canonical-ci-engineering team mailing list archive

[Martin Pitt <martin.pitt@xxxxxxxxxx>]

Hello Francis,

Francis Ginther [2013-10-29  9:48 -0500]:
> The piece you are most likely missing is the script inside the jenkins
> job itself, [4].

Indeed, that was it. Fortunately it turned out that it wasn't a
problem with the infrastructure at all, but that something in the
unity-daily PPA breaks messaging-app and the test discovers that.
That's why I had so much trouble reproducing it, until I saw that
Jenkins job I didn't think of adding the PPA and just tested with
saucy and trusty.

> Now the proper solution is that we should not have any logic in the
> jenkins jobs itself. It should all be included in the test runner
> branch or we should be using a standard test runner approach. Both
> would allow us to write simpler documentation to enable developers to
> run this at home. These are things that we need to clean-up this
> cycle. These test runners are not private, so there is no reason not
> do document them on the public wiki.

Sounds great!

Thank you,

Martin
-- 
Martin Pitt                        | http://www.piware.de
Ubuntu Developer (www.ubuntu.com)  | Debian Developer  (www.debian.org)