Re: bug-buddy and your frequently duplicated bugs (how to reduce your bugzilla spam!)

[Elijah Newren <newren gmail com>]

On Tue, 04 Jan 2005 11:34:19 +1100, Wayne Schuller
<k_wayne schuller id au> wrote:

> Something I tried to do when setting up the mostfrequent/bugbuddy
> interface, was try to get bug buddy to append the report to the most
> frequent bug (and add the reporter to the cc list for the bug), rather
> than discourage submitting anything for a popular duplicate bug.

Interesting idea.  Note that bug-buddy does include a link directly to
the bug report in its interface, so that users can go and read what
information there already is.  A simple, though less elegant solution
to this problem could be to just change the wording in bug-buddy to
encourage reports to add information directly to the report.

> At the time, I couldn't get bugzilla or its interface to do this.
> 
> Maybe this is possible with the new gnome bugzilla?

I'm not aware of anything in the new bugzilla that would make this any easier.

> If this could work, it would mean that people COULD submit their report
> for a highly duplicate bug, but all the info would go in the one report,
> and people working on the bug would have an immediate audience of people
> suffering from the bug to interact with.
> 
> Duplicate bug reports are actually useful data, we just need to allow
> them to not cause inordinate time to triage or administrate.

Yes, they are useful, though their usefulness often decreases as the
number of reports increase (and they can be useless if the bug is
already fixed).  But, one thing that makes this not as big of an issue
with the current way things are set up is the fact that lots of people
don't take the time to look at the list that bug-buddy displays and we
will thus continue to get duplicates anyway (note that there have been
around 500-600 duplicates of bug 94625 filed AFTER the bugbuddy
keyword was added; it's still in the list of most frequently
duplicated bugs over the last month despite the fact that it was fixed
over two years ago).

> Of course I am too busy to help with this. :)

Understood--so am I.  :-)  If we were to do this, though, there's a
logistical problem that would need to be thought out.  Having lots of
stack traces appended to a bug report would make it near impossible to
follow the train of conversation when trying to gather info and figure
out what's going wrong--any sufficiently long explanation typed up by
a reporter would do likewise.  So we'd need a way to avoid this.


Cheers,
Elijah