Patches!!! [was Re: GNOME 2.7/2.8 Proposed Schedule]

[Luis Villa <louie ximian com>]

On Sat, 2004-04-24 at 10:50 -0400, Luis Villa wrote:
> I agree that generally bugzilla cleaning is an ongoing process, and I'm
> not sure that enshrining it in the schedule makes sense- especially
> since new bugs come in all the time; people should (as Jeff says) always
> be doing these things.
> 
> Patches are maybe a different problem, though- there are so many, and
> not reading them is /so/ detrimental to what we are trying to do, that
> maybe a week or two near the beginning of the schedule where bugsquad
> devotes some days, and where maintainers get really actively nagged
> about them, wouldn't be a bad thing. I worry, though, that saying 'this
> is patch review time' means they'll be neglected the rest of the cycle-
> maybe making it counterproductive.
> 
> At any rate, I'd say this can't be scheduled right now, though- someone
> first needs to write an SQL query/report cgi that provides a good query
> for maintainers for patches. Hopefully I will find time for that soon
> but I'm very busy ATM so I'm hoping someone else will find the time :/
> If anyone wants to help with that, the rough details of what is needed
> in the query are here:
> http://bugzilla.gnome.org/show_bug.cgi?id=138144#c6

So... I've whipped up a better query: 
http://bugzilla.gnome.org/patch-status.cgi

This page should give you all patches in a product. To get a bug off
this query, mark the bug obsolete, needs-work, or 'commit' in the patch
editor (e.g.,
http://bugzilla.gnome.org/attachment.cgi?id=10601&action=edit )

Notes:

* Gman already rocks for having re-reviewed and marked all of his
outstanding patches.

* PATCH keyword is going away as soon as I can audit that PATCH ~=
having at least one attachment marked as a patch on the bug.

* It is likely that this query is buggy- it is my first experiment with
joins in SQL and works as much by chance as anything else. If you find
bugs with it, file them against bugzilla.gnome.org.

* I realize there are a lot of old patches here that are here because
there didn't used to be a way to mark patches needs-work/obsolete. I'll
be working with the bugsquad this week to kill as many of those as
possible.

* There is now a 'committed' status for patches that have been committed
but for which the bug cannot/should not be closed for some reason.

I hope this query is useful- let me know if not.

Luis