My vote is that a patch can go into 0.8 if (1) it fixes a critical issue or
(2) the change is trivial and it makes the 0.8 experience better (e.g.,
improving log4j readability). kafka-946 may fall into (2).
On Tue, Jul 2, 2013 at 7:34 PM, Joe Stein <[EMAIL PROTECTED]> wrote:
> << How about for now: by default we will not incorporate fixes into 0.8
> unless there is a compelling argument (e.g., regression/clear bug with no
> good workaround) to do so.
> On Tue, Jul 2, 2013 at 8:48 PM, Joel Koshy <[EMAIL PROTECTED]> wrote:
> > Good question. Some fixes are clearly critical (e.g., consumer
> > deadlocks) that would impact everyone and need to go into 0.8.
> > Unfortunately the criticality of most other fixes is subjective and
> > I'm not sure how feasible it is to develop a global criteria. It
> > probably needs to be determined through consensus whether it needs to
> > go into 0.8 or not. How about for now: by default we will not
> > incorporate fixes into 0.8 unless there is a compelling argument
> > (e.g., regression/clear bug with no good workaround) to do so.
> > Joel
> > On Tue, Jul 2, 2013 at 5:11 PM, Jay Kreps <[EMAIL PROTECTED]> wrote:
> > > What should the criteria for fixes on 0.8 be? This seems like a
> > reasonable
> > > candidate but I don't think we discussed what we would be taking so I
> > > thought I would ask...
> > >
> > > https://issues.apache.org/jira/browse/KAFKA-946
> > >
> > > -Jay
> Joe Stein
> Twitter: @allthingshadoop <http://www.twitter.com/allthingshadoop>