Small bug in last mail :) 2013/3/22 Sebastian Schaffert <[email protected]>
> Hi Fabian and Jakob, > > as Jakob says, it is necessary to have the release candidate version and > tag pushed upstream, because otherwise other people are not able to check > the release. > > Since I also did this revert already twice, here is the process (it is > simple and harmless): > - remove the release tag from GIT (this is possible even for a remote > repository) > see http://nathanhoad.net/how-to-delete-a-remote-git-tag > - update the version number in all POM files back to the previous snapshot > version with the following command: > > mvn --batch-mode > release:update-versions -DdevelopmentVersion=3.1.0-incubating-SNAPSHOT > -DautoVersionSubmodules=true > sorry, that should be mvn --batch-mode release:update-versions -DdevelopmentVersion=3.0.0-incubating-SNAPSHOT -DautoVersionSubmodules=true i.e. the previous version :) > > - commit again and push > > The advantage of this approach is that in this case the whole process is > documented in the GIT repository. Reverting in GIT is IMO only necessary in > case the release process itself fails (which can happen e.g. if you enter a > wrong password or some other kind of system failure). > > ;-) > > Greetings, > > Sebastian > > > 2013/3/22 Fabian Christ <[email protected]> > >> 2013/3/22 Jakob Frank <[email protected]>: >> > Not sure whether this is possible: >> > >> > As Andy said [1] one of the checks to do before voting should be: >> >> if there is a tag in the SCM, does it contain reproduceable sources? >> >> So is Step 6 for publishing the tag or for publishing the increased >> development versions of the artifacts - or both? >> >> >> -- >> Fabian >> http://twitter.com/fctwitt >> > >
