[Numpy-tickets] [NumPy] #338: Valgrind warning

NumPy numpy-tickets at scipy.net
Thu Oct 12 07:50:46 CDT 2006


#338: Valgrind warning
-----------------------------+----------------------------------------------
 Reporter:  albertstrasheim  |       Owner:  somebody   
     Type:  defect           |      Status:  new        
 Priority:  normal           |   Milestone:  1.0 Release
Component:  numpy.core       |     Version:  devel      
 Severity:  normal           |    Keywords:             
-----------------------------+----------------------------------------------
 When running the following code:

 {{{
 #!python
 import numpy as np
 import scipy as sp
 from scipy.interpolate import interp1d
 print np.__version__
 print sp.__version__
 x10 = np.arange(10.)
 y210 = np.arange(20.).reshape((2, 10))
 interp210 = interp1d(x10, y210)
 interp210(1.5)
 }}}

 Valgrind emits the following warning on 32-bit Linux:
 {{{
 ==1076== Conditional jump or move depends on uninitialised value(s)
 ==1076==    at 0x4433E14: PyArray_MapIterReset (arrayobject.c:9684)
 ==1076==    by 0x44679F3: array_ass_sub (arrayobject.c:2442)
 ==1076==    by 0x4A1EA6B: PyObject_SetItem (in
 /usr/lib/libpython2.4.so.1.0)
 ==1076==    by 0x4A7D92C: PyEval_EvalFrame (in
 /usr/lib/libpython2.4.so.1.0)
 ==1076==    by 0x4A81757: PyEval_EvalCodeEx (in
 /usr/lib/libpython2.4.so.1.0)
 ...
 }}}

 Similar warning on 64-bit Linux:

 {{{
 ==31596== Conditional jump or move depends on uninitialised value(s)
 ==31596==    at 0x571ACD0: PyArray_MapIterReset (arrayobject.c:9650)
 ==31596==    by 0x575CED0: array_ass_sub (arrayobject.c:2433)
 ==31596==    by 0x474080: PyEval_EvalFrame (in /usr/bin/python2.4)
 ==31596==    by 0x4783FE: PyEval_EvalCodeEx (in /usr/bin/python2.4)
 ...
 }}}

 Tested with Valgrind 3.1.0 and 3.2.1. Used !NumPy 1.0.dev3217 and !SciPy
 0.5.2.dev2254 on 64-bit system. Used  !NumPy 1.0.dev3315 and !SciPy
 0.5.2.dev2241 on 32-bit system.

-- 
Ticket URL: <http://projects.scipy.org/scipy/numpy/ticket/338>
NumPy <http://projects.scipy.org/scipy/numpy>
The fundamental package needed for scientific computing with Python.


More information about the Numpy-tickets mailing list