Author: Rusty Russell 2019-03-20 00:22:05
Published on: 2019-03-20T00:22:05+00:00
In a discussion about publishing to the blockchain, Anthony Towns suggests that there is still some risk with eltoo channels as it allows participants to choose the CSV delay, which could result in non-cooperative closure of a channel. He then goes on to question the differentiation between malicious and non-malicious double-spends and how it would affect handling coins from a trusted source. The discussion then moves on to the potential problems caused by NOINPUT signatures, with one of the possible nightmare scenarios being that naive users do silly things with the signature and end up losing funds due to replays. However, Rusty does not find this scenario credible. Although he is inclined to err on the side of putting more safety measures in for NOINPUT, he believes that these measures should be practical and helpful in failure scenarios. Finally, he suggests making "must have a sig that commits to the input tx" a standardness rule rather than consensus if it is deemed necessary.
Updated on: 2023-06-02T17:51:59.386254+00:00