britney

Merge lp://staging/~paelzer/britney/hints-ubuntu-groovy-mmdebstrap into lp://staging/~ubuntu-release/britney/hints-ubuntu

hints-ubuntu-groovy-mmdebstrap
Merge into hints-ubuntu

Proposed by Christian Ehrhardt  on 2020-08-20

Status:	Merged
Merged at revision:	4953
Proposed branch:	lp://staging/~paelzer/britney/hints-ubuntu-groovy-mmdebstrap
Merge into:	lp://staging/~ubuntu-release/britney/hints-ubuntu
Diff against target:	11 lines (+4/-0) 1 file modified ubuntu-release (+4/-0)
To merge this branch:	bzr merge lp://staging/~paelzer/britney/hints-ubuntu-groovy-mmdebstrap
Related bugs:	Link a bug report

Reviewer	Review Type	Date Requested	Status
Iain Lane (community)		2020-08-20	Approve on 2020-08-20
Review via email: mp+389586@code.staging.launchpad.net

Commit message

single positive outlier breaks migraton
https://autopkgtest.ubuntu.com/packages/m/mmdebstrap/groovy/armhf
This was due to skipping all tests

All other arches never worked at all

Currently blocks mmdebstrap bdebstrap and qemu in groovy.

Revision history for this message

Iain Lane (laney) wrote on 2020-08-20:

thanks

review: Approve

Revision history for this message

Steve Langasek (vorlon) wrote on 2020-08-20:

fwiw you say the other archs "never worked at all", but I see that they went from "neutral" results to "fail", which last I knew wasn't supposed to be allowed for autopkgtests; so I'm unclear how this ever was allowed to migrate, or why other archs weren't a blocker for qemu as well...

Revision history for this message

Iain Lane (laney) wrote on 2020-08-20:

Nah, it was a bug that neutral was treated as pass before, because we didn't have the proper set of states available in britney. Fixing this was actually one of the major reasons I had for wanting to update to latest upstream. neutral results are supposed to express nothing about whether the package is good or bad, and so they are "invisible" wrt. migration.

That's why this became a regression on armhf only; the copied-up result was a pass, and the fails made it regression there. For all the other arches, there have only ever been fail or neutral, which is "always failed" as far as migration is concerned.