path: root/to.do/flex.rmail

BABYL OPTIONS: -*- rmail -*-
Version: 5
Labels:
Note:   This is the header of an rmail file.
Note:   If you are seeing it in rmail,
Note:    it means the file has no messages in it.

1,,
X-Coding-System: nil
Mail-from: From vern@daffy.ee.lbl.gov  Wed Mar 14 04:01:40 2001
Return-Path: <vern@daffy.ee.lbl.gov>
Received: from localhost (localhost [127.0.0.1])
	by michael.uncg.edu (8.9.3/8.9.3) with ESMTP id EAA09959
	for <wlestes@localhost>; Wed, 14 Mar 2001 04:01:40 -0500
Received: from imap.uncg.edu
	by localhost with IMAP (fetchmail-5.1.0)
	for wlestes@localhost (single-drop); Wed, 14 Mar 2001 04:01:40 -0500 (EST)
Received: from external-gw.uncg.edu (external-gw.uncg.edu [152.13.2.70])
	by hermes.email.uncg.edu (8.11.0/8.11.0) with ESMTP id f2E8pme19617
	for <wlestes@hermes.email.uncg.edu>; Wed, 14 Mar 2001 03:51:48 -0500 (EST)
Received: from daffy.ee.lbl.gov (daffy.ee.lbl.gov [131.243.1.31])
	by external-gw.uncg.edu (8.9.3/8.9.3) with ESMTP id DAA20394
	for <wlestes@uncg.edu>; Wed, 14 Mar 2001 03:51:47 -0500 (EST)
Received: (from vern@localhost)
	by daffy.ee.lbl.gov (8.10.0/8.10.0) id f2E8pkM07870;
	Wed, 14 Mar 2001 00:51:46 -0800 (PST)
Message-Id: <200103140851.f2E8pkM07870@daffy.ee.lbl.gov>
To: "W. L. Estes" <wlestes@uncg.edu>
Subject: Re: possibly taking over maintenance of flex 
In-reply-to: Your message of Tue, 13 Mar 2001 09:00:58 EST.
Date: Wed, 14 Mar 2001 00:51:46 PST
From: Vern Paxson <vern@ee.lbl.gov>

*** EOOH ***
To: "W. L. Estes" <wlestes@uncg.edu>
Subject: Re: possibly taking over maintenance of flex 
In-reply-to: Your message of Tue, 13 Mar 2001 09:00:58 EST.
Date: Wed, 14 Mar 2001 00:51:46 PST
From: Vern Paxson <vern@ee.lbl.gov>

> probably best to have a trial period up front as you suggest.

Okay, you can get a copy of the master sources from

        ftp://ftp.ee.lbl.gov/.vp-flex-sources.tar.gz

The TODO/ subdirectory has the general wish list plus a number of contributed
patches.  The faqs/ subdirectory has a bunch of email messages (in MH
folder format, i.e., one message per file, files numbered sequentially)
for commonly asked questions - worth taking a look through to see what
sorts of things people frequently ask about.

Next messages are various stuff I sent to Dick King.

Thanks for giving it a try - naturally, by all means let me know when you
have questions ...

		Vern


1,,
X-Coding-System: nil
Mail-from: From vern@daffy.ee.lbl.gov  Wed Mar 14 04:01:40 2001
Return-Path: <vern@daffy.ee.lbl.gov>
Received: from localhost (localhost [127.0.0.1])
	by michael.uncg.edu (8.9.3/8.9.3) with ESMTP id EAA09961
	for <wlestes@localhost>; Wed, 14 Mar 2001 04:01:40 -0500
Received: from imap.uncg.edu
	by localhost with IMAP (fetchmail-5.1.0)
	for wlestes@localhost (single-drop); Wed, 14 Mar 2001 04:01:40 -0500 (EST)
Received: from external-gw.uncg.edu (external-gw.uncg.edu [152.13.2.70])
	by hermes.email.uncg.edu (8.11.0/8.11.0) with ESMTP id f2E8ppe19633
	for <wlestes@hermes.email.uncg.edu>; Wed, 14 Mar 2001 03:51:51 -0500 (EST)
Received: from daffy.ee.lbl.gov (daffy.ee.lbl.gov [131.243.1.31])
	by external-gw.uncg.edu (8.9.3/8.9.3) with ESMTP id DAA20402
	for <wlestes@uncg.edu>; Wed, 14 Mar 2001 03:51:50 -0500 (EST)
Received: (from vern@localhost)
	by daffy.ee.lbl.gov (8.10.0/8.10.0) id f2E8poV07877;
	Wed, 14 Mar 2001 00:51:50 -0800 (PST)
Message-Id: <200103140851.f2E8poV07877@daffy.ee.lbl.gov>
To: wlestes@uncg.edu
Subject: messages sent to Dick King (1 of 2)
Date: Wed, 14 Mar 2001 00:51:49 PST
From: Vern Paxson <vern@ee.lbl.gov>

*** EOOH ***
To: wlestes@uncg.edu
Subject: messages sent to Dick King (1 of 2)
Date: Wed, 14 Mar 2001 00:51:49 PST
From: Vern Paxson <vern@ee.lbl.gov>


------- Forwarded Messages

Date:  Mon, 21 Aug 2000 23:03:54 PDT
From:  Vern Paxson <vern>
Subject:  Re: Some new Flex maintainer volunteers
To:  Dick King <king@reasoning.com>
Cc:  bkuhn@ebb.org
In-reply-to:  Your message of Tue, 15 Aug 2000 14:03:58 PDT.

> Vern, i'm ready when you are.

Okay, I've put the master sources in

	ftp://ftp.ee.lbl.gov/.vp-flex-sources.tar.gz

The TODO/ subdirectory has a the general wish list plus a number of
contributed patches (per the mail I'll be cc'ing you on regarding reentrant
scanners).  The faqs/ subdirectory has a bunch of email messages (in MH
folder format, i.e., one message per file, files numbered sequentially) for
commonly asked questions - worth taking a look through to see what sorts of
things people frequently ask about.

I very much appreciate your taking over maintaining/developing flex, and
apologize for how long it's taken me to bundle stuff up to give to you (all
too symptomatic of why I'm no longer able to work on flex :-( ).  By all
means, let me know when you have questions ...

		Vern

------- Message 2

Date:  Mon, 21 Aug 2000 23:38:00 PDT
From:  Vern Paxson <vern>
Subject:  Re: I volunteer to modify flex to generate reentrant C code.
To:  "John W. Millaway" <jmillawa@nimbus.ocis.temple.edu>
Cc:  help-flex@gnu.org, king@reasoning.com
In-reply-to:  Your message of Mon, 21 Aug 2000 16:23:18 PDT.

> I'm requesting this feature and at the same time volunteering to implement
> it:	

John Bossom contributed patches a while ago for adding reentrant scanners.
I never managed to try them out or integrate them (note that I'm no longer
the flex maintainer/developer, having just transferred the honors to Dick
King, whom I've cc'd), but I've made them available from:

	ftp://ftp.ee.lbl.gov/.vp-flex-reentrant.tar.gz

I imagine Dick will be interested in hearing of problems/improvements/etc.

		Vern

------- Message 3

Date:  Wed, 30 Aug 2000 07:23:07 PDT
From:  Vern Paxson <vern>
Subject:  Re: bug in flex 2.5.4? 
To:  Keith McGuigan <keith.mcguigan@ecora.com>
Cc:  king@reasoning.com
In-reply-to:  Your message of Tue, 29 Aug 2000 12:21:02 EDT.

This does indeed sound like a bug.  Dick King (cc'd) recently took over
development & maintenance for flex; I've sent him a copy of your message.

		Vern

------- Message 4

Date:  Thu, 31 Aug 2000 15:41:00 PDT
From:  Vern Paxson <vern>
Subject:  Re: Flex 
To:  "Skifstrom, Eric" <SKIFSTRE@Mattel.com>
Cc:  king@reasoning.com
In-reply-to:  Your message of Wed, 30 Aug 2000 09:35:58 PDT.

> Thanks for all the work on Flex.  It is appreciated.  How do I get a copy of
> flex for Windows 98 platform.

I've only directly support flex for Unix, and have always distributed it
in source-code form only.  So I don't know where you get a copy, though
I bet with some web surfing you can find one.

Also, please note, Dick King (cc'd) has taken over flex development
and maintenance.

		Vern

------- Message 5

Date:  Thu, 31 Aug 2000 15:43:40 -0700 (PDT)
From:  Dick King <king@reasoning.com>
Subject:  Re: Flex 
To:  vern@ee.lbl.gov
MIME-Version:  1.0
Content-Type:  text/plain; charset=us-ascii
Content-Transfer-Encoding:  7bit
In-Reply-To:  <200008312241.e7VMf0G14335@daffy.ee.lbl.gov>
References:  <200008312241.e7VMf0G14335@daffy.ee.lbl.gov>
X-Mailer:  VM 6.61 under 19.14 XEmacs Lucid

Vern Paxson writes:
 >> Thanks for all the work on Flex.  It is appreciated.  How do I get a copy of
 >> flex for Windows 98 platform.
 >
 >I've only directly support flex for Unix, and have always distributed it
 >in source-code form only.  So I don't know where you get a copy, though
 >I bet with some web surfing you can find one.
 >
 >Also, please note, Dick King (cc'd) has taken over flex development
 >and maintenance.
 >
 >		Vern
 >

I have a couple of how-to-be-a-gnu-product-maintainer questions.

First, how do i get to the wish list?  I assume people mail the stuff and it
accumulates somewhere.

Second, if i decide i have a fix for a group of problems what do i do to get it
distributed?

- -dk

------- Message 6

Date:  Sun, 03 Sep 2000 12:22:48 PDT
From:  Vern Paxson <vern>
Subject:  Re: Flex 
To:  Dick King <king@reasoning.com>
In-reply-to:  Your message of Thu, 31 Aug 2000 15:43:40 PDT.

> First, how do i get to the wish list?  I assume people mail the stuff and it
> accumulates somewhere.

There's a honed-down copy in the distribution I put together.  I also have
several hundred email messages, which will be rather tedious to wade through,
but I can put together for you if you wish (this is non-trivial, though,
since there's some personal mail intermixed in it which I'll have to locate
and remove).

> Second, if i decide i have a fix for a group of problems what do i do to get it
> distributed?

Put together a new release distribution (either a minor version change or
a major one, depennding on the scope of the changes).  Make sure it works
with "make bigcheck", and make sure you can copy the distribution to a
virgin machine, unpack it & successfully build it (including "make bigcheck").

Next, make it available somewhere for anonymous ftp and/or Web access.
(If this is a problem, I can still host the distribution at ftp.ee.lbl.gov.)

Finally, send mail like the appended example.  For your first message,
you should explain that you've taken over development of flex from me.
It would probably be good to cc me, in case a thread develops from the
note that it would be helpful to have me contribute to.  No need to cc
me on later notes (though I don't mind, either).

I used to send the note to compilers@iecc.com (the USENET comp.compilers
group) and info-gnu@prep.ai.mit.edu.  I don't know if the former is still
the correct address - you could find out from the USENET group's archives,
presumably.  The latter has likely changed to info-gnu@gnu.org.

		Vern


Date:  Tue, 28 Mar 95 11:58:04 PST
From:  Vern Paxson <vern>
Subject:  flex release 2.5.1 now available
To:  compilers@iecc.com, info-gnu@prep.ai.mit.edu

Release 2.5.1 of flex, a POSIX-compliant "lex" scanner generator, is now
available for anonymous ftp to ftp.ee.lbl.gov.  Retrieve flex-2.5.1.tar.Z
or flex-2.5.1.tar.gz.  flex-2.5.1.tar.gz should be showing up shortly at
the usual GNU mirror sites.

The main changes between 2.5 and 2.4 are:

	- A new concept of "start condition" scope lets you group together
	  rules sharing the same start conditions into a syntactic unit.
	- Flex now includes a general mechanism for specifying scanner options.
	- Routines have been added for scanning strings instead of files.
	- Routines have been added for manipulating stacks of start conditions.
	- Fledgling support for POSIX character class expressions.
	- Enhanced portability to MS-DOS, VMS, NT, Macintosh, Amiga.
	- A number of bugs have been fixed.
	- input() no longer trashes yytext.
	- Interactive scanners now run significantly faster.
	- C++ scanner objects now work with the -P option.
	- The promised rewrite of the C++ FlexLexer class has not been done,
	  but is still planned.

See the "NEWS" file for more details.

Please report bugs to: vern@ee.lbl.gov or bug-gnu-utils@prep.ai.mit.edu

		Vern

Vern Paxson
Lawrence Berkeley Laboratory
vern@ee.lbl.gov

------- Message 7

Date:  Wed, 13 Sep 2000 00:59:33 PDT
From:  Vern Paxson <vern>
Subject:  Re: Does the FSF maintain Flex?
To:  "Leanard Lin" <leanard@syntest.com.tw>
Cc:  king@reasoning.com
In-reply-to:  Your message of Wed, 13 Sep 2000 00:55:38 PDT.

> Dear Vern,
> 
> I have a problem about flex of linux version.
> If comment is very long, flex will not re-malloc
> for it.
> It will display a error message and exit program.
> 
> Thanks,
> 
> Leanard

Well, clearly if a comment is larger than the available memory, then flex
can't realloc to accommodate it.  But presumably that's not what's going on
or you wouldn't have bothered writing; so I don't know what the problem
might be.

Flex is now developed & maintained by Dick King, king@reasoning.com, whom
I've cc'd.

		Vern

------- Message 8

Date:  Wed, 04 Oct 2000 23:36:59 PDT
From:  Vern Paxson <vern>
Subject:  Re: Flex 2.5.4.1 Bugs
To:  Hans Aberg <haberg@matematik.su.se>
Cc:  king@reasoning.com
In-reply-to:  Your message of Sun, 01 Oct 2000 20:32:48 PDT.

> I made a Metrowerks MacOS CW Pro 5 plugin out of the Flex 2.5.4.1. Here are
> some bugs:
> 
> -- flexerror() calls flexend() which becomes an infinite loop in the case
> of an error in flexend(). So, flexerror(), replace flexend() by exit().
> 
> -- If there is no output file error, flexend() fails to close the output
> file. This causes a problem as a plugin, because the file remains locked
> the next time the compiler runs, and the file cannot be reopened.
> 
> This bug also occurs if, one first has created the .c file, then runs it
> with a -S option for a skeleton filet Flex cannot find, in which case
> flexend() fails to close the output file.

Please note, Dick King (king@reasoning.com) has taken over flex maintenance
and development.  I've cc'd him.

		Vern

------- Message 9

Date:  Wed, 04 Oct 2000 23:38:07 PDT
From:  Vern Paxson <vern>
Subject:  Re: flex -+ option and ISO C++ IOStreams
To:  Harri Pasanen <harri.pasanen@trema.com>
Cc:  king@reasoning.com
In-reply-to:  Your message of Mon, 02 Oct 2000 10:15:59 PDT.

> You probably already know this: 
> 
> flex version 2.5.4 generates with -+ flag a forward declaration of class
> istream.
> Problem is that it is incompatible with the new ANSI IOStreams.  This
> becomes apparent if I try to use KAI C++ (http://www.kai.com/) to
> compile the generated scanner.
> 
> Lexing Scanner.l...
> Compiling Scanner.o...
> "/usr/local/KAI/KCC.pu-4.0b-1/KCC_BASE/include/istream.h", line 12:
> error #101: 
>           "istream" has already been declared in the current scope
>   using std::istream;
> 
> I wonder if an updated version of flex is in the works?  In the time
> being I can compile my own...

Dick King (king@reasoning.com) has recently taken over flex maintenance and
development (I've cc'd him).  Since there will be a fair amount of getting
up to speed for him, I think building your own version for now is the
best course.

		Vern

------- Message 10

Date:  Thu, 05 Oct 2000 00:27:15 PDT
From:  Vern Paxson <vern>
Subject:  Re: Flex Feature Idea 
To:  Davy Durham <david.durham@wcom.com>
Cc:  king@reasoning.com
In-reply-to:  Your message of Thu, 28 Sep 2000 11:28:52 EDT.

>     First I just want to say how much flex has been helpful to me in
> writing my parser, and that I really appreciate the free use of the
> program...

You're welcome!

Please note, though, that Dick King (cc'd) has taken over flex maintenance
and development.

>     I can't re-write YY_INPUT which could  detect and remove the '\' and
> \n  and go ahead and return the _next_ char, because I'm  keeping up
> with the line-number and column-number in YY_USER_ACTION for each
> token... Since flex buffer's the data retrieved from YY_INPUT, _my_
> YY_INPUT would need to increment the line-number, but it would happen
> possibly too soon...  Plus I'm using yy_switch_buffer to do macro
> replacements quite a bit which I think would complicate things even
> more....

Well, that's how you have to do it, I'm afraid.  I can't see any practical
way to add this as a feature to flex at a later stage in the input analysis,
since the semantics really are "pretend this slash and this newline don't
exist (mod line numbers)".  What you can probably do for line numbers is
the following: create a data structure that maps byte offsets in the current
input stream to line numbers.  YY_INPUT would fill this in as it returns
new (possibly munged) lines, and your actions would consult it in order to
generate correct location information for error messages.  (To deal with
multiple input buffers, you maintain one data structure per buffer.)

		Vern

------- Message 11

Date:  Thu, 05 Oct 2000 01:03:01 PDT
From:  Vern Paxson <vern>
Subject:  Re: [minor] obscure error messages if YY_DECL ends in semicolon 
To:  "Eric R. Buddington" <ebuddington@mail.wesleyan.edu>
Cc:  king@reasoning.com
In-reply-to:  Your message of Thu, 14 Sep 2000 13:39:52 EDT.

> Your address was listed in the flex info pages, so I send this to you; if
> you aren't the currnet maintainer, please let me know.

I'm not longer the maintainer - Dick King (cc'd) has taken over.

> First, I would like to add an '%option reentrant", such that the flex code
> would have the lexer function as its only global symbol. Would you
> consider such a patch for inclusion in the official flex?

Someone has already contributed patches for this - you can get them
from ftp://ftp.ee.lbl.gov/.vp-flex-reentrant.tar.gz .

> Secondly, an erroneous input that might give better error messages:
> 
> If YY_DECL ends with a semicolon, the generated error messages refer to
> wrong (nonexistant in this case) line numbers in the source file. The
> user error is only clear after turning off #line directives and looking at
> the generated .c file.
> 
> It would be good to have this error (a fairly easy one to make) reported
> clearly, if that can be done elegantly. This I what I get now:
> 
> flex -t -Pcap_string_ cap_string.flex > cap_string.c
> cc -Wall   -c -o cap_string.o cap_string.c
> cap_string.flex:163: parse error before `{'
> cap_string.flex:165: register name not specified for `yy_cp'
> cap_string.flex:165: register name not specified for `yy_bp'
> cap_string.flex:166: register name not specified for `yy_act'
> cap_string.flex:39: parse error before `if'
> ...

I don't offhand see an easy way to fix this, other than scanning the source
for a YY_DECL definition and looking for a semi-colon (urgh, what a hack).
Dick will have to toss it onto the to-do list.

		Vern

------- Message 12

Date:  Thu, 5 Oct 2000 10:51:53 +0200
From:  Hans Aberg <haberg@matematik.su.se>
Subject:  Re: Flex 2.5.4.1 Bugs
To:  Vern Paxson <vern@ee.lbl.gov>
Cc:  king@reasoning.com
X-Sender:  haberg@pop.matematik.su.se (Unverified)
In-Reply-To:  <200010050636.e956axT12828@daffy.ee.lbl.gov>
References:  Your message of Sun, 01 Oct 2000 20:32:48 PDT.
Mime-Version:  1.0
Content-Type:  text/plain; charset="us-ascii"
Status:  U

At 23.36 -0700 0-10-04, Vern Paxson wrote:
>Please note, Dick King (king@reasoning.com) has taken over flex maintenance
>and development.  I've cc'd him.

OK.

To: Dick King
Bison seems to under revision right now (I'm working hard with the Bison
development team just these days on pinpointing some bugs). -- Perhaps Flex
& Bison development should somehow be synchronized.

I have labelled Flex as BSD. Is that correct?

  Hans Aberg



------- Message 13

Date:  Fri, 06 Oct 2000 02:04:11 PDT
From:  Vern Paxson <vern>
Subject:  Re: C++, flex under cygwin and Visual C++
To:  "Simon J. Julier" <sjulier@erols.com>
Cc:  king@reasoning.com
In-reply-to:  Your message of Thu, 05 Oct 2000 07:15:17 PDT.

> Many apologies if either (a) it's a stupid question or (b) if I'm emailing 
> the wrong person.

It's (b) :-).  Dick King (king@reasoning.com) has taken over flex
maintenance and development.  I've cc'd him.

> I have been using flex (v. impressive) to write a parser 
> for a C++ program. The program is being compiled using a Windows 
> 2000/Windows NT machine using the cygwin tool set and the Microsoft Visual 
> C++ compiler. Because the cygwin distribution does not include FlexLexer.h, 
> I took the advice from the man page and simply renamed lex.yy.c as 
> lex.yy.cxx and compiled this as a C++ program. This almost works except for 
> the fact that, in C++ compile mode, there is a dependency on unistd.h due 
> to the following code snippet:
> 
> #ifdef __cplusplus
> 
> #include <stdlib.h>
> #include <unistd.h>
> 
> ... (other stuff) ...
> 
> Since the Visual C++ compiler does not come with the header file unistd.h, 
> I manually modified the code to:
> 
> #ifdef __cplusplus
> 
> #include <stdlib.h>
> #ifndef _WIN32
> #include <unistd.h>
> #endif /* ! _WIN32 */
> 
> Is this the "right way" to solve the problem?

I imagine it is.  This problem has come up a bunch of times, and,
much as I hate adding #ifdef's to the skeleton, I haven't thought of
any other way to do it.

> If so, would it be possible 
> to consider making the appropriate changes to flex.skl?

Sounds okay to me, but this is now Dick's call.

		Vern

------- Message 14

Date:  Fri, 6 Oct 2000 09:31:08 -0700 (PDT)
From:  Dick King <king@reasoning.com>
Subject:  Re: C++, flex under cygwin and Visual C++
To:  vern@ee.lbl.gov
Cc:  sjulier@erols.com
MIME-Version:  1.0
Content-Type:  text/plain; charset=us-ascii
Content-Transfer-Encoding:  7bit
In-Reply-To:  <200010060904.e9694BU17584@daffy.ee.lbl.gov>
References:  <200010060904.e9694BU17584@daffy.ee.lbl.gov>
X-Mailer:  VM 6.61 under 19.14 XEmacs Lucid



I'm getting my flex support operations organized, finally.

I've established a new eddress, flex@2xtreme.net .

I'll start looking at the small bug fixes flex has been asked for in a week or
so.  Then i will get to meaty changes.

One meaty change i've been asked for which i would like to do [i did it for my
current company's lexer product so i think i understand the issues] is a
character level prefilter infrastructure, that would allow you to integrate a
cpp-style preprocessor [rather than making it a separate pass as is the norm]
or to write a flex lexer for a language like FORTRAN or COBOL or that language
i've heard about whose name escapes me where indentation is syntactically
significant.

Vern Paxson writes:
 >> Many apologies if either (a) it's a stupid question or (b) if I'm emailing 
 >> the wrong person.
 >
 >It's (b) :-).  Dick King (king@reasoning.com) has taken over flex
 >maintenance and development.  I've cc'd him.
 >
 >> I have been using flex (v. impressive) to write a parser 
 >> for a C++ program. The program is being compiled using a Windows 
 >> 2000/Windows NT machine using the cygwin tool set and the Microsoft Visual 
 >> C++ compiler. Because the cygwin distribution does not include FlexLexer.h, 
 >> I took the advice from the man page and simply renamed lex.yy.c as 
 >> lex.yy.cxx and compiled this as a C++ program. This almost works except for 
 >> the fact that, in C++ compile mode, there is a dependency on unistd.h due 
 >> to the following code snippet:
 >> 
 >> #ifdef __cplusplus
 >> 
 >> #include <stdlib.h>
 >> #include <unistd.h>
 >> 
 >> ... (other stuff) ...
 >> 
 >> Since the Visual C++ compiler does not come with the header file unistd.h, 
 >> I manually modified the code to:
 >> 
 >> #ifdef __cplusplus
 >> 
 >> #include <stdlib.h>
 >> #ifndef _WIN32
 >> #include <unistd.h>
 >> #endif /* ! _WIN32 */
 >> 
 >> Is this the "right way" to solve the problem?
 >
 >I imagine it is.  This problem has come up a bunch of times, and,
 >much as I hate adding #ifdef's to the skeleton, I haven't thought of
 >any other way to do it.
 >
 >> If so, would it be possible 
 >> to consider making the appropriate changes to flex.skl?
 >
 >Sounds okay to me, but this is now Dick's call.
 >
 >		Vern
 >

------- Message 15

Date:  Sun, 08 Oct 2000 01:19:08 PDT
From:  Vern Paxson <vern>
Subject:  Re: C++, flex under cygwin and Visual C++
To:  Dick King <king@reasoning.com>
In-reply-to:  Your message of Fri, 06 Oct 2000 09:31:08 PDT.

> One meaty change i've been asked for which i would like to do [i did it for my
> current company's lexer product so i think i understand the issues] is a
> character level prefilter infrastructure, that would allow you to integrate a
> cpp-style preprocessor [rather than making it a separate pass as is the norm]
> or to write a flex lexer for a language like FORTRAN or COBOL or that language
> i've heard about whose name escapes me where indentation is syntactically
> significant.

That would be really nice to have!

		Vern

------- Message 16

Date:  Mon, 9 Oct 2000 12:27:13 -0400 (EDT)
From:  Andrew Droffner <adroffne@versus.dmz.advance.net>
Subject:  Re: Flex & Multiple Threads 
To:  Vern Paxson <vern@ee.lbl.gov>
Cc:  help-flex@gnu.org, king@reasoning.com
In-Reply-To:  <200010050752.e957qGq13639@daffy.ee.lbl.gov>
MIME-Version:  1.0
Content-Type:  TEXT/PLAIN; charset=US-ASCII

I'm unable to compile the scan.l LEX specification using the bootstrap
scanner, initscan.c. I have tried it on the latest flex source from GNU,
and the tarball from your page at ee.lbl.gov.

If I drop the "%option reentrant" from scan.l, then it all compiles. I
doubt it works right then. What else do I need here, a new initscan.c?

Thanks for the help. 

On Thu, 5 Oct 2000, Vern Paxson wrote:

> > I'm interested in building a multi-threaded server using YACC & Lex.
> > The server & client parse the data they pass to each other.
> > 
> > The Bison dialect of YACC uses a "%pure_parser" option to force re-entrant
> > code. I would like a Flex equivalent; is there one already?
> 
> There are patches to flex to support reentrant scanners, which I've
> put in ftp://ftp.ee.lbl.gov/.vp-flex-reentrant.tar.gz .
> 
> Also, please note, flex maintenance has been taken over by Dick King
> <king@reasoning.com>.
> 
> 		Vern
> 

- -- 
[ Andrew Droffner
[ Advance Publications Internet
[
[ adroffne@advance.net


------- Message 17

Date:  Tue, 10 Oct 2000 22:40:12 PDT
From:  Vern Paxson <vern>
Subject:  Re: Flex & Multiple Threads 
To:  Andrew Droffner <adroffne@versus.dmz.advance.net>
Cc:  help-flex@gnu.org, king@reasoning.com
In-reply-to:  Your message of Mon, 09 Oct 2000 12:27:13 PDT.

> I'm unable to compile the scan.l LEX specification using the bootstrap
> scanner, initscan.c. I have tried it on the latest flex source from GNU,
> and the tarball from your page at ee.lbl.gov.
> 
> If I drop the "%option reentrant" from scan.l, then it all compiles. I
> doubt it works right then. What else do I need here, a new initscan.c?

I don't have any firsthand experience with those patches, unfortunately,
so all I can offer is to try building without the %option reentrant;
then run flex on scan.l to build a new scan.c; recompile using that;
run "make check" to make sure it's working okay; then try putting the
%option reentrant back in.  Because I believe the modified scan.l doesn't
actually need %option reentrant, it's just there to exercise the scanning
during "make check".

		Vern

------- Message 18

Date:  Wed, 11 Oct 2000 11:42:35 PDT
From:  Vern Paxson <vern>
Subject:  Re: Flex & Multiple Threads 
To:  Dick King <king@reasoning.com>
In-reply-to:  Your message of Wed, 11 Oct 2000 11:29:49 PDT.

> How do i read help-flex@gnu.org?

Try sending mail to help-flex-request@gnu.org with a message body of
"subscribe".  Let me know if that doesn't work.

		Ven

------- End of Forwarded Messages


1,,
X-Coding-System: nil
Mail-from: From vern@daffy.ee.lbl.gov  Wed Mar 14 04:01:40 2001
Return-Path: <vern@daffy.ee.lbl.gov>
Received: from localhost (localhost [127.0.0.1])
	by michael.uncg.edu (8.9.3/8.9.3) with ESMTP id EAA09964
	for <wlestes@localhost>; Wed, 14 Mar 2001 04:01:40 -0500
Received: from imap.uncg.edu
	by localhost with IMAP (fetchmail-5.1.0)
	for wlestes@localhost (single-drop); Wed, 14 Mar 2001 04:01:40 -0500 (EST)
Received: from external-gw.uncg.edu (external-gw.uncg.edu [152.13.2.70])
	by hermes.email.uncg.edu (8.11.0/8.11.0) with ESMTP id f2E8pse19643
	for <wlestes@hermes.email.uncg.edu>; Wed, 14 Mar 2001 03:51:54 -0500 (EST)
Received: from daffy.ee.lbl.gov (daffy.ee.lbl.gov [131.243.1.31])
	by external-gw.uncg.edu (8.9.3/8.9.3) with ESMTP id DAA20408
	for <wlestes@uncg.edu>; Wed, 14 Mar 2001 03:51:53 -0500 (EST)
Received: (from vern@localhost)
	by daffy.ee.lbl.gov (8.10.0/8.10.0) id f2E8pqR07884;
	Wed, 14 Mar 2001 00:51:52 -0800 (PST)
Message-Id: <200103140851.f2E8pqR07884@daffy.ee.lbl.gov>
To: wlestes@uncg.edu
Subject: messages sent to Dick King (2 of 2)
Date: Wed, 14 Mar 2001 00:51:52 PST
From: Vern Paxson <vern@ee.lbl.gov>

*** EOOH ***
To: wlestes@uncg.edu
Subject: messages sent to Dick King (2 of 2)
Date: Wed, 14 Mar 2001 00:51:52 PST
From: Vern Paxson <vern@ee.lbl.gov>


------- Forwarded Messages

Date:  Sat, 07 Oct 2000 10:31:22 -0700
From:  "Dick King, flex maintainer" <flex@2xtreme.net>
Subject:  Re: the source tree
To:  vern@ee.lbl.gov
Cc:  flex@2xtreme.net
In-reply-to:  <14814.15467.441052.399457@fiddle.mtview.reasoning.com>
Reply-to:  flex@2xtreme.net
MIME-version:  1.0
X-Mailer:  Forte Agent 1.8/32.548
Content-type:  text/plain; charset=us-ascii
Content-transfer-encoding:  7bit
References:  <14814.15467.441052.399457@fiddle.mtview.reasoning.com>

On Fri, 6 Oct 2000 13:56:11 -0700 (PDT), you wrote:

> > I think i'm prepared to give this a try.
> 
> Cool!, thanks.  Here's a snapshot of the current source tree.  I'm about
> to leave for two week's vacation, not reading email, but upon my return
> I'll put together the stuff-to-do email archive, etc.
> 
> 		Vern
> 

What is the current release?

You sent me 2.5.4a to my office computer on June 20, but i found 2.5.5a on
June 20 and downloaded it to my home computer.

- -dk

PS: when anyone writes to you, tell them about my new eddress
[flex@2xtreme.net]?  Thanks.

- -dk


------- Message 2

Date:  Sun, 08 Oct 2000 01:26:26 PDT
From:  Vern Paxson <vern>
Subject:  Re: a sort of flex-related question :-)
To:  Alan Donovan <alan.donovan@arm.com>
Cc:  flex@2xtreme.net
In-reply-to:  Your message of Thu, 05 Oct 2000 18:49:51 PDT.

> Vern, sorry to write out of the blue like this but I figure that as the
> maintainer of flex, you could probably answer my question in a second.

I'm actually no longer the maintainer.  Dick King has taken over.  He's set
up the address flex@2xtreme.net for flex correspondence.  I've cc'd him.

> I'm writing a tool somewhere between flex and yacc, in which you specify
> a set of rules and actions (as in yacc) but the language is that of
> regular expressions not context-free grammars.
> 
> Each rule may include literals, operators, and simpler (i.e. previously
> declared) rules as components, and have an action.  e.g.
> 
> foo    :[0-9]+   { func1(); }
> 
> bar    :"x"foo   { func2(); }
>        |"y"foo   { func3(); }
> 
> So the NDFA for "bar" actually contains two copies of the NDFA for
> "foo".  My problem arises when trying to determine where in the DFA
> (generated by subset construction) to invoke the actions associated with
> each rule.  All my DFA states are annotated as to which rules (if any)
> they start and which rules (if any) they accept. However this is not
> enough to correctly invoke the actions.
> 
> Consider:
> 
>    foo: ab
>    bar: a
> 
> in the context of some larger expression zip:(foo|bar)"c".
> 
> The DFA for "zip" looks like:
> 
>     a      b      c
>  P ---> Q ---> R ---> T
>         |
>         +----> S
>            c
> 
> P starts rules "foo",  "bar" and "zip". Q accepts "bar" and R accepts
> rule "foo". S and T accept zip.
> 
> On input "abc", how do you know _not_ to invoke the action for rule
> "bar" when you reach state Q (or more correctly, when you receive the
> "b" and move to state R)?
> 
> So the question is: do you think what I am trying to do is possible?

What you are doing looks like a parse, and you can only parse LR(0)
languages with DFAs (if I recall correctly), i.e., those without
lookahead.  The above example requires lookahead to differentiate
matching foo from bar.  So I think you're out of luck here, if
I understand the example correctly.

		Vern

------- Message 3

Date:  Sun, 08 Oct 2000 07:47:01 -0700
From:  "Dick King, flex maintainer" <flex@2xtreme.net>
Subject:  Re: a sort of flex-related question :-)
To:  Vern Paxson <vern@ee.lbl.gov>
Cc:  Alan Donovan <alan.donovan@arm.com>
In-reply-to:  <200010080826.e988QQC23650@daffy.ee.lbl.gov>
Reply-to:  flex@2xtreme.net
MIME-version:  1.0
X-Mailer:  Forte Agent 1.8/32.548
Content-type:  text/plain; charset=us-ascii
Content-transfer-encoding:  7bit
References:  <200010080826.e988QQC23650@daffy.ee.lbl.gov>
Status:  U

On Sun, 08 Oct 2000 01:26:26 PDT, you wrote:

> > Vern, sorry to write out of the blue like this but I figure that as the
> > maintainer of flex, you could probably answer my question in a second.
> 
> I'm actually no longer the maintainer.  Dick King has taken over.  He's set
> up the address flex@2xtreme.net for flex correspondence.  I've cc'd him.
> 
> > I'm writing a tool somewhere between flex and yacc, in which you specify
> > a set of rules and actions (as in yacc) but the language is that of
> > regular expressions not context-free grammars.
> > 
> > Each rule may include literals, operators, and simpler (i.e. previously
> > declared) rules as components, and have an action.  e.g.
> > 
> > foo    :[0-9]+   { func1(); }
> > 
> > bar    :"x"foo   { func2(); }
> >        |"y"foo   { func3(); }
> > 
> > So the NDFA for "bar" actually contains two copies of the NDFA for
> > "foo".  My problem arises when trying to determine where in the DFA
> > (generated by subset construction) to invoke the actions associated with
> > each rule.  All my DFA states are annotated as to which rules (if any)
> > they start and which rules (if any) they accept. However this is not
> > enough to correctly invoke the actions.
> > 
> > Consider:
> > 
> >    foo: ab
> >    bar: a
> > 
> > in the context of some larger expression zip:(foo|bar)"c".
> > 
> > The DFA for "zip" looks like:
> > 
> >     a      b      c
> >  P ---> Q ---> R ---> T
> >         |
> >         +----> S
> >            c
> > 
> > P starts rules "foo",  "bar" and "zip". Q accepts "bar" and R accepts
> > rule "foo". S and T accept zip.
> > 
> > On input "abc", how do you know _not_ to invoke the action for rule
> > "bar" when you reach state Q (or more correctly, when you receive the
> > "b" and move to state R)?
> > 
> > So the question is: do you think what I am trying to do is possible?
> 
> What you are doing looks like a parse, and you can only parse LR(0)
> languages with DFAs (if I recall correctly), i.e., those without
> lookahead.  The above example requires lookahead to differentiate
> matching foo from bar.  So I think you're out of luck here, if
> I understand the example correctly.
> 
> 		Vern

Vern is mostly correct.  Small extensions can give you a single token
[character, here] lookahead but that is not enough to deliver the
functionality you want.  Consider

foo: a    { foo_fn(); }
bar: ab   { bar_fn(); }

mumble:    foo bcde.....yZ   { any }
         | bar  cde.....yW   { any }

- -dk


------- Message 4

Date:  Tue, 10 Oct 2000 02:36:57 PDT
From:  Vern Paxson <vern>
Subject:  Re: Bison's semantic parsers
To:  Akim Demaille <akim@epita.fr>
Cc:  Hans Aberg <haberg@matematik.su.se>,
        Davy Durham <david.durham@wcom.com>, Bison Help <help-bison@gnu.org>,
        flex@2xtreme.net
In-reply-to:  Your message of 10 Oct 2000 11:43:07 PDT.

> Nope, indeed Vern seems also to say he won't work on Flex again.

Unfortunately, yes.

> Vern, could we imagine that Flex be put on the CVS server of the FSF,
> subversions.gnu.org?

Thank you very much for the offer, I certainly appreciate it.  However,
Dick King has already volunteered to take over maitenance and development.
I've cc'd the address he's using for flex-related mail, flex@2xtreme.net.

		Vern

------- Message 5

Date:  Tue, 10 Oct 2000 13:30:02 +0200
From:  Hans Aberg <haberg@matematik.su.se>
Subject:  Re: Bison's semantic parsers
To:  Vern Paxson <vern@ee.lbl.gov>
Cc:  Akim Demaille <akim@epita.fr>, Davy Durham <david.durham@wcom.com>,
       Bison Help <help-bison@gnu.org>, flex@2xtreme.net
X-Sender:  haberg@pop.matematik.su.se
In-Reply-To:  <200010100936.e9A9avg01250@daffy.ee.lbl.gov>
References:  Your message of 10 Oct 2000 11:43:07 PDT.
Mime-Version:  1.0
Content-Type:  text/plain; charset="us-ascii"
Status:  U

At 02.36 -0700 0-10-10, Vern Paxson wrote:
>> Vern, could we imagine that Flex be put on the CVS server of the FSF,
>> subversions.gnu.org?
>
>Thank you very much for the offer, I certainly appreciate it.  However,
>Dick King has already volunteered to take over maitenance and development.
>I've cc'd the address he's using for flex-related mail, flex@2xtreme.net.

At 11.43 +0200 0-10-10, Akim Demaille wrote:
>Hans> Perhaps Bison and Flex should be bundled, or tuned up together.
>
>It's a good but bad idea.  For instance you could also argue that
>Autoconf, Automake and Libtool be merged, but if you just observe that
>they have totally different evolutions, it becomes infeasible.  While
>definitely a seducing idea, it wouldn't work IMHO.

My experimenting with C++ though suggests that Bison and Flex needs some
synchronizing if one should take full advantage of that language:

- -- One thing that come to my mind is that instead of letting the Bison
output .cc file include the Flex .cc file, they communicate the information
via header files.

- -- When experimenting C++ manipulator style I mentioned before
    Parser parser;
    parser_value val;
    ifstream ifs("...", ios_base::in | ios_base::binary);
    ofstream ofs("...");
    if (ifs) try { for (;;) {
        ifs >> parser >> val;
        cout << val << endl;
    }
I experience problems with how Flex resets its state, and how streams and
such are initialized.

For example, if I try to gobble up UNIX, MacOS & DOS line separators in Flex by
    \n|\r|\r\n		 { return '\n'; }
then, when Flex sees a \r it must first look ahead to verify the next
character isn't a \n. However, if a successful value "val" is produced,
then the next time  one ends up at the line ifs >> parser >> val, Flex is
reset, and as a consequence, the extra character needed to the \n check is
gobbled up.

One can get around this by not allowing this ifs >> parser >> val
construct, then the advantage of freer, and therefore easier user syntaxes
are lost.

- -- I think that there might be some other issues, such as how to properly
localize all data that Bison and Flex uses, and how they should communicate
that. These are more questions of hand-shaking than bundling them together
as a single program.

As for the comparisons of Bison and Flex versus other types of
compiler-compilers, I think the main disadvantage with them (the former) is
that they are extremely difficult to debug. It's rather hopeless trying to
debug the C/C++ files they produce as output. If one could hook up an
debugger though, so that one could step throw the .l/.y source files, just
as in the case of visual debugging of a C/C++ file, then this would change,
and it would not make any difference of what method is used, as long as it
is efficient and does the job.

  Hans Aberg



------- Message 6

Date:  10 Oct 2000 13:51:33 +0200
From:  Akim Demaille <akim@epita.fr>
Subject:  Re: Bison's semantic parsers
To:  Vern Paxson <vern@ee.lbl.gov>
Cc:  Hans Aberg <haberg@matematik.su.se>, Davy Durham <david.durham@wcom.com>,
       Bison Help <help-bison@gnu.org>, flex@2xtreme.net,
       Didier Verna <didier@epita.fr>
References:  <200010100936.e9A9avg01250@daffy.ee.lbl.gov>
In-Reply-To:  Vern Paxson's message of "Tue, 10 Oct 2000 02:36:57 PDT"
Lines:  27
User-Agent:  Gnus/5.0807 (Gnus v5.8.7) XEmacs/21.1 (Channel Islands)
MIME-Version:  1.0
Content-Type:  text/plain; charset=us-ascii
Status:  U

>>>>> "Vern" == Vern Paxson <vern@ee.lbl.gov> writes:

>> Nope, indeed Vern seems also to say he won't work on Flex again.

Vern> Unfortunately, yes.

Just by curiosity: you're no longer interested, or you don't have time
enough?

>> Vern, could we imagine that Flex be put on the CVS server of the
>> FSF, subversions.gnu.org?

Vern> Thank you very much for the offer, I certainly appreciate it.

I'm also quite pleased to see my offer pleased you so much :)

Vern> However, Dick King has already volunteered to take over
Vern> maitenance and development.  I've cc'd the address he's using
Vern> for flex-related mail, flex@2xtreme.net.

Thanks!

Then, I shall restate my question for flex@2xtreme: is there a CVS
server or something which would make it possible to participate the
development of Flex?

        Akim

------- Message 7

Date:  Tue, 10 Oct 2000 14:15:53 PDT
From:  Vern Paxson <vern>
Subject:  Re: Bison's semantic parsers
To:  Akim Demaille <akim@epita.fr>
Cc:  Hans Aberg <haberg@matematik.su.se>,
        Davy Durham <david.durham@wcom.com>, Bison Help <help-bison@gnu.org>,
        flex@2xtreme.net, Didier Verna <didier@epita.fr>
In-reply-to:  Your message of 10 Oct 2000 13:51:33 PDT.

> Just by curiosity: you're no longer interested, or you don't have time
> enough?

No time ...

		Vern

------- Message 8

Date:  Thu, 12 Oct 2000 20:07:57 PDT
From:  Vern Paxson <vern>
Subject:  Re: Flex 2.5.4 Bug 
To:  Cliff Sarginson <csarginson@descartes.com>
Cc:  flex@2xtreme.net
In-reply-to:  Your message of Thu, 12 Oct 2000 13:52:06 +0200.

> I dont know if you still receive bug reports about flex ..

I'm not, Dick King (cc'd) has taken over maintenance & development.

> but here is one.. !
> 
> In a scanner we have ..
> 
> %start this that msg another
> 
> The field "msg" gets turned into a
> 
> #define msg 3
> 
> We do not define a  xxx_YY_FATAL_ERROR
> 
> So flex supplies one, as a macro, with a definition:
> 
> xxx_yy_fatal_error (const msg [])
> 
> The "C" pre-processor merrily turns this into
> 
> xx_yy_fatal_error ( const 3  [] )
> 
> Are we really the first to come across this ? Flex should maybe
> be a bit more creative with it's paramater naming for macro's
> or do some other name space checking maybe ?

Yep, this is a known problem.  The difficulty is that when lex was
originally designed, it allowed you to switch start states using
"BEGIN msg" rather than "BEGIN(msg)".  This means that either you
do a major analysis of the source to only patch up true BEGIN's
(rather than, say, one that occurs inside a string literal); or you
use a gross hack in defining BEGIN, like flex does:

	#define BEGIN yy_start = 1 + 2 *

With this sort of definition, there's no way to modify the namespace
consumed by start states.  If instead you could require the ()'s, then
you could use

	#define BEGIN(state) yy_start = 1 + 2 * (YY_STATE_ ## state)

and all would be happy ...

		Vern

------- Message 9

Date:  Thu, 12 Oct 2000 20:08:03 PDT
From:  Vern Paxson <vern>
Subject:  Re: a sort of flex-related question :-) 
To:  Alan.Donovan@arm.com
Cc:  flex@2xtreme.net
In-reply-to:  Your message of Thu, 12 Oct 2000 15:36:38 BST.

> Thank you both for your help.  I can probably rearrange my grammars to be
> LR(0). Can you give me a reference for the algorithm for LR(0) parsing with
> a DFA?

Dunno off hand, but I imagine it must be in the Dragon Book.

		Vern

------- Message 10

Date:  Thu, 02 Nov 2000 11:56:35 PST
From:  Vern Paxson <vern>
Subject:  Re: flex documentation enhancement request
To:  Tracy Camp <campt@thalvors.miralink.com>
Cc:  flex@2xtreme.net
In-reply-to:  Your message of Thu, 02 Nov 2000 10:17:17 PST.

Yes, a lot of people get bitten by this, unfortunately.

Note, Dick King has taken over flex maintenance and development, and has
set up the address flex@2xtreme.net for flex correspondence.  I've cc'd him.

		Vern


> As a new user to flex I found the following situation quite confusing and
> think a breif caviate note in the manual would have helped out (or at
> least an emphisis if there was one) :)
> 
> when using external variables when patching a pattern such as:
> 
> %x buffer_dev
> %{
> extern struct mystruct astruct; /* struct containing member buffer_dev */
> %}
> %%
> buffer_dev=	BEGIN(buffer_dev);
> 
> <buffer_dev>.*$ {
> 	strncpy(astruct.buffer_dev,yytext,255);
> 	}
> %%
> 
> this causes all sorts of problems because 'buffer_dev' is defined by flex
> at some point and I used buffer_dev as a member of an external structure
> that I'm trying to fill in with flex.  This causes all sorts of compile
> problems.  Renaming my 'buffer_dev' condition to something else works
> around this, but it was not immediately obvious what the problem was from
> the compile error.  For all else I found your manual to be of GREAT help -
> thanks!
> 
> t.
> 
> Tracy Camp
> Product Development
> Miralink Corp.PDX
> Portland OR
> 503-223-3140
> 

------- Message 11

Date:  Thu, 09 Nov 2000 23:03:38 PST
From:  Vern Paxson <vern>
Subject:  Re: Standard C++ compatibility for flex++ 
To:  Volker Simonis <simonis@informatik.uni-tuebingen.de>
Cc:  help-flex@gnu.org, flex@2xtreme.net
In-reply-to:  Your message of Wed, 08 Nov 2000 15:50:36 +0100.

> Is there any chance for a new release of flex? Is there a maintainer right now?
> Is it possible to submit some bug-fixes or to contribute in some way?

Dick King has taken over flex development and maintenance.  You can
reach him at flex@2xtreme.net (cc'd).

		Vern

------- Message 12

Date:  Wed, 14 Feb 2001 12:57:37 PST
From:  Vern Paxson <vern>
Subject:  Re: Bug#84780: Fix for info documentation (fwd)
To:  Santiago Vila <sanvila@unex.es>
Cc:  Gordon Sadler <gbsadler1@lcisp.com>, flex@2xtreme.net
In-reply-to:  Your message of Wed, 14 Feb 2001 15:40:30 PST.

Please note, Dick King has taken over flex maintenance and development,
and has set up the address flex@2xtreme.net for flex correspondence.

		Vern


> Hello.
> 
> I received this from the Debian Bug System.
> 
> ---------- Forwarded message ----------
> Date: Sun, 4 Feb 2001 02:41:30 -0600
> From: Gordon Sadler <gbsadler1@lcisp.com>
> To: Debian Bug Tracking System <submit@bugs.debian.org>
> Subject: #84780: Fix for info documentation
> 
> Package: flex
> Version: 2.5.4a-9
> Severity: wishlist
> 
> Attached find a diff to MISC/texinfo/flex.texi
> I noticed during the texi2html run it complains of first section not
> being toplevel. I opened the texi file and changed the toplevel
> sections to chapters. While I was there noticed the comment about
> fixing all of the See... to make them @xref, so I did.
> 
> Looks like upstream is pretty much done ... but you seem to be in
> touch, last changelog mentions manpage patch from author?
> 
> If you use this texi to regen the info the x-refs make it a little more
> useable. Maybe I'll add the index in next -)
> 
> Gordon Sadler
> ------------------------------------------------------------------------
> This is the patch in unified format:
> 
> diff -r -u flex-2.5.4.orig/MISC/texinfo/flex.texi flex-2.5.4/MISC/texinfo/flex.texi
> --- flex-2.5.4.orig/MISC/texinfo/flex.texi	Sun Jul 27 04:47:21 1997
> +++ flex-2.5.4/MISC/texinfo/flex.texi	Wed Feb 14 15:34:49 2001
> @@ -151,12 +151,12 @@
>  @end ifinfo
> 
>  @node Name, Synopsis, Top, Top
> -@section Name
> +@chapter Name
> 
>  flex - fast lexical analyzer generator
> 
>  @node Synopsis, Overview, Name, Top
> -@section Synopsis
> +@chapter Synopsis
> 
>  @example
>  flex [-bcdfhilnpstvwBFILTV78+? -C[aefFmr] -ooutput -Pprefix -Sskeleton]
> @@ -164,7 +164,7 @@
>  @end example
> 
>  @node Overview, Description, Synopsis, Top
> -@section Overview
> +@chapter Overview
> 
>  This manual describes @code{flex}, a tool for generating programs
>  that perform pattern-matching on text.  The manual
> @@ -244,7 +244,7 @@
>  @end table
> 
>  @node Description, Examples, Overview, Top
> -@section Description
> +@chapter Description
> 
>  @code{flex} is a tool for generating @dfn{scanners}: programs which
>  recognized lexical patterns in text.  @code{flex} reads the given
> @@ -260,7 +260,7 @@
>  corresponding C code.
> 
>  @node Examples, Format, Description, Top
> -@section Some simple examples
> +@chapter Some simple examples
> 
>  First some simple examples to get the flavor of how one
>  uses @code{flex}.  The following @code{flex} input specifies a scanner
> @@ -371,7 +371,7 @@
>  following sections.
> 
>  @node Format, Patterns, Examples, Top
> -@section Format of the input file
> +@chapter Format of the input file
> 
>  The @code{flex} input file consists of three sections, separated
>  by a line with just @samp{%%} in it:
> @@ -439,7 +439,7 @@
>  where the pattern must be unindented and the action must
>  begin on the same line.
> 
> -See below for a further description of patterns and
> +@xref{Patterns}, for a further description of patterns and
>  actions.
> 
>  Finally, the user code section is simply copied to
> @@ -468,7 +468,7 @@
>  is also copied verbatim to the output up to the next "*/".
> 
>  @node Patterns, Matching, Format, Top
> -@section Patterns
> +@chapter Patterns
> 
>  The patterns in the input are written using an extended
>  set of regular expressions.  These are:
> @@ -697,7 +697,7 @@
>  @end itemize
> 
>  @node Matching, Actions, Patterns, Top
> -@section How the input is matched
> +@chapter How the input is matched
> 
>  When the generated scanner is run, it analyzes its input
>  looking for strings which match any of its patterns.  If
> @@ -773,7 +773,7 @@
>  classes (the @code{c++} option; see below).
> 
>  @node Actions, Generated scanner, Matching, Top
> -@section Actions
> +@chapter Actions
> 
>  Each pattern in a rule has a corresponding action, which
>  can be any arbitrary C statement.  The pattern ends at the
> @@ -1035,7 +1035,7 @@
>  @end itemize
> 
>  @node Generated scanner, Start conditions, Actions, Top
> -@section The generated scanner
> +@chapter The generated scanner
> 
>  The output of @code{flex} is the file @file{lex.yy.c}, which contains
>  the scanning routine @samp{yylex()}, a number of tables used by
> @@ -1136,15 +1136,15 @@
> 
>  Three routines are available for scanning from in-memory
>  buffers rather than files: @samp{yy_scan_string()},
> -@samp{yy_scan_bytes()}, and @samp{yy_scan_buffer()}.  See the discussion
> -of them below in the section Multiple Input Buffers.
> +@samp{yy_scan_bytes()}, and @samp{yy_scan_buffer()}.
> +@xref{Multiple buffers, ,Multiple Input Buffers}.
> 
>  The scanner writes its @samp{ECHO} output to the @code{yyout} global
>  (default, stdout), which may be redefined by the user
>  simply by assigning it to some other @code{FILE} pointer.
> 
>  @node Start conditions, Multiple buffers, Generated scanner, Top
> -@section Start conditions
> +@chapter Start conditions
> 
>  @code{flex} provides a mechanism for conditionally activating
>  rules.  Any rule whose pattern is prefixed with "<sc>"
> @@ -1500,7 +1500,7 @@
>  @samp{%option stack} directive (see Options below).
> 
>  @node Multiple buffers, End-of-file rules, Start conditions, Top
> -@section Multiple input buffers
> +@chapter Multiple input buffers
> 
>  Some scanners (such as those which support "include"
>  files) require reading from several input streams.  As
> @@ -1675,7 +1675,7 @@
>  @end table
> 
>  @node End-of-file rules, Miscellaneous, Multiple buffers, Top
> -@section End-of-file rules
> +@chapter End-of-file rules
> 
>  The special rule "<<EOF>>" indicates actions which are to
>  be taken when an end-of-file is encountered and yywrap()
> @@ -1735,7 +1735,7 @@
>  @end example
> 
>  @node Miscellaneous, User variables, End-of-file rules, Top
> -@section Miscellaneous macros
> +@chapter Miscellaneous macros
> 
>  The macro @code{YY_USER_ACTION} can be defined to provide an
>  action which is always executed prior to the matched
> @@ -1800,7 +1800,7 @@
>  the @code{YY_BREAK} is inaccessible.
> 
>  @node User variables, YACC interface, Miscellaneous, Top
> -@section Values available to the user
> +@chapter Values available to the user
> 
>  This section summarizes the various values available to
>  the user in the rule actions.
> @@ -1865,7 +1865,7 @@
>  @end itemize
> 
>  @node YACC interface, Options, User variables, Top
> -@section Interfacing with @code{yacc}
> +@chapter Interfacing with @code{yacc}
> 
>  One of the main uses of @code{flex} is as a companion to the @code{yacc}
>  parser-generator.  @code{yacc} parsers expect to call a routine
> @@ -1890,7 +1890,7 @@
>  @end example
> 
>  @node Options, Performance, YACC interface, Top
> -@section Options
> +@chapter Options
>  @code{flex} has the following options:
> 
>  @table @samp
> @@ -1903,8 +1903,8 @@
>  and @samp{-Cf} or @samp{-CF} is used, the generated scanner will
>  run faster (see the @samp{-p} flag).  Only users who wish
>  to squeeze every last cycle out of their scanners
> -need worry about this option.  (See the section on
> -Performance Considerations below.)
> +need worry about this option.  (@pxref{Performance, ,
> +Performance Considerations})
> 
>  @item -c
>  is a do-nothing, deprecated option included for
> @@ -2138,8 +2138,7 @@
> 
>  @item -+
>  specifies that you want flex to generate a C++
> -scanner class.  See the section on Generating C++
> -Scanners below for details.
> +scanner class.  @xref{C++, ,Generating C++ Scanners}.
> 
>  @item -C[aefFmr]
>  controls the degree of table compression and, more
> @@ -2419,8 +2418,7 @@
>  function @samp{foo::yylex()} instead of @samp{yyFlexLexer::yylex()}.
>  It also generates a @samp{yyFlexLexer::yylex()} member function that
>  emits a run-time error (by invoking @samp{yyFlexLexer::LexerError()})
> -if called.  See Generating C++ Scanners, below, for additional
> -information.
> +if called.  @xref{C++, ,Generating C++ Scanners}.
> 
>  A number of options are available for lint purists who
>  want to suppress the appearance of unneeded routines in
> @@ -2439,7 +2437,7 @@
>  unless you use @samp{%option stack}).
> 
>  @node Performance, C++, Options, Top
> -@section Performance considerations
> +@chapter Performance considerations
> 
>  The main design goal of @code{flex} is that it generate
>  high-performance scanners.  It has been optimized for dealing
> @@ -2756,7 +2754,7 @@
>  the two is at about 8K characters/token.
> 
>  @node C++, Incompatibilities, Performance, Top
> -@section Generating C++ scanners
> +@chapter Generating C++ scanners
> 
>  @code{flex} provides two different ways to generate scanners for
>  use with C++.  The first way is to simply compile a
> @@ -2975,7 +2973,7 @@
>  releases.
> 
>  @node Incompatibilities, Diagnostics, C++, Top
> -@section Incompatibilities with @code{lex} and POSIX
> +@chapter Incompatibilities with @code{lex} and POSIX
> 
>  @code{flex} is a rewrite of the AT&T Unix @code{lex} tool (the two
>  implementations do not share any code, though), with some
> @@ -3051,8 +3049,8 @@
> 
>  Also note that flex C++ scanner classes @emph{are}
>  reentrant, so if using C++ is an option for you, you
> -should use them instead.  See "Generating C++
> -Scanners" above for details.
> +should use them instead.  @xref{C++, ,Generating C++
> +Scanners}.
> 
>  @item
>  @samp{output()} is not supported.  Output from the @samp{ECHO}
> @@ -3198,7 +3196,7 @@
>  line.
> 
>  @node Diagnostics, Files, Incompatibilities, Top
> -@section Diagnostics
> +@chapter Diagnostics
> 
>  @table @samp
>  @item warning, rule cannot be matched
> @@ -3251,8 +3249,7 @@
>  scanner specification includes recognizing the 8-bit
>  character @var{x} and you did not specify the -8 flag, and your
>  scanner defaulted to 7-bit because you used the @samp{-Cf} or @samp{-CF}
> -table compression options.  See the discussion of the @samp{-7}
> -flag for details.
> +table compression options. @xref{Options, ,@samp{-7}flag}.
> 
>  @item flex scanner push-back overflow
>  you used @samp{unput()} to push
> @@ -3285,7 +3282,7 @@
>  @end table
> 
>  @node Files, Deficiencies, Diagnostics, Top
> -@section Files
> +@chapter Files
> 
>  @table @file
>  @item -lfl
> @@ -3311,7 +3308,7 @@
>  @end table
> 
>  @node Deficiencies, See also, Files, Top
> -@section Deficiencies / Bugs
> +@chapter Deficiencies / Bugs
> 
>  Some trailing context patterns cannot be properly matched
>  and generate warning messages ("dangerous trailing
> @@ -3365,7 +3362,7 @@
>  The @code{flex} internal algorithms need documentation.
> 
>  @node See also, Author, Deficiencies, Top
> -@section See also
> +@chapter See also
> 
>  @code{lex}(1), @code{yacc}(1), @code{sed}(1), @code{awk}(1).
> 
> @@ -3380,7 +3377,7 @@
>  (deterministic finite automata).
> 
>  @node Author,  , See also, Top
> -@section Author
> +@chapter Author
> 
>  Vern Paxson, with the help of many ideas and much inspiration from
>  Van Jacobson.  Original version by Jef Poskanzer.  The fast table
> 

------- End of Forwarded Messages


1, answered,,
X-Coding-System: nil
Mail-from: From vern@daffy.ee.lbl.gov  Mon Mar 26 18:48:05 2001
Return-Path: <vern@daffy.ee.lbl.gov>
Received: from localhost (localhost [127.0.0.1])
	by michael.uncg.edu (8.9.3/8.9.3) with ESMTP id SAA00754
	for <wlestes@localhost>; Mon, 26 Mar 2001 18:48:05 -0500
Received: from imap.uncg.edu
	by localhost with IMAP (fetchmail-5.1.0)
	for wlestes@localhost (single-drop); Mon, 26 Mar 2001 18:48:05 -0500 (EST)
Received: from external-gw.uncg.edu (external-gw.uncg.edu [152.13.2.70])
	by hermes.email.uncg.edu (8.11.0/8.11.0) with ESMTP id f2QNVUe03862
	for <wlestes@hermes.email.uncg.edu>; Mon, 26 Mar 2001 18:31:30 -0500 (EST)
Received: from daffy.ee.lbl.gov (daffy.ee.lbl.gov [131.243.1.31])
	by external-gw.uncg.edu (8.9.3/8.9.3) with ESMTP id SAA18856
	for <wlestes@uncg.edu>; Mon, 26 Mar 2001 18:31:28 -0500 (EST)
Received: (from vern@localhost)
	by daffy.ee.lbl.gov (8.10.0/8.10.0) id f2QNVPc23647;
	Mon, 26 Mar 2001 15:31:25 -0800 (PST)
Message-Id: <200103262331.f2QNVPc23647@daffy.ee.lbl.gov>
To: wlestes@uncg.edu
Subject: Fwd: release 2.5 of flex...
Date: Mon, 26 Mar 2001 15:31:25 PST
From: Vern Paxson <vern@ee.lbl.gov>

*** EOOH ***
To: wlestes@uncg.edu
Subject: Fwd: release 2.5 of flex...
Date: Mon, 26 Mar 2001 15:31:25 PST
From: Vern Paxson <vern@ee.lbl.gov>

(Any luck finding some flex cycles?)

------- Forwarded Message

Date:  Mon, 26 Mar 2001 00:06:31 +0200
From:  Kai Hennig <khennig@pobox.com>
Subject:  release 2.5 of flex...
To:  vern@ee.lbl.gov
X-Spam-Filter:  check_local@studserv.stud.uni-hannover.de by digitalanswers.org
Reply-To:  khennig@pobox.com
Organization:  TRI Systems
X-Mailer:  Mozilla 4.76 (Macintosh; U; PPC)
X-Accept-Language:  en
MIME-Version:  1.0
Content-Type:  text/plain; charset=iso-8859-1
Content-Transfer-Encoding:  8bit

> Extract from file 'INSTALL' from flex-2.5.4
>
> ...
> If you need to do unusual things to compile the package, we encourage
> you to figure out how `configure' could check whether to do them, and
> mail diffs or instructions to the address given in the README so we
> can include them in the next release.
> ...

Hi Vern,

I'm using your flex package 2.5.4 on a Macintosh with Apple's Macintosh
Programmers Workshop 3.5 (MPW). There is a description comming along
with your package for the THINK C version 6.0 from Scott Hofmann
(23-JUL-94) which is not helpfull for using the package with MPW. I
enclosed a brief instruction for using the flex package with MPW which
might be more helpfull since Think C is no longer officially available
(as far as I know).

greetings,
Kai




Notes on the MPW 3.5 version of Flex 2.5.4
Kai Hennig 25-Mar-2001
Internet: khennig@pobox.com

To compile and run Flex 2.5.4 as an MPW Tool no changes to any file have
to be done. Instead add a file called 'config.h' to the files listed in
the command lines given below with the content listed below the
commands. To compile and link all necessary files execute the following
commands in the MPW shell:

MrC -noMapCR -alloca -ansi strict main.c
MrC -noMapCR -alloca -ansi strict ccl.c
MrC -noMapCR -alloca -ansi strict dfa.c
MrC -noMapCR -alloca -ansi strict ecs.c
MrC -noMapCR -alloca -ansi strict gen.c
MrC -noMapCR -alloca -ansi strict misc.c
MrC -noMapCR -alloca -ansi strict nfa.c
MrC -noMapCR -alloca -ansi strict -w off parse.c
MrC -noMapCR -alloca -ansi strict initscan.c
MrC -noMapCR -alloca -ansi strict tblcmp.c
MrC -noMapCR -alloca -ansi strict sym.c
MrC -noMapCR -alloca -ansi strict skel.c
MrC -noMapCR -alloca -ansi strict yylex.c
PPCLink -t 'MPST' -c 'MPS ' -o flex *
main.c.o *
ccl.c.o *
dfa.c.o *
ecs.c.o *
gen.c.o *
misc.c.o *
nfa.c.o *
parse.c.o *
initscan.c.o *
tblcmp.c.o *
sym.c.o *
skel.c.o *
yylex.c.o *
"{SharedLibraries}InterfaceLib" *
"{SharedLibraries}StdCLib" *
"{SharedLibraries}MathLib" *
"{PPCLibraries}StdCRuntime.o" *
"{PPCLibraries}PPCCRuntime.o" *
"{PPCLibraries}MrCIOStreams.o" *
"{PPCLibraries}PPCToolLibs.o"


/* File 'config.h' */
/* Define to `unsigned' if <sys/types.h> doesn't define.  */
#undef size_t

/* Define if you have the ANSI C header files.  */
#define STDC_HEADERS 1

/* Define if you have the <string.h> header file.  */
#define HAVE_STRING_H 1

/* Define if you have <alloca.h> and it should be used (not on 
Ultrix).  */
#define HAVE_ALLOCA_H 1

/* Define if you use FAT file system, leave undefined for NTFS */
#undef SHORT_FILE_NAMES 

/* Never do interactive input */
#define YY_NEVER_INTERACTIVE 1
/* EOF File 'config.h' */

------- End of Forwarded Message


1,,
X-Coding-System: nil
Mail-from: From vern@daffy.ee.lbl.gov  Wed Mar 28 04:07:49 2001
Return-Path: <vern@daffy.ee.lbl.gov>
Received: from localhost (localhost [127.0.0.1])
	by michael.uncg.edu (8.9.3/8.9.3) with ESMTP id EAA02765
	for <wlestes@localhost>; Wed, 28 Mar 2001 04:07:49 -0500
Received: from imap.uncg.edu
	by localhost with IMAP (fetchmail-5.1.0)
	for wlestes@localhost (single-drop); Wed, 28 Mar 2001 04:07:49 -0500 (EST)
Received: from external-gw.uncg.edu (external-gw.uncg.edu [152.13.2.70])
	by hermes.email.uncg.edu (8.11.0/8.11.0) with ESMTP id f2S8o2e15934
	for <wlestes@hermes.email.uncg.edu>; Wed, 28 Mar 2001 03:50:02 -0500 (EST)
Received: from daffy.ee.lbl.gov (daffy.ee.lbl.gov [131.243.1.31])
	by external-gw.uncg.edu (8.9.3/8.9.3) with ESMTP id DAA23213
	for <wlestes@uncg.edu>; Wed, 28 Mar 2001 03:50:00 -0500 (EST)
Received: (from vern@localhost)
	by daffy.ee.lbl.gov (8.10.0/8.10.0) id f2S8nr529885;
	Wed, 28 Mar 2001 00:49:53 -0800 (PST)
Message-Id: <200103280849.f2S8nr529885@daffy.ee.lbl.gov>
To: "John W. Millaway" <johnmillaway@yahoo.com>
Cc: help-flex@gnu.org, wlestes@uncg.edu
Subject: Re: reentrant C scanner
In-reply-to: Your message of Tue, 27 Mar 2001 12:49:46 PST.
Date: Wed, 28 Mar 2001 00:49:53 PST
From: Vern Paxson <vern@ee.lbl.gov>

*** EOOH ***
To: "John W. Millaway" <johnmillaway@yahoo.com>
Cc: help-flex@gnu.org, wlestes@uncg.edu
Subject: Re: reentrant C scanner
In-reply-to: Your message of Tue, 27 Mar 2001 12:49:46 PST.
Date: Wed, 28 Mar 2001 00:49:53 PST
From: Vern Paxson <vern@ee.lbl.gov>

> I contacted you (or someone there at flex H.Q.) about modifying flex to generate a
> reentrant scanner. I found some time recently, and the modifications are nearly
> complete. It wasn't as bad as I had originally thought it would be.
> 
> I have two questions for you:
> 
> 1. What is the procedure for regression testing?
> 2. What is the procedure for merging my modifications with the current
> distribution?

Both of these are not well defined at the moment.  Will Estes (cc'd) has
volunteered to give taking over flex maintenance/development a try, and
we're right now in a preliminary period during which he's gauging whether
he'll indeed have the cycles for doing so.

		Vern


1,,
X-Coding-System: nil
Mail-from: From vern@daffy.ee.lbl.gov  Thu Mar 29 18:05:46 2001
Return-Path: <vern@daffy.ee.lbl.gov>
Received: from localhost (localhost [127.0.0.1])
	by michael.uncg.edu (8.9.3/8.9.3) with ESMTP id SAA02631
	for <wlestes@localhost>; Thu, 29 Mar 2001 18:05:46 -0500
Received: from imap.uncg.edu
	by localhost with IMAP (fetchmail-5.1.0)
	for wlestes@localhost (single-drop); Thu, 29 Mar 2001 18:05:46 -0500 (EST)
Received: from external-gw.uncg.edu (external-gw.uncg.edu [152.13.2.70])
	by hermes.email.uncg.edu (8.11.0/8.11.0) with ESMTP id f2TMmHe28250
	for <wlestes@hermes.email.uncg.edu>; Thu, 29 Mar 2001 17:48:17 -0500 (EST)
Received: from daffy.ee.lbl.gov (daffy.ee.lbl.gov [131.243.1.31])
	by external-gw.uncg.edu (8.9.3/8.9.3) with ESMTP id RAA02561
	for <wlestes@uncg.edu>; Thu, 29 Mar 2001 17:48:16 -0500 (EST)
Received: (from vern@localhost)
	by daffy.ee.lbl.gov (8.10.0/8.10.0) id f2TMmCD10050;
	Thu, 29 Mar 2001 14:48:12 -0800 (PST)
Message-Id: <200103292248.f2TMmCD10050@daffy.ee.lbl.gov>
To: John Tupper <john_tupper@tenornetworks.com>
Subject: Re: flex bug
Cc: wlestes@uncg.edu
In-reply-to: Your message of Thu, 29 Mar 2001 13:52:38 PST.
Date: Thu, 29 Mar 2001 14:48:12 PST
From: Vern Paxson <vern@ee.lbl.gov>

*** EOOH ***
To: John Tupper <john_tupper@tenornetworks.com>
Subject: Re: flex bug
Cc: wlestes@uncg.edu
In-reply-to: Your message of Thu, 29 Mar 2001 13:52:38 PST.
Date: Thu, 29 Mar 2001 14:48:12 PST
From: Vern Paxson <vern@ee.lbl.gov>

Argh, yes, you're right.  This is (mis)feature interaction between
%option yylineno and trailing context.

I've cc'd Will Estes, who is trying out taking over flex maintenance /
development.  Will, I think the place to fix this is in finish_rule() in
nfa.c, making it generate extra code to loop through the characters being
put back and decrement yylineno accordingly.

		Vern


> 
> --------------DC4CBAA5BA86BA0906373DD6
> Content-Type: text/plain; charset=us-ascii
> Content-Transfer-Encoding: 7bit
> 
> Vern,
>   Are you the right person to send this to?  It's a groaner. . .
> 
> If you define a rule with a trailing context and the trailing context
> matches a newline, the newline gets counted twice (in yylineno), once in
> the trailing context and once when its matched for real.
> 
> Sample grammer:
> 
> \/\*               { begin(Comment); }
> <Comment>\*/[^/]   { ; /* this rule is problematic */ }
> <Comment>[^\*]+    { ; }
> <Comment>\*\/      { begin(0); }
> 
> The above grammer strips out C style comments.  If a comment contains a
> star at the end of a line, yylineno gets incremented twice.
> 
> We used the obvious work around by adding a seperate rule to explicitly
> match *\n without any trailing context, so we're not waiting for a fix.
> 
> Oh yeah, we're using version 2.5.4.
> 
> Thanks,
> John Tupper
> (john_tupper@tenornetworks.com)
> 
> 
> --------------DC4CBAA5BA86BA0906373DD6
> Content-Type: text/html; charset=us-ascii
> Content-Transfer-Encoding: 7bit
> 
> <!doctype html public "-//w3c//dtd html 4.0 transitional//en">
> <html>
> Vern,
> <br>&nbsp; Are you the right person to send this to?&nbsp; It's a groaner.
> . .
> <p>If you define a rule with a trailing context and the trailing context
> matches a newline, the newline gets counted twice (in yylineno), once in
> the trailing context and once when its matched for real.
> <p>Sample grammer:
> <p><tt>\/\*&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
> { begin(Comment); }</tt>
> <br><tt>&lt;Comment>\*/[^/]&nbsp;&nbsp; { ; /* this rule is problematic
> */ }</tt>
> <br><tt>&lt;Comment>[^\*]+&nbsp;&nbsp;&nbsp; { ; }</tt>
> <br><tt>&lt;Comment>\*\/&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; { begin(0); }</tt>
> <p>The above grammer strips out C style comments.&nbsp; If a comment contains
> a star at the end of a line, yylineno gets incremented twice.
> <p>We used the obvious work around by adding a seperate rule to explicitly
> match *\n without any trailing context, so we're not waiting for a fix.
> <p>Oh yeah, we're using version 2.5.4.
> <p>Thanks,
> <br>John Tupper
> <br>(john_tupper@tenornetworks.com)
> <br>&nbsp;</html>
> 
> --------------DC4CBAA5BA86BA0906373DD6--
> 


1, answered,,
X-Coding-System: nil
Mail-from: From johnmillaway@yahoo.com  Sun Apr  8 18:10:46 2001
Return-Path: <johnmillaway@yahoo.com>
Received: from localhost (localhost [127.0.0.1])
	by michael.uncg.edu (8.9.3/8.9.3) with ESMTP id SAA06099
	for <wlestes@localhost>; Sun, 8 Apr 2001 18:10:46 -0400
Received: from imap.uncg.edu
	by localhost with IMAP (fetchmail-5.1.0)
	for wlestes@localhost (single-drop); Sun, 08 Apr 2001 18:10:46 -0400 (EDT)
Received: from external-gw.uncg.edu (external-gw.uncg.edu [152.13.2.70])
	by hermes.email.uncg.edu (8.11.0/8.11.0) with ESMTP id f38Lr7e10030
	for <wlestes@hermes.email.uncg.edu>; Sun, 8 Apr 2001 17:53:07 -0400 (EDT)
Received: from web9407.mail.yahoo.com (web9407.mail.yahoo.com [216.136.129.23])
	by external-gw.uncg.edu (8.9.3/8.9.3) with SMTP id RAA28292
	for <wlestes@uncg.edu>; Sun, 8 Apr 2001 17:53:05 -0400 (EDT)
Message-ID: <20010408215303.56462.qmail@web9407.mail.yahoo.com>
Received: from [216.254.82.102] by web9407.mail.yahoo.com; Sun, 08 Apr 2001 14:53:03 PDT
Date: Sun, 8 Apr 2001 14:53:03 -0700 (PDT)
From: "John W. Millaway" <johnmillaway@yahoo.com>
Subject: Re: reentrant C scanner
To: "W. L. Estes" <wlestes@uncg.edu>
In-Reply-To: <200104082000.QAA05847@michael.uncg.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii

*** EOOH ***
Date: Sun, 8 Apr 2001 14:53:03 -0700 (PDT)
From: "John W. Millaway" <johnmillaway@yahoo.com>
Subject: Re: reentrant C scanner
To: "W. L. Estes" <wlestes@uncg.edu>
In-Reply-To: <200104082000.QAA05847@michael.uncg.edu>

Hi,

(Sorry it took a while, my DSL went down again.)

Here is a link to the diff, because my browser (Opera) doesn't do file-uploads
properly.

http://astro.temple.edu/~john43/flex_r.diff.gz

I did not include a diff of skel.c, because it is built from flex.skl. I did a GNU
diff, instead of CVS diff, because I had already modified flex before importing it
into CVS... which means `cvs diff' won't work without some massaging. I'm no CVS
expert, let me know if you have a solution.

This all works, but is not finished. I want to test it with as many scanners as I
can dig up. At some point, I'll submit a test directory and a "README.reenrant" as
well as a modification of the texinfo docs.

Briefly, the changes are these:

1. Added option "reentrant" (-R).
2. Added preprocessor directive: YY_REENTRANT. (and a few more related to this).
3. Added internal "struct yy_globals_t"; user-visible as "void*".
4. All global variables are conditionally eliminated and placed in the struct.
5. Nearly all functions conditionally take an additional argument.
6. yylex() is now yylex_r(void*)
7. All access to globals must be wrapped in YY_G() macro.
8. User may attach arbitrary data (void*) to the scanner, to maintain state.
9. New functions for using scanner:  yy_init_r, yy_free_r;
10. Accessor functions to extract values from struct or globals.
11. Frequently accessed variables (yytext, yyleng, etc..) have macro shortcuts,
(yytext_r, yyleng_r, etc.) which hide the internals from users building reentrant
scanners.

-John Millaway


__________________________________________________
Do You Yahoo!?
Get email at your own domain with Yahoo! Mail. 
http://personal.mail.yahoo.com/


1,,
X-Coding-System: nil
Mail-from: From johnmillaway@yahoo.com  Sun Apr  8 19:16:58 2001
Return-Path: <johnmillaway@yahoo.com>
Received: from localhost (localhost [127.0.0.1])
	by michael.uncg.edu (8.9.3/8.9.3) with ESMTP id TAA06140
	for <wlestes@localhost>; Sun, 8 Apr 2001 19:16:58 -0400
Received: from imap.uncg.edu
	by localhost with IMAP (fetchmail-5.1.0)
	for wlestes@localhost (single-drop); Sun, 08 Apr 2001 19:16:58 -0400 (EDT)
Received: from external-gw.uncg.edu (external-gw.uncg.edu [152.13.2.70])
	by hermes.email.uncg.edu (8.11.0/8.11.0) with ESMTP id f38MwXe18411
	for <wlestes@hermes.email.uncg.edu>; Sun, 8 Apr 2001 18:58:33 -0400 (EDT)
Received: from web9402.mail.yahoo.com (web9402.mail.yahoo.com [216.136.129.108])
	by external-gw.uncg.edu (8.9.3/8.9.3) with SMTP id SAA01092
	for <wlestes@uncg.edu>; Sun, 8 Apr 2001 18:58:32 -0400 (EDT)
Message-ID: <20010408225832.81903.qmail@web9402.mail.yahoo.com>
Received: from [216.254.82.102] by web9402.mail.yahoo.com; Sun, 08 Apr 2001 15:58:32 PDT
Date: Sun, 8 Apr 2001 15:58:32 -0700 (PDT)
From: "John W. Millaway" <johnmillaway@yahoo.com>
Subject: Re: reentrant C scanner
To: "W. L. Estes" <wlestes@uncg.edu>
In-Reply-To: <200104082224.SAA06112@michael.uncg.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii

*** EOOH ***
Date: Sun, 8 Apr 2001 15:58:32 -0700 (PDT)
From: "John W. Millaway" <johnmillaway@yahoo.com>
Subject: Re: reentrant C scanner
To: "W. L. Estes" <wlestes@uncg.edu>
In-Reply-To: <200104082224.SAA06112@michael.uncg.edu>

> have downloaded this and will set aside time to read your patch this
> week.
Please download again, I fixed another bug.

> > I did not include a diff of skel.c, because it is built from flex.skl. I did a
> GNU> 
> ok. will remember this...

Just do `./mkskel.sh <flex.skl >skel.c` to rebuild it.

> tag your current sources.
> import the standard sources and give them another tag/branch name.
> do a cvs diff explicitly between the two above.

Will do.


> that's the gut reaction. let me know if that works or if you have
> troubles. But I need to prepare a publicly readable cvs archive anyway.

Have you thought about sourceforge?

> the scanner stuff is finished, yes? (sure, there will be bugs and
> things to work out, but i am asking if the reentrant stuff is in place.)
Yes, finished. But the changes were so widespread and there are so many options and
directives that I'm sure something will break very soon. That's why a
regression-test directory would be nice.

> Go ahead and write the README. This is a good exercise--and usually
> fairly easy to knock off. It can serve as a guide later on down the
> road.
OK.

> To clarify, reentrant should probably not be a default behavior for
> now.
It is not. In fact, yacc/bison parsers won't call yylex_r()... something I'll work
on later.


> If it's ok with you, I'm going to forward your post on to
> help-flex@gnu.org so the "community" can dig in too.

Great. Please include my email address.

Also, the community should really decide on the new API. I am wide open to
suggestions. Currently, the new API allows a lexer to be backwards compatible, so
that a working reentrant scanner should be able to be recompiled as a non-reentrant
scanner, and everything will work just dandy on the inside. However, the syntax is
not so pretty for the user. On the other hand, a reentrant scanner that is not
meant to be backwards compatible has a cleaner API (fewer oddball macros).

An example to illustrate:

%option reentrant
%s FOOBAR

%%

  /* The line in each pair below do the same thing.
     But second of each pair is backwards-compatible. */

"call-func"  { yy_push_state ( FOOBAR , yy_globals ); }
"call-func"  { yy_push_state ( FOOBAR  YYCALL_LAST_ARG ); }

"user-data"  { printf( (char*) yyusr_r );}
"user-data"  { printf( (char*) yy_get_usr(YYCALL_ONLY_ARG));}

"yyleng"     { printf( "%s %d" , yy_text_r, yyleng_r ); }
"yyleng"     { printf( "%s %d" , YY_G(yy_text), YY_G(yyleng)); }

%%
main()
{
    char* stuff = "user specific data.";
    void * mylexer;
   
    yy_init_r ( & mylexer );       /* Create the scanner. */
    yy_set_usr( stuff, mylexer );  /* set user-defined stuff. */

    while( 1 )
        yylex_r ( lexer );

    yy_free_r ( lexer );
}

-John Millaway


__________________________________________________
Do You Yahoo!?
Get email at your own domain with Yahoo! Mail. 
http://personal.mail.yahoo.com/


1,,
X-Coding-System: nil
Mail-from: From johnmillaway@yahoo.com  Sun Apr  8 20:12:39 2001
Return-Path: <johnmillaway@yahoo.com>
Received: from localhost (localhost [127.0.0.1])
	by michael.uncg.edu (8.9.3/8.9.3) with ESMTP id UAA06162
	for <wlestes@localhost>; Sun, 8 Apr 2001 20:12:39 -0400
Received: from imap.uncg.edu
	by localhost with IMAP (fetchmail-5.1.0)
	for wlestes@localhost (single-drop); Sun, 08 Apr 2001 20:12:39 -0400 (EDT)
Received: from external-gw.uncg.edu (external-gw.uncg.edu [152.13.2.70])
	by hermes.email.uncg.edu (8.11.0/8.11.0) with ESMTP id f38NsQe25613
	for <wlestes@hermes.email.uncg.edu>; Sun, 8 Apr 2001 19:54:26 -0400 (EDT)
Received: from web9406.mail.yahoo.com (web9406.mail.yahoo.com [216.136.129.22])
	by external-gw.uncg.edu (8.9.3/8.9.3) with SMTP id TAA03391
	for <wlestes@uncg.edu>; Sun, 8 Apr 2001 19:54:24 -0400 (EDT)
Message-ID: <20010408235423.12509.qmail@web9406.mail.yahoo.com>
Received: from [216.254.82.102] by web9406.mail.yahoo.com; Sun, 08 Apr 2001 16:54:23 PDT
Date: Sun, 8 Apr 2001 16:54:23 -0700 (PDT)
From: "John W. Millaway" <johnmillaway@yahoo.com>
Subject: Re: reentrant C scanner
To: "W. L. Estes" <wlestes@uncg.edu>
In-Reply-To: <200104082224.SAA06112@michael.uncg.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii

*** EOOH ***
Date: Sun, 8 Apr 2001 16:54:23 -0700 (PDT)
From: "John W. Millaway" <johnmillaway@yahoo.com>
Subject: Re: reentrant C scanner
To: "W. L. Estes" <wlestes@uncg.edu>
In-Reply-To: <200104082224.SAA06112@michael.uncg.edu>

I'm sure you haven't had time to look at it yet, but there are two other points I
wanted to mention:

1. Once the reentrant scanner is working, there is no need for the non-reentrant
scanner. The non-reentrant API can be built as a wrapper around the reentrant
scanner. Much, much less code maintenance. I will happily do the conversion.

2. The reentrancy can be verified, in part, by running `nm -A -l' on a compiled
scanner. `nm` is part of binutils and shows, among other things, the global
variables. In the reentrant scanner, there should be none.

-John Millaway

PS - What is the mailing list? I guess I should subscribe! The gnu site says that
there is currently no list for flex:
http://www.gnu.org/software/flex/flex.html#lists


__________________________________________________
Do You Yahoo!?
Get email at your own domain with Yahoo! Mail. 
http://personal.mail.yahoo.com/


1, answered,,
X-Coding-System: nil
Mail-from: From vern@daffy.ee.lbl.gov  Tue Apr 17 04:00:22 2001
Return-Path: <vern@daffy.ee.lbl.gov>
Received: from localhost (localhost [127.0.0.1])
	by michael.uncg.edu (8.9.3/8.9.3) with ESMTP id EAA31012
	for <wlestes@localhost>; Tue, 17 Apr 2001 04:00:21 -0400
Received: from imap.uncg.edu
	by localhost with IMAP (fetchmail-5.1.0)
	for wlestes@localhost (single-drop); Tue, 17 Apr 2001 04:00:21 -0400 (EDT)
Received: from external-gw.uncg.edu (external-gw.uncg.edu [152.13.2.70])
	by hermes.email.uncg.edu (8.11.0/8.11.0) with ESMTP id f3H7eqe08321
	for <wlestes@hermes.email.uncg.edu>; Tue, 17 Apr 2001 03:40:52 -0400 (EDT)
Received: from daffy.ee.lbl.gov (daffy.ee.lbl.gov [131.243.1.31])
	by external-gw.uncg.edu (8.9.3/8.9.3) with ESMTP id DAA08895
	for <wlestes@uncg.edu>; Tue, 17 Apr 2001 03:40:50 -0400 (EDT)
Received: (from vern@localhost)
	by daffy.ee.lbl.gov (8.10.0/8.10.0) id f3H7eoW18946;
	Tue, 17 Apr 2001 00:40:50 -0700 (PDT)
Message-Id: <200104170740.f3H7eoW18946@daffy.ee.lbl.gov>
To: "Yuri Victorovich" <yurivict@hotmail.com>
Subject: Re: Flex comments
Cc: wlestes@uncg.edu
In-reply-to: Your message of Sun, 15 Apr 2001 22:47:37 PDT.
Date: Tue, 17 Apr 2001 00:40:50 PDT
From: Vern Paxson <vern@ee.lbl.gov>

*** EOOH ***
To: "Yuri Victorovich" <yurivict@hotmail.com>
Subject: Re: Flex comments
Cc: wlestes@uncg.edu
In-reply-to: Your message of Sun, 15 Apr 2001 22:47:37 PDT.
Date: Tue, 17 Apr 2001 00:40:50 PDT
From: Vern Paxson <vern@ee.lbl.gov>

> One problem: how to eliminate the static variables without making it C++ 
> scanner with option -+? I do not want to introduce any of C++ into my 
> project and at the same time I use pthreads on FreeBSD.
> 
> That would be a nice feature to have such an option.

I agree it would be nice.  I've cc'd Will Estes, who is trying out taking
over flex maintenance / development.

		Vern


1, answered,,
X-Coding-System: nil
Mail-from: From yurivict@hotmail.com  Tue Apr 17 10:42:56 2001
Return-Path: <yurivict@hotmail.com>
Received: from localhost (localhost [127.0.0.1])
	by michael.uncg.edu (8.9.3/8.9.3) with ESMTP id KAA01266
	for <wlestes@localhost>; Tue, 17 Apr 2001 10:42:56 -0400
Received: from imap.uncg.edu
	by localhost with IMAP (fetchmail-5.1.0)
	for wlestes@localhost (single-drop); Tue, 17 Apr 2001 10:42:56 -0400 (EDT)
Received: from external-gw.uncg.edu (external-gw.uncg.edu [152.13.2.70])
	by hermes.email.uncg.edu (8.11.0/8.11.0) with ESMTP id f3HENMe01358
	for <wlestes@hermes.email.uncg.edu>; Tue, 17 Apr 2001 10:23:22 -0400 (EDT)
Received: from hotmail.com (f210.law12.hotmail.com [64.4.19.210])
	by external-gw.uncg.edu (8.9.3/8.9.3) with ESMTP id KAA06325
	for <wlestes@uncg.edu>; Tue, 17 Apr 2001 10:23:21 -0400 (EDT)
Received: from mail pickup service by hotmail.com with Microsoft SMTPSVC;
	 Tue, 17 Apr 2001 07:22:51 -0700
Received: from 194.158.212.99 by lw12fd.law12.hotmail.msn.com with HTTP;	Tue, 17 Apr 2001 14:22:51 GMT
X-Originating-IP: [194.158.212.99]
From: "Yuri Victorovich" <yurivict@hotmail.com>
To: wlestes@uncg.edu
Subject: Re: Flex comments
Date: Tue, 17 Apr 2001 07:22:51 -0700
Mime-Version: 1.0
Content-Type: text/plain; format=flowed
Message-ID: <F210b98KtdhQYZkxOjm00009556@hotmail.com>
X-OriginalArrivalTime: 17 Apr 2001 14:22:51.0694 (UTC) FILETIME=[E2BCD4E0:01C0C749]

*** EOOH ***
From: "Yuri Victorovich" <yurivict@hotmail.com>
To: wlestes@uncg.edu
Subject: Re: Flex comments
Date: Tue, 17 Apr 2001 07:22:51 -0700

Will, Thank you for reply!

I spent the whole day yesterday writing awk script to
convert normal flex output to thread-safe form.

If you do not have time or will need help with this
work on flex -- I can do anything from research, programming
to testing. Just let me know -- I digged into it to some
extent already and will have some time in the coming weeks.

Kind Regards,
Yuri.





>From: "W. L. Estes" <wlestes@uncg.edu>
>To: vern@ee.lbl.gov
>CC: yurivict@hotmail.com
>Subject: Re: Flex comments
>Date: Tue, 17 Apr 2001 08:33:35 -0400
>
> > > One problem: how to eliminate the static variables without making it 
>C++
> > > scanner with option -+? I do not want to introduce any of C++ into my
> > > project and at the same time I use pthreads on FreeBSD.
> > >
> > > That would be a nice feature to have such an option.
> >
> > I agree it would be nice.  I've cc'd Will Estes, who is trying out 
>taking
> > over flex maintenance / development.
>
>Thanks for the suggestion. I'll look into what it will take to do
>this.
>
>--Will

_________________________________________________________________
Get your FREE download of MSN Explorer at http://explorer.msn.com


1, answered,,
X-Coding-System: nil
Mail-from: From vern@daffy.ee.lbl.gov  Tue Apr 17 12:34:26 2001
Return-Path: <vern@daffy.ee.lbl.gov>
Received: from localhost (localhost [127.0.0.1])
	by michael.uncg.edu (8.9.3/8.9.3) with ESMTP id MAA01409
	for <wlestes@localhost>; Tue, 17 Apr 2001 12:34:26 -0400
Received: from imap.uncg.edu
	by localhost with IMAP (fetchmail-5.1.0)
	for wlestes@localhost (single-drop); Tue, 17 Apr 2001 12:34:26 -0400 (EDT)
Received: from external-gw.uncg.edu (external-gw.uncg.edu [152.13.2.70])
	by hermes.email.uncg.edu (8.11.0/8.11.0) with ESMTP id f3HGFpe27457
	for <wlestes@hermes.email.uncg.edu>; Tue, 17 Apr 2001 12:15:51 -0400 (EDT)
Received: from daffy.ee.lbl.gov (daffy.ee.lbl.gov [131.243.1.31])
	by external-gw.uncg.edu (8.9.3/8.9.3) with ESMTP id MAA20041
	for <wlestes@uncg.edu>; Tue, 17 Apr 2001 12:15:49 -0400 (EDT)
Received: (from vern@localhost)
	by daffy.ee.lbl.gov (8.10.0/8.10.0) id f3HGFoP19785;
	Tue, 17 Apr 2001 09:15:50 -0700 (PDT)
Message-Id: <200104171615.f3HGFoP19785@daffy.ee.lbl.gov>
To: "W. L. Estes" <wlestes@uncg.edu>
Subject: Re: Flex comments
In-reply-to: Your message of Tue, 17 Apr 2001 08:57:41 PDT.
Date: Tue, 17 Apr 2001 09:15:50 PDT
From: Vern Paxson <vern@ee.lbl.gov>

*** EOOH ***
To: "W. L. Estes" <wlestes@uncg.edu>
Subject: Re: Flex comments
In-reply-to: Your message of Tue, 17 Apr 2001 08:57:41 PDT.
Date: Tue, 17 Apr 2001 09:15:50 PDT
From: Vern Paxson <vern@ee.lbl.gov>

> If there is more detail in the original message, could you forward it
> to me?

That was the whole message.  The basic problem that people run into
is they want reentrant scanners, so no statics or global variables.
The general strategy for this is to have a new yylex() interface in
which the caller passes in a struct that has all of the otherwise-global
variables.  This is a bit tricky to do cleanly because you can't switch
to that as *the* interface, of course, because it would break the
existing API; which means you have to be able to generate two forms of
the skeleton file, one for the existing API with the statics/globals,
and one without them and the struct definition instead.

There's probably some more discussion on this in the mail tarball I sent
you - search for "reentrant".

		Vern


1, filed, answered,,
X-Coding-System: nil
Mail-from: From vern@daffy.ee.lbl.gov  Tue Apr 17 12:34:26 2001
Return-Path: <vern@daffy.ee.lbl.gov>
Received: from localhost (localhost [127.0.0.1])
	by michael.uncg.edu (8.9.3/8.9.3) with ESMTP id MAA01409
	for <wlestes@localhost>; Tue, 17 Apr 2001 12:34:26 -0400
Received: from imap.uncg.edu
	by localhost with IMAP (fetchmail-5.1.0)
	for wlestes@localhost (single-drop); Tue, 17 Apr 2001 12:34:26 -0400 (EDT)
Received: from external-gw.uncg.edu (external-gw.uncg.edu [152.13.2.70])
	by hermes.email.uncg.edu (8.11.0/8.11.0) with ESMTP id f3HGFpe27457
	for <wlestes@hermes.email.uncg.edu>; Tue, 17 Apr 2001 12:15:51 -0400 (EDT)
Received: from daffy.ee.lbl.gov (daffy.ee.lbl.gov [131.243.1.31])
	by external-gw.uncg.edu (8.9.3/8.9.3) with ESMTP id MAA20041
	for <wlestes@uncg.edu>; Tue, 17 Apr 2001 12:15:49 -0400 (EDT)
Received: (from vern@localhost)
	by daffy.ee.lbl.gov (8.10.0/8.10.0) id f3HGFoP19785;
	Tue, 17 Apr 2001 09:15:50 -0700 (PDT)
Message-Id: <200104171615.f3HGFoP19785@daffy.ee.lbl.gov>
To: "W. L. Estes" <wlestes@uncg.edu>
Subject: Re: Flex comments
In-reply-to: Your message of Tue, 17 Apr 2001 08:57:41 PDT.
Date: Tue, 17 Apr 2001 09:15:50 PDT
From: Vern Paxson <vern@ee.lbl.gov>

*** EOOH ***
To: "W. L. Estes" <wlestes@uncg.edu>
Subject: Re: Flex comments
In-reply-to: Your message of Tue, 17 Apr 2001 08:57:41 PDT.
Date: Tue, 17 Apr 2001 09:15:50 PDT
From: Vern Paxson <vern@ee.lbl.gov>

> If there is more detail in the original message, could you forward it
> to me?

That was the whole message.  The basic problem that people run into
is they want reentrant scanners, so no statics or global variables.
The general strategy for this is to have a new yylex() interface in
which the caller passes in a struct that has all of the otherwise-global
variables.  This is a bit tricky to do cleanly because you can't switch
to that as *the* interface, of course, because it would break the
existing API; which means you have to be able to generate two forms of
the skeleton file, one for the existing API with the statics/globals,
and one without them and the struct definition instead.

There's probably some more discussion on this in the mail tarball I sent
you - search for "reentrant".

		Vern


1,,
X-Coding-System: nil
Mail-from: From help-flex-admin@gnu.org  Wed Apr 25 00:20:40 2001
Return-Path: <help-flex-admin@gnu.org>
Received: from localhost (localhost [127.0.0.1])
	by michael.uncg.edu (8.9.3/8.9.3) with ESMTP id AAA22703
	for <wlestes@localhost>; Wed, 25 Apr 2001 00:20:40 -0400
Received: from imap.uncg.edu
	by localhost with IMAP (fetchmail-5.1.0)
	for wlestes@localhost (single-drop); Wed, 25 Apr 2001 00:20:40 -0400 (EDT)
Received: from external-gw.uncg.edu (external-gw.uncg.edu [152.13.2.70])
	by hermes.email.uncg.edu (8.11.0/8.11.0) with ESMTP id f3P403e01034
	for <wlestes@hermes.email.uncg.edu>; Wed, 25 Apr 2001 00:00:03 -0400 (EDT)
Received: from fencepost.gnu.org (we-refuse-to-spy-on-our-users@fencepost.gnu.org [199.232.76.164])
	by external-gw.uncg.edu (8.9.3/8.9.3) with ESMTP id AAA05202
	for <wlestes@uncg.edu>; Wed, 25 Apr 2001 00:00:02 -0400 (EDT)
Received: from localhost ([127.0.0.1] helo=fencepost.gnu.org)
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 14sGTC-0006Tm-00
	for <wlestes@uncg.edu>; Wed, 25 Apr 2001 00:00:02 -0400
Received: from [203.199.199.50] (helo=banyannetworks.com)
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 14sGRt-0006Sr-00
	for <help-flex@gnu.org>; Tue, 24 Apr 2001 23:58:44 -0400
Received: from localhost (arasi@localhost) by banyannetworks.com (8.9.3/8.8.3) with ESMTP id JAA03208 for <help-flex@gnu.org>; Wed, 25 Apr 2001 09:25:36 +0530
From: Avai Arasi <arasi@banyanNetworks.com>
To: help-flex@gnu.org
Subject: Clarification Regarding Porting Flex and Bison to RTOS
Message-ID: <Pine.LNX.4.10.10104250920140.3092-100000@banyannetworks.com>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: help-flex-admin@gnu.org
Errors-To: help-flex-admin@gnu.org
X-BeenThere: help-flex@gnu.org
X-Mailman-Version: 2.0.3
Precedence: bulk
List-Help: <mailto:help-flex-request@gnu.org?subject=help>
List-Post: <mailto:help-flex@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/help-flex>,
	<mailto:help-flex-request@gnu.org?subject=subscribe>
List-Id: Users list for Flex,
	the GNU lexical analyser generator <help-flex.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/help-flex>,
	<mailto:help-flex-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/help-flex/>
Date: Wed, 25 Apr 2001 09:25:36 +0530 (IST)

*** EOOH ***
From: Avai Arasi <arasi@banyanNetworks.com>
To: help-flex@gnu.org
Subject: Clarification Regarding Porting Flex and Bison to RTOS
Sender: help-flex-admin@gnu.org
Precedence: bulk
List-Help: <mailto:help-flex-request@gnu.org?subject=help>
List-Post: <mailto:help-flex@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/help-flex>,
	<mailto:help-flex-request@gnu.org?subject=subscribe>
List-Id: Users list for Flex,
	the GNU lexical analyser generator <help-flex.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/help-flex>,
	<mailto:help-flex-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/help-flex/>
Date: Wed, 25 Apr 2001 09:25:36 +0530 (IST)

Hello,

   I am trying to port both Flex and Bison to a Real Time OS. I don't have
the device options like the file which the Flex expects as input. In the
place of the file i have a character pointer. How should I proceed with
the parsing ? 

  Is there any standard procedure for this porting ? If somebody has
already done such a work can i have some hints please.

Thank You.

Regards,
Avaiarasi
Senior Engineer R & D
Banyan Networks Pvt. Ltd.
Chennai.


_______________________________________________
Help-flex mailing list
Help-flex@gnu.org
http://mail.gnu.org/mailman/listinfo/help-flex


1,,
X-Coding-System: nil
Mail-from: From help-flex-admin@gnu.org  Wed Apr 25 13:17:31 2001
Return-Path: <help-flex-admin@gnu.org>
Received: from localhost (localhost [127.0.0.1])
	by michael.uncg.edu (8.9.3/8.9.3) with ESMTP id NAA23790
	for <wlestes@localhost>; Wed, 25 Apr 2001 13:17:31 -0400
Received: from imap.uncg.edu
	by localhost with IMAP (fetchmail-5.1.0)
	for wlestes@localhost (single-drop); Wed, 25 Apr 2001 13:17:31 -0400 (EDT)
Received: from external-gw.uncg.edu (external-gw.uncg.edu [152.13.2.70])
	by hermes.email.uncg.edu (8.11.0/8.11.0) with ESMTP id f3PGv2e04258
	for <wlestes@hermes.email.uncg.edu>; Wed, 25 Apr 2001 12:57:02 -0400 (EDT)
Received: from fencepost.gnu.org (fencepost.gnu.org [199.232.76.164])
	by external-gw.uncg.edu (8.9.3/8.9.3) with ESMTP id MAA11157
	for <wlestes@uncg.edu>; Wed, 25 Apr 2001 12:57:01 -0400 (EDT)
Received: from localhost ([127.0.0.1] helo=fencepost.gnu.org)
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 14sSb8-00056g-00
	for <wlestes@uncg.edu>; Wed, 25 Apr 2001 12:57:02 -0400
Received: from web9403.mail.yahoo.com ([216.136.129.109])
	by fencepost.gnu.org with smtp (Exim 3.16 #1 (Debian))
	id 14sSYq-00053W-00
	for <help-flex@gnu.org>; Wed, 25 Apr 2001 12:54:41 -0400
Message-ID: <20010425165439.64856.qmail@web9403.mail.yahoo.com>
Received: from [216.254.82.102] by web9403.mail.yahoo.com; Wed, 25 Apr 2001 09:54:39 PDT
From: "John W. Millaway" <johnmillaway@yahoo.com>
Subject: Bug: %option main overrides %option yywrap
To: help-flex@gnu.org
In-Reply-To: <200104251228.IAA23213@michael.uncg.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Sender: help-flex-admin@gnu.org
Errors-To: help-flex-admin@gnu.org
X-BeenThere: help-flex@gnu.org
X-Mailman-Version: 2.0.3
Precedence: bulk
List-Help: <mailto:help-flex-request@gnu.org?subject=help>
List-Post: <mailto:help-flex@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/help-flex>,
	<mailto:help-flex-request@gnu.org?subject=subscribe>
List-Id: Users list for Flex,
	the GNU lexical analyser generator <help-flex.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/help-flex>,
	<mailto:help-flex-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/help-flex/>
Date: Wed, 25 Apr 2001 09:54:39 -0700 (PDT)

*** EOOH ***
From: "John W. Millaway" <johnmillaway@yahoo.com>
Subject: Bug: %option main overrides %option yywrap
To: help-flex@gnu.org
In-Reply-To: <200104251228.IAA23213@michael.uncg.edu>
Sender: help-flex-admin@gnu.org
Precedence: bulk
List-Help: <mailto:help-flex-request@gnu.org?subject=help>
List-Post: <mailto:help-flex@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/help-flex>,
	<mailto:help-flex-request@gnu.org?subject=subscribe>
List-Id: Users list for Flex,
	the GNU lexical analyser generator <help-flex.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/help-flex>,
	<mailto:help-flex-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/help-flex/>
Date: Wed, 25 Apr 2001 09:54:39 -0700 (PDT)

Hello,

There is a "bug" in the current version of flex (2.5.4) in which the %option
[no]main unconditionally sets the %option [no]yywrap. I use the term "bug"
hesitantly, because this behavior appears to be intentional. However, I can't think
of any use for it other than to cause unexpected link errors when yywrap() is not
found.

Example:

  %option noyywrap  /* Disable option yywrap. */
  %option nomain    /* Option yywrap is now enabled! */

One workaround is to always specify %option yywrap AFTER %option main. In general,
if the user explicitly sets an option, then flex should either honor that setting,
or report a conflict.

The bug is in "scan.l" at line 247, and is undocumented.

244:	lex-compat	lex_compat = option_sense;
245:	main		{
246:			action_define( "YY_MAIN", option_sense );
247:			do_yywrap = ! option_sense;
248:			}
249:	meta-ecs	usemecs = option_sense;

-John Millaway


__________________________________________________
Do You Yahoo!?
Yahoo! Auctions - buy the things you want at great prices
http://auctions.yahoo.com/

_______________________________________________
Help-flex mailing list
Help-flex@gnu.org
http://mail.gnu.org/mailman/listinfo/help-flex


1,,
X-Coding-System: nil
Mail-from: From help-flex-admin@gnu.org  Wed Apr 25 14:11:45 2001
Return-Path: <help-flex-admin@gnu.org>
Received: from localhost (localhost [127.0.0.1])
	by michael.uncg.edu (8.9.3/8.9.3) with ESMTP id OAA23826
	for <wlestes@localhost>; Wed, 25 Apr 2001 14:11:44 -0400
Received: from imap.uncg.edu
	by localhost with IMAP (fetchmail-5.1.0)
	for wlestes@localhost (single-drop); Wed, 25 Apr 2001 14:11:44 -0400 (EDT)
Received: from external-gw.uncg.edu (external-gw.uncg.edu [152.13.2.70])
	by hermes.email.uncg.edu (8.11.0/8.11.0) with ESMTP id f3PHp7e27019
	for <wlestes@hermes.email.uncg.edu>; Wed, 25 Apr 2001 13:51:07 -0400 (EDT)
Received: from fencepost.gnu.org (fencepost.gnu.org [199.232.76.164])
	by external-gw.uncg.edu (8.9.3/8.9.3) with ESMTP id NAA17880
	for <wlestes@uncg.edu>; Wed, 25 Apr 2001 13:51:08 -0400 (EDT)
Received: from localhost ([127.0.0.1] helo=fencepost.gnu.org)
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 14sTRT-0003QA-00
	for <wlestes@uncg.edu>; Wed, 25 Apr 2001 13:51:07 -0400
Received: from web9404.mail.yahoo.com ([216.136.129.110])
	by fencepost.gnu.org with smtp (Exim 3.16 #1 (Debian))
	id 14sTQy-0003Pm-00
	for <help-flex@gnu.org>; Wed, 25 Apr 2001 13:50:36 -0400
Message-ID: <20010425175036.43329.qmail@web9404.mail.yahoo.com>
Received: from [216.254.82.102] by web9404.mail.yahoo.com; Wed, 25 Apr 2001 10:50:36 PDT
From: "John W. Millaway" <johnmillaway@yahoo.com>
Subject: Re: scaning from memory
To: "alexandre.gouraud" <alexandre.gouraud@laposte.net>, help-flex@gnu.org
In-Reply-To: <GCCZ8R$IiFF_2Id0FNAO2kAXNPACqX5slPk_GQyEsHVyKKjYXVpqFi3j4Ntc@laposte.net>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Sender: help-flex-admin@gnu.org
Errors-To: help-flex-admin@gnu.org
X-BeenThere: help-flex@gnu.org
X-Mailman-Version: 2.0.3
Precedence: bulk
List-Help: <mailto:help-flex-request@gnu.org?subject=help>
List-Post: <mailto:help-flex@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/help-flex>,
	<mailto:help-flex-request@gnu.org?subject=subscribe>
List-Id: Users list for Flex,
	the GNU lexical analyser generator <help-flex.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/help-flex>,
	<mailto:help-flex-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/help-flex/>
Date: Wed, 25 Apr 2001 10:50:36 -0700 (PDT)

*** EOOH ***
From: "John W. Millaway" <johnmillaway@yahoo.com>
Subject: Re: scaning from memory
To: "alexandre.gouraud" <alexandre.gouraud@laposte.net>, help-flex@gnu.org
In-Reply-To: <GCCZ8R$IiFF_2Id0FNAO2kAXNPACqX5slPk_GQyEsHVyKKjYXVpqFi3j4Ntc@laposte.net>
Sender: help-flex-admin@gnu.org
Precedence: bulk
List-Help: <mailto:help-flex-request@gnu.org?subject=help>
List-Post: <mailto:help-flex@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/help-flex>,
	<mailto:help-flex-request@gnu.org?subject=subscribe>
List-Id: Users list for Flex,
	the GNU lexical analyser generator <help-flex.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/help-flex>,
	<mailto:help-flex-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/help-flex/>
Date: Wed, 25 Apr 2001 10:50:36 -0700 (PDT)

> I am trying to write a parser with flex/bison scaning strings
> in memory, but I can't make it work. I am pretty sure it's a
> common way of using flex, but I didn't find any example on the
> web. (I have read the manual, of course). I have found several
> times this question, but no answer.
> I don't know where to put the yy_scan_string() statement. I
> use the <<EOF>> rule as last rule. And I have also modified
> the prototype of the yylex function with
> #define YY_DECL int yylex(YYSTYPE *lvalp,char *name)
> where name is the string I want to parse.

You should call yy_scan_string before calling yylex. Here are two examples:

1. In general:

  YY_BUFFER_STATE buf_state = yy_scan_string( str );
  while( yylex() != 0 )
      ;
  yy_delete_buffer(buf_state);


2. Your particular scanner:

%{
#define YY_USER_INIT  buf_state = yy_scan_string( name );
%}
%%
    static YY_BUFFER_STATE buf_state;

<<EOF>>   { yy_buffer_delete( buf_state ); yyterminate(); }
%%

-John


__________________________________________________
Do You Yahoo!?
Yahoo! Auctions - buy the things you want at great prices
http://auctions.yahoo.com/

_______________________________________________
Help-flex mailing list
Help-flex@gnu.org
http://mail.gnu.org/mailman/listinfo/help-flex


1,,
X-Coding-System: nil
Mail-from: From help-flex-admin@gnu.org  Wed Apr 25 14:41:58 2001
Return-Path: <help-flex-admin@gnu.org>
Received: from localhost (localhost [127.0.0.1])
	by michael.uncg.edu (8.9.3/8.9.3) with ESMTP id OAA23852
	for <wlestes@localhost>; Wed, 25 Apr 2001 14:41:58 -0400
Received: from imap.uncg.edu
	by localhost with IMAP (fetchmail-5.1.0)
	for wlestes@localhost (single-drop); Wed, 25 Apr 2001 14:41:58 -0400 (EDT)
Received: from external-gw.uncg.edu (external-gw.uncg.edu [152.13.2.70])
	by hermes.email.uncg.edu (8.11.0/8.11.0) with ESMTP id f3PILLe10211
	for <wlestes@hermes.email.uncg.edu>; Wed, 25 Apr 2001 14:21:21 -0400 (EDT)
Received: from fencepost.gnu.org (fencepost.gnu.org [199.232.76.164])
	by external-gw.uncg.edu (8.9.3/8.9.3) with ESMTP id OAA21692
	for <wlestes@uncg.edu>; Wed, 25 Apr 2001 14:21:21 -0400 (EDT)
Received: from localhost ([127.0.0.1] helo=fencepost.gnu.org)
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 14sTuj-0007Yw-00
	for <wlestes@uncg.edu>; Wed, 25 Apr 2001 14:21:21 -0400
Received: from daffy.ee.lbl.gov ([131.243.1.31])
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 14sTtm-0007Xg-00
	for <help-flex@gnu.org>; Wed, 25 Apr 2001 14:20:22 -0400
Received: (from vern@localhost)
	by daffy.ee.lbl.gov (8.10.0/8.10.0) id f3PIKJa19998;
	Wed, 25 Apr 2001 11:20:19 -0700 (PDT)
Message-Id: <200104251820.f3PIKJa19998@daffy.ee.lbl.gov>
To: "John W. Millaway" <johnmillaway@yahoo.com>
Cc: help-flex@gnu.org
Subject: Re: Bug: %option main overrides %option yywrap
In-reply-to: Your message of Wed, 25 Apr 2001 09:54:39 PDT.
From: Vern Paxson <vern@ee.lbl.gov>
Sender: help-flex-admin@gnu.org
Errors-To: help-flex-admin@gnu.org
X-BeenThere: help-flex@gnu.org
X-Mailman-Version: 2.0.3
Precedence: bulk
List-Help: <mailto:help-flex-request@gnu.org?subject=help>
List-Post: <mailto:help-flex@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/help-flex>,
	<mailto:help-flex-request@gnu.org?subject=subscribe>
List-Id: Users list for Flex,
	the GNU lexical analyser generator <help-flex.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/help-flex>,
	<mailto:help-flex-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/help-flex/>
Date: Wed, 25 Apr 2001 11:20:19 PDT

*** EOOH ***
To: "John W. Millaway" <johnmillaway@yahoo.com>
Cc: help-flex@gnu.org
Subject: Re: Bug: %option main overrides %option yywrap
In-reply-to: Your message of Wed, 25 Apr 2001 09:54:39 PDT.
From: Vern Paxson <vern@ee.lbl.gov>
Sender: help-flex-admin@gnu.org
Precedence: bulk
List-Help: <mailto:help-flex-request@gnu.org?subject=help>
List-Post: <mailto:help-flex@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/help-flex>,
	<mailto:help-flex-request@gnu.org?subject=subscribe>
List-Id: Users list for Flex,
	the GNU lexical analyser generator <help-flex.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/help-flex>,
	<mailto:help-flex-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/help-flex/>
Date: Wed, 25 Apr 2001 11:20:19 PDT

> 245:	main		{
> 246:			action_define( "YY_MAIN", option_sense );
> 247:			do_yywrap = ! option_sense;
> 248:			}

I think I did that (which is documented in the manual, by the way) so that
"%option main" would be a quick way to specify you want a bare-bones
program for which you don't have to define any additional routines.
I'd be disinclined to change it (though it's Will's call) because it's
so easy to work around - just don't use "%option main", which is only
saving you a few lines of code anyway.

		Vern

_______________________________________________
Help-flex mailing list
Help-flex@gnu.org
http://mail.gnu.org/mailman/listinfo/help-flex


1,,
X-Coding-System: nil
Mail-from: From help-flex-admin@gnu.org  Wed Apr 25 16:28:59 2001
Return-Path: <help-flex-admin@gnu.org>
Received: from localhost (localhost [127.0.0.1])
	by michael.uncg.edu (8.9.3/8.9.3) with ESMTP id QAA23955
	for <wlestes@localhost>; Wed, 25 Apr 2001 16:28:59 -0400
Received: from imap.uncg.edu
	by localhost with IMAP (fetchmail-5.1.0)
	for wlestes@localhost (single-drop); Wed, 25 Apr 2001 16:28:59 -0400 (EDT)
Received: from external-gw.uncg.edu (external-gw.uncg.edu [152.13.2.70])
	by hermes.email.uncg.edu (8.11.0/8.11.0) with ESMTP id f3PK99e25951
	for <wlestes@hermes.email.uncg.edu>; Wed, 25 Apr 2001 16:09:09 -0400 (EDT)
Received: from fencepost.gnu.org (fencepost.gnu.org [199.232.76.164])
	by external-gw.uncg.edu (8.9.3/8.9.3) with ESMTP id QAA05277
	for <wlestes@uncg.edu>; Wed, 25 Apr 2001 16:09:08 -0400 (EDT)
Received: from localhost ([127.0.0.1] helo=fencepost.gnu.org)
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 14sVb1-0005jV-00
	for <wlestes@uncg.edu>; Wed, 25 Apr 2001 16:09:07 -0400
Received: from michael.uncg.edu ([152.13.5.20])
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 14sVaJ-0005iu-00
	for <help-flex@gnu.org>; Wed, 25 Apr 2001 16:08:23 -0400
Received: (from wlestes@localhost)
	by michael.uncg.edu (8.9.3/8.9.3) id QAA23947;
	Wed, 25 Apr 2001 16:27:46 -0400
Message-Id: <200104252027.QAA23947@michael.uncg.edu>
X-Authentication-Warning: michael.uncg.edu: wlestes set sender to wlestes@michael.uncg.edu using -f
From: "W. L. Estes" <wlestes@uncg.edu>
To: vern@ee.lbl.gov
CC: johnmillaway@yahoo.com, help-flex@gnu.org
In-reply-to: <200104251820.f3PIKJa19998@daffy.ee.lbl.gov> (message from Vern
	Paxson on Wed, 25 Apr 2001 11:20:19 PDT)
Subject: Re: Bug: %option main overrides %option yywrap
References:  <200104251820.f3PIKJa19998@daffy.ee.lbl.gov>
Sender: help-flex-admin@gnu.org
Errors-To: help-flex-admin@gnu.org
X-BeenThere: help-flex@gnu.org
X-Mailman-Version: 2.0.3
Precedence: bulk
List-Help: <mailto:help-flex-request@gnu.org?subject=help>
List-Post: <mailto:help-flex@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/help-flex>,
	<mailto:help-flex-request@gnu.org?subject=subscribe>
List-Id: Users list for Flex,
	the GNU lexical analyser generator <help-flex.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/help-flex>,
	<mailto:help-flex-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/help-flex/>
Date: Wed, 25 Apr 2001 16:27:46 -0400

*** EOOH ***
From: "W. L. Estes" <wlestes@uncg.edu>
To: vern@ee.lbl.gov
CC: johnmillaway@yahoo.com, help-flex@gnu.org
In-reply-to: <200104251820.f3PIKJa19998@daffy.ee.lbl.gov> (message from Vern
	Paxson on Wed, 25 Apr 2001 11:20:19 PDT)
Subject: Re: Bug: %option main overrides %option yywrap
Sender: help-flex-admin@gnu.org
Precedence: bulk
List-Help: <mailto:help-flex-request@gnu.org?subject=help>
List-Post: <mailto:help-flex@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/help-flex>,
	<mailto:help-flex-request@gnu.org?subject=subscribe>
List-Id: Users list for Flex,
	the GNU lexical analyser generator <help-flex.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/help-flex>,
	<mailto:help-flex-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/help-flex/>
Date: Wed, 25 Apr 2001 16:27:46 -0400

> I think I did that (which is documented in the manual, by the way) so that
> "%option main" would be a quick way to specify you want a bare-bones
> program for which you don't have to define any additional routines.
> I'd be disinclined to change it (though it's Will's call) because it's
> so easy to work around - just don't use "%option main", which is only
> saving you a few lines of code anyway.

Additionally, flex.1 says the following:

       main   directs  flex  to  provide a default main() program
              for the scanner, which simply calls yylex().   This
              option implies noyywrap (see below).

My sense is that Vern is correct: The right thing to do is to not use
%option main if you want yywrap() to be under your control.

Thanks,

--Will

_______________________________________________
Help-flex mailing list
Help-flex@gnu.org
http://mail.gnu.org/mailman/listinfo/help-flex


1,,
X-Coding-System: nil
Mail-from: From help-flex-admin@gnu.org  Thu Apr 26 12:11:23 2001
Return-Path: <help-flex-admin@gnu.org>
Received: from localhost (localhost [127.0.0.1])
	by michael.uncg.edu (8.9.3/8.9.3) with ESMTP id MAA25162
	for <wlestes@localhost>; Thu, 26 Apr 2001 12:11:22 -0400
Received: from imap.uncg.edu
	by localhost with IMAP (fetchmail-5.1.0)
	for wlestes@localhost (single-drop); Thu, 26 Apr 2001 12:11:22 -0400 (EDT)
Received: from external-gw.uncg.edu (external-gw.uncg.edu [152.13.2.70])
	by hermes.email.uncg.edu (8.11.0/8.11.0) with ESMTP id f3QFp6e21203
	for <wlestes@hermes.email.uncg.edu>; Thu, 26 Apr 2001 11:51:06 -0400 (EDT)
Received: from fencepost.gnu.org (fencepost.gnu.org [199.232.76.164])
	by external-gw.uncg.edu (8.9.3/8.9.3) with ESMTP id LAA21348
	for <wlestes@uncg.edu>; Thu, 26 Apr 2001 11:51:05 -0400 (EDT)
Received: from localhost ([127.0.0.1] helo=fencepost.gnu.org)
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 14so2r-00043k-00
	for <wlestes@uncg.edu>; Thu, 26 Apr 2001 11:51:05 -0400
Received: from [208.155.173.144] (helo=symbology.com)
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 14so2e-0003Dl-00
	for <help-flex@gnu.org>; Thu, 26 Apr 2001 11:50:53 -0400
Received: from Spooler by symbology.com (Mercury/32 v3.21c) ID MO003737;
    26 Apr 01 10:52:43 -0500
Received: from spooler by symbology.com (Mercury/32 v3.21c); 26 Apr 01 10:52:29 -0500
Received: from grommit (10.0.0.74) by Symbology (Mercury/32 v3.21c) ID MG003736;
   26 Apr 01 10:52:22 -0500
From: "Dave Dutcher" <ddutcher@symbology.com>
To: <help-flex@gnu.org>
Subject: Suppress copying characters to yytext?
Message-ID: <002501c0ce68$99b3ec40$4a00000a@symbology.com>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
X-Priority: 3 (Normal)
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook CWS, Build 9.0.2416 (9.0.2910.0)
X-MimeOLE: Produced By Microsoft MimeOLE V5.50.4133.2400
Importance: Normal
Sender: help-flex-admin@gnu.org
Errors-To: help-flex-admin@gnu.org
X-BeenThere: help-flex@gnu.org
X-Mailman-Version: 2.0.3
Precedence: bulk
List-Help: <mailto:help-flex-request@gnu.org?subject=help>
List-Post: <mailto:help-flex@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/help-flex>,
	<mailto:help-flex-request@gnu.org?subject=subscribe>
List-Id: Users list for Flex,
	the GNU lexical analyser generator <help-flex.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/help-flex>,
	<mailto:help-flex-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/help-flex/>
Date: Thu, 26 Apr 2001 10:50:20 -0500

*** EOOH ***
From: "Dave Dutcher" <ddutcher@symbology.com>
To: <help-flex@gnu.org>
Subject: Suppress copying characters to yytext?
Importance: Normal
Sender: help-flex-admin@gnu.org
Precedence: bulk
List-Help: <mailto:help-flex-request@gnu.org?subject=help>
List-Post: <mailto:help-flex@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/help-flex>,
	<mailto:help-flex-request@gnu.org?subject=subscribe>
List-Id: Users list for Flex,
	the GNU lexical analyser generator <help-flex.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/help-flex>,
	<mailto:help-flex-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/help-flex/>
Date: Thu, 26 Apr 2001 10:50:20 -0500

Hi,

First off I was wondering if there is an official web site for Flex?

Second, I would like to be able to have Flex use certain characters in a
regular expression to match the expression but suppress copying the
characters to yytext.  A simple example would be if I was trying to match
text surrounded by quotes, but I don't want to use the quotes. i.e.

\"[^\n]\"		{ yylval = strdup(yytext); //This gives me the text and quotes. }

or

\"[^\n]\"		{ yylval = strdup(yytext+1); yylval[strlen(yylval)-1] = '\0';
//Now I've removed the quotes. }

But since I'm lazy ;)  I would like to be able to tell Flex not to copy the
quotes.  What I would like is an operator that would tell flex to match an
expression but to not copy it to yytext.  Off the top of my head maybe the #
symbol could work for such an operator, so I could then do:

\"#[^\n]\"#	{ yylval = strdup(yytext); //This would give me the text without
the quotes. }

Does anyone know if such a feature or something similar exists?  If this
doesn't exists would it be practical?

Thanks,

Dave Dutcher
Software Engineer
Symbology, Inc.
Minneapolis, MN


_______________________________________________
Help-flex mailing list
Help-flex@gnu.org
http://mail.gnu.org/mailman/listinfo/help-flex


1,,
X-Coding-System: nil
Mail-from: From help-flex-admin@gnu.org  Thu Apr 26 12:26:26 2001
Return-Path: <help-flex-admin@gnu.org>
Received: from localhost (localhost [127.0.0.1])
	by michael.uncg.edu (8.9.3/8.9.3) with ESMTP id MAA25203
	for <wlestes@localhost>; Thu, 26 Apr 2001 12:26:26 -0400
Received: from imap.uncg.edu
	by localhost with IMAP (fetchmail-5.1.0)
	for wlestes@localhost (single-drop); Thu, 26 Apr 2001 12:26:26 -0400 (EDT)
Received: from external-gw.uncg.edu (external-gw.uncg.edu [152.13.2.70])
	by hermes.email.uncg.edu (8.11.0/8.11.0) with ESMTP id f3QG63e27511
	for <wlestes@hermes.email.uncg.edu>; Thu, 26 Apr 2001 12:06:03 -0400 (EDT)
Received: from fencepost.gnu.org (fencepost.gnu.org [199.232.76.164])
	by external-gw.uncg.edu (8.9.3/8.9.3) with ESMTP id MAA23192
	for <wlestes@uncg.edu>; Thu, 26 Apr 2001 12:06:02 -0400 (EDT)
Received: from localhost ([127.0.0.1] helo=fencepost.gnu.org)
	by fencepost.gnu.org with esmtp (Exim 3.16 #1 (Debian))
	id 14soHK-0004gM-00
	for <wlestes@uncg.edu>; Thu, 26 Apr 2001 12:06:02 -0400
Received: from acp3bf.physik.rwth-aachen.de ([137.226.32.75])
	by fencepost.gnu.org with smtp (Exim 3.16 #1 (Debian))
	id 14soFe-0004cw-00
	for <help-flex@gnu.org>; Thu, 26 Apr 2001 12:04:18 -0400
Received: from localhost (broeker@localhost) by acp3bf.physik.rwth-aachen.de (8.7.5/v3.2) with ESMTP id SAA00797; Thu, 26 Apr 2001 18:04:09 +0200
X-Authentication-Warning: acp3bf.physik.rwth-aachen.de: broeker owned process doing -bs
From: Hans-Bernhard Broeker <broeker@physik.rwth-aachen.de>
X-Sender: broeker@acp3bf
To: Dave Dutcher <ddutcher@symbology.com>
cc: help-flex@gnu.org
Subject: Re: Suppress copying characters to yytext?
In-Reply-To: <002501c0ce68$99b3ec40$4a00000a@symbology.com>
Message-ID: <Pine.LNX.4.10.10104261759540.592-100000@acp3bf>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: help-flex-admin@gnu.org
Errors-To: help-flex-admin@gnu.org
X-BeenThere: help-flex@gnu.org
X-Mailman-Version: 2.0.3
Precedence: bulk
List-Help: <mailto:help-flex-request@gnu.org?subject=help>
List-Post: <mailto:help-flex@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/help-flex>,
	<mailto:help-flex-request@gnu.org?subject=subscribe>
List-Id: Users list for Flex,
	the GNU lexical analyser generator <help-flex.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/help-flex>,
	<mailto:help-flex-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/help-flex/>
Date: Thu, 26 Apr 2001 18:04:06 +0200 (MET DST)

*** EOOH ***
From: Hans-Bernhard Broeker <broeker@physik.rwth-aachen.de>
To: Dave Dutcher <ddutcher@symbology.com>
cc: help-flex@gnu.org
Subject: Re: Suppress copying characters to yytext?
In-Reply-To: <002501c0ce68$99b3ec40$4a00000a@symbology.com>
Sender: help-flex-admin@gnu.org
Precedence: bulk
List-Help: <mailto:help-flex-request@gnu.org?subject=help>
List-Post: <mailto:help-flex@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/help-flex>,
	<mailto:help-flex-request@gnu.org?subject=subscribe>
List-Id: Users list for Flex,
	the GNU lexical analyser generator <help-flex.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/help-flex>,
	<mailto:help-flex-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/help-flex/>
Date: Thu, 26 Apr 2001 18:04:06 +0200 (MET DST)

On Thu, 26 Apr 2001, Dave Dutcher wrote:

> Second, I would like to be able to have Flex use certain characters in a
> regular expression to match the expression but suppress copying the
> characters to yytext.  

You can do that with a start condition and trailing context rules. I.e.
roughly like this:

<INITIAL>\"	BEGIN(IN_STRING)

<IN_STRING>[^\n]*/\"	{ 
			/* yytext is the interior of the string, now */
			BEGIN(INITIAL)
			}

-- 
Hans-Bernhard Broeker (broeker@physik.rwth-aachen.de)
Even if all the snow were burnt, ashes would remain.


_______________________________________________
Help-flex mailing list
Help-flex@gnu.org
http://mail.gnu.org/mailman/listinfo/help-flex


1, answered,,
X-Coding-System: nil
Mail-from: From johnmillaway@yahoo.com  Wed Apr 25 18:52:08 2001
Return-Path: <johnmillaway@yahoo.com>
Received: from localhost (localhost [127.0.0.1])
	by michael.uncg.edu (8.9.3/8.9.3) with ESMTP id SAA24103
	for <wlestes@localhost>; Wed, 25 Apr 2001 18:52:08 -0400
Received: from imap.uncg.edu
	by localhost with IMAP (fetchmail-5.1.0)
	for wlestes@localhost (single-drop); Wed, 25 Apr 2001 18:52:08 -0400 (EDT)
Received: from external-gw.uncg.edu (external-gw.uncg.edu [152.13.2.70])
	by hermes.email.uncg.edu (8.11.0/8.11.0) with ESMTP id f3PMWQe07299
	for <wlestes@hermes.email.uncg.edu>; Wed, 25 Apr 2001 18:32:26 -0400 (EDT)
Received: from web9404.mail.yahoo.com (web9404.mail.yahoo.com [216.136.129.110])
	by external-gw.uncg.edu (8.9.3/8.9.3) with SMTP id SAA23019
	for <wlestes@uncg.edu>; Wed, 25 Apr 2001 18:32:24 -0400 (EDT)
Message-ID: <20010425223059.65780.qmail@web9404.mail.yahoo.com>
Received: from [216.254.82.102] by web9404.mail.yahoo.com; Wed, 25 Apr 2001 15:30:59 PDT
Date: Wed, 25 Apr 2001 15:30:59 -0700 (PDT)
From: "John W. Millaway" <johnmillaway@yahoo.com>
Subject: Re: Bug: %option main overrides %option yywrap
To: "W. L. Estes" <wlestes@uncg.edu>, vern@ee.lbl.gov
Cc: help-flex@gnu.org
In-Reply-To: <200104252027.QAA23947@michael.uncg.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii

*** EOOH ***
Date: Wed, 25 Apr 2001 15:30:59 -0700 (PDT)
From: "John W. Millaway" <johnmillaway@yahoo.com>
Subject: Re: Bug: %option main overrides %option yywrap
To: "W. L. Estes" <wlestes@uncg.edu>, vern@ee.lbl.gov
Cc: help-flex@gnu.org
In-Reply-To: <200104252027.QAA23947@michael.uncg.edu>

> Additionally, flex.1 says the following:
> 
>        main   directs  flex  to  provide a default main() program
>               for the scanner, which simply calls yylex().   This
>               option implies noyywrap (see below).

I understand, and it makes sense. Just be aware that the inverse of this rule is
also true --  %option nomain implies yywrap -- which makes much less sense.

Of course, I only complain because it had me on one heck of a bug hunt. I had
specified %option noyywrap and %option nomain, yet the linker complained that it
couldn't find yywrap. I thought I had broken something! :)

-John


__________________________________________________
Do You Yahoo!?
Yahoo! Auctions - buy the things you want at great prices
http://auctions.yahoo.com/