[vtk-developers] Strange Python bug in latest CVS tree?

[Prabhu Ramachandran]

hi,

I just ran into this strange bug with the Python wrappers for VTK.

>>> from vtkpython import *
>>> a = vtkScalarBarActor()
>>> a.GetBounds()
Program received signal SIGSEGV, Segmentation fault.
0x41b211d9 in PyvtkProp_GetBounds ()
   from /skratch/prabhu/vtk/cvs/vtk/common/libVTKCommonPython.so
(gdb) back
#0  0x41b211d9 in PyvtkProp_GetBounds ()
   from /skratch/prabhu/vtk/cvs/vtk/common/libVTKCommonPython.so
#1  0x8059da6 in call_cfunction (func=0x8122118, arg=0x80d2454, kw=0x0)
    at Python/ceval.c:2838
#2  0x8059c98 in call_object (func=0x8122118, arg=0x80d2454, kw=0x0)
    at Python/ceval.c:2801
<...>

The segfault seems to happen with the vtkProp class's method.

>>> from vtkpython import *
>>> a = vtkProp()
>>> a.GetBounds()
Segmentation fault

Now, the build that I have was not made after a 'make clean'.  I just
reconfigured and rebuilt it.  Can someone confirm if this is a bug or
do I just need a make clean and a fresh rebuild?

Also, how do you folks know when to do a fresh rebuild and when not
to?

BTW, this problem did not occur earlier with the CVS tree (I cant
remember how long ago I checked) but I can confirm that it doesnt
happen with the 3.2 release.

thanks,
prabhu