force-autopkgtest semantics

[Colin Watson]

On Wed, Jun 26, 2013 at 05:58:11PM -0600, Adam Conrad wrote:
> On Wed, Jun 26, 2013 at 07:48:10PM -0400, Scott Kitterman wrote:
> > Would it be possible to support force-autopkgtest on either the package under 
> > test or the triggering package.  Then, for today's case, we could have done:
> > 
> > force-autopkgtest kde-baseapps/4:4.10.80-0ubuntu1
> > 
> > That would then skip all autopkgtests that might be triggered by the 
> > transition.
> 
> I would find it pretty confusing if a force-foo had different semantics,
> depending on which package you applied it to.  What might be slightly
> more intuitive here would be to break it into the two actual usecases:
> 
> force-badtest pango1.0/1.2.3-1       # hints that the test is broken
> force-skiptest kde-baseapps/4.1.2-3  # hints that we don't want to wait
> 
> That gives us two clear and defined ways to deal with either "this
> package's tests are busted, but we don't want the world to hang on
> that" and "this package is urgent and we want to skip all the tests
> that were triggered by it".

Makes sense to me.  I've implemented that, and renamed the existing
hints to force-badtest.

Thanks,

-- 
Colin Watson                                       [cjwatson at ubuntu.com]