From:	Linley Henzell <zel AT olis DOT net DOT au>
Newsgroups:	comp.os.msdos.djgpp
Subject:	Re: DJGPP vs Borland C++
Date:	Mon, 27 Jan 1997 22:47:10 -0800
Organization:	Microtronics Information Systems
Lines:	47
Message-ID:	<32EDA0EE.676F@olis.net.au>
References:	<5c1qik$9sh AT lion DOT cs DOT latrobe DOT edu DOT au> <32E98087 DOT 6A50 AT cs DOT com> <32ECF799 DOT 74EF AT olis DOT net DOT au> <5ci12n$fbd AT star DOT cs DOT vu DOT nl>
Reply-To:	zel AT olis DOT net DOT au
NNTP-Posting-Host:	ppp07.olis.net.au
Mime-Version:	1.0
To:	djgpp AT delorie DOT com
DJ-Gateway:	from newsgroup comp.os.msdos.djgpp

Ruiter de M wrote:
> 
> Linley Henzell (zel AT olis DOT net DOT au) wrote:
> : Well, I'm having some problems with a C program written with Borland C++
> : 3.1 which refuses to work with DJGPP v2. It compiles fine, but every
> : time I run it, it aborts with a SIGSEGV GPF. I have no idea why this
> : happens, because it seems to choose a completely arbitrary point at
> : which to abort (I checked with Symify), and my program contains nothing
> : which I can imagine being objectionable (like pointers or dynamic memory
> : allocation).
> : Meanwhile, a little test program works okay with DJGPP.
> :
> : Borland's compiler has no problems whatsoever with it, and I've rarely
> : had any problems with BC++. If anybody knows about something which
> : Borland C++ 3.1 copes with, but which DJGPP just doesn't like, I would
> : really appreciate them letting me know so that I can remove it from my
> : program! Any assistance in this problem would be greatly appreciated.
> 
> Assure you compile with -Wall (and -W) and you have the newest DJGPP-
> version 2.01. Several bugs have been fixed since the first v2.
> I personally have experienced problems like these, also with gcc on UNIX.
> Somewhere I did 'var = 0;' and the program crashed... But it was because
> of an error I made and I didn't notice the warnings about it.
> 
> Maybe you could post some code? I bet someone will be able to help you.
> 
> I am sure every compiler has its bugs, but I have found gcc and DJGPP to be
> very stable.
> --

I eventually found the problem by hacking chunks of code from my source
until it worked. Because the program crashed at apparently innocuous
places, the printf debugging method (which I usually use) wasn't much
help, but I finally managed to isolate my problem to a call to atoi. For
some reason, djgpp just doesn't seem to like my programs calling this
function. And I am using v2; is this one of its documented bugs? 

Sometimes, although only sometimes, symify showed the GPF to have
ocurred at a call to strtol, which I never use. Because this sounded a
bit like atoi (it means "string to long" or something like that), I
removed the atoi and it worked almost fine. I still have to cope with a
few minor irritations, like the immense volume of warnings which djgpp
generates (even without -Wall), and the doubling of my .exe file from
~150K to over 300K, but I can cope with that.

Thanks for the help,
Linley Henzell

- Raw text -