[SciPy-User] Building Scipy for Mac OS X 10.6
Sat Jun 5 22:49:38 CDT 2010
On Sat, Jun 5, 2010 at 8:20 AM, Scott Stephens <email@example.com>wrote:
> On Fri, Jun 4, 2010 at 7:31 AM, Ralf Gommers
> <firstname.lastname@example.org> wrote:
> > On Fri, Jun 4, 2010 at 8:16 PM, Scott Stephens <email@example.com>
> > wrote:
> >> Looks like all of the created files have the right architectures, and
> >> nothing in the unit tests fail because of import errors. I'm still not
> >> getting a flawless scipy.test() run though (6 errors and 1 failure).
> >> I haven't yet had a chance to do a deep dive and figure out what
> >> exactly is failing and if it might be related to build problems.
> > Can you post the errors and failure? Currently one test gives an error,
> > test_lapack_misaligned. Others that have been failing in recent checkouts
> > are one lambertw test and several matlab ones, so no need to investigate
> > those.
> Here they are:
> ERROR: test_complex_nonsymmetric_modes
> ERROR: test_nonsymmetric_modes (test_arpack.TestEigenNonSymmetric)
> ERROR: test_starting_vector (test_arpack.TestEigenNonSymmetric)
> ERROR: test_add_function_ordered (test_catalog.TestCatalog)
> ERROR: Test persisting a function in the default catalog
> (test_add_function_persistent1 from
> ERROR: Shouldn't get a single file from the temp dir.
> (test_get_existing_files2 from scipy/weave/tests/test_catalog.py)
> FAIL: test_complex_symmetric_modes (test_arpack.TestEigenComplexSymmetric)
> This is scipy-0.7.1 from a tarball.
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.
> I'm wondering if if some of the
> nonsymmetric failures are related to me not installing UMFPACK. At
> some point in my build adventure I got some messages about UMFPACK not
> being found, but nothing about it appears in the installation
> instructions, so I haven't done anything with it.
> UMFPACK is not required for scipy and it not being installed shouldn't
result in test failures.
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the SciPy-User