[Numpy-discussion] recompiling needed for binary module after numpy 1.0

belinda thom bthom at cs.hmc.edu
Mon Jan 8 01:04:01 EST 2007

On Jan 5, 2007, at 6:01 PM, Robert Kern wrote:

> belinda thom wrote:
>> On Jan 5, 2007, at 5:32 PM, Sebastian Haase wrote:
>>> Hi,
>>> All I did is recompiling my (on source code file) C extension. I  
>>> made
>>> sure that it was including the current numpy header files.
>> Where are these files located? What command did you use? (The gorey
>> details would help me quite a bit, as I'm to compile-level installs,
>> having usually relied on things like macports port to do the work for
>> me)...
> Actually, it's more likely to confuse the matter since Sebastian is  
> using a
> quite nonstandard way of building Python extension modules.
> What packages do you want help with building from source? Which build
> instructions have you read? What issues still remain unclear for  
> you after
> having read them? What platforms are you concerned with?


I am running on Macs, Intels, G4s, and G5s using OS X 10.4.8.

I am using MacPython 2.4.4 (whose universal installer successfully  
installs on both types of machines). I might consider using 2.5, but  
a core of some externally developed coursework doesn't run in 2.5.

Currently I have kluged some disparate parts together to get a  
working system (by which I mean interactive matplotlib, numpy, and  
scipy are all working and pass their respective tests [matplotlib  
doesn't appear to come with a test suite, so I couldn't test that one  
as extensively]). I understand that Erin Sheldon has had no problems  
using Fink to set up this integrated configuration, but I tend to  
avoid Fink b/c I've heard it doesn't always play nice w/other Mac  
internals. Besides, I like macports.

More info on my kluges can be found at: http://www.mail-archive.com/ 
numpy-discussion at scipy.org/msg00481.html, where Chris had mentioned  
the need to provide an integrated set of Mac OS X bundle(s?) to do  
this job. I am not in a position to do this because I don't know how  
to install these packages from source. Others have told me only  
expert users tend to foray into such an area. At the same time, if I  
wish to have my students set up their computers and use this  
development environment, I'd like it to be as easy as possible to do.

There's other reasons --- mostly curiousity driven --- that motivate  
me to try a "rawer" (e.g. source, non-pkg based) install:

1) My current kludge only supports TkAgg, so I can't play w/the wx  
backend at all. Have no idea why
2) My current kludge only works w/numpy; under Numeric, matplotlib  
plotting dies.
3) If I used the numpy/scipy bundles from http://www.macpython.org/,  
I'd get the C API error Sebastian was originally talking about. I  
have no idea how to fix this.

I have just completed installing ipython from source by running:

   python setup.py install

in a directory from which I downloaded ipython via subversion. That  
was pretty painless (this is my first such install of a relatively  
big program, BTW). If scipy, numpy and matplotlib could be just as  
painlessly derived, I'd give it a try.

Thanks for any comments,


More information about the NumPy-Discussion mailing list