Open Stack

Le lundi 23 novembre 2015, à 14:00 +0100, Ihar Hrachyshka a écrit :
> Vincent Untz <vuntz at suse.com> wrote:
> 
> >Le lundi 23 novembre 2015, à 13:17 +0100, Ihar Hrachyshka a écrit :
> >>Vincent Untz <vuntz at suse.com> wrote:
> >>
> >>>Hi,
> >>>
> >>>I know 2014.2.4 is (in theory) the last juno release, but I'd still
> >>>expect to see a post-release bump to 2014.2.5 in git, to avoid any
> >>>confusion as to what lives in git. This is especially useful if people
> >>>build new tarballs from git.
> >>>
> >>>Any objection against this, before I send patches? :-)
> >>>
> >>>Cheers,
> >>>
> >>>Vincent
> >>
> >>I probably miss something, but why do we care about what is in the
> >>branch now that we don’t plan to merge anything more there? Is it to
> >>accommodate for downstream consumers that may want to introduce more
> >>patches on top of the upstream tag?
> >
> >As I said: because people might keep generating tarballs from git, and
> >they'd expect to have a version that is correct. Yes, this is mostly
> >downstreams. (And not necessarily to introduce more patches, but just to
> >reflect the reality).
> >
> >I would also argue that we care because we're leaving git in a state
> >that is kind of wrong (since the version is not correct).
> 
> Can you elaborate why it’s incorrect? It’s 2014.2.4 right? So that’s
> indeed what you get if you generate tarballs from latest commits in
> those branches; it should totally reflect the contents that are in
> official tarballs and hence have the same version.

Oh, indeed, you're correct.

But to illustrate better the issue I'm seeing:
http://tarballs.openstack.org/cinder/cinder-stable-juno.tar.gz contains
a directory cinder-2014.2.4.dev24, which is kind of wrong. That's the
bit that I'd like to see fixed.

Cheers,

Vincent

-- 
Les gens heureux ne sont pas pressés.