Message-ID: <19990201113120.A28031@mc31.merton.ox.ac.uk>
Date: Mon, 1 Feb 1999 11:31:20 +0000
From: George Foot <mert0407 AT sable DOT ox DOT ac DOT uk>
To: djgpp AT delorie DOT com
Subject: Re: Creating 16-bit code?
References: <199901291338 DOT OAA09477 AT acp3bf DOT physik DOT rwth-aachen DOT de> <Pine DOT GSO DOT 4 DOT 05 DOT 9901291454000 DOT 22923-100000 AT cbl DOT umces DOT edu>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
X-Mailer: Mutt 0.91.1
In-Reply-To: <Pine.GSO.4.05.9901291454000.22923-100000@cbl.umces.edu>; from Fleet Teachout on Fri, Jan 29, 1999 at 03:05:58PM -0500
Reply-To: djgpp AT delorie DOT com

On Fri, Jan 29, 1999 at 03:05:58PM -0500, Fleet Teachout wrote:
> The scanner contains a 386 CPU, 640K RAM (470K useable as a RAM disk).
> When I attempted to run the program I developed in the scanner, I got
> "load error: no DPMI". (or something real close to that).

DJGPP programs require a DPMI server to run.  You need one that 
uses low memory though, since there's no high memory in the system.  
I don't know much about that.

> The assumption I have made (and I'll freely admit I may be WAAAAY off) is
> that I need to use code that will run in a 16-bit environment.

The number of bits isn't the problem, it is that djgpp programs
use DPMI to get their memory.

> The question(s) is/are:  Can I generate 16-bit code using DJGPP operating
> in a 32-bit environment - and if so, how?

You can generate 16 bit code but I don't recommend it and doubt
that it will help you much.  For one thing you'd have to write
in assembler and not use the standard stub or any of the C 
library...  In the end you wouldn't really be using djgpp for 
anything other than development.

Since you'd have to write in assebler anyway, if you know Intel
syntax you might be better off playing with NASM.  Not having any
high memory, you might be better off sticking to real mode if you 
can't get a DPMI server to run your program in low memory.

-- 
George Foot