From: kunst AT prl DOT philips DOT nl
Subject: Re: OEmacs 4.1 and DPMI bugs (???)
To: darrylo AT sr DOT hp DOT com (Darryl Okahata)
Date: Mon, 9 May 1994 15:47:15 +0100 (METDST)
Cc: djgpp AT sun DOT soe DOT clarkson DOT edu (DJGPP users list)


Darryl Okahata (darrylo AT sr DOT hp DOT com) writes:

> [stuff deleted]
> 
>      I've got OEmacs undumpable under DJGPP 1.11, but it's been modified
> to use DJGPP's DPMI support (so that a special GO32.EXE is not needed),
> and, unfortunately, 

> DJGPP's DPMI support is currently too buggy to use

How do you mean 'DPMI support is too buggy' ?

I have been using GCC and running GCC compiled programs
in Windows DOS boxes for several months with little or no trouble.

The only 'buggy' part is that DPMI memory is not always properly
released in the case of a program crash, which can be verified
using a Windows memory allocation monitor.

If you know of any more DPMI bugs, please share your knowledge
with the readers of this list. Perhaps we can help to fix them!

> with OEmacs.  I've been meaning to back out the changes (i.e., again
> require a special GO32.EXE), but I haven't had the time.

Wouldn't it be more useful for the long term if you or some other
knowledgeable OEmacs user tried to fix things that are broken in 
GO32/DJGPP instead of providing your own version of GO32 each time ?

Pieter Kunst (kunst AT prl DOT philips DOT nl)