Re: Fwd: 30+ patches in BugZilla.

["Kevin Kubasik" <kevin kubasik net>]

On 5/17/06, Joe Shaw <joeshaw novell com> wrote:
> Hi,
>
> On Wed, 2006-05-17 at 08:30 -0400, Kevin Kubasik wrote:
> > Hey, I dunno if everyone here watches the f-spot devel lists, but I
> > saw this, and found it to hit a little close to home. Beagle is in
> > much the same position with 31 unreviewed patches in Bugzilla, while
> > more of those patches have some sort of comment, several have
> > bitrotten past the point of easy revival, despite their usefulness. A
> > few (such as child indexables in the file system and an archive
> > filter) have been outstanding with some time now, and have no clear
> > definition of what still needs to be done. I know I have submitted
> > several patches over the past week or so and have not heard much about
> > them. Overall, the message bellow just seemed relevant.
>
> As I mentioned before, I am just overwhelmed with Novell stuff at the
> moment to do an adequate job of reviewing non-trivial patches.
> Unfortunately Bera is gone for the summer, Daniel and Fredrik are pretty
> busy with school and summer work, and Jon has moved on, so right now I'm
> the only one I feel comfortable about individually reviewing patches.
> I'm sorry about this.  Fortunately, this won't to be an issue forever,
> just while we push through our release, but I do suspect it will be the
> rest of the month of May.

Understandably so, I didn't mean to come off as a nag.
> What would really help me out a lot is if the adventurous among you who
> build from CVS tried out some of these patches.  In particular, Kevin's
> compressed text cache (#339470), using a regexp for parsing query
> strings from Max (#339844), and the thunderbird backend (#323065) from
> Pierre would all be great to test.  If a lot of people are really using
> them without major problems, that makes me feel a lot better about
> committing them without going over them in great detail.

Agreed, I have been running all three without issue, and keep a
'bleeding edge' SVN with a couple pending patches running (which I
know a couple people who really want the thunderbird backend have been
using without major issue) I'll post some info about the repo when I
get home and tidy it up a bit if anyone else wants to give it a try.

> A list of outstanding patches can always be found in bugzilla here:
>
>         http://bugzilla.gnome.org/reports/patch-report.cgi?product=beagle&patch-status=none
>
> Maybe we should put that on the wiki somewhere?
I added a little note on the development page.
> In particular from the forwarded message:
> > I see some comments that the patch is not solving it to 100% or is just a
> > workaround for a fault in another product. But the fact that there is a
> > bug in BugZilla indicates that there is a problem, and a possible patch
> > fixes it (hopefully). If it is not perfect, lets include it to HEAD, and
> > fix it over time (new bug reports).
>
> While I understand the idea that not including (or being late reviewing)
> a patch can be somewhat discouraging for potential developers, there is
> also the possibility that a bunch of buggy and untested code is dropped
> into HEAD and then someone (me) has to deal with a torrent of bug
> reports and fix code that is unfamiliar.  If these are showstopper bugs,
> then I can't do releases until these are fixed.  This is less of an
> issue for people who have stuck around the community like you, Kevin,
> and who can be trusted to fix these bugs, but this is can be a major
> problem for very helpful and well-intentioned individuals who do one-off
> patches to fix problems but who aren't interested in becoming regular
> members of the community.

Understood and I agree completely. The forwarded e-mail didn't present
a perfect solution or argument, but It just seemed to strike a little
close to home. Thanks for the response.

--
Cheers,
Kevin Kubasik
http://kubasik.net/blog