Understood that you don't want any more API changes between now and 2.0. However, since there are already API changes in 1.7, what is the harm in adding more to finish the release. Why not come to this agreement for 1.8? Users / Integrators will already have to make changes for 1.7.
----- Original Message ----- From: "Sean Busbey" <[email protected]> To: "dev@accumulo apache. org" <[email protected]> Sent: Wednesday, December 3, 2014 10:52:48 AM Subject: Re: [VOTE] API release policy for 1.7/2.0 On Wed, Dec 3, 2014 at 9:45 AM, Josh Elser <[email protected]> wrote: > > And, for context, if this no-new-API freeze does stick, you would be > posthumously forcing us to go back and rip out a bunch of already committed > code for the replication feature and (possibly) the port to HTrace. That > would be pointless and a waste of time IMO. > > I think I've been consistent in saying this isn't a black-and-white issue. I know we've already made some additions. The branch that became 1.7 has been around a long time. While I'd love to revisit what changes we've added in that time, I'm not going to demand that that happen if I don't think I have the time to help make it happen. Since the replication feature was one of the things driving the desire to have 1.7, I think it qualifies as a substantial enough gain to justify the additions. Not to go off topic on this thread, but does the HTrace port touch the client API? I was under the impression that it was all internals. -- Sean
