Update contributing guidelines?
Created by: rhfogh
Actually reading the contributing guidelines I was surprised by the contents.
- Reviewers are apparently required to test the PR before approving.
- It is all organised around reviewers getting requested. They cannot be bypassed, so a missing review will hold everything up. Once the last reviewer has spoken, he should merge the PR, leaving no time for third parties to contribute.
On 1) a lot of PRs cannot be tested, except at the right beamline. Many of my PRs cannot be tested unless you have global phasing installed. Also, running a test is a lot more demanding than looking at the code, particularly if you are supposed to exercise all the various code paths. I would be more than happy if the submitter of the PR had tested it, preferably both on his own beamline and in mock mode.
I submit that the way we currently work does not match the contributing guidelines. Being rather literal-minded, I would prefer if we adapted the guidelines to the way we work, or vice versa.