Open Stack

-----Original Message-----
From: Jeremy Stanley <fungi at yuggoth.org>
Reply: OpenStack Development Mailing List (not for usage questions) <openstack-dev at lists.openstack.org>
Date: March 24, 2016 at 16:15:38
To: OpenStack Development Mailing List (not for usage questions) <openstack-dev at lists.openstack.org>
Subject:  Re: [openstack-dev] [release] [pbr] semver on master branches after RC WAS Re: How do I calculate the semantic version prior to a release?

> On 2016-03-25 08:28:16 +1300 (+1300), Robert Collins wrote:
> > However, Doug and I just uncovered a concern on IRC: having PyPI be
> > ahead of master for more than very short periods seems like a poor
> > idea.
> >
> > When we tag a final release on a branch, without master being higher
> > than that final release, folk running things installed from master,
> > will now be 'upgradable' to whats on PyPI, even though thats older
> > from a human perspective.
> >
> > So, I think we do need some aggressive means to bump master up past
> > the versions reserved for a stable branch, at the stable branches
> > creation.
>  
> I could be misunderstanding, but isn't this why we have a release
> pipeline job which merges the release tag into master? So that ~as
> soon as there is a new release, post-versioning in the master branch
> is parsed by PBR as being at or later than the most recent release?

Also this only affects libraries and not the server projects. Besides, hasn't this community always asserted that no one installs from PyPI anyway (inspite of evidence of projects allowing just that based on user feedback)?

--  
Ian Cordasco