This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
Re: [PATCH v5 6/8] gdbserver build-id attribute generator
- From: Tom Tromey <tromey at redhat dot com>
- To: Jan Kratochvil <jan dot kratochvil at redhat dot com>
- Cc: gdb-patches at sourceware dot org, Aleksandar Ristovski <ARistovski at qnx dot com>
- Date: Wed, 21 May 2014 10:47:43 -0600
- Subject: Re: [PATCH v5 6/8] gdbserver build-id attribute generator
- Authentication-results: sourceware.org; auth=none
- References: <20140319223004 dot 14668 dot 20989 dot stgit at host1 dot jankratochvil dot net> <20140319223115 dot 14668 dot 73456 dot stgit at host1 dot jankratochvil dot net> <87ppj8tswu dot fsf at fleche dot redhat dot com> <20140521122414 dot GA14744 at host2 dot jankratochvil dot net>
Jan> I wrote a never replied essay + [patch] fix about why the current handling
Jan> of 'version' is wrong:
Jan> [patch] gdbserver <library-list> and its #FIXED version="1.0"
Jan> https://sourceware.org/ml/gdb-patches/2011-11/msg00099.html
I read that note. The patch seems correct to me.
Tom> I really don't know; nor did I find anything in the documentation.
Tom> So at the very least this needs an update to the manual to warn
Tom> future users.
Jan> I do not understand this. The only changed (newly added) attribute is
Jan> build-id #IMPLIED - therefore optional - therefore backward/forward
Jan> compatible. Who needs to be warned about it?
I thought all attributes had to be specified in the DTD. Is that not
so?
If they must be then the problem comes when using an older gdb with a
newer gdbserver.
Anyway the point of my "warn" comment is that, ideally, our versioning
approach ought to be documented.
Tom