I would appreciate if we separated fixes and tests on Pull Requests.
A lot of times i will revert the fixes to validate if the test is good (if it fails without a fix) and how it failed. (not that I don't trust the committer, just part of the validation as sometimes I want to see what was the semantic change and fix). I may eventually play a better fix in the process.. and I am sure that would apply to anyone else helping on reviewing commits. I had at some point gone back in history and needed to apply the test without a fix to find a better fix. I know eventually it's not possible to separate these.. but if you could as much as possible separate them:? I recently did that into PR #2004... https://github.com/apache/activemq-artemis/commit/a72046a0e32fd47cad988a8d71512927f74c8585 https://github.com/apache/activemq-artemis/commit/a72046a0e32fd47cad988a8d71512927f74c8585 I may update the hacking guide with this.. WDYT? -- Clebert Suconic
