Rusty Russell

Results 308 comments of Rusty Russell

Merged in that patch, testing now. Ack 368dcc3a9a3d5be28227d5f191b6621b1cc30a6e

You're actually only allowed a single connection to each peer, so we need to pick one in this case. We'll fall back if one doesn't work at all, but if...

> Yep, fixed: this is actually a quote from the route-blinding branch only.

Trivial rebase on master, those simple typo fixes, and added (as first two commits) a new API system for hsmd_wire.csv

Ack https://github.com/ElementsProject/lightning/pull/5646/commits/4c25d7c0d9e638df18962c8a6bb0b795e64b0ae1 Since @cdecker approved and I need to get moving on the next PR :)

This is a regression, I had not appreciated that we would now treat names as literals to announce. Until DNS-name address records become widely supported, we should always turn it...

Yes, this is my understanding too. If you ever end up constructing a commitment tx which has more than max_accepted_htlcs, your peer has violated the constraint. Thus, you *could* choose...

> Yeah so I guess the spec doesn't specifically mention that. > In lnd the following is not allowed (but is technically valid): > > ``` > 6. ---add---> (max_accepted_htlcs...

Actually, I was wrong: we have that language for update_fee (...`but MAY delay this check until the `update_fee` is committed.`) but the other checks are immediate: ``` ### Adding an...

Main points from discussion: 1. This is a premature optimization. 2. It's not something we want to encourage, as it's antisocial (such a node isn't useful for other nodes). 3....