2015-05-20 17:50 GMT+02:00 Andrew Bayer <[email protected]>: > Yeah, any commit needs to be done by an actual committer. What I'd advise > for this release is that we do the TeamCity release build against a > personal fork of the repo, where TeamCity has creds to push etc, and then > the release manager pushes the commits/branches/tags to the canonical repo > manually. It's a little awkward, but meets the requirements for ASF > policies.
I see, this is probably the best thing to do at least for the first release. > > Can I get access to the TeamCity setup and a walk through of the process? > For TeamCity I will send you an email with credentials. Release process is described here [1] and here [2]. For [1] we should probably backup this because it is going to be shutdown anytime soon. We can back it up through a PR to [3]. > I'd like to start working on how I can (wearing my Infra hat here) figure > out how to make this process smoother and better integrated with Apache > processes going forward - and admittedly, how I can port it to Jenkins, > given my personal biases. =) I am also interested in seeing how the Groovy > release process may be able to provide a model for other ASF projects > looking for a more automated release build process. > > A. > > [1] http://docs.codehaus.org/display/GROOVY/Release+Process [2] http://groovy-lang.org/wiki/groovy-release.html [3] https://github.com/groovy/groovy-website/tree/master/site/src/site/wiki
