[Numpy-discussion] long double woes on win32

David Cournapeau cournape@gmail....
Sat Aug 16 12:10:58 CDT 2008


On Sat, Aug 16, 2008 at 11:15 AM, Charles R Harris
<charlesr.harris@gmail.com> wrote:

>
> That's my opinion also, I just thought that -DNOLONGDOUBLE was an easy way
> to force that choice. David thinks that the function detection in the ufunc
> module will be a problem.

Forget what I said, I think I used a broken mingw toolchain, which
caused the problems I got. In any case, I cannot reproduce the problem
anymore.

But forcing SIZEOF_LONG_DOUBLE to 8 at configuration stage caused test
failures too (at different points, though):

http://scipy.org/scipy/numpy/ticket/890

The first one looks strange.


More information about the Numpy-discussion mailing list