Thread: CommitFest progress

CommitFest progress

From
Robert Haas
Date:
Of the 83 patches in this CommitFest, there are currently 35 that are
marked as needing review, 23 that are waiting on author, 7 that are
ready for committer, 11 that are committed, 5 that are returned with
feedback, and 2 that are rejected.  Since we're now supposedly in the
last week of this month-long CommitFest, that doesn't bode well.  I
have no problem continuing to work on the patches that are ready for
committer, and the people whose patches are marked as in need of
review because they have never been reviewed (such as, cough, my
background workers round three patch) surely deserve to get one.  But
I do have a problem with the fact that no concerted effort has been
made to mark patches that are clearly nowhere near ready for commit,
or which have been waiting on the author for extended periods of time,
as returned with feedback.  The CommitFest is supposed to be a time to
*commit the patches that are ready to be committed*, not to wait
indefinitely for them to become ready to be committed.  If we make it
into the latter, then nobody should be surprised when CommitFests
never end.

I therefore propose that we start by marking all of the patches that
are currently Waiting on Author as Returned with Feedback.  Most of
them have been that way for a long time.

Then, I think all of the people who are listed as reviewers need to
take a look at the current state of their patches and decide whether
or not they are reasonably ready to be committed.  If they are, then
they should be marked Ready for Committer.  If they're not, but
they've had at least one good thorough review, then they should also
be marked Returned with Feedback; they can be resubmitted for the next
CommitFest for further review.  If they haven't had a good thorough
review yet, then the people who signed up to review should, uh, hurry
up and do that.  If they can't, they should remove their name from the
patch and we should seek other volunteers.

Finally, we need to find reviewers for the patches that still don't
have them, or that lack them once the names of people who aren't
really going to review get removed.  I have certainly reviewed more
patches this CommitFest than I submitted by quite a bit, but I'm still
willing to help with a few more if that's needed.  I can't, however,
do all of them, especially if I'm on the hook to also commit
everything I review once it becomes ready.

Thanks,

-- 
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company



Re: CommitFest progress

From
Dimitri Fontaine
Date:
Robert Haas <robertmhaas@gmail.com> writes:
>   The CommitFest is supposed to be a time to
> *commit the patches that are ready to be committed*, not to wait
> indefinitely for them to become ready to be committed.

I beg to differ. Commit Fests are the time when patch authors know they
can get feedback from the community and in particular committers.

So as a patch author it's best if you can arrange your schedule and be
ready to submit new versions as asked, or comment on your design choices
and trade-offs, etc.

Patch commit can happen whenever in the cycle at the discretion of the
committer. Commit Fest are all about *review* and *feedback*.

> I therefore propose that we start by marking all of the patches that
> are currently Waiting on Author as Returned with Feedback.  Most of
> them have been that way for a long time.

That seems fair.

> Then, I think all of the people who are listed as reviewers need to
> take a look at the current state of their patches and decide whether
> or not they are reasonably ready to be committed.  If they are, then

I've been distracted away from this commit fest but should be able to
get back to it now. Will post soon about the patches I enrolled myself
with.

Regards,
-- 
Dimitri Fontaine
http://2ndQuadrant.fr     PostgreSQL : Expertise, Formation et Support



Re: CommitFest progress

From
Robert Haas
Date:
On Thu, Oct 10, 2013 at 5:21 AM, Dimitri Fontaine
<dimitri@2ndquadrant.fr> wrote:
> Robert Haas <robertmhaas@gmail.com> writes:
>>   The CommitFest is supposed to be a time to
>> *commit the patches that are ready to be committed*, not to wait
>> indefinitely for them to become ready to be committed.
>
> I beg to differ. Commit Fests are the time when patch authors know they
> can get feedback from the community and in particular committers.
>
> So as a patch author it's best if you can arrange your schedule and be
> ready to submit new versions as asked, or comment on your design choices
> and trade-offs, etc.
>
> Patch commit can happen whenever in the cycle at the discretion of the
> committer. Commit Fest are all about *review* and *feedback*.

Sure, I don't disagree with any of that.

>> I therefore propose that we start by marking all of the patches that
>> are currently Waiting on Author as Returned with Feedback.  Most of
>> them have been that way for a long time.
>
> That seems fair.
>
>> Then, I think all of the people who are listed as reviewers need to
>> take a look at the current state of their patches and decide whether
>> or not they are reasonably ready to be committed.  If they are, then
>
> I've been distracted away from this commit fest but should be able to
> get back to it now. Will post soon about the patches I enrolled myself
> with.

Thanks.

-- 
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company



Re: CommitFest progress

From
Robert Haas
Date:
On Wed, Oct 9, 2013 at 2:03 PM, Robert Haas <robertmhaas@gmail.com> wrote:
> I therefore propose that we start by marking all of the patches that
> are currently Waiting on Author as Returned with Feedback.  Most of
> them have been that way for a long time.

Hearing no objections, I went through and did this, but skipped some
that had recent activity, and instead marked one as ready for
committer on the basis that the reviewer thought it was in good shape
except for needing test case revision.  Together with today's rash of
commits, this means that we've now disposed of half the patches in the
CommitFest.  IOW, we have a lot of work left to do.

-- 
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company