[Numpy-discussion] arrayobject.h (and friends) install location

[pearu at cens.ioc.ee]

On Mon, 16 Jan 2006, Sasha wrote:

> > Yes, this has been discussed before and the discussion is in scipy-dev
> > archives somewhere.. The location of numpy header files is well-documented
> > (we have numpy/doc/CAPI.txt and numpy.get_numpy_include()), so I don't
> > understand what is the problem with the current situation.
> 
> The problem with the current situation is that the include files end
> up in a rather obscure
> place. Is there any other python extension package that installs its
> include files in site-packages?  Can someone summarize the reasons for
> the change so that this problem
> can find its resting place in a FAQ somewhere? What problem does the
> new location solve that cannot be solved with the --prefix option?

See the threads starting at

 http://www.scipy.org/mailinglists/mailman?fn=scipy-dev/2005-October/003420.html
 http://www.scipy.org/mailinglists/mailman?fn=scipy-user/2005-October/005454.html

Basically, the main argument behind the current decision is that there are
systems where people don't have permissions to install header files to std
python include directory but they can install to python site-packages
(e.g. when installing prebuilt distributions). 

Also note that not all potential numpy users are able/willing/.. to use
--prefix option. Why we should scare them away when we can provide
defaults that work for all situations, though the defaults are not
standard and require some tolerance from developers to follow numpy docs..

Pearu