[Numpy-tickets] [NumPy] #671: linalg ImportError on Solaris 10 systems

NumPy numpy-tickets@scipy....
Tue Feb 26 10:56:42 CST 2008


#671: linalg ImportError on Solaris 10 systems
--------------------------------+-------------------------------------------
 Reporter:  chanley             |        Owner:  somebody
     Type:  defect              |       Status:  new     
 Priority:  normal              |    Milestone:  1.0.5   
Component:  numpy.linalg        |      Version:  none    
 Severity:  normal              |   Resolution:          
 Keywords:  linalg,ImportError  |  
--------------------------------+-------------------------------------------
Comment (by oliphant):

 {{{
 /opt/SUNWspro-6u2/bin/f90 -Bdynamic -G -Bdynamic -G
 build/temp.solaris-2.10-sun4u-2.5/numpy/linalg/lapack_litemodule.o
 -L/usr/stsci/pyssgx/Python-2.5.1/lib -llapack -lblas -lfsu -lsunmath
 -lmvec -lf77compat -o
 build/lib.solaris-2.10-sun4u-2.5/numpy/linalg/lapack_lite.so
 }}}

 I think this is the trouble.   Is this f90 compiler the same one that was
 used to build the lapack and blas libraries that were found by
 numpy.distutils?

 {{{
 blas_info:
 31        FOUND:
 32          libraries = ['blas']
 33          library_dirs = ['/usr/stsci/pyssgx/Python-2.5.1/lib']
 34          language = f77
 35
 36        FOUND:
 37          libraries = ['blas']
 38          library_dirs = ['/usr/stsci/pyssgx/Python-2.5.1/lib']
 39          define_macros = [('NO_ATLAS_INFO', 1)]
 40          language = f77
 }}}

 {{{
 lapack_info:
 80        FOUND:
 81          libraries = ['lapack']
 82          library_dirs = ['/usr/stsci/pyssgx/Python-2.5.1/lib']
 83          language = f77
 84
 85        FOUND:
 86          libraries = ['lapack', 'blas']
 87          library_dirs = ['/usr/stsci/pyssgx/Python-2.5.1/lib']
 88          define_macros = [('NO_ATLAS_INFO', 1)]
 89          language = f77
 90
 }}}

-- 
Ticket URL: <http://scipy.org/scipy/numpy/ticket/671#comment:3>
NumPy <http://projects.scipy.org/scipy/numpy>
The fundamental package needed for scientific computing with Python.


More information about the Numpy-tickets mailing list