Date: Sun, 6 Dec 1998 10:44:28 +0200 (IST)
From: Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
X-Sender: eliz AT is
To: Jason <jason AT datek DOT com DOT my>
cc: djgpp AT delorie DOT com
Subject: Re: Page fault on 486
In-Reply-To: <3668B30E.C8F35DA9@datek.com.my>
Message-ID: <Pine.SUN.3.91.981206104407.28436J-100000@is>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Reply-To: djgpp AT delorie DOT com


On Sat, 5 Dec 1998, Jason wrote:

> I developed an application in C++ using my Pentium machine but it was
> meant to be implemented on a 486 with 4Mb RAM, MSDOS ver 5. Somehow,
> when implemented on the 486, it just crashes at some point in the
> application complaining about page fault, etc, etc. It was all fine
> running on my Pentium Win 95 machine.

Most probably, your program dereferences a null pointer, and Windows
95 just lets it get away with that.  CWSDPMI catches such bugs and
aborts your program, and the traceback pinpoints exactly where that
dereference happened.

> I'm thinking if it could be because of the cwsdmpi server which I need
> to run along with my application. How do I run my completed application
> without having to use the cwsdpmi server?

You are trying to kill the messenger instead of finding the bug in
your program.  I suggest that you debug the program near the locus
reported in the traceback (see section 12.2 of the FAQ, for some clues
and techniques) and find and fix the bug, instead of sweeping it under
the carpet ;-).