Bug tracker: next steps

[Stephen J. Turnbull]

Wulf C. Krueger writes:

 > On Friday, 07. September 2007 20:19:00 Stephen J. Turnbull wrote:
 > > systems I've worked with on a reasonably frequent basis (Debian,
 > > Gentoo, 
 > 
 > You have reported exactly *one* bug on bugs.gentoo.org with your 
 > @xemacs.org address. Any other addresses you want me to check?

Um, I wrote "worked with", not "reported to".  It's not about "my"
bugs, it's about the effectiveness of the whole system.

 > I've reported more bugs by mail to this list than you ever reported on 
 > Gentoo and the biggest black hole I've seen around is *this* very list.

Would you like to reread my post, and in the process discover that
that's *my* point?  I suspect that I'm above average for Gentoo users
in terms of frequency of use of the ITS, and bug reporting in general,
if not to Gentoo.  I'm certainly way above average for XEmacs users.
If I'm discouraged from reporting by what I see when looking up Gentoo
packages' issue histories, how much more so will users be discouraged
from reporting to an XEmacs ITS that behaves like xemacs-beta?

If you would like to explain how XEmacs responsiveness is going change
by installing a bug tracker, that would be of great interest.  What I
see on Gentoo as well as the others I mentioned is that it depends on
the maintainer, not on the tracker.  No?

As for the need for an email interface, yes, I believe that's a
genuine need, and I think *you* are being uninformed to deprecate it.
We have infrastructure to support email reports that is not going to
be reproduced for web reports for some time, and it's proven to help
in resolving problems.

 > That would be a waste of time. Nobody on this list is much likely to work 
 > on years old stuff anyway if it hasn't been done already.

In general I believe you are correct (again, that's one of my points),
but "nobody" is unfair to Aidan Kehoe, who has done exactly that on
(at least) one patch recently, and to Steve Youngs who actually fixed
(at least) one of the issues on jwz's wish list, which is from the
last millennium.  There are also a number of things that have been
hanging fire for years now (GTK, for one) but I don't know if that
should count.

 > Do you *really* think *years* later users still care about the stuff they 
 > reported way back then?

Yes, I do, because they send us URLs to the archives where they
reported the problem or sent a patch years ago.

 > (Yes, I'm very much annoyed about your uninformed rambling
 > about other people's bugtrackers being "black holes".)

Er, those bug trackers are open to the public.  It's easy enough to
find plenty of examples that XEmacs should not emulate.

But really, it's not about bug trackers, it's about the difference
between active maintainers who acknowledge responsibility for a well-
defined project, and XEmacs's current crew with "areas of interest"
they're reasonably good about, but few well-defined modules so that we
can automatically direct "pings" to appropriate places, and vast
tracts of unclaimed code.  All of the projects I've mentioned (except
maybe MacPorts) have generally good reputations, but trackers for
modules without active maintainers still end up being black holes.  Do
you deny that, even for Gentoo?

Now, the three posts in the thread preceding mine all disclaimed, at
least implicitly, responsibility for making sure that bugs get handled
at least with a WONTFIX and apology.  The point here is to find a way
to do that given our very limited resources, or admit that we're not
going to do so.  If not, well, enough XEmacs hackers think it will
make their work easier that an ITS is worth doing.  But I don't want
it advertised as a major breakthrough in responsiveness, especially
not to people like you and Hans and Mike Fabian, if nothing major is
going to change.


So, back to the thread.  Is a tracker going to help us to cover more
of the bug reports in a timely, useful way?  If so, how, and what else
do we need to do to make that work?