From: sandmann AT clio DOT rice DOT edu (Charles Sandmann)
Message-Id: <9710151551.AA15725@clio.rice.edu>
Subject: Re: cwsdpmi r4 beta
To: eliz AT is DOT elta DOT co DOT il (Eli Zaretskii)
Date: Wed, 15 Oct 1997 10:50:59 -0600 (CDT)
Cc: billc AT blackmagic DOT tait DOT co DOT nz, djgpp-workers AT delorie DOT com
In-Reply-To: <Pine.SUN.3.91.971015125149.17868B-100000@is> from "Eli Zaretskii" at Oct 15, 97 12:56:52 pm
Content-Type: text
Precedence: bulk

> I suggest to make two (three, counting CWSDPR0) binaries, and only have 
> the VDS support in one of them, even when it is fully tested.  That way, 
> only those programs/users which need DMA will load the VDS-supporting 
> host or stubedit the .exe to load it.  It seems reasonable, since VDS 
> support probably enlarges the footprint, and many programs, particularly 
> those which run deeply-nested children, like Make, don't need DMA at all.

The way I would plan to work it, the overhead would only exist if you used
it, so only the 3Kb of code would be extra.  In that case, when it's fully
tested it would go into the binaries and not be much of an issue.