Lightning Specification Meeting 2023/02/27

[t-bast]

Lightning Specification Meeting 2023/02/27

[Roasbeef]

route blinding:

• lnd has interop now in a PR, not yet reviewed
• main PR has been squashed down, ppl to do final pass overs now

onion messaging:

• nearly universal interop?
• questions re how ppl handle the state machine of the non-happy path:
  • fall back if someone critical on the route doesn't support
    • ideally direct connect from there then
  • what about timeouts?
    • not direct signal if you send and packets get dropped, or timedout, etc
    • will need co properly test/control to ensure UX is up to par w/ other solutions

fat errors:

• there, in a holding pattern
• someone from eclair making prototype

splicing:

• eclair working on splicing stuff, made a lot of progress
  • things get hairy when you combine this w/ zero conf....IYKYK (when you get there)
• msat rounding
  • issue where one party needs to round up and the other needs to round down
  • current protocol uses the delta in msat, which then needs to work out properly on chain, or you risk bleeding out a sat or two on chain
• can the msat issue be avoided by just not sending the final amount?
  • just send the inputs+outputs and then let the rest be implicit
  • or you just need to express the fee or fee rate
• will continue to brainstorm what other issue might pop up, etc

taproot channels:

• lnd working thru PR review, in a holding pattern re interop
• secp-zkp PR now approved: BlockstreamResearch/secp256k1-zkp#211
  • others can be unblocked now re giving feedback re APIs, etc

dyn commit

• lnd wanting to revisit, as a way to upgrade ppl to taproot channels, and also maybe across versions if something changes in the future
• CL wanting to revisit as well, since wanting to move ppl over to zero fee anchors
  • also have something custom already for static remote key
• revisiting question of the overlap with splicing
  • eventually splicing has a channel type on it