Re: New Bitbucket/Mercurial copy of the SVN repository

On 5 Jul, 2011, at 17:50, Leonardo Santagada wrote:

> On Tue, Jul 5, 2011 at 12:21 PM, Lawrence Akka
> <[email protected]> wrote:
>> 4) Clang: interesting. I have been playing with clang compilation of
>>   PyObjC generally, today. Had a couple of problems with libxml though.
> 
> Any chance that while rewriting you do a pure python version of
> pyobjc? pyglet is using a pure python bridge to objective-c because
> getting pyobjc compiled is a pain. And specially now that pypy can run
> ctypes faster than python c extensions I would say it is time to think
> about it.

It is unlikely that I'll rewrite PyObjC using ctypes anytime soon. Porting to PyPy would need some rearchitecturing anyway as the current implementation is pretty tightly bound to the CPython implementation anyway.

To be honest I'm more likely to work on a fork of CPython where a PyObject* is toll-free bridged to an Objective-C class, that would solve some problems were currently having and might open up other interesting avenues.

BTW. What problems are you haveing with compiling PyObjC? Would those be solved by having a single archive with an install script that compiles the various bit of PyObjC in the right order?

Ronald

------------------------------------------------------------------------------
All of the data generated in your IT infrastructure is seriously valuable.
Why? It contains a definitive record of application performance, security 
threats, fraudulent activity, and more. Splunk takes this data and makes 
sense of it. IT sense. And common sense.
http://p.sf.net/sfu/splunk-d2d-c2