Author: ZmnSCPxj 2019-03-20 08:07:00
Published on: 2019-03-20T08:07:00+00:00
In a Lightning-dev mailing list, ZmnSCPxj raised a concern regarding the implementation of eltoo watchtowers. He questioned how the current plan for eltoo watchtowers would be able to create a non-SIGHASH_NOINPUT signature if one of the signatures must be non-`SIGHASH_NOINPUT`. Rusty Russell suggested that the proposal of "must have a non-SIGHASH_NOINPUT" rule might address the reuse scenario and can be a standardness rule instead of consensus. ZmnSCPxj also proposed an alternative method where every script is required to have a valid signature that commits to the input instead of using SIGHASH_NOINPUT. He then discussed a possible eltoo construction consisting of a setup transaction and update transactions that pays out to OP_IF and OP_ELSE scripts. The `SIGHASH_NOINPUT` signature for `muSig(A_u,B_u)` would unlock the setup transaction or any update transaction with lower `nLockTime`. The watchtower would generate the signature for `Q`, committing to a particular UTXO.
Updated on: 2023-06-02T17:52:59.690675+00:00