Submitting new features to Nova

I just wrote down a few pieces advice for someone submitting a large feature patch to Nova, so I thought I’d re-post it here:

  • Think about what it is like to be a nova-core reviewer looking at a list of 40 to 60 reviews and having maybe 2 hours today to do reviews. Think about how to make it more likely that such a reviewer will choose your code to review.
  • Small patches are easier to consume. The smaller you make the patch, the more likely it is that it will get reviewed quickly.
  • Break your changes into a series of small, self-contained changes.
  • The earlier in the release cycle you can begin submitting some of your changes the better. Don’t wait until all of your changes are finished before submitting.
  • Do as much of your design discussion in the open, preferably on the openstack-dev mailing list. If you have discussions on the phone or IRC, try and post a summary of those discussions to the
    mailing list.
  • Holding a design summit session to discuss your changes in advance is a great idea, but don’t assume that everybody who may have an opinion on your changes is present. Also, bear in mind that someone’s quick opinion offered at a design summit session may be very different from their considered opinion after reviewing the code in detail.
  • Finally, try and participate in the project beyond just making the changes you need. Review other peoples’ changes in gerrit and offer your opinions, participate in design discussions on the mailing lists, fix bugs you come across, triage bug reports, etc. etc. All of this will allow other developers to get to know you, trust your judgement and review your changes more quickly. You will also learn more by interacting with other developers.

3 Responses to “Submitting new features to Nova”

  1. Daniel Berrange says:

    Your point “Break your changes into a series of small, self-contained changes.” deserves a much bigger focus – it is perhaps the most important aspect when it comes to getting effective code review. For more info see

    http://wiki.openstack.org/GitCommitMessages#Structural_split_of_changes

  2. markmc says:

    Thanks Dan, I added your link

  3. [...] Submitting new features to Nova [...]

    [WORDPRESS HASHCASH] The comment’s server IP (174.143.11.196) doesn’t match the comment’s URL host IP (50.56.129.2) and so is spam.