Re: [Nautilus-list] Handling Nautilus in bugzilla

[Luis Villa <louie ximian com>]

On Tue, 2002-03-05 at 10:13, Alex Larsson wrote:
> > First, I would say that whatever makes bugzilla work for you in
> > getting Nautilus under control is what we should do. However, we
> > already have two overlapping designations to distinguish GNOME 2
> > Nautilus bugs that we've already tried to implement (obviously with
> > only limited success). First, Darin has long used the milestone label
> > "1.1.x" to distinguish bugs that needed to be fixed for GNOME
> > 2. Second, Louie added the "GNOME2" keyword. Those are the two
> > designators that I've been trying to use as I categorize incoming
> > Nautilus bugs.[1]

I'd like to have a 1.2 and 1.2.x milestone as well (as per some email
with Darin that slipped through the cracks this weekend and did not get
followed up on.) 

> > Is there some reason we shouldn't use those two designations to
> > distinguish GNOME 2 Nautilus bugs?
> 
> Hmmm. As long as someone keeps setting the GNOME2 keyword on new bugs I
> guess I can work only with the GNOME2 ones. I prefer that to the 
> milestone, because then we can have several milestones for gnome 2.

I'm trying to keep up and set the keyword as much as possible; I'm far
from perfect on that ATM, but definitely working on it as fast as I can,
and I think (generally) doing a fairly good job. It needs to be added to
all the 1.1.x bugs.

Would this be sufficient? I'm not really opposed to splitting things up
as you originally proposed, but I think adding yet a third way of
identifying/separating out the bugs (beyond keywords and milestones)
might be getting overly confusing :)

> Is there any chance of adding an X-Bugzilla-Keywords: email header in 
> addition to the ones we have now so I can filter natuilus-maint into two 
> folders?

I've written a patch for this against 2.16, but it is (basically) broken
wrt 2.10 mailhandling. If it'll help, I'll definitely try to backport it
sometime this week.

Related to your original thought, but not directly, Alex:

The main thrust of the original email [ignore the old bugs for now, go
through them later] brings up a question that I think John raised
before: why don't we just close all the 'old' milestone bugs that have
not been touched in > 4 months? I'd like to close them with something
like this:

'These bugs are against extremely old versions of nautilus, or were
incorrectly marked as such at one point. We ask that you help the
development of nautilus 1.2 by reopening these bugs if they are still
valid; if you cannot reopen the bug for some reason, but still feel it
is a problem, /please/ contact louie ximian com and tell him the bug
number.'

99% are never going to be responded to, but the ones that are reopened
or responded too are going to be a very high hit rate for 1.2.0/1.2.x
stuff. As it stands, if the original reporters still care, we get
feedback (they likely do still care for the important bugs) and if they
don't, we've just been spared reading 1,000 bugs, many of which are
basically 'there is a FIXME in the code here'. [This uber-specific
approach to bug filing made sense when nautilus had a QA staff and lots
of devels, but not so much anymore, IMHO.]

Just a thought-
Luis