[Python-modules-team] Bug#466868: python-scipy -- please transition to use gfortran based lapack, blas

[Ondrej Certik]

On Thu, Feb 21, 2008 at 4:31 PM, Ondrej Certik <ondrej at certik.cz> wrote:
> On Thu, Feb 21, 2008 at 4:11 PM, Kumar Appaiah <akumar at ee.iitm.ac.in> wrote:
>  > Package: python-scipy
>  >  Version: 0.6.0-5.1
>  >  Severity: important
>  >  User: debian-toolchain at lists.debian.org
>  >  Usertags: gfortran
>  >
>  >  Hi!
>  >
>  >  I would request you to please shift Build-Depends to use the new
>  >  gfortran based Lapack and Blas packages in order to phase out packages
>  >  dependent on the old g2c (g77). In this regard, I would request you to
>  >  refer to:
>  >
>  >  http://wiki.debian.org/GfortranTransition
>  >
>  >  Thank you.
>
>  Thanks for the bug report. We just uploaded a new revision, that makes scipy
>  installable again with python-numpy. Also, new gfortran based
>  suitesparse, which scipy depends on,
>  will be uploaded to unstable either today or tomorrow. When it enters unstable,
>  we'll rebuild scipy against this new suitesparse and upload a new
>  revision of scipy.
>
>  This should fix this bug.

Updated and fixed package sits in our svn. Please review, if it's ok,
I'll upload tomorrow.

Pontential problems:

    0.6.0-7:

    FAILED (failures=2)
    Out[2]: <unittest._TextTestResult run=1728 errors=0 failures=2>

0.6.0-8:

  In [2]: scipy.test()
  [...]
  FAILED (failures=5)
  Out[2]: <unittest._TextTestResult run=1728 errors=0 failures=5>


I think those errors are not serious, as the scipy package have some
tests failing anyway and its more important to get the gfortran based
package
first. We can then report all the bugs upstream. What do you think?

Ondrej