On Saturday, October 01, 2016 4:01:04 AM Rusty Russell wrote:
> Prefer a three-arg version (gbits-to-compare, blocknum, hash):
> - If <bits> is 0 or > 256, invalid.
> - If the hash length is not (<bits> + 7) / 8, invalid.

This means zero padding on-chain, which would be undesirable.
Rather "at most" and have the consensus implementation do the padding.

> - If the hash unused bits are not 0, invalid.

Why?

> - Otherwise <bits> of hash is compared to lower <bits> of blockhash.

Lower in what endian? Why only that endian? Why only lower? I can see a 
possible use case where one wants to look at only the high bits to ensure 
their transaction is only valid in a block with at least a certain 
difficulty...

> This version also lets you play gambling games on-chain!
> 
> Or maybe I've just put another nail in CBAH's coffin?

Or maybe resurrected it...

Luke
_______________________________________________
bitcoin-dev mailing list
[email protected]
https://lists.linuxfoundation.org/mailman/listinfo/bitcoin-dev

Reply via email to