From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 1987 invoked by alias); 7 Mar 2013 09:52:54 -0000 Received: (qmail 1979 invoked by uid 22791); 7 Mar 2013 09:52:53 -0000 X-SWARE-Spam-Status: No, hits=-6.5 required=5.0 tests=AWL,BAYES_00,KHOP_RCVD_UNTRUST,KHOP_SPAMHAUS_DROP,RCVD_IN_DNSWL_HI,RCVD_IN_HOSTKARMA_W,RP_MATCHES_RCVD,SPF_HELO_PASS X-Spam-Check-By: sourceware.org Received: from mx1.redhat.com (HELO mx1.redhat.com) (209.132.183.28) by sourceware.org (qpsmtpd/0.43rc1) with ESMTP; Thu, 07 Mar 2013 09:52:41 +0000 Received: from int-mx11.intmail.prod.int.phx2.redhat.com (int-mx11.intmail.prod.int.phx2.redhat.com [10.5.11.24]) by mx1.redhat.com (8.14.4/8.14.4) with ESMTP id r279qe9m018282 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Thu, 7 Mar 2013 04:52:40 -0500 Received: from host2.jankratochvil.net (ovpn-116-50.ams2.redhat.com [10.36.116.50]) by int-mx11.intmail.prod.int.phx2.redhat.com (8.14.4/8.14.4) with ESMTP id r279qRWD025433 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES128-SHA bits=128 verify=NO); Thu, 7 Mar 2013 04:52:31 -0500 Date: Thu, 07 Mar 2013 09:52:00 -0000 From: Jan Kratochvil To: Doug Evans Cc: gdb-patches Subject: Re: [patch+doc 1/2] filename-display: 1->4 options {inferior,libs}{,-sepdebug} Message-ID: <20130307095224.GA12606@host2.jankratochvil.net> References: <20130215202536.GA20435@host2.jankratochvil.net> <20130227185345.GA21375@host2.jankratochvil.net> <20130227195251.GA29891@host2.jankratochvil.net> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: User-Agent: Mutt/1.5.21 (2010-09-15) X-IsSubscribed: yes Mailing-List: contact gdb-patches-help@sourceware.org; run by ezmlm Precedence: bulk List-Id: List-Subscribe: List-Archive: List-Post: List-Help: , Sender: gdb-patches-owner@sourceware.org X-SW-Source: 2013-03/txt/msg00269.txt.bz2 On Wed, 06 Mar 2013 20:54:16 +0100, Doug Evans wrote: > We have various parameters that together specify where to find separate debug info. > Is it possible to piggyback on that? > E.g., Something minimal for now like specifying which ones are > displayed as relative with the default being none? > [And thus by default everything with separate debug info will > be displayed using absolute file names.] > Or something like that. It all comes from the problem both /usr/local and $HOME directories look the same for GDB but they should behaves differentely. Some 'where to find separate debug info' based decisions were proposed in: http://sourceware.org/ml/gdb-patches/2013-02/msg00123.html Message-ID: <20130205182245.GA19307@host2.jankratochvil.net> such as: no sepdebug => relative filenames sepdebug in /usr/lib/debug => absolute filenames sepdebug from .debug or .debug/ => relative filenames That will work for your stated binary with sepdebug in $HOME (although I still cannot imagine how that looks / why is it built that way). But that will still not work for /usr/local/lib/mumble.so not using sepdebug where absolute filenames are appropriate but GDB will use relative filenames. > Oops. I also meant to ask about executable-vs-library. > IOW, the distinction between executables and shared libs in your patch. > There's a 2x2 matrix of {executable,shared-lib} x {embedded-debug-info,separate-debug-info}. I find shared libraries uncommon in $HOME so using set filename-display libraries absolute will fix the /usr/local/lib/mumble.so problem while it will still keep relative filenames for hello.c examples. I did not intend to make the "libraries absolute" setting a default but maybe we can. That still will not fix /usr/local/bin/mumble problem where we would also like absolute filenames. There could be very many complicated settings, there could be done some per-directory setting like there is 'set auto-load safe-path'. But I find that needlessly complicated, in fact I believe any experienced GDB user should just do: echo >>~/.gdbinit set filename-display absolute and be done with it. Just I did not want to break first GDB experience with hello.c example code by displaying absolute pathname for ~/hello.c. Rather than those 2x2 settings personally I still prefer the former [patchv2 12/11] New options {relative,basename}-with-system-absolute http://sourceware.org/ml/gdb-patches/2013-01/msg00700.html Message-ID: <20130129221019.GA27463@host2.jankratochvil.net> + [patchv2 13/11] Make relative-with-system-absolute the default http://sourceware.org/ml/gdb-patches/2013-01/msg00701.html Message-ID: <20130129221118.GB27463@host2.jankratochvil.net> possibly just renaming it s/system/separate-debuginfo/. Thanks, Jan