home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: gnu.gcc.bug
- Path: sparky!uunet!cis.ohio-state.edu!feynman.t30.physik.tu-muenchen.de!lauten
- From: lauten@feynman.t30.physik.tu-muenchen.de (Markus Lautenbacher)
- Subject: gcc code dumpes core, other compilers do not
- Message-ID: <199208120757.AA18156@feynman.t30.physik.tu-muenchen.de>
- Sender: gnulists@ai.mit.edu
- Reply-To: lauten@feynman.t30.physik.tu-muenchen.de
- Organization: GNUs Not Usenet
- Distribution: gnu
- Date: Wed, 12 Aug 1992 11:57:32 GMT
- Approved: bug-gcc@prep.ai.mit.edu
- Lines: 49
-
- Dear Sir,
-
- whether this really is a bug report or not I am not a computer expert
- enough to decide. So let me just bring the following to your attention.
-
- I am writing right now a C program to get some theoretical predictions
- for an experimentally measured quantity in particle physics.
- Numerically this basically means to solve for the eigensystem of real,
- non-symmetric matrices. Compiling my C-sources (basically K&R I style)
- with various compilers, in particular GNU gcc, I have run into the
- following problem:
-
- --------------------------------------------------------
- Compiler | Flags
- | default | -g | -O | -g-O
- --------------------------------------------------------
- Turbo-C 1.0 | OK | | |
- cc | | OK | OK |
- gcc-1.40 | | OK | OK |
- gcc-2.1 | | OK | CD | CD
- gcc-2.2.2 | | CD | CD | CD
-
- where OK means "compiles and links fine, runs and produces sensible
- data" and CD means "compiles and links fine, but dumps core".
- The workstation I am using is a Sun SPRCstation 1+ running SunOS 4.1.2.
- The PC with the Turbo-C is a standard AT 286/7 clone under DOS 3.3.
-
- I have found out that the problem is in a routine I have copied from
- the program-listing disk of the wellknown "Numerical Recipes" book. So
- it seems unlikely that I have made a programming error that now causes
- the core dump. Strange enough the problem has come about with the
- major release change 1.* -> 2.* and there first only in optimized code,
- but now with the latest gcc in any kind of code.
-
- Do you want me to send you my sources (in *.tar.Z form) to have a look
- at the problem ? If you're interested let me know (please not that you
- probably won't be able to reach me from 22/8/92 until 22/9/92 since I
- am abroad for a research visit during that period).
-
-
- Best regards,
- MARKUS
-
- -----
- MARKUS E. LAUTENBACHER,
- Technical University Munich, | phone: +49/89/3209-2398
- Physics Dept.-Theo. Physics T31,| Internet:
- D-8046 Garching, FRG. | lauten@feynman.t30.physik.tu-muenchen.de
-
-