This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
Re: [PATCH v5 8/8] Tests for validate symbol file using build-id
- From: Jan Kratochvil <jan dot kratochvil at redhat dot com>
- To: Tom Tromey <tromey at redhat dot com>
- Cc: gdb-patches at sourceware dot org, Aleksandar Ristovski <ARistovski at qnx dot com>
- Date: Tue, 20 May 2014 17:29:34 +0200
- Subject: Re: [PATCH v5 8/8] Tests for validate symbol file using build-id
- Authentication-results: sourceware.org; auth=none
- References: <20140319223004 dot 14668 dot 20989 dot stgit at host1 dot jankratochvil dot net> <20140319223131 dot 14668 dot 9029 dot stgit at host1 dot jankratochvil dot net> <87ha4kts54 dot fsf at fleche dot redhat dot com>
On Tue, 20 May 2014 16:57:11 +0200, Tom Tromey wrote:
> >>>>> "Jan" == Jan Kratochvil <jan.kratochvil@redhat.com> writes:
>
> Jan> 2014-02-26 Aleksandar Ristovski <aristovski@qnx.com
> Jan> Tests for validate symbol file using build-id.
> Jan> * gdb.server/solib-mismatch-lib.c: New file.
> Jan> * gdb.server/solib-mismatch-libmod.c: New file.
> Jan> * gdb.server/solib-mismatch.c: New file.
> Jan> * gdb.server/solib-mismatch.exp: New file.
>
> I thought Pedro had wanted these not in gdb.server.
> Or am I confusing that with some other patch?
In a local copy they are moved back to gdb.base/ .
But I haven't re-post the whole series just because of it.
Planning to check it in into gdb.base/ .
> Jan> +if ![is_remote target] {
> Jan> + untested "only gdbserver supports build-id reporting"
> Jan> + return -1
>
> I was mildly confused to read this.
> Isn't build-id also supported natively?
For this case of build-id it is not. It was discussed in:
Re: [patchv3 7/8] Validate symbol file using build-id
Message-ID: <53108EF7.3000708@redhat.com>
https://sourceware.org/ml/gdb-patches/2014-02/msg00862.html
https://sourceware.org/ml/gdb-patches/2014-03/msg00011.html
> How does the new functionality interact with the existing build-id
> functionality?
Currently only the separate debug info file is located and validated by
build-id. This patch is about validating the primary file (without .debug
extension - in fact symbol file as GDB uses that file only for symbols).
Besides all of these there are additional patches:
http://pkgs.fedoraproject.org/cgit/gdb.git/tree/
gdb-6.6-buildid-locate-*
which are also about locating the primary files but those patches are for
local (NAT) mode.
Also this series does not validate / locate the main executable, the Fedora
patches above work also for the main executable.
It got all a bit messy so I decided to merge it all and preparing it as an
update (technically add-on) on this patch series. So also contrary to my
original plans in
Re: [patchv3 7/8] Validate symbol file using build-id
above I am going to implement the local (NAT) mode for locating the files.
As I expect GDB is not going to unify LocalRemoteFeatureParity soon enough.
It will be about upstreaming the Fedora patchset which is the last largest
Fedora-specific patchsets kept.
Jan