Package: python-scipy
Version: 0.3.2-6
A known bug in scipy is still in the stable package of python-scipy
on Sarge. The following script will terminate the python interpreter
with a segmentation fault:
import scipy
scipy.apply_along_axis(scipy.linalg.norm, 1, scipy.zeros((128,3000)))
What it should do is to create a larger matrix with 3000 lines of 128
zeros each and then calculate the 'norm' for every line. Using norm
is not part of the problem, the segmentation fault is only due to the
size of the array (it is a dereferencing issue). Please see the scipy
bug tracker:
http://scipy.net/roundup/scipy/issue218
http://scipy.net/roundup/scipy/issue202
Obviously, a fix is available (issue 202) and already in the CVS
version of scipy.
Yours, Mark
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
