[RFC] Bugzilla Cleanup

[Matt Turner]

On Mon, Sep 19, 2011 at 3:17 AM, Jeremy Huddleston
<jeremyhu at freedesktop.org> wrote:
> So, one of the things talked about at XDC was just how icky bugzilla has become.  There are a ton of open bugs which haven't been touched in ages and either don't apply any more or don't have sufficient information to act upon.  I would like to spend some effort cleaning these up, but before I go and make thousands of changes, I wanted to discuss my plans in case anyone has objections, suggestions, and critiques.
>
> My thoughts:
>
> 1) Close all NEEDINFO and notourbug bugs which haven't been touched in over a year.
>
> 2) Clean up keywords (how do I do this?)
>   2a) Remove "cleanup0407"
>   2b) Remove movetoxkc (and move those bugs to xkeyboard-config if applicable)
>   2c) Add "HITLIST" for top issues that we should focus on addressing soon (ie: this will be a nag list).  Is there a way to restrict who can add certain keywords?
>   2d) Add "NOMINATE" to allow reporters to nominate the issue as a hitlist bug.  Maintainers can go through and review nominated bugs to see if they want to HITLIST them or not.
>   2e) Add "STALE" keyword which will be used to automatically close old, stale bugs (below)
>
> 3) Add the STALE keyword to all bugs with a note that they will be closed if nobody comments in the bug within 2 weeks if they satisfy each of the following:
>   3a) haven't been touched in >= 2 years
>   3b) Have no other keywords
>
> 4) 2 weeks later, I will act on #3 and close all bugs which received no comments.  Those which did have a comment will be reviewed and screened appropriately.
>
>
> 2d and 2e will allow us to keep track of top issues without using the tracker-* metabug.  If preferred, we can still keep the tracker-* metabugs and just add a "NOMINATE" keyword to nominate a bug for the tracker.
>
> It's unfortunate that the keywords list includes a column for the total number of bugs with such a keyword rather than the total number of *OPEN* bugs with that keyword.  Is this configurable somewhere (other than hacking the bugzilla source)?

Sounds good to me. I did some clean ups a few months ago, but barely
made a dent.

What do you suggest we do about bugs where the reporter provides a
bogus patch which fixes the problem for him, is unable to reproduce
his own report, and doesn't have relevant hardware anymore, but still
demands that the bug be kept open?

... ie, https://bugs.freedesktop.org/show_bug.cgi?id=31538

Matt