[Numpy-tickets] [NumPy] #373: Yet another valgrind warning
NumPy
numpy-tickets at scipy.net
Sat Nov 4 07:30:15 CST 2006
#373: Yet another valgrind warning
--------------------+-------------------------------------------------------
Reporter: faltet | Owner: somebody
Type: defect | Status: new
Priority: normal | Milestone:
Component: Other | Version:
Severity: normal | Keywords:
--------------------+-------------------------------------------------------
Hi,
Valgrind is detecting this in some tests of mine:
{{{
==12109== Invalid read of size 4
==12109== at 0x446374: PyObject_Free (obmalloc.c:919)
==12109== by 0x6010F87: array_dealloc (arrayobject.c:1930)
==12109== by 0x4CD67B: frame_dealloc (frameobject.c:416)
==12109== by 0x4833ED: PyEval_EvalFrameEx (ceval.c:3654)
==12109== by 0x4833CA: PyEval_EvalFrameEx (ceval.c:3652)
==12109== by 0x4845B0: PyEval_EvalCodeEx (ceval.c:2833)
[snip]
==12109== Conditional jump or move depends on uninitialised value(s)
==12109== at 0x44637D: PyObject_Free (obmalloc.c:919)
==12109== by 0x6010F87: array_dealloc (arrayobject.c:1930)
==12109== by 0x6010F87: array_dealloc (arrayobject.c:1930)
==12109== by 0x4CD67B: frame_dealloc (frameobject.c:416)
==12109== by 0x4833ED: PyEval_EvalFrameEx (ceval.c:3654)
==12109== by 0x4845B0: PyEval_EvalCodeEx (ceval.c:2833)
==12109== by 0x4CF0D3: function_call (funcobject.c:517)
==12109== by 0x415C32: PyObject_Call (abstract.c:1860)
==12109== by 0x4817F8: PyEval_EvalFrameEx (ceval.c:3846)
==12109== by 0x4845B0: PyEval_EvalCodeEx (ceval.c:2833)
==12109== by 0x4CF0D3: function_call (funcobject.c:517)
==12109== by 0x415C32: PyObject_Call (abstract.c:1860)
==12109==
==12109== Use of uninitialised value of size 8
==12109== at 0x446396: PyObject_Free (obmalloc.c:919)
==12109== by 0x6010F87: array_dealloc (arrayobject.c:1930)
==12109== by 0x6010F87: array_dealloc (arrayobject.c:1930)
==12109== by 0x4CD67B: frame_dealloc (frameobject.c:416)
==12109== by 0x4833ED: PyEval_EvalFrameEx (ceval.c:3654)
==12109== by 0x4845B0: PyEval_EvalCodeEx (ceval.c:2833)
}}}
Is self.base being always initialized properly? This is using pure !NumPy
(i.e. without doing conversions with array protocol). I'm using r3428, so
the affected line should be seen in:
source:/trunk/numpy/core/src/arrayobject.c#3428
Sorry, but again, this come without a code snippet :(
--
Ticket URL: <http://projects.scipy.org/scipy/numpy/ticket/373>
NumPy <http://projects.scipy.org/scipy/numpy>
The fundamental package needed for scientific computing with Python.
More information about the Numpy-tickets
mailing list