Date: Mon, 2 Feb 1998 19:14:26 +0200 (IST)
From: Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
To: Shawn Hargreaves <Shawn AT talula DOT demon DOT co DOT uk>
cc: djgpp AT delorie DOT com
Subject: Re: Physical memory, virtual memory. DJGPP+CWSDPMI, how they work?
In-Reply-To: <75hd6DAmrM10EwlZ@talula.demon.co.uk>
Message-ID: <Pine.SUN.3.91.980202191127.19322U-100000@is>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Precedence: bulk


On Sun, 1 Feb 1998, Shawn Hargreaves wrote:

> >This is not entirely accurate.  There *is* a bit in the selector
> >descriptor which says whether the segment is code or data.
>
> True, but that doesn't help with a djgpp program where the code and data
> segments both cover the entire address space of your program!

Well, I suppose the DPMI host could watch addresses referenced with the 
CS selector and assume that at least the 4KB page around that is code.

But this is far-fetched; I never meant to contradict your original 
comment on this.

> The problem is that even if someone spent the time to upgrade CWSDPMI to
> a full DPMI 1.0 implementation, it would be pretty much useless because
> we still wouldn't have these features under win95, and I don't think
> many people these days are willing to write programs that will only work
> in a clean DOS environment!

I agree.

> What do you think it would it take to make Microsoft start supporting
> DPMI 1.0? It is ridiculous that they are still only using version 0.9 of
> an official standard like this, especially since the real thing offers
> so many major improvements, but I find it hard to imagine them taking
> much notice of our complaints :-)

I don't think there's anything that can persuade them, either.  The
DPMI server built into NT is a clear evidence, since it is even worse
than the one in Windows 95.