Re: Getting Bugzilla support into Bug-buddy

[Jody Goldberg <jgoldberg home com>]

On Tue, Feb 06, 2001 at 07:29:20PM +0100, Martin Baulig wrote:
> 
> Rejecting such bug reports (and even silently dropping them if the sender's
> system is misconfigured (no valid email address)) is the only way to do it.
> 
> Look at the current situation in bugs.gnome.org; such bugs are currently
> accepted by the system but nobody ever looks at them. They're just slowing
> the whole system down and eating up resources.

I don't agree.  Some bug reports with invalid emails are useful some
are not.  Part of the responsibility of being 'maintainer' is
diligently handing incoming reports.  Going through vast majority
of ancient fixed problems as they arrive is part of the job.

> Telling people "your bug report has been rejected for reason blah, blah, blubb"
> is way better to just let it go to some file somewhere in a spool directory
> where nobody ever looks at it.

The only reason to potentially toss a report would be no package and
no email,  and even then just tossing the report can potentially
lose information.  If we can identify a package (even without valid
reply address) we should keep it.
 
> mailing system before sending out bug reports. Applications like bug-buddy
> can help to enforce this, for instance by refusing to send out bug reports
> from addresses like * * localdomain and by telling people that they must
> specify a working email address to submit a bug report.

Excellent idea.
 
> To summarize, this is a bug report for a package which has a Bugtracker
> somewhere and with a maintainer who is actually interested in bug reports
> for his package and who's actually fixing such bugs. The submitter has a
> working From: line so that he'll get and read a reply.
> 
> In this case, we _must_ reject the bug report and tell the submitter where
> to send it. Not doing this effectively means that this bug report will be
> discarded (since it'll end up somewhere in bugzilla.gnome.org where the
> maintainer of the package will never look) while it would have been fixed
> if the submitter sent it to the correct address.

Agreed.
 
> We cannot expect package maintainers to keep track of more than one bug
> tracking system for their package, so it's only fair to help them to avoid
> that their bug reports are cluttered all over x different bug tracking
> systems while they should all be in one system.

Not entirely true.  The various distributions are fertile sources of
reports (eg debian).  Some are distro specific, some need to be
nudged 'upstream' into bugzilla.gnome.