[SciPy-user] failed easy_install on OSX

Robert Kern robert.kern@gmail....
Sat Feb 7 16:27:33 CST 2009

On Sat, Feb 7, 2009 at 16:13, Dav Clark <dav@alum.mit.edu> wrote:
> Hi,
> I found a small bug (more with OSX than with SciPy) but worth
> mentioning.
> If you upgrade setuptools on OS X without changing your path, for some
> reason /usr/bin/easy_install (system setuptools 0.6c7) will remain
> ahead of /usr/local/bin/easy_install (your current install).  Then, if
> you try to do an easy_install of scipy, it fails because setuptools
> 0.6c7 doesn't provide the proper fcompiler attribute.

No version of setuptools provides an fcompiler attribute. That's all
numpy.distutils. I suspect there is a different problem going on. The
system's Python comes with a 1.0.x series numpy. I think that is the
root of the problem.

> Two solutions:
> 1) download and run setup.py manually - this will use the most recent
> setuptools via python
> or
> 2) Change your PATH so that /usr/local/bin comes before /usr/bin.  Why
> this isn't the case already, I have no idea.  I guess it's apples way
> of insulating casual users from hackers like us.

I believe all of the Python binaries I am aware of (www.python.org,
Activestate, and EPD) will modify your .bashrc or .bash_profile to
place the appropriate bin/ path (not always /usr/local/bin/!) at the
front of your $PATH. If you are using a different shell, you may have
to do this manually. Additionally, the default installation location
for scripts is not /usr/local/bin/ but
/Library/Frameworks/Python.framework/Versions/Current/bin/, so I
suspect you have modified your .pydistutilsrc file to point there.
When you modify things, you are on your own.  :-)

> I would like to put this advice here:
> http://www.scipy.org/Installing_SciPy/Mac_OS_X
> But I don't have permission.  If you want to give me permission, I am
> DavClark on the scipy.org wiki.

You have now been added to the EditorsGroup.

Robert Kern

"I have come to believe that the whole world is an enigma, a harmless
enigma that is made terrible by our own mad attempt to interpret it as
though it had an underlying truth."
  -- Umberto Eco

More information about the SciPy-user mailing list