Xref:	news2.mv.net comp.os.msdos.djgpp:2809
From:	ao950 AT FreeNet DOT Carleton DOT CA (Paul Derbyshire)
Newsgroups:	comp.os.msdos.djgpp
Subject:	Re: malloc crash
Date:	17 Apr 1996 01:23:07 GMT
Organization:	The National Capital FreeNet
Lines:	30
Sender:	ao950 AT freenet2 DOT carleton DOT ca (Paul Derbyshire)
Message-ID:	<4l1h5r$o7v@freenet-news.carleton.ca>
References:	<Pine DOT SUN DOT 3 DOT 91 DOT 960415130304 DOT 4189J-100000 AT is>
Reply-To:	ao950 AT FreeNet DOT Carleton DOT CA (Paul Derbyshire)
NNTP-Posting-Host:	freenet2.carleton.ca
To:	djgpp AT delorie DOT com
DJ-Gateway:	from newsgroup comp.os.msdos.djgpp

Eli Zaretskii (eliz AT is DOT elta DOT co DOT il) writes:
> On 14 Apr 1996, Paul Derbyshire wrote:
> 
>> My program can't be causing it that way: it uses realloc to replace a
>> smaller chunk with a larger chunk.
> 
> Yes, it can.  Each time you realloc, a different region might be 
> allocated by the library `sbrk'.

Yeah, but when this happens the old region is deallocated. So
multiply-reallocked blocks still mean that CWSDPMI is keeping track
of one block at any given time... so this can't contribute to its running
out of handles or whatever they are for tracking allocated blocks.

>> In any case, I don't think I have this CWSDPMI thing... or is it part of
>> djgpp?
> 
> You should have CWSDPMI.EXE in your bin/ subdirectory.  If you do, and 
> don't run under Windows or QDPMI, CWSDPMI is loaded by the startup code 
> of every DJGPP program (in v2).

I have 1.2, and regular GO32, so the whole thing is moot anyways. I'm not
using CWSDPMI...

--
    .*.  "Clouds are not spheres, mountains are not cones, coastlines are not
 -()  <  circles, and bark is not smooth, nor does lightning travel in a
    `*'  straight line."    ,------------------------------------------------
         -- B. Mandelbrot  |  Paul Derbyshire (PGD) ao950 AT freenet DOT carleton DOT ca

- Raw text -