[haiku] Re: State of Haiku QA
- From: Niels Reedijk <niels.reedijk@xxxxxxxxx>
- To: haiku@xxxxxxxxxxxxx
- Date: Mon, 4 Jan 2010 19:55:08 +0100
Hi,
2010/1/4 PulkoMandy <pulkomandy@xxxxxxxxx>:
>> Again - how exactly will a better QA process bring immediate benefit to
>> the
>> project, if we fail to draw the big potential benefit out of the process
>> we
>> alreay have? By big potential benefit I mean that Haiku would already be
>> so
>> much more polished if a big chunk of all existing tickets were fixed right
>> now. The main problems are lack of time and/or motivation. How will a
>> better QA process help with that? More likely (IMHO), it has the potential
>> to cause even more frustration when tickets just linger if the one who
>> prepared it put even more effort into it.
>
> Better QA is not necessarily about finding even more bugs. It can be
> checking that old bugs are still reproductible in recent builds,
> investigating way to trigger the bug, and perhaps even searching for the
> source of the problem, linking tickets that seem related, and so on. About
> anything we already more or less do on trac besides writiing code, actually.
> I think the QA team could get permissions to close/assign/... tickets as
> developpers do, but no access to the svn, for example. On the other side,
> the QA is done on code cleanliness on the mailing list after someone commits
> something. It's about checking if the code runs and does not break the
> build, but also checking it fits the coding guidelines.
>
> Most of these things the developpers are already doing in more or less
> formalized and automated ways. But a formal team could help even if they
> don't know much about the code.
Well, instead of doing all-or-nothing, I was planning on organizing a
'Bug Day' soon. We can try such a day to see how that works.
In short, there are currently 910 open tickets in our bug database
that have the version 'R1/pre-alpha1.' There are a total of 1431 open
tickets at the moment. I would like to try to make a community effort
to go through all the pre alpha 1 tickets and test whether they still
exist. The goal is to have the tickets closed (because they are fixed)
or reassigned to R1/alpha1 to point out that they still exist in alpha
1 (and beyond).
My goal would be tap into the lurkers on this list and try to engage
them into spending some time on the project.
I wanted to ask some developers to be available to guide the hunters
on IRC, as well as manage all the contributions (like setting
versions, as normal users cannot do this).
I think it is worth it to work out the idea and to give it a try to
see whether this helps keeping the ticket database on dev.haiku-os.org
clean, and whether it increases the number of good bug reports.
Perhaps this can evolve to a more structured testing of bug reports.
Kind regards,
N>
Other related posts:
