From:	"John M. Aldrich" <fighteer AT cs DOT com>
Newsgroups:	comp.os.msdos.djgpp
Subject:	Re: Win95 DPMI/CWSDPMI problems
Date:	Mon, 14 Jul 1997 22:19:48 +0000
Organization:	Two pounds of chaos and a pinch of salt
Lines:	28
Message-ID:	<33CAA604.2D2F@cs.com>
References:	<5qe2tj$khu$1 AT news DOT utdallas DOT edu>
Reply-To:	fighteer AT cs DOT com
NNTP-Posting-Host:	ppp202.cs.com
Mime-Version:	1.0
To:	djgpp AT delorie DOT com
DJ-Gateway:	from newsgroup comp.os.msdos.djgpp

David G Farrell wrote:
> 
> this. I've tried to get it to crash under Win95 but it won't, ironically. Is
> there anything that a Win95 DOS box will let you get away with that
> CWSDPMI won't?

In fact, yes.  CWSDPMI supports certain features of the DPMI 1.0
specification that Windows 95 does not, specifically, the linear
framebuffer and NULL page protection.  It is the latter that is probably
tripping you:  CWSDPMI is detecting that you are trying to dereference a
pointer to an invalid address and causes a page fault as a result.  This
is a problem that should be corrected, even if it doesn't lead to
aberrant program behavior.

The best I can recommend is to save the stack traceback that is
displayed when your program crashes and run the 'symify' program on it. 
This will tell you the function, file, and line number where it crashed,
and you can use this to deduce the cause of the crash.  If you need more
complex debugging, try using the built in debugger of RHIDE, RHGDB.  It
supports debugging of graphics programs.

-- 
---------------------------------------------------------------------
|      John M. Aldrich       | "Always listen to experts.  They'll  |
|       aka Fighteer I       | tell you what can't be done, and why.|
|   mailto:fighteer AT cs DOT com   | Then do it."                         |
| http://www.cs.com/fighteer |                - Lazarus Long        |
---------------------------------------------------------------------

- Raw text -