Date:	Fri, 7 Oct 94 16:21:31 JST
From:	Stephen Turnbull <turnbull AT shako DOT sk DOT tsukuba DOT ac DOT jp>
To:	Jeremy_Norbury DOT LAX1B AT xerox DOT com
Cc:	djgpp AT sun DOT soe DOT clarkson DOT edu
Subject:	DJGPP flex247.exe vs flex238.exe - generated lexer hanging.

   From: Jeremy_Norbury DOT LAX1B AT xerox DOT com

   I have a flex/yacc parser which I compile and run both on an RS/6000
   AIX 3.2.5 and MSDOS. I use the pre-built MSDOS version in DJGPP (
   flx2xxbn.zip ) for testing at home. About a month or so ago I went from
   flex 2.3.8 to 2.4.7 simply because it was the newest version.  Everything

You rebuilt flex with DJGPP?  Or you're using flex on the RS/6000 and
taking the lexyy.c home?  Is the lexyy.c the same on both systems if
you're using both?  If so, this isn't a DJGPP problem.

   went OK, ran flex on the parser.l, compiled and it ran fine under AIX.
   Yesterday, went to do this at home, using flex247.exe and the generated
   program, when compiled same as usual with Borland C++ 3.0 (large memory
   model) hangs my machine dead. 
   I switched back to flex238.exe, everything OK.
   I compared the lexyy.c's generated and there are, obviously,
   differences. Nothing radical though.

I would say a successful run vs "hangs my machine dead" is a radical
difference.  Probably Borland C++ is broken.  DJGPP and flex certainly
can't be :-)

   The parser.l file is 315 lines BTW, hardly huge.
   Does this ring any bells with anyone ? Anyone with the same/similar
   problem ? 
   It should be noted that I can generate smaller flex programs
   without problems. This suggests some sort of memory problem.

No, this suggests that your smaller flex programs don't have enough
code to exercise the particular arcane bug in the flex algorithm or
(more likely IMHO) the 2.4.7 implementation.  Where are the C code
diffs!!  That's what the problem is.  (I am, obviously, excluding
uninitialized and other bogus pointer behavior from "memory problems."
But if that's what you meant by memory problems, you shoulda sent the
diffs.)
    Read the GNU and DJGPP bug reporting instructions; next time I may
be able to help you, instead of flame you.

-- 
+-----------------------------------------------------------------------+
|                           Stephen Turnbull                            |
|     University of Tsukuba, Institute of Socio-Economic Planning       |
|          Tennodai 1-chome 1--1, Tsukuba, Ibaraki 305 JAPAN            |
|        Phone:  +81 (298) 53-5091     Fax:  +81 (298) 55-3849          |
|               Email:  turnbull AT shako DOT sk DOT tsukuba DOT ac DOT jp                 |
|                                                                       |
|                Founder and CEO, Skinny Boy Associates                 |
|               Mechanism Design and Social Engineering                 |
| REAL solutions to REAL problems of REAL people in REAL time!  REALLY. |
|                      Phone:  +81 (298) 56-2703                        |
+-----------------------------------------------------------------------+

- Raw text -