[Numpy-discussion] numpy broken by r8077?

Fernando Perez fperez.net@gmail....
Thu Jan 21 20:07:30 CST 2010


Hi all,

This simple-looking change:

http://projects.scipy.org/numpy/changeset/8077

is giving me a wholly broken (unimportable) numpy on ubuntu 9.10, 64-bit:

uqbar[junk]> python -c 'import numpy'
Traceback (most recent call last):
  File "<string>", line 1, in <module>
  File "/home/fperez/usr/opt/lib/python2.6/site-packages/numpy/__init__.py",
line 136, in <module>
    import add_newdocs
  File "/home/fperez/usr/opt/lib/python2.6/site-packages/numpy/add_newdocs.py",
line 9, in <module>
    from numpy.lib import add_newdoc
  File "/home/fperez/usr/opt/lib/python2.6/site-packages/numpy/lib/__init__.py",
line 4, in <module>
    from type_check import *
  File "/home/fperez/usr/opt/lib/python2.6/site-packages/numpy/lib/type_check.py",
line 8, in <module>
    import numpy.core.numeric as _nx
  File "/home/fperez/usr/opt/lib/python2.6/site-packages/numpy/core/__init__.py",
line 5, in <module>
    import multiarray
ImportError: /home/fperez/usr/opt/lib/python2.6/site-packages/numpy/core/multiarray.so:
undefined symbol: PyUnicodeUCS2_AsASCIIString

I reverted to 8076 and it's ok now:

uqbar[junk]> python -c 'import numpy;numpy.test()'
Running unit tests for numpy
NumPy version 1.5.0.dev8076
NumPy is installed in /home/fperez/usr/opt/lib/python2.6/site-packages/numpy
Python version 2.6.4 (r264:75706, Dec  7 2009, 18:43:55) [GCC 4.4.1]
nose version 0.11.1
..........................................................................
----------------------------------------------------------------------
Ran 2506 tests in 7.083s

OK (KNOWNFAIL=5, SKIP=4)


The change is very small and was logged as fixing another bug, but it
seems something is amiss now.  If it was committed like that I imagine
it works on some systems, so is it a problem on my end?  I'm using the
system python (2.6) and it's an otherwise fairly standard ubuntu box,
were I only run ipython, numpy, scipy and matplotlib from dev trees
and everything else is stock from the distro.

Any help much appreciated,

f


More information about the NumPy-Discussion mailing list