Rajani, we can't cut a new master from 4.4
Leo's work on comparing the branches showed us that. So the new flow will
be limited To master land 4.5+

biligual spelling checker used.read at your own risk
Op 31 jul. 2014 06:22 schreef "Rajani Karuturi" <rajani.karut...@citrix.com
>:

> For the git flow:
> 1. We agreed to follow git-flow explained here
> http://nvie.com/posts/a-successful-git-branching-model/
> 2. This is the proposal for first cut
> 2a. rename 'master' to 'develop’
> 2b. branch a new 'master' from ‘4.4’ and update tags with release/4.4.0
> 2c. branch ‘release/4.5' from the develop
> 2d. merge ‘release/4.5' to master once the release voting is done.
> 3. This would be the flow for a hot fix
> 3a. branch off from the release tag on master. in this case it would be
> release/4.4.0
> 3b. commit the fixes in hotfix/4.4.1
> 3c. do the release
> 3d. merge to develop
> 3e. merge to master and update tags
> 3f. delete hot fix branch
> 4. for any LTS release create a support branch when required using
> git-flow support
> 4a. http://stackoverflow.com/a/16866118/201514
>
> using the git-flow git extension available at
> https://github.com/nvie/gitflow can reduce the number of commands/errors
>
> In addition:
> 1. Every commit should have unit tests
> 2. every feature/merge request should have unit and marvin integration
> tests
> 3. A commit should not have check style or find bugs issues
> 4. any coverity issues reported in the new code should be addressed
> immediately
> 5. every developer should run the BVT on the simulator before doing a
> checkin (
> https://cwiki.apache.org/confluence/display/CLOUDSTACK/Validating+check-ins+for+your+local+changes%2C+using+Simulator
> <
> https://cwiki.apache.org/confluence/display/CLOUDSTACK/Validating+check-ins+for+your+local+changes,+using+Simulator
> >)
>  (This I am not very sure. May be we should let jenkins handle it and
> report integration failures if any?)
> Please add/amend if I missed anything.
>
> Can we call for a vote on this and freeze this without further delay?
>
>
> ~Rajani
>
>
>
>

Reply via email to