[Numpy-discussion] Python 3.2 crashes numpy with undefined symbol PyCObject_AsVoidPtr

Bruce Southey bsouthey at gmail.com
Tue Aug 3 17:10:47 EDT 2010


  Hi,
Since I was testing the 1.5 beta, I also tested the alpha release of 
Python 3.2 on Linux 64-bit (gcc version 4.4.4 20100630 (Red Hat 
4.4.4-10) (GCC)). While my other Python versions passed the tests (once 
I copied the necessary tools directory for Python3.1), Python3.2 has 
problems so I created ticket 1571:
http://projects.scipy.org/numpy/ticket/1571

The minor change is that string.maketrans function, depreciated in 
Python 3.1, has been removed.
http://docs.python.org/dev/py3k/whatsnew/3.2.html

However the more serious problem is that both 1.5 beta and the current 
svn version crashed:

$ python3.2 -c "import numpy; numpy.test()"
Traceback (most recent call last):
File "<string>", line 1, in <module>
File "/usr/local/lib/python3.2/site-packages/numpy/__init__.py", line 
136, in <module>
from . import add_newdocs
File "/usr/local/lib/python3.2/site-packages/numpy/add_newdocs.py", line 
9, in <module>
from numpy.lib import add_newdoc
File "/usr/local/lib/python3.2/site-packages/numpy/lib/__init__.py", 
line 4, in <module>
from .type_check import *
File "/usr/local/lib/python3.2/site-packages/numpy/lib/type_check.py", 
line 8, in <module>
import numpy.core.numeric as _nx
File "/usr/local/lib/python3.2/site-packages/numpy/core/__init__.py", 
line 5, in <module>
from . import multiarray
ImportError: 
/usr/local/lib/python3.2/site-packages/numpy/core/multiarray.so: 
undefined symbol: PyCObject_AsVoidPtr


This warning that does not appears in Python3.1:
In file included from 
numpy/core/src/multiarray/multiarraymodule_onefile.c:28:
numpy/core/src/multiarray/convert_datatype.c: In function 
‘PyArray_GetCastFunc’:
numpy/core/src/multiarray/convert_datatype.c:106: warning: implicit 
declaration of function ‘PyCObject_AsVoidPtr’
numpy/core/src/multiarray/convert_datatype.c:106: warning: assignment 
makes pointer from integer without a cast

Bruce



More information about the NumPy-Discussion mailing list