Date: Sat, 28 Dec 1996 21:57:18 -0500
Message-Id: <199612290257.VAA05915@delorie.com>
From: DJ Delorie <dj AT delorie DOT com>
To: grendel AT ananke DOT amu DOT edu DOT pl
CC: djgpp AT delorie DOT com
In-reply-to: <199612282157.WAA10028@math.amu.edu.pl>
	(grendel AT ananke DOT amu DOT edu DOT pl)
Subject: Re: DPMI incorporation...


> True. I must've overlooked this statement. Nevertheless, I have never 
> seen any successful implementation of DPMI for 286 machines (perhaps 
> Windoze 3.xx was such?). Even if there were a server for 286 it would 
> provide only a limited set of DPMI 0.9 functions.There is no 
> provision in 286 to implement any of the page-related functions, nor the 
> debbuging-related ones.

Borland's programs are 286-dpmi.  *All* 16-bit programs *today* are
286, because nobody (but us and 386max) support DPMI 1.0.

> My fault. I used a wrong expression. What I meant was that DPMI uses 
> i386 memory pages to implement virtual memory. This is opposed to the 286 
> VM scheme where memory is swapped out in segments, not pages.

DPMI 0.9 (Windows, OS/2, QEMM) does NOT support the paging functions.
There is nothing in the DPMI 0.9 API that suggests that paging must be
supported, or that virtual memory will be available.