Author: Rusty Russell 2016-10-13 19:15:36
Published on: 2016-10-13T19:15:36+00:00
In a mailing list post, Rusty is seeking feedback on writing the specification for the propose/commitsig/revoke cycle. Changes to the commitsig message have been agreed upon, which include an additional field indicating updates included from the receiver. This has two implications that Rusty had not previously considered. Firstly, the state of updates is no longer linear, and forwarding is triggered on the final state as both sides are fully committed. Secondly, fail-fast cannot be implemented if the receiver pushes uncommitted changes into the sender's commit transaction, rendering the scheme ineffective. Rusty is struggling to write the spec clearly with these additions and is looking for thoughts and feedback.
Updated on: 2023-05-24T00:37:35.790677+00:00