[Zope3-dev] Re: Python version for Zope 3.4 ?

[Gary Poster]

On Sep 29, 2006, at 10:03 AM, Jim Fulton wrote:

> Tres Seaver wrote:
> ...
>> The change from 2.4 to 2.5 is *massively* disruptive for a framework
>> like Zope:  much more so than any change since 2.2, I think (maybe  
>> even
>> 2.0/2.1).  The hardest bit is the change to the way the compiler  
>> works:
>>  RestrictedPython is completely incompatible with the new AST-based
>> strategy.
>
> Yup.  Also, everything I've tried (I haven't tried much) has had test
> breakage.  Most of this is probably shallow, but it will take some  
> time to
> sort it all out.
>
> As we move toward splitting things up into eggs, people can help  
> out by
> getting things working on an egg-by-egg basis.

I think buildbot makes supporting Zope 3 on py2.4 and py2.5  
reasonable, and a first step to making Zope 2 support py2.5.  I like  
the idea of dividing things up also, to help the constraints (for  
instance, I don't know if anyone is going to have the time and energy  
to move the zodbcode module forward to py2.5).

I glanced through some of the problems today.  Most of them didn't  
seem too bad, and I have RestrictedPython foo (and interest/need) to  
help there on the Zope 3 side.  However,

../bin/python2.5 test.py -m persistent and ../bin/python2.5 test.py - 
m ZODB

both result in numerous errors culminating in segfaults, at least on  
my AMD A64.  (BTree has a bunch of errors too, but no segfault).   
These might be relatively shallow too, but superficially they seem  
scary.

Gary