Hey all, I know everyone is busy, but I’d love to get some form of 2.0 release out before the holidays.
Here’s a proposal: ## Preface 2.0 is in decent shape, but we know a few things are missing*. We also know that unless we give this to a wider audience, we’ll never find all of the issues. * https://issues.apache.org/jira/browse/COUCHDB-2876?jql=project%20%3D%20COUCHDB%20AND%20priority%20%3D%20Blocker%20AND%20resolution%20%3D%20Unresolved%20ORDER%20BY%20key%20DESC + the ones rnewson mentioned in his previous mail. ## The Plan 1. Find someone who’s willing to run the release process. - At this point this should be along the lines of: 1. Build CouchDB locally. 2. Run `make dist`. 3. Upload resulting tarball somewhere. (I have this semi-automated on my old Jenkins install, maybe this is enough for the time being.) 2. Announce the tarball as “Apache CouchDB 2.0 alpha-git-hash”** in these places: - here on dev@ - user@ - on the website - weekly news - social media Along with the tarball, we explain where people can report issues and get help (user@, dev@ or IRC) 3. fix issues as people report them Repeat. Once we have addressed all known issues, we start calling the release “Apache CouchDB 2.0 beta”**. During the beta period, we should start talking to client library maintainers and OS package managers, to start the work required to support 2.0, if they haven’t done this already. ** Note: these are just tarballs we put out, they are not official Apache releases. Those need a vote and everything. That comes after the betas. Once we are satisfied with the betas, we should start doing proper, voted-on Apache releases for “Apache CouchDB 2.0 release candidate [1-N]” and get widest possible testing for those and eventually designate one of them as the final release. * * * How does this sound? Am I missing something? Is this too simple, or too complicated? Thanks for your feedback, bikesheds welcome! Best Jan --
