[SciPy-User] ANN: NumPy 1.6.0 - FAIL: test_expon (test_morestats.TestAnderson)

[Yury V. Zaytsev]

Hi!

On Sat, 2011-05-14 at 11:54 +0200, Ralf Gommers wrote:
> 
> I am pleased to announce the release of NumPy 1.6.0. This release is
> the result of 9 months of work, and includes many new features,
> performance improvements and bug fixes. Some highlights are:

Congratulations! Absolutely great news!

I have started to update my private builds on Ubuntu Maverick 64-bit:

$ uname -a 
Linux newpride 2.6.35-28-generic #50-Ubuntu SMP Fri Mar 18 18:42:20 UTC 2011 x86_64 GNU/Linux

I am building against ATLAS optimized for Core i7:

sudo apt-get install libatlas3gf-corei7sse3 libatlas3gf-corei7sse3-dev

export BLAS=/usr/lib/libblas.so
export LAPACK=/usr/lib/liblapack.so
export ATLAS=/usr/lib/libatlas.so

pip install numpy
pip install scipy

The NumPy builds fine and passes the test suite, however, when I rebuild
latest SciPy 0.9.0 against latest NumPy 1.6.0 I am getting one test
failure: 

======================================================================
FAIL: test_expon (test_morestats.TestAnderson)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/srv/virtualenv/mle/lib/python2.7/site-packages/scipy/stats/tests/test_morestats.py", line 72, in test_expon
    assert_array_less(crit[:-1], A)
  File "/srv/virtualenv/mle/lib/python2.7/site-packages/numpy/testing/utils.py", line 869, in assert_array_less
    header='Arrays are not less-ordered')
  File "/srv/virtualenv/mle/lib/python2.7/site-packages/numpy/testing/utils.py", line 613, in assert_array_compare
    chk_same_position(x_id, y_id, hasval='inf')
  File "/srv/virtualenv/mle/lib/python2.7/site-packages/numpy/testing/utils.py", line 588, in chk_same_position
    raise AssertionError(msg)
AssertionError: 
Arrays are not less-ordered

x and y inf location mismatch:
 x: array([ 0.911,  1.065,  1.325,  1.587])
 y: array(inf)

----------------------------------------------------------------------

This doesn't happen if I build SciPy 0.9.0 against NumPy 1.5.0, all the
tests pass.

Hope that helps to figure out the problem,

-- 
Sincerely yours,
Yury V. Zaytsev