Gentoo Archives: gentoo-python

From: Sjujsckij Nikolaj <sterkrig@×××××××.com>
To: gentoo-python@l.g.o
Subject: Re: [gentoo-python] Bug 375257: python and >=linux-3 don't play well
Date: Wed, 21 Sep 2011 19:37:57
Message-Id: op.v16biccx5oldpl@gentoobook.trollsnaetverk
In Reply to: Re: [gentoo-python] Bug 375257: python and >=linux-3 don't play well by Matthew Summers
Den 2011-09-16 17:51:09 skrev Matthew Summers <quantumsummers@g.o>:

> On Fri, Sep 16, 2011 at 4:14 AM, Dirkjan Ochtman <djc@g.o> wrote: >> I think it would be good to take the upstream patch and apply it to >> our packages. We might even do this for all the old versions. However, >> in the simple case that would mean revbumping 6 slots and requesting >> restabilization, for something excessively tiny (which doesn't cause >> issues for any Python built while running a pre-3.0 kernel). Is there >> a way we can slip this in without revbumping for older versions, >> keeping the stable keywords? Should we consult gentoo-dev about it? >> >> Cheers, >> >> Dirkjan >> >> > > Glad this is fixed upstream and in Gentoo! As far as slipping this > into older, stable packages without a revbump, I strongly urge against > this course of action. Worst case and since its perceived as a pain, > just don't bother with stabilizing the older versions until there is > another release, if ever. > > If you have questions about policy, this seems to be QA related, so > ask one of the team.
But when Linux 3.x goes stable, stable Python versions suddenly became "broken". I don't think it's reasonable to have Python 2.4-2.6 in Portage tree keyworded stable (and therefore considered to be supported) and know for sure that it's just a matter of time when somebody have to recompile it and boom! half of Python modules stop working. I daresay these cases should be handled too.