AutoQA: distro congestion?

[Michael Schwendt]

On Tue, 19 Apr 2011 10:40:22 -0700, AW wrote:

> On Tue, 2011-04-19 at 18:30 +0200, Thomas Spura wrote:
> > On Tue, 19 Apr 2011 08:37:25 -0700
> > Adam Williamson wrote:
> > 
> > > On Tue, 2011-04-19 at 11:05 +0200, Michael Schwendt wrote:
> > > 
> > > > Some packagers have been observed circumventing the system by
> > > > configuring a karma threshold of 1, so their own +1 vote or the
> > > > first one from an arbitrary tester make it possible to mark the
> > > > update stable.
> > > 
> > > Not...really. The update submitter's own vote should count as 0 (I
> > > can't remember if this has landed in current Bodhi yet). Setting the
> > > karma threshold to 1 cannot circumvent the 'proventesters +1 and any
> > > +1' requirement for stable releases; just try it, it doesn't work. The
> > > correct requirements are enforced whatever you set the autopush
> > > threshold to.
> > 
> > I just tried it (ok after the threshold of 3 days in f15) and it worked:
> > https://admin.fedoraproject.org/updates/ipython-0.10.2-1.fc15
> 
> F15 is a pre-release and does not have that requirement. Its requirement
> is +1 from anyone - so you're not circumventing anything. :)

We've always had the rule of thumb that packagers should not vote for
their own updates. It is assumed that the packagers test their own updates
and don't need to be explicit about their confidence in the update with
karma points in bodhi.

One goal of the update acceptance criteria for pre-releases is also that
_anyone_ gets an opportunity to try out a test-update before it is marked
stable.

Anyway, I think I've seen counted "self-votes" also for older dists, but not
done by many packagers.