Date: Fri, 31 Oct 1997 13:50:33 +0100 (MET)
From: Hans-Bernhard Broeker <broeker AT physik DOT rwth-aachen DOT de>
Subject: Re: flat pointer
In-reply-to: <3459CE68.390E@trash.lip6.fr>
To: Fabrice ILPONSE <fabrice AT asim DOT lip6 DOT fr>
Cc: djgpp workers group <djgpp-workers AT delorie DOT com>
Message-id: <Pine.LNX.3.91.971031134349.445G-100000@acp3bf>
MIME-version: 1.0
Content-type: TEXT/PLAIN; charset=US-ASCII
Content-transfer-encoding: 7BIT
Precedence: bulk

On Fri, 31 Oct 1997, Fabrice ILPONSE wrote:

> Fabrice ILPONSE wrote:
> > 
> >         I am thinking about adding another kind of pointer in djgpp.
> >         In djgpp, the pointers are flat but the lowest adress is limited by the
> > DS selector.
> >         Why not use a "<type> flat *" in djgpp that would make easier to access
> > the < 640Ko memory for the video memory for example.

By all means: *no*! This would serve no useful purpose (you can always get
the same effect with the various existing methods), but re-introduce a
concept that's been haunting previous MS-DOS programmers porting their
programs to other systems for ages. Adding additional keywords like 'far'
to the core language is about as bad a design as you can possibly come up
with... 

Coincidentally, your 'flat' is exactly what 16-bit DOS compilers used to
call 'far' (or maybe 'huge'). Re-ntroducing it would be like opening
pandora's box: users would immediately cry for 'flat' versions of malloc()
and friends, the string functions and whatnot. 

HBB