From: "Charles Sandmann" <sandmann AT clio DOT rice DOT edu>
Newsgroups: comp.os.msdos.djgpp
Subject: Re: CWSDPMI problem with Allegro Was:Problems with Allegro (followup)
Date: Thu, 15 May 1997 22:42:20
Organization: Aspen Technology, Inc.
Lines: 22
Message-ID: <337b914c.sandmann@clio.rice.edu>
References: <Pine DOT SUN DOT 3 DOT 91 DOT 970515191128 DOT 10056K-100000 AT is>
Reply-To: sandmann AT clio DOT rice DOT edu
NNTP-Posting-Host: spica.dmccorp.com
To: djgpp AT delorie DOT com
DJ-Gateway: from newsgroup comp.os.msdos.djgpp
Precedence: bulk

> If you do think that CWSDPMI has a bug, I suggest to write to Charles 
> Sandmann <sandmann AT clio DOT rice DOT edu> and tell him about the problem (please 
> also include the crash traceback with your message).  Charles will 
> probably be able to give you some hints about what might be the reason.

When submitting a bug report, I require:
1) The program be trimmed to the absolute minimum which shows the error.
   Any source which is over 100 lines will be ignored - I just don't have
   the time to diagnose such problems anymore, sorry.
2) It needs to be repeatable on my machine in a standard DOS environment.
3) It needs to run by itself (not under some other program) or include a
   very simple nesting image (same 100 line rule).

While this sounds very harsh, I've found 99.9% of the supposed CWSDPMI bugs 
to be caused by bugs in the original code or in something loaded at the 
same time.  Usually when trimming, you'll find the source of the problem.
If it's a small program, it's easy for me to diagnose if there is DPMI rule
violation, or if it's a real bug.  If it's a real bug, the small source is
much faster to load, transfer, and analyze.

Even if you follow the rules and find a bug - no guarantees on a timely fix.
But the more simple the code, the higher the odds :-)