Closed garyyu closed 6 years ago
This is a good find, it should be as you say... so at the moment it's only putting the first 8 bytes of the key into the sighash whereas it should be writing the full 32 bytes.
However, changing this here will break consensus since all transactions thus far have been using the incorrect hash function. We'll need to hard fork before we can include this.
secp256k1_compute_sighash_single()
function internally has a wrong sizeof() usage on a pointer which will be always 4 or 8.