hadoop-general mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Arun C Murthy <...@hortonworks.com>
Subject Commit then Review (was Re: HDFS-1623 branching strategy)
Date Mon, 11 Jul 2011 22:52:29 GMT
On Jul 11, 2011, at 3:05 PM, Jakob Homan wrote:

> The question at hand is does the community want to go ahead with the
> 1073 model, or not?  My preference is not, but since some in the
> community like it, I'd suggest making a small tweak to bring it closer
> in line with the intent of our bylaws, which clearly are RTC: "The
> code can be committed after the first +1."  It's not productive to
> re-litigate the individual commits on 1073.

Ok, I've tried hard to stay out - but, I'll bite.

It's clear that Hadoop needs to move forward, and fast.

Trying to litigate our progress on feature branches seems backwards. IMO, it's more important
to get things done than trying forever to get it perfect*.

I've done CTR on MR-279 with great results. I know of *several* ASF projects who do the same.

Everyone 'committing' to the feature branch is a committer anyway - which means we implicitly
trust them. Anyone is welcome to review patches and provide feedback on the CTR branch. Please
do review the branch as often as you wish - heck, feel free to revert a patch if you want.
If not, we owe a simple 'thank you' to the people who are doing the work.

However, blocking RTC en-masse is just counter-productive. We should let folks doing the work
decide how best they want to do the work.

To be clear: I support RTC for trunk and more stringent reviews for merge. Yet, all these
need to be time-bound, we cannot afford silent vetoes to block progress.

I don't really know enough about HDFS-1623, so I'm not qualified to speak about the mechanics
for this branch.


* http://www.jwz.org/doc/worse-is-better.html
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message