[SciPy-dev] NAN in special/cephes breaks CVS build

Travis Oliphant oliphant.travis at ieee.org
Sat Feb 23 11:40:20 CST 2002

On Saturday 23 February 2002 09:45 am, you wrote:
> On Sat, 23 Feb 2002, Prabhu Ramachandran wrote:
> > >>>>> "PP" == Pearu Peterson <pearu at cens.ioc.ee> writes:
> >
> >     PP> Ok, I found that one needs to use -D_ISOC99_SOURCE when
> >     PP> compiling.  See /usr/include/features.h for comments and
> >     PP> /usr/include/bits/nan.h.
> >
> > I just updated my CVS copy and ran a python setup.py build and it
> > seems to build fine on woody.

Actually, const.c defines NAN already and so I don't understand why you would 
get this error.  The file iv.c includes an extern double NAN line.

> Yes, because travo already fixed this. However, now I get

Sorry, I added distributions.py to the __init__ file of stats but forgot to 
add the file to the CVS tree.  It's there now (or you can just delete the 
reference to distributions in stats/__init__.py  --- the problem is not fft.


More information about the Scipy-dev mailing list