ubuntu-bugcontrol team mailing list archive

[Steve Beattie <mp+427845@xxxxxxxxxxxxxxxxxx>]

On Thu, Aug 04, 2022 at 09:32:24PM -0000, Alex Murray wrote:
> > +def find_milestone(ubuntu, release):
> 
> Perhaps a better name would be pending_milestone()?

Oh yeah, I like that. Will convert.

> > +        if milestone is None:
> >              unembargo.append(seen[series_name][0])
> > +        else:
> > +            print("WARNING: %s is approaching milestone %s (due %s)" %
> > +                  (series_name, milestone.name,
> > +                  time.strftime("%Y-%m-%d", milestone.date_targeted.timetuple())))
> > +            print("NOTE: Please coordinate with the #ubuntu-release team before releasing.")
> > +            if opt.force:
> > +                print("NOTE: unembargo for %s will continue due to use of --force." % (series_name))
> > +                unembargo.append(seen[series_name][0])
> > +            else:
> > +                print("NOTE: To override this check and publish anyway please use the --force.")
> 
> Since Marc feels strongly about it, we may as well change this to add a sys.exit(1) here then as well as part of this MR.

I'm not sure that satisfies the objection completely, in the case like
the publication that triggered the bug.

For example, imagine we didn't publish openjdk-8 updates for jammy.
Then, if the sequencing was right, the script could go:

  inspect bionic and focal for openjdk-8:
  - no milestones found, continues on and publishes

  inspect bionic, focal, and jammy for openjdk-lts/11
  - milestone found, with the exit located above, abort

To do it properly, a little more refactoring needs to occur to find all
the releases being published for all the source packages, perform the
milestone check, and then publish.

-- 
Steve Beattie
<sbeattie@xxxxxxxxxx>

https://code.launchpad.net/~sbeattie/ubuntu-qa-tools/+git/ubuntu-qa-tools-1/+merge/427845
Your team Ubuntu Bug Control is subscribed to branch ubuntu-qa-tools:master.