On Tue, Jan 24, 2023 at 1:11 AM Jeff Jirsa <jji...@gmail.com> wrote: > .... But it's not merge-than-review, because they've already been > reviewed, before being merged to the feature branch, by committers > (actually PMC members)? > > There's no question that the feature branch already meets the minimum requirement for review. If we look at https://cwiki.apache.org/confluence/display/CASSANDRA/Cassandra+Project+Governance that would be point #3 in the process for code contributions.
On the other hand it seems clear that there is still active discussion about the code (including at least one change request: better commented code) and at least one committer has requested reasonable time to review. These would be points #4 and #5 in the governance process, so clearly we are in the state that code "must not be committed". Presumably it makes sense to review only after the code has been better documented, so the reasonable time might not have started yet? This is just my understanding of the email discussion, I'm not participating in the review myself. I can also attest that at least for Jacek and Mick, as well as Ekaterina if she participates, reviewing Accord will be the top priority of their day job, so I believe the "reasonable time" criteria is clearly met in this case. If we step back from the process a bit, the above could also be summarized as: It's not reasonable to expect that every committer would or should track work happening in every feature branch. It's only natural that there will need to be time for review at the point of merging to trunk. PS: Personally I don't really believe in commit-then-merge. If the author of the patch is committed to respond to review comments with high priority, it shouldn't make a difference to them whether the code is committed before or after the review. And of course if they aren't committed to work with the reviewer with high priority, then what's the point of reviewing at all? Since a reviewer is already obligated to do their part in a reasonable time, it follows that merging after the review is the process where incentives are aligned, AND there's no downside to the patch author. You want code that's been written by one PMC member and reviewed by 2 other > PMC members to be put up for review by some random 4th party? For how long? > > Why is there a difference whether the reviewer is a committer or a PMC member? henrik -- Henrik Ingo c. +358 40 569 7354 w. www.datastax.com <https://www.facebook.com/datastax> <https://twitter.com/datastax> <https://www.linkedin.com/company/datastax/> <https://github.com/datastax/>
