[Pkg-gmagick-im-team] Bug#518771: libmagickcore1: AcquireImageColormap and other functions hang on gomp_team_end

Guillaume Gimenez ggimenez at free.fr
Mon Mar 9 10:03:45 UTC 2009


Bastien ROUCARIES a écrit :
> On Sun, Mar 8, 2009 at 6:34 PM, Guillaume Gimenez
> <guillaume at blackmilk.fr> wrote:
>> Bastien ROUCARIES a écrit :
>>> On Sun, Mar 8, 2009 at 3:32 PM, Guillaume Gimenez <ggimenez at free.fr>
>>> wrote:
>>>> Package: libmagickcore1
>>>> Version: 7:6.4.5.4.dfsg1-1
>>>> Severity: normal
>>>>
> 
> Thank you
> 
> I work on my testing version however. So I suppose you are working on
> experimental :) According to changelog they are no work on this area
> of code in recent version. Could I have your arch.
> 
> Regards
> 
> Bastien

Yes, packages libmagick++-dev,libmagick++1 and libmagickcore1 come from 
experimental.
on x86_64, testcase hangs in futex_wait called from gomp
on i386, testcase segfaults somewhere in gomp.

However, with the following packages, testcase works perfectly on both 
architectures

$ dpkg -l | grep libmagick
ii  libmagick++10                          7:6.3.7.9.dfsg1-3          ...
ii  libmagick++9-dev                       7:6.3.7.9.dfsg1-3          ...
ii  libmagick10                            7:6.3.7.9.dfsg1-3          ...
ii  libmagick9-dev                         7:6.3.7.9.dfsg1-3          ...

I wonder if race condition is caused by the use of gomp.

Also, is there an inconsistency in the naming convention ? I don't understand 
very well what means 1, 9 and 10.

Best regards






More information about the Pkg-gmagick-im-team mailing list