT_INODES@ + +# if @WINDOWS_STAT_INODES@ == 2 +/* Experimental, not useful in Windows 10. */ + +/* Define dev_t to a 64-bit type. */ +# if !defined GNULIB_defined_dev_t +typedef unsigned long long int rpl_dev_t; +# undef dev_t +# define dev_t rpl_dev_t +# define GNULIB_defined_dev_t 1 +# endif + +/* Define ino_t to a 128-bit type. */ +# if !defined GNULIB_defined_ino_t +/* MSVC does not have a 128-bit integer type. + GCC has a 128-bit integer type __int128, but only on 64-bit targets. */ +typedef struct { unsigned long long int _gl_ino[2]; } rpl_ino_t; +# undef ino_t +# define ino_t rpl_ino_t +# define GNULIB_defined_ino_t 1 +# endif + +# else /* @WINDOWS_STAT_INODES@ == 1 */ + +/* Define ino_t to a 64-bit type. */ +# if !defined GNULIB_defined_ino_t +typedef unsigned long long int rpl_ino_t; +# undef ino_t +# define ino_t rpl_ino_t +# define GNULIB_defined_ino_t 1 +# endif + +# endif + +/* Indicator, for gnulib internal purposes. */ +# define _GL_WINDOWS_STAT_INODES @WINDOWS_STAT_INODES@ + +#endif + +/* MSVC 9 defines size_t in <stddef.h>, not in <sys/types.h>. */ +/* But avoid namespace pollution on glibc systems. */ +#if (defined _WIN32 && ! defined __CYGWIN__) && ! defined __GLIBC__ +# include <stddef.h> +#endif + +#endif /* _@GUARD_PREFIX@_SYS_TYPES_H */ +#endif /* _@GUARD_PREFIX@_SYS_TYPES_H */ +#endif /* __need_XXX */ diff --git a/lib/tempname.c b/lib/tempname.c new file mode 100644 index 0000000..da81263 --- /dev/null +++ b/lib/tempname.c @@ -0,0 +1,323 @@ +/* tempname.c - generate the name of a temporary file. + + Copyright (C) 1991-2003, 2005-2007, 2009-2018 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +/* Extracted from glibc sysdeps/posix/tempname.c. See also tmpdir.c. */ + +#if !_LIBC +# include <config.h> +# include "tempname.h" +#endif + +#include <sys/types.h> +#include <assert.h> + +#include <errno.h> +#ifndef __set_errno +# define __set_errno(Val) errno = (Val) +#endif + +#include <stdio.h> +#ifndef P_tmpdir +# define P_tmpdir "/tmp" +#endif +#ifndef TMP_MAX +# define TMP_MAX 238328 +#endif +#ifndef __GT_FILE +# define __GT_FILE 0 +# define __GT_DIR 1 +# define __GT_NOCREATE 2 +#endif +#if !_LIBC && (GT_FILE != __GT_FILE || GT_DIR != __GT_DIR \ + || GT_NOCREATE != __GT_NOCREATE) +# error report this to bug-gnulib@gnu.org +#endif + +#include <stddef.h> +#include <stdlib.h> +#include <string.h> + +#include <fcntl.h> +#include <sys/time.h> +#include <stdint.h> +#include <unistd.h> + +#include <sys/stat.h> + +#if _LIBC +# define struct_stat64 struct stat64 +#else +# define struct_stat64 struct stat +# define __try_tempname try_tempname +# define __gen_tempname gen_tempname +# define __getpid getpid +# define __gettimeofday gettimeofday +# define __mkdir mkdir +# define __open open +# define __lxstat64(version, file, buf) lstat (file, buf) +#endif + +#ifdef _LIBC +# include <hp-timing.h> +# if HP_TIMING_AVAIL +# define RANDOM_BITS(Var) \ + if (__builtin_expect (value == UINT64_C (0), 0)) \ + { \ + /* If this is the first time this function is used initialize \ + the variable we accumulate the value in to some somewhat \ + random value. If we'd not do this programs at startup time \ + might have a reduced set of possible names, at least on slow \ + machines. */ \ + struct timeval tv; \ + __gettimeofday (&tv, NULL); \ + value = ((uint64_t) tv.tv_usec << 16) ^ tv.tv_sec; \ + } \ + HP_TIMING_NOW (Var) +# endif +#endif + +/* Use the widest available unsigned type if uint64_t is not + available. The algorithm below extracts a number less than 62**6 + (approximately 2**35.725) from uint64_t, so ancient hosts where + uintmax_t is only 32 bits lose about 3.725 bits of randomness, + which is better than not having mkstemp at all. */ +#if !defined UINT64_MAX && !defined uint64_t +# define uint64_t uintmax_t +#endif + +#if _LIBC +/* Return nonzero if DIR is an existent directory. */ +static int +direxists (const char *dir) +{ + struct_stat64 buf; + return __xstat64 (_STAT_VER, dir, &buf) == 0 && S_ISDIR (buf.st_mode); +} + +/* Path search algorithm, for tmpnam, tmpfile, etc. If DIR is + non-null and exists, uses it; otherwise uses the first of $TMPDIR, + P_tmpdir, /tmp that exists. Copies into TMPL a template suitable + for use with mk[s]temp. Will fail (-1) if DIR is non-null and + doesn't exist, none of the searched dirs exists, or there's not + enough space in TMPL. */ +int +__path_search (char *tmpl, size_t tmpl_len, const char *dir, const char *pfx, + int try_tmpdir) +{ + const char *d; + size_t dlen, plen; + + if (!pfx || !pfx[0]) + { + pfx = "file"; + plen = 4; + } + else + { + plen = strlen (pfx); + if (plen > 5) + plen = 5; + } + + if (try_tmpdir) + { + d = __secure_getenv ("TMPDIR"); + if (d != NULL && direxists (d)) + dir = d; + else if (dir != NULL && direxists (dir)) + /* nothing */ ; + else + dir = NULL; + } + if (dir == NULL) + { + if (direxists (P_tmpdir)) + dir = P_tmpdir; + else if (strcmp (P_tmpdir, "/tmp") != 0 && direxists ("/tmp")) + dir = "/tmp"; + else + { + __set_errno (ENOENT); + return -1; + } + } + + dlen = strlen (dir); + while (dlen > 1 && dir[dlen - 1] == '/') + dlen--; /* remove trailing slashes */ + + /* check we have room for "${dir}/${pfx}XXXXXX\0" */ + if (tmpl_len < dlen + 1 + plen + 6 + 1) + { + __set_errno (EINVAL); + return -1; + } + + sprintf (tmpl, "%.*s/%.*sXXXXXX", (int) dlen, dir, (int) plen, pfx); + return 0; +} +#endif /* _LIBC */ + +/* These are the characters used in temporary file names. */ +static const char letters[] = +"abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789"; + +int +__try_tempname (char *tmpl, int suffixlen, void *args, + int (*tryfunc) (char *, void *)) +{ + int len; + char *XXXXXX; + static uint64_t value; + uint64_t random_time_bits; + unsigned int count; + int fd = -1; + int save_errno = errno; + + /* A lower bound on the number of temporary files to attempt to + generate. The maximum total number of temporary file names that + can exist for a given template is 62**6. It should never be + necessary to try all of these combinations. Instead if a reasonable + number of names is tried (we define reasonable as 62**3) fail to + give the system administrator the chance to remove the problems. */ +#define ATTEMPTS_MIN (62 * 62 * 62) + + /* The number of times to attempt to generate a temporary file. To + conform to POSIX, this must be no smaller than TMP_MAX. */ +#if ATTEMPTS_MIN < TMP_MAX + unsigned int attempts = TMP_MAX; +#else + unsigned int attempts = ATTEMPTS_MIN; +#endif + + len = strlen (tmpl); + if (len < 6 + suffixlen || memcmp (&tmpl[len - 6 - suffixlen], "XXXXXX", 6)) + { + __set_errno (EINVAL); + return -1; + } + + /* This is where the Xs start. */ + XXXXXX = &tmpl[len - 6 - suffixlen]; + + /* Get some more or less random data. */ +#ifdef RANDOM_BITS + RANDOM_BITS (random_time_bits); +#else + { + struct timeval tv; + __gettimeofday (&tv, NULL); + random_time_bits = ((uint64_t) tv.tv_usec << 16) ^ tv.tv_sec; + } +#endif + value += random_time_bits ^ __getpid (); + + for (count = 0; count < attempts; value += 7777, ++count) + { + uint64_t v = value; + + /* Fill in the random bits. */ + XXXXXX[0] = letters[v % 62]; + v /= 62; + XXXXXX[1] = letters[v % 62]; + v /= 62; + XXXXXX[2] = letters[v % 62]; + v /= 62; + XXXXXX[3] = letters[v % 62]; + v /= 62; + XXXXXX[4] = letters[v % 62]; + v /= 62; + XXXXXX[5] = letters[v % 62]; + + fd = tryfunc (tmpl, args); + if (fd >= 0) + { + __set_errno (save_errno); + return fd; + } + else if (errno != EEXIST) + return -1; + } + + /* We got out of the loop because we ran out of combinations to try. */ + __set_errno (EEXIST); + return -1; +} + +static int +try_file (char *tmpl, void *flags) +{ + int *openflags = flags; + return __open (tmpl, + (*openflags & ~O_ACCMODE) + | O_RDWR | O_CREAT | O_EXCL, S_IRUSR | S_IWUSR); +} + +static int +try_dir (char *tmpl, void *flags _GL_UNUSED) +{ + return __mkdir (tmpl, S_IRUSR | S_IWUSR | S_IXUSR); +} + +static int +try_nocreate (char *tmpl, void *flags _GL_UNUSED) +{ + struct_stat64 st; + + if (__lxstat64 (_STAT_VER, tmpl, &st) == 0 || errno == EOVERFLOW) + __set_errno (EEXIST); + return errno == ENOENT ? 0 : -1; +} + +/* Generate a temporary file name based on TMPL. TMPL must match the + rules for mk[s]temp (i.e. end in "XXXXXX", possibly with a suffix). + The name constructed does not exist at the time of the call to + __gen_tempname. TMPL is overwritten with the result. + + KIND may be one of: + __GT_NOCREATE: simply verify that the name does not exist + at the time of the call. + __GT_FILE: create the file using open(O_CREAT|O_EXCL) + and return a read-write fd. The file is mode 0600. + __GT_DIR: create a directory, which will be mode 0700. + + We use a clever algorithm to get hard-to-predict names. */ +int +__gen_tempname (char *tmpl, int suffixlen, int flags, int kind) +{ + int (*tryfunc) (char *, void *); + + switch (kind) + { + case __GT_FILE: + tryfunc = try_file; + break; + + case __GT_DIR: + tryfunc = try_dir; + break; + + case __GT_NOCREATE: + tryfunc = try_nocreate; + break; + + default: + assert (! "invalid KIND in __gen_tempname"); + abort (); + } + return __try_tempname (tmpl, suffixlen, &flags, tryfunc); +} diff --git a/lib/tempname.h b/lib/tempname.h new file mode 100644 index 0000000..4020c73 --- /dev/null +++ b/lib/tempname.h @@ -0,0 +1,65 @@ +/* Create a temporary file or directory. + + Copyright (C) 2006, 2009-2018 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +/* header written by Eric Blake */ + +#ifndef GL_TEMPNAME_H +# define GL_TEMPNAME_H + +# include <stdio.h> + +# ifdef __GT_FILE +# define GT_FILE __GT_FILE +# define GT_DIR __GT_DIR +# define GT_NOCREATE __GT_NOCREATE +# else +# define GT_FILE 0 +# define GT_DIR 1 +# define GT_NOCREATE 2 +# endif + +#ifdef __cplusplus +extern "C" { +#endif + +/* Generate a temporary file name based on TMPL. TMPL must match the + rules for mk[s]temp (i.e. end in "XXXXXX", possibly with a suffix). + The name constructed does not exist at the time of the call to + gen_tempname. TMPL is overwritten with the result. + + KIND may be one of: + GT_NOCREATE: simply verify that the name does not exist + at the time of the call. + GT_FILE: create a large file using open(O_CREAT|O_EXCL) + and return a read-write fd. The file is mode 0600. + GT_DIR: create a directory, which will be mode 0700. + + We use a clever algorithm to get hard-to-predict names. */ +extern int gen_tempname (char *tmpl, int suffixlen, int flags, int kind); + +/* Similar to gen_tempname, but TRYFUNC is called for each temporary + name to try. If TRYFUNC returns a non-negative number, TRY_GEN_TEMPNAME + returns with this value. Otherwise, if errno is set to EEXIST, another + name is tried, or else TRY_GEN_TEMPNAME returns -1. */ +extern int try_tempname (char *tmpl, int suffixlen, void *args, + int (*tryfunc) (char *, void *)); + +#ifdef __cplusplus +} +#endif + +#endif /* GL_TEMPNAME_H */ diff --git a/lib/time.in.h b/lib/time.in.h new file mode 100644 index 0000000..3128f44 --- /dev/null +++ b/lib/time.in.h @@ -0,0 +1,350 @@ +/* A more-standard <time.h>. + + Copyright (C) 2007-2018 Free Software Foundation, Inc. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3, or (at your option) + any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, see <https://www.gnu.org/licenses/>. */ + +#if __GNUC__ >= 3 +@PRAGMA_SYSTEM_HEADER@ +#endif +@PRAGMA_COLUMNS@ + +/* Don't get in the way of glibc when it includes time.h merely to + declare a few standard symbols, rather than to declare all the + symbols. (However, skip this for MinGW as it treats __need_time_t + incompatibly.) Also, Solaris 8 <time.h> eventually includes itself + recursively; if that is happening, just include the system <time.h> + without adding our own declarations. */ +#if (((defined __need_time_t || defined __need_clock_t \ + || defined __need_timespec) \ + && !defined __MINGW32__) \ + || defined _@GUARD_PREFIX@_TIME_H) + +# @INCLUDE_NEXT@ @NEXT_TIME_H@ + +#else + +# define _@GUARD_PREFIX@_TIME_H + +# @INCLUDE_NEXT@ @NEXT_TIME_H@ + +/* NetBSD 5.0 mis-defines NULL. */ +# include <stddef.h> + +/* The definitions of _GL_FUNCDECL_RPL etc. are copied here. */ + +/* The definition of _GL_ARG_NONNULL is copied here. */ + +/* The definition of _GL_WARN_ON_USE is copied here. */ + +/* Some systems don't define struct timespec (e.g., AIX 4.1). + Or they define it with the wrong member names or define it in <sys/time.h> + (e.g., FreeBSD circa 1997). Stock Mingw prior to 3.0 does not define it, + but the pthreads-win32 library defines it in <pthread.h>. */ +# if ! @TIME_H_DEFINES_STRUCT_TIMESPEC@ +# if @SYS_TIME_H_DEFINES_STRUCT_TIMESPEC@ +# include <sys/time.h> +# elif @PTHREAD_H_DEFINES_STRUCT_TIMESPEC@ +# include <pthread.h> +# elif @UNISTD_H_DEFINES_STRUCT_TIMESPEC@ +# include <unistd.h> +# else + +# ifdef __cplusplus +extern "C" { +# endif + +# if !GNULIB_defined_struct_timespec +# undef timespec +# define timespec rpl_timespec +struct timespec +{ + time_t tv_sec; + long int tv_nsec; +}; +# define GNULIB_defined_struct_timespec 1 +# endif + +# ifdef __cplusplus +} +# endif + +# endif +# endif + +# if !GNULIB_defined_struct_time_t_must_be_integral +/* Per http://austingroupbugs.net/view.php?id=327, POSIX requires + time_t to be an integer type, even though C99 permits floating + point. We don't know of any implementation that uses floating + point, and it is much easier to write code that doesn't have to + worry about that corner case, so we force the issue. */ +struct __time_t_must_be_integral { + unsigned int __floating_time_t_unsupported : (time_t) 1; +}; +# define GNULIB_defined_struct_time_t_must_be_integral 1 +# endif + +/* Sleep for at least RQTP seconds unless interrupted, If interrupted, + return -1 and store the remaining time into RMTP. See + <http://www.opengroup.org/susv3xsh/nanosleep.html>. */ +# if @GNULIB_NANOSLEEP@ +# if @REPLACE_NANOSLEEP@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# define nanosleep rpl_nanosleep +# endif +_GL_FUNCDECL_RPL (nanosleep, int, + (struct timespec const *__rqtp, struct timespec *__rmtp) + _GL_ARG_NONNULL ((1))); +_GL_CXXALIAS_RPL (nanosleep, int, + (struct timespec const *__rqtp, struct timespec *__rmtp)); +# else +# if ! @HAVE_NANOSLEEP@ +_GL_FUNCDECL_SYS (nanosleep, int, + (struct timespec const *__rqtp, struct timespec *__rmtp) + _GL_ARG_NONNULL ((1))); +# endif +_GL_CXXALIAS_SYS (nanosleep, int, + (struct timespec const *__rqtp, struct timespec *__rmtp)); +# endif +_GL_CXXALIASWARN (nanosleep); +# endif + +/* Initialize time conversion information. */ +# if @GNULIB_TZSET@ +# if @REPLACE_TZSET@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef tzset +# define tzset rpl_tzset +# endif +_GL_FUNCDECL_RPL (tzset, void, (void)); +_GL_CXXALIAS_RPL (tzset, void, (void)); +# else +# if ! @HAVE_TZSET@ +_GL_FUNCDECL_SYS (tzset, void, (void)); +# endif +_GL_CXXALIAS_SYS (tzset, void, (void)); +# endif +_GL_CXXALIASWARN (tzset); +# endif + +/* Return the 'time_t' representation of TP and normalize TP. */ +# if @GNULIB_MKTIME@ +# if @REPLACE_MKTIME@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# define mktime rpl_mktime +# endif +_GL_FUNCDECL_RPL (mktime, time_t, (struct tm *__tp) _GL_ARG_NONNULL ((1))); +_GL_CXXALIAS_RPL (mktime, time_t, (struct tm *__tp)); +# else +_GL_CXXALIAS_SYS (mktime, time_t, (struct tm *__tp)); +# endif +_GL_CXXALIASWARN (mktime); +# endif + +/* Convert TIMER to RESULT, assuming local time and UTC respectively. See + <http://www.opengroup.org/susv3xsh/localtime_r.html> and + <http://www.opengroup.org/susv3xsh/gmtime_r.html>. */ +# if @GNULIB_TIME_R@ +# if @REPLACE_LOCALTIME_R@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef localtime_r +# define localtime_r rpl_localtime_r +# endif +_GL_FUNCDECL_RPL (localtime_r, struct tm *, (time_t const *restrict __timer, + struct tm *restrict __result) + _GL_ARG_NONNULL ((1, 2))); +_GL_CXXALIAS_RPL (localtime_r, struct tm *, (time_t const *restrict __timer, + struct tm *restrict __result)); +# else +# if ! @HAVE_DECL_LOCALTIME_R@ +_GL_FUNCDECL_SYS (localtime_r, struct tm *, (time_t const *restrict __timer, + struct tm *restrict __result) + _GL_ARG_NONNULL ((1, 2))); +# endif +_GL_CXXALIAS_SYS (localtime_r, struct tm *, (time_t const *restrict __timer, + struct tm *restrict __result)); +# endif +# if @HAVE_DECL_LOCALTIME_R@ +_GL_CXXALIASWARN (localtime_r); +# endif +# if @REPLACE_LOCALTIME_R@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef gmtime_r +# define gmtime_r rpl_gmtime_r +# endif +_GL_FUNCDECL_RPL (gmtime_r, struct tm *, (time_t const *restrict __timer, + struct tm *restrict __result) + _GL_ARG_NONNULL ((1, 2))); +_GL_CXXALIAS_RPL (gmtime_r, struct tm *, (time_t const *restrict __timer, + struct tm *restrict __result)); +# else +# if ! @HAVE_DECL_LOCALTIME_R@ +_GL_FUNCDECL_SYS (gmtime_r, struct tm *, (time_t const *restrict __timer, + struct tm *restrict __result) + _GL_ARG_NONNULL ((1, 2))); +# endif +_GL_CXXALIAS_SYS (gmtime_r, struct tm *, (time_t const *restrict __timer, + struct tm *restrict __result)); +# endif +# if @HAVE_DECL_LOCALTIME_R@ +_GL_CXXALIASWARN (gmtime_r); +# endif +# endif + +/* Convert TIMER to RESULT, assuming local time and UTC respectively. See + <http://www.opengroup.org/susv3xsh/localtime.html> and + <http://www.opengroup.org/susv3xsh/gmtime.html>. */ +# if @GNULIB_LOCALTIME@ || @REPLACE_LOCALTIME@ +# if @REPLACE_LOCALTIME@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef localtime +# define localtime rpl_localtime +# endif +_GL_FUNCDECL_RPL (localtime, struct tm *, (time_t const *__timer) + _GL_ARG_NONNULL ((1))); +_GL_CXXALIAS_RPL (localtime, struct tm *, (time_t const *__timer)); +# else +_GL_CXXALIAS_SYS (localtime, struct tm *, (time_t const *__timer)); +# endif +_GL_CXXALIASWARN (localtime); +# endif + +# if 0 || @REPLACE_GMTIME@ +# if @REPLACE_GMTIME@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef gmtime +# define gmtime rpl_gmtime +# endif +_GL_FUNCDECL_RPL (gmtime, struct tm *, (time_t const *__timer) + _GL_ARG_NONNULL ((1))); +_GL_CXXALIAS_RPL (gmtime, struct tm *, (time_t const *__timer)); +# else +_GL_CXXALIAS_SYS (gmtime, struct tm *, (time_t const *__timer)); +# endif +_GL_CXXALIASWARN (gmtime); +# endif + +/* Parse BUF as a timestamp, assuming FORMAT specifies its layout, and store + the resulting broken-down time into TM. See + <http://www.opengroup.org/susv3xsh/strptime.html>. */ +# if @GNULIB_STRPTIME@ +# if ! @HAVE_STRPTIME@ +_GL_FUNCDECL_SYS (strptime, char *, (char const *restrict __buf, + char const *restrict __format, + struct tm *restrict __tm) + _GL_ARG_NONNULL ((1, 2, 3))); +# endif +_GL_CXXALIAS_SYS (strptime, char *, (char const *restrict __buf, + char const *restrict __format, + struct tm *restrict __tm)); +_GL_CXXALIASWARN (strptime); +# endif + +/* Convert *TP to a date and time string. See + <http://pubs.opengroup.org/onlinepubs/9699919799/functions/ctime.html>. */ +# if @GNULIB_CTIME@ +# if @REPLACE_CTIME@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# define ctime rpl_ctime +# endif +_GL_FUNCDECL_RPL (ctime, char *, (time_t const *__tp) + _GL_ARG_NONNULL ((1))); +_GL_CXXALIAS_RPL (ctime, char *, (time_t const *__tp)); +# else +_GL_CXXALIAS_SYS (ctime, char *, (time_t const *__tp)); +# endif +_GL_CXXALIASWARN (ctime); +# endif + +/* Convert *TP to a date and time string. See + <http://pubs.opengroup.org/onlinepubs/9699919799/functions/strftime.html>. */ +# if @GNULIB_STRFTIME@ +# if @REPLACE_STRFTIME@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# define strftime rpl_strftime +# endif +_GL_FUNCDECL_RPL (strftime, size_t, (char *__buf, size_t __bufsize, + const char *__fmt, const struct tm *__tp) + _GL_ARG_NONNULL ((1, 3, 4))); +_GL_CXXALIAS_RPL (strftime, size_t, (char *__buf, size_t __bufsize, + const char *__fmt, const struct tm *__tp)); +# else +_GL_CXXALIAS_SYS (strftime, size_t, (char *__buf, size_t __bufsize, + const char *__fmt, const struct tm *__tp)); +# endif +_GL_CXXALIASWARN (strftime); +# endif + +# if defined _GNU_SOURCE && @GNULIB_TIME_RZ@ && ! @HAVE_TIMEZONE_T@ +typedef struct tm_zone *timezone_t; +_GL_FUNCDECL_SYS (tzalloc, timezone_t, (char const *__name)); +_GL_CXXALIAS_SYS (tzalloc, timezone_t, (char const *__name)); +_GL_FUNCDECL_SYS (tzfree, void, (timezone_t __tz)); +_GL_CXXALIAS_SYS (tzfree, void, (timezone_t __tz)); +_GL_FUNCDECL_SYS (localtime_rz, struct tm *, + (timezone_t __tz, time_t const *restrict __timer, + struct tm *restrict __result) _GL_ARG_NONNULL ((2, 3))); +_GL_CXXALIAS_SYS (localtime_rz, struct tm *, + (timezone_t __tz, time_t const *restrict __timer, + struct tm *restrict __result)); +_GL_FUNCDECL_SYS (mktime_z, time_t, + (timezone_t __tz, struct tm *restrict __result) + _GL_ARG_NONNULL ((2))); +_GL_CXXALIAS_SYS (mktime_z, time_t, + (timezone_t __tz, struct tm *restrict __result)); +# endif + +/* Convert TM to a time_t value, assuming UTC. */ +# if @GNULIB_TIMEGM@ +# if @REPLACE_TIMEGM@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef timegm +# define timegm rpl_timegm +# endif +_GL_FUNCDECL_RPL (timegm, time_t, (struct tm *__tm) _GL_ARG_NONNULL ((1))); +_GL_CXXALIAS_RPL (timegm, time_t, (struct tm *__tm)); +# else +# if ! @HAVE_TIMEGM@ +_GL_FUNCDECL_SYS (timegm, time_t, (struct tm *__tm) _GL_ARG_NONNULL ((1))); +# endif +_GL_CXXALIAS_SYS (timegm, time_t, (struct tm *__tm)); +# endif +_GL_CXXALIASWARN (timegm); +# endif + +/* Encourage applications to avoid unsafe functions that can overrun + buffers when given outlandish struct tm values. Portable + applications should use strftime (or even sprintf) instead. */ +# if defined GNULIB_POSIXCHECK +# undef asctime +_GL_WARN_ON_USE (asctime, "asctime can overrun buffers in some cases - " + "better use strftime (or even sprintf) instead"); +# endif +# if defined GNULIB_POSIXCHECK +# undef asctime_r +_GL_WARN_ON_USE (asctime, "asctime_r can overrun buffers in some cases - " + "better use strftime (or even sprintf) instead"); +# endif +# if defined GNULIB_POSIXCHECK +# undef ctime +_GL_WARN_ON_USE (asctime, "ctime can overrun buffers in some cases - " + "better use strftime (or even sprintf) instead"); +# endif +# if defined GNULIB_POSIXCHECK +# undef ctime_r +_GL_WARN_ON_USE (asctime, "ctime_r can overrun buffers in some cases - " + "better use strftime (or even sprintf) instead"); +# endif + +#endif diff --git a/lib/unistd.c b/lib/unistd.c new file mode 100644 index 0000000..72bad1c --- /dev/null +++ b/lib/unistd.c @@ -0,0 +1,4 @@ +#include <config.h> +#define _GL_UNISTD_INLINE _GL_EXTERN_INLINE +#include "unistd.h" +typedef int dummy; diff --git a/lib/unistd.in.h b/lib/unistd.in.h new file mode 100644 index 0000000..66f254d --- /dev/null +++ b/lib/unistd.in.h @@ -0,0 +1,1668 @@ +/* Substitute for and wrapper around <unistd.h>. + Copyright (C) 2003-2018 Free Software Foundation, Inc. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3, or (at your option) + any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, see <https://www.gnu.org/licenses/>. */ + +#ifndef _@GUARD_PREFIX@_UNISTD_H + +#if __GNUC__ >= 3 +@PRAGMA_SYSTEM_HEADER@ +#endif +@PRAGMA_COLUMNS@ + +#ifdef _GL_INCLUDING_UNISTD_H +/* Special invocation convention: + - On Mac OS X 10.3.9 we have a sequence of nested includes + <unistd.h> -> <signal.h> -> <pthread.h> -> <unistd.h> + In this situation, the functions are not yet declared, therefore we cannot + provide the C++ aliases. */ + +#@INCLUDE_NEXT@ @NEXT_UNISTD_H@ + +#else +/* Normal invocation convention. */ + +/* The include_next requires a split double-inclusion guard. */ +#if @HAVE_UNISTD_H@ +# define _GL_INCLUDING_UNISTD_H +# @INCLUDE_NEXT@ @NEXT_UNISTD_H@ +# undef _GL_INCLUDING_UNISTD_H +#endif + +/* Get all possible declarations of gethostname(). */ +#if @GNULIB_GETHOSTNAME@ && @UNISTD_H_HAVE_WINSOCK2_H@ \ + && !defined _GL_INCLUDING_WINSOCK2_H +# define _GL_INCLUDING_WINSOCK2_H +# include <winsock2.h> +# undef _GL_INCLUDING_WINSOCK2_H +#endif + +#if !defined _@GUARD_PREFIX@_UNISTD_H && !defined _GL_INCLUDING_WINSOCK2_H +#define _@GUARD_PREFIX@_UNISTD_H + +/* NetBSD 5.0 mis-defines NULL. Also get size_t. */ +#include <stddef.h> + +/* mingw doesn't define the SEEK_* or *_FILENO macros in <unistd.h>. */ +/* MSVC declares 'unlink' in <stdio.h>, not in <unistd.h>. We must include + it before we #define unlink rpl_unlink. */ +/* Cygwin 1.7.1 declares symlinkat in <stdio.h>, not in <unistd.h>. */ +/* But avoid namespace pollution on glibc systems. */ +#if (!(defined SEEK_CUR && defined SEEK_END && defined SEEK_SET) \ + || ((@GNULIB_UNLINK@ || defined GNULIB_POSIXCHECK) \ + && (defined _WIN32 && ! defined __CYGWIN__)) \ + || ((@GNULIB_SYMLINKAT@ || defined GNULIB_POSIXCHECK) \ + && defined __CYGWIN__)) \ + && ! defined __GLIBC__ +# include <stdio.h> +#endif + +/* Cygwin 1.7.1 declares unlinkat in <fcntl.h>, not in <unistd.h>. */ +/* But avoid namespace pollution on glibc systems. */ +#if (@GNULIB_UNLINKAT@ || defined GNULIB_POSIXCHECK) && defined __CYGWIN__ \ + && ! defined __GLIBC__ +# include <fcntl.h> +#endif + +/* mingw fails to declare _exit in <unistd.h>. */ +/* mingw, MSVC, BeOS, Haiku declare environ in <stdlib.h>, not in + <unistd.h>. */ +/* Solaris declares getcwd not only in <unistd.h> but also in <stdlib.h>. */ +/* OSF Tru64 Unix cannot see gnulib rpl_strtod when system <stdlib.h> is + included here. */ +/* But avoid namespace pollution on glibc systems. */ +#if !defined __GLIBC__ && !defined __osf__ +# define __need_system_stdlib_h +# include <stdlib.h> +# undef __need_system_stdlib_h +#endif + +/* Native Windows platforms declare chdir, getcwd, rmdir in + <io.h> and/or <direct.h>, not in <unistd.h>. + They also declare access(), chmod(), close(), dup(), dup2(), isatty(), + lseek(), read(), unlink(), write() in <io.h>. */ +#if ((@GNULIB_CHDIR@ || @GNULIB_GETCWD@ || @GNULIB_RMDIR@ \ + || defined GNULIB_POSIXCHECK) \ + && (defined _WIN32 && ! defined __CYGWIN__)) +# include <io.h> /* mingw32, mingw64 */ +# include <direct.h> /* mingw64, MSVC 9 */ +#elif (@GNULIB_CLOSE@ || @GNULIB_DUP@ || @GNULIB_DUP2@ || @GNULIB_ISATTY@ \ + || @GNULIB_LSEEK@ || @GNULIB_READ@ || @GNULIB_UNLINK@ || @GNULIB_WRITE@ \ + || defined GNULIB_POSIXCHECK) \ + && (defined _WIN32 && ! defined __CYGWIN__) +# include <io.h> +#endif + +/* AIX and OSF/1 5.1 declare getdomainname in <netdb.h>, not in <unistd.h>. + NonStop Kernel declares gethostname in <netdb.h>, not in <unistd.h>. */ +/* But avoid namespace pollution on glibc systems. */ +#if ((@GNULIB_GETDOMAINNAME@ && (defined _AIX || defined __osf__)) \ + || (@GNULIB_GETHOSTNAME@ && defined __TANDEM)) \ + && !defined __GLIBC__ +# include <netdb.h> +#endif + +/* MSVC defines off_t in <sys/types.h>. + May also define off_t to a 64-bit type on native Windows. */ +#if !@HAVE_UNISTD_H@ || @WINDOWS_64_BIT_OFF_T@ +/* Get off_t. */ +# include <sys/types.h> +#endif + +#if (@GNULIB_READ@ || @GNULIB_WRITE@ \ + || @GNULIB_READLINK@ || @GNULIB_READLINKAT@ \ + || @GNULIB_PREAD@ || @GNULIB_PWRITE@ || defined GNULIB_POSIXCHECK) +/* Get ssize_t. */ +# include <sys/types.h> +#endif + +/* The definitions of _GL_FUNCDECL_RPL etc. are copied here. */ + +/* The definition of _GL_ARG_NONNULL is copied here. */ + +/* The definition of _GL_WARN_ON_USE is copied here. */ + + +/* Get getopt(), optarg, optind, opterr, optopt. */ +#if @GNULIB_UNISTD_H_GETOPT@ && !defined _GL_SYSTEM_GETOPT +# include <getopt-cdefs.h> +# include <getopt-pfx-core.h> +#endif + +#ifndef _GL_INLINE_HEADER_BEGIN + #error "Please include config.h first." +#endif +_GL_INLINE_HEADER_BEGIN +#ifndef _GL_UNISTD_INLINE +# define _GL_UNISTD_INLINE _GL_INLINE +#endif + +/* Hide some function declarations from <winsock2.h>. */ + +#if @GNULIB_GETHOSTNAME@ && @UNISTD_H_HAVE_WINSOCK2_H@ +# if !defined _@GUARD_PREFIX@_SYS_SOCKET_H +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef socket +# define socket socket_used_without_including_sys_socket_h +# undef connect +# define connect connect_used_without_including_sys_socket_h +# undef accept +# define accept accept_used_without_including_sys_socket_h +# undef bind +# define bind bind_used_without_including_sys_socket_h +# undef getpeername +# define getpeername getpeername_used_without_including_sys_socket_h +# undef getsockname +# define getsockname getsockname_used_without_including_sys_socket_h +# undef getsockopt +# define getsockopt getsockopt_used_without_including_sys_socket_h +# undef listen +# define listen listen_used_without_including_sys_socket_h +# undef recv +# define recv recv_used_without_including_sys_socket_h +# undef send +# define send send_used_without_including_sys_socket_h +# undef recvfrom +# define recvfrom recvfrom_used_without_including_sys_socket_h +# undef sendto +# define sendto sendto_used_without_including_sys_socket_h +# undef setsockopt +# define setsockopt setsockopt_used_without_including_sys_socket_h +# undef shutdown +# define shutdown shutdown_used_without_including_sys_socket_h +# else + _GL_WARN_ON_USE (socket, + "socket() used without including <sys/socket.h>"); + _GL_WARN_ON_USE (connect, + "connect() used without including <sys/socket.h>"); + _GL_WARN_ON_USE (accept, + "accept() used without including <sys/socket.h>"); + _GL_WARN_ON_USE (bind, + "bind() used without including <sys/socket.h>"); + _GL_WARN_ON_USE (getpeername, + "getpeername() used without including <sys/socket.h>"); + _GL_WARN_ON_USE (getsockname, + "getsockname() used without including <sys/socket.h>"); + _GL_WARN_ON_USE (getsockopt, + "getsockopt() used without including <sys/socket.h>"); + _GL_WARN_ON_USE (listen, + "listen() used without including <sys/socket.h>"); + _GL_WARN_ON_USE (recv, + "recv() used without including <sys/socket.h>"); + _GL_WARN_ON_USE (send, + "send() used without including <sys/socket.h>"); + _GL_WARN_ON_USE (recvfrom, + "recvfrom() used without including <sys/socket.h>"); + _GL_WARN_ON_USE (sendto, + "sendto() used without including <sys/socket.h>"); + _GL_WARN_ON_USE (setsockopt, + "setsockopt() used without including <sys/socket.h>"); + _GL_WARN_ON_USE (shutdown, + "shutdown() used without including <sys/socket.h>"); +# endif +# endif +# if !defined _@GUARD_PREFIX@_SYS_SELECT_H +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef select +# define select select_used_without_including_sys_select_h +# else + _GL_WARN_ON_USE (select, + "select() used without including <sys/select.h>"); +# endif +# endif +#endif + + +/* OS/2 EMX lacks these macros. */ +#ifndef STDIN_FILENO +# define STDIN_FILENO 0 +#endif +#ifndef STDOUT_FILENO +# define STDOUT_FILENO 1 +#endif +#ifndef STDERR_FILENO +# define STDERR_FILENO 2 +#endif + +/* Ensure *_OK macros exist. */ +#ifndef F_OK +# define F_OK 0 +# define X_OK 1 +# define W_OK 2 +# define R_OK 4 +#endif + + +/* Declare overridden functions. */ + + +#if defined GNULIB_POSIXCHECK +/* The access() function is a security risk. */ +_GL_WARN_ON_USE (access, "the access function is a security risk - " + "use the gnulib module faccessat instead"); +#endif + + +#if @GNULIB_CHDIR@ +_GL_CXXALIAS_SYS (chdir, int, (const char *file) _GL_ARG_NONNULL ((1))); +_GL_CXXALIASWARN (chdir); +#elif defined GNULIB_POSIXCHECK +# undef chdir +# if HAVE_RAW_DECL_CHDIR +_GL_WARN_ON_USE (chown, "chdir is not always in <unistd.h> - " + "use gnulib module chdir for portability"); +# endif +#endif + + +#if @GNULIB_CHOWN@ +/* Change the owner of FILE to UID (if UID is not -1) and the group of FILE + to GID (if GID is not -1). Follow symbolic links. + Return 0 if successful, otherwise -1 and errno set. + See the POSIX:2008 specification + <http://pubs.opengroup.org/onlinepubs/9699919799/functions/chown.html. */ +# if @REPLACE_CHOWN@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef chown +# define chown rpl_chown +# endif +_GL_FUNCDECL_RPL (chown, int, (const char *file, uid_t uid, gid_t gid) + _GL_ARG_NONNULL ((1))); +_GL_CXXALIAS_RPL (chown, int, (const char *file, uid_t uid, gid_t gid)); +# else +# if !@HAVE_CHOWN@ +_GL_FUNCDECL_SYS (chown, int, (const char *file, uid_t uid, gid_t gid) + _GL_ARG_NONNULL ((1))); +# endif +_GL_CXXALIAS_SYS (chown, int, (const char *file, uid_t uid, gid_t gid)); +# endif +_GL_CXXALIASWARN (chown); +#elif defined GNULIB_POSIXCHECK +# undef chown +# if HAVE_RAW_DECL_CHOWN +_GL_WARN_ON_USE (chown, "chown fails to follow symlinks on some systems and " + "doesn't treat a uid or gid of -1 on some systems - " + "use gnulib module chown for portability"); +# endif +#endif + + +#if @GNULIB_CLOSE@ +# if @REPLACE_CLOSE@ +/* Automatically included by modules that need a replacement for close. */ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef close +# define close rpl_close +# endif +_GL_FUNCDECL_RPL (close, int, (int fd)); +_GL_CXXALIAS_RPL (close, int, (int fd)); +# else +_GL_CXXALIAS_SYS (close, int, (int fd)); +# endif +_GL_CXXALIASWARN (close); +#elif @UNISTD_H_HAVE_WINSOCK2_H_AND_USE_SOCKETS@ +# undef close +# define close close_used_without_requesting_gnulib_module_close +#elif defined GNULIB_POSIXCHECK +# undef close +/* Assume close is always declared. */ +_GL_WARN_ON_USE (close, "close does not portably work on sockets - " + "use gnulib module close for portability"); +#endif + + +#if @GNULIB_DUP@ +# if @REPLACE_DUP@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# define dup rpl_dup +# endif +_GL_FUNCDECL_RPL (dup, int, (int oldfd)); +_GL_CXXALIAS_RPL (dup, int, (int oldfd)); +# else +_GL_CXXALIAS_SYS (dup, int, (int oldfd)); +# endif +_GL_CXXALIASWARN (dup); +#elif defined GNULIB_POSIXCHECK +# undef dup +# if HAVE_RAW_DECL_DUP +_GL_WARN_ON_USE (dup, "dup is unportable - " + "use gnulib module dup for portability"); +# endif +#endif + + +#if @GNULIB_DUP2@ +/* Copy the file descriptor OLDFD into file descriptor NEWFD. Do nothing if + NEWFD = OLDFD, otherwise close NEWFD first if it is open. + Return newfd if successful, otherwise -1 and errno set. + See the POSIX:2008 specification + <http://pubs.opengroup.org/onlinepubs/9699919799/functions/dup2.html>. */ +# if @REPLACE_DUP2@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# define dup2 rpl_dup2 +# endif +_GL_FUNCDECL_RPL (dup2, int, (int oldfd, int newfd)); +_GL_CXXALIAS_RPL (dup2, int, (int oldfd, int newfd)); +# else +# if !@HAVE_DUP2@ +_GL_FUNCDECL_SYS (dup2, int, (int oldfd, int newfd)); +# endif +_GL_CXXALIAS_SYS (dup2, int, (int oldfd, int newfd)); +# endif +_GL_CXXALIASWARN (dup2); +#elif defined GNULIB_POSIXCHECK +# undef dup2 +# if HAVE_RAW_DECL_DUP2 +_GL_WARN_ON_USE (dup2, "dup2 is unportable - " + "use gnulib module dup2 for portability"); +# endif +#endif + + +#if @GNULIB_DUP3@ +/* Copy the file descriptor OLDFD into file descriptor NEWFD, with the + specified flags. + The flags are a bitmask, possibly including O_CLOEXEC (defined in <fcntl.h>) + and O_TEXT, O_BINARY (defined in "binary-io.h"). + Close NEWFD first if it is open. + Return newfd if successful, otherwise -1 and errno set. + See the Linux man page at + <https://www.kernel.org/doc/man-pages/online/pages/man2/dup3.2.html>. */ +# if @HAVE_DUP3@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# define dup3 rpl_dup3 +# endif +_GL_FUNCDECL_RPL (dup3, int, (int oldfd, int newfd, int flags)); +_GL_CXXALIAS_RPL (dup3, int, (int oldfd, int newfd, int flags)); +# else +_GL_FUNCDECL_SYS (dup3, int, (int oldfd, int newfd, int flags)); +_GL_CXXALIAS_SYS (dup3, int, (int oldfd, int newfd, int flags)); +# endif +_GL_CXXALIASWARN (dup3); +#elif defined GNULIB_POSIXCHECK +# undef dup3 +# if HAVE_RAW_DECL_DUP3 +_GL_WARN_ON_USE (dup3, "dup3 is unportable - " + "use gnulib module dup3 for portability"); +# endif +#endif + + +#if @GNULIB_ENVIRON@ +# if defined __CYGWIN__ && !defined __i386__ +/* The 'environ' variable is defined in a DLL. Therefore its declaration needs + the '__declspec(dllimport)' attribute, but the system's <unistd.h> lacks it. + This leads to a link error on 64-bit Cygwin when the option + -Wl,--disable-auto-import is in use. */ +_GL_EXTERN_C __declspec(dllimport) char **environ; +# endif +# if !@HAVE_DECL_ENVIRON@ +/* Set of environment variables and values. An array of strings of the form + "VARIABLE=VALUE", terminated with a NULL. */ +# if defined __APPLE__ && defined __MACH__ +# include <TargetConditionals.h> +# if !TARGET_OS_IPHONE && !TARGET_IPHONE_SIMULATOR +# define _GL_USE_CRT_EXTERNS +# endif +# endif +# ifdef _GL_USE_CRT_EXTERNS +# include <crt_externs.h> +# define environ (*_NSGetEnviron ()) +# else +# ifdef __cplusplus +extern "C" { +# endif +extern char **environ; +# ifdef __cplusplus +} +# endif +# endif +# endif +#elif defined GNULIB_POSIXCHECK +# if HAVE_RAW_DECL_ENVIRON +_GL_UNISTD_INLINE char *** +_GL_WARN_ON_USE_ATTRIBUTE ("environ is unportable - " + "use gnulib module environ for portability") +rpl_environ (void) +{ + return &environ; +} +# undef environ +# define environ (*rpl_environ ()) +# endif +#endif + + +#if @GNULIB_EUIDACCESS@ +/* Like access(), except that it uses the effective user id and group id of + the current process. */ +# if !@HAVE_EUIDACCESS@ +_GL_FUNCDECL_SYS (euidaccess, int, (const char *filename, int mode) + _GL_ARG_NONNULL ((1))); +# endif +_GL_CXXALIAS_SYS (euidaccess, int, (const char *filename, int mode)); +_GL_CXXALIASWARN (euidaccess); +# if defined GNULIB_POSIXCHECK +/* Like access(), this function is a security risk. */ +_GL_WARN_ON_USE (euidaccess, "the euidaccess function is a security risk - " + "use the gnulib module faccessat instead"); +# endif +#elif defined GNULIB_POSIXCHECK +# undef euidaccess +# if HAVE_RAW_DECL_EUIDACCESS +_GL_WARN_ON_USE (euidaccess, "euidaccess is unportable - " + "use gnulib module euidaccess for portability"); +# endif +#endif + + +#if @GNULIB_FACCESSAT@ +# if @REPLACE_FACCESSAT@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef faccessat +# define faccessat rpl_faccessat +# endif +_GL_FUNCDECL_RPL (faccessat, int, + (int fd, char const *name, int mode, int flag) + _GL_ARG_NONNULL ((2))); +_GL_CXXALIAS_RPL (faccessat, int, + (int fd, char const *name, int mode, int flag)); +# else +# if !@HAVE_FACCESSAT@ +_GL_FUNCDECL_SYS (faccessat, int, + (int fd, char const *file, int mode, int flag) + _GL_ARG_NONNULL ((2))); +# endif +_GL_CXXALIAS_SYS (faccessat, int, + (int fd, char const *file, int mode, int flag)); +# endif +_GL_CXXALIASWARN (faccessat); +#elif defined GNULIB_POSIXCHECK +# undef faccessat +# if HAVE_RAW_DECL_FACCESSAT +_GL_WARN_ON_USE (faccessat, "faccessat is not portable - " + "use gnulib module faccessat for portability"); +# endif +#endif + + +#if @GNULIB_FCHDIR@ +/* Change the process' current working directory to the directory on which + the given file descriptor is open. + Return 0 if successful, otherwise -1 and errno set. + See the POSIX:2008 specification + <http://pubs.opengroup.org/onlinepubs/9699919799/functions/fchdir.html>. */ +# if ! @HAVE_FCHDIR@ +_GL_FUNCDECL_SYS (fchdir, int, (int /*fd*/)); + +/* Gnulib internal hooks needed to maintain the fchdir metadata. */ +_GL_EXTERN_C int _gl_register_fd (int fd, const char *filename) + _GL_ARG_NONNULL ((2)); +_GL_EXTERN_C void _gl_unregister_fd (int fd); +_GL_EXTERN_C int _gl_register_dup (int oldfd, int newfd); +_GL_EXTERN_C const char *_gl_directory_name (int fd); + +# else +# if !@HAVE_DECL_FCHDIR@ +_GL_FUNCDECL_SYS (fchdir, int, (int /*fd*/)); +# endif +# endif +_GL_CXXALIAS_SYS (fchdir, int, (int /*fd*/)); +_GL_CXXALIASWARN (fchdir); +#elif defined GNULIB_POSIXCHECK +# undef fchdir +# if HAVE_RAW_DECL_FCHDIR +_GL_WARN_ON_USE (fchdir, "fchdir is unportable - " + "use gnulib module fchdir for portability"); +# endif +#endif + + +#if @GNULIB_FCHOWNAT@ +# if @REPLACE_FCHOWNAT@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef fchownat +# define fchownat rpl_fchownat +# endif +_GL_FUNCDECL_RPL (fchownat, int, (int fd, char const *file, + uid_t owner, gid_t group, int flag) + _GL_ARG_NONNULL ((2))); +_GL_CXXALIAS_RPL (fchownat, int, (int fd, char const *file, + uid_t owner, gid_t group, int flag)); +# else +# if !@HAVE_FCHOWNAT@ +_GL_FUNCDECL_SYS (fchownat, int, (int fd, char const *file, + uid_t owner, gid_t group, int flag) + _GL_ARG_NONNULL ((2))); +# endif +_GL_CXXALIAS_SYS (fchownat, int, (int fd, char const *file, + uid_t owner, gid_t group, int flag)); +# endif +_GL_CXXALIASWARN (fchownat); +#elif defined GNULIB_POSIXCHECK +# undef fchownat +# if HAVE_RAW_DECL_FCHOWNAT +_GL_WARN_ON_USE (fchownat, "fchownat is not portable - " + "use gnulib module openat for portability"); +# endif +#endif + + +#if @GNULIB_FDATASYNC@ +/* Synchronize changes to a file. + Return 0 if successful, otherwise -1 and errno set. + See POSIX:2008 specification + <http://pubs.opengroup.org/onlinepubs/9699919799/functions/fdatasync.html>. */ +# if !@HAVE_FDATASYNC@ || !@HAVE_DECL_FDATASYNC@ +_GL_FUNCDECL_SYS (fdatasync, int, (int fd)); +# endif +_GL_CXXALIAS_SYS (fdatasync, int, (int fd)); +_GL_CXXALIASWARN (fdatasync); +#elif defined GNULIB_POSIXCHECK +# undef fdatasync +# if HAVE_RAW_DECL_FDATASYNC +_GL_WARN_ON_USE (fdatasync, "fdatasync is unportable - " + "use gnulib module fdatasync for portability"); +# endif +#endif + + +#if @GNULIB_FSYNC@ +/* Synchronize changes, including metadata, to a file. + Return 0 if successful, otherwise -1 and errno set. + See POSIX:2008 specification + <http://pubs.opengroup.org/onlinepubs/9699919799/functions/fsync.html>. */ +# if !@HAVE_FSYNC@ +_GL_FUNCDECL_SYS (fsync, int, (int fd)); +# endif +_GL_CXXALIAS_SYS (fsync, int, (int fd)); +_GL_CXXALIASWARN (fsync); +#elif defined GNULIB_POSIXCHECK +# undef fsync +# if HAVE_RAW_DECL_FSYNC +_GL_WARN_ON_USE (fsync, "fsync is unportable - " + "use gnulib module fsync for portability"); +# endif +#endif + + +#if @GNULIB_FTRUNCATE@ +/* Change the size of the file to which FD is opened to become equal to LENGTH. + Return 0 if successful, otherwise -1 and errno set. + See the POSIX:2008 specification + <http://pubs.opengroup.org/onlinepubs/9699919799/functions/ftruncate.html>. */ +# if @REPLACE_FTRUNCATE@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef ftruncate +# define ftruncate rpl_ftruncate +# endif +_GL_FUNCDECL_RPL (ftruncate, int, (int fd, off_t length)); +_GL_CXXALIAS_RPL (ftruncate, int, (int fd, off_t length)); +# else +# if !@HAVE_FTRUNCATE@ +_GL_FUNCDECL_SYS (ftruncate, int, (int fd, off_t length)); +# endif +_GL_CXXALIAS_SYS (ftruncate, int, (int fd, off_t length)); +# endif +_GL_CXXALIASWARN (ftruncate); +#elif defined GNULIB_POSIXCHECK +# undef ftruncate +# if HAVE_RAW_DECL_FTRUNCATE +_GL_WARN_ON_USE (ftruncate, "ftruncate is unportable - " + "use gnulib module ftruncate for portability"); +# endif +#endif + + +#if @GNULIB_GETCWD@ +/* Get the name of the current working directory, and put it in SIZE bytes + of BUF. + Return BUF if successful, or NULL if the directory couldn't be determined + or SIZE was too small. + See the POSIX:2008 specification + <http://pubs.opengroup.org/onlinepubs/9699919799/functions/getcwd.html>. + Additionally, the gnulib module 'getcwd' guarantees the following GNU + extension: If BUF is NULL, an array is allocated with 'malloc'; the array + is SIZE bytes long, unless SIZE == 0, in which case it is as big as + necessary. */ +# if @REPLACE_GETCWD@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# define getcwd rpl_getcwd +# endif +_GL_FUNCDECL_RPL (getcwd, char *, (char *buf, size_t size)); +_GL_CXXALIAS_RPL (getcwd, char *, (char *buf, size_t size)); +# else +/* Need to cast, because on mingw, the second parameter is + int size. */ +_GL_CXXALIAS_SYS_CAST (getcwd, char *, (char *buf, size_t size)); +# endif +_GL_CXXALIASWARN (getcwd); +#elif defined GNULIB_POSIXCHECK +# undef getcwd +# if HAVE_RAW_DECL_GETCWD +_GL_WARN_ON_USE (getcwd, "getcwd is unportable - " + "use gnulib module getcwd for portability"); +# endif +#endif + + +#if @GNULIB_GETDOMAINNAME@ +/* Return the NIS domain name of the machine. + WARNING! The NIS domain name is unrelated to the fully qualified host name + of the machine. It is also unrelated to email addresses. + WARNING! The NIS domain name is usually the empty string or "(none)" when + not using NIS. + + Put up to LEN bytes of the NIS domain name into NAME. + Null terminate it if the name is shorter than LEN. + If the NIS domain name is longer than LEN, set errno = EINVAL and return -1. + Return 0 if successful, otherwise set errno and return -1. */ +# if @REPLACE_GETDOMAINNAME@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef getdomainname +# define getdomainname rpl_getdomainname +# endif +_GL_FUNCDECL_RPL (getdomainname, int, (char *name, size_t len) + _GL_ARG_NONNULL ((1))); +_GL_CXXALIAS_RPL (getdomainname, int, (char *name, size_t len)); +# else +# if !@HAVE_DECL_GETDOMAINNAME@ +_GL_FUNCDECL_SYS (getdomainname, int, (char *name, size_t len) + _GL_ARG_NONNULL ((1))); +# endif +_GL_CXXALIAS_SYS (getdomainname, int, (char *name, size_t len)); +# endif +_GL_CXXALIASWARN (getdomainname); +#elif defined GNULIB_POSIXCHECK +# undef getdomainname +# if HAVE_RAW_DECL_GETDOMAINNAME +_GL_WARN_ON_USE (getdomainname, "getdomainname is unportable - " + "use gnulib module getdomainname for portability"); +# endif +#endif + + +#if @GNULIB_GETDTABLESIZE@ +/* Return the maximum number of file descriptors in the current process. + In POSIX, this is same as sysconf (_SC_OPEN_MAX). */ +# if @REPLACE_GETDTABLESIZE@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef getdtablesize +# define getdtablesize rpl_getdtablesize +# endif +_GL_FUNCDECL_RPL (getdtablesize, int, (void)); +_GL_CXXALIAS_RPL (getdtablesize, int, (void)); +# else +# if !@HAVE_GETDTABLESIZE@ +_GL_FUNCDECL_SYS (getdtablesize, int, (void)); +# endif +_GL_CXXALIAS_SYS (getdtablesize, int, (void)); +# endif +_GL_CXXALIASWARN (getdtablesize); +#elif defined GNULIB_POSIXCHECK +# undef getdtablesize +# if HAVE_RAW_DECL_GETDTABLESIZE +_GL_WARN_ON_USE (getdtablesize, "getdtablesize is unportable - " + "use gnulib module getdtablesize for portability"); +# endif +#endif + + +#if @GNULIB_GETGROUPS@ +/* Return the supplemental groups that the current process belongs to. + It is unspecified whether the effective group id is in the list. + If N is 0, return the group count; otherwise, N describes how many + entries are available in GROUPS. Return -1 and set errno if N is + not 0 and not large enough. Fails with ENOSYS on some systems. */ +# if @REPLACE_GETGROUPS@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef getgroups +# define getgroups rpl_getgroups +# endif +_GL_FUNCDECL_RPL (getgroups, int, (int n, gid_t *groups)); +_GL_CXXALIAS_RPL (getgroups, int, (int n, gid_t *groups)); +# else +# if !@HAVE_GETGROUPS@ +_GL_FUNCDECL_SYS (getgroups, int, (int n, gid_t *groups)); +# endif +_GL_CXXALIAS_SYS (getgroups, int, (int n, gid_t *groups)); +# endif +_GL_CXXALIASWARN (getgroups); +#elif defined GNULIB_POSIXCHECK +# undef getgroups +# if HAVE_RAW_DECL_GETGROUPS +_GL_WARN_ON_USE (getgroups, "getgroups is unportable - " + "use gnulib module getgroups for portability"); +# endif +#endif + + +#if @GNULIB_GETHOSTNAME@ +/* Return the standard host name of the machine. + WARNING! The host name may or may not be fully qualified. + + Put up to LEN bytes of the host name into NAME. + Null terminate it if the name is shorter than LEN. + If the host name is longer than LEN, set errno = EINVAL and return -1. + Return 0 if successful, otherwise set errno and return -1. */ +# if @UNISTD_H_HAVE_WINSOCK2_H@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef gethostname +# define gethostname rpl_gethostname +# endif +_GL_FUNCDECL_RPL (gethostname, int, (char *name, size_t len) + _GL_ARG_NONNULL ((1))); +_GL_CXXALIAS_RPL (gethostname, int, (char *name, size_t len)); +# else +# if !@HAVE_GETHOSTNAME@ +_GL_FUNCDECL_SYS (gethostname, int, (char *name, size_t len) + _GL_ARG_NONNULL ((1))); +# endif +/* Need to cast, because on Solaris 10 and OSF/1 5.1 systems, the second + parameter is + int len. */ +_GL_CXXALIAS_SYS_CAST (gethostname, int, (char *name, size_t len)); +# endif +_GL_CXXALIASWARN (gethostname); +#elif @UNISTD_H_HAVE_WINSOCK2_H@ +# undef gethostname +# define gethostname gethostname_used_without_requesting_gnulib_module_gethostname +#elif defined GNULIB_POSIXCHECK +# undef gethostname +# if HAVE_RAW_DECL_GETHOSTNAME +_GL_WARN_ON_USE (gethostname, "gethostname is unportable - " + "use gnulib module gethostname for portability"); +# endif +#endif + + +#if @GNULIB_GETLOGIN@ +/* Returns the user's login name, or NULL if it cannot be found. Upon error, + returns NULL with errno set. + + See <http://www.opengroup.org/susv3xsh/getlogin.html>. + + Most programs don't need to use this function, because the information is + available through environment variables: + ${LOGNAME-$USER} on Unix platforms, + $USERNAME on native Windows platforms. + */ +# if !@HAVE_DECL_GETLOGIN@ +_GL_FUNCDECL_SYS (getlogin, char *, (void)); +# endif +_GL_CXXALIAS_SYS (getlogin, char *, (void)); +_GL_CXXALIASWARN (getlogin); +#elif defined GNULIB_POSIXCHECK +# undef getlogin +# if HAVE_RAW_DECL_GETLOGIN +_GL_WARN_ON_USE (getlogin, "getlogin is unportable - " + "use gnulib module getlogin for portability"); +# endif +#endif + + +#if @GNULIB_GETLOGIN_R@ +/* Copies the user's login name to NAME. + The array pointed to by NAME has room for SIZE bytes. + + Returns 0 if successful. Upon error, an error number is returned, or -1 in + the case that the login name cannot be found but no specific error is + provided (this case is hopefully rare but is left open by the POSIX spec). + + See <http://www.opengroup.org/susv3xsh/getlogin.html>. + + Most programs don't need to use this function, because the information is + available through environment variables: + ${LOGNAME-$USER} on Unix platforms, + $USERNAME on native Windows platforms. + */ +# if @REPLACE_GETLOGIN_R@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# define getlogin_r rpl_getlogin_r +# endif +_GL_FUNCDECL_RPL (getlogin_r, int, (char *name, size_t size) + _GL_ARG_NONNULL ((1))); +_GL_CXXALIAS_RPL (getlogin_r, int, (char *name, size_t size)); +# else +# if !@HAVE_DECL_GETLOGIN_R@ +_GL_FUNCDECL_SYS (getlogin_r, int, (char *name, size_t size) + _GL_ARG_NONNULL ((1))); +# endif +/* Need to cast, because on Solaris 10 systems, the second argument is + int size. */ +_GL_CXXALIAS_SYS_CAST (getlogin_r, int, (char *name, size_t size)); +# endif +_GL_CXXALIASWARN (getlogin_r); +#elif defined GNULIB_POSIXCHECK +# undef getlogin_r +# if HAVE_RAW_DECL_GETLOGIN_R +_GL_WARN_ON_USE (getlogin_r, "getlogin_r is unportable - " + "use gnulib module getlogin_r for portability"); +# endif +#endif + + +#if @GNULIB_GETPAGESIZE@ +# if @REPLACE_GETPAGESIZE@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# define getpagesize rpl_getpagesize +# endif +_GL_FUNCDECL_RPL (getpagesize, int, (void)); +_GL_CXXALIAS_RPL (getpagesize, int, (void)); +# else +# if !@HAVE_GETPAGESIZE@ +# if !defined getpagesize +/* This is for POSIX systems. */ +# if !defined _gl_getpagesize && defined _SC_PAGESIZE +# if ! (defined __VMS && __VMS_VER < 70000000) +# define _gl_getpagesize() sysconf (_SC_PAGESIZE) +# endif +# endif +/* This is for older VMS. */ +# if !defined _gl_getpagesize && defined __VMS +# ifdef __ALPHA +# define _gl_getpagesize() 8192 +# else +# define _gl_getpagesize() 512 +# endif +# endif +/* This is for BeOS. */ +# if !defined _gl_getpagesize && @HAVE_OS_H@ +# include <OS.h> +# if defined B_PAGE_SIZE +# define _gl_getpagesize() B_PAGE_SIZE +# endif +# endif +/* This is for AmigaOS4.0. */ +# if !defined _gl_getpagesize && defined __amigaos4__ +# define _gl_getpagesize() 2048 +# endif +/* This is for older Unix systems. */ +# if !defined _gl_getpagesize && @HAVE_SYS_PARAM_H@ +# include <sys/param.h> +# ifdef EXEC_PAGESIZE +# define _gl_getpagesize() EXEC_PAGESIZE +# else +# ifdef NBPG +# ifndef CLSIZE +# define CLSIZE 1 +# endif +# define _gl_getpagesize() (NBPG * CLSIZE) +# else +# ifdef NBPC +# define _gl_getpagesize() NBPC +# endif +# endif +# endif +# endif +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# define getpagesize() _gl_getpagesize () +# else +# if !GNULIB_defined_getpagesize_function +_GL_UNISTD_INLINE int +getpagesize () +{ + return _gl_getpagesize (); +} +# define GNULIB_defined_getpagesize_function 1 +# endif +# endif +# endif +# endif +/* Need to cast, because on Cygwin 1.5.x systems, the return type is size_t. */ +_GL_CXXALIAS_SYS_CAST (getpagesize, int, (void)); +# endif +# if @HAVE_DECL_GETPAGESIZE@ +_GL_CXXALIASWARN (getpagesize); +# endif +#elif defined GNULIB_POSIXCHECK +# undef getpagesize +# if HAVE_RAW_DECL_GETPAGESIZE +_GL_WARN_ON_USE (getpagesize, "getpagesize is unportable - " + "use gnulib module getpagesize for portability"); +# endif +#endif + + +#if @GNULIB_GETPASS@ +/* Function getpass() from module 'getpass': + Read a password from /dev/tty or stdin. + Function getpass() from module 'getpass-gnu': + Read a password of arbitrary length from /dev/tty or stdin. */ +# if @REPLACE_GETPASS@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef getpass +# define getpass rpl_getpass +# endif +_GL_FUNCDECL_RPL (getpass, char *, (const char *prompt) + _GL_ARG_NONNULL ((1))); +_GL_CXXALIAS_RPL (getpass, char *, (const char *prompt)); +# else +# if !@HAVE_GETPASS@ +_GL_FUNCDECL_SYS (getpass, char *, (const char *prompt) + _GL_ARG_NONNULL ((1))); +# endif +_GL_CXXALIAS_SYS (getpass, char *, (const char *prompt)); +# endif +_GL_CXXALIASWARN (getpass); +#elif defined GNULIB_POSIXCHECK +# undef getpass +# if HAVE_RAW_DECL_GETPASS +_GL_WARN_ON_USE (getpass, "getpass is unportable - " + "use gnulib module getpass or getpass-gnu for portability"); +# endif +#endif + + +#if @GNULIB_GETUSERSHELL@ +/* Return the next valid login shell on the system, or NULL when the end of + the list has been reached. */ +# if !@HAVE_DECL_GETUSERSHELL@ +_GL_FUNCDECL_SYS (getusershell, char *, (void)); +# endif +_GL_CXXALIAS_SYS (getusershell, char *, (void)); +_GL_CXXALIASWARN (getusershell); +#elif defined GNULIB_POSIXCHECK +# undef getusershell +# if HAVE_RAW_DECL_GETUSERSHELL +_GL_WARN_ON_USE (getusershell, "getusershell is unportable - " + "use gnulib module getusershell for portability"); +# endif +#endif + +#if @GNULIB_GETUSERSHELL@ +/* Rewind to pointer that is advanced at each getusershell() call. */ +# if !@HAVE_DECL_GETUSERSHELL@ +_GL_FUNCDECL_SYS (setusershell, void, (void)); +# endif +_GL_CXXALIAS_SYS (setusershell, void, (void)); +_GL_CXXALIASWARN (setusershell); +#elif defined GNULIB_POSIXCHECK +# undef setusershell +# if HAVE_RAW_DECL_SETUSERSHELL +_GL_WARN_ON_USE (setusershell, "setusershell is unportable - " + "use gnulib module getusershell for portability"); +# endif +#endif + +#if @GNULIB_GETUSERSHELL@ +/* Free the pointer that is advanced at each getusershell() call and + associated resources. */ +# if !@HAVE_DECL_GETUSERSHELL@ +_GL_FUNCDECL_SYS (endusershell, void, (void)); +# endif +_GL_CXXALIAS_SYS (endusershell, void, (void)); +_GL_CXXALIASWARN (endusershell); +#elif defined GNULIB_POSIXCHECK +# undef endusershell +# if HAVE_RAW_DECL_ENDUSERSHELL +_GL_WARN_ON_USE (endusershell, "endusershell is unportable - " + "use gnulib module getusershell for portability"); +# endif +#endif + + +#if @GNULIB_GROUP_MEMBER@ +/* Determine whether group id is in calling user's group list. */ +# if !@HAVE_GROUP_MEMBER@ +_GL_FUNCDECL_SYS (group_member, int, (gid_t gid)); +# endif +_GL_CXXALIAS_SYS (group_member, int, (gid_t gid)); +_GL_CXXALIASWARN (group_member); +#elif defined GNULIB_POSIXCHECK +# undef group_member +# if HAVE_RAW_DECL_GROUP_MEMBER +_GL_WARN_ON_USE (group_member, "group_member is unportable - " + "use gnulib module group-member for portability"); +# endif +#endif + + +#if @GNULIB_ISATTY@ +# if @REPLACE_ISATTY@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef isatty +# define isatty rpl_isatty +# endif +_GL_FUNCDECL_RPL (isatty, int, (int fd)); +_GL_CXXALIAS_RPL (isatty, int, (int fd)); +# else +_GL_CXXALIAS_SYS (isatty, int, (int fd)); +# endif +_GL_CXXALIASWARN (isatty); +#elif defined GNULIB_POSIXCHECK +# undef isatty +# if HAVE_RAW_DECL_ISATTY +_GL_WARN_ON_USE (isatty, "isatty has portability problems on native Windows - " + "use gnulib module isatty for portability"); +# endif +#endif + + +#if @GNULIB_LCHOWN@ +/* Change the owner of FILE to UID (if UID is not -1) and the group of FILE + to GID (if GID is not -1). Do not follow symbolic links. + Return 0 if successful, otherwise -1 and errno set. + See the POSIX:2008 specification + <http://pubs.opengroup.org/onlinepubs/9699919799/functions/lchown.html>. */ +# if @REPLACE_LCHOWN@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef lchown +# define lchown rpl_lchown +# endif +_GL_FUNCDECL_RPL (lchown, int, (char const *file, uid_t owner, gid_t group) + _GL_ARG_NONNULL ((1))); +_GL_CXXALIAS_RPL (lchown, int, (char const *file, uid_t owner, gid_t group)); +# else +# if !@HAVE_LCHOWN@ +_GL_FUNCDECL_SYS (lchown, int, (char const *file, uid_t owner, gid_t group) + _GL_ARG_NONNULL ((1))); +# endif +_GL_CXXALIAS_SYS (lchown, int, (char const *file, uid_t owner, gid_t group)); +# endif +_GL_CXXALIASWARN (lchown); +#elif defined GNULIB_POSIXCHECK +# undef lchown +# if HAVE_RAW_DECL_LCHOWN +_GL_WARN_ON_USE (lchown, "lchown is unportable to pre-POSIX.1-2001 systems - " + "use gnulib module lchown for portability"); +# endif +#endif + + +#if @GNULIB_LINK@ +/* Create a new hard link for an existing file. + Return 0 if successful, otherwise -1 and errno set. + See POSIX:2008 specification + <http://pubs.opengroup.org/onlinepubs/9699919799/functions/link.html>. */ +# if @REPLACE_LINK@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# define link rpl_link +# endif +_GL_FUNCDECL_RPL (link, int, (const char *path1, const char *path2) + _GL_ARG_NONNULL ((1, 2))); +_GL_CXXALIAS_RPL (link, int, (const char *path1, const char *path2)); +# else +# if !@HAVE_LINK@ +_GL_FUNCDECL_SYS (link, int, (const char *path1, const char *path2) + _GL_ARG_NONNULL ((1, 2))); +# endif +_GL_CXXALIAS_SYS (link, int, (const char *path1, const char *path2)); +# endif +_GL_CXXALIASWARN (link); +#elif defined GNULIB_POSIXCHECK +# undef link +# if HAVE_RAW_DECL_LINK +_GL_WARN_ON_USE (link, "link is unportable - " + "use gnulib module link for portability"); +# endif +#endif + + +#if @GNULIB_LINKAT@ +/* Create a new hard link for an existing file, relative to two + directories. FLAG controls whether symlinks are followed. + Return 0 if successful, otherwise -1 and errno set. */ +# if @REPLACE_LINKAT@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef linkat +# define linkat rpl_linkat +# endif +_GL_FUNCDECL_RPL (linkat, int, + (int fd1, const char *path1, int fd2, const char *path2, + int flag) + _GL_ARG_NONNULL ((2, 4))); +_GL_CXXALIAS_RPL (linkat, int, + (int fd1, const char *path1, int fd2, const char *path2, + int flag)); +# else +# if !@HAVE_LINKAT@ +_GL_FUNCDECL_SYS (linkat, int, + (int fd1, const char *path1, int fd2, const char *path2, + int flag) + _GL_ARG_NONNULL ((2, 4))); +# endif +_GL_CXXALIAS_SYS (linkat, int, + (int fd1, const char *path1, int fd2, const char *path2, + int flag)); +# endif +_GL_CXXALIASWARN (linkat); +#elif defined GNULIB_POSIXCHECK +# undef linkat +# if HAVE_RAW_DECL_LINKAT +_GL_WARN_ON_USE (linkat, "linkat is unportable - " + "use gnulib module linkat for portability"); +# endif +#endif + + +#if @GNULIB_LSEEK@ +/* Set the offset of FD relative to SEEK_SET, SEEK_CUR, or SEEK_END. + Return the new offset if successful, otherwise -1 and errno set. + See the POSIX:2008 specification + <http://pubs.opengroup.org/onlinepubs/9699919799/functions/lseek.html>. */ +# if @REPLACE_LSEEK@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# define lseek rpl_lseek +# endif +_GL_FUNCDECL_RPL (lseek, off_t, (int fd, off_t offset, int whence)); +_GL_CXXALIAS_RPL (lseek, off_t, (int fd, off_t offset, int whence)); +# else +_GL_CXXALIAS_SYS (lseek, off_t, (int fd, off_t offset, int whence)); +# endif +_GL_CXXALIASWARN (lseek); +#elif defined GNULIB_POSIXCHECK +# undef lseek +# if HAVE_RAW_DECL_LSEEK +_GL_WARN_ON_USE (lseek, "lseek does not fail with ESPIPE on pipes on some " + "systems - use gnulib module lseek for portability"); +# endif +#endif + + +#if @GNULIB_PIPE@ +/* Create a pipe, defaulting to O_BINARY mode. + Store the read-end as fd[0] and the write-end as fd[1]. + Return 0 upon success, or -1 with errno set upon failure. */ +# if !@HAVE_PIPE@ +_GL_FUNCDECL_SYS (pipe, int, (int fd[2]) _GL_ARG_NONNULL ((1))); +# endif +_GL_CXXALIAS_SYS (pipe, int, (int fd[2])); +_GL_CXXALIASWARN (pipe); +#elif defined GNULIB_POSIXCHECK +# undef pipe +# if HAVE_RAW_DECL_PIPE +_GL_WARN_ON_USE (pipe, "pipe is unportable - " + "use gnulib module pipe-posix for portability"); +# endif +#endif + + +#if @GNULIB_PIPE2@ +/* Create a pipe, applying the given flags when opening the read-end of the + pipe and the write-end of the pipe. + The flags are a bitmask, possibly including O_CLOEXEC (defined in <fcntl.h>) + and O_TEXT, O_BINARY (defined in "binary-io.h"). + Store the read-end as fd[0] and the write-end as fd[1]. + Return 0 upon success, or -1 with errno set upon failure. + See also the Linux man page at + <https://www.kernel.org/doc/man-pages/online/pages/man2/pipe2.2.html>. */ +# if @HAVE_PIPE2@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# define pipe2 rpl_pipe2 +# endif +_GL_FUNCDECL_RPL (pipe2, int, (int fd[2], int flags) _GL_ARG_NONNULL ((1))); +_GL_CXXALIAS_RPL (pipe2, int, (int fd[2], int flags)); +# else +_GL_FUNCDECL_SYS (pipe2, int, (int fd[2], int flags) _GL_ARG_NONNULL ((1))); +_GL_CXXALIAS_SYS (pipe2, int, (int fd[2], int flags)); +# endif +_GL_CXXALIASWARN (pipe2); +#elif defined GNULIB_POSIXCHECK +# undef pipe2 +# if HAVE_RAW_DECL_PIPE2 +_GL_WARN_ON_USE (pipe2, "pipe2 is unportable - " + "use gnulib module pipe2 for portability"); +# endif +#endif + + +#if @GNULIB_PREAD@ +/* Read at most BUFSIZE bytes from FD into BUF, starting at OFFSET. + Return the number of bytes placed into BUF if successful, otherwise + set errno and return -1. 0 indicates EOF. + See the POSIX:2008 specification + <http://pubs.opengroup.org/onlinepubs/9699919799/functions/pread.html>. */ +# if @REPLACE_PREAD@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef pread +# define pread rpl_pread +# endif +_GL_FUNCDECL_RPL (pread, ssize_t, + (int fd, void *buf, size_t bufsize, off_t offset) + _GL_ARG_NONNULL ((2))); +_GL_CXXALIAS_RPL (pread, ssize_t, + (int fd, void *buf, size_t bufsize, off_t offset)); +# else +# if !@HAVE_PREAD@ +_GL_FUNCDECL_SYS (pread, ssize_t, + (int fd, void *buf, size_t bufsize, off_t offset) + _GL_ARG_NONNULL ((2))); +# endif +_GL_CXXALIAS_SYS (pread, ssize_t, + (int fd, void *buf, size_t bufsize, off_t offset)); +# endif +_GL_CXXALIASWARN (pread); +#elif defined GNULIB_POSIXCHECK +# undef pread +# if HAVE_RAW_DECL_PREAD +_GL_WARN_ON_USE (pread, "pread is unportable - " + "use gnulib module pread for portability"); +# endif +#endif + + +#if @GNULIB_PWRITE@ +/* Write at most BUFSIZE bytes from BUF into FD, starting at OFFSET. + Return the number of bytes written if successful, otherwise + set errno and return -1. 0 indicates nothing written. See the + POSIX:2008 specification + <http://pubs.opengroup.org/onlinepubs/9699919799/functions/pwrite.html>. */ +# if @REPLACE_PWRITE@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef pwrite +# define pwrite rpl_pwrite +# endif +_GL_FUNCDECL_RPL (pwrite, ssize_t, + (int fd, const void *buf, size_t bufsize, off_t offset) + _GL_ARG_NONNULL ((2))); +_GL_CXXALIAS_RPL (pwrite, ssize_t, + (int fd, const void *buf, size_t bufsize, off_t offset)); +# else +# if !@HAVE_PWRITE@ +_GL_FUNCDECL_SYS (pwrite, ssize_t, + (int fd, const void *buf, size_t bufsize, off_t offset) + _GL_ARG_NONNULL ((2))); +# endif +_GL_CXXALIAS_SYS (pwrite, ssize_t, + (int fd, const void *buf, size_t bufsize, off_t offset)); +# endif +_GL_CXXALIASWARN (pwrite); +#elif defined GNULIB_POSIXCHECK +# undef pwrite +# if HAVE_RAW_DECL_PWRITE +_GL_WARN_ON_USE (pwrite, "pwrite is unportable - " + "use gnulib module pwrite for portability"); +# endif +#endif + + +#if @GNULIB_READ@ +/* Read up to COUNT bytes from file descriptor FD into the buffer starting + at BUF. See the POSIX:2008 specification + <http://pubs.opengroup.org/onlinepubs/9699919799/functions/read.html>. */ +# if @REPLACE_READ@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef read +# define read rpl_read +# endif +_GL_FUNCDECL_RPL (read, ssize_t, (int fd, void *buf, size_t count) + _GL_ARG_NONNULL ((2))); +_GL_CXXALIAS_RPL (read, ssize_t, (int fd, void *buf, size_t count)); +# else +/* Need to cast, because on mingw, the third parameter is + unsigned int count + and the return type is 'int'. */ +_GL_CXXALIAS_SYS_CAST (read, ssize_t, (int fd, void *buf, size_t count)); +# endif +_GL_CXXALIASWARN (read); +#endif + + +#if @GNULIB_READLINK@ +/* Read the contents of the symbolic link FILE and place the first BUFSIZE + bytes of it into BUF. Return the number of bytes placed into BUF if + successful, otherwise -1 and errno set. + See the POSIX:2008 specification + <http://pubs.opengroup.org/onlinepubs/9699919799/functions/readlink.html>. */ +# if @REPLACE_READLINK@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# define readlink rpl_readlink +# endif +_GL_FUNCDECL_RPL (readlink, ssize_t, + (const char *file, char *buf, size_t bufsize) + _GL_ARG_NONNULL ((1, 2))); +_GL_CXXALIAS_RPL (readlink, ssize_t, + (const char *file, char *buf, size_t bufsize)); +# else +# if !@HAVE_READLINK@ +_GL_FUNCDECL_SYS (readlink, ssize_t, + (const char *file, char *buf, size_t bufsize) + _GL_ARG_NONNULL ((1, 2))); +# endif +_GL_CXXALIAS_SYS (readlink, ssize_t, + (const char *file, char *buf, size_t bufsize)); +# endif +_GL_CXXALIASWARN (readlink); +#elif defined GNULIB_POSIXCHECK +# undef readlink +# if HAVE_RAW_DECL_READLINK +_GL_WARN_ON_USE (readlink, "readlink is unportable - " + "use gnulib module readlink for portability"); +# endif +#endif + + +#if @GNULIB_READLINKAT@ +# if @REPLACE_READLINKAT@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# define readlinkat rpl_readlinkat +# endif +_GL_FUNCDECL_RPL (readlinkat, ssize_t, + (int fd, char const *file, char *buf, size_t len) + _GL_ARG_NONNULL ((2, 3))); +_GL_CXXALIAS_RPL (readlinkat, ssize_t, + (int fd, char const *file, char *buf, size_t len)); +# else +# if !@HAVE_READLINKAT@ +_GL_FUNCDECL_SYS (readlinkat, ssize_t, + (int fd, char const *file, char *buf, size_t len) + _GL_ARG_NONNULL ((2, 3))); +# endif +_GL_CXXALIAS_SYS (readlinkat, ssize_t, + (int fd, char const *file, char *buf, size_t len)); +# endif +_GL_CXXALIASWARN (readlinkat); +#elif defined GNULIB_POSIXCHECK +# undef readlinkat +# if HAVE_RAW_DECL_READLINKAT +_GL_WARN_ON_USE (readlinkat, "readlinkat is not portable - " + "use gnulib module readlinkat for portability"); +# endif +#endif + + +#if @GNULIB_RMDIR@ +/* Remove the directory DIR. */ +# if @REPLACE_RMDIR@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# define rmdir rpl_rmdir +# endif +_GL_FUNCDECL_RPL (rmdir, int, (char const *name) _GL_ARG_NONNULL ((1))); +_GL_CXXALIAS_RPL (rmdir, int, (char const *name)); +# else +_GL_CXXALIAS_SYS (rmdir, int, (char const *name)); +# endif +_GL_CXXALIASWARN (rmdir); +#elif defined GNULIB_POSIXCHECK +# undef rmdir +# if HAVE_RAW_DECL_RMDIR +_GL_WARN_ON_USE (rmdir, "rmdir is unportable - " + "use gnulib module rmdir for portability"); +# endif +#endif + + +#if @GNULIB_SETHOSTNAME@ +/* Set the host name of the machine. + The host name may or may not be fully qualified. + + Put LEN bytes of NAME into the host name. + Return 0 if successful, otherwise, set errno and return -1. + + Platforms with no ability to set the hostname return -1 and set + errno = ENOSYS. */ +# if !@HAVE_SETHOSTNAME@ || !@HAVE_DECL_SETHOSTNAME@ +_GL_FUNCDECL_SYS (sethostname, int, (const char *name, size_t len) + _GL_ARG_NONNULL ((1))); +# endif +/* Need to cast, because on Solaris 11 2011-10, Mac OS X 10.5, IRIX 6.5 + and FreeBSD 6.4 the second parameter is int. On Solaris 11 + 2011-10, the first parameter is not const. */ +_GL_CXXALIAS_SYS_CAST (sethostname, int, (const char *name, size_t len)); +_GL_CXXALIASWARN (sethostname); +#elif defined GNULIB_POSIXCHECK +# undef sethostname +# if HAVE_RAW_DECL_SETHOSTNAME +_GL_WARN_ON_USE (sethostname, "sethostname is unportable - " + "use gnulib module sethostname for portability"); +# endif +#endif + + +#if @GNULIB_SLEEP@ +/* Pause the execution of the current thread for N seconds. + Returns the number of seconds left to sleep. + See the POSIX:2008 specification + <http://pubs.opengroup.org/onlinepubs/9699919799/functions/sleep.html>. */ +# if @REPLACE_SLEEP@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef sleep +# define sleep rpl_sleep +# endif +_GL_FUNCDECL_RPL (sleep, unsigned int, (unsigned int n)); +_GL_CXXALIAS_RPL (sleep, unsigned int, (unsigned int n)); +# else +# if !@HAVE_SLEEP@ +_GL_FUNCDECL_SYS (sleep, unsigned int, (unsigned int n)); +# endif +_GL_CXXALIAS_SYS (sleep, unsigned int, (unsigned int n)); +# endif +_GL_CXXALIASWARN (sleep); +#elif defined GNULIB_POSIXCHECK +# undef sleep +# if HAVE_RAW_DECL_SLEEP +_GL_WARN_ON_USE (sleep, "sleep is unportable - " + "use gnulib module sleep for portability"); +# endif +#endif + + +#if @GNULIB_SYMLINK@ +# if @REPLACE_SYMLINK@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef symlink +# define symlink rpl_symlink +# endif +_GL_FUNCDECL_RPL (symlink, int, (char const *contents, char const *file) + _GL_ARG_NONNULL ((1, 2))); +_GL_CXXALIAS_RPL (symlink, int, (char const *contents, char const *file)); +# else +# if !@HAVE_SYMLINK@ +_GL_FUNCDECL_SYS (symlink, int, (char const *contents, char const *file) + _GL_ARG_NONNULL ((1, 2))); +# endif +_GL_CXXALIAS_SYS (symlink, int, (char const *contents, char const *file)); +# endif +_GL_CXXALIASWARN (symlink); +#elif defined GNULIB_POSIXCHECK +# undef symlink +# if HAVE_RAW_DECL_SYMLINK +_GL_WARN_ON_USE (symlink, "symlink is not portable - " + "use gnulib module symlink for portability"); +# endif +#endif + + +#if @GNULIB_SYMLINKAT@ +# if @REPLACE_SYMLINKAT@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef symlinkat +# define symlinkat rpl_symlinkat +# endif +_GL_FUNCDECL_RPL (symlinkat, int, + (char const *contents, int fd, char const *file) + _GL_ARG_NONNULL ((1, 3))); +_GL_CXXALIAS_RPL (symlinkat, int, + (char const *contents, int fd, char const *file)); +# else +# if !@HAVE_SYMLINKAT@ +_GL_FUNCDECL_SYS (symlinkat, int, + (char const *contents, int fd, char const *file) + _GL_ARG_NONNULL ((1, 3))); +# endif +_GL_CXXALIAS_SYS (symlinkat, int, + (char const *contents, int fd, char const *file)); +# endif +_GL_CXXALIASWARN (symlinkat); +#elif defined GNULIB_POSIXCHECK +# undef symlinkat +# if HAVE_RAW_DECL_SYMLINKAT +_GL_WARN_ON_USE (symlinkat, "symlinkat is not portable - " + "use gnulib module symlinkat for portability"); +# endif +#endif + + +#if @GNULIB_TRUNCATE@ +/* Change the size of the file designated by FILENAME to become equal to LENGTH. + Return 0 if successful, otherwise -1 and errno set. + See the POSIX:2008 specification + <http://pubs.opengroup.org/onlinepubs/9699919799/functions/truncate.html>. */ +# if @REPLACE_TRUNCATE@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef truncate +# define truncate rpl_truncate +# endif +_GL_FUNCDECL_RPL (truncate, int, (const char *filename, off_t length) + _GL_ARG_NONNULL ((1))); +_GL_CXXALIAS_RPL (truncate, int, (const char *filename, off_t length)); +# else +# if !@HAVE_DECL_TRUNCATE@ +_GL_FUNCDECL_SYS (truncate, int, (const char *filename, off_t length) + _GL_ARG_NONNULL ((1))); +# endif +_GL_CXXALIAS_SYS (truncate, int, (const char *filename, off_t length)); +# endif +_GL_CXXALIASWARN (truncate); +#elif defined GNULIB_POSIXCHECK +# undef truncate +# if HAVE_RAW_DECL_TRUNCATE +_GL_WARN_ON_USE (truncate, "truncate is unportable - " + "use gnulib module truncate for portability"); +# endif +#endif + + +#if @GNULIB_TTYNAME_R@ +/* Store at most BUFLEN characters of the pathname of the terminal FD is + open on in BUF. Return 0 on success, otherwise an error number. */ +# if @REPLACE_TTYNAME_R@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef ttyname_r +# define ttyname_r rpl_ttyname_r +# endif +_GL_FUNCDECL_RPL (ttyname_r, int, + (int fd, char *buf, size_t buflen) _GL_ARG_NONNULL ((2))); +_GL_CXXALIAS_RPL (ttyname_r, int, + (int fd, char *buf, size_t buflen)); +# else +# if !@HAVE_DECL_TTYNAME_R@ +_GL_FUNCDECL_SYS (ttyname_r, int, + (int fd, char *buf, size_t buflen) _GL_ARG_NONNULL ((2))); +# endif +_GL_CXXALIAS_SYS (ttyname_r, int, + (int fd, char *buf, size_t buflen)); +# endif +_GL_CXXALIASWARN (ttyname_r); +#elif defined GNULIB_POSIXCHECK +# undef ttyname_r +# if HAVE_RAW_DECL_TTYNAME_R +_GL_WARN_ON_USE (ttyname_r, "ttyname_r is not portable - " + "use gnulib module ttyname_r for portability"); +# endif +#endif + + +#if @GNULIB_UNLINK@ +# if @REPLACE_UNLINK@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef unlink +# define unlink rpl_unlink +# endif +_GL_FUNCDECL_RPL (unlink, int, (char const *file) _GL_ARG_NONNULL ((1))); +_GL_CXXALIAS_RPL (unlink, int, (char const *file)); +# else +_GL_CXXALIAS_SYS (unlink, int, (char const *file)); +# endif +_GL_CXXALIASWARN (unlink); +#elif defined GNULIB_POSIXCHECK +# undef unlink +# if HAVE_RAW_DECL_UNLINK +_GL_WARN_ON_USE (unlink, "unlink is not portable - " + "use gnulib module unlink for portability"); +# endif +#endif + + +#if @GNULIB_UNLINKAT@ +# if @REPLACE_UNLINKAT@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef unlinkat +# define unlinkat rpl_unlinkat +# endif +_GL_FUNCDECL_RPL (unlinkat, int, (int fd, char const *file, int flag) + _GL_ARG_NONNULL ((2))); +_GL_CXXALIAS_RPL (unlinkat, int, (int fd, char const *file, int flag)); +# else +# if !@HAVE_UNLINKAT@ +_GL_FUNCDECL_SYS (unlinkat, int, (int fd, char const *file, int flag) + _GL_ARG_NONNULL ((2))); +# endif +_GL_CXXALIAS_SYS (unlinkat, int, (int fd, char const *file, int flag)); +# endif +_GL_CXXALIASWARN (unlinkat); +#elif defined GNULIB_POSIXCHECK +# undef unlinkat +# if HAVE_RAW_DECL_UNLINKAT +_GL_WARN_ON_USE (unlinkat, "unlinkat is not portable - " + "use gnulib module openat for portability"); +# endif +#endif + + +#if @GNULIB_USLEEP@ +/* Pause the execution of the current thread for N microseconds. + Returns 0 on completion, or -1 on range error. + See the POSIX:2001 specification + <http://www.opengroup.org/susv3xsh/usleep.html>. */ +# if @REPLACE_USLEEP@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef usleep +# define usleep rpl_usleep +# endif +_GL_FUNCDECL_RPL (usleep, int, (useconds_t n)); +_GL_CXXALIAS_RPL (usleep, int, (useconds_t n)); +# else +# if !@HAVE_USLEEP@ +_GL_FUNCDECL_SYS (usleep, int, (useconds_t n)); +# endif +_GL_CXXALIAS_SYS (usleep, int, (useconds_t n)); +# endif +_GL_CXXALIASWARN (usleep); +#elif defined GNULIB_POSIXCHECK +# undef usleep +# if HAVE_RAW_DECL_USLEEP +_GL_WARN_ON_USE (usleep, "usleep is unportable - " + "use gnulib module usleep for portability"); +# endif +#endif + + +#if @GNULIB_WRITE@ +/* Write up to COUNT bytes starting at BUF to file descriptor FD. + See the POSIX:2008 specification + <http://pubs.opengroup.org/onlinepubs/9699919799/functions/write.html>. */ +# if @REPLACE_WRITE@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef write +# define write rpl_write +# endif +_GL_FUNCDECL_RPL (write, ssize_t, (int fd, const void *buf, size_t count) + _GL_ARG_NONNULL ((2))); +_GL_CXXALIAS_RPL (write, ssize_t, (int fd, const void *buf, size_t count)); +# else +/* Need to cast, because on mingw, the third parameter is + unsigned int count + and the return type is 'int'. */ +_GL_CXXALIAS_SYS_CAST (write, ssize_t, (int fd, const void *buf, size_t count)); +# endif +_GL_CXXALIASWARN (write); +#endif + +_GL_INLINE_HEADER_END + +#endif /* _@GUARD_PREFIX@_UNISTD_H */ +#endif /* _GL_INCLUDING_UNISTD_H */ +#endif /* _@GUARD_PREFIX@_UNISTD_H */ diff --git a/lib/unlocked-io.h b/lib/unlocked-io.h new file mode 100644 index 0000000..e3f0fcf --- /dev/null +++ b/lib/unlocked-io.h @@ -0,0 +1,136 @@ +/* Prefer faster, non-thread-safe stdio functions if available. + + Copyright (C) 2001-2004, 2009-2018 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +/* Written by Jim Meyering. */ + +#ifndef UNLOCKED_IO_H +# define UNLOCKED_IO_H 1 + +/* These are wrappers for functions/macros from the GNU C library, and + from other C libraries supporting POSIX's optional thread-safe functions. + + The standard I/O functions are thread-safe. These *_unlocked ones are + more efficient but not thread-safe. That they're not thread-safe is + fine since all of the applications in this package are single threaded. + + Also, some code that is shared with the GNU C library may invoke + the *_unlocked functions directly. On hosts that lack those + functions, invoke the non-thread-safe versions instead. */ + +# include <stdio.h> + +# if HAVE_DECL_CLEARERR_UNLOCKED +# undef clearerr +# define clearerr(x) clearerr_unlocked (x) +# else +# define clearerr_unlocked(x) clearerr (x) +# endif + +# if HAVE_DECL_FEOF_UNLOCKED +# undef feof +# define feof(x) feof_unlocked (x) +# else +# define feof_unlocked(x) feof (x) +# endif + +# if HAVE_DECL_FERROR_UNLOCKED +# undef ferror +# define ferror(x) ferror_unlocked (x) +# else +# define ferror_unlocked(x) ferror (x) +# endif + +# if HAVE_DECL_FFLUSH_UNLOCKED +# undef fflush +# define fflush(x) fflush_unlocked (x) +# else +# define fflush_unlocked(x) fflush (x) +# endif + +# if HAVE_DECL_FGETS_UNLOCKED +# undef fgets +# define fgets(x,y,z) fgets_unlocked (x,y,z) +# else +# define fgets_unlocked(x,y,z) fgets (x,y,z) +# endif + +# if HAVE_DECL_FPUTC_UNLOCKED +# undef fputc +# define fputc(x,y) fputc_unlocked (x,y) +# else +# define fputc_unlocked(x,y) fputc (x,y) +# endif + +# if HAVE_DECL_FPUTS_UNLOCKED +# undef fputs +# define fputs(x,y) fputs_unlocked (x,y) +# else +# define fputs_unlocked(x,y) fputs (x,y) +# endif + +# if HAVE_DECL_FREAD_UNLOCKED +# undef fread +# define fread(w,x,y,z) fread_unlocked (w,x,y,z) +# else +# define fread_unlocked(w,x,y,z) fread (w,x,y,z) +# endif + +# if HAVE_DECL_FWRITE_UNLOCKED +# undef fwrite +# define fwrite(w,x,y,z) fwrite_unlocked (w,x,y,z) +# else +# define fwrite_unlocked(w,x,y,z) fwrite (w,x,y,z) +# endif + +# if HAVE_DECL_GETC_UNLOCKED +# undef getc +# define getc(x) getc_unlocked (x) +# else +# define getc_unlocked(x) getc (x) +# endif + +# if HAVE_DECL_GETCHAR_UNLOCKED +# undef getchar +# define getchar() getchar_unlocked () +# else +# define getchar_unlocked() getchar () +# endif + +# if HAVE_DECL_PUTC_UNLOCKED +# undef putc +# define putc(x,y) putc_unlocked (x,y) +# else +# define putc_unlocked(x,y) putc (x,y) +# endif + +# if HAVE_DECL_PUTCHAR_UNLOCKED +# undef putchar +# define putchar(x) putchar_unlocked (x) +# else +# define putchar_unlocked(x) putchar (x) +# endif + +# undef flockfile +# define flockfile(x) ((void) 0) + +# undef ftrylockfile +# define ftrylockfile(x) 0 + +# undef funlockfile +# define funlockfile(x) ((void) 0) + +#endif /* UNLOCKED_IO_H */ diff --git a/lib/unused-parameter.h b/lib/unused-parameter.h new file mode 100644 index 0000000..40a0ff8 --- /dev/null +++ b/lib/unused-parameter.h @@ -0,0 +1,36 @@ +/* A C macro for declaring that specific function parameters are not used. + Copyright (C) 2008-2018 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify it + under the terms of the GNU General Public License as published + by the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +/* _GL_UNUSED_PARAMETER is a marker that can be appended to function parameter + declarations for parameters that are not used. This helps to reduce + warnings, such as from GCC -Wunused-parameter. The syntax is as follows: + type param _GL_UNUSED_PARAMETER + or more generally + param_decl _GL_UNUSED_PARAMETER + For example: + int param _GL_UNUSED_PARAMETER + int *(*param)(void) _GL_UNUSED_PARAMETER + Other possible, but obscure and discouraged syntaxes: + int _GL_UNUSED_PARAMETER *(*param)(void) + _GL_UNUSED_PARAMETER int *(*param)(void) + */ +#ifndef _GL_UNUSED_PARAMETER +# if __GNUC__ >= 3 || (__GNUC__ == 2 && __GNUC_MINOR__ >= 7) +# define _GL_UNUSED_PARAMETER __attribute__ ((__unused__)) +# else +# define _GL_UNUSED_PARAMETER +# endif +#endif diff --git a/lib/verify.h b/lib/verify.h new file mode 100644 index 0000000..3b57dde --- /dev/null +++ b/lib/verify.h @@ -0,0 +1,285 @@ +/* Compile-time assert-like macros. + + Copyright (C) 2005-2006, 2009-2018 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +/* Written by Paul Eggert, Bruno Haible, and Jim Meyering. */ + +#ifndef _GL_VERIFY_H +#define _GL_VERIFY_H + + +/* Define _GL_HAVE__STATIC_ASSERT to 1 if _Static_assert works as per C11. + This is supported by GCC 4.6.0 and later, in C mode, and its use + here generates easier-to-read diagnostics when verify (R) fails. + + Define _GL_HAVE_STATIC_ASSERT to 1 if static_assert works as per C++11. + This will likely be supported by future GCC versions, in C++ mode. + + Use this only with GCC. If we were willing to slow 'configure' + down we could also use it with other compilers, but since this + affects only the quality of diagnostics, why bother? */ +#if (4 < __GNUC__ + (6 <= __GNUC_MINOR__) \ + && (201112L <= __STDC_VERSION__ || !defined __STRICT_ANSI__) \ + && !defined __cplusplus) +# define _GL_HAVE__STATIC_ASSERT 1 +#endif +/* The condition (99 < __GNUC__) is temporary, until we know about the + first G++ release that supports static_assert. */ +#if (99 < __GNUC__) && defined __cplusplus +# define _GL_HAVE_STATIC_ASSERT 1 +#endif + +/* FreeBSD 9.1 <sys/cdefs.h>, included by <stddef.h> and lots of other + system headers, defines a conflicting _Static_assert that is no + better than ours; override it. */ +#ifndef _GL_HAVE_STATIC_ASSERT +# include <stddef.h> +# undef _Static_assert +#endif + +/* Each of these macros verifies that its argument R is nonzero. To + be portable, R should be an integer constant expression. Unlike + assert (R), there is no run-time overhead. + + If _Static_assert works, verify (R) uses it directly. Similarly, + _GL_VERIFY_TRUE works by packaging a _Static_assert inside a struct + that is an operand of sizeof. + + The code below uses several ideas for C++ compilers, and for C + compilers that do not support _Static_assert: + + * The first step is ((R) ? 1 : -1). Given an expression R, of + integral or boolean or floating-point type, this yields an + expression of integral type, whose value is later verified to be + constant and nonnegative. + + * Next this expression W is wrapped in a type + struct _gl_verify_type { + unsigned int _gl_verify_error_if_negative: W; + }. + If W is negative, this yields a compile-time error. No compiler can + deal with a bit-field of negative size. + + One might think that an array size check would have the same + effect, that is, that the type struct { unsigned int dummy[W]; } + would work as well. However, inside a function, some compilers + (such as C++ compilers and GNU C) allow local parameters and + variables inside array size expressions. With these compilers, + an array size check would not properly diagnose this misuse of + the verify macro: + + void function (int n) { verify (n < 0); } + + * For the verify macro, the struct _gl_verify_type will need to + somehow be embedded into a declaration. To be portable, this + declaration must declare an object, a constant, a function, or a + typedef name. If the declared entity uses the type directly, + such as in + + struct dummy {...}; + typedef struct {...} dummy; + extern struct {...} *dummy; + extern void dummy (struct {...} *); + extern struct {...} *dummy (void); + + two uses of the verify macro would yield colliding declarations + if the entity names are not disambiguated. A workaround is to + attach the current line number to the entity name: + + #define _GL_CONCAT0(x, y) x##y + #define _GL_CONCAT(x, y) _GL_CONCAT0 (x, y) + extern struct {...} * _GL_CONCAT (dummy, __LINE__); + + But this has the problem that two invocations of verify from + within the same macro would collide, since the __LINE__ value + would be the same for both invocations. (The GCC __COUNTER__ + macro solves this problem, but is not portable.) + + A solution is to use the sizeof operator. It yields a number, + getting rid of the identity of the type. Declarations like + + extern int dummy [sizeof (struct {...})]; + extern void dummy (int [sizeof (struct {...})]); + extern int (*dummy (void)) [sizeof (struct {...})]; + + can be repeated. + + * Should the implementation use a named struct or an unnamed struct? + Which of the following alternatives can be used? + + extern int dummy [sizeof (struct {...})]; + extern int dummy [sizeof (struct _gl_verify_type {...})]; + extern void dummy (int [sizeof (struct {...})]); + extern void dummy (int [sizeof (struct _gl_verify_type {...})]); + extern int (*dummy (void)) [sizeof (struct {...})]; + extern int (*dummy (void)) [sizeof (struct _gl_verify_type {...})]; + + In the second and sixth case, the struct type is exported to the + outer scope; two such declarations therefore collide. GCC warns + about the first, third, and fourth cases. So the only remaining + possibility is the fifth case: + + extern int (*dummy (void)) [sizeof (struct {...})]; + + * GCC warns about duplicate declarations of the dummy function if + -Wredundant-decls is used. GCC 4.3 and later have a builtin + __COUNTER__ macro that can let us generate unique identifiers for + each dummy function, to suppress this warning. + + * This implementation exploits the fact that older versions of GCC, + which do not support _Static_assert, also do not warn about the + last declaration mentioned above. + + * GCC warns if -Wnested-externs is enabled and verify() is used + within a function body; but inside a function, you can always + arrange to use verify_expr() instead. + + * In C++, any struct definition inside sizeof is invalid. + Use a template type to work around the problem. */ + +/* Concatenate two preprocessor tokens. */ +#define _GL_CONCAT(x, y) _GL_CONCAT0 (x, y) +#define _GL_CONCAT0(x, y) x##y + +/* _GL_COUNTER is an integer, preferably one that changes each time we + use it. Use __COUNTER__ if it works, falling back on __LINE__ + otherwise. __LINE__ isn't perfect, but it's better than a + constant. */ +#if defined __COUNTER__ && __COUNTER__ != __COUNTER__ +# define _GL_COUNTER __COUNTER__ +#else +# define _GL_COUNTER __LINE__ +#endif + +/* Generate a symbol with the given prefix, making it unique if + possible. */ +#define _GL_GENSYM(prefix) _GL_CONCAT (prefix, _GL_COUNTER) + +/* Verify requirement R at compile-time, as an integer constant expression + that returns 1. If R is false, fail at compile-time, preferably + with a diagnostic that includes the string-literal DIAGNOSTIC. */ + +#define _GL_VERIFY_TRUE(R, DIAGNOSTIC) \ + (!!sizeof (_GL_VERIFY_TYPE (R, DIAGNOSTIC))) + +#ifdef __cplusplus +# if !GNULIB_defined_struct__gl_verify_type +template <int w> + struct _gl_verify_type { + unsigned int _gl_verify_error_if_negative: w; + }; +# define GNULIB_defined_struct__gl_verify_type 1 +# endif +# define _GL_VERIFY_TYPE(R, DIAGNOSTIC) \ + _gl_verify_type<(R) ? 1 : -1> +#elif defined _GL_HAVE__STATIC_ASSERT +# define _GL_VERIFY_TYPE(R, DIAGNOSTIC) \ + struct { \ + _Static_assert (R, DIAGNOSTIC); \ + int _gl_dummy; \ + } +#else +# define _GL_VERIFY_TYPE(R, DIAGNOSTIC) \ + struct { unsigned int _gl_verify_error_if_negative: (R) ? 1 : -1; } +#endif + +/* Verify requirement R at compile-time, as a declaration without a + trailing ';'. If R is false, fail at compile-time, preferably + with a diagnostic that includes the string-literal DIAGNOSTIC. + + Unfortunately, unlike C11, this implementation must appear as an + ordinary declaration, and cannot appear inside struct { ... }. */ + +#ifdef _GL_HAVE__STATIC_ASSERT +# define _GL_VERIFY _Static_assert +#else +# define _GL_VERIFY(R, DIAGNOSTIC) \ + extern int (*_GL_GENSYM (_gl_verify_function) (void)) \ + [_GL_VERIFY_TRUE (R, DIAGNOSTIC)] +#endif + +/* _GL_STATIC_ASSERT_H is defined if this code is copied into assert.h. */ +#ifdef _GL_STATIC_ASSERT_H +# if !defined _GL_HAVE__STATIC_ASSERT && !defined _Static_assert +# define _Static_assert(R, DIAGNOSTIC) _GL_VERIFY (R, DIAGNOSTIC) +# endif +# if !defined _GL_HAVE_STATIC_ASSERT && !defined static_assert +# define static_assert _Static_assert /* C11 requires this #define. */ +# endif +#endif + +/* @assert.h omit start@ */ + +/* Each of these macros verifies that its argument R is nonzero. To + be portable, R should be an integer constant expression. Unlike + assert (R), there is no run-time overhead. + + There are two macros, since no single macro can be used in all + contexts in C. verify_true (R) is for scalar contexts, including + integer constant expression contexts. verify (R) is for declaration + contexts, e.g., the top level. */ + +/* Verify requirement R at compile-time, as an integer constant expression. + Return 1. This is equivalent to verify_expr (R, 1). + + verify_true is obsolescent; please use verify_expr instead. */ + +#define verify_true(R) _GL_VERIFY_TRUE (R, "verify_true (" #R ")") + +/* Verify requirement R at compile-time. Return the value of the + expression E. */ + +#define verify_expr(R, E) \ + (_GL_VERIFY_TRUE (R, "verify_expr (" #R ", " #E ")") ? (E) : (E)) + +/* Verify requirement R at compile-time, as a declaration without a + trailing ';'. */ + +#ifdef __GNUC__ +# define verify(R) _GL_VERIFY (R, "verify (" #R ")") +#else +/* PGI barfs if R is long. Play it safe. */ +# define verify(R) _GL_VERIFY (R, "verify (...)") +#endif + +#ifndef __has_builtin +# define __has_builtin(x) 0 +#endif + +/* Assume that R always holds. This lets the compiler optimize + accordingly. R should not have side-effects; it may or may not be + evaluated. Behavior is undefined if R is false. */ + +#if (__has_builtin (__builtin_unreachable) \ + || 4 < __GNUC__ + (5 <= __GNUC_MINOR__)) +# define assume(R) ((R) ? (void) 0 : __builtin_unreachable ()) +#elif 1200 <= _MSC_VER +# define assume(R) __assume (R) +#elif ((defined GCC_LINT || defined lint) \ + && (__has_builtin (__builtin_trap) \ + || 3 < __GNUC__ + (3 < __GNUC_MINOR__ + (4 <= __GNUC_PATCHLEVEL__)))) + /* Doing it this way helps various packages when configured with + --enable-gcc-warnings, which compiles with -Dlint. It's nicer + when 'assume' silences warnings even with older GCCs. */ +# define assume(R) ((R) ? (void) 0 : __builtin_trap ()) +#else + /* Some tools grok NOTREACHED, e.g., Oracle Studio 12.6. */ +# define assume(R) ((R) ? (void) 0 : /*NOTREACHED*/ (void) 0) +#endif + +/* @assert.h omit end@ */ + +#endif diff --git a/lib/version-etc-fsf.c b/lib/version-etc-fsf.c new file mode 100644 index 0000000..098a634 --- /dev/null +++ b/lib/version-etc-fsf.c @@ -0,0 +1,30 @@ +/* Variable with FSF copyright information, for version-etc. + Copyright (C) 1999-2006, 2009-2018 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +/* Written by Jim Meyering. */ + +#include <config.h> + +/* Specification. */ +#include "version-etc.h" + +/* Default copyright goes to the FSF. */ + +const char version_etc_copyright[] = + /* Do *not* mark this string for translation. %s is a copyright + symbol suitable for this locale, and %d is the copyright + year. */ + "Copyright %s %d Free Software Foundation, Inc."; diff --git a/lib/version-etc.c b/lib/version-etc.c new file mode 100644 index 0000000..dfec8ad --- /dev/null +++ b/lib/version-etc.c @@ -0,0 +1,258 @@ +/* Print --version and bug-reporting information in a consistent format. + Copyright (C) 1999-2018 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +/* Written by Jim Meyering. */ + +#include <config.h> + +/* Specification. */ +#include "version-etc.h" + +#include <stdarg.h> +#include <stdio.h> + +#if USE_UNLOCKED_IO +# include "unlocked-io.h" +#endif + +#include "gettext.h" +#define _(msgid) gettext (msgid) + +/* If you use AM_INIT_AUTOMAKE's no-define option, + PACKAGE is not defined. Use PACKAGE_TARNAME instead. */ +#if ! defined PACKAGE && defined PACKAGE_TARNAME +# define PACKAGE PACKAGE_TARNAME +#endif + +enum { COPYRIGHT_YEAR = 2018 }; + +/* The three functions below display the --version information the + standard way. + + If COMMAND_NAME is NULL, the PACKAGE is assumed to be the name of + the program. The formats are therefore: + + PACKAGE VERSION + + or + + COMMAND_NAME (PACKAGE) VERSION. + + The functions differ in the way they are passed author names. */ + +/* Display the --version information the standard way. + + Author names are given in the array AUTHORS. N_AUTHORS is the + number of elements in the array. */ +void +version_etc_arn (FILE *stream, + const char *command_name, const char *package, + const char *version, + const char * const * authors, size_t n_authors) +{ + if (command_name) + fprintf (stream, "%s (%s) %s\n", command_name, package, version); + else + fprintf (stream, "%s %s\n", package, version); + +#ifdef PACKAGE_PACKAGER +# ifdef PACKAGE_PACKAGER_VERSION + fprintf (stream, _("Packaged by %s (%s)\n"), PACKAGE_PACKAGER, + PACKAGE_PACKAGER_VERSION); +# else + fprintf (stream, _("Packaged by %s\n"), PACKAGE_PACKAGER); +# endif +#endif + + /* TRANSLATORS: Translate "(C)" to the copyright symbol + (C-in-a-circle), if this symbol is available in the user's + locale. Otherwise, do not translate "(C)"; leave it as-is. */ + fprintf (stream, version_etc_copyright, _("(C)"), COPYRIGHT_YEAR); + + fputs (_("\ +\n\ +License GPLv3+: GNU GPL version 3 or later <https://gnu.org/licenses/gpl.html>.\n\ +This is free software: you are free to change and redistribute it.\n\ +There is NO WARRANTY, to the extent permitted by law.\n\ +\n\ +"), + stream); + + switch (n_authors) + { + case 0: + /* No authors are given. The caller should output authorship + info after calling this function. */ + break; + case 1: + /* TRANSLATORS: %s denotes an author name. */ + fprintf (stream, _("Written by %s.\n"), authors[0]); + break; + case 2: + /* TRANSLATORS: Each %s denotes an author name. */ + fprintf (stream, _("Written by %s and %s.\n"), authors[0], authors[1]); + break; + case 3: + /* TRANSLATORS: Each %s denotes an author name. */ + fprintf (stream, _("Written by %s, %s, and %s.\n"), + authors[0], authors[1], authors[2]); + break; + case 4: + /* TRANSLATORS: Each %s denotes an author name. + You can use line breaks, estimating that each author name occupies + ca. 16 screen columns and that a screen line has ca. 80 columns. */ + fprintf (stream, _("Written by %s, %s, %s,\nand %s.\n"), + authors[0], authors[1], authors[2], authors[3]); + break; + case 5: + /* TRANSLATORS: Each %s denotes an author name. + You can use line breaks, estimating that each author name occupies + ca. 16 screen columns and that a screen line has ca. 80 columns. */ + fprintf (stream, _("Written by %s, %s, %s,\n%s, and %s.\n"), + authors[0], authors[1], authors[2], authors[3], authors[4]); + break; + case 6: + /* TRANSLATORS: Each %s denotes an author name. + You can use line breaks, estimating that each author name occupies + ca. 16 screen columns and that a screen line has ca. 80 columns. */ + fprintf (stream, _("Written by %s, %s, %s,\n%s, %s, and %s.\n"), + authors[0], authors[1], authors[2], authors[3], authors[4], + authors[5]); + break; + case 7: + /* TRANSLATORS: Each %s denotes an author name. + You can use line breaks, estimating that each author name occupies + ca. 16 screen columns and that a screen line has ca. 80 columns. */ + fprintf (stream, _("Written by %s, %s, %s,\n%s, %s, %s, and %s.\n"), + authors[0], authors[1], authors[2], authors[3], authors[4], + authors[5], authors[6]); + break; + case 8: + /* TRANSLATORS: Each %s denotes an author name. + You can use line breaks, estimating that each author name occupies + ca. 16 screen columns and that a screen line has ca. 80 columns. */ + fprintf (stream, _("\ +Written by %s, %s, %s,\n%s, %s, %s, %s,\nand %s.\n"), + authors[0], authors[1], authors[2], authors[3], authors[4], + authors[5], authors[6], authors[7]); + break; + case 9: + /* TRANSLATORS: Each %s denotes an author name. + You can use line breaks, estimating that each author name occupies + ca. 16 screen columns and that a screen line has ca. 80 columns. */ + fprintf (stream, _("\ +Written by %s, %s, %s,\n%s, %s, %s, %s,\n%s, and %s.\n"), + authors[0], authors[1], authors[2], authors[3], authors[4], + authors[5], authors[6], authors[7], authors[8]); + break; + default: + /* 10 or more authors. Use an abbreviation, since the human reader + will probably not want to read the entire list anyway. */ + /* TRANSLATORS: Each %s denotes an author name. + You can use line breaks, estimating that each author name occupies + ca. 16 screen columns and that a screen line has ca. 80 columns. */ + fprintf (stream, _("\ +Written by %s, %s, %s,\n%s, %s, %s, %s,\n%s, %s, and others.\n"), + authors[0], authors[1], authors[2], authors[3], authors[4], + authors[5], authors[6], authors[7], authors[8]); + break; + } +} + +/* Display the --version information the standard way. See the initial + comment to this module, for more information. + + Author names are given in the NULL-terminated array AUTHORS. */ +void +version_etc_ar (FILE *stream, + const char *command_name, const char *package, + const char *version, const char * const * authors) +{ + size_t n_authors; + + for (n_authors = 0; authors[n_authors]; n_authors++) + ; + version_etc_arn (stream, command_name, package, version, authors, n_authors); +} + +/* Display the --version information the standard way. See the initial + comment to this module, for more information. + + Author names are given in the NULL-terminated va_list AUTHORS. */ +void +version_etc_va (FILE *stream, + const char *command_name, const char *package, + const char *version, va_list authors) +{ + size_t n_authors; + const char *authtab[10]; + + for (n_authors = 0; + n_authors < 10 + && (authtab[n_authors] = va_arg (authors, const char *)) != NULL; + n_authors++) + ; + version_etc_arn (stream, command_name, package, version, + authtab, n_authors); +} + + +/* Display the --version information the standard way. + + If COMMAND_NAME is NULL, the PACKAGE is assumed to be the name of + the program. The formats are therefore: + + PACKAGE VERSION + + or + + COMMAND_NAME (PACKAGE) VERSION. + + The authors names are passed as separate arguments, with an additional + NULL argument at the end. */ +void +version_etc (FILE *stream, + const char *command_name, const char *package, + const char *version, /* const char *author1, ...*/ ...) +{ + va_list authors; + + va_start (authors, version); + version_etc_va (stream, command_name, package, version, authors); + va_end (authors); +} + +void +emit_bug_reporting_address (void) +{ + /* TRANSLATORS: The placeholder indicates the bug-reporting address + for this package. Please add _another line_ saying + "Report translation bugs to <...>\n" with the address for translation + bugs (typically your translation team's web or email address). */ + printf (_("\nReport bugs to: %s\n"), PACKAGE_BUGREPORT); +#ifdef PACKAGE_PACKAGER_BUG_REPORTS + printf (_("Report %s bugs to: %s\n"), PACKAGE_PACKAGER, + PACKAGE_PACKAGER_BUG_REPORTS); +#endif +#ifdef PACKAGE_URL + printf (_("%s home page: <%s>\n"), PACKAGE_NAME, PACKAGE_URL); +#else + printf (_("%s home page: <https://www.gnu.org/software/%s/>\n"), + PACKAGE_NAME, PACKAGE); +#endif + fputs (_("General help using GNU software: <https://www.gnu.org/gethelp/>\n"), + stdout); +} diff --git a/lib/version-etc.h b/lib/version-etc.h new file mode 100644 index 0000000..3b06bf9 --- /dev/null +++ b/lib/version-etc.h @@ -0,0 +1,78 @@ +/* Print --version and bug-reporting information in a consistent format. + Copyright (C) 1999, 2003, 2005, 2009-2018 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +/* Written by Jim Meyering. */ + +#ifndef VERSION_ETC_H +# define VERSION_ETC_H 1 + +# include <stdarg.h> +# include <stdio.h> + +/* The 'sentinel' attribute was added in gcc 4.0. */ +#ifndef _GL_ATTRIBUTE_SENTINEL +# if 4 <= __GNUC__ +# define _GL_ATTRIBUTE_SENTINEL __attribute__ ((__sentinel__)) +# else +# define _GL_ATTRIBUTE_SENTINEL /* empty */ +# endif +#endif + +extern const char version_etc_copyright[]; + +/* The three functions below display the --version information in the + standard way: command and package names, package version, followed + by a short GPLv3+ notice and a list of up to 10 author names. + + If COMMAND_NAME is NULL, the PACKAGE is assumed to be the name of + the program. The formats are therefore: + + PACKAGE VERSION + + or + + COMMAND_NAME (PACKAGE) VERSION. + + The functions differ in the way they are passed author names: */ + +/* N_AUTHORS names are supplied in array AUTHORS. */ +extern void version_etc_arn (FILE *stream, + const char *command_name, const char *package, + const char *version, + const char * const * authors, size_t n_authors); + +/* Names are passed in the NULL-terminated array AUTHORS. */ +extern void version_etc_ar (FILE *stream, + const char *command_name, const char *package, + const char *version, const char * const * authors); + +/* Names are passed in the NULL-terminated va_list. */ +extern void version_etc_va (FILE *stream, + const char *command_name, const char *package, + const char *version, va_list authors); + +/* Names are passed as separate arguments, with an additional + NULL argument at the end. */ +extern void version_etc (FILE *stream, + const char *command_name, const char *package, + const char *version, + /* const char *author1, ..., NULL */ ...) + _GL_ATTRIBUTE_SENTINEL; + +/* Display the usual "Report bugs to" stanza. */ +extern void emit_bug_reporting_address (void); + +#endif /* VERSION_ETC_H */ diff --git a/lib/warn-on-use.h b/lib/warn-on-use.h new file mode 100644 index 0000000..72d67cc --- /dev/null +++ b/lib/warn-on-use.h @@ -0,0 +1,131 @@ +/* A C macro for emitting warnings if a function is used. + Copyright (C) 2010-2018 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify it + under the terms of the GNU General Public License as published + by the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +/* _GL_WARN_ON_USE (function, "literal string") issues a declaration + for FUNCTION which will then trigger a compiler warning containing + the text of "literal string" anywhere that function is called, if + supported by the compiler. If the compiler does not support this + feature, the macro expands to an unused extern declaration. + + _GL_WARN_ON_USE_ATTRIBUTE ("literal string") expands to the + attribute used in _GL_WARN_ON_USE. If the compiler does not support + this feature, it expands to empty. + + These macros are useful for marking a function as a potential + portability trap, with the intent that "literal string" include + instructions on the replacement function that should be used + instead. + _GL_WARN_ON_USE is for functions with 'extern' linkage. + _GL_WARN_ON_USE_ATTRIBUTE is for functions with 'static' or 'inline' + linkage. + + However, one of the reasons that a function is a portability trap is + if it has the wrong signature. Declaring FUNCTION with a different + signature in C is a compilation error, so this macro must use the + same type as any existing declaration so that programs that avoid + the problematic FUNCTION do not fail to compile merely because they + included a header that poisoned the function. But this implies that + _GL_WARN_ON_USE is only safe to use if FUNCTION is known to already + have a declaration. Use of this macro implies that there must not + be any other macro hiding the declaration of FUNCTION; but + undefining FUNCTION first is part of the poisoning process anyway + (although for symbols that are provided only via a macro, the result + is a compilation error rather than a warning containing + "literal string"). Also note that in C++, it is only safe to use if + FUNCTION has no overloads. + + For an example, it is possible to poison 'getline' by: + - adding a call to gl_WARN_ON_USE_PREPARE([[#include <stdio.h>]], + [getline]) in configure.ac, which potentially defines + HAVE_RAW_DECL_GETLINE + - adding this code to a header that wraps the system <stdio.h>: + #undef getline + #if HAVE_RAW_DECL_GETLINE + _GL_WARN_ON_USE (getline, "getline is required by POSIX 2008, but" + "not universally present; use the gnulib module getline"); + #endif + + It is not possible to directly poison global variables. But it is + possible to write a wrapper accessor function, and poison that + (less common usage, like &environ, will cause a compilation error + rather than issue the nice warning, but the end result of informing + the developer about their portability problem is still achieved): + #if HAVE_RAW_DECL_ENVIRON + static char *** + rpl_environ (void) { return &environ; } + _GL_WARN_ON_USE (rpl_environ, "environ is not always properly declared"); + # undef environ + # define environ (*rpl_environ ()) + #endif + or better (avoiding contradictory use of 'static' and 'extern'): + #if HAVE_RAW_DECL_ENVIRON + static char *** + _GL_WARN_ON_USE_ATTRIBUTE ("environ is not always properly declared") + rpl_environ (void) { return &environ; } + # undef environ + # define environ (*rpl_environ ()) + #endif + */ +#ifndef _GL_WARN_ON_USE + +# if 4 < __GNUC__ || (__GNUC__ == 4 && 3 <= __GNUC_MINOR__) +/* A compiler attribute is available in gcc versions 4.3.0 and later. */ +# define _GL_WARN_ON_USE(function, message) \ +extern __typeof__ (function) function __attribute__ ((__warning__ (message))) +# define _GL_WARN_ON_USE_ATTRIBUTE(message) \ + __attribute__ ((__warning__ (message))) +# elif __GNUC__ >= 3 && GNULIB_STRICT_CHECKING +/* Verify the existence of the function. */ +# define _GL_WARN_ON_USE(function, message) \ +extern __typeof__ (function) function +# define _GL_WARN_ON_USE_ATTRIBUTE(message) +# else /* Unsupported. */ +# define _GL_WARN_ON_USE(function, message) \ +_GL_WARN_EXTERN_C int _gl_warn_on_use +# define _GL_WARN_ON_USE_ATTRIBUTE(message) +# endif +#endif + +/* _GL_WARN_ON_USE_CXX (function, rettype, parameters_and_attributes, "string") + is like _GL_WARN_ON_USE (function, "string"), except that the function is + declared with the given prototype, consisting of return type, parameters, + and attributes. + This variant is useful for overloaded functions in C++. _GL_WARN_ON_USE does + not work in this case. */ +#ifndef _GL_WARN_ON_USE_CXX +# if 4 < __GNUC__ || (__GNUC__ == 4 && 3 <= __GNUC_MINOR__) +# define _GL_WARN_ON_USE_CXX(function,rettype,parameters_and_attributes,msg) \ +extern rettype function parameters_and_attributes \ + __attribute__ ((__warning__ (msg))) +# elif __GNUC__ >= 3 && GNULIB_STRICT_CHECKING +/* Verify the existence of the function. */ +# define _GL_WARN_ON_USE_CXX(function,rettype,parameters_and_attributes,msg) \ +extern rettype function parameters_and_attributes +# else /* Unsupported. */ +# define _GL_WARN_ON_USE_CXX(function,rettype,parameters_and_attributes,msg) \ +_GL_WARN_EXTERN_C int _gl_warn_on_use +# endif +#endif + +/* _GL_WARN_EXTERN_C declaration; + performs the declaration with C linkage. */ +#ifndef _GL_WARN_EXTERN_C +# if defined __cplusplus +# define _GL_WARN_EXTERN_C extern "C" +# else +# define _GL_WARN_EXTERN_C extern +# endif +#endif diff --git a/lib/wchar.in.h b/lib/wchar.in.h new file mode 100644 index 0000000..51a5dda --- /dev/null +++ b/lib/wchar.in.h @@ -0,0 +1,1072 @@ +/* A substitute for ISO C99 <wchar.h>, for platforms that have issues. + + Copyright (C) 2007-2018 Free Software Foundation, Inc. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3, or (at your option) + any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, see <https://www.gnu.org/licenses/>. */ + +/* Written by Eric Blake. */ + +/* + * ISO C 99 <wchar.h> for platforms that have issues. + * <http://www.opengroup.org/susv3xbd/wchar.h.html> + * + * For now, this just ensures proper prerequisite inclusion order and + * the declaration of wcwidth(). + */ + +#if __GNUC__ >= 3 +@PRAGMA_SYSTEM_HEADER@ +#endif +@PRAGMA_COLUMNS@ + +#if (((defined __need_mbstate_t || defined __need_wint_t) \ + && !defined __MINGW32__) \ + || (defined __hpux \ + && ((defined _INTTYPES_INCLUDED && !defined strtoimax) \ + || defined _GL_JUST_INCLUDE_SYSTEM_WCHAR_H)) \ + || (defined __MINGW32__ && defined __STRING_H_SOURCED__) \ + || defined _GL_ALREADY_INCLUDING_WCHAR_H) +/* Special invocation convention: + - Inside glibc and uClibc header files, but not MinGW. + - On HP-UX 11.00 we have a sequence of nested includes + <wchar.h> -> <stdlib.h> -> <stdint.h>, and the latter includes <wchar.h>, + once indirectly <stdint.h> -> <sys/types.h> -> <inttypes.h> -> <wchar.h> + and once directly. In both situations 'wint_t' is not yet defined, + therefore we cannot provide the function overrides; instead include only + the system's <wchar.h>. + - With MinGW 3.22, when <string.h> includes <wchar.h>, only some part of + <wchar.h> is actually processed, and that doesn't include 'mbstate_t'. + - On IRIX 6.5, similarly, we have an include <wchar.h> -> <wctype.h>, and + the latter includes <wchar.h>. But here, we have no way to detect whether + <wctype.h> is completely included or is still being included. */ + +#@INCLUDE_NEXT@ @NEXT_WCHAR_H@ + +#else +/* Normal invocation convention. */ + +#ifndef _@GUARD_PREFIX@_WCHAR_H + +#define _GL_ALREADY_INCLUDING_WCHAR_H + +#if @HAVE_FEATURES_H@ +# include <features.h> /* for __GLIBC__ */ +#endif + +/* Tru64 with Desktop Toolkit C has a bug: <stdio.h> must be included before + <wchar.h>. + BSD/OS 4.0.1 has a bug: <stddef.h>, <stdio.h> and <time.h> must be + included before <wchar.h>. + In some builds of uClibc, <wchar.h> is nonexistent and wchar_t is defined + by <stddef.h>. + But avoid namespace pollution on glibc systems. */ +#if !(defined __GLIBC__ && !defined __UCLIBC__) +# include <stddef.h> +#endif +#ifndef __GLIBC__ +# include <stdio.h> +# include <time.h> +#endif + +/* Include the original <wchar.h> if it exists. + Some builds of uClibc lack it. */ +/* The include_next requires a split double-inclusion guard. */ +#if @HAVE_WCHAR_H@ +# @INCLUDE_NEXT@ @NEXT_WCHAR_H@ +#endif + +#undef _GL_ALREADY_INCLUDING_WCHAR_H + +#ifndef _@GUARD_PREFIX@_WCHAR_H +#define _@GUARD_PREFIX@_WCHAR_H + +/* The __attribute__ feature is available in gcc versions 2.5 and later. + The attribute __pure__ was added in gcc 2.96. */ +#if __GNUC__ > 2 || (__GNUC__ == 2 && __GNUC_MINOR__ >= 96) +# define _GL_ATTRIBUTE_PURE __attribute__ ((__pure__)) +#else +# define _GL_ATTRIBUTE_PURE /* empty */ +#endif + +/* The definitions of _GL_FUNCDECL_RPL etc. are copied here. */ + +/* The definition of _GL_ARG_NONNULL is copied here. */ + +/* The definition of _GL_WARN_ON_USE is copied here. */ + + +/* Define wint_t and WEOF. (Also done in wctype.in.h.) */ +#if !@HAVE_WINT_T@ && !defined wint_t +# define wint_t int +# ifndef WEOF +# define WEOF -1 +# endif +#else +/* mingw and MSVC define wint_t as 'unsigned short' in <crtdefs.h> or + <stddef.h>. This is too small: ISO C 99 section 7.24.1.(2) says that + wint_t must be "unchanged by default argument promotions". Override it. */ +# if @GNULIB_OVERRIDES_WINT_T@ +# if !GNULIB_defined_wint_t +# if @HAVE_CRTDEFS_H@ +# include <crtdefs.h> +# else +# include <stddef.h> +# endif +typedef unsigned int rpl_wint_t; +# undef wint_t +# define wint_t rpl_wint_t +# define GNULIB_defined_wint_t 1 +# endif +# endif +# ifndef WEOF +# define WEOF ((wint_t) -1) +# endif +#endif + + +/* Override mbstate_t if it is too small. + On IRIX 6.5, sizeof (mbstate_t) == 1, which is not sufficient for + implementing mbrtowc for encodings like UTF-8. */ +#if !(@HAVE_MBSINIT@ && @HAVE_MBRTOWC@) || @REPLACE_MBSTATE_T@ +# if !GNULIB_defined_mbstate_t +typedef int rpl_mbstate_t; +# undef mbstate_t +# define mbstate_t rpl_mbstate_t +# define GNULIB_defined_mbstate_t 1 +# endif +#endif + + +/* Convert a single-byte character to a wide character. */ +#if @GNULIB_BTOWC@ +# if @REPLACE_BTOWC@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef btowc +# define btowc rpl_btowc +# endif +_GL_FUNCDECL_RPL (btowc, wint_t, (int c) _GL_ATTRIBUTE_PURE); +_GL_CXXALIAS_RPL (btowc, wint_t, (int c)); +# else +# if !@HAVE_BTOWC@ +_GL_FUNCDECL_SYS (btowc, wint_t, (int c) _GL_ATTRIBUTE_PURE); +# endif +_GL_CXXALIAS_SYS (btowc, wint_t, (int c)); +# endif +_GL_CXXALIASWARN (btowc); +#elif defined GNULIB_POSIXCHECK +# undef btowc +# if HAVE_RAW_DECL_BTOWC +_GL_WARN_ON_USE (btowc, "btowc is unportable - " + "use gnulib module btowc for portability"); +# endif +#endif + + +/* Convert a wide character to a single-byte character. */ +#if @GNULIB_WCTOB@ +# if @REPLACE_WCTOB@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef wctob +# define wctob rpl_wctob +# endif +_GL_FUNCDECL_RPL (wctob, int, (wint_t wc) _GL_ATTRIBUTE_PURE); +_GL_CXXALIAS_RPL (wctob, int, (wint_t wc)); +# else +# if !defined wctob && !@HAVE_DECL_WCTOB@ +/* wctob is provided by gnulib, or wctob exists but is not declared. */ +_GL_FUNCDECL_SYS (wctob, int, (wint_t wc) _GL_ATTRIBUTE_PURE); +# endif +_GL_CXXALIAS_SYS (wctob, int, (wint_t wc)); +# endif +_GL_CXXALIASWARN (wctob); +#elif defined GNULIB_POSIXCHECK +# undef wctob +# if HAVE_RAW_DECL_WCTOB +_GL_WARN_ON_USE (wctob, "wctob is unportable - " + "use gnulib module wctob for portability"); +# endif +#endif + + +/* Test whether *PS is in the initial state. */ +#if @GNULIB_MBSINIT@ +# if @REPLACE_MBSINIT@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef mbsinit +# define mbsinit rpl_mbsinit +# endif +_GL_FUNCDECL_RPL (mbsinit, int, (const mbstate_t *ps)); +_GL_CXXALIAS_RPL (mbsinit, int, (const mbstate_t *ps)); +# else +# if !@HAVE_MBSINIT@ +_GL_FUNCDECL_SYS (mbsinit, int, (const mbstate_t *ps)); +# endif +_GL_CXXALIAS_SYS (mbsinit, int, (const mbstate_t *ps)); +# endif +_GL_CXXALIASWARN (mbsinit); +#elif defined GNULIB_POSIXCHECK +# undef mbsinit +# if HAVE_RAW_DECL_MBSINIT +_GL_WARN_ON_USE (mbsinit, "mbsinit is unportable - " + "use gnulib module mbsinit for portability"); +# endif +#endif + + +/* Convert a multibyte character to a wide character. */ +#if @GNULIB_MBRTOWC@ +# if @REPLACE_MBRTOWC@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef mbrtowc +# define mbrtowc rpl_mbrtowc +# endif +_GL_FUNCDECL_RPL (mbrtowc, size_t, + (wchar_t *pwc, const char *s, size_t n, mbstate_t *ps)); +_GL_CXXALIAS_RPL (mbrtowc, size_t, + (wchar_t *pwc, const char *s, size_t n, mbstate_t *ps)); +# else +# if !@HAVE_MBRTOWC@ +_GL_FUNCDECL_SYS (mbrtowc, size_t, + (wchar_t *pwc, const char *s, size_t n, mbstate_t *ps)); +# endif +_GL_CXXALIAS_SYS (mbrtowc, size_t, + (wchar_t *pwc, const char *s, size_t n, mbstate_t *ps)); +# endif +_GL_CXXALIASWARN (mbrtowc); +#elif defined GNULIB_POSIXCHECK +# undef mbrtowc +# if HAVE_RAW_DECL_MBRTOWC +_GL_WARN_ON_USE (mbrtowc, "mbrtowc is unportable - " + "use gnulib module mbrtowc for portability"); +# endif +#endif + + +/* Recognize a multibyte character. */ +#if @GNULIB_MBRLEN@ +# if @REPLACE_MBRLEN@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef mbrlen +# define mbrlen rpl_mbrlen +# endif +_GL_FUNCDECL_RPL (mbrlen, size_t, (const char *s, size_t n, mbstate_t *ps)); +_GL_CXXALIAS_RPL (mbrlen, size_t, (const char *s, size_t n, mbstate_t *ps)); +# else +# if !@HAVE_MBRLEN@ +_GL_FUNCDECL_SYS (mbrlen, size_t, (const char *s, size_t n, mbstate_t *ps)); +# endif +_GL_CXXALIAS_SYS (mbrlen, size_t, (const char *s, size_t n, mbstate_t *ps)); +# endif +_GL_CXXALIASWARN (mbrlen); +#elif defined GNULIB_POSIXCHECK +# undef mbrlen +# if HAVE_RAW_DECL_MBRLEN +_GL_WARN_ON_USE (mbrlen, "mbrlen is unportable - " + "use gnulib module mbrlen for portability"); +# endif +#endif + + +/* Convert a string to a wide string. */ +#if @GNULIB_MBSRTOWCS@ +# if @REPLACE_MBSRTOWCS@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef mbsrtowcs +# define mbsrtowcs rpl_mbsrtowcs +# endif +_GL_FUNCDECL_RPL (mbsrtowcs, size_t, + (wchar_t *dest, const char **srcp, size_t len, mbstate_t *ps) + _GL_ARG_NONNULL ((2))); +_GL_CXXALIAS_RPL (mbsrtowcs, size_t, + (wchar_t *dest, const char **srcp, size_t len, + mbstate_t *ps)); +# else +# if !@HAVE_MBSRTOWCS@ +_GL_FUNCDECL_SYS (mbsrtowcs, size_t, + (wchar_t *dest, const char **srcp, size_t len, mbstate_t *ps) + _GL_ARG_NONNULL ((2))); +# endif +_GL_CXXALIAS_SYS (mbsrtowcs, size_t, + (wchar_t *dest, const char **srcp, size_t len, + mbstate_t *ps)); +# endif +_GL_CXXALIASWARN (mbsrtowcs); +#elif defined GNULIB_POSIXCHECK +# undef mbsrtowcs +# if HAVE_RAW_DECL_MBSRTOWCS +_GL_WARN_ON_USE (mbsrtowcs, "mbsrtowcs is unportable - " + "use gnulib module mbsrtowcs for portability"); +# endif +#endif + + +/* Convert a string to a wide string. */ +#if @GNULIB_MBSNRTOWCS@ +# if @REPLACE_MBSNRTOWCS@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef mbsnrtowcs +# define mbsnrtowcs rpl_mbsnrtowcs +# endif +_GL_FUNCDECL_RPL (mbsnrtowcs, size_t, + (wchar_t *dest, const char **srcp, size_t srclen, size_t len, + mbstate_t *ps) + _GL_ARG_NONNULL ((2))); +_GL_CXXALIAS_RPL (mbsnrtowcs, size_t, + (wchar_t *dest, const char **srcp, size_t srclen, size_t len, + mbstate_t *ps)); +# else +# if !@HAVE_MBSNRTOWCS@ +_GL_FUNCDECL_SYS (mbsnrtowcs, size_t, + (wchar_t *dest, const char **srcp, size_t srclen, size_t len, + mbstate_t *ps) + _GL_ARG_NONNULL ((2))); +# endif +_GL_CXXALIAS_SYS (mbsnrtowcs, size_t, + (wchar_t *dest, const char **srcp, size_t srclen, size_t len, + mbstate_t *ps)); +# endif +_GL_CXXALIASWARN (mbsnrtowcs); +#elif defined GNULIB_POSIXCHECK +# undef mbsnrtowcs +# if HAVE_RAW_DECL_MBSNRTOWCS +_GL_WARN_ON_USE (mbsnrtowcs, "mbsnrtowcs is unportable - " + "use gnulib module mbsnrtowcs for portability"); +# endif +#endif + + +/* Convert a wide character to a multibyte character. */ +#if @GNULIB_WCRTOMB@ +# if @REPLACE_WCRTOMB@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef wcrtomb +# define wcrtomb rpl_wcrtomb +# endif +_GL_FUNCDECL_RPL (wcrtomb, size_t, (char *s, wchar_t wc, mbstate_t *ps)); +_GL_CXXALIAS_RPL (wcrtomb, size_t, (char *s, wchar_t wc, mbstate_t *ps)); +# else +# if !@HAVE_WCRTOMB@ +_GL_FUNCDECL_SYS (wcrtomb, size_t, (char *s, wchar_t wc, mbstate_t *ps)); +# endif +_GL_CXXALIAS_SYS (wcrtomb, size_t, (char *s, wchar_t wc, mbstate_t *ps)); +# endif +_GL_CXXALIASWARN (wcrtomb); +#elif defined GNULIB_POSIXCHECK +# undef wcrtomb +# if HAVE_RAW_DECL_WCRTOMB +_GL_WARN_ON_USE (wcrtomb, "wcrtomb is unportable - " + "use gnulib module wcrtomb for portability"); +# endif +#endif + + +/* Convert a wide string to a string. */ +#if @GNULIB_WCSRTOMBS@ +# if @REPLACE_WCSRTOMBS@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef wcsrtombs +# define wcsrtombs rpl_wcsrtombs +# endif +_GL_FUNCDECL_RPL (wcsrtombs, size_t, + (char *dest, const wchar_t **srcp, size_t len, mbstate_t *ps) + _GL_ARG_NONNULL ((2))); +_GL_CXXALIAS_RPL (wcsrtombs, size_t, + (char *dest, const wchar_t **srcp, size_t len, + mbstate_t *ps)); +# else +# if !@HAVE_WCSRTOMBS@ +_GL_FUNCDECL_SYS (wcsrtombs, size_t, + (char *dest, const wchar_t **srcp, size_t len, mbstate_t *ps) + _GL_ARG_NONNULL ((2))); +# endif +_GL_CXXALIAS_SYS (wcsrtombs, size_t, + (char *dest, const wchar_t **srcp, size_t len, + mbstate_t *ps)); +# endif +_GL_CXXALIASWARN (wcsrtombs); +#elif defined GNULIB_POSIXCHECK +# undef wcsrtombs +# if HAVE_RAW_DECL_WCSRTOMBS +_GL_WARN_ON_USE (wcsrtombs, "wcsrtombs is unportable - " + "use gnulib module wcsrtombs for portability"); +# endif +#endif + + +/* Convert a wide string to a string. */ +#if @GNULIB_WCSNRTOMBS@ +# if @REPLACE_WCSNRTOMBS@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef wcsnrtombs +# define wcsnrtombs rpl_wcsnrtombs +# endif +_GL_FUNCDECL_RPL (wcsnrtombs, size_t, + (char *dest, const wchar_t **srcp, size_t srclen, size_t len, + mbstate_t *ps) + _GL_ARG_NONNULL ((2))); +_GL_CXXALIAS_RPL (wcsnrtombs, size_t, + (char *dest, const wchar_t **srcp, size_t srclen, size_t len, + mbstate_t *ps)); +# else +# if !@HAVE_WCSNRTOMBS@ +_GL_FUNCDECL_SYS (wcsnrtombs, size_t, + (char *dest, const wchar_t **srcp, size_t srclen, size_t len, + mbstate_t *ps) + _GL_ARG_NONNULL ((2))); +# endif +_GL_CXXALIAS_SYS (wcsnrtombs, size_t, + (char *dest, const wchar_t **srcp, size_t srclen, size_t len, + mbstate_t *ps)); +# endif +_GL_CXXALIASWARN (wcsnrtombs); +#elif defined GNULIB_POSIXCHECK +# undef wcsnrtombs +# if HAVE_RAW_DECL_WCSNRTOMBS +_GL_WARN_ON_USE (wcsnrtombs, "wcsnrtombs is unportable - " + "use gnulib module wcsnrtombs for portability"); +# endif +#endif + + +/* Return the number of screen columns needed for WC. */ +#if @GNULIB_WCWIDTH@ +# if @REPLACE_WCWIDTH@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef wcwidth +# define wcwidth rpl_wcwidth +# endif +_GL_FUNCDECL_RPL (wcwidth, int, (wchar_t) _GL_ATTRIBUTE_PURE); +_GL_CXXALIAS_RPL (wcwidth, int, (wchar_t)); +# else +# if !@HAVE_DECL_WCWIDTH@ +/* wcwidth exists but is not declared. */ +_GL_FUNCDECL_SYS (wcwidth, int, (wchar_t) _GL_ATTRIBUTE_PURE); +# endif +_GL_CXXALIAS_SYS (wcwidth, int, (wchar_t)); +# endif +_GL_CXXALIASWARN (wcwidth); +#elif defined GNULIB_POSIXCHECK +# undef wcwidth +# if HAVE_RAW_DECL_WCWIDTH +_GL_WARN_ON_USE (wcwidth, "wcwidth is unportable - " + "use gnulib module wcwidth for portability"); +# endif +#endif + + +/* Search N wide characters of S for C. */ +#if @GNULIB_WMEMCHR@ +# if !@HAVE_WMEMCHR@ +_GL_FUNCDECL_SYS (wmemchr, wchar_t *, (const wchar_t *s, wchar_t c, size_t n) + _GL_ATTRIBUTE_PURE); +# endif + /* On some systems, this function is defined as an overloaded function: + extern "C++" { + const wchar_t * std::wmemchr (const wchar_t *, wchar_t, size_t); + wchar_t * std::wmemchr (wchar_t *, wchar_t, size_t); + } */ +_GL_CXXALIAS_SYS_CAST2 (wmemchr, + wchar_t *, (const wchar_t *, wchar_t, size_t), + const wchar_t *, (const wchar_t *, wchar_t, size_t)); +# if ((__GLIBC__ == 2 && __GLIBC_MINOR__ >= 10) && !defined __UCLIBC__) \ + && (__GNUC__ > 4 || (__GNUC__ == 4 && __GNUC_MINOR__ >= 4)) +_GL_CXXALIASWARN1 (wmemchr, wchar_t *, (wchar_t *s, wchar_t c, size_t n)); +_GL_CXXALIASWARN1 (wmemchr, const wchar_t *, + (const wchar_t *s, wchar_t c, size_t n)); +# else +_GL_CXXALIASWARN (wmemchr); +# endif +#elif defined GNULIB_POSIXCHECK +# undef wmemchr +# if HAVE_RAW_DECL_WMEMCHR +_GL_WARN_ON_USE (wmemchr, "wmemchr is unportable - " + "use gnulib module wmemchr for portability"); +# endif +#endif + + +/* Compare N wide characters of S1 and S2. */ +#if @GNULIB_WMEMCMP@ +# if !@HAVE_WMEMCMP@ +_GL_FUNCDECL_SYS (wmemcmp, int, + (const wchar_t *s1, const wchar_t *s2, size_t n) + _GL_ATTRIBUTE_PURE); +# endif +_GL_CXXALIAS_SYS (wmemcmp, int, + (const wchar_t *s1, const wchar_t *s2, size_t n)); +_GL_CXXALIASWARN (wmemcmp); +#elif defined GNULIB_POSIXCHECK +# undef wmemcmp +# if HAVE_RAW_DECL_WMEMCMP +_GL_WARN_ON_USE (wmemcmp, "wmemcmp is unportable - " + "use gnulib module wmemcmp for portability"); +# endif +#endif + + +/* Copy N wide characters of SRC to DEST. */ +#if @GNULIB_WMEMCPY@ +# if !@HAVE_WMEMCPY@ +_GL_FUNCDECL_SYS (wmemcpy, wchar_t *, + (wchar_t *dest, const wchar_t *src, size_t n)); +# endif +_GL_CXXALIAS_SYS (wmemcpy, wchar_t *, + (wchar_t *dest, const wchar_t *src, size_t n)); +_GL_CXXALIASWARN (wmemcpy); +#elif defined GNULIB_POSIXCHECK +# undef wmemcpy +# if HAVE_RAW_DECL_WMEMCPY +_GL_WARN_ON_USE (wmemcpy, "wmemcpy is unportable - " + "use gnulib module wmemcpy for portability"); +# endif +#endif + + +/* Copy N wide characters of SRC to DEST, guaranteeing correct behavior for + overlapping memory areas. */ +#if @GNULIB_WMEMMOVE@ +# if !@HAVE_WMEMMOVE@ +_GL_FUNCDECL_SYS (wmemmove, wchar_t *, + (wchar_t *dest, const wchar_t *src, size_t n)); +# endif +_GL_CXXALIAS_SYS (wmemmove, wchar_t *, + (wchar_t *dest, const wchar_t *src, size_t n)); +_GL_CXXALIASWARN (wmemmove); +#elif defined GNULIB_POSIXCHECK +# undef wmemmove +# if HAVE_RAW_DECL_WMEMMOVE +_GL_WARN_ON_USE (wmemmove, "wmemmove is unportable - " + "use gnulib module wmemmove for portability"); +# endif +#endif + + +/* Set N wide characters of S to C. */ +#if @GNULIB_WMEMSET@ +# if !@HAVE_WMEMSET@ +_GL_FUNCDECL_SYS (wmemset, wchar_t *, (wchar_t *s, wchar_t c, size_t n)); +# endif +_GL_CXXALIAS_SYS (wmemset, wchar_t *, (wchar_t *s, wchar_t c, size_t n)); +_GL_CXXALIASWARN (wmemset); +#elif defined GNULIB_POSIXCHECK +# undef wmemset +# if HAVE_RAW_DECL_WMEMSET +_GL_WARN_ON_USE (wmemset, "wmemset is unportable - " + "use gnulib module wmemset for portability"); +# endif +#endif + + +/* Return the number of wide characters in S. */ +#if @GNULIB_WCSLEN@ +# if !@HAVE_WCSLEN@ +_GL_FUNCDECL_SYS (wcslen, size_t, (const wchar_t *s) _GL_ATTRIBUTE_PURE); +# endif +_GL_CXXALIAS_SYS (wcslen, size_t, (const wchar_t *s)); +_GL_CXXALIASWARN (wcslen); +#elif defined GNULIB_POSIXCHECK +# undef wcslen +# if HAVE_RAW_DECL_WCSLEN +_GL_WARN_ON_USE (wcslen, "wcslen is unportable - " + "use gnulib module wcslen for portability"); +# endif +#endif + + +/* Return the number of wide characters in S, but at most MAXLEN. */ +#if @GNULIB_WCSNLEN@ +# if !@HAVE_WCSNLEN@ +_GL_FUNCDECL_SYS (wcsnlen, size_t, (const wchar_t *s, size_t maxlen) + _GL_ATTRIBUTE_PURE); +# endif +_GL_CXXALIAS_SYS (wcsnlen, size_t, (const wchar_t *s, size_t maxlen)); +_GL_CXXALIASWARN (wcsnlen); +#elif defined GNULIB_POSIXCHECK +# undef wcsnlen +# if HAVE_RAW_DECL_WCSNLEN +_GL_WARN_ON_USE (wcsnlen, "wcsnlen is unportable - " + "use gnulib module wcsnlen for portability"); +# endif +#endif + + +/* Copy SRC to DEST. */ +#if @GNULIB_WCSCPY@ +# if !@HAVE_WCSCPY@ +_GL_FUNCDECL_SYS (wcscpy, wchar_t *, (wchar_t *dest, const wchar_t *src)); +# endif +_GL_CXXALIAS_SYS (wcscpy, wchar_t *, (wchar_t *dest, const wchar_t *src)); +_GL_CXXALIASWARN (wcscpy); +#elif defined GNULIB_POSIXCHECK +# undef wcscpy +# if HAVE_RAW_DECL_WCSCPY +_GL_WARN_ON_USE (wcscpy, "wcscpy is unportable - " + "use gnulib module wcscpy for portability"); +# endif +#endif + + +/* Copy SRC to DEST, returning the address of the terminating L'\0' in DEST. */ +#if @GNULIB_WCPCPY@ +# if !@HAVE_WCPCPY@ +_GL_FUNCDECL_SYS (wcpcpy, wchar_t *, (wchar_t *dest, const wchar_t *src)); +# endif +_GL_CXXALIAS_SYS (wcpcpy, wchar_t *, (wchar_t *dest, const wchar_t *src)); +_GL_CXXALIASWARN (wcpcpy); +#elif defined GNULIB_POSIXCHECK +# undef wcpcpy +# if HAVE_RAW_DECL_WCPCPY +_GL_WARN_ON_USE (wcpcpy, "wcpcpy is unportable - " + "use gnulib module wcpcpy for portability"); +# endif +#endif + + +/* Copy no more than N wide characters of SRC to DEST. */ +#if @GNULIB_WCSNCPY@ +# if !@HAVE_WCSNCPY@ +_GL_FUNCDECL_SYS (wcsncpy, wchar_t *, + (wchar_t *dest, const wchar_t *src, size_t n)); +# endif +_GL_CXXALIAS_SYS (wcsncpy, wchar_t *, + (wchar_t *dest, const wchar_t *src, size_t n)); +_GL_CXXALIASWARN (wcsncpy); +#elif defined GNULIB_POSIXCHECK +# undef wcsncpy +# if HAVE_RAW_DECL_WCSNCPY +_GL_WARN_ON_USE (wcsncpy, "wcsncpy is unportable - " + "use gnulib module wcsncpy for portability"); +# endif +#endif + + +/* Copy no more than N characters of SRC to DEST, returning the address of + the last character written into DEST. */ +#if @GNULIB_WCPNCPY@ +# if !@HAVE_WCPNCPY@ +_GL_FUNCDECL_SYS (wcpncpy, wchar_t *, + (wchar_t *dest, const wchar_t *src, size_t n)); +# endif +_GL_CXXALIAS_SYS (wcpncpy, wchar_t *, + (wchar_t *dest, const wchar_t *src, size_t n)); +_GL_CXXALIASWARN (wcpncpy); +#elif defined GNULIB_POSIXCHECK +# undef wcpncpy +# if HAVE_RAW_DECL_WCPNCPY +_GL_WARN_ON_USE (wcpncpy, "wcpncpy is unportable - " + "use gnulib module wcpncpy for portability"); +# endif +#endif + + +/* Append SRC onto DEST. */ +#if @GNULIB_WCSCAT@ +# if !@HAVE_WCSCAT@ +_GL_FUNCDECL_SYS (wcscat, wchar_t *, (wchar_t *dest, const wchar_t *src)); +# endif +_GL_CXXALIAS_SYS (wcscat, wchar_t *, (wchar_t *dest, const wchar_t *src)); +_GL_CXXALIASWARN (wcscat); +#elif defined GNULIB_POSIXCHECK +# undef wcscat +# if HAVE_RAW_DECL_WCSCAT +_GL_WARN_ON_USE (wcscat, "wcscat is unportable - " + "use gnulib module wcscat for portability"); +# endif +#endif + + +/* Append no more than N wide characters of SRC onto DEST. */ +#if @GNULIB_WCSNCAT@ +# if !@HAVE_WCSNCAT@ +_GL_FUNCDECL_SYS (wcsncat, wchar_t *, + (wchar_t *dest, const wchar_t *src, size_t n)); +# endif +_GL_CXXALIAS_SYS (wcsncat, wchar_t *, + (wchar_t *dest, const wchar_t *src, size_t n)); +_GL_CXXALIASWARN (wcsncat); +#elif defined GNULIB_POSIXCHECK +# undef wcsncat +# if HAVE_RAW_DECL_WCSNCAT +_GL_WARN_ON_USE (wcsncat, "wcsncat is unportable - " + "use gnulib module wcsncat for portability"); +# endif +#endif + + +/* Compare S1 and S2. */ +#if @GNULIB_WCSCMP@ +# if !@HAVE_WCSCMP@ +_GL_FUNCDECL_SYS (wcscmp, int, (const wchar_t *s1, const wchar_t *s2) + _GL_ATTRIBUTE_PURE); +# endif +_GL_CXXALIAS_SYS (wcscmp, int, (const wchar_t *s1, const wchar_t *s2)); +_GL_CXXALIASWARN (wcscmp); +#elif defined GNULIB_POSIXCHECK +# undef wcscmp +# if HAVE_RAW_DECL_WCSCMP +_GL_WARN_ON_USE (wcscmp, "wcscmp is unportable - " + "use gnulib module wcscmp for portability"); +# endif +#endif + + +/* Compare no more than N wide characters of S1 and S2. */ +#if @GNULIB_WCSNCMP@ +# if !@HAVE_WCSNCMP@ +_GL_FUNCDECL_SYS (wcsncmp, int, + (const wchar_t *s1, const wchar_t *s2, size_t n) + _GL_ATTRIBUTE_PURE); +# endif +_GL_CXXALIAS_SYS (wcsncmp, int, + (const wchar_t *s1, const wchar_t *s2, size_t n)); +_GL_CXXALIASWARN (wcsncmp); +#elif defined GNULIB_POSIXCHECK +# undef wcsncmp +# if HAVE_RAW_DECL_WCSNCMP +_GL_WARN_ON_USE (wcsncmp, "wcsncmp is unportable - " + "use gnulib module wcsncmp for portability"); +# endif +#endif + + +/* Compare S1 and S2, ignoring case. */ +#if @GNULIB_WCSCASECMP@ +# if !@HAVE_WCSCASECMP@ +_GL_FUNCDECL_SYS (wcscasecmp, int, (const wchar_t *s1, const wchar_t *s2) + _GL_ATTRIBUTE_PURE); +# endif +_GL_CXXALIAS_SYS (wcscasecmp, int, (const wchar_t *s1, const wchar_t *s2)); +_GL_CXXALIASWARN (wcscasecmp); +#elif defined GNULIB_POSIXCHECK +# undef wcscasecmp +# if HAVE_RAW_DECL_WCSCASECMP +_GL_WARN_ON_USE (wcscasecmp, "wcscasecmp is unportable - " + "use gnulib module wcscasecmp for portability"); +# endif +#endif + + +/* Compare no more than N chars of S1 and S2, ignoring case. */ +#if @GNULIB_WCSNCASECMP@ +# if !@HAVE_WCSNCASECMP@ +_GL_FUNCDECL_SYS (wcsncasecmp, int, + (const wchar_t *s1, const wchar_t *s2, size_t n) + _GL_ATTRIBUTE_PURE); +# endif +_GL_CXXALIAS_SYS (wcsncasecmp, int, + (const wchar_t *s1, const wchar_t *s2, size_t n)); +_GL_CXXALIASWARN (wcsncasecmp); +#elif defined GNULIB_POSIXCHECK +# undef wcsncasecmp +# if HAVE_RAW_DECL_WCSNCASECMP +_GL_WARN_ON_USE (wcsncasecmp, "wcsncasecmp is unportable - " + "use gnulib module wcsncasecmp for portability"); +# endif +#endif + + +/* Compare S1 and S2, both interpreted as appropriate to the LC_COLLATE + category of the current locale. */ +#if @GNULIB_WCSCOLL@ +# if !@HAVE_WCSCOLL@ +_GL_FUNCDECL_SYS (wcscoll, int, (const wchar_t *s1, const wchar_t *s2)); +# endif +_GL_CXXALIAS_SYS (wcscoll, int, (const wchar_t *s1, const wchar_t *s2)); +_GL_CXXALIASWARN (wcscoll); +#elif defined GNULIB_POSIXCHECK +# undef wcscoll +# if HAVE_RAW_DECL_WCSCOLL +_GL_WARN_ON_USE (wcscoll, "wcscoll is unportable - " + "use gnulib module wcscoll for portability"); +# endif +#endif + + +/* Transform S2 into array pointed to by S1 such that if wcscmp is applied + to two transformed strings the result is the as applying 'wcscoll' to the + original strings. */ +#if @GNULIB_WCSXFRM@ +# if !@HAVE_WCSXFRM@ +_GL_FUNCDECL_SYS (wcsxfrm, size_t, (wchar_t *s1, const wchar_t *s2, size_t n)); +# endif +_GL_CXXALIAS_SYS (wcsxfrm, size_t, (wchar_t *s1, const wchar_t *s2, size_t n)); +_GL_CXXALIASWARN (wcsxfrm); +#elif defined GNULIB_POSIXCHECK +# undef wcsxfrm +# if HAVE_RAW_DECL_WCSXFRM +_GL_WARN_ON_USE (wcsxfrm, "wcsxfrm is unportable - " + "use gnulib module wcsxfrm for portability"); +# endif +#endif + + +/* Duplicate S, returning an identical malloc'd string. */ +#if @GNULIB_WCSDUP@ +# if !@HAVE_WCSDUP@ +_GL_FUNCDECL_SYS (wcsdup, wchar_t *, (const wchar_t *s)); +# endif +_GL_CXXALIAS_SYS (wcsdup, wchar_t *, (const wchar_t *s)); +_GL_CXXALIASWARN (wcsdup); +#elif defined GNULIB_POSIXCHECK +# undef wcsdup +# if HAVE_RAW_DECL_WCSDUP +_GL_WARN_ON_USE (wcsdup, "wcsdup is unportable - " + "use gnulib module wcsdup for portability"); +# endif +#endif + + +/* Find the first occurrence of WC in WCS. */ +#if @GNULIB_WCSCHR@ +# if !@HAVE_WCSCHR@ +_GL_FUNCDECL_SYS (wcschr, wchar_t *, (const wchar_t *wcs, wchar_t wc) + _GL_ATTRIBUTE_PURE); +# endif + /* On some systems, this function is defined as an overloaded function: + extern "C++" { + const wchar_t * std::wcschr (const wchar_t *, wchar_t); + wchar_t * std::wcschr (wchar_t *, wchar_t); + } */ +_GL_CXXALIAS_SYS_CAST2 (wcschr, + wchar_t *, (const wchar_t *, wchar_t), + const wchar_t *, (const wchar_t *, wchar_t)); +# if ((__GLIBC__ == 2 && __GLIBC_MINOR__ >= 10) && !defined __UCLIBC__) \ + && (__GNUC__ > 4 || (__GNUC__ == 4 && __GNUC_MINOR__ >= 4)) +_GL_CXXALIASWARN1 (wcschr, wchar_t *, (wchar_t *wcs, wchar_t wc)); +_GL_CXXALIASWARN1 (wcschr, const wchar_t *, (const wchar_t *wcs, wchar_t wc)); +# else +_GL_CXXALIASWARN (wcschr); +# endif +#elif defined GNULIB_POSIXCHECK +# undef wcschr +# if HAVE_RAW_DECL_WCSCHR +_GL_WARN_ON_USE (wcschr, "wcschr is unportable - " + "use gnulib module wcschr for portability"); +# endif +#endif + + +/* Find the last occurrence of WC in WCS. */ +#if @GNULIB_WCSRCHR@ +# if !@HAVE_WCSRCHR@ +_GL_FUNCDECL_SYS (wcsrchr, wchar_t *, (const wchar_t *wcs, wchar_t wc) + _GL_ATTRIBUTE_PURE); +# endif + /* On some systems, this function is defined as an overloaded function: + extern "C++" { + const wchar_t * std::wcsrchr (const wchar_t *, wchar_t); + wchar_t * std::wcsrchr (wchar_t *, wchar_t); + } */ +_GL_CXXALIAS_SYS_CAST2 (wcsrchr, + wchar_t *, (const wchar_t *, wchar_t), + const wchar_t *, (const wchar_t *, wchar_t)); +# if ((__GLIBC__ == 2 && __GLIBC_MINOR__ >= 10) && !defined __UCLIBC__) \ + && (__GNUC__ > 4 || (__GNUC__ == 4 && __GNUC_MINOR__ >= 4)) +_GL_CXXALIASWARN1 (wcsrchr, wchar_t *, (wchar_t *wcs, wchar_t wc)); +_GL_CXXALIASWARN1 (wcsrchr, const wchar_t *, (const wchar_t *wcs, wchar_t wc)); +# else +_GL_CXXALIASWARN (wcsrchr); +# endif +#elif defined GNULIB_POSIXCHECK +# undef wcsrchr +# if HAVE_RAW_DECL_WCSRCHR +_GL_WARN_ON_USE (wcsrchr, "wcsrchr is unportable - " + "use gnulib module wcsrchr for portability"); +# endif +#endif + + +/* Return the length of the initial segmet of WCS which consists entirely + of wide characters not in REJECT. */ +#if @GNULIB_WCSCSPN@ +# if !@HAVE_WCSCSPN@ +_GL_FUNCDECL_SYS (wcscspn, size_t, (const wchar_t *wcs, const wchar_t *reject) + _GL_ATTRIBUTE_PURE); +# endif +_GL_CXXALIAS_SYS (wcscspn, size_t, (const wchar_t *wcs, const wchar_t *reject)); +_GL_CXXALIASWARN (wcscspn); +#elif defined GNULIB_POSIXCHECK +# undef wcscspn +# if HAVE_RAW_DECL_WCSCSPN +_GL_WARN_ON_USE (wcscspn, "wcscspn is unportable - " + "use gnulib module wcscspn for portability"); +# endif +#endif + + +/* Return the length of the initial segmet of WCS which consists entirely + of wide characters in ACCEPT. */ +#if @GNULIB_WCSSPN@ +# if !@HAVE_WCSSPN@ +_GL_FUNCDECL_SYS (wcsspn, size_t, (const wchar_t *wcs, const wchar_t *accept) + _GL_ATTRIBUTE_PURE); +# endif +_GL_CXXALIAS_SYS (wcsspn, size_t, (const wchar_t *wcs, const wchar_t *accept)); +_GL_CXXALIASWARN (wcsspn); +#elif defined GNULIB_POSIXCHECK +# undef wcsspn +# if HAVE_RAW_DECL_WCSSPN +_GL_WARN_ON_USE (wcsspn, "wcsspn is unportable - " + "use gnulib module wcsspn for portability"); +# endif +#endif + + +/* Find the first occurrence in WCS of any character in ACCEPT. */ +#if @GNULIB_WCSPBRK@ +# if !@HAVE_WCSPBRK@ +_GL_FUNCDECL_SYS (wcspbrk, wchar_t *, + (const wchar_t *wcs, const wchar_t *accept) + _GL_ATTRIBUTE_PURE); +# endif + /* On some systems, this function is defined as an overloaded function: + extern "C++" { + const wchar_t * std::wcspbrk (const wchar_t *, const wchar_t *); + wchar_t * std::wcspbrk (wchar_t *, const wchar_t *); + } */ +_GL_CXXALIAS_SYS_CAST2 (wcspbrk, + wchar_t *, (const wchar_t *, const wchar_t *), + const wchar_t *, (const wchar_t *, const wchar_t *)); +# if ((__GLIBC__ == 2 && __GLIBC_MINOR__ >= 10) && !defined __UCLIBC__) \ + && (__GNUC__ > 4 || (__GNUC__ == 4 && __GNUC_MINOR__ >= 4)) +_GL_CXXALIASWARN1 (wcspbrk, wchar_t *, + (wchar_t *wcs, const wchar_t *accept)); +_GL_CXXALIASWARN1 (wcspbrk, const wchar_t *, + (const wchar_t *wcs, const wchar_t *accept)); +# else +_GL_CXXALIASWARN (wcspbrk); +# endif +#elif defined GNULIB_POSIXCHECK +# undef wcspbrk +# if HAVE_RAW_DECL_WCSPBRK +_GL_WARN_ON_USE (wcspbrk, "wcspbrk is unportable - " + "use gnulib module wcspbrk for portability"); +# endif +#endif + + +/* Find the first occurrence of NEEDLE in HAYSTACK. */ +#if @GNULIB_WCSSTR@ +# if !@HAVE_WCSSTR@ +_GL_FUNCDECL_SYS (wcsstr, wchar_t *, + (const wchar_t *haystack, const wchar_t *needle) + _GL_ATTRIBUTE_PURE); +# endif + /* On some systems, this function is defined as an overloaded function: + extern "C++" { + const wchar_t * std::wcsstr (const wchar_t *, const wchar_t *); + wchar_t * std::wcsstr (wchar_t *, const wchar_t *); + } */ +_GL_CXXALIAS_SYS_CAST2 (wcsstr, + wchar_t *, (const wchar_t *, const wchar_t *), + const wchar_t *, (const wchar_t *, const wchar_t *)); +# if ((__GLIBC__ == 2 && __GLIBC_MINOR__ >= 10) && !defined __UCLIBC__) \ + && (__GNUC__ > 4 || (__GNUC__ == 4 && __GNUC_MINOR__ >= 4)) +_GL_CXXALIASWARN1 (wcsstr, wchar_t *, + (wchar_t *haystack, const wchar_t *needle)); +_GL_CXXALIASWARN1 (wcsstr, const wchar_t *, + (const wchar_t *haystack, const wchar_t *needle)); +# else +_GL_CXXALIASWARN (wcsstr); +# endif +#elif defined GNULIB_POSIXCHECK +# undef wcsstr +# if HAVE_RAW_DECL_WCSSTR +_GL_WARN_ON_USE (wcsstr, "wcsstr is unportable - " + "use gnulib module wcsstr for portability"); +# endif +#endif + + +/* Divide WCS into tokens separated by characters in DELIM. */ +#if @GNULIB_WCSTOK@ +# if !@HAVE_WCSTOK@ +_GL_FUNCDECL_SYS (wcstok, wchar_t *, + (wchar_t *wcs, const wchar_t *delim, wchar_t **ptr)); +# endif +_GL_CXXALIAS_SYS (wcstok, wchar_t *, + (wchar_t *wcs, const wchar_t *delim, wchar_t **ptr)); +_GL_CXXALIASWARN (wcstok); +#elif defined GNULIB_POSIXCHECK +# undef wcstok +# if HAVE_RAW_DECL_WCSTOK +_GL_WARN_ON_USE (wcstok, "wcstok is unportable - " + "use gnulib module wcstok for portability"); +# endif +#endif + + +/* Determine number of column positions required for first N wide + characters (or fewer if S ends before this) in S. */ +#if @GNULIB_WCSWIDTH@ +# if @REPLACE_WCSWIDTH@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef wcswidth +# define wcswidth rpl_wcswidth +# endif +_GL_FUNCDECL_RPL (wcswidth, int, (const wchar_t *s, size_t n) + _GL_ATTRIBUTE_PURE); +_GL_CXXALIAS_RPL (wcswidth, int, (const wchar_t *s, size_t n)); +# else +# if !@HAVE_WCSWIDTH@ +_GL_FUNCDECL_SYS (wcswidth, int, (const wchar_t *s, size_t n) + _GL_ATTRIBUTE_PURE); +# endif +_GL_CXXALIAS_SYS (wcswidth, int, (const wchar_t *s, size_t n)); +# endif +_GL_CXXALIASWARN (wcswidth); +#elif defined GNULIB_POSIXCHECK +# undef wcswidth +# if HAVE_RAW_DECL_WCSWIDTH +_GL_WARN_ON_USE (wcswidth, "wcswidth is unportable - " + "use gnulib module wcswidth for portability"); +# endif +#endif + + +/* Convert *TP to a date and time wide string. See + <http://pubs.opengroup.org/onlinepubs/9699919799/functions/wcsftime.html>. */ +#if @GNULIB_WCSFTIME@ +# if @REPLACE_WCSFTIME@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# undef wcsftime +# define wcsftime rpl_wcsftime +# endif +_GL_FUNCDECL_RPL (wcsftime, size_t, (wchar_t *__buf, size_t __bufsize, + const wchar_t *__fmt, const struct tm *__tp) + _GL_ARG_NONNULL ((1, 3, 4))); +_GL_CXXALIAS_RPL (wcsftime, size_t, (wchar_t *__buf, size_t __bufsize, + const wchar_t *__fmt, const struct tm *__tp)); +# else +# if !@HAVE_WCSFTIME@ +_GL_FUNCDECL_SYS (wcsftime, size_t, (wchar_t *__buf, size_t __bufsize, + const wchar_t *__fmt, const struct tm *__tp) + _GL_ARG_NONNULL ((1, 3, 4))); +# endif +_GL_CXXALIAS_SYS (wcsftime, size_t, (wchar_t *__buf, size_t __bufsize, + const wchar_t *__fmt, const struct tm *__tp)); +# endif +_GL_CXXALIASWARN (wcsftime); +#elif defined GNULIB_POSIXCHECK +# undef wcsftime +# if HAVE_RAW_DECL_WCSFTIME +_GL_WARN_ON_USE (wcsftime, "wcsftime is unportable - " + "use gnulib module wcsftime for portability"); +# endif +#endif + + +#endif /* _@GUARD_PREFIX@_WCHAR_H */ +#endif /* _@GUARD_PREFIX@_WCHAR_H */ +#endif diff --git a/lib/wcrtomb.c b/lib/wcrtomb.c new file mode 100644 index 0000000..8f01972 --- /dev/null +++ b/lib/wcrtomb.c @@ -0,0 +1,53 @@ +/* Convert wide character to multibyte character. + Copyright (C) 2008-2018 Free Software Foundation, Inc. + Written by Bruno Haible <bruno@clisp.org>, 2008. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +#include <config.h> + +/* Specification. */ +#include <wchar.h> + +#include <errno.h> +#include <stdlib.h> + + +size_t +wcrtomb (char *s, wchar_t wc, mbstate_t *ps) +{ + /* This implementation of wcrtomb on top of wctomb() supports only + stateless encodings. ps must be in the initial state. */ + if (ps != NULL && !mbsinit (ps)) + { + errno = EINVAL; + return (size_t)(-1); + } + + if (s == NULL) + /* We know the NUL wide character corresponds to the NUL character. */ + return 1; + else + { + int ret = wctomb (s, wc); + + if (ret >= 0) + return ret; + else + { + errno = EILSEQ; + return (size_t)(-1); + } + } +} diff --git a/lib/wctob.c b/lib/wctob.c new file mode 100644 index 0000000..070f97a --- /dev/null +++ b/lib/wctob.c @@ -0,0 +1,38 @@ +/* Convert wide character to unibyte character. + Copyright (C) 2008, 2010-2018 Free Software Foundation, Inc. + Written by Bruno Haible <bruno@clisp.org>, 2008. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +#include <config.h> + +/* Specification. */ +#include <wchar.h> + +#include <stdio.h> +#include <stdlib.h> + +int +wctob (wint_t wc) +{ + char buf[64]; + + if (!(MB_CUR_MAX <= sizeof (buf))) + abort (); + /* Handle the case where WEOF is a value that does not fit in a wchar_t. */ + if (wc == (wchar_t)wc) + if (wctomb (buf, (wchar_t)wc) == 1) + return (unsigned char) buf[0]; + return EOF; +} diff --git a/lib/wctomb-impl.h b/lib/wctomb-impl.h new file mode 100644 index 0000000..58c3c0c --- /dev/null +++ b/lib/wctomb-impl.h @@ -0,0 +1,34 @@ +/* Convert wide character to multibyte character. + Copyright (C) 2011-2018 Free Software Foundation, Inc. + Written by Bruno Haible <bruno@clisp.org>, 2011. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +int +wctomb (char *s, wchar_t wc) +{ + if (s == NULL) + return 0; + else + { + mbstate_t state; + size_t result; + + memset (&state, 0, sizeof (mbstate_t)); + result = wcrtomb (s, wc, &state); + if (result == (size_t)-1) + return -1; + return result; + } +} diff --git a/lib/wctomb.c b/lib/wctomb.c new file mode 100644 index 0000000..47addc4 --- /dev/null +++ b/lib/wctomb.c @@ -0,0 +1,25 @@ +/* Convert wide character to multibyte character. + Copyright (C) 2011-2018 Free Software Foundation, Inc. + Written by Bruno Haible <bruno@clisp.org>, 2011. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +#include <config.h> + +#include <stdlib.h> + +#include <string.h> +#include <wchar.h> + +#include "wctomb-impl.h" diff --git a/lib/wctype-h.c b/lib/wctype-h.c new file mode 100644 index 0000000..bb5f847 --- /dev/null +++ b/lib/wctype-h.c @@ -0,0 +1,4 @@ +/* Normally this would be wctype.c, but that name's already taken. */ +#include <config.h> +#define _GL_WCTYPE_INLINE _GL_EXTERN_INLINE +#include "wctype.h" diff --git a/lib/wctype.in.h b/lib/wctype.in.h new file mode 100644 index 0000000..3c6e5f2 --- /dev/null +++ b/lib/wctype.in.h @@ -0,0 +1,533 @@ +/* A substitute for ISO C99 <wctype.h>, for platforms that lack it. + + Copyright (C) 2006-2018 Free Software Foundation, Inc. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3, or (at your option) + any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, see <https://www.gnu.org/licenses/>. */ + +/* Written by Bruno Haible and Paul Eggert. */ + +/* + * ISO C 99 <wctype.h> for platforms that lack it. + * <http://www.opengroup.org/susv3xbd/wctype.h.html> + * + * iswctype, towctrans, towlower, towupper, wctrans, wctype, + * wctrans_t, and wctype_t are not yet implemented. + */ + +#if __GNUC__ >= 3 +@PRAGMA_SYSTEM_HEADER@ +#endif +@PRAGMA_COLUMNS@ + +#if (defined __MINGW32__ && defined __CTYPE_H_SOURCED__) + +/* Special invocation convention: + - With MinGW 3.22, when <ctype.h> includes <wctype.h>, only some part of + <wctype.h> is being processed, which doesn't include the idempotency + guard. */ + +#@INCLUDE_NEXT@ @NEXT_WCTYPE_H@ + +#else +/* Normal invocation convention. */ + +#ifndef _@GUARD_PREFIX@_WCTYPE_H + +#if @HAVE_WINT_T@ +/* Solaris 2.5 has a bug: <wchar.h> must be included before <wctype.h>. + Tru64 with Desktop Toolkit C has a bug: <stdio.h> must be included before + <wchar.h>. + BSD/OS 4.0.1 has a bug: <stddef.h>, <stdio.h> and <time.h> must be + included before <wchar.h>. */ +# include <stddef.h> +# include <stdio.h> +# include <time.h> +# include <wchar.h> +#endif + +/* Native Windows (mingw, MSVC) have declarations of towupper, towlower, and + isw* functions in <ctype.h>, <wchar.h> as well as in <wctype.h>. Include + <ctype.h>, <wchar.h> in advance to avoid rpl_ prefix being added to the + declarations. */ +#if defined _WIN32 && ! defined __CYGWIN__ +# include <ctype.h> +# include <wchar.h> +#endif + +/* Include the original <wctype.h> if it exists. + BeOS 5 has the functions but no <wctype.h>. */ +/* The include_next requires a split double-inclusion guard. */ +#if @HAVE_WCTYPE_H@ +# @INCLUDE_NEXT@ @NEXT_WCTYPE_H@ +#endif + +#ifndef _@GUARD_PREFIX@_WCTYPE_H +#define _@GUARD_PREFIX@_WCTYPE_H + +#ifndef _GL_INLINE_HEADER_BEGIN + #error "Please include config.h first." +#endif +_GL_INLINE_HEADER_BEGIN +#ifndef _GL_WCTYPE_INLINE +# define _GL_WCTYPE_INLINE _GL_INLINE +#endif + +/* The definitions of _GL_FUNCDECL_RPL etc. are copied here. */ + +/* The definition of _GL_WARN_ON_USE is copied here. */ + +/* Solaris 2.6 <wctype.h> includes <widec.h> which includes <euc.h> which + #defines a number of identifiers in the application namespace. Revert + these #defines. */ +#ifdef __sun +# undef multibyte +# undef eucw1 +# undef eucw2 +# undef eucw3 +# undef scrw1 +# undef scrw2 +# undef scrw3 +#endif + +/* Define wint_t and WEOF. (Also done in wchar.in.h.) */ +#if !@HAVE_WINT_T@ && !defined wint_t +# define wint_t int +# ifndef WEOF +# define WEOF -1 +# endif +#else +/* mingw and MSVC define wint_t as 'unsigned short' in <crtdefs.h> or + <stddef.h>. This is too small: ISO C 99 section 7.24.1.(2) says that + wint_t must be "unchanged by default argument promotions". Override it. */ +# if @GNULIB_OVERRIDES_WINT_T@ +# if !GNULIB_defined_wint_t +# if @HAVE_CRTDEFS_H@ +# include <crtdefs.h> +# else +# include <stddef.h> +# endif +typedef unsigned int rpl_wint_t; +# undef wint_t +# define wint_t rpl_wint_t +# define GNULIB_defined_wint_t 1 +# endif +# endif +# ifndef WEOF +# define WEOF ((wint_t) -1) +# endif +#endif + + +#if !GNULIB_defined_wctype_functions + +/* FreeBSD 4.4 to 4.11 has <wctype.h> but lacks the functions. + Linux libc5 has <wctype.h> and the functions but they are broken. + Assume all 11 functions (all isw* except iswblank) are implemented the + same way, or not at all. */ +# if ! @HAVE_ISWCNTRL@ || @REPLACE_ISWCNTRL@ + +/* IRIX 5.3 has macros but no functions, its isw* macros refer to an + undefined variable _ctmp_ and to <ctype.h> macros like _P, and they + refer to system functions like _iswctype that are not in the + standard C library. Rather than try to get ancient buggy + implementations like this to work, just disable them. */ +# undef iswalnum +# undef iswalpha +# undef iswblank +# undef iswcntrl +# undef iswdigit +# undef iswgraph +# undef iswlower +# undef iswprint +# undef iswpunct +# undef iswspace +# undef iswupper +# undef iswxdigit +# undef towlower +# undef towupper + +/* Linux libc5 has <wctype.h> and the functions but they are broken. */ +# if @REPLACE_ISWCNTRL@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# define iswalnum rpl_iswalnum +# define iswalpha rpl_iswalpha +# define iswblank rpl_iswblank +# define iswcntrl rpl_iswcntrl +# define iswdigit rpl_iswdigit +# define iswgraph rpl_iswgraph +# define iswlower rpl_iswlower +# define iswprint rpl_iswprint +# define iswpunct rpl_iswpunct +# define iswspace rpl_iswspace +# define iswupper rpl_iswupper +# define iswxdigit rpl_iswxdigit +# endif +# endif +# if @REPLACE_TOWLOWER@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# define towlower rpl_towlower +# define towupper rpl_towupper +# endif +# endif + +_GL_WCTYPE_INLINE int +# if @REPLACE_ISWCNTRL@ +rpl_iswalnum +# else +iswalnum +# endif + (wint_t wc) +{ + return ((wc >= '0' && wc <= '9') + || ((wc & ~0x20) >= 'A' && (wc & ~0x20) <= 'Z')); +} + +_GL_WCTYPE_INLINE int +# if @REPLACE_ISWCNTRL@ +rpl_iswalpha +# else +iswalpha +# endif + (wint_t wc) +{ + return (wc & ~0x20) >= 'A' && (wc & ~0x20) <= 'Z'; +} + +_GL_WCTYPE_INLINE int +# if @REPLACE_ISWCNTRL@ +rpl_iswblank +# else +iswblank +# endif + (wint_t wc) +{ + return wc == ' ' || wc == '\t'; +} + +_GL_WCTYPE_INLINE int +# if @REPLACE_ISWCNTRL@ +rpl_iswcntrl +# else +iswcntrl +# endif + (wint_t wc) +{ + return (wc & ~0x1f) == 0 || wc == 0x7f; +} + +_GL_WCTYPE_INLINE int +# if @REPLACE_ISWCNTRL@ +rpl_iswdigit +# else +iswdigit +# endif + (wint_t wc) +{ + return wc >= '0' && wc <= '9'; +} + +_GL_WCTYPE_INLINE int +# if @REPLACE_ISWCNTRL@ +rpl_iswgraph +# else +iswgraph +# endif + (wint_t wc) +{ + return wc >= '!' && wc <= '~'; +} + +_GL_WCTYPE_INLINE int +# if @REPLACE_ISWCNTRL@ +rpl_iswlower +# else +iswlower +# endif + (wint_t wc) +{ + return wc >= 'a' && wc <= 'z'; +} + +_GL_WCTYPE_INLINE int +# if @REPLACE_ISWCNTRL@ +rpl_iswprint +# else +iswprint +# endif + (wint_t wc) +{ + return wc >= ' ' && wc <= '~'; +} + +_GL_WCTYPE_INLINE int +# if @REPLACE_ISWCNTRL@ +rpl_iswpunct +# else +iswpunct +# endif + (wint_t wc) +{ + return (wc >= '!' && wc <= '~' + && !((wc >= '0' && wc <= '9') + || ((wc & ~0x20) >= 'A' && (wc & ~0x20) <= 'Z'))); +} + +_GL_WCTYPE_INLINE int +# if @REPLACE_ISWCNTRL@ +rpl_iswspace +# else +iswspace +# endif + (wint_t wc) +{ + return (wc == ' ' || wc == '\t' + || wc == '\n' || wc == '\v' || wc == '\f' || wc == '\r'); +} + +_GL_WCTYPE_INLINE int +# if @REPLACE_ISWCNTRL@ +rpl_iswupper +# else +iswupper +# endif + (wint_t wc) +{ + return wc >= 'A' && wc <= 'Z'; +} + +_GL_WCTYPE_INLINE int +# if @REPLACE_ISWCNTRL@ +rpl_iswxdigit +# else +iswxdigit +# endif + (wint_t wc) +{ + return ((wc >= '0' && wc <= '9') + || ((wc & ~0x20) >= 'A' && (wc & ~0x20) <= 'F')); +} + +_GL_WCTYPE_INLINE wint_t +# if @REPLACE_TOWLOWER@ +rpl_towlower +# else +towlower +# endif + (wint_t wc) +{ + return (wc >= 'A' && wc <= 'Z' ? wc - 'A' + 'a' : wc); +} + +_GL_WCTYPE_INLINE wint_t +# if @REPLACE_TOWLOWER@ +rpl_towupper +# else +towupper +# endif + (wint_t wc) +{ + return (wc >= 'a' && wc <= 'z' ? wc - 'a' + 'A' : wc); +} + +# elif @GNULIB_ISWBLANK@ && (! @HAVE_ISWBLANK@ || @REPLACE_ISWBLANK@) +/* Only the iswblank function is missing. */ + +# if @REPLACE_ISWBLANK@ +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# define iswblank rpl_iswblank +# endif +_GL_FUNCDECL_RPL (iswblank, int, (wint_t wc)); +# else +_GL_FUNCDECL_SYS (iswblank, int, (wint_t wc)); +# endif + +# endif + +# if defined __MINGW32__ + +/* On native Windows, wchar_t is uint16_t, and wint_t is uint32_t. + The functions towlower and towupper are implemented in the MSVCRT library + to take a wchar_t argument and return a wchar_t result. mingw declares + these functions to take a wint_t argument and return a wint_t result. + This means that: + 1. When the user passes an argument outside the range 0x0000..0xFFFF, the + function will look only at the lower 16 bits. This is allowed according + to POSIX. + 2. The return value is returned in the lower 16 bits of the result register. + The upper 16 bits are random: whatever happened to be in that part of the + result register. We need to fix this by adding a zero-extend from + wchar_t to wint_t after the call. */ + +_GL_WCTYPE_INLINE wint_t +rpl_towlower (wint_t wc) +{ + return (wint_t) (wchar_t) towlower (wc); +} +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# define towlower rpl_towlower +# endif + +_GL_WCTYPE_INLINE wint_t +rpl_towupper (wint_t wc) +{ + return (wint_t) (wchar_t) towupper (wc); +} +# if !(defined __cplusplus && defined GNULIB_NAMESPACE) +# define towupper rpl_towupper +# endif + +# endif /* __MINGW32__ */ + +# define GNULIB_defined_wctype_functions 1 +#endif + +#if @REPLACE_ISWCNTRL@ +_GL_CXXALIAS_RPL (iswalnum, int, (wint_t wc)); +_GL_CXXALIAS_RPL (iswalpha, int, (wint_t wc)); +_GL_CXXALIAS_RPL (iswcntrl, int, (wint_t wc)); +_GL_CXXALIAS_RPL (iswdigit, int, (wint_t wc)); +_GL_CXXALIAS_RPL (iswgraph, int, (wint_t wc)); +_GL_CXXALIAS_RPL (iswlower, int, (wint_t wc)); +_GL_CXXALIAS_RPL (iswprint, int, (wint_t wc)); +_GL_CXXALIAS_RPL (iswpunct, int, (wint_t wc)); +_GL_CXXALIAS_RPL (iswspace, int, (wint_t wc)); +_GL_CXXALIAS_RPL (iswupper, int, (wint_t wc)); +_GL_CXXALIAS_RPL (iswxdigit, int, (wint_t wc)); +#else +_GL_CXXALIAS_SYS (iswalnum, int, (wint_t wc)); +_GL_CXXALIAS_SYS (iswalpha, int, (wint_t wc)); +_GL_CXXALIAS_SYS (iswcntrl, int, (wint_t wc)); +_GL_CXXALIAS_SYS (iswdigit, int, (wint_t wc)); +_GL_CXXALIAS_SYS (iswgraph, int, (wint_t wc)); +_GL_CXXALIAS_SYS (iswlower, int, (wint_t wc)); +_GL_CXXALIAS_SYS (iswprint, int, (wint_t wc)); +_GL_CXXALIAS_SYS (iswpunct, int, (wint_t wc)); +_GL_CXXALIAS_SYS (iswspace, int, (wint_t wc)); +_GL_CXXALIAS_SYS (iswupper, int, (wint_t wc)); +_GL_CXXALIAS_SYS (iswxdigit, int, (wint_t wc)); +#endif +_GL_CXXALIASWARN (iswalnum); +_GL_CXXALIASWARN (iswalpha); +_GL_CXXALIASWARN (iswcntrl); +_GL_CXXALIASWARN (iswdigit); +_GL_CXXALIASWARN (iswgraph); +_GL_CXXALIASWARN (iswlower); +_GL_CXXALIASWARN (iswprint); +_GL_CXXALIASWARN (iswpunct); +_GL_CXXALIASWARN (iswspace); +_GL_CXXALIASWARN (iswupper); +_GL_CXXALIASWARN (iswxdigit); + +#if @GNULIB_ISWBLANK@ +# if @REPLACE_ISWCNTRL@ || @REPLACE_ISWBLANK@ +_GL_CXXALIAS_RPL (iswblank, int, (wint_t wc)); +# else +_GL_CXXALIAS_SYS (iswblank, int, (wint_t wc)); +# endif +_GL_CXXALIASWARN (iswblank); +#endif + +#if !@HAVE_WCTYPE_T@ +# if !GNULIB_defined_wctype_t +typedef void * wctype_t; +# define GNULIB_defined_wctype_t 1 +# endif +#endif + +/* Get a descriptor for a wide character property. */ +#if @GNULIB_WCTYPE@ +# if !@HAVE_WCTYPE_T@ +_GL_FUNCDECL_SYS (wctype, wctype_t, (const char *name)); +# endif +_GL_CXXALIAS_SYS (wctype, wctype_t, (const char *name)); +_GL_CXXALIASWARN (wctype); +#elif defined GNULIB_POSIXCHECK +# undef wctype +# if HAVE_RAW_DECL_WCTYPE +_GL_WARN_ON_USE (wctype, "wctype is unportable - " + "use gnulib module wctype for portability"); +# endif +#endif + +/* Test whether a wide character has a given property. + The argument WC must be either a wchar_t value or WEOF. + The argument DESC must have been returned by the wctype() function. */ +#if @GNULIB_ISWCTYPE@ +# if !@HAVE_WCTYPE_T@ +_GL_FUNCDECL_SYS (iswctype, int, (wint_t wc, wctype_t desc)); +# endif +_GL_CXXALIAS_SYS (iswctype, int, (wint_t wc, wctype_t desc)); +_GL_CXXALIASWARN (iswctype); +#elif defined GNULIB_POSIXCHECK +# undef iswctype +# if HAVE_RAW_DECL_ISWCTYPE +_GL_WARN_ON_USE (iswctype, "iswctype is unportable - " + "use gnulib module iswctype for portability"); +# endif +#endif + +#if @REPLACE_TOWLOWER@ || defined __MINGW32__ +_GL_CXXALIAS_RPL (towlower, wint_t, (wint_t wc)); +_GL_CXXALIAS_RPL (towupper, wint_t, (wint_t wc)); +#else +_GL_CXXALIAS_SYS (towlower, wint_t, (wint_t wc)); +_GL_CXXALIAS_SYS (towupper, wint_t, (wint_t wc)); +#endif +_GL_CXXALIASWARN (towlower); +_GL_CXXALIASWARN (towupper); + +#if !@HAVE_WCTRANS_T@ +# if !GNULIB_defined_wctrans_t +typedef void * wctrans_t; +# define GNULIB_defined_wctrans_t 1 +# endif +#endif + +/* Get a descriptor for a wide character case conversion. */ +#if @GNULIB_WCTRANS@ +# if !@HAVE_WCTRANS_T@ +_GL_FUNCDECL_SYS (wctrans, wctrans_t, (const char *name)); +# endif +_GL_CXXALIAS_SYS (wctrans, wctrans_t, (const char *name)); +_GL_CXXALIASWARN (wctrans); +#elif defined GNULIB_POSIXCHECK +# undef wctrans +# if HAVE_RAW_DECL_WCTRANS +_GL_WARN_ON_USE (wctrans, "wctrans is unportable - " + "use gnulib module wctrans for portability"); +# endif +#endif + +/* Perform a given case conversion on a wide character. + The argument WC must be either a wchar_t value or WEOF. + The argument DESC must have been returned by the wctrans() function. */ +#if @GNULIB_TOWCTRANS@ +# if !@HAVE_WCTRANS_T@ +_GL_FUNCDECL_SYS (towctrans, wint_t, (wint_t wc, wctrans_t desc)); +# endif +_GL_CXXALIAS_SYS (towctrans, wint_t, (wint_t wc, wctrans_t desc)); +_GL_CXXALIASWARN (towctrans); +#elif defined GNULIB_POSIXCHECK +# undef towctrans +# if HAVE_RAW_DECL_TOWCTRANS +_GL_WARN_ON_USE (towctrans, "towctrans is unportable - " + "use gnulib module towctrans for portability"); +# endif +#endif + +_GL_INLINE_HEADER_END + +#endif /* _@GUARD_PREFIX@_WCTYPE_H */ +#endif /* _@GUARD_PREFIX@_WCTYPE_H */ +#endif diff --git a/lib/xalloc-die.c b/lib/xalloc-die.c new file mode 100644 index 0000000..bfe6109 --- /dev/null +++ b/lib/xalloc-die.c @@ -0,0 +1,41 @@ +/* Report a memory allocation failure and exit. + + Copyright (C) 1997-2000, 2002-2004, 2006, 2009-2018 Free Software + Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +#include <config.h> + +#include "xalloc.h" + +#include <stdlib.h> + +#include "error.h" +#include "exitfail.h" + +#include "gettext.h" +#define _(msgid) gettext (msgid) + +void +xalloc_die (void) +{ + error (exit_failure, 0, "%s", _("memory exhausted")); + + /* _Noreturn cannot be given to error, since it may return if + its first argument is 0. To help compilers understand the + xalloc_die does not return, call abort. Also, the abort is a + safety feature if exit_failure is 0 (which shouldn't happen). */ + abort (); +} diff --git a/lib/xalloc-oversized.h b/lib/xalloc-oversized.h new file mode 100644 index 0000000..3426e10 --- /dev/null +++ b/lib/xalloc-oversized.h @@ -0,0 +1,60 @@ +/* xalloc-oversized.h -- memory allocation size checking + + Copyright (C) 1990-2000, 2003-2004, 2006-2018 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +#ifndef XALLOC_OVERSIZED_H_ +#define XALLOC_OVERSIZED_H_ + +#include <stddef.h> +#include <stdint.h> + +/* True if N * S would overflow in a size_t calculation, + or would generate a value larger than PTRDIFF_MAX. + This expands to a constant expression if N and S are both constants. + By gnulib convention, SIZE_MAX represents overflow in size + calculations, so the conservative size_t-based dividend to use here + is SIZE_MAX - 1. */ +#define __xalloc_oversized(n, s) \ + ((size_t) (PTRDIFF_MAX < SIZE_MAX ? PTRDIFF_MAX : SIZE_MAX - 1) / (s) < (n)) + +#if PTRDIFF_MAX < SIZE_MAX +typedef ptrdiff_t __xalloc_count_type; +#else +typedef size_t __xalloc_count_type; +#endif + +/* Return 1 if an array of N objects, each of size S, cannot exist + reliably due to size or ptrdiff_t arithmetic overflow. S must be + positive and N must be nonnegative. This is a macro, not a + function, so that it works correctly even when SIZE_MAX < N. */ + +#if 7 <= __GNUC__ +# define xalloc_oversized(n, s) \ + __builtin_mul_overflow_p (n, s, (__xalloc_count_type) 1) +#elif 5 <= __GNUC__ && !defined __ICC && !__STRICT_ANSI__ +# define xalloc_oversized(n, s) \ + (__builtin_constant_p (n) && __builtin_constant_p (s) \ + ? __xalloc_oversized (n, s) \ + : ({ __xalloc_count_type __xalloc_count; \ + __builtin_mul_overflow (n, s, &__xalloc_count); })) + +/* Other compilers use integer division; this may be slower but is + more portable. */ +#else +# define xalloc_oversized(n, s) __xalloc_oversized (n, s) +#endif + +#endif /* !XALLOC_OVERSIZED_H_ */ diff --git a/lib/xalloc.h b/lib/xalloc.h new file mode 100644 index 0000000..c419a2d --- /dev/null +++ b/lib/xalloc.h @@ -0,0 +1,262 @@ +/* xalloc.h -- malloc with out-of-memory checking + + Copyright (C) 1990-2000, 2003-2004, 2006-2018 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +#ifndef XALLOC_H_ +#define XALLOC_H_ + +#include <stddef.h> +#include <stdint.h> + +#include "xalloc-oversized.h" + +#ifndef _GL_INLINE_HEADER_BEGIN + #error "Please include config.h first." +#endif +_GL_INLINE_HEADER_BEGIN +#ifndef XALLOC_INLINE +# define XALLOC_INLINE _GL_INLINE +#endif + +#ifdef __cplusplus +extern "C" { +#endif + + +#if ! defined __clang__ && \ + (__GNUC__ > 4 || (__GNUC__ == 4 && __GNUC_MINOR__ >= 3)) +# define _GL_ATTRIBUTE_ALLOC_SIZE(args) __attribute__ ((__alloc_size__ args)) +#else +# define _GL_ATTRIBUTE_ALLOC_SIZE(args) +#endif + +/* This function is always triggered when memory is exhausted. + It must be defined by the application, either explicitly + or by using gnulib's xalloc-die module. This is the + function to call when one wants the program to die because of a + memory allocation failure. */ +extern _Noreturn void xalloc_die (void); + +void *xmalloc (size_t s) + _GL_ATTRIBUTE_MALLOC _GL_ATTRIBUTE_ALLOC_SIZE ((1)); +void *xzalloc (size_t s) + _GL_ATTRIBUTE_MALLOC _GL_ATTRIBUTE_ALLOC_SIZE ((1)); +void *xcalloc (size_t n, size_t s) + _GL_ATTRIBUTE_MALLOC _GL_ATTRIBUTE_ALLOC_SIZE ((1, 2)); +void *xrealloc (void *p, size_t s) + _GL_ATTRIBUTE_ALLOC_SIZE ((2)); +void *x2realloc (void *p, size_t *pn); +void *xmemdup (void const *p, size_t s) + _GL_ATTRIBUTE_ALLOC_SIZE ((2)); +char *xstrdup (char const *str) + _GL_ATTRIBUTE_MALLOC; + +/* In the following macros, T must be an elementary or structure/union or + typedef'ed type, or a pointer to such a type. To apply one of the + following macros to a function pointer or array type, you need to typedef + it first and use the typedef name. */ + +/* Allocate an object of type T dynamically, with error checking. */ +/* extern t *XMALLOC (typename t); */ +#define XMALLOC(t) ((t *) xmalloc (sizeof (t))) + +/* Allocate memory for N elements of type T, with error checking. */ +/* extern t *XNMALLOC (size_t n, typename t); */ +#define XNMALLOC(n, t) \ + ((t *) (sizeof (t) == 1 ? xmalloc (n) : xnmalloc (n, sizeof (t)))) + +/* Allocate an object of type T dynamically, with error checking, + and zero it. */ +/* extern t *XZALLOC (typename t); */ +#define XZALLOC(t) ((t *) xzalloc (sizeof (t))) + +/* Allocate memory for N elements of type T, with error checking, + and zero it. */ +/* extern t *XCALLOC (size_t n, typename t); */ +#define XCALLOC(n, t) \ + ((t *) (sizeof (t) == 1 ? xzalloc (n) : xcalloc (n, sizeof (t)))) + + +/* Allocate an array of N objects, each with S bytes of memory, + dynamically, with error checking. S must be nonzero. */ + +XALLOC_INLINE void *xnmalloc (size_t n, size_t s) + _GL_ATTRIBUTE_MALLOC _GL_ATTRIBUTE_ALLOC_SIZE ((1, 2)); +XALLOC_INLINE void * +xnmalloc (size_t n, size_t s) +{ + if (xalloc_oversized (n, s)) + xalloc_die (); + return xmalloc (n * s); +} + +/* Change the size of an allocated block of memory P to an array of N + objects each of S bytes, with error checking. S must be nonzero. */ + +XALLOC_INLINE void *xnrealloc (void *p, size_t n, size_t s) + _GL_ATTRIBUTE_ALLOC_SIZE ((2, 3)); +XALLOC_INLINE void * +xnrealloc (void *p, size_t n, size_t s) +{ + if (xalloc_oversized (n, s)) + xalloc_die (); + return xrealloc (p, n * s); +} + +/* If P is null, allocate a block of at least *PN such objects; + otherwise, reallocate P so that it contains more than *PN objects + each of S bytes. S must be nonzero. Set *PN to the new number of + objects, and return the pointer to the new block. *PN is never set + to zero, and the returned pointer is never null. + + Repeated reallocations are guaranteed to make progress, either by + allocating an initial block with a nonzero size, or by allocating a + larger block. + + In the following implementation, nonzero sizes are increased by a + factor of approximately 1.5 so that repeated reallocations have + O(N) overall cost rather than O(N**2) cost, but the + specification for this function does not guarantee that rate. + + Here is an example of use: + + int *p = NULL; + size_t used = 0; + size_t allocated = 0; + + void + append_int (int value) + { + if (used == allocated) + p = x2nrealloc (p, &allocated, sizeof *p); + p[used++] = value; + } + + This causes x2nrealloc to allocate a block of some nonzero size the + first time it is called. + + To have finer-grained control over the initial size, set *PN to a + nonzero value before calling this function with P == NULL. For + example: + + int *p = NULL; + size_t used = 0; + size_t allocated = 0; + size_t allocated1 = 1000; + + void + append_int (int value) + { + if (used == allocated) + { + p = x2nrealloc (p, &allocated1, sizeof *p); + allocated = allocated1; + } + p[used++] = value; + } + + */ + +XALLOC_INLINE void * +x2nrealloc (void *p, size_t *pn, size_t s) +{ + size_t n = *pn; + + if (! p) + { + if (! n) + { + /* The approximate size to use for initial small allocation + requests, when the invoking code specifies an old size of + zero. This is the largest "small" request for the GNU C + library malloc. */ + enum { DEFAULT_MXFAST = 64 * sizeof (size_t) / 4 }; + + n = DEFAULT_MXFAST / s; + n += !n; + } + if (xalloc_oversized (n, s)) + xalloc_die (); + } + else + { + /* Set N = floor (1.5 * N) + 1 so that progress is made even if N == 0. + Check for overflow, so that N * S stays in both ptrdiff_t and + size_t range. The check may be slightly conservative, but an + exact check isn't worth the trouble. */ + if ((PTRDIFF_MAX < SIZE_MAX ? PTRDIFF_MAX : SIZE_MAX) / 3 * 2 / s + <= n) + xalloc_die (); + n += n / 2 + 1; + } + + *pn = n; + return xrealloc (p, n * s); +} + +/* Return a pointer to a new buffer of N bytes. This is like xmalloc, + except it returns char *. */ + +XALLOC_INLINE char *xcharalloc (size_t n) + _GL_ATTRIBUTE_MALLOC _GL_ATTRIBUTE_ALLOC_SIZE ((1)); +XALLOC_INLINE char * +xcharalloc (size_t n) +{ + return XNMALLOC (n, char); +} + +#ifdef __cplusplus +} + +/* C++ does not allow conversions from void * to other pointer types + without a cast. Use templates to work around the problem when + possible. */ + +template <typename T> inline T * +xrealloc (T *p, size_t s) +{ + return (T *) xrealloc ((void *) p, s); +} + +template <typename T> inline T * +xnrealloc (T *p, size_t n, size_t s) +{ + return (T *) xnrealloc ((void *) p, n, s); +} + +template <typename T> inline T * +x2realloc (T *p, size_t *pn) +{ + return (T *) x2realloc ((void *) p, pn); +} + +template <typename T> inline T * +x2nrealloc (T *p, size_t *pn, size_t s) +{ + return (T *) x2nrealloc ((void *) p, pn, s); +} + +template <typename T> inline T * +xmemdup (T const *p, size_t s) +{ + return (T *) xmemdup ((void const *) p, s); +} + +#endif + +_GL_INLINE_HEADER_END + +#endif /* !XALLOC_H_ */ diff --git a/lib/xmalloc.c b/lib/xmalloc.c new file mode 100644 index 0000000..9a30d67 --- /dev/null +++ b/lib/xmalloc.c @@ -0,0 +1,122 @@ +/* xmalloc.c -- malloc with out of memory checking + + Copyright (C) 1990-2000, 2002-2006, 2008-2018 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <https://www.gnu.org/licenses/>. */ + +#include <config.h> + +#define XALLOC_INLINE _GL_EXTERN_INLINE + +#include "xalloc.h" + +#include <stdlib.h> +#include <string.h> + +/* 1 if calloc is known to be compatible with GNU calloc. This + matters if we are not also using the calloc module, which defines + HAVE_CALLOC_GNU and supports the GNU API even on non-GNU platforms. */ +#if defined HAVE_CALLOC_GNU || (defined __GLIBC__ && !defined __UCLIBC__) +enum { HAVE_GNU_CALLOC = 1 }; +#else +enum { HAVE_GNU_CALLOC = 0 }; +#endif + +/* Allocate N bytes of memory dynamically, with error checking. */ + +void * +xmalloc (size_t n) +{ + void *p = malloc (n); + if (!p && n != 0) + xalloc_die (); + return p; +} + +/* Change the size of an allocated block of memory P to N bytes, + with error checking. */ + +void * +xrealloc (void *p, size_t n) +{ + if (!n && p) + { + /* The GNU and C99 realloc behaviors disagree here. Act like + GNU, even if the underlying realloc is C99. */ + free (p); + return NULL; + } + + p = realloc (p, n); + if (!p && n) + xalloc_die (); + return p; +} + +/* If P is null, allocate a block of at least *PN bytes; otherwise, + reallocate P so that it contains more than *PN bytes. *PN must be + nonzero unless P is null. Set *PN to the new block's size, and + return the pointer to the new block. *PN is never set to zero, and + the returned pointer is never null. */ + +void * +x2realloc (void *p, size_t *pn) +{ + return x2nrealloc (p, pn, 1); +} + +/* Allocate S bytes of zeroed memory dynamically, with error checking. + There's no need for xnzalloc (N, S), since it would be equivalent + to xcalloc (N, S). */ + +void * +xzalloc (size_t s) +{ + return memset (xmalloc (s), 0, s); +} + +/* Allocate zeroed memory for N elements of S bytes, with error + checking. S must be nonzero. */ + +void * +xcalloc (size_t n, size_t s) +{ + void *p; + /* Test for overflow, since objects with size greater than + PTRDIFF_MAX cause pointer subtraction to go awry. Omit size-zero + tests if HAVE_GNU_CALLOC, since GNU calloc never returns NULL if + successful. */ + if (xalloc_oversized (n, s) + || (! (p = calloc (n, s)) && (HAVE_GNU_CALLOC || n != 0))) + xalloc_die (); + return p; +} + +/* Clone an object P of size S, with error checking. There's no need + for xnmemdup (P, N, S), since xmemdup (P, N * S) works without any + need for an arithmetic overflow check. */ + +void * +xmemdup (void const *p, size_t s) +{ + return memcpy (xmalloc (s), p, s); +} + +/* Clone STRING. */ + +char * +xstrdup (char const *string) +{ + return xmemdup (string, strlen (string) + 1); +} |