Date: Tue, 24 Feb 1998 09:54:17 +0200 (IST)
From: Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
To: "Richard M. Myers" <myersr AT nosc DOT mil>
cc: djgpp AT delorie DOT com
Subject: Re: FP_SEG() and FP_OFF() macros
In-Reply-To: <34F24DA5.BF697D29@nosc.mil>
Message-ID: <Pine.SUN.3.91.980224095401.28179G-100000@is>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Precedence: bulk


On Mon, 23 Feb 1998, Richard M. Myers wrote:

> I studied the faq and could  find nothing *useful* related to this
> DOS-ism, though there are some vaque references to FP_SEG() and
> FP_OFF().

Please explain what is ``vague'' and ``unuseful'' in section 17.7 of
the FAQ where it explains how to port FP_XXX macros.  I cannot change
anything in the FAQ unless people tell me why the current text is not
good enough, beacuse I always do my best to make it as clear and as
useful as I can possibly get.

The problem with these macros is that there is no single way of
converting them which will work in all cases.  That's what the FAQ
says, and that is why it refers you to section 18.2 which explains in
detail how to address real-mode buffers from DJGPP programs.  You will
need to read section 18.2 and the library docs for the related
functions, and then decide how to convert your code.  Since you didn't
even say what do these macros do in the original code, I cannot advise
about a solution suitable for your case.

>           (To the faq maintainer:  Examples can be indispensible to
> someone not versed in the x86 segmentation follies. :)

The FAQ is full of examples where I could invent them, including in
section 18.2.  I don't know how to make an example for a case which
could have half a dozen different solutions, depending on what the
code needs to do.  If you (or anybody else) have a suggestion for an
example of how to convert FP_XXX, please tell me and I will gladly add
it to the FAQ.