Ah, okay. Personally, it seems to me the real problem is in the documentation's claim that `reload_delta` creates separation between "RPC reloads", vs. no clear conception of what an "RPC reload" is. In thinking about it generally, I would expect such a reload to encompass both tables.