[Libreoffice-qa] MAB. 3.6 to 4.0 migration. When?

Petr Mladek pmladek at suse.cz
Fri Jul 26 04:47:29 PDT 2013

(I am sorry for duplicit posting. There was a problem in the mailing
list setting and my mail had been eaten.)

Tommy píše v St 24. 07. 2013 v 20:39 +0200:
> 3.6.7 is out and will be the last 3.6 release.
> I'm just curious to know when still living 3.6 bugs will be moved to the  
> 4.0 list.

IMHO, we could start at any time now. But please, double check it with

> I offer my help to revise some of those bugs and move them to the new list  
> if they are still valid in 4.0.4.

Thanks a lot for your offer. It will be a big help!

I think that we should use this opportunity also for some clean up. As
you mention, it would be great to double check if the bugs are still
valid. Also we should double check that they really deserve to be MABs,
see https://wiki.documentfoundation.org/Most_Annoying_Bugs

The problem is that we are not that strict all the time. It helped to
reduce useless fights about the priority. On the other hand, we should
not keep the wrongly sorted bugs for years. If a bug is not fixed soon
there is something smelly about it ;-)

First, let's repeat how MABs work. The severity/priority is not set
correctly for most bugs, so it is hard to see the important bugs in the
jungle. MABs help to highlight the most critical bugs and bring
developers attention. The comments from the meta bug are sent to the
developer mailing list and the bugs are advertised there. Developers
monitor it and give these bugs high priority. The nearly blockers are
fixed almost immediately. The other _very_ critical bugs are fixed
within days or weeks. Some bugs are somehow ignored, see below.

The main purpose of MAB is to get TOP priority attention. It is for
blockers, nearly blockers, something that really should get fixed very
soon. These are usually regressions that affect many people, do not have
any workaround, ...

The magic will work only if we keep the flow on reasonable level (max 10
bugs per week?). Otherwise, developers won't be able to keep up with the
flood, the list of non-fixed bug will grow and we will get another
swamp. It will lose the dynamics and all parts will get frustrated.

Now, what are the MABs that are not fixed for months and how to deal
with them? I would split them into several groups:

     + annoying problems that newer worked or did not worked more
       than two years; users are already somehow used to live with

       => they does not belong to MAB because they are not
       urgent by definition

     + annoying problems that affect only limited group of users
       and have workaround; they were usually added by the reporter;
       it was an attempt to get attention; they were ignored by
       developers because they see more important bugs around

       => they does not belong to MABs as well

     + hard to fix problems; some things are hard to debug and it
       is hard to find a volunteer to work on them; if they are
       really critical and need to be fixed soon

       => ESC should get noticed and find a volunteer ;-)

OK, it was theory. Now, how to do this practically? I would use the
following approach:

     1. check if the bug is still valid; if not
        => close it

     2. check if it was newly nominated (<2 months) and looks critical
        and urgent

        => simply move it to 4.0 MABs

     3. it does not belong to MAB as described above

        => I would remove it with some kind message about that we
        understand the annoyance but the bug does not fit the top 50
        most critical bugs

        also please, set the severity and priority more reasonable as
        suggested by
        finally, if it looks really annoying, you might consider adding
        some experts into CC, see
        So, it is left on the radar.

      4. If you are in doubts; of the bug looks really critical and
         ignored for more than 2 months; please add a comment about that
         you are not sure about the status, if it belongs to MABs, etc.,
         and add "Need_Advice" whiteboard flag, see

Please, do not get scared to remove bugs from MAB. About 3000 bugs were
fixed for LO 4.1 as mentioned at
Only small part of them was fixed via the MABs process. By other words,
if you remove a bug from MAB, it does not mean that it is dead ;-)

I am sorry for the long mail but I wanted to provide a more global view.
I hope that it will help.

Best Regards,

More information about the Libreoffice-qa mailing list