How about letting developers ack PRs?

These PRs can go in a separate queue for maintainers and will definitely not contain trivial errors but may contain design changes. Gitmate can have one more state if PRs are acked by a developer, somewhat like pseudo passed but still not approved for merge.

The more developers ack a PR, the PR can get to one side of the queue, that means it has more probability of being error free. So if someone is not quite in the mood to review, he can start from this end and keep merging.

While those PRs which are acked by lesser developers or (some requested changes and some acked) should go to the other end, as a maintainer will have to look closely and then see why is the PR being controversial.

Start by reading the discussion on issue #112, including the comments about developer acknowledgements, queue ordering, and maintainer review. Done would require an agreed design for the acknowledgement state and queue behavior, followed by the corresponding implementation or documented decision.

Written by the indexing model from the issue text.