This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: [PATCH] Polished patch (crash after pthread_exit)
- From: "Carlos O'Donell" <carlos at redhat dot com>
- To: Roland McGrath <roland at hack dot frob dot com>
- Cc: libc-alpha at sourceware dot org
- Date: Wed, 26 Jun 2013 16:20:11 -0400
- Subject: Re: [PATCH] Polished patch (crash after pthread_exit)
- References: <CAKt5dD1-XUyCSdhys+J5NJbCYQTMunwpcmECvRm1F0VSR4Jh+w at mail dot gmail dot com> <51C89EEB dot 3030604 at suse dot com> <51C8B629 dot 5090206 at redhat dot com> <20130624214304 dot EC63D2C09B at topped-with-meat dot com> <51C9B434 dot 5000303 at redhat dot com> <20130626200615 dot CB3F12C081 at topped-with-meat dot com>
On 06/26/2013 04:06 PM, Roland McGrath wrote:
>> I had forgotten it. Are you opposed to me starting a new
>> set of nptl tests called `tst-pthreadX-static.c' which tracks
>> all of the sensible static things you might want to do with
>> libpthread.a?
>
> Sounds fine.
Excellent, I'll do that.
>> Although there are lots of permutations of these functions
>> there are probably only a few useful orderings we need to
>> check to make libpthread.a work correctly (without having
>> to resort to something like Fedora's linking of the entire
>> library into one .o file).
>
> I didn't realize Fedora was doing that. The downstream bugs that led to
> that change should be reported upstream so we can have regression tests for
> those cases and find the best solution.
I didn't either, I just discovered it, and it's probably hiding
a lot of problems. I think I'm going to revert this behaviour and
just start fixing the resulting bugs instead.
>> Unless you think that Fedora's large static .o file is
>> perhaps a more maintainable solution? I'd rather we try
>> to keep the bloat down if at all possible.
>
> I agree.
Glad to hear it.
Cheers,
Carlos.