Re: branching for 9.2devel

On Sun, May 1, 2011 at 1:14 PM, Kevin Grittner
<Kevin(dot)Grittner(at)wicourts(dot)gov> wrote:
> Robert Treat <rob(at)xzilla(dot)net> wrote:
>> Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
>
>>>        CF #1: June 1-30
>>>        CF #2: August 1-31
>>>        CF #3: October 1-31
>>>        CF #4 (one week shortened CF): December 1-7
>>>        CF #5: January 1-31
>>>
>>> I think the main thing we have to think about before choosing is
>>> whether we believe that we can shorten the CFs at all.  Josh's
>>> proposal had 3-week CFs after the first one, which makes it a lot
>>> easier to have a fest in November or December, but only if you
>>> really can end it on time.
>>
>> If we made the "deadline" for patch acceptance into 9.2 CF#4, then
>> shortening that to a two week cycle whose main goal was simply to
>> sanity check patches for 9.2 would probably work. Most would
>> probably still need further work, which we would expect to get
>> handled in the final, full CF#5, but we wouldn't let anything new
>> come into CF#5. This way when we get the 100 patch pile up in
>> CF#4, there's no expectation that those patches will be committed,
>> just that they can be sanity checked for the 9.2 release.
>
> Which makes it not really a CommitFest, but rather ... a SanityFest?
>
> To make sure I understand you, you're suggesting no WIP patch review
> in the last two CFs?  (Of course nothing stops someone from looking
> at someone else's WIP between fests.)  Would a patch submitted to
> #4, the sanity of which was questioned, be eligible for another try
> in #5.
>

I think you can have WIP patches for both CF#4 and CF#5. What we're
hoping to get from CF#4 is a better scope on the number of patches we
might have to get 9.2 out the door. WRT patches whose sanity is
questioned, I'd presume that questioning would have a list of
specific complaints, so if you address those between CF#4 and CF#5, I
don't see why you can't try again.

Robert Treat
play: xzilla.net
work: omniti.com