From: Hans-Bernhard Broeker <broeker AT physik DOT rwth-aachen DOT de>
Newsgroups: comp.os.msdos.djgpp
Subject: Re: silly question: SIGSEGV
Date: 22 Jan 2001 11:39:21 GMT
Organization: Aachen University of Technology (RWTH)
Lines: 44
Message-ID: <94h659$gnf$1@nets3.rz.RWTH-Aachen.DE>
References: <Pine DOT SUN DOT 3 DOT 91 DOT 1010121111642 DOT 13920S-100000 AT is> <3a6ab865$0$25126 AT SSP1NO25 DOT highway DOT telekom DOT at>
NNTP-Posting-Host: acp3bf.physik.rwth-aachen.de
X-Trace: nets3.rz.RWTH-Aachen.DE 980163561 17135 137.226.32.75 (22 Jan 2001 11:39:21 GMT)
X-Complaints-To: abuse AT rwth-aachen DOT de
NNTP-Posting-Date: 22 Jan 2001 11:39:21 GMT
Originator: broeker@
To: djgpp AT delorie DOT com
DJ-Gateway: from newsgroup comp.os.msdos.djgpp
Reply-To: djgpp AT delorie DOT com

In addition to the hints given by Eli already, let me point at a few
more issues that jumped to my eye:

Florian X <dos DOT fire AT aon DOT at> wrote:

> Call frame traceback EIPs:
>   0x000a9b9f _free+247
>   0x000016c4 _sf_free+24, line 81 of program.c
>   0x00003d38 _dispose+52, line 1478 of program.c
>   0xffd94c14 0xffd94c14   <------------- funtktion in the DLX program
>   0x00003d23 _dispose+31, line 1476 of program.c

This *may* be part of your problem. Note that dispose() is being
called recursively, with only that slightly volatile 'DLX program' in
between. Iff that routine uses any kind of non-volatile storage in a
way DLX doesn't expect, this could get you in trouble.

> I used it, but they couldn't show me more infos, because of the DLXs
> and because some *.c files are included into the main file program.c
> via fe. #include"object.c".

Two solutions to that:

1) use -gstabs+ or -ggdb instead of -g in GCC

2) Don't #include *.c files into one another. It's bad style, anyway.

> Yes, I tried it. I think it tries to remove a memory block which wasn't
> right allocated via malloc.

Or which got its internal bookkeeping datastructures damaged. For this
type of problem, YAMD would be my recommendation. If it's possible to
run your app under YAMD, it'll find the bug much more quickly than you
can do it yourself using code-snippeting. It'll crash your app exactly
at the point where it tries overwrites crucial stuff.

> All free() commands where replaced by a macro like this:

That macro does nothing useful on DJGPP. You might want to get rid of
it.

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