I often find myself wishing for this block cloning thing, but with rsync hooks. Like suppose you're copying a large genome from A to B, and you already have a

I often find myself wishing for this block cloning thing, but with rsync hooks. Like suppose you're copying a large genome from A to B, and you already have a genome from that species (but different organism) lying around on B. Sure, you could clone and then rsync on top of the clone to avoid transferring the common bits again. But that's forethought that users often don't have. Better to use the rolling hash related metadata that rsync generates as a query into all possible targets and pick one automatically so that the user doesn't have to think about it and just sees it as a really fast copy.