sometimes triggers tests for one architecture only

[Martin Pitt]

Seen on excuses right now:

python-passlib (- to 1.6.5-1)

    Maintainer: Debian Python Modules Team
    0 days old
    autopkgtest for keystone 2:8.0.0~b2-0ubuntu2: armhf: Test in progress
    Not considered

[Martin Pitt (pitti)]

Corresponding log: http://people.canonical.com/~ubuntu-archive/proposed-migration/log/wily/2015-09-11/14:21:40.log

Relevant:

I: [Fri Sep 11 14:30:33 2015] - Reverse dependency keystone of python-passlib/1.6.5-1 is unbuilt or uninstallable, running test ag
ainst testing version 2:8.0.0~b2-0ubuntu2
I: [Fri Sep 11 14:30:33 2015] - Reverse dependency keystone of python-passlib/1.6.5-1 is unbuilt or uninstallable, running test against testing version 2:8.0.0~b2-0ubuntu2
I: [Fri Sep 11 14:30:33 2015] - Reverse dependency keystone of python-passlib/1.6.5-1 is unbuilt or uninstallable, running test against testing version 2:8.0.0~b2-0ubuntu2
I: [Fri Sep 11 14:30:33 2015] - Reverse dependency keystone of python-passlib/1.6.5-1 is unbuilt or uninstallable, running test against testing version 2:8.0.0~b2-0ubuntu2
I: [Fri Sep 11 14:30:33 2015] - Reverse dependency keystone of python-passlib/1.6.5-1 is unbuilt or uninstallable, running test against testing version 2:8.0.0~b2-0ubuntu2
I: [Fri Sep 11 14:30:33 2015] - Reverse dependency keystone of python-passlib/1.6.5-1 is unbuilt or uninstallable, running test against testing version 2:8.0.0~b2-0ubuntu2
I: [Fri Sep 11 14:30:34 2015] - Requesting keystone/2:8.0.0~b2-0ubuntu2/ppc64el autopkgtest to verify python-passlib/1.6.5-1, routes/2.2-1ubuntu1
I: [Fri Sep 11 14:30:34 2015] - Requesting keystone/2:8.0.0~b2-0ubuntu2/i386 autopkgtest to verify python-passlib/1.6.5-1, routes/2.2-1ubuntu1
I: [Fri Sep 11 14:30:34 2015] - Requesting keystone/2:8.0.0~b2-0ubuntu2/amd64 autopkgtest to verify python-passlib/1.6.5-1, routes/2.2-1ubuntu1
I: [Fri Sep 11 14:30:34 2015] - Requesting keystone/2:8.0.0~b2-0ubuntu2/armhf autopkgtest to verify python-passlib/1.6.5-1, routes/2.2-1ubuntu1
I: [Fri Sep 11 14:30:44 2015] - -> matches pending request for triggers {('python-passlib', '1.6.5-1'), ('routes', '2.2-1ubuntu1')}
I: [Fri Sep 11 14:30:44 2015] - -> matches pending request for triggers {('python-passlib', '1.6.5-1'), ('routes', '2.2-1ubuntu1')}
I: [Fri Sep 11 14:30:44 2015] - -> matches pending request for triggers {('python-passlib', '1.6.5-1'), ('routes', '2.2-1ubuntu1')}

E: [Fri Sep 11 14:30:54 2015] - FIXME: Result for keystone/2:8.0.0~b2-0ubuntu2/amd64 (triggered by python-passlib/1.6.5-1) is neither known nor pending!
E: [Fri Sep 11 14:30:54 2015] - FIXME: Result for keystone/2:8.0.0~b2-0ubuntu2/i386 (triggered by python-passlib/1.6.5-1) is neither known nor pending!
E: [Fri Sep 11 14:30:54 2015] - FIXME: Result for keystone/2:8.0.0~b2-0ubuntu2/ppc64el (triggered by python-passlib/1.6.5-1) is neither known nor pending!

So apparently the routes tests were triggered slightly earlier, and the armhf result was already there. The three "matches pending request" for the other arches look wrong, as in the previous run (http://people.canonical.com/~ubuntu-archive/proposed-migration/log/wily/2015-09-11/14:11:17.log) was not present at all yet.

[Martin Pitt (pitti)]

This should be fixed with http://bazaar.launchpad.net/~ubuntu-release/britney/britney2-ubuntu/revision/547 . If not, britney will now crash which makes this kind of error much more obvious.