On Wed, 2006-09-27 at 12:52 +0200, Adrian Brock wrote:
On Mon, 2006-09-25 at 10:40 -0600, Andrig T Miller wrote:
> The idea behind this is to keep releases from filling with
> everything, and them not getting reviewed. When things are directly
> assigned to releases, everyone just assumes it should be done, and
> there is no review. By having them initially unscheduled, and
> instituting a mandatory review, we can intelligently assign them to
> the proper release with the proper priority. They shouldn't remain in
> unscheduled for any length of time.
> Before we did this, releases just grow in scope
unchecked, and we
> cannot continue to do that.
You can't fix the problem of bugs not getting reviewed by
hiding them under the rug. :-)
Having them unscheduled doesn't hid them under any rug. They are
just as visible in JIRA.
Assigning to a release, forces somebody to remove
them from that release. So they are least looked at.
All it really does is force someone to go into JIRA and change the
fix version, not to actually evaluate the issue.
Where this failed before, is that they had no assignees.
So Dimitris just bumped to the next release without
pinging the assignee to tell them to
"to pull his finger out ..." :-)
If don't want to assign a bug to release,
e.g. it is too complicated to fix in the near future
there is a "No Release" dummy version.
But doing this, will probably lead to them
getting forgotten about!
To really solve this, is we need a formal review process. It
doesn't matter whether something is assigned to a release or not. We
have to have formal, mandatory review as a team.
Andrig (Andy) Miller
JBoss, a division of Red Hat