Hello, 

I've been following OpenERP for 2 years now, with several successful 
implementations. My take is that the OpenERP team resources are limited and 
IMHO there are lots of other priorities than switching to Python 3k for now.

A few things I put before moving to Python 3k:
- avoiding regressions with the new buildbot server and a larger testsuite, 
also more rigorous process at Tiny and when interacting with the community 
(open planning and strategy for API changes, bug prioritization, clean 
back-ports...)
- refactoring some code, especially large methods, old style wizards and 
methods that create/alter records without returning the proper ids to 
overriders.
- polishing some feature to get basic things done right out of the box
- finishing the accounting for the main targeted countries (closing periods, 
financial reports...)
- make admin customization a first class citizen accross migrations
- run on Python 2.6 out of the box (including on Ubuntu, some cheating is still 
required I think).
- make sure all the bugfix from 5.0.x is ported to trunk properly in order not 
to frustrate the community with double bug fixing in the next versions


Considering the current dev power at Tiny and community, I think this postpone 
migration effort to Python 3k to mid 2010 at best. I see no big deal here as 
lots of libraries won't move faster. I've also been reading that maintaining 
several Python target distro seems a headache. Indeed, for such a large project 
such as OpenERP, I think automatic migration tools (like 2to3) won't make it.

So it means Tiny will have to:
a) either maintain 2 branches, which I believe is not reasonnable considering 
how hard it is currently to maintain the 4.2.x + 5.0.x + trunk branches already.
b) switch from 2.5/2.6 to Python 3k without further support for older Python 
versions. I think this is unfortunately the best option, But stopping 2.5/2.6 
support will also frustrate a large part of the community. Currently this is 
not doable but we might have a better time window for that switch next year.

I should say I'm more eager to have OpenERP running on Jython and Java servers 
(see my blog for details). Jython won't move to Python 3.x support before next 
year, but quicker Jythons and compatible OpenERP code (no more obsolete C libs) 
are to expect by the end of 2009.

Finally @oradoe, if you change your pool to include a deadline, then I might 
vote yes or no; currently I can't.


Best regards,

Raphaël Valyi




-------------------- m2f --------------------

--
http://www.openobject.com/forum/viewtopic.php?p=40402#40402

-------------------- m2f --------------------


_______________________________________________
Tinyerp-users mailing list
http://tiny.be/mailman2/listinfo/tinyerp-users

Reply via email to