From:	"Michael Allison" <michael DOT allison2 AT sympatico DOT ca>
Newsgroups:	comp.os.msdos.djgpp
References:	<Pine DOT SUN DOT 3 DOT 91 DOT 1010226084225 DOT 21073C-100000 AT is>
Subject:	Re: Eradicating djgpp W2000 problem
Lines:	47
X-Priority:	3
X-MSMail-Priority:	Normal
X-Newsreader:	Microsoft Outlook Express 5.50.4133.2400
X-MimeOLE:	Produced By Microsoft MimeOLE V5.50.4133.2400
Message-ID:	<f8sm6.234740$Pm2.3776410@news20.bellglobal.com>
Date:	Mon, 26 Feb 2001 12:39:39 GMT
NNTP-Posting-Host:	216.209.119.164
X-Trace:	news20.bellglobal.com 983191179 216.209.119.164 (Mon, 26 Feb 2001 07:39:39 EST)
NNTP-Posting-Date:	Mon, 26 Feb 2001 07:39:39 EST
Organization:	Sympatico
To:	djgpp AT delorie DOT com
DJ-Gateway:	from newsgroup comp.os.msdos.djgpp
Reply-To:	djgpp AT delorie DOT com

"Eli Zaretskii" <eliz AT is DOT elta DOT co DOT il> wrote in message
news:Pine DOT SUN DOT 3 DOT 91 DOT 1010226084225 DOT 21073C-100000 AT is...
> One possible reason for this is that the child program(s) actually
> need to do something non-trivial to reveal the bug.  I don't know what
> that is, but the real problems happen with Make and GCC which are
> highly non-trivial programs.
>
> I suggest to stop wasting time on this and start with a scenario where
> the crashes are reproducible.

You will need to be able to produce a small sample that reproduces the
problem if you want to file a report with Microsoft to have them
actually fix it, which is what started this thread.  I don't think you're
going to get much action from them if you tell them to download gcc and
GNU make and then start compiling some non-trivial application.

As far as using Make itself, has anyone been able to identify the
last Make source line that was executing when the NTVDM crash took
place?

> If you look at the report produced by Dr Watson, you will clearly see
> that it's NTVDM's own code which crashes, and it looks like it is
> using the application's stack to do something when it crashes.  Using
> an application's stack is an absolute no-no for system-level software,
> because that stack might be invalid at any given moment.

I'm not disputing that the problem is due to a bug in NTVDM, it is. I
am, however, interested in what is required to set it off.  I'm also
interested in getting it reported once the circumstances are understood.
If the circumstances are indeed understood, it should in fact be
possible to create a test sample to pass on to the vendor, wouldn't
you agree?

Have all of the following reported NTVDM problems been ruled out:

http://support.microsoft.com/support/kb/articles/Q265/8/01.ASP

http://support.microsoft.com/support/kb/articles/Q264/3/20.ASP

http://support.microsoft.com/support/kb/articles/Q255/5/70.ASP

http://support.microsoft.com/support/kb/articles/Q156/6/87.asp

- Raw text -