They were saving computational horsepower. But their cashed values had a sort of architectural ambiguity and could be faked out. Once the attacker learned how to take advantage of the cache he was able to bypass the range proof and enter a negative value for LBTC.

Replies (1)

From my knowledge there is not much to cache? The range-proof would be TX specific anywayz The only caee it might benefit is mempool / block validation; which is so dumb for a network that could proceed all txs on a toaster
↑