pnmimage crashes if x/y coodinate outside image

[Hypnos]

crashes on: panda3d 1.6.1 on jaunty:

example code:
from pandac.PandaModules import *
i = PNMImage(256,256)
for x in xrange(1024):
  for y in xrange(1024):
    i.setXel(x,y,1)

can eighter segfault or "*** glibc detected *** python: malloc(): memory corruption: 0x0956cd70 ***", depending which values are written, in which order.

for some values it crashes for others it doesnt. it's really not predictable.

[Treeform (starplant)]

I wonder what it should return instead? Should it wrap the image or
raise exception?

On Tue, Jun 2, 2009 at 12:06 AM, pro-rsoft <email address hidden> wrote:
> *** This bug is a duplicate of bug 372253 ***
>    https://bugs.launchpad.net/bugs/372253
>
> ** This bug has been marked a duplicate of bug 372253
>   pnmimage crashes if x/y coodinate outside image
>
> --
> pnmimage crashes if x/y coodinate outside image
> https://bugs.launchpad.net/bugs/372254
> You received this bug notification because you are subscribed to
> Panda3D.
>

[Hypnos (rspoerri)]

I'd make it raise a exception. It's bad that it sometimes goes
undetected and other times completely segfaults python.

on the other hand i've talked with pro-rsoft, he said pnmimage is
implemented to be as fast as possible. While adding a checking mechanism
in c/c++ would slow down the implementation where it needs to be fast,
it definately should be checked when accessing from python (if that's
possible).

Treeform wrote:
> *** This bug is a duplicate of bug 372253 ***
> https://bugs.launchpad.net/bugs/372253
>
> I wonder what it should return instead? Should it wrap the image or
> raise exception?
>
> On Tue, Jun 2, 2009 at 12:06 AM, pro-rsoft <email address hidden> wrote:
>
>> *** This bug is a duplicate of bug 372253 ***
>> https://bugs.launchpad.net/bugs/372253
>>
>> ** This bug has been marked a duplicate of bug 372253
>> pnmimage crashes if x/y coodinate outside image
>>
>> --
>> pnmimage crashes if x/y coodinate outside image
>> https://bugs.launchpad.net/bugs/372254
>> You received this bug notification because you are subscribed to
>> Panda3D.
>>
>>
>
>