Re: Do we have a CF manager for November? - Mailing list pgsql-hackers

From Michael Paquier
Subject Re: Do we have a CF manager for November?
Date
Msg-id 20191128065916.GA259486@paquier.xyz
Whole thread Raw
In response to Re: Do we have a CF manager for November?  (Michael Paquier <michael@paquier.xyz>)
Responses Re: Do we have a CF manager for November?  (Michael Paquier <michael@paquier.xyz>)
List pgsql-hackers
On Tue, Nov 05, 2019 at 11:18:00AM +0900, Michael Paquier wrote:
> That may have been me.  I can take this one if there is nobody else
> around.
>
> Note: I have switched the app as in progress a couple of days ago,
> after AoE was on the 1st of November of course.

So, we are close to the end of this commit fest, and I have done a
first pass on something like one third of the entries, mainly updating
incorrect patch status, bumping them into next CF or closing stale
items waiting on author.  As things stand, the progress is not that
good taking the total number of patches:
Needs review: 109.
Ready for Committer: 7.
Committed: 36.
Moved to next CF: 40.
Withdrawn: 4. Rejected: 6.
Returned with Feedback: 19
Total: 221.

So we still have a lot of patches in need of review, and most of them
will likely get bumped to the next CF.  The not-so-good news is that
numbers tend to be comparable with the last last CF.  An actual bad
news is that, based on what I have looked at until now, I have noticed
that a certain number of patches had an incorrect status for *weeks*
so it is actually necessary to go through each patch to make sure that
things are in a correct state in the CF app, which increases quite a
bit the classification work burden.

When looking at a patch, I usually try to use the following rules for
classification in a current CF:
- If patch is in "Needs Review" state, bump it to the next CF.
- If patch is in "Waiting on Author", with a state not updated for at
least two weeks, mark it as returned with feedback.
- If patch is in "Waiting on Author", with a state updated recently
(aka within two weeks), bump it to next CF with the same state.  This
requires more manual steps than the rest as the CF app does not allow
moving a patch to next CF waiting on author, but that's not right
either to bump out a patch without giving time to the author to answer
back.  Using half of the CF period looks rather right regarding that.

If you are registered as a patch author, reviewer or even committer,
it would be nice to look at each item you are involved in, and make
then sure that the patch you are looking at is in a correct state to
prevent any errors.  If you can move it to next CF or mark it as RwF
by yourself, this also saves cycles to the CFMs.  (Thanks Ibrar I
have noticed your activity!).

Thanks,
--
Michael

Attachment

pgsql-hackers by date:

Previous
From: Pavel Stehule
Date:
Subject: Re: proposal: type info support functions for functions that use"any" type
Next
From: Konstantin Knizhnik
Date:
Subject: Re: Why JIT speed improvement is so modest?