Daniel Kahn Gillmor <dkg@fifthhorseman.net> writes: > I don't think this reasoning is sensible. If the entire thread is > deleted, and a new message comes in, it should *not* get the same mesage > ID. ghosts should only exist in the database when other messages point > to them. > > So i'd be fine with killing this entire last test, unless someone can > propose a good reason to keep it. I think I buy your reasoning, but I'd be happy if Austin (who introduced ghost messages, and wrote that test, could comment. d