I'd also like to know how to do this. The current approaches seem, afaict, to involve making on both end of the connection. Even given the inherently racy nature of the issue, that seems unwieldy to me.
Let's say I've got a scenario where I'm doing a pg_dump replication rather than online streaming, e.g. due to air-gap or whatever.
Is there a scriptable way to validate the restore ? e.g. using doing something clever with ctid or something to ensure both the schema and all its rows were restored to the same point at which the dump was taken ?