Should not unapprove the reviewer when he update the PR himself

[Julien Herr]

No description provided.

[Robert Noack]

Personally I disagree. In my opinion the PR is a nice way to view the diff, and there are certainly times when I made a commit only to realize I messed something up and need to make another update. Of course it's not supposed to happen - but that is the point of the approval process.

[Bryan Turner]

Whether a reviewer updated the pull request or any other user, if the update changes the effective diff and what would be merged, reviewers (including the one that made the change) should still verify those changes before merging.