This is the mail archive of the
binutils@sourceware.cygnus.com
mailing list for the binutils project.
Re: patch to fix nm.c(print_object_filename_bsd) on FreeBSD
- To: Michael Sokolov <msokolov at ivan dot Harhan dot ORG>
- Subject: Re: patch to fix nm.c(print_object_filename_bsd) on FreeBSD
- From: "David O'Brien" <obrien at NUXI dot ucdavis dot edu>
- Date: Wed, 19 Apr 2000 11:00:13 -0700
- Cc: binutils at sourceware dot cygnus dot com
- Organization: The NUXI BSD group
- References: <0004191238.AA06748@ivan.Harhan.ORG>
- Reply-To: obrien at NUXI dot ucdavis dot edu
On Wed, Apr 19, 2000 at 07:38:57AM -0500, Michael Sokolov wrote:
>
> I guess I'm the authority on this one, being the current maintainer of the
> traditional VAX BSD.
I disagree. I would say the CSRG SCCS files are the authority. The CSRG
defined what BSD is. If you have CSRG sources, you will see that the
output my patch causes has been depended on by BSD since 1990/03/20.
Previous to 1990/03/20, lorder.sh did not use the "-o" ``nm'' option.
I have checked the CSRG nm.c revisions and found that the current
Binutils output is not in agreement since rev 5.1.
----------------------------
revision 5.1
date: 1989/09/10 17:47:01; author: bostic; state: Exp; lines: +491 -308
new version from Hans Huebner
hans@garp.mit.edu, huebner@db0tui6.BITNET
----------------------------
CSRG rev 4.8 output agrees with your output.
----------------------------
revision 4.8
date: 1987/04/07 17:25:35; author: bostic; state: Exp; lines: +264 -231
bug report 4.2BSD/bin/31 (well, that's how it started...)
----------------------------
It is my understanding that ``nm'' going all the way back to the AT&T
code on the PDP-11 behaved as per my patch. Possibly that is why
hans@garp.mit.edu changed the output.
> The nm in the current CVS binutils behaves exactly as the
> native 4.3BSD nm in this respect
True, but my patch does not break ``lorder'' on 4.3BSD as the sed
expression doesn't use nm's "-o" option. [I am assuming CSRG rev 5.1
(date: 1988/10/22 16:08:00; author: bostic; state: Exp; Initial
revision) is the 4.3BSD version]
> so don't change anything Ian.
I disagree. The author of the BSD bits (Keith Bostic) defined BSD's nm
to output as per my patch. Thus it should be fixed in the Souceware CVS
sources.
--
-- David (obrien@NUXI.com)