Re: Is there a committer in the house?

All,

Currently we have 8 patches "Ready for Committer" in the current CF.
Some of them have been that status for some time.

From traffic on this list, I'm getting the impression that nobody other
than Robert, Heikki and Tom are committing other people's patches. I
know we have more committers than that. Bruce? Simon? Itagaki? Bueller?

--
Josh Berkus
PostgreSQL Experts Inc.
http://pgexperts.com

On Thu, Jul 14, 2011 at 8:01 PM, Josh Berkus <josh(at)agliodbs(dot)com> wrote:

> Currently we have 8 patches "Ready for Committer" in the current CF.
> Some of them have been that status for some time.
>
> From traffic on this list, I'm getting the impression that nobody other
> than Robert, Heikki and Tom are committing other people's patches.  I
> know we have more committers than that.  Bruce?  Simon?  Itagaki?  Bueller?

Someone called Simon Riggs has committed one patch by another author
and reviewed 3 others, as well as spending many days working on bugs
for beta.

It would be sensible if people based their comments on actual events
rather than negative impressions.

ISTM that you are right that there are other committers that have not
done anything. How strange that you name myself, yet not others.

--
 Simon Riggs                   http://www.2ndQuadrant.com/
 PostgreSQL Development, 24x7 Support, Training & Services

> ISTM that you are right that there are other committers that have not
> done anything. How strange that you name myself, yet not others.

Touchy today, eh?

And I do name others, read the email again.

Anyway, my question is: the list of committers I know of who have
general knowledge of the codebase and can commit a wide variety of other
people's patches are:

Tom
Robert
Heikki
Bruce
Simon

Who else?

--
Josh Berkus
PostgreSQL Experts Inc.
http://pgexperts.com

Josh Berkus wrote:
>
> > ISTM that you are right that there are other committers that have not
> > done anything. How strange that you name myself, yet not others.
>
> Touchy today, eh?
>
> And I do name others, read the email again.
>
> Anyway, my question is: the list of committers I know of who have
> general knowledge of the codebase and can commit a wide variety of other
> people's patches are:
>
> Tom
> Robert
> Heikki
> Bruce
> Simon

I have found my reading of the email lists is too delayed to effectively
commit other's patches.

--
Bruce Momjian <bruce(at)momjian(dot)us> http://momjian.us
EnterpriseDB http://enterprisedb.com

+ It's impossible for everything to be true. +

On tor, 2011-07-14 at 12:01 -0700, Josh Berkus wrote:
> Currently we have 8 patches "Ready for Committer" in the current CF.
> Some of them have been that status for some time.
>
> >From traffic on this list, I'm getting the impression that nobody other
> than Robert, Heikki and Tom are committing other people's patches. I
> know we have more committers than that. Bruce? Simon? Itagaki? Bueller?

I'm still working on preparing 9.1. I'm not following 9.2 development
yet.

Excerpts from Josh Berkus's message of jue jul 14 15:01:10 -0400 2011:
> All,
>
> Currently we have 8 patches "Ready for Committer" in the current CF.
> Some of them have been that status for some time.
>
> From traffic on this list, I'm getting the impression that nobody other
> than Robert, Heikki and Tom are committing other people's patches. I
> know we have more committers than that. Bruce? Simon? Itagaki? Bueller?

It seems that by mentioning some people but not all, you offended both
the people you mentioned (at least some of them, because they are
already actively helping) and those that you didn't (at least some of
them, because they are already actively helping; those that are not
completely inactive in the project, that is).

--
Álvaro Herrera <alvherre(at)commandprompt(dot)com>
The PostgreSQL Company - Command Prompt, Inc.
PostgreSQL Replication, Consulting, Custom Development, 24x7 support

Alvaro,

> It seems that by mentioning some people but not all, you offended both
> the people you mentioned (at least some of them, because they are
> already actively helping) and those that you didn't (at least some of
> them, because they are already actively helping; those that are not
> completely inactive in the project, that is).

Yeah, everybody's super-touchy this week. Must be the weather.

Speaking of which, is there anything you could commit? ;-)

--
Josh Berkus
PostgreSQL Experts Inc.
http://pgexperts.com

Excerpts from Josh Berkus's message of vie jul 15 16:32:42 -0400 2011:
> Alvaro,
>
> > It seems that by mentioning some people but not all, you offended both
> > the people you mentioned (at least some of them, because they are
> > already actively helping) and those that you didn't (at least some of
> > them, because they are already actively helping; those that are not
> > completely inactive in the project, that is).
>
> Yeah, everybody's super-touchy this week. Must be the weather.

It's been regularly horrible here, so yeah, that might explain it.

> Speaking of which, is there anything you could commit? ;-)

I intend to give the finishing touches to Pavel's plpgsql patch and
commit, so that's one item to take off the list. Everything else seems
to be in somebody else's hands, or is larger than I can grab ATM.

--
Álvaro Herrera <alvherre(at)commandprompt(dot)com>
The PostgreSQL Company - Command Prompt, Inc.
PostgreSQL Replication, Consulting, Custom Development, 24x7 support

Josh Berkus wrote:
> Alvaro,
>
> > It seems that by mentioning some people but not all, you offended both
> > the people you mentioned (at least some of them, because they are
> > already actively helping) and those that you didn't (at least some of
> > them, because they are already actively helping; those that are not
> > completely inactive in the project, that is).
>
> Yeah, everybody's super-touchy this week. Must be the weather.

Somehow blaming everyone else doesn't seem like the proper reaction. :-(

--
Bruce Momjian <bruce(at)momjian(dot)us> http://momjian.us
EnterpriseDB http://enterprisedb.com

+ It's impossible for everything to be true. +

>> Yeah, everybody's super-touchy this week. Must be the weather.
>
> Somehow blaming everyone else doesn't seem like the proper reaction. :-(

Look, it wasn't meant to be a complete list, or even a representative
one. That's why I tagged a "Bueller?" at the end. Even for those who
don't get the reference, there should have been a "Who the hell is Bueller?"

You might notice that we don't publish a list of committers anywhere.
In fact, *I* don't have one. So when I want to say "hey, who could be
doing committing and isn't currently?" I don't have any way to answer
that question. Which is why I posted the e-mail.

So, I reiterate: who do we have, as committers, who are capable of
committing a fairly wide array of patches? This would be good
information for every CF Manager to have, when things start getting
stuck in "ready for committer".

--
Josh Berkus
PostgreSQL Experts Inc.
http://pgexperts.com

Josh Berkus <josh(at)agliodbs(dot)com> writes:
> You might notice that we don't publish a list of committers anywhere.
> In fact, *I* don't have one.

http://wiki.postgresql.org/wiki/Committers

regards, tom lane

> http://wiki.postgresql.org/wiki/Committers

Oh, thanks! I didn't know that existed.

So, if any of the following people could possibly pick up even one patch
from the current commitfest and commit it, it would clear out our
pending commit list:

* Bruce Momjian
* Tatsuo Ishii
* Andrew Dunstan
* Magnus Hagander
* Greg Stark
* Itagaki Takahiro

... noting that I have already heard from Simon, Joe, Alvaro, Tom,
Robert, Peter and Heikki. I'm basing the above list on (a) some
knowledge of which folks only seem to work on very specific areas of
code, and (b) looking at pgsql-committers for the past 3 weeks.

In any case, it seems like we have a pool of 13 people would could be
committing general patches every commitfest. Not a huge number
considering the number of patches we get, but more than I see actually
committing stuff for most CFs.

--
Josh Berkus
PostgreSQL Experts Inc.
http://pgexperts.com

Excerpts from Josh Berkus's message of vie jul 15 18:33:14 -0400 2011:
>
> > http://wiki.postgresql.org/wiki/Committers
>
> Oh, thanks! I didn't know that existed.
>
> So, if any of the following people could possibly pick up even one patch
> from the current commitfest and commit it, it would clear out our
> pending commit list:

The "ready for committer" state does not mean that the committer can
grab the patch and apply it. Last time I checked, one was still
expected to review it and take full responsibility for any breakage
caused by it. Given this, I cannot responsibly grab more than a couple
patches, because then I'd be swamped when bug reports started coming in.

--
Álvaro Herrera <alvherre(at)commandprompt(dot)com>
The PostgreSQL Company - Command Prompt, Inc.
PostgreSQL Replication, Consulting, Custom Development, 24x7 support

Alvaro,

> The "ready for committer" state does not mean that the committer can
> grab the patch and apply it. Last time I checked, one was still
> expected to review it and take full responsibility for any breakage
> caused by it.

You're absolutely correct. Which is why committer bandwidth is such a
choke point.

--
Josh Berkus
PostgreSQL Experts Inc.
http://pgexperts.com

On Fri, Jul 15, 2011 at 5:37 PM, Bruce Momjian <bruce(at)momjian(dot)us> wrote:
> Josh Berkus wrote:
>> Alvaro,
>>
>> > It seems that by mentioning some people but not all, you offended both
>> > the people you mentioned (at least some of them, because they are
>> > already actively helping) and those that you didn't (at least some of
>> > them, because they are already actively helping; those that are not
>> > completely inactive in the project, that is).
>>
>> Yeah, everybody's super-touchy this week.   Must be the weather.
>
> Somehow blaming everyone else doesn't seem like the proper reaction.  :-(

I don't think Josh's tone is really the problem we should be worrying
about here. He's pointing out a legitimate problem. If you go back
and look at the CF app for 9.1, you'll see that Tom, Peter, and I
committed the overwhelming majority of the patches which were
submitted to CFs and went on to get committed. So if we have a
CommitFest where Tom is on vacation and Peter is devoting his time to
polishing release N-1 rather than new development on release N, then
we're either going to need a much larger investment of time by one or
more other committers, or we're not really going to get through
everything. When you lose the efforts of somebody who might commit 10
or 20 patches in a CF and comment usefully on another 10 or 20, it
leaves a big hole.

I don't believe Josh's intent was to disparage your contributions, or
Simon's, or Alvaro's, and it certainly isn't mine. I appreciate all
the work that has been done on this CommitFest by everyone who has
participated, reviewers as well as committers. At the same time, part
of the thankless task of being CF manager is asking people to step up
to the plate and do more. It takes a heck of a lot of work to get 70
patches reviewed and committed, and it is unlikely that we will ever
have enough people spontaneously step up to the plate to make that
happen. Since we can't call up people's bosses and complain that they
aren't doing enough work on the CommitFest, the CF manager is left
with the options of (1) trying to review (and commit?) all 30 or 40
remaining patches single-handedly or (2) begging. If we're then going
to complain when they do one or both of those things, well, I think
that's a bit unfair.

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

Robert Haas wrote:
> On Fri, Jul 15, 2011 at 5:37 PM, Bruce Momjian <bruce(at)momjian(dot)us> wrote:
> > Josh Berkus wrote:
> >> Alvaro,
> >>
> >> > It seems that by mentioning some people but not all, you offended both
> >> > the people you mentioned (at least some of them, because they are
> >> > already actively helping) and those that you didn't (at least some of
> >> > them, because they are already actively helping; those that are not
> >> > completely inactive in the project, that is).
> >>
> >> Yeah, everybody's super-touchy this week. ? Must be the weather.
> >
> > Somehow blaming everyone else doesn't seem like the proper reaction. ?:-(
>
> I don't think Josh's tone is really the problem we should be worrying
> about here. He's pointing out a legitimate problem. If you go back
> and look at the CF app for 9.1, you'll see that Tom, Peter, and I
> committed the overwhelming majority of the patches which were
> submitted to CFs and went on to get committed. So if we have a
> CommitFest where Tom is on vacation and Peter is devoting his time to
> polishing release N-1 rather than new development on release N, then
> we're either going to need a much larger investment of time by one or
> more other committers, or we're not really going to get through
> everything. When you lose the efforts of somebody who might commit 10
> or 20 patches in a CF and comment usefully on another 10 or 20, it
> leaves a big hole.

This mostly revoles around the problem of trying to finalize 9.1 while
applying 9.2 patches --- no surprise we don't have enough cycles to do
that.

--
Bruce Momjian <bruce(at)momjian(dot)us> http://momjian.us
EnterpriseDB http://enterprisedb.com

+ It's impossible for everything to be true. +

On Sat, Jul 16, 2011 at 10:04 PM, Bruce Momjian <bruce(at)momjian(dot)us> wrote:
> This mostly revoles around the problem of trying to finalize 9.1 while
> applying 9.2 patches --- no surprise we don't have enough cycles to do
> that.

Well, sorta. The fact that Josh got his head bitten off for
suggesting that we weren't going to get this CommitFest finished
without some more committer involvement could have happened at any
time in the release cycle (and has, in the past).

I think the question of whether we can overlap release cycles is a bit
of a red herring. We had our first CommitFest for 9.1 in July 2010,
and that ran quite smoothly, though it was also a release-cycle
overlap. One big difference this time is that Tom and Peter haven't
participated in this CommitFest very much at all (and I've done less
as well, due to other commitments), whereas they did last time around.
So I think the real question is not "how much bandwidth do we have as
a community?" but rather "what works for the key people who make the
process function?" and maybe "how can we induce other people to make
the kind of time commitment that Tom and Peter have in the past?".

I think the fact that we've managed to get 18 patches committed - and
will probably squeeze in a few more - despite Tom and Peter being busy
is pretty good. But it certainly emphasizes the extent to which we
depend on a relatively small number of contributors to do an awful lot
of the work.

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