[SciPy-User] Building Scipy for Mac OS X 10.6

Ralf Gommers ralf.gommers@googlemail....
Wed Jun 16 10:40:45 CDT 2010


On Tue, Jun 8, 2010 at 11:22 AM, Scott Stephens <stephens.js@gmail.com>wrote:

> On Mon, Jun 7, 2010 at 1:23 AM, David <david@silveregg.co.jp> wrote:
> > On 06/07/2010 12:42 PM, Scott Stephens wrote:
> >> On Sat, Jun 5, 2010 at 10:49 PM, Ralf Gommers
> >> <ralf.gommers@googlemail.com>  wrote:
> >>> Could you try the 0.8.0 beta (
> http://sourceforge.net/projects/scipy/files/)?
> >>> Some of this may be fixed, and you also get another 18 months worth or
> so of
> >>> new features and bug fixes.
> >>
> >> 0.8.0b1 doesn't build for me using numscons.  Build log is attached.
> >> I'm using numpy 1.4.1 built from source (and tested, no errors or
> >> unknown failures) and numscons freshly checked out from the git
> >> repository (at least numscons-0.11 is required to build scipy-0.8.0b1,
> >> and only numscons-0.10 is available via easy_install; couldn't find
> >> source for numscons-0.11).
> >
> > Could you see whether r6487 fixes it for you ?
> >
>
> Checked out and built the trunk (r6490).  r6487 fixed the build
> problem, and some of the test failures I was getting in 0.7.1, but not
> all.  The errors and failures I get now are:
>
> ERROR: test_decomp.test_lapack_misaligned(<function solve at
> 0x103da4f50>, (array([[  1.734e-255,   8.189e-217,   4.025e-178,
> 1.903e-139,   9.344e-101,
>
ERROR: test_complex_nonsymmetric_modes
> (test_arpack.TestEigenComplexNonSymmetric)
> ERROR: test_nonsymmetric_modes (test_arpack.TestEigenNonSymmetric)
> ERROR: test_starting_vector (test_arpack.TestEigenNonSymmetric)
> ERROR:
> test_continuous_basic.test_cont_basic(<scipy.stats.distributions.wald_gen
> object at 0x104608e50>, (), 'wald')
> ERROR:
> test_continuous_basic.test_cont_basic(<scipy.stats.distributions.wald_gen
> object at 0x104608e50>, (), 'wald')
> ERROR:
> test_continuous_basic.test_cont_basic(<scipy.stats.distributions.wald_gen
> object at 0x104608e50>, (), 'wald')
> FAIL: test_complex_symmetric_modes (test_arpack.TestEigenComplexSymmetric)
>
> Full text of the test run is attached.
>

The lapack_misaligned one is a known error. The wald errors were fixed in
r6495. The arpack errors seem to be due to a lapack issue, see
http://thread.gmane.org/gmane.comp.python.scientific.devel/8551

Cheers,
Ralf
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.scipy.org/pipermail/scipy-user/attachments/20100616/daf6a913/attachment.html 


More information about the SciPy-User mailing list