[Libreoffice-qa] Bug Status NEW but Assigned to Someone - Please Change

bfo bfo.bugmail at spamgourmet.com
Wed Jan 16 08:07:02 PST 2013

Pedro wrote
> There is no point in triaging again. The QA work is done. Now only Devs
> can fix or don't fix confirmed bugs.
> Assigning to someone is not a good idea. It seems like that person is
> going to act on the problem when in fact he/she didn't select the issue.

David Tardon wrote
> I think explicitly setting a bug to ASSIGNED is pretty much pointless.
> You cannot force a developer to start working on a bug just because he
> has been assigned to it; on the other side, seeing that a bug is already
> assigned might discourage others from taking it.

Seems my post was misunderstood. My intention wasn't assigning bugs to
anyone by QA people. 
Simply developers should change status to ASSIGNED when they are actively
working on fix, leaving Assigned to as default when bug is in NEW state. Not
everyone do it, some bugs have sudden NEW>FIXED transition. 
Also bugs are fixed without any report in the Bugzilla. You have to watch
commit messages and changelogs to get the idea. This complicates work while

Pedro wrote
> You can not force people (especially volunteers) to fix anything they
> don't feel like.
> That is the nature of this project. It is frustrating but this is how it
> works.
> Unless you know how to code and fix the problem yourself, this is as far
> as you can go.

Wrote it few times already, will write it once more - you can manage bugs
even in the open source project. This will be more wishful thinking (more
like Please fix this in next maintenance release), but it is possible. 
Having release fever (core developers as firemen) is not good policy for bug
fixing.  I believe people would consider to fix things when there is clear
indication that a fix is needed and expected sooner than later. Just put
yourself in the shoes of new developer who browse 3700 NEW bugs. Impossible
to pick anything.

David Tardon wrote
> There are only so many developers. And to fix a bug typically takes more
> time than to triage it.

I have a few examples already that well triaged bug is fixable within hours
when it got dev attention somehow. Not general rule of course.

David Tardon wrote
> Similarly, setting a
> confirmed bug to WONTFIX just before nobody has fixed it in a certain
> period of time (if we really do not want to/cannot fix it, it should
> have already been marked appropriately during triage) brings us no
> positive effect (except having smaller number of opened bugs--maybe).
> One negative effect that immediately comes to mind is that all the
> triaging work will need to be repeated from scratch the next time
> someone reports the same bug. 

In real life you have to triage every single bug anyway. Check for
duplicates first. If you have WONTIXed bugs already (good reasoned in
comments) you can triage faster by simply marking another bug as a
duplicate. I am not an expert to mark any bugs as WONTFIX. This should be
done by core developers. They would need to discuss new NEW bugs regularly.

David Tardon wrote
> Therefore, the only possible reason for a
> triaging marathon like you suggested is to identify (and dispose of)
> bugs already fixed in the newest release (or, preferably, master).

Exactly this should be done some day. Unfortunately QA guys are too busy
with Unconfirmed backlog.
Best regards.

View this message in context: http://nabble.documentfoundation.org/Re-Bug-Status-NEW-but-Assigned-to-Someone-Please-Change-tp4029440p4029666.html
Sent from the QA mailing list archive at Nabble.com.

More information about the Libreoffice-qa mailing list