Transition-Regression Tag

[Mike Rooney]

On Tue, Jun 2, 2009 at 11:32 AM, Brian Curtis <briancurtis.wx at gmail.com> wrote:
> On Tue, Jun 2, 2009 at 2:05 PM, Mike Rooney <mrooney at gmail.com> wrote:
>> On Tue, Jun 2, 2009 at 10:41 AM, Brian Murray <brian at ubuntu.com> wrote:
>>> On Mon, Jun 01, 2009 at 09:44:48PM -0400, Brian Curtis wrote:
>>>> Hi all,
>>>>
>>>> Upon discussions with mrooney about potential regressions due to
>>>> programs (like empathy and banshee) being transitioned to default.
>>>> There is a new tag, which is also listed in
>>>> https://wiki.ubuntu.com/Bugs/Tags called "transition-regression" that
>>>> is meant to mark bugs that may need to be addressed so users can make
>>>> an easy transition from one default to another.
>>>>
>>>> An example of such a bug is
>>>> https://bugs.edge.launchpad.net/ubuntu/+source/banshee/+bug/235529
>>>>
>>>> Discussion?
>>>
>>> It seems to me that this type of a bug, a functional regression caused
>>> by changing the default application from one to another, is actually a
>>> subset of the regression-potential tag.  Subsequently, I think tagging
>>> these as 'regression-potential' is the best idea but that a secondary
>>> tag, perhaps 'default-application', would be highly beneficial.  By
>>> having this secondary tag we could then separate the
>>> 'default-application' potential regressions from other potential
>>> regressions.
>>>
>>
>> This definitely makes some sense. This is already an understand
>> mechanism and any of these bugs that are released could be
>> transitioned to regression-release just like the others. I'd like to
>> talk to some people on the desktop team about integration with their
>> QA process and how this can help them as well as the upstream
>> projects, tracking issues and their importance. It seems like a formal
>> process and tracking of these issues could really help polish such
>> transitions, by fixing the bugs (ideally upstream with their help) and
>> documenting the ones we can't. Do you know who is good to talk to
>> here, perhaps rickspencer3, pitti, or seb?
>>
>> --
>> Michael Rooney
>> mrooney at ubuntu.com
>>
>> --
>> Ubuntu-bugsquad mailing list
>> Ubuntu-bugsquad at lists.ubuntu.com
>> https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugsquad
>>
>
> Hi everyone,
>
> We have added a "default-application" tag that will be a "child tag"
> to the regression-potential tag.
> https://wiki.ubuntu.com/Bugs/Tags#Regression%20specific
>
> ~Brian Curtis
>

I talked with seb in #ubuntu-bugs and he prefers that we use a
specific tag for each transition, such as rhythmbox-to-banshee or
pidgin-to-empathy. This can make it easier to search by the specific
transition, since perhaps a transition could span multiple packages.
It was also his opinion that we don't want to track these regressions
the same way we do with regression-* since they aren't regressions in
the applications and don't represent a blocker in the same way (it
seems to me like sometimes they would but I guess we handle that case
by case via the importance/nomination of the Ubuntu task).

I suppose this might mean that such a tag is generally incompatible
with regression-* (unless of course an application we are switching to
also has a regression itself that the previous default performs), and
the default-application tag isn't necessary?

Lots of agile bug work going on here; what does everyone think? I'm
glad we are converging on a formal approach to make testers' work more
trackable and actionable!

-- 
Michael Rooney
mrooney at gmail.com