path: root/infrastructure/m4
Commit message (Collapse)AuthorAge
* Detect which debugger to use in t-gdb scripts: lldb or gdbChris Wilson2018-01-01
| | | | | | | | | | | Should make debugging on Mac and BSD less painful. Make t-gdb fail with an error when no debugger was detected. (cherry picked from commit da98a8183737fb1fcb658cebd9612791a4750c4c) (cherry picked from commit ea6bc722131222d1fad08521e61f552f16941b01) (cherry picked from commit 38ba7510f2cad9accf445d5633c64e98a637a139) (cherry picked from commit 02bfd7e5bdd2abc185b12f37e1f536269db6817a)
* Fix intercept of SYS_open on platforms that only have SYS_openat (e.g. arm64)Chris Wilson2017-06-22
* Use AX_CHECK_COMPILE_FLAG, add more checksChris Wilson2017-06-22
| | | | Disable pointless deprecation warnings for std::auto_ptr.
* Detect whether -Werror=return-type is supported by compiler.Chris Wilson2016-09-20
| | | | Fixes compilation on CentOS 5.
* Update platform feature detection to work with CMake.Chris Wilson2016-09-03
| | | | | | | Rewrite some macros in forms that can be parsed by our parser written in CMake language (fake m4 with regexps). Fix incorrect conditional includes and incorrect #if/#ifdef assumptions.
* Fix -rdynamic checkMoritz 'Morty' Strübe2015-10-21
| | | | | The flag -rdynamic must be removed from LDFLAGS after the check. Otherwise its also set for the following checks and they fail.
* Make wrong return type an error, not a warning.Chris Wilson2015-08-29
| | | | We should never fail to return something from a non-void function.
* We must check for ftruncate if we intend to use its feature macro.Chris Wilson2014-12-26
* Add a note about why fixing NetBSD's kqueue support is harder than it looks.Chris Wilson2014-12-15
* Fix detection of filesystems without extended attribute support on NetBSD.Chris Wilson2014-12-06
| | | | | | | | | | | NetBSD (version 6) uses ENOTSUP as the errno code to indicate missing support for extended attribute in the filesystem. This appears to be at odds with other Unixes: We need to detect and handle ENOTSUP to stop the backup daemon from killing itself while trying to read extended attributes from the first file in the backup set.
* Fix missing include of sys/socket.h to find struct ucred on Linux.Chris Wilson2014-12-06
* Check for existence of netdb.h for conditional inclusion.Chris Wilson2014-12-06
* Workaround for old gnu tar, failing to restore timestamps on symlinks.Chris Wilson2014-12-06
| | | | | | | | | | | | | | | Older versions of GNU tar fail to set the timestamps on symlinks, which makes them appear too recent/new to be backed up immediately, causing test_bbackupd_uploads_files() for example to fail. Fixed by restoring the timestamps manually after extracting the fixture archive. For more details about the issue in tar, please see: This resulted in symlinks in fixture test files
* Add support for getsockopt(SO_PEERCRED) on OpenBSD 5.2.Chris Wilson2014-12-06
| | | | | | | | | | | OpenBSD requires different headers to be included, including a bug which makes it essential to include sys/param.h before sys/ucred.h, and gives the fields of struct ucred different names than Linux does. This fixes compilation on OpenBSD and allows the user connecting to a UNIX socket to be checked, increasing security of the command socket on this platform.
* Generate configure script in qdbm directory.Chris Wilson2014-12-01
| | | | | | Now that we've removed it from version control, we need to build it from as part of bootstrapping Box Backup.
* Support microsecond timestamps for files on Linux.Chris Wilson2014-04-09
* Use AC_SEARCH_LIBS instead of AC_CHECK_LIB when searching for SSLChris Wilson2012-11-14
| | | | | libraries, to avoid adding duplicate libraries to LIBS.
* Use AC_SEARCH_LIBS instead of AC_CHECK_LIB to search for EVP_CipherInit_ex,Chris Wilson2012-11-14
| | | | | | | | | | | | as it should already be in LIBS by this point and we don't really want to add a duplicate if we don't need it. also recommends the use of AC_SEARCH_LIBS instead. Check whether we have getpeername() and sys/un.h to fix compile errors on Windows with /dev/log fd leak test.
* Check for -lcrypto before -lssl, to ensure correct order in LIBS and ↵Chris Wilson2012-04-28
| | | | successful link.
* Move libdl tests before boxbackup tests in to fix OpenSSL link ↵Chris Wilson2012-04-28
| | | | errors.
* Hopefully fix TcpNice setsockopt/struct tcp_info compile errors on Windows.Chris Wilson2012-03-16
* Add experimental "TCP Nice" mode, disabled by default.Chris Wilson2012-02-12
* Patch to support building on MinGW, thanks to Paolo ToscoChris Wilson2011-10-17
| | | | | (,
* Silence warnings from new MinGW headers that expect __MINGW_FEATURES__Chris Wilson2011-10-08
| | | | | | | | to be defined. Check for fcntl.h and include it if we have it, not just on MSVC, now that MinGW also defines O_BINARY in newer versions.
* Rephrase -rdynamic message.Chris Wilson2011-10-07
* Skip -rdynamic on cygwin as well as Solaris.Chris Wilson2011-10-05
* Move Box Backup's custom tests into own file, to share with Boxi.Chris Wilson2011-09-25
* Restore the newline in the includes variable, it is actually requiredChris Wilson2011-02-28
| | | | | even though it causes a warning.
* Test whether pdcurses is required to compile with readline support.Chris Wilson2011-02-27
* Workaround for Mac OSX where readline.h doesn't compile if stdio.h isChris Wilson2010-09-29
| | | | | not included first.
* Detect whether rl_completion_matches or completion_matches is declared Chris Wilson2010-02-26
| | | | | | | by readline headers at configure time, and use the appropriate one in bbackupquery. Thanks to Melissa Jenkins for reporting the compilation failure on OSX and providing additional details.
* Blind fix for autoconf problems introduced by earlier attempt to supportChris Wilson2009-04-04
| | | | | cross-compiling.
* Assume that we have no way to determine whether struct dirent has aChris Wilson2009-03-31
| | | | | valid d_type when cross-compiling, as this should be the safest option.
* Assume that lseek does not take a "dummy parameter" when cross-compiling,Chris Wilson2009-03-31
| | | | | | as this is only relevent for intercept tests anyway, and was probably fixed by casting the offset parameter to off_t anyway.
* Reverse the test for __syscall(), remove the test for syscall(),Chris Wilson2008-10-26
| | | | | | | | | | | | | remove the definition and reverse the sense in ax_func_syscall.m4 (which checks for __syscall() needing definition). Autoconf's AC_CHECK_FUNC defines it when testing for its presence, so HAVE___SYSCALL will be true even if __syscall has no definition in the system libraries, and this is precisely the case that we want to test for, so now we test whether the test program compiles with no explicit definition (only the system headers) and if that fails, we set HAVE___SYSCALL_NEED_DEFN to 1.
* Update autoconf to silence warnings on autoconf >= 2.62James O'Gorman2008-10-05
| | | | | | * AC_SUBST now only works with one variable per call * When using AC_CACHE the variable must contain the string '_cv_', thus prefix all cache variables with box_cv_
* Fix raidfile tests on FreeBSD 7Chris Wilson2008-01-23
* Stricter check for dirent.d_type validity (possible problems withChris Wilson2007-12-18
| | | | | | large file support resulting in reading the wrong part of the structure on Slackware?)
* Remove --with-bdb-dir option, to avoid confusion and be consistent with theChris Wilson2007-07-01
| | | | | SSL flags. (merges [1717] by James O'Gorman, refs #3)
* Add --with-bdb-headers and --with-bdb-lib flags to configure to fix BerkeleyChris Wilson2007-07-01
| | | | | | | DB detection issues on FreeBSD. The original --with-bdb-dir flag is still present for systems that install BDB in a "standard" location. (merges [1714] from trunk by James O'Gorman, refs #3)
* Merge new m4 from trunk. (refs #3, merges [1453])Chris Wilson2007-03-24
* Apply patch from Stuart Hickinbottom <> to ↵Ben Summers2006-06-19
| | | | fix complication problems with gcc 4.1.1 where configure didn't detect LLONG_MAX properly. Fix by removing configure check completely, and using STL interface.
* Fix execute permission on substituted perl scriptsMartin Ebourne2006-03-12
* Patch from Per Thomsen to detect statvfs. This is used by recent versions of ↵Martin Ebourne2006-02-23
| | | | NetBSD in preference to statfs.
* Beef up configure checks for 64 bit endian swapping function.Martin Ebourne2006-02-03
| | | | NOTE: Needs testing on other platforms, especially BSD.
* Rewrite configure check for readline lib to match libraries to headers and ↵Martin Ebourne2006-01-04
| | | | | | be more selective. Also changes --enable-gnu-readline which intentionally used to prefer editline over readline even when specified (for licence goodness), to now prefer readline if requested. This is probably less confusing to a user.
* The last commit fixed the header file check, hopefully this will fix the ↵Martin Ebourne2006-01-02
| | | | library check on platforms with only db version 1.
* Try and bodge round the scary mess Berkeley DB has left us. This possible ↵Martin Ebourne2006-01-02
| | | | fix is for FreeBSD which fails to compile db.h
* Now supports Berkeley DB versions >= 4.1 in addition to 1.x. The versions ↵Martin Ebourne2005-12-28
| | | | inbetween are not supported because they require code changes and I don't have them available to test against.
* Fix FreeBSD not finding sys/mount.hMartin Ebourne2005-12-25