Hi Julian, Thanks for your feedback and suggestion. It would be probably my bad. What we want is, by thru some RC releases, the code base can be mature and the major 1.0.0 release then be out. I thought we should have used 'M'(milestone) instead of 'RC' for this purpose? Emmanuel did mention this idea some time before, but considering the change may cause messy we don't do it. I guess we can consider different release name strategy after 1.0.0 accordingly?
Regards, Kai -----Original Message----- From: Emmanuel Lécharny [mailto:elecha...@gmail.com] Sent: Tuesday, April 05, 2016 6:24 AM To: kerby@directory.apache.org Subject: Re: Release names Le 05/04/16 00:04, Julian Hyde a écrit : > Has anyone commented that on your release naming policy? Maybe you’ve got a > good reason for including “RC” in release names, but I wanted to point out > that they look strange to people. > > Josh Elser just included Kerby 1.0.0-RC2 in Apache Calcite’s Avatica > sub-project[1], and my first reaction was, “Why is he including a > release candidate as a dependency?” We have a policy of not depending > on release candidates or snapshots. In Apache, a release candidate has > not necessarily passed a vote, At The ASF *EVERY* single release passed a vote. Always. That of course includes RC. A Release Candidate is just a pre-GA version : we think we are almost done, but we would be sure not to miss obvious bugs before going to GA, just for the sake of providing a stable and production ready product. > so is not necessarily “clean” from an IP standpoint. They *are* clean. We don't release if it's not clean from an IP standpoint.