openstack team mailing list archive

[Jay Pipes <jaypipes@xxxxxxxxx>]

On Wed, 2011-11-09 at 14:57 +0100, Thierry Carrez wrote:
> Soren Hansen wrote:
> > 2011/11/9 Nachi Ueno <ueno.nachi@xxxxxxxxxxxxxxxxxxxx>:
> >> I understand your point. Stop QAing stable/diablo and focus on Essex.
> > 
> > Oh, no no. That's not the point. I'm thrilled to have you work on
> > QAing Diablo. The only issue is that the fixes you come up with should
> > be pushed to Essex first. There are two reasons for this:
> > 
> >  * If we don't push the fixes to Essex, the problems will still be
> > present in Essex and every release after that.
> > 
> >  * Having them in Essex lets us try them out, vet them and validate
> > them more thoroughly before we let them into the stable branch. When a
> > patch lands in the stable branch it has to be well tested already
> > (unless of course Essex has deviated too much, in which case we'll
> > have to accept the risk of getting it into Diablo directly).
> 
> +1
> 
> You should submit patches to master and then backport them to
> stable/diablo, rather than proposing them for stable/diablo directly.
> That ensures your work benefits both branches: making diablo better
> without making essex worse than diablo.
> 
> If that's just too much work, maybe you should raise the issue at the
> next QA meeting to try to get some outside help ?

At the QA meeting yesterday, I offered my help to Nati. I will handle
proposing his patches to Essex up to a future date where Nati and his
team will switch to code against Essex, not Diablo/stable and propose
first to master, then others will backport to diablo/stable.

Nati and I will decide on that future date for his team to switch their
focus to Essex trunk and not have to have someone manually
"forward-port" these patches to trunk.

Cheers,
-jay