[PEAK] ruledispatch and python 2.5

Phillip J. Eby pje at telecommunity.com
Tue Oct 31 12:16:33 EST 2006

At 05:58 PM 10/30/2006 -0500, Kevin Dangoor wrote:
>On Oct 29, 2006, at 6:02 PM, Phillip J. Eby wrote:
>>At 01:29 PM 10/29/2006 -0500, Kevin Dangoor wrote:
>>>The pyrex list has people posting their unofficial versions of pyrex
>>>for 2.5. I haven't seen anyone speak up with a response about an
>>>"official" release for 2.5. Can we use one of these alternative pyrex
>>>versions to make ruledispatch work with python 2.5? Are there any
>>>tradeoffs to consider (or any other incompatibilities beyond pyrex)?
>>There are a couple of grammar changes in 2.5 that affect the
>>ast_builder module; I don't have a patch at this point and don't
>>know when I will get to it.  Tested patches welcome.  :)
>Would it be usable without those changes? If it can be used just
>without 2.5 language features, that would certainly be an improvement
>over not being usable at all.

I've just checked in the following:

"""Crude hacks to make tests pass on Python 2.5, at least when 
--without-speedups (because Pyrex-generated code doesn't work right
for Python 2.5).  This code will probably also behave badly if you
use conditional expressions in a when() string.  But, if you avoid
conditional expressions and either leave out the speedups or compile
them with a Pyrex hacked for Python 2.5, you'll probably be okay for

Let me know if you have any problems.

More information about the PEAK mailing list