Обсуждение: Commitfest 2018-11

Hi,

Commitfest 2018-11 is closed. Thanks everyone who participated by posting
patches, reviewing them, occasionally committing them, commenting on ideas or
just raising one eyebrow in surprise while reading the hackers mailing list.

I've turned off my spam machine and wanted to share some results:

* Obligatory graph with the numbers of committed, moved, returned and rejected
  CF patches over the time you can find in [cf_items_status.png]

* It was a bit frustrating to see a significant number of stale items, where an
  author was not interested in it anymore (and sometimes both author and
  assigned reviewer). Just out of curiosity I quickly looked at the following
  statistics - in the set of all the CF patches authors (I was considering an
  author as a person, who send a first email in thread), how many commitfests
  happened between the first and the last submitted patch (it could be the same
  patch going through several commitfests). The histogramm of the distribution
  for this value you can find in [cf_authors_span.png] - as you can
see the majority
  of authors submitted only once. Of course this statistics may not be exactly
  correct, since there are other ways of contributing patches besides just
  registering in the application, and there could be a lot of reasons for this
  kind of trend. But at the end of the day it looks like the scenario, when a
  person managed to overcome "entrance" complexity, read the code, sometimes
  understand and modify it to achieve a goal - and then stopped. I hope it's
  not a sign that the community is loosing potential contributors.

* I also noticed, that a lot of people add themselves as reviewers without
  actually doing any review. I started to think that probably it's some sort of
  "bookmarking", so probably it would be a good feature request for CF
  application - to allow mark some items as "interesting" for you.

Dmitry Dolgov <9erthalion6@gmail.com> writes:
> Commitfest 2018-11 is closed. Thanks everyone who participated by posting
> patches, reviewing them, occasionally committing them, commenting on ideas or
> just raising one eyebrow in surprise while reading the hackers mailing list.

Thanks for all the hard work of CF managing!

> * I also noticed, that a lot of people add themselves as reviewers without
>   actually doing any review. I started to think that probably it's some sort of
>   "bookmarking", so probably it would be a good feature request for CF
>   application - to allow mark some items as "interesting" for you.

Yeah, I've been annoyed by that too (not that I've not been guilty of
it myself).  We should probably tell people not to add themselves as
reviewers unless they're actively planning to review soon, because it
discourages other people from reviewing the same item.  Maybe also,
when a patch gets moved to the next CF, we should drop all reviewers
(ie make them sign up afresh if they're still interested).

            regards, tom lane

On Sat, 1 Dec 2018 at 12:38, Tom Lane <tgl@sss.pgh.pa.us> wrote:
>
> Maybe also,
> when a patch gets moved to the next CF, we should drop all reviewers
> (ie make them sign up afresh if they're still interested).
>

i like that idea, fwiw

-- 
Jaime Casanova                      www.2ndQuadrant.com
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services

On Sat, Dec 01, 2018 at 06:24:47PM +0100, Dmitry Dolgov wrote:
> I've turned off my spam machine and wanted to share some results:

Thanks for doing the work of this commit fest, Dmitry!  This is a
difficult task.

> * I also noticed, that a lot of people add themselves as reviewers without
>   actually doing any review. I started to think that probably it's some sort of
>   "bookmarking", so probably it would be a good feature request for CF
>   application - to allow mark some items as "interesting" for you.

This is not a new pattern unfortunately :(
--
Michael

On 12/1/18 1:46 PM, Lætitia Avrot wrote:
> 
> 
>     Yeah, I've been annoyed by that too (not that I've not been guilty of
>     it myself).  We should probably tell people not to add themselves as
>     reviewers unless they're actively planning to review soon, because it
>     discourages other people from reviewing the same item.  Maybe also,
>     when a patch gets moved to the next CF, we should drop all reviewers
>     (ie make them sign up afresh if they're still interested).
> 
> 
> +1 on that one. It's not because you had time on the previous commit 
> fest that you'll have time for the next.

+1.  I've been in this situation myself.

-- 
-David
david@pgmasters.net