From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <archer-return-2027-listarch-archer=sourceware.org@sourceware.org>
Received: (qmail 14119 invoked by alias); 6 Jul 2010 20:28:36 -0000
Mailing-List: contact archer-help@sourceware.org; run by ezmlm
Sender: <archer@sourceware.org>
Precedence: bulk
List-Post: <mailto:archer@sourceware.org>
List-Help: <mailto:archer-help@sourceware.org>
List-Subscribe: <mailto:archer-subscribe@sourceware.org>
List-Id: <archer.sourceware.org>
Received: (qmail 14110 invoked by uid 22791); 6 Jul 2010 20:28:35 -0000
X-SWARE-Spam-Status: No, hits=-4.2 required=5.0
	tests=AWL,BAYES_00,RCVD_IN_DNSWL_HI,SPF_HELO_PASS,TW_BJ,TW_JC,T_RP_MATCHES_RCVD
X-Spam-Check-By: sourceware.org
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
From: Roland McGrath <roland@redhat.com>
To: Tom Tromey <tromey@redhat.com>
Cc: Panu Matilainen <pmatilai@redhat.com>,
        Project Archer <archer@sourceware.org>
Subject: Re: find-debuginfo.sh change for gdb index
In-Reply-To: Tom Tromey's message of  Tuesday, 6 July 2010 13:41:41 -0600 <m3wrt8bdoq.fsf@fleche.redhat.com>
References: <m3vd91tqpl.fsf@fleche.redhat.com>
	<20100629232147.C019548255@magilla.sf.frob.com>
	<m3r5joscvs.fsf@fleche.redhat.com>
	<20100630181436.518364C33C@magilla.sf.frob.com>
	<m31vbos7lc.fsf@fleche.redhat.com>
	<m3y6dws151.fsf@fleche.redhat.com>
	<20100630204424.3DCE34C33C@magilla.sf.frob.com>
	<m3lj9wrz60.fsf@fleche.redhat.com>
	<20100630221406.254AC4C33E@magilla.sf.frob.com>
	<m339w1lkuu.fsf@fleche.redhat.com>
	<alpine.LFD.2.00.1007051223290.8757@localhost>
	<m38w5ocvcj.fsf@fleche.redhat.com>
	<20100706191407.535874824F@magilla.sf.frob.com>
	<m3wrt8bdoq.fsf@fleche.redhat.com>
Message-Id: <20100706202828.8ABFA4824F@magilla.sf.frob.com>
Date: Tue, 06 Jul 2010 20:28:00 -0000
X-SW-Source: 2010-q3/txt/msg00010.txt.bz2

> I was thinking we would just objcopy the data into the .debug file in
> find-debuginfo.sh.
> 
> Is there an advantage to doing it before the stripping?

I stated two:

1. CRC32 in .gnu_debuglink will be correct.
   
2. Do it to those that don't get stripped.  
   (AFAIK the only case is vmlinux.)

I am not especially concerned about either of those issues.  But they point
to it probably being the more right or more clean approach.

I am far more concerned about the unknown potential problems.
For example, making sure that eu-unstrip can still recombine the files.
(I think that will be fine, but it's something to worry about.)

In the past we have had various problems with objcopy/bfd deciding to
gratuitously regenerate things and breaking the data along the way.
I believe all the old issues we ever noticed have been fixed by now.
But it certainly makes me nervous.  I trust eu-strip far more.


Thanks,
Roland