Mildest of hot takes: just like clicking ‘Update branch’ does not dismiss a GitHub PR approval, applying a suggestion made by a reviewer should not dismiss that same reviewer’s approval.
— Philippe Serhal (@philippeserhal.com) on Bluesky
When I started contributing to npmx.dev, it surprised me that pushing commits to my PR after it already had been reviewed, did not dismiss that very review.
I always assumed dismissing it was the safer default.
But why should updating the PR against the main branch dismiss the review? Why should implementing the suggestions by the reviewer dismiss the review? Those questions made me realize that I had a fundamental misunderstanding of the review process.
Forcing your maintainers to re-review on every change will only train them to skim through the changes. They will potentially miss important details and ultimately it’ll have the opposite effect of what you intended.
A trusted maintainer merges the PR either way. If the initial review was thorough and the follow-up changes are minor, there’s no reason to restart the process from scratch.