Sender: g595160 AT fs1 DOT mar DOT lmco DOT com
Message-Id: <3199C1F7.1E3B@fs1.mar.lmco.com>
Date: Wed, 15 May 1996 07:37:27 -0400
From: Shawn Lee <slee AT fs1 DOT mar DOT lmco DOT com>
Organization: Lockheed Martin Aeronautical Systems
Mime-Version: 1.0
To: The DJGPP Mailing List <djgpp AT delorie DOT com>
Subject: Re: djgpp 2 DPMI
References: <Pine DOT SGI DOT 3 DOT 91 DOT 960514115828 DOT 22385B-100000 AT tower DOT york DOT ac DOT uk>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

Shawn Hargreaves wrote:
> 
> On Tue, 14 May 1996, Eli Zaretskii wrote:
> 
> > > >   What is the progress of making (programs compiled and assembled and
> > > > linked by djgpp v2) fully portable and independent of any faults or
> > > > absence of whatever DPMI is in the PC that it is run on? One way might
> > >
> > > I think you are missing the point. It's not possible to avoid the DPMI
> > > on the machine. That is the whole reason DPMI was required in the first
> >
> > I think what Anthony meant was a possibility to bind a DPMI host with the
> > program, so you'd get a stand-alone executable that serves itself, so to
> > speak.
> 
> That would cover "absence of DPMI", but what about "faults"? Binding a
> DPMI host into the program is no good if the machine is already running a
> buggy DPMI: you can't load up and use cwsdpmi in a win95 DOS box, for
> example.
> 
> I don't think binding a DPMI host into the executable would be a
> particularly useful thing to do, in any case. The process of loading
> cwsdpmi.exe is totally automatic, and end users don't need to know about
> it. Standalone exe's are kind of appealing, but most programs have
> other support files of their own. I've given lots of djgpp-compiled
> programs to computer illiterate people, and none of them have run into
> the slightest trouble with DPMI issues...
> 

Is that true?  You can't run cwsdpmi under a Win95 Dos window?  Why
would I need to keep cwsdpmi?  If I write a robust code, any host would
do, right?