Minutes of the Bugzilla meeting

[Frédéric Buclin]

We had a Bugzilla meeting today. You can read the log at 
http://bugzilla.glob.com.au/irc/?c=bugzilla-meeting&a=date&s=11+Jul+2006&e=11+Jul+2006

* Related to the recurrent lack of reviewers, getting reviews can be 
somewhat slow, even for trivial patches. It appears that most of the 
time, patches need on average 3-5 revisions, sometimes more (especially 
for new contributors), but also sometimes less (for some reviewers, most 
likely).

* The questions are then: would it help more than it hurts to let some 
reviewers go and commit patches without requiring review first? Would 
this speed up development? Would this affect code quality and stability?

* The Bugzilla product needs to run its own code (dogfood), i.e. we 
should have a separate bugzilla.bugzilla.org installation running the 
tip code.

* We are going to have module owners. They will be allowed to commit to 
their own area without review (on the trunk only, and outside freezing 
periods. Patches for branches still require review + approval). This 
discussion will be internal to reviewers (the discussion won't be 
public). Actions to be taken are:

1. get a list of components that need owners on the wiki (LpSolit)
2. nominate people for ownership of said modules (reviewers + justdave 
for the final decision)
3. update dependencies for bug 127876 - dogfood (justdave)
4. write the new review policy (justdave)
5. developer documentation is going to have to be updated to reflect all 
of this (mkanat)

* Around September 1, we will decide if we postpone things which aren't 
ready for 3.0 or if we delay the freeze date to get them in.

* Support for Oracle has been removed from the roadmap as the guy 
working on it has left the Oracle Corp.

LpSolit