[Numpy-discussion] extension module with swig

Matthieu Brucher matthieu.brucher@gmail....
Tue Nov 4 05:09:55 CST 2008


The issue with the LD_LIBRARY_PATH would come up in any case. You have
to put your gsl library folder before the system one if you want your
gsl library to be used.

For the cblas issue, it seems from Google you have to link against a
CBLAS library as well to use the GSL (for instance blas or atlas
shoule be enough).

Matthieu

2008/11/4 Giovanni Samaey <giovanni.samaey@cs.kuleuven.be>:
> And, additionally setting the environment variable LD_LIBRARY_PATH to
> start with /data/home/u0038151/lib instead of ending with it, it picks
> up my own gsl, and gives the error message
>
> 0 u0038151@lo-03-02 dot2 $ python -c "import dot"
> Traceback (most recent call last):
>  File "<string>", line 1, in ?
>  File "dot.py", line 7, in ?
>    import _dot
> ImportError: /data/home/u0038151/lib/libgsl.so.0: undefined symbol:
> cblas_ctrmv
>
> Again something different...
>
> Giovanni
>
> On 04 Nov 2008, at 11:46, Giovanni Samaey wrote:
>
>> Hi Matthieu,
>>
>> thank you for your prompt reply.
>>
>> On 04 Nov 2008, at 11:24, Matthieu Brucher wrote:
>>
>>>> # dot extension module
>>>> _test = Extension("_test",
>>>>                   ["test_wrap.c",
>>>>                    "test.c"],
>>>>                   include_dirs = [numpy_include,'/data/home/
>>>> u0038151/include'],
>>>>                  library_dirs = ['/data/home/u0038151/lib']
>>>>                 )
>>
>> This is where is specify the directory where my header files are, as
>> well as the directory of the library.  If I add, from your
>> suggestion libraries = ['gsl'] to that list, it tries to pick up a
>> gsl that is installed in /usr/lib (but there are no headers there.)
>> Then I get the message:
>>
>> python -c "import dot"
>> Traceback (most recent call last):
>>  File "<string>", line 1, in ?
>>  File "dot.py", line 7, in ?
>>    import _dot
>> ImportError: /usr/lib/libgsl.so.0: undefined symbol: cblas_dsdot
>>
>> So it finds the gsl symbols in the system library (not mine),
>> probably combined with my headers, and then has a different error.
>>
>> Giovanni
>
> _______________________________________________
> Numpy-discussion mailing list
> Numpy-discussion@scipy.org
> http://projects.scipy.org/mailman/listinfo/numpy-discussion
>



-- 
Information System Engineer, Ph.D.
Website: http://matthieu-brucher.developpez.com/
Blogs: http://matt.eifelle.com and http://blog.developpez.com/?blog=92
LinkedIn: http://www.linkedin.com/in/matthieubrucher


More information about the Numpy-discussion mailing list