Mail Archives: cygwin-developers/1998/02/03/14:03:01
Andy Piper wrote:
From andyp AT parallax DOT co DOT uk Tue Feb 3 00:48:46 1998
Message-Id: <3 DOT 0 DOT 3 DOT 32 DOT 19980203084902 DOT 00957540 AT mailhost>
X-Sender: andyp AT mailhost
X-Mailer: QUALCOMM Windows Eudora Light Version 3.0.3 (32)
Date: Tue, 03 Feb 1998 08:49:02 +0000
To: Geoffrey Noer <noer AT cygnus DOT com>
From: Andy Piper <andyp AT parallax DOT co DOT uk>
Subject: Re: Another coff.h question
In-Reply-To: <199802030146 DOT RAA27413 AT skaro DOT cygnus DOT com>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
At 05:46 PM 2/2/98 -0800, you wrote:
>>I no longer have access to a COFF based native system, but I don't
>>recall a coff.h header file. I recall header files like <a.out.h>,
>><sym.h> and <reloc.h>.
>
>That's right. Digital UNIX (an ECOFF system) has all of these files and
>more. This is what is included in a.out.h:
>
> #include <filehdr.h>
> #include <aouthdr.h>
> #include <scnhdr.h>
> #include <reloc.h>
> #include <syms.h> /* includes storclass.h */
>
>I think a.out.h is pretty standard. Possibly we could get by with just
>calling it a.out.h.
on AT&T 3B* there were:
filehdr.h
aouthdr.h
scnhdr.h
reloc.h
syms.h
storclass.h
(at least according to the document I have).
emacs' coff unexec includes one of coff.h, coff-encap/a.out.encap.h, a.out.h
I am of the opinion that:
a) there should be a single header, multiple headers seems like overkill to
me.
b) coff.h is more descriptive, but there obviously is precedent for a.out.h
including coff information.
c) there appears to be variation from machine to machine, so whatever is
decided is probably not going to conform to something or other.
If a.out.h is agreeable to everyone then call it a.out.h. Its certainly no
less obvious what it is for than coff.h.
Can you let me know of your decision since I will have to change XEmacs.
cheers,
andy
--
Geoffrey Noer
noer AT cygnus DOT com
- Raw text -