if there's going to be more than one experimental downstream, I'd
prefer we start voting on release artifacts (say, a tarball of our
source checkout) so that we can get a release cadence going. I'd be
fine with doing this while e.g. Hadoop's adoption is being driven in a
branch.

For getting Hadoop working this weekend, I think a working
commit-without-review branch is reasonable. I'd like there to be a
yetus jira that tracks the branch's lifecycle, with the branch named
for it.

On Thu, Oct 15, 2015 at 10:18 PM, Stack <[email protected]> wrote:
> Use hbase as your guinea pig?
> St.Ack
>
> On Thu, Oct 15, 2015 at 6:30 PM, Allen Wittenauer <[email protected]> wrote:
>
>>
>> On Oct 15, 2015, at 6:28 PM, Sean Busbey <[email protected]> wrote:
>>
>> > I presume the concern for just e.g. tagging master is turn around time
>> > for reviewing changes needed?
>>
>>
>> Yeah.  I'd much rather move forward if the fixes are simple than have to
>> back out (unless absolutely necessary, of course).

Reply via email to