sqoop-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Arvind Prabhakar <arv...@apache.org>
Subject Re: [DISCUSS] RTC / CTR
Date Fri, 12 Aug 2011 19:03:34 GMT
On Fri, Aug 12, 2011 at 11:47 AM, Ahmed Radwan <ahmed@apache.org> wrote:
> Thanks Arvind,
> I prefer RTC with timeout. If we decide on timeout, what is the suitable
> timeout period and how can we manage it for different patches (some patches
> may require more time than others for review)? Is the timeout measured from
> review submission or from last activity on the review? Any ideas?

Good point Ahmed. I don't think there is any one scheme that will
address all such concerns. Given that someone may view a change as
trivial but it may be very complex for someone else, I think that the
idea of a set timeout will not be fair in all cases.

Another option to consider is to be like Hive project, where a
committer's patch must be +1'd by another committer who actually
commits it. For patches coming from non-committers, any committer who
reviews it can commit it.


> Best Regards
> Ahmed
> On Fri, Aug 12, 2011 at 10:59 AM, Arvind Prabhakar <arvind@apache.org>wrote:
>> Hi All,
>> We are starting to see some traction in JIRA and patch activity. I
>> believe now is a good time for us to put a formal policy in place that
>> guides the overall review and commit process. Different projects have
>> adopted different ways of addressing this, but at a high level there
>> are two - Review-Then-Commit (RTC) style, and Commit-Then-Review
>> (CTR).
>> Lets discuss this to bring out various point of views and then do a
>> formal vote on the candidate policy that is acceptable to the
>> majority.
>> My thoughts: I prefer RTC with timeout provisions. Specifically, I
>> feel that every change must get reviewed and if the reviewers do not
>> respond within a certain time, the change can be committed.
>> Please share your thoughts, comments and concerns on this.
>> Thanks,
>> Arvind

View raw message