zookeeper-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Enrico Olivelli <eolive...@gmail.com>
Subject Re: [Suggestion] Use Co-authored-by in commit messages
Date Wed, 08 May 2019 18:17:19 GMT
Yes, it is a good idea to have a common practice for tracking the original
author.

IMHO this is up to the person who is picking up an old patch, it is his own
responsibility.

IIRC In Bookkeeper we keep the original author of the patch if the patch is
a straight port from another private company fork with minimal changes.

Having a Co-author is good from my side. I am not sure we can force it

My 2cents

Enrico

Il mer 8 mag 2019, 19:31 Brian Nixon <brian.nixon.cs@gmail.com> ha scritto:

> +1 to the idea of multiple authors, particularly for rescued code
>
> -1 to including all reviewers in the commit proper, this information is
> easily enough found from poking at the mail archive where "original author"
> requires studying a ticket on jira
>
> awesome idea!
>
>
> On Wed, May 8, 2019 at 6:32 AM Norbert Kalmar <nkalmar@cloudera.com.invalid
> >
> wrote:
>
> > Sorry everyone for the multiple emails...
> > So, I get your suggestion now Maoling, sorry for the confusion.
> > We already indicate the reviewer if it's from an apache email, as it
> looks
> > to me. (Doesn't have to be ZooKeeper committer). We should add external
> > emails as well.
> >
> > So I just clarified this with Andor, looks like this is a manual entry
> (the
> > names/emails itself) during the commit (script).
> >
> > Let's hear what others think :)
> >
> >
> > On Wed, May 8, 2019 at 3:24 PM Norbert Kalmar <nkalmar@cloudera.com>
> > wrote:
> >
> > > Well, HBase does it for example, commits have a "Signed-off-by: ..."
> > line.
> > >
> > > All right, votes on for co-author and signed-off-by :)
> > >
> > >
> > >
> > > On Wed, May 8, 2019 at 2:58 PM Norbert Kalmar <nkalmar@cloudera.com>
> > > wrote:
> > >
> > >> Thanks Maoling, I also think encouraging code review as well is a good
> > >> idea, but, unfortunately I have a "but" :)
> > >> I see two issues with including reviewers in the commit message.
> > >> First, I don't think there is a method to automate this, although I
> > think
> > >> the commit script the committers are using can be modified to include
> > it.
> > >> Otherwise doing manually would complicate merging PRs for committers.
> > >> My other, bigger issue is that there is nothing to track this
> > >> information. At least I am not aware of anything. What I mean is
> Github
> > >> tracks authors of the commits. But what would we use the reviewers
> > >> information? If you just want to check reviewers for whatever reason,
> > there
> > >> is a filter for that already on github, in the Pull Request view. And
> > this
> > >> would also make the commit message more "bloated".
> > >>
> > >> I'm not saying we shouldn't do this (not a -1 from my side), I just
> have
> > >> my concerns mentioned above.
> > >>
> > >> Is there any Apache project doing this? Just out of curiosity.
> > >>
> > >> Regards,
> > >> Norbert
> > >>
> > >>
> > >> On Wed, May 8, 2019 at 2:34 PM Justin Ling Mao <
> > maoling199210191@sina.com>
> > >> wrote:
> > >>
> > >>> +1,A very good Suggestion.Thanks Norbert.I also suggest about the
> > >>> sign-off of the Reviewers' name.For the incentive, if someone
> > participate
> > >>> in the review of PR, no matter whether he/she is a committer, we all
> > need
> > >>> include his/her name?
> > >>>
> > >>> ----- Original Message -----
> > >>> From: Norbert Kalmar <nkalmar@cloudera.com.INVALID>
> > >>> To: dev@zookeeper.apache.org
> > >>> Subject: [Suggestion] Use Co-authored-by in commit messages
> > >>> Date: 2019-05-08 17:36
> > >>>
> > >>> Hi Devs,
> > >>> I've got this idea from HBase.
> > >>> So: when there is a patch that is abandoned by its original author
> for
> > >>> any
> > >>> reason, and it can no longer be merged, someone comes by, and asks
to
> > >>> continue to work on it. Usually the reply is to use the change freely
> > or
> > >>> no
> > >>> reply at all. Either way, what people end up doing is a new pull
> > request,
> > >>> and (correct me if I'm wrong) we do not have a standardized method
> how
> > to
> > >>> indicate, or even to indicate at all the original author.
> > >>> My proposal is to use github's feature of Co-author, which is a way
> of
> > >>> attributing multiple authors of a given commit. See more details
> here:
> > >>>
> > >>>
> >
> https://help.github.com/en/articles/creating-a-commit-with-multiple-authors
> > >>> I wouldn't think this needs to be forced or anything on future PRs,
> but
> > >>> it's a nice thing to have. And if someone sees an old patch, this
> could
> > >>> give more incentive to continue to work on it, knowing there's a
> > >>> guideline
> > >>> in the HowToContribute guide to credit him/her.
> > >>> I can update the guide at
> > >>>
> https://cwiki.apache.org/confluence/display/ZOOKEEPER/HowToContribute
> > if
> > >>> the reception is positive.
> > >>> Regards,
> > >>> Norbert
> > >>>
> > >>
> >
>

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