[pypy-dev] GSOC Ideas

[Stefan Behnel]

Maciej Fijalkowski, 28.03.2012 10:28:
> On Wed, Mar 28, 2012 at 9:33 AM, Stefan Behnel wrote:
>> Armin Rigo, 23.03.2012 07:55:
>>> On Fri, Mar 23, 2012 at 00:13, Lynne Wallace wrote:
>>>> I was wondering if anyone else had ideas or suggestions for gsoc projects?
>>>
>>> We have a list of projects (which might or might not fit a GSoC size):
>>> http://doc.pypy.org/en/latest/project-ideas.html
>>
>> I'm missing the project to fix up and optimise cpyext from that list. I
>> think that would be one of the most widely rewarding in the short run. And
>> it has the advantage for a GSoC that optimisation is generally open-ended,
>> so it can stop at any time with a positive result.
> 
> Indeed.
> 
> Can you write down the necessary paragraph? You can just commit it
> (it's pypy/doc/project-ideas.rst) or I'll commit it if you don't have
> commit privs.

I don't, so here's some plain text:

"""
Optimising cpyext (CPython C-API compatibility layer)

A lot of work has gone into PyPy's implementation of CPython's C-API over
the last years to let it reach a practical level of compatibility, so that
C extensions for CPython work on PyPy without major rewrites. However,
there are still many edges and corner cases where it misbehaves, and it has
not received any substantial optimisation so far.

The objective of this project is to fix bugs in cpyext and to optimise
several performance critical parts of it, such as the reference counting
support and other heavily used C-API functions. The net result would be to
have CPython extensions run much faster on PyPy than they currently do, or
to make them work at all if they currently don't. A part of this work would
be to get cpyext into a shape where it supports running Cython generated
extensions.
"""

Stefan