"Re-request review" should dismiss the reviewer's approval
Proposal
If I re-request someone's review and they'd previously approved the MR, I'd expect their approval to be dismissed (as mentioned here: #345364 (comment 875691526)). Otherwise, why would I be re-requesting a review from them? Usually I'd use this when making a large change to the contents of the MR that I think constitutes requesting another round of review from my code reviewers, so having their approval stick around is really unintuitive.
Right now, with the way approvals and reviews work, the interactions between the two features feel very disconnected.
I'm aware that there's a setting for automatically dismissing approvals on any code push, but I think that's overkill as it'd also mean merging main
into my branch would trigger all the approvals to be dismissed, and that'd be very frustrating when trying to resolve merge conflicts so you could merge an MR.