Date: Sun, 20 May 2001 17:46:48 +0300 (IDT)
From: Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
X-Sender: eliz AT is
To: Martin Str|mberg <ams AT ludd DOT luth DOT se>
cc: djgpp-workers AT delorie DOT com
Subject: Re: realloc enhancement
In-Reply-To: <200105201203.OAA21918@mother.ludd.luth.se>
Message-ID: <Pine.SUN.3.91.1010520174407.23429C-100000@is>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Reply-To: djgpp-workers AT delorie DOT com
Errors-To: nobody AT delorie DOT com
X-Mailing-List: djgpp-workers AT delorie DOT com
X-Unsubscribes-To: listserv AT delorie DOT com
Precedence: bulk


On Sun, 20 May 2001, Martin Str|mberg wrote:

> According to Mark E.:
> > I found and fixed the problem in _write.c:_write_int that fouled up 
> > profiling. Turns out the transfer buffer was being overfilled with data.
> 
> How did you trace it down?

In addition to what Mark described (which is a technique I use myself 
when hunting down similar problems), another clue in the same direction 
was the info you, Martin, posted: if COMMAND.COM is unable to regain 
control after a program exits, it usually means that something overwrote 
DOS memory chain or DOS itself.  So any code that uses conventional 
memory becomes the immediate suspect.