[Numpy-discussion] Generalised ufuncs
Sun Nov 9 23:15:37 CST 2008
On Mon, Nov 10, 2008 at 9:46 AM, Stéfan van der Walt <email@example.com> wrote:
> Hi everyone,
> I finally merged the generalised ufunc patches that awaited comment as
> the gen_ufuncs branch. Please review the changes on trunk, and let me
> know if problems occur.
> One of the tests fail on FreeBSD 64:
> ERROR: Failure: ImportError
> Undefined symbol "cblas_dgemm")
> Is there any reason why blasdot should work fine, but that this should
> be broken?
The obvious question is what is umath_test.so :) We don't generate
this file AFAIK, and I certainly cannot see it mentioned anywhere in
numpy source code (and neither in my mac os X or linux installation).
FreeBSD would also require some minor modifications to
numpy.distutils, I believe. I have just noticed a bug which may be
related in numpy.distutils on a fresh ubuntu install, where numpy uses
the C compiler to link atlas when no fortran compiler is available (it
should fail instead, since we use the fortran interface in that case,
and that cannot work without some fortran support, obviously).
More information about the Numpy-discussion