calcite-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Gian Merlino <g...@imply.io>
Subject Re: [DISCUSS] Reviewing pull requests
Date Thu, 24 Aug 2017 20:38:34 GMT
We have a similar problem in the Druid community. PR review backlogs are
long and not all committers help out equally. We're grateful when anyone
does show an interest in developing the Druid community rather than just
using Druid or contributing patches. But that interest in the community is
not universal. I don't think there are easy solutions, since it's a free
rider problem at heart, and it's tough to force people to help out.

My experience in Druid is that productive reviewers are usually either
doing it out of love for the project, or because of corporate sponsorship
(i.e. they're doing it at work because community development is one of
their job responsibilities). I'm not sure there are any other ways to get
people to contribute, short of making it mandatory (you can't get your
patch reviewed unless you review someone else's) or at least a strong
community norm (community members hassle you if you don't help out). I
guess that starting this thread, Julian, may end up making this more of a
community norm.

Gian

On Thu, Aug 24, 2017 at 12:40 PM, Julian Hyde <jhyde@apache.org> wrote:

> Thanks for your support, Michael.
>
> Not a peep from the rest of you. I suppose there’s no incentive to speak
> up while the free lunch keeps on coming.
>
> What should I do to get some assistance from the many people that benefit
> from this being a healthy community? Do I have to threaten to go on strike?
>
> Julian
>
>
>
> > On Aug 22, 2017, at 8:01 AM, Michael Mior <mmior@uwaterloo.ca> wrote:
> >
> > I agree this is an important problem to solve. The biggest barrier for me
> > is that I'm not familiar with most of the Calcite code base and I don't
> > really have the time to invest right now in order to become familiar to
> the
> > point I'd be comfortable reviewing the majority of PRs. I do have
> > notifications of PRs enabled for the repo and will try to make an effort
> to
> > review when I feel qualified.
> >
> > --
> > Michael Mior
> > mmior@apache.org <mailto:mmior@apache.org>
> >
> > 2017-08-21 15:33 GMT-04:00 Julian Hyde <jhyde@apache.org <mailto:
> jhyde@apache.org>>:
> >
> >> I was on vacation last week. Before I went on vacation I sent an email
> to
> >> this list[0] asking the community of committers to stay on top of pull
> >> requests. My rationale, not explicitly stated in the email but hopefully
> >> clear to everyone, is that we need to respond to contributors in a
> timely
> >> fashion, otherwise they will not join the community. “Timely” means a
> day
> >> or two, maximum.
> >>
> >> Since I went on vacation 8 days ago, an existing PR was committed, but
> 10
> >> PRs have been created[1], and only one of them received feedback [2].
> >> (Thanks, Jesus, for the commit and the review.) Several JIRA cases have
> >> been logged, also.
> >>
> >> I don’t think this is an acceptable amount of feedback to sustain the
> >> community.
> >>
> >> The Calcite community is successful and growing, but it takes work to
> keep
> >> it going. I’m tired of being the main person who does that work.
> Committers
> >> and PMC members, you all benefit from the community. You (rightly)
> expect
> >> that your own PRs are reviewed and committed in a timely manner. How
> can I
> >> encourage you to take a greater share of the work? Do I need to take
> more
> >> vacation?
> >>
> >> Julian
> >>
> >> [0] http://mail-archives.apache.org/mod_mbox/calcite-dev/
> >> 201708.mbox/%3CFF047CF1-7B7D-4F4B-BB05-764BB7F9E70D%40apache.org%3E <
> >> http://mail-archives.apache.org/mod_mbox/calcite-dev/ <
> http://mail-archives.apache.org/mod_mbox/calcite-dev/>
> >> 201708.mbox/%3CFF047CF1-7B7D-4F4B-BB05-764BB7F9E70D@apache.org <mailto:
> 3CFF047CF1-7B7D-4F4B-BB05-764BB7F9E70D@apache.org>%3E>
> >>
> >> [1] https://github.com/apache/calcite/pulls <
> https://github.com/apache/calcite/pulls> <https://github.com/apache/ <
> https://github.com/apache/>
> >> calcite/pulls>
> >>
> >> [2] https://github.com/apache/calcite/pull/523#pullrequestreview
> -57275412 <https://github.com/apache/calcite/pull/523#pullrequestrevie
> w-57275412>
> >> <https://github.com/apache/calcite/pull/523#pullrequestreview-57275412
> <https://github.com/apache/calcite/pull/523#pullrequestreview-57275412>>
>
>

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message