I do not think we should have done a force push to rewrite two commit messages. The problematic messages had commit hashes in their names, so it was easy enough to figure out what went wrong and how to find the original commit. If it was really important to fix the messages, I think we should have reverted the problematic commits and re-done the cherry pick with the desired commit messages.
I also do not think we should allow force pushing to release branches or to master branch in the future. In my opinion, the git history should be immutable. Thanks, Michael On Fri, Nov 25, 2022 at 11:39 PM 丛搏 <bog...@apache.org> wrote: > > +3 binding(Penghui Li, JiWei guo, Hang Chen) +2 nonbing(Zike Yang, > Yubiao Feng), so I will close this discussion and merge > https://github.com/apache/pulsar/pull/18623. thanks for your votes. > > Thanks, > bo > > 丛搏 <congbobo...@gmail.com> 于2022年11月25日周五 18:55写道： > > > > > > > > One thing I'd like to make clear: this is a temporary solution, and we'll > > > re-enable branch protection after the actions are performed (or relax to > > > 2.9.4 released). > > yes, after this pr: https://github.com/apache/pulsar/pull/18623 merged > > and reset branch-2.9 to commit 5ed247de3a, I will push a new PR to the > > master branch to open the branch-2.9 force push > > > > > > If it's the case, please file an issue and assign yourself to revert on > > > time :) > > if no one is against it in 24 hours, this discussion will be approved > > then merge https://github.com/apache/pulsar/pull/18623 then reset > > branch-2.9 to commit 5ed247de3a
