[Scipy-tickets] [SciPy] #785: scipy.spatial problem for 64 bits python2.4

SciPy scipy-tickets@scipy....
Sat Nov 8 01:32:51 CST 2008


#785: scipy.spatial problem for 64 bits python2.4
--------------------+-------------------------------------------------------
 Reporter:  cdavid  |        Owner:  peridot
     Type:  defect  |       Status:  new    
 Priority:  normal  |    Milestone:  0.7.0  
Component:  Other   |      Version:         
 Severity:  normal  |   Resolution:         
 Keywords:          |  
--------------------+-------------------------------------------------------
Old description:

> Several scipy.spatial tests fail on 64 bits arch under python 2.4 (this
> is different from the other problem reported in #779):
>
> {{{#!
> ======================================================================
> ERROR: test_kdtree.test_small_nonleaf_compiled.test_points_near_linf
> ----------------------------------------------------------------------
> Traceback (most recent call last):
>   File "/usr/data/david/local/stow/nose-0.10.4/lib/python2.4/site-
> packages/nose/case.py", line 363, in setUp
>     try_run(self.inst, ('setup', 'setUp'))
>   File "/usr/data/david/local/stow/nose-0.10.4/lib/python2.4/site-
> packages/nose/util.py", line 453, in try_run
>     return func()
>   File "/usr/data/david/local/lib64/python2.4/site-
> packages/scipy/spatial/tests/test_kdtree.py", line 129, in setUp
>     self.kdtree = cKDTree(self.data,leafsize=1)
>   File "ckdtree.pyx", line 223, in scipy.spatial.ckdtree.cKDTree.__init__
> (scipy/spatial/ckdtree.c:973)
>   File "numpy.pxd", line 43, in numpy.ndarray.__getbuffer__
> (scipy/spatial/ckdtree.c:2082)
> RuntimeError: Py_intptr_t and Py_ssize_t differs in size, numpy.pxd does
> not support this
> }}
>
> I am not sure, but I believe Py_ssize_t is an int in python 2.4, even on
> 64 bits arch (whereas Py_intrptr_t is just an alias to intptr_t, that is
> has to be 8 bytes to 64 bits arch), which would explain the problem. I
> don't know if that's a cython problem or scipy.spatial.

New description:

 Several scipy.spatial tests fail on 64 bits arch under python 2.4 (this is
 different from the other problem reported in #779):

 {{{
 #!
 ======================================================================
 ERROR: test_kdtree.test_small_nonleaf_compiled.test_points_near_linf
 ----------------------------------------------------------------------
 Traceback (most recent call last):
   File "/usr/data/david/local/stow/nose-0.10.4/lib/python2.4/site-
 packages/nose/case.py", line 363, in setUp
     try_run(self.inst, ('setup', 'setUp'))
   File "/usr/data/david/local/stow/nose-0.10.4/lib/python2.4/site-
 packages/nose/util.py", line 453, in try_run
     return func()
   File "/usr/data/david/local/lib64/python2.4/site-
 packages/scipy/spatial/tests/test_kdtree.py", line 129, in setUp
     self.kdtree = cKDTree(self.data,leafsize=1)
   File "ckdtree.pyx", line 223, in scipy.spatial.ckdtree.cKDTree.__init__
 (scipy/spatial/ckdtree.c:973)
   File "numpy.pxd", line 43, in numpy.ndarray.__getbuffer__
 (scipy/spatial/ckdtree.c:2082)
 RuntimeError: Py_intptr_t and Py_ssize_t differs in size, numpy.pxd does
 not support this
 }}}

 I am not sure, but I believe Py_ssize_t is an int in python 2.4, even on
 64 bits arch (whereas Py_intrptr_t is just an alias to intptr_t, that is
 has to be 8 bytes to 64 bits arch), which would explain the problem. I
 don't know if that's a cython problem or scipy.spatial.

-- 
Ticket URL: <http://scipy.org/scipy/scipy/ticket/785#comment:1>
SciPy <http://www.scipy.org/>
SciPy is open-source software for mathematics, science, and engineering.


More information about the Scipy-tickets mailing list