Date: Mon, 21 Apr 1997 10:07:53 +0300 (IDT)
From: Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
To: Orlando Andico <orly AT gibson DOT eee DOT upd DOT edu DOT ph>
cc: Chris Croughton <crough45 AT amc DOT de>, djgpp AT delorie DOT com
Subject: Re: CWSDPMI problem
In-Reply-To: <Pine.SGI.3.93.970421094839.4139C-100000@gibson.eee.upd.edu.ph>
Message-ID: <Pine.SUN.3.91.970421100728.2883C-100000@is>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Precedence: bulk


On Mon, 21 Apr 1997, Orlando Andico wrote:

> > Did you try to use QDPMI?  I don't know if it has such problems (I
> > don't use DesqView), but given that it is from the same shop, it might
> > work better.
> 
> Err, I seem to have read somewhere that as DPMI hosts go, QDPMI
> sucks.

That's changed with the latest QDPMI versions (from QEMM 8).  The only
problem with that one is in nested DJGPP programs: it aborts both the
child *and* the parent(s) when the child gets Ctrl-C.  Seems to me
like a lesser evil than to have a sharing violation which effectively
prohibits multi-processing in the case in point.

> if there's no QDPMI, DV will use VCPI, which breaks CWSDPMI quite
> nicely.. :) 

Huh?  I use CWSDPMI in VCPI mode all the time, and it never did any
trouble.  In fact, almost every DJGPP user runs CWSDPMI in VCPI mode,
if they have EMM386 or QEMM386 installed.