[Pyrex] [Cython-dev] Test case for Pyrex and Cython

[Jim Kleckner]

Jim Kleckner wrote:
> Greg Ewing wrote:
>> Jim Kleckner wrote:
>>> This could be done either by Pyrex permitting naming
>>> of the form package/module.pyx (instead of requiring
>>> the form package/package.module.pyx)
>> That can just be package.module.pyx -- it doesn't need
>> to be in a subdirectory as well.
> 
> To minimize confusion of other readers of the list,
> it was stated earlier in the thread that Pyrex works
> with fully-qualified file names.  This suggestion is
> in the context of how to obtain better Pyrex and Cython
> compatibility.  Adding this form permits testing of
> both without rework of file layout and has the added
> benefit that David pointed out of being consistent with
> other languages (and somewhat less redundant).

On the other hand, the advantage to have Cython permit
the form of "package.module.pyx" makes it possible to
contain all the files for a package in one directory.
You can put a Makefile, headers, and supporting C code
all in one place, possibly even not using setup.py
at all (which I think was one of Greg's stated goals
with this naming method).  Cython's forcing of an
extra directory somewhat dictates the use of setup.py.