This is a proposed way of working, under discussion on the mailing list. See http://markmail.org/message/xoyjw2sduenlytwm
Commit to master through PR only was decided on here: http://comments.gmane.org/gmane.comp.apache.cloudstack.devel/62223
Apache CloudStack is a broad and encompassing project covering many areas that require in-depth technical knowledge. As such, it is not possible for one person to know and/or test all components of CloudStack in depth. Miscommunications and misunderstandings between developers can lead to poor quality in CloudStack's code and cause community to deliver a unstable release. Knowledge lost from one release to the next can cause regressions in the current release. This page details the development process the CloudStack community must follow in order to ensure the community delivers high quality releases. This page is a must read for all developers and testers.
This process is developed with the following principles.
The development infrastructure is setup as follows.
The process is very simple.
The BVT will be run on master and the current release branch on a continuous basis. If the BVT fails, the commits submitted between the last successful BVT run and the failed BVT runs are reverted. The developers who submitted the commits are notified of the revert and can merge their changes again when they have figured out the problem on their own branch.
As a code contributor to Apache CloudStack, regardless of if you have committer or contributor status, your responsibilities are as follows:
Communicating what issues you're working on and how you're working on it is an important part of participating in the community. We do that on the mailing list. However, you should use JIRA to update issue status and add comments with conclusions drawn from mailing list discussions. This allows release managers, users, and testers to follow the issue better, to understand if the issue is being worked on, to get a better feel of whether an issue will be resolved in time for the release.
Automated tests ensures that your feature will not be regressed. Anyone who submits a code change that breaks an existing automated test is responsible for fixing that automated test or revert the code change.
Q: I'm a contributor and not a committer. How do I participate in this process?
A: You can send a pull request on github. When two other collaborators give a LGTM, a committer will merge your pull request.
Q: Who is providing this jenkins?
A: Currently, the jenkins instance will be provided by Citrix, mainly because we're still working out an infrastructure testing lab in ASF. When that is ready, Citrix is willing to donate the setup necessary to run continuous integration to ASF.
Q: I like to set this up in my own organization so that we can run CI on our own development branches. Can I do this?
A: The setup itself is publicly available. The people who made this possible: Amogh, Bharat, Santhosh, are all on the mailing list. You can communicate with them via the mailing list and ask them help you with the setup. We will try to document this setup as much as possible once it's available.
Q: I currently have a Jenkins slave that runs tests for my own code on every build. Can you add my slave to your jenkins?
A: Currently, no. There's a lot of issues involving jenkins in Citrix making call to outside of Citrix that really I'm not willing to go resolve. For now, you just have to run that with the ASF jenkins.