On 2014-09-05 14:09:36 +0200 (+0200), Alan Pevec wrote:

I assume adding py34 env on stable branches was a mistake or is there more to this than what's in the commit message?

https://review.openstack.org/#/q/branch:stable/havana+topic:py34,n,z https://review.openstack.org/#/q/branch:stable/icehouse+topic:py34,n,z

These are mainly belt-and-suspenders in case any projects are running unit tests for Python 3.3 on their stable release branches, since we will soon (within a week or two hopefully) be dropping the custom platform we've been using for that and switching all those jobs to run with Python 3.4 instead. I was already mostly machine-generating the patches to projects to increase 3.4 visibility to developers anywhere 3.3 was being used, but you can feel free to ignore the proposed stable branch changes if you like. Just be aware that any stable branches of projects previously tested on 3.3 (if there are any) will either need to work on 3.4 or have their Py3K tests dropped entirely since we won't be supporting any test infrastructure which can run 3.3-based jobs. Luckily, so far most of the projects which work on 3.3 and have voting jobs for it are clients and libraries which don't participate in integrated releases or have stable branches, so this may be a complete non-issue. The proposed patches are there in case you want them, but use your best judgement and I encourage you to double-check whether you have gate-{name}-python33 jobs currently voting on any potentially affected projects. Since we've not, as a project, declared victory on Py3K compliance, it is of course entirely legitimate to just blacklist those jobs from any stable branch commits through Icehouse so that we'll never run them there (and in a month or so we can discuss whether there's any point in running them on Juno either). -- Jeremy Stanley