From: charles AT homer DOT pentek DOT com (Charles Krug)
Newsgroups: comp.os.msdos.djgpp
Subject: Re: Runtime mystery
References: <OE2s4nLLW94lJ7tUUQD00003408 AT hotmail DOT com> <003001c0435b$3e965b60$1e806f83 AT quns DOT cam DOT ac DOT uk> <OE34F58tP9DLf4bg5Xm000008cc AT hotmail DOT com>
Message-ID: <slrn900s3e.lf5.charles@homer.pentek.com>
User-Agent: slrn/0.9.6.3 (SunOS)
NNTP-Posting-Host: mail.pentek.com
Date: 1 Nov 2000 14:35:11 -0500
X-Trace: 1 Nov 2000 14:35:11 -0500, mail.pentek.com
Lines: 25
To: djgpp AT delorie DOT com
DJ-Gateway: from newsgroup comp.os.msdos.djgpp
Reply-To: djgpp AT delorie DOT com

On Wed, 1 Nov 2000 11:24:01 -0600, Tom Hutto <huttothomas AT hotmail DOT com> wrote:
>The source is proprietary.  I can't show it.  I would have to cobble
>something that duplicates the problem.  Kind of makes it tough to ask
>intellegent questions :-( 

Um ya

> Furthermore, if I could duplicate the problem
>in a small program then I could fix it myself, nicht war?
>

Not really.  That's often the case anyway--in the course of creating the
"smallest buildable code that consistantly reproduces the problem," I usually
find that I've botched something.

It really does two things.  First, it gets me thinking about my code and how
things work and why writing to that uninitialized pointer is an
extraordinarily bad idea . . .

Second, if it turns out that everything is good, it isolates the fault from a
billion other possibilities that are equally likely.  That way, when I post it
on the NG, someone goes, "Oh well of course, this works like . . " fairly
quickly.

Charles