On 2/17/10 7:47 AM, Patrick Riesen wrote:
hi all of you
just made a pull for dorsal + all packages and get the following:
In file included from
/scratch/priesen/FEniCS/build/lib/python2.5/site-packages/numpy/core/include/numpy/ndarrayobject.h:68,
from
/scratch/priesen/FEniCS/build/lib/python2.5/site-packages/numpy/core/include/numpy/arrayobject.h:14,
from dolfin/swig/cpp_wrap.cc:3519:
/scratch/priesen/FEniCS/build/lib/python2.5/site-packages/numpy/core/include/numpy/npy_common.h:120:2:
error: #error npy_cdouble definition is not compatible with C99 complex
definition ! Please contact Numpy maintainers and give detailed
information about your compiler and platform
/scratch/priesen/FEniCS/build/lib/python2.5/site-packages/numpy/core/include/numpy/npy_common.h:127:2:
error: #error npy_cfloat definition is not compatible with C99 complex
definition ! Please contact Numpy maintainers and give detailed
information about your compiler and platform
/scratch/priesen/FEniCS/build/lib/python2.5/site-packages/numpy/core/include/numpy/npy_common.h:134:2:
error: #error npy_clongdouble definition is not compatible with C99
complex definition ! Please contact Numpy maintainers and give detailed
information about your compiler and platform
In file included from
/scratch/priesen/FEniCS/build/lib/python2.5/site-packages/numpy/core/include/numpy/ndarrayobject.h:1258,
from
/scratch/priesen/FEniCS/build/lib/python2.5/site-packages/numpy/core/include/numpy/arrayobject.h:14,
from dolfin/swig/cpp_wrap.cc:3519:
/scratch/priesen/FEniCS/build/lib/python2.5/site-packages/numpy/core/include/numpy/__multiarray_api.h:
In function 'int _import_array()':
/scratch/priesen/FEniCS/build/lib/python2.5/site-packages/numpy/core/include/numpy/__multiarray_api.h:1202:
error: 'NPY_ABI_VERSION' was not declared in this scope
/scratch/priesen/FEniCS/build/lib/python2.5/site-packages/numpy/core/include/numpy/__multiarray_api.h:1208:
error: 'NPY_API_VERSION' was not declared in this scope
scons: *** [dolfin/swig/cpp_wrap.os] Error 1
scons: building terminated because of errors.
In dorsal numpy1.4 is used, do you all use the same?
I am not sure if anyone uses numpy from dorsal. The package exists there
only for really really old platforms (like Red Hat Enterprise).
I guess Dorsal's new policy is going to be to get people building as
little as possible, and people who need additional stuff should know
what they are doing. Try building numpy manually, and see if the problem
persists.