[Python-modules-team] Bug#544274: python-numpy: conflict with libatlas3gf-base

[Johan Grönqvist]

Package: python-numpy
Version: 1:1.3.0-3
Severity: normal


I believe I have the same issue.

My test script looks like:

import numpy as np
for ix in range(2, 300):
    print ix
    m = np.random.random((ix, ix))
    (x, v) = np.linalg.eig(m)
    print v.shape

Running the above code with python-numpy _and_ libatlas3gf-base installed 
generates a glibc memory management error message. Running the same
script with python-numpy installed but _without_ libatlas3gf-base
installed is unproblematic for me. 

Both unstable and testing show the same behaviour.

Exactly at what stage in the python program the error appears seems
non-repeatable, but it always appears for the code above.


See also <http://article.gmane.org/gmane.comp.python.numeric.general/33548/match=>

/ johan


-- System Information:
Debian Release: squeeze/sid
  APT prefers testing
  APT policy: (900, 'testing'), (600, 'stable')
Architecture: amd64 (x86_64)

Kernel: Linux 2.6.30-2-amd64 (SMP w/2 CPU cores)
Locale: LANG=sv_SE.UTF-8, LC_CTYPE=sv_SE.UTF-8 (charmap=UTF-8)
Shell: /bin/sh linked to /bin/dash

Versions of packages python-numpy depends on:
ii  libblas3gf [libblas.so.3gf]   1.2-2      Basic Linear Algebra Subroutines 3
ii  libc6                         2.9-25     GNU C Library: Shared libraries
ii  libgcc1                       1:4.4.1-4  GCC support library
ii  libgfortran3                  4.4.1-4    Runtime library for GNU Fortran ap
ii  liblapack3gf [liblapack.so.3g 3.2.1-2    library of linear algebra routines
ii  python                        2.5.4-2    An interactive high-level object-o
ii  python-central                0.6.11     register and build utility for Pyt

python-numpy recommends no packages.

Versions of packages python-numpy suggests:
pn  python-nose                   <none>     (no description available)
ii  python-numpy-dbg              1:1.3.0-3  Fast array facility to the Python 
ii  python-numpy-doc              1:1.3.0-3  NumPy documentation

-- no debconf information