From: elf AT netcom DOT com (Marc Singer)
Message-Id: <199604231603.JAA04633@netcom9.netcom.com>
Subject: Re: Standard 32-bit libraries
To: orly AT abigail DOT eee DOT upd DOT edu DOT ph (Orlando Andico)
Date: Tue, 23 Apr 1996 09:03:27 -0700 (PDT)
Cc: djgpp AT sun DOT soe DOT clarkson DOT edu (DJGPP List Alias)
In-Reply-To: <Pine.LNX.3.91.960422183936.8203B-100000@abigail.eee.upd.edu.ph> from "Orlando Andico" at Apr 22, 96 06:47:25 pm
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Content-Length: 1026

> I have a further question  :)
> 
> It's not my intention to question the design decisions made by the DJGPP 
> developer(s), but what advantages exactly does COFF have over AOUT 
> (forgive my ignorance)? I mean, with COFF I lose an extremely desirable 
> feature -- that of being able to debug/step-through code in other source 
> files, i.e. headers (I know you've answered this question of mine before, 
> Eli).

COFF is the method of implementation for AOUT used by the GNU
compiler.  This is not a choice made by the DJGPP developers.

> 
> Also: in the source code for gcc, there is an option to insert stabs in 
> the compiled file, kind of to get around the limitations of COFF. Does 
> this *work* for DJGPP? I don't know nearly enough to change the 
> compiler's output back to AOUT, but will this stabs-insertion that the 
> compiler (if rebuilt...) re-enable the desired GDB features?

Good question.  It seems to me that going to ELF would be better since
it is a better designed format than COFF.

Marc Singer