On Nov 10, 2015 1:16 PM, "Adam Conrad" <firstname.lastname@example.org> wrote:
> On Tue, Nov 10, 2015 at 09:31:54AM -0800, Steve Langasek wrote:
> > That works for a silo, because with a silo you want to install *all* of the
> > packages from the ppa together, and pull any additional dependencies from
> > the main archive. For -proposed, we explicitly want to pick and choose
> > *which* packages we are pulling from -proposed vs. the release pocket,
> > because -proposed always contains multiple unrelated "landings" at the same
> > time and we want to be able to disambiguate the test results so we know
> > which package introduces the regression.
> Robert's still on the right track here, though. This is just poor use of
> pinning. See the following:
+1, it's true that autopkgtest has no way to know that groups of packages came through together from a single silo, but it's not true that its impossible for apt pinning to install "just one package and its dependencies" from proposed, as Adam demonstrates.