[SciPy-user] SciPy on OS X installation problem: ImportError: No module named scipy_distutils

Michael Williams williams at thphys.ox.ac.uk
Sun Oct 23 08:02:22 CDT 2005


I'd like to install SciPy on OS X. I'm following a combination of the  
instructions on these two pages:


I'm running OS X 10.4 and have previously upgraded from Tiger's stock  
Python 2.3.5 to Python 2.4.1 using Bob Ippolito's MacPython installer:


Most of the steps described on the two guides work as advertised.  
However, I had a warning while building FFTW (described below), and a  
show-stopping failure in SciPy's setup.py build.

I checked SciPy out from SVN around twelve hours ago. When building I  
get the following error:

> ~/Desktop/scipy mike$ python setup.py build
> Traceback (most recent call last):
>   File "setup.py", line 47, in ?
>     import scipy_distutils
> ImportError: No module named scipy_distutils

 From Googling, I understand the scipy_distutils module comes with  
F2PY. I've double checked my installation of this package, and I'm  
sure it's there (it's in /Library/Frameworks/Python.framework/ 
Versions/2.4/lib/python2.4/site-packages/f2py2e, as expected). The  
version of F2PY I've installed is 2.45.241_1926.

Does anyone have any suggestions? Should I try an earlier version of  

By the way, as mentioned above, when I ran ./configure on FFTW I got  
the following warning:

checking for Fortran 77 libraries...  -lcrt1.o -lcrt2.o -L/usr/local/ 
lib/gcc/powerpc-apple-darwin7.9.0/3.4.4 -L/usr/local/lib/gcc/powerpc- 
apple-darwin7.9.0/3.4.4/../../.. -lfrtbegin -lg2c -lgcc_s -lSystem
checking for dummy main to link with Fortran 77 libraries... unknown
configure: WARNING: *** Couldn't figure out how to link C and  
Fortran; switching to --disable-fortran.

FFTW went on to compile and install without further problems. I don't  
think this is related to SciPy's failure to build, but I describe it  
on the off chance that it's related, or anyone knows how to fix it,  
or whether it even needs fixing.

Thanks for reading this far,
Mike Williams
Theoretical Physics, University of Oxford

More information about the SciPy-user mailing list