[SciPy-dev] build error on OSX 10.3

John Hunter jdh2358@gmail....
Sun Oct 14 07:18:38 CDT 2007


On 10/14/07, David Cournapeau <david@ar.media.kyoto-u.ac.jp> wrote:

> This seems to be 10.3 specific (it does not appear on 10.4). The problem
> seems to be that on panther, veclib does not define the complex type. If
> you know where to get information on the veclib, I would be grateful,
> because I didn't find a lot of information (in general, I find the
> documentation on apple frameworks/libs really sparse, or maybe I just do
> not know where to look).

OK, I added the complex.h header as described there to
scipy/lib/blas/fblaswrap_veclib_c.c.src and
scipy/linalg/src/fblaswrap_veclib_c.c and was able to get scipy
compiled.  I read a few of the links in the page you cited, and the
only thing I could find about the brokenness of complex in veclib was:

   > Complex numbers support may be broken (some problems with
passing/returning
   > complex values on 64-bit targets, and not checked against the
requirements of the
   > C99 standard). However, the _Complex keyword now works.

Are any 10.3 platforms 64 bit?  If not, there may be little danger in
simply adding it, or conditionally adding it for non-64 bit platforms.

The other question is: what has recently changed vis-a-vis veclib that
is now causing the compile on 10.3 to break.  I compiled scipy from
svn several months ago, though I don't have the  revision number off
the top of head.

I had 9 errors on my test, but all were related to loadmat.  My guess
is this is that endianess is not being handled properly in the test.
Everything else appears to pass.

I'm attaching a patch against r3435 if you decide this should be added.


JDH
-------------- next part --------------
A non-text attachment was scrubbed...
Name: veclib.patch
Type: text/x-patch
Size: 671 bytes
Desc: not available
Url : http://projects.scipy.org/pipermail/scipy-dev/attachments/20071014/4a069f67/attachment.bin 


More information about the Scipy-dev mailing list