Home | About | Sematext search-lucene.com search-hadoop.com
 Search Hadoop and all its subprojects:

Switch to Threaded View
HBase >> mail # dev >> HBASE-7088 ready to commit ;)


Copy link to this message
-
Re: HBASE-7088 ready to commit ;)
My point is that it forces more eyes to look at what is done and might
allow (force) more discussions?

On your side Andrew, regarding 0.98, can you please commit on not taking
any vacations for the next 12 months please? ;)
2013/12/18 Andrew Purtell <[EMAIL PROTECTED]>

> I don't see how more votes will increase the quality of design. Your
> participation in the community and providing of feedback will increase the
> quality of design, and code too if you volunteer time to do code review on
> JIRAs and reviews.apache.org.
>
> I can understand the motivation of adding friction. We already provide
> substantial friction as part of the normal Apache process of having Release
> Managers. If LarsH or Stack go away for a month, we probably won't have
> commits to 0.94 or 0.96 during that time. I'm not convinced we need more
> friction than this so I'm not in favor of the policy as it is currently
> written in the online manual. That said, I certainly have an open mind
> about this subject. Perhaps there is some evidence that more friction is
> warranted?
>
>
> On Wed, Dec 18, 2013 at 4:59 PM, Jean-Marc Spaggiari <
> [EMAIL PROTECTED]> wrote:
>
> > Honestly I'm pretty fine with the policy described at
> > http://hbase.apache.org/book.html#decisions
> >
> > "Patches that fit within the scope of a single Apache HBase component
> > require, at least, a +1 by one of the component's owners before commit.
> If
> > owners are absent -- busy or otherwise -- two +1s by non-owners will
> > suffice. " that mean usually one +1 is enough, or sometime 2x +1.
> >
> > But I will not complain if I need just one +1 for this patch ;)
> >
> > I think having 2x +1 will increase the quality of the code/design.
> >
> >
> >
> > 2013/12/18 Andrew Purtell <[EMAIL PROTECTED]>
> >
> > > Also let me clarify something: A while back we discussed the Hadoop
> > policy
> > > of requiring 3 +1s for a branch merge. That sounds reasonable to me. I
> > > don't see this in the current online manual text.
> > >
> > >
> > > On Wed, Dec 18, 2013 at 4:47 PM, Andrew Purtell <[EMAIL PROTECTED]>
> > > wrote:
> > >
> > > > Eh, that must have been discussed when I wasn't there or on the phone
> > and
> > > > unable to hear clearly. I'm not in favor of that policy as stated.
> > > > Ownership isn't working out as far as I can see. Owners are not
> around
> > > > enough. In fact I would say many people are relatively absent from
> the
> > > > community for long stretches of time. That's fine, this is a
> volunteer
> > > > society. We can't gate on an owner +1. I am not in favor of requiring
> > > more
> > > > than one +1 except for the obvious case where a committer should not
> +1
> > > and
> > > > commit their own work. I am in favor of continuing our informal
> policy
> > of
> > > > CTR for trivial changes.
> > > >
> > > >
> > > >
> > > >
> > > > On Wed, Dec 18, 2013 at 4:42 PM, Stack <[EMAIL PROTECTED]> wrote:
> > > >
> > > >> Here is what we decided as 'policy' on +1s:
> > > >>
> > > >> http://hbase.apache.org/book.html#decisions
> > > >>
> > > >> At our last meetup, we talked of upping the commit friction some to
> > give
> > > >> chance for more review before commit but this suggestion did not
> > > progress
> > > >> beyond discussion.
> > > >>
> > > >> St.Ack
> > > >>
> > > >>
> > > >> On Wed, Dec 18, 2013 at 3:32 PM, Andrew Purtell <
> [EMAIL PROTECTED]>
> > > >> wrote:
> > > >>
> > > >> > AFAIK, we just don't want a committer to +1 their own work.
> > > >> >
> > > >> >
> > > >> > On Wed, Dec 18, 2013 at 3:30 PM, Andrew Purtell <
> > [EMAIL PROTECTED]>
> > > >> > wrote:
> > > >> >
> > > >> > > No
> > > >> > >
> > > >> > >
> > > >> > > On Wed, Dec 18, 2013 at 3:27 PM, Jean-Marc Spaggiari <
> > > >> > > [EMAIL PROTECTED]> wrote:
> > > >> > >
> > > >> > >> Don't we need 2 commiters  +1 per JIRA?
> > > >> > >>  Le 2013-12-18 18:23, "Andrew Purtell" <[EMAIL PROTECTED]> a
> > > >> écrit :
> > > >> > >>
> > > >> > >> > Why is one +1 not good enough for commit?