I might be over thinking this, but I'm trying to figure out the best way to build an unspent output pool. Should I just naively parse every block? Every input gets removed from the hash set, and every output gets added back in? Seems terribly inefficient. The thing is it doesn't seem like there is anyway to accurately build this utxo pool without traversing the block chain at least once and constructing it. Is there a well defined short-cut I should be aware of? Pointers welcome.
I guess I can iterate through the on-disk blocks first, and then update as blocks come in.