Date: Sun, 28 Jul 1996 08:28:16 +0200 (IST)
From: Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
To: Charles Sandmann <sandmann AT clio DOT rice DOT edu>
Cc: djgpp AT delorie DOT com, Skip Collins <collins AT aplcomm DOT jhuapl DOT edu>
Subject: Re: emacs problem under WfWG 3.11
In-Reply-To: <31f8cd4d.sandmann@clio.rice.edu>
Message-Id: <Pine.SUN.3.91.960728081949.23430D-100000@is>
Mime-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII


On Fri, 26 Jul 1996, Charles Sandmann wrote:

> it probably doesn't get tested as much), or that Windows with Win32s is
> broken (the re-allocate dpmi call is not used by most other extenders so
> dpmi providers tend to break it and not notice) or that the failure at
> this point is an artifact of running under the debugger.
> It could also be that hooking a hardware interrupt (like the keyboard)
> with the unixy sbrk() causes problems in this environment if the arena
> moves on you.

Based on what Charles says, I'd suggest trying to build Emacs with the 
default sbrk behavior.  Just edit src/msdos.c and remove the 
_CRT0_FLAG_UNIX_SBRK bit from the initialization of the variable
`__crt0_startup_flags' (it's at the beginning of the file).  Warning: I 
don't think I tested Emacs thoroughly with the non-unix sbrk, so please 
tell me if you see anything worth attention.

>  We really need
> the register listing at the crash point to figure out what's going on.

Charles, I think he will have hard time providing the registers, because 
as far as I understand, Emacs doesn't crash, it gets aborted by Windows.  
Windows just pops that box telling "This application has violated system 
integrity" etc.  Is there any procedure to get registers' listing in such 
cases?