[yt-users] error launching yt: __gnu_cxx::__concurrence_lock_error
Matthew Turk
matthewturk at gmail.com
Tue Jan 17 11:12:57 PST 2012
Hi JC,
Ah, okay. This sounds more and more like a problem with building
matplotlib. I'm going to recommend a potentially unpopular solution:
wipe your yt-whatever directory, make sure you set CC and CXX, and
build from scratch with a fresh checkout of the install script. My
suspicion is that there is some library -- in this case, likely one of
the matplotlib .so's -- that uses C++, is incorrectly built/linked,
and is causing these funky import errors. This may not fix your
problem, but if it does, that's a happy day. If it does not, we'll
sit down next week and I'll take a look.
-Matt
On Tue, Jan 17, 2012 at 2:15 PM, Jean-Claude Passy <jcpassy at gmail.com> wrote:
> Hi Matt,
>
> I rebuilt the directory but it did not fix it (I did not have to manually
> edit yt/frontends/ramses/setup.py).
> I also found out that an 'import pylab' lin the ipython session leads to the
> same error.
>
> Thanks,
>
> JC
>
>
> On 17/01/12 13:09, Matthew Turk wrote:
>>
>> Hi JC,
>>
>> This is suspicious. Can you try forcing a rebuild of your yt-hg
>> directory, and make sure:
>>
>> CC=gcc-4.2
>> CXX=g++-4.2
>>
>> Rebuild with:
>>
>> python2.7 setup.py build_ext -i -f
>>
>> This might fix it. You may also have to manually edit
>> yt/frontends/ramses/setup.py and ensure that stdc++ is in the
>> libraries=[] argument for _ramses_reader.pyx.
>>
>> -Matt
>>
>> On Tue, Jan 17, 2012 at 12:21 PM, Jean-Claude Passy<jcpassy at gmail.com>
>> wrote:
>>>
>>> Hi John,
>>>>
>>>> Does ipython load fine?
>>>
>>> yes, it does.
>>>
>>>> Can you manually import yt.mods in a python session?
>>>
>>> No, I can't. The error occurs when I try to import it.
>>>
>>>> Could you check whether a standalone python instance fails with the same
>>>> error when you do
>>>>
>>>> import matplotlib
>>>> import h5py
>>>> import numpy
>>>
>>> No problem, I can import all these packages.
>>>
>>> Thanks a lot for you help,
>>>
>>> JC
>>>
>>>
>>>>> Hi,
>>>>>
>>>>> I have been using the install script with the regular gcc-4.2 as
>>>>> suggested:
>>>>>
>>>>> ------------------------------------------------------------------
>>>>> [ 9:12:39] Obiwan:$ CC --version
>>>>> i686-apple-darwin10-gcc-4.2.1 (GCC) 4.2.1 (Apple Inc. build 5666) (dot
>>>>> 3)
>>>>> ------------------------------------------------------------------
>>>>>
>>>>> It is so frustrating... yt was working great on my laptop a couple of
>>>>> months ago and I did not use it until 2 days ago, mainly to check that
>>>>> I
>>>>> was ready for the workshop. May be this shows that one should not spend
>>>>> a single week without using yt! ;-)
>>>>>
>>>>> Thanks for your help guys, I appreciate.
>>>>> JC
>>>>>
>>>>>
>>>>> On 16/01/12 21:01, Casey W. Stark wrote:
>>>>>>
>>>>>> I agree, it sounds like a gcc build issue. Which version did you build
>>>>>> against JC?
>>>>>>
>>>>>> It might take me a while to get to it, but I can try to reproduce it
>>>>>> this week.
>>>>>>
>>>>>> Best,
>>>>>> Casey
>>>>>>
>>>>>>
>>>>>> On Mon, Jan 16, 2012 at 4:53 PM, Matthew Turk<matthewturk at gmail.com
>>>>>> <mailto:matthewturk at gmail.com>> wrote:
>>>>>>
>>>>>> Hi JC,
>>>>>>
>>>>>> Unfortunately I'm not really sure -- I no longer have a Mac to
>>>>>> test
>>>>>> against, and this error looks like some deep compilation bug,
>>>>>> likely
>>>>>> with matplotlib. It might be something about your installation
>>>>>> (did
>>>>>> you use the install script? were you using a non-standard GCC?
>>>>>> etc)
>>>>>> but I'm not really sure.
>>>>>>
>>>>>> Anyone else out there, with a Mac, who might be able to give JC a
>>>>>> pointer?
>>>>>>
>>>>>> -Matt
>>>>>>
>>>>>> On Sun, Jan 15, 2012 at 1:56 PM, Jean-Claude Passy
>>>>>> <jcpassy at gmail.com<mailto:jcpassy at gmail.com>> wrote:
>>>>>> > Hey,
>>>>>> >
>>>>>> > unfortunately, it does not fix the thing...
>>>>>> > Any other idea?
>>>>>> >
>>>>>> > Thanks,
>>>>>> >
>>>>>> > JC
>>>>>> >
>>>>>> >
>>>>>> > On 15/01/12 13:30, Matthew Turk wrote:
>>>>>> >>
>>>>>> >> Hi JC,
>>>>>> >>
>>>>>> >> My guess is that this is related to using the OSX backend
>>>>>> for
>>>>>> >> Matplotlib. If you can, could you set (in ~/.matplotlibrc)
>>>>>> the
>>>>>> >> variable "backend" to be "Agg" it might fix it.
>>>>>> >>
>>>>>> >> -Matt
>>>>>> >>
>>>>>> >> On Sun, Jan 15, 2012 at 1:33 PM, Jean-Claude
>>>>>> Passy<jcpassy at gmail.com<mailto:jcpassy at gmail.com>>
>>>>>> >> wrote:
>>>>>> >>>
>>>>>> >>> Hi all,
>>>>>> >>>
>>>>>> >>> I am encountering a weird error when trying to execute
>>>>>> yt-unstable on my
>>>>>> >>> MacBook Pro (10.6.8):
>>>>>> >>>
>>>>>> >>>
>>>>>> ----------------------------------------------------------------
>>>>>> >>> [13:26:06] Obiwan:$ iyt
>>>>>> >>> terminate called after throwing an instance of
>>>>>> >>> '__gnu_cxx::__concurrence_lock_error'
>>>>>> >>> what(): __gnu_cxx::__concurrence_lock_error
>>>>>> >>> Abort trap
>>>>>> >>>
>>>>>> ----------------------------------------------------------------
>>>>>> >>>
>>>>>> >>> I re-installed everything successfully but the error is
>>>>>> still
>>>>>> there.
>>>>>> >>> Any help will be much appreciated!
>>>>>> >>>
>>>>>> >>> Thanks a lot,
>>>>>> >>>
>>>>>> >>> JC
>>>>>> >>> _______________________________________________
>>>>>> >>> yt-users mailing list
>>>>>> >>>
>>>>>> yt-users at lists.spacepope.org<mailto:yt-users at lists.spacepope.org>
>>>>>> >>>
>>>>>> http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
>>>>>> _______________________________________________
>>>>>> yt-users mailing list
>>>>>> yt-users at lists.spacepope.org<mailto:yt-users at lists.spacepope.org>
>>>>>> http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
>>>>>>
>>>>>>
>>>>> _______________________________________________
>>>>> yt-users mailing list
>>>>> yt-users at lists.spacepope.org
>>>>> http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
>>>
>>> _______________________________________________
>>> yt-users mailing list
>>> yt-users at lists.spacepope.org
>>> http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
More information about the yt-users
mailing list