For those following the Mailing list, you would have seen the discussion about moving to Github. This is mainly due to the fact that google code will be shutting down.
Read on for some of the changes/challenges
I have already migrated the code, issues and google wiki over to github, which you can visit at http://www.github.com/OpenZWave/ but this change also means migrating from Subversion to Git.
Some things that have changed:
Obviously you need to switch to git if you like living on the bleeding edge of OZW development. Or you can use the snapshots to download source or binary packages of the most recent versions. I’ve updated the old google SVN repository with a file saying we moved to github, so those that don’t follow our mailing list would see it changed when they try to update next time.
The change to git has also resulted in a few breakages, mainly around our version numbers, and some scripts we use to help package up OZW. This should be fixed now in the current version, but only time will tell. As a result of the change, the version numbers have taken a step back. Where as previously we would do our versioning as X.Y.SVN where SVN was the svn revision number, now we do X.Y.GIT where git is the number of commits since the most recent "tag", in this case, 1.3. So our version numbers have jumped back from 1.3.1027 to 1.3.103 (currently). This might cause problems for people using the binary packages, but at least our versioning is automatic still.
Also, some of the backend infrastructure is affected as well. Mainly our CI and Build Tools. Again, these should be working correctly again now, but only time will tell.
As for Development flow, we will mimick what we have been doing on SVN, that is, most updates will go into the master branch directly. After we release 1.4, we might discuss switching to a github flow style of development.
One thing I am disapointed with is the Github issue tracker. It lacks a feature we used extensively, namely the ability to attach files to issues. As you know, we often as for Log Files etc when trying to diagnose issues. For now, I think we will ask users to upload Logs to the Log Analyser here and reference it in the issue tracker. In future, I might see how we can integrate with the GitHub API to automate some of that.
Overall, I see this as a postive move although unexpected. Hopefully the "social development" thats so ripe on Github (Forking, Pull Requests etc) will mean a bit more contributions from other developers and help us speed up development a bit.
If you do encounter any problems, please raise them on the mailing list or issue tracker on github.