Re: Status of Bugzilla?

["Joseph S. Myers" <jsm28 at cam dot ac dot uk>]

Observations on the states of bugs in Bugzilla:

>From the definitions, it looks like "UNCONFIRMED" means what GNATS has as 
"open", and "NEW" means "analyzed".  So new email submissions should start 
in "UNCONFIRMED".

We don't have the QA system for verifying bug fixes implicit in the 
difference between "RESOLVED", "VERIFIED" and "CLOSED".  Thus, these 
should all be merged into one "CLOSED".

"WONTFIX" isn't a proper way to resolve a bug; we shouldn't have it.  
"LATER" and "REMIND" are both analogues of "suspended"; I think just one
would suffice.  Furthermore, "suspended" indicates an unfixed bug, not a
fixed one, so such bugs should show up on default searches, as should
UNCONFIRMED.  This may mean not marking them RESOLVED/..., but having a
new status value to cover nonresolutions that we now have as "feedback"  
and "suspended", or treating those as status values rather than
resolutions (which I think would be the better option, meaning that
UNCONFIRMED, NEW, ASSIGNED, REOPENED, SUSPENDED, FEEDBACK and CLOSED would
be the statuses, and FIXED, INVALID, DUPLICATE, WORKSFORME the
resolutions).

Can we have some milestones to play with when the database is next
converted from GNATS?  (At least, a 3.1 milestone, to which "high"  
priority bugs are assigned, but maybe some more.  Note that incoming
"high" priority bugs - from old gccbug versions, current gccbug doesn't
allow them - shouldn't get milestones assigned to them, just those
currently in GNATS.)

(Previously noted:) the summary version list should distinguish 2.96 
(redhat) from other 2.96, and 3.0-pre from 3.0 release.

(Previously noted:) the full version string from bug reports (new and in
GNATS) should be preserved, as well as the summary version for searches.

(Previously noted:) the Class field from bug reports is useful and should
be preserved.  (OTOH, I doubt that Organization is, and Submitter-Id
surely is not.)

-- 
Joseph S. Myers
jsm28@cam.ac.uk