[MERGE] Create a new hook Branch.open. (Robert Collins)
Robert Collins
robertc at robertcollins.net
Fri Sep 5 06:31:32 BST 2008
On Fri, 2008-09-05 at 00:04 -0500, John Arbash Meinel wrote:
> I can agree that this particular instance did not. However, if we
> generally actually *encourage* feature freezes during FF week, that
> helps focus people to actually finish the things that they started, and
> get some time on the bug tracker.
I think FF is too late to be starting to look at the bug tracker. bug
management is an ongoing process, not one that spotty day-or-week
activities do much for.
> I think bzr as a project is getting pretty bad about pushing things
> through to completion. I see 29 patches on the "semi approved but not
> merged yet" pages:
> http://bundlebuggy.aaronbentley.com/?selected=pending&unreviewed=n
> http://bundlebuggy.aaronbentley.com/?selected=pending&start=24&unreviewed=n
>
> These are things that at least one person has approved, and often more
> than one. That is a sign that we are doing a lot of work that isn't
> getting merged. If it needs to be rejected/resubmitted fine. But it is
> sitting not getting reviews, or being approved and not getting merged.
> I've certainly been asking people to focus on reviewing (and I've gotten
> "My Pending" down to only 5 items).
So, we have 29 things open. Given 7-8 'core' devs, thats under 4 things
in flight each - but its actually spread out over a fair number of other
people. I don't think thats bad at all.
What I think is undesirable is the length of the tail - things sitting
around for ages.
> > The feature
> > freeze is about time to catch regressions IMO, not about stopping
> > everything. My memory of the release cycle may be spotty, but I thought
>
> That is true, I've explictly chosen for 1.7 to make it about getting
> things that have been done completed and out the door. I've mentioned it
> in most of my "1.7" update emails.
>
> > it was:
> > 3 weeks open
> > 1 week no disruptive changes
>
> We've had differing levels of what that week means. Including bringing
> in *tons* of disruptive changes during that week. (Witness 0.15, some of
> 1.6, etc.)
Both times we made an explicit decision to do so. And both times we've
regretted it. I don't think we need to get overly proscriptive in normal
times.
...
> We are just discussing what "frozen" means. We have generally been
> rather soft here, and often too soft. I'm probably being too strict for
> 1.7. I'll let others decide how clear I've been about wanting people to
> get things out of the review queue and actually merged. I know I've sent
> an email including that sentiment at least 2 or 3 times.
I think you're being very clear. But, its coming across as nagging and
badgering. I know Aaron has commented on it, and now, so am I. When I
was doing releases I really felt it was about just making sure that we
got a release out the door, without regressions - the rest is business
as normal, as we try to keep trunk unbroken and up to date with docs
etc. There should be anything more to the release, if there is our basic
process is broken and we should be working on that.
-Rob
--
GPG key available at: <http://www.robertcollins.net/keys.txt>.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part
Url : https://lists.ubuntu.com/archives/bazaar/attachments/20080905/6575246b/attachment.pgp
More information about the bazaar
mailing list