Splice with htlcs

[remyers]

This PR builds on splices #430 to account for pending htlcs that may exist if the channel has first negotiated quiescence #558. This PR is modeled after Eclair PR # 2720.

The helper function makeCommitTxsWithoutHtlcs is renamed to makeCommitTxs and uses the set of committed htlcs that exist when InteractiveTxSigningSession is instantiated for a splice.

A new attribute htlcsAmount is added to InteractiveTxOutput.Shared to account for the value committed to htlcs from the shared funding input but not assigned to localAmount or remoteAmount.

This PR assumes the quiescence feature is mandatory rather than handle splices without exchanging stfu.

Updated tests for #421 with a pending splice with 5 instead of 6 htlcs maximum on each side.

[remyers]

@t-bast sorry for the delay in getting the rebased version. I'm getting an error with the liquidity ads lease backwards compatibility test where it compares the current state with the checkpoints on lines L119 and L138. When I review the difference with the checkpoint I notice what must be the additional discriminator and zero value added in 9aa3be5.

I don't think it makes sense here that the checkpoint should exactly match all future states. I believe we should remove these assert lines otherwise it will always fail when we change the state data structures, even if backwards compatibility is preserved.

[t-bast]

I don't think it makes sense here that the checkpoint should exactly match all future states. I believe we should remove these assert lines otherwise it will always fail when we change the state data structures, even if backwards compatibility is preserved.

Sure, modify what you need to modify to get the rebase working, and then we'll see during the review.

[t-bast]

Looking good, I improved the tests and found a small bug in b7b4364, which you can probably apply to your branch and then squash with the existing commit?

[remyers]

Looking good, I improved the tests and found a small bug in b7b4364, which you can probably apply to your branch and then squash with the existing commit?

Thanks for the review/fixes. I've squashed them into 699d52f.

[t-bast]

ACK 699d52f, let's wait to make sure liquidity ads is stable before merging this to master, but it's looking good to me 👍

[t-bast]

LGTM!