On 04/01/14 20:01, Javier Lopez wrote:
Hi,
Happy new year to all QA ubuntu members =). Since some time ago, launchpad has
been used to track the progress on manual|automatic tests for Ubuntu and derived
flavours (kubuntu, lubuntu, xubuntu, etc).
https://bugs.launchpad.net/ubuntu-manual-tests
https://launchpad.net/ubuntu-autopilot-tests/
Recently however, Paul White has pointed out that some of the flavours use
other mediums to track progress on this aspect, eg. Kubuntu uses
https://trello.com/kubuntu.
https://bugs.launchpad.net/ubuntu-manual-tests/+bug/1183934
Should keep using launchpad to track manual|automatic test for ubuntu derived
flavours?
Greetings
Happy new year to you too :)
My position is simple enough.
*Manual Testing.*
If a flavour team wants to use the trackers then they should.
If I'd seen this discussion before the bugs for manual testcases were
created, I would at that point have been -1 to a mass creation of
testcases for flavours without any discussion with them.
It should be up to a team how to deal with their QA infrastructure.
Different teams have differing resources - I'm positive that the only
way that we (Xubuntu) managed to get the testcases we did written, onto
the package tracker and actually tested was down to a lot of work by a
very few people at the beginning.
It was of some concern to find ~50 bugs all tagged Xubuntu - I didn't
see any discussion on any mailing list about this.
Now that we have the testcases - the tracker itself mostly looks after
itself. Though actually getting tests done and reported is a different
story.
As far as Kubuntu goes - if they aren't using, and seem to have no
inclination to do so - why shouldn't the Kubuntu testcase bugs be marked
as invalid - as they surely are that ;) [1]
I see that ubuntu-studio has 53 outstanding testcase bugs - they have
less people than we do.
Edubuntu has 65 of them - though I've no idea how many people there are
available to do their testcases.
*Autopilot testing*
I can see the value in there being testcase bugs raised here - once the
testcase is done then the need for people to get involved reduces
remarkably - I assume.
I'm surprised to be honest that there isn't a massive list of these.
Perhaps because it would seem that autopilot was set up purely for
Ubuntu - which is understandable - we are having massive problems even
getting anything to run for us in our environment, it would certainly
have been of enormous help for us to have automatic testing - it is
looking increasingly likely that we'll not ever be in a position to
really take advantage of this.
*Summary*
In summary - it seems to me that we're looking at discussing something
now that should have been discussed earlier.
But the main thing we should look at is whether a flavour has any
intention of using the package tracker. If QA as a team hasn't touched
base with them - then perhaps we should be.
Apologies if I did miss mailings.
regards
Elfy
[1] https://lists.ubuntu.com/archives/kubuntu-devel/2014-January/007653.html
--
Ubuntu Forum Council Member
Xubuntu QA Lead
--
Ubuntu-quality mailing list
[email protected]
Modify settings or unsubscribe at:
https://lists.ubuntu.com/mailman/listinfo/ubuntu-quality
