[Numpy-discussion] Stop generating include files?

[Charles R Harris]

Hi All,

I would like to suggest that it is time to stop generating the include files
by scanning the (c) sources. I think this would be a good thing to do for
the following reasons.

   1. The Numpy API has settled down and we shouldn't be changing is much.
   2. It is good practice to separate the declarations and definitions so
   that they act as checks on one another.
   3. It simplifies the build system. Instead of lists of scanned files,
   lists of API order, and a generator, we simply have *.h files.
   4. A simplified build system makes it easier to break up files and
   generate API functions from code templates.
   5. The *.h files will be available in Numpy *before* a build. This should
   make it easier for newbies to figure out what is going on.

I don't suggest we remove the various API tags, they are useful bits of
documentation when reading the code.

Chuck
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/numpy-discussion/attachments/20081123/c10faf96/attachment.html>