svn commit: r312209 - in head/devel: allegro apr1 apr2 bennugd-core bennugd-modules boost-all boost-libs boost-python-libs bullet ccons cgit clanlib1 clanlib22 clint cocaine-core commoncpp cppcheck...

Jeremy Messenger mezz.freebsd at gmail.com
Tue Apr 16 13:12:52 UTC 2013


On Mon, Apr 15, 2013 at 7:49 PM, Steve Wills <steve at mouf.net> wrote:
>> On 2013-Apr-14, 22:29, Steve Wills wrote:
>>> I may be wrong, but some of these changes seem wrong. At least for glib
>>> and any of the others that completely removed PTHREAD_LIBS entirely. I
>>> think it's fine to replace PTHREAD_LIBS with -pthread, but removing
>>> PTHREAD_LIBS or -pthread entirely could be problematic. These two PRs:
>>>
>>> http://www.freebsd.org/cgi/query-pr.cgi?pr=177800
>>> http://www.freebsd.org/cgi/query-pr.cgi?pr=176888
>>>
>>> seem to be caused by this. Am I wrong? If not, we may need to go back
>>> through these changes and put -pthread back where it was removed
>>> entirely.
>>
>> I would be surprised if they were. glib20 configures fine without the
>> explicit pthread flags. See config.log. The shlib ends up correctly
>> linked against the threading library:
>>
>>> ldd /usr/local/lib/libglib-2.0.so.0
>>     libintl.so.9 => /usr/local/lib/libintl.so.9 (0x801702000)
>>     libiconv.so.3 => /usr/local/lib/libiconv.so.3 (0x80190b000)
>>     libpcre.so.3 => /usr/local/lib/libpcre.so.3 (0x801c04000)
>>     libthr.so.3 => /lib/libthr.so.3 (0x801e67000)
>>     libc.so.7 => /lib/libc.so.7 (0x80081c000)
>>
>> Moreover, the symptoms aren't quite what I'd expect from an application
>> missing a pthread link, as they usually die in initialization phase,
>> even before they reach main.
>>
>
> Good points, thanks, I knew I was missing something. I do indeed see that
> my ldd output looks like yours above.
>
>> Moreover, if glib20 was affected, I'd expect a far larger number of
>> ports to start behaving strangely ;)
>>
>> anyway, I'm building xpra as we speak and I'll let you know what I find
>> out.
>
> I think you're right, it's an upstream issue with xpra and sonata, I'll
> try to fix what I can.

It could be a bug in the pygtk2 or it's just in both xpra and sonata.
I have no idea as I am not good with the thread stuff.

> Thanks,
> Steve
>
>>
>>
>>>
>>> Steve
<snip>


--
mezz.freebsd at gmail.com - mezz at FreeBSD.org
FreeBSD GNOME Team
http://www.FreeBSD.org/gnome/ - gnome at FreeBSD.org


More information about the svn-ports-head mailing list