Joost Jager
Joost Jager
Marked blocked in anticipation of anchor output commitment format developments. Anchor outputs also have a value that is possibly taken out of a reserve.
I don't know. Also, maybe B can do this to any node, not just the ones that they have a channel with.
Didn't actually wait for that, but very likely that it is still bounded.
>Is there perhaps any unconsidered complexity for situations where multiple HTLCs exist in the commitment transaction for a single channel? There should not be such complexity. The code as is...
Waiting for design ack for the direction set out in this PR. In particular storing the settled htlc ids in a new bucket `settled-htlcs` and doing so atomically in the...
> I think this patch shouldn't use the shortchanid, and should instead use the funding outpoint - we shouldn't add more code that needs reorg handling. Are reorgs even supported...
> No, my point is that (I think) eventually we'll be moving away from SCIDs to identify a channel, so it may not make sense to make another place where...
@Crypt-iQ if you talk about the funding outpoint, do you mean txid:index, or the "compressed" 32-byte format that xors the index into the txid? I don't think we use the...
Another thing to be aware of is that we need to keep the mapping around for closed channels too. For example when you `LookupInvoice` and want to get the final...
Did a simple benchmark to check the performance implications of saving the final htlc outcome to disk. * Alice and Bob, single channel * Call `lncli sendpayment` 20 times to...