@me no, it's not that hard at all. Currently the relay interface sends "fat pings" -- it includes the full content object. If it instead used "thin pings", with just the URL of the content, the relay and all the downstream servers have to fetch the content from the source server. That lets the source server enforce server blocks, including user server blocks.
@evan To be honest, I haven't looked that deeply into the protocol. I'm sure what you're telling me is true. After all, you literally wrote the book on ActivityPub. Also, this is perhaps not exactly what you're talking about, but as soon as a post federates to another server (because the poster has a legit follower there) the originating server is no longer in exclusive control of the content of that post, is it not?
Unless I'm fundamentally misunderstanding something here—which I'll admit is a possibility—a single misbehaving server could circumvent this entire system, could it not? Obviously, there's the option of defederating such a server, but you'd have to know it's happening first. It's not just about trusting your own server, but also every server you federate your posts to.
Am I way off base here? If so, I'd love to understand how.