On Fri, Jul 22, 2022 at 11:21:39AM +0200, Omar Polo wrote:
> here's the updated diff, but at this point i'm even fine with dropping
> it.

Thinking about this some more: How about we update the author timestamp
in case author name+email == committer name+email, and keep the existing
author timestamp otherwise?

This way, we would stop updating the author timestamp only once someone
else "takes over" a commit object. This is not foolproof, of course (a
given person could be using multiple email addresses). But a commit with
equivalent author and committer name+email, and different timestamps
for each, would look a bit strange.