Sunday, 26 January 2014

Re: Ran autopkgtests with python3.4 as default

Am 26.01.2014 11:30, schrieb Martin Pitt:
> Matthias Klose [2014-01-26 9:53 +0100]:
>> Unfortunately this includes some tests which did fail before too
>> (python-misaka). Marking these as python3.4 is "interesting".
>
> As I wrote, I didn't file bugs about "broken tests" for packages which
> already fail with 3.3. For misaka I filed LP#1272372 that merely says
> "needs to build a python 3.4 extension". The test shows clearly that
> it doesn't currently have one.

https://jenkins.qa.ubuntu.com/view/Trusty/view/AutoPkgTest/job/trusty-adt-python-misaka/

clearly shows that the test never did succeed. Yes, it did need a fixed build
dependency, but it still fails the test as before, even after adding the module
as a dependency for the autopkg test.

>> In general I don't like the way how the introduction of autopkg tests forces
>> work on maintainers to do for the migration, when you don't have the time to
>> spend resources or want to use this time to improve packages in other areas.
>> Developer time is limited, and I'm tired being forced to spend time on broken
>> autopkg tests and to spend time to hunt down people to ignore failing or
>> stalling tests on the autopkg test infrastructure.
>
> Yes, that's a fair point. I believe we should only consider a failed
> autopkgtest in britney if it ever succeeded, so that this doesn't
> happen.

that would help for the failing tests, but not for the timeouts(?) or still
"RUNNING" tests in the infrastructure.

Matthias


--
ubuntu-devel mailing list
[email protected]
Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel