Skip to content

2077 Followups #2360

Closed
Closed

Description

@dunxen

There are a bunch of followups to #2077 that need to be included before 116.

See: #2077 (review)

Reproduced below:

  • close_channel_internal should fallback to force-closure if the channel is in the inbound/outbound sets, though I'm not sure its super critical,
  • update_partial_channel_config should probably be updated to work on any channel
  • internal_shutdown should FC any pending channels
  • do_chain_event needs to call best_block_updated even on unfunded channels so we time out channels if they aren't funded in 2 weeks
  • peer_connected tries to remove outbound unfunded channels on reconnect but doesn't look at the new outbound map (which, now that I look at it, should be generating a ChannelClosed event, but isn't...ugh we need to consolidate the channel close codepaths)
  • Rename get_funding_outbound_created to get_funding_created

Activity

  1. added this to the 0.0.116 milestone on Jun 16, 2023
  2. dunxen commented on Jun 26, 2023

    @dunxen
    ContributorAuthor

    @TheBlueMatt So for this feedback:

    do_chain_event needs to call best_block_updated even on unfunded channels so we time out channels if they aren't funded in 2 weeks

    It seems and I feel that the timeout is distinct in the sense that it's for established, but not yet confirmed inbound channels which seems right. (i.e. specifically waiting for funding to confirm) These are necessarily already Channels.

    Question: If we've always just relied on this as a timeout for receiving a funding_created as well? I feel like that should be handled differently and probably have a shorter timeout honestly. As we're not waiting on a funding tx to confirm, we're actually waiting on our peer to just send a message.

  3. wpaulino commented on Jun 27, 2023

    @wpaulino
    Contributor

    Question: If we've always just relied on this as a timeout for receiving a funding_created as well? I feel like that should be handled differently and probably have a shorter timeout honestly. As we're not waiting on a funding tx to confirm, we're actually waiting on our peer to just send a message.

    Agreed, I don't see any reason to not impose a shorter deadline then. Mostly we just want to revert back to the behavior we had previously, which is giving up after some time of still being unconfirmed, regardless of receiving funding_created or not.

  4. TheBlueMatt commented on Jun 27, 2023

    @TheBlueMatt
    Collaborator

    I'm happy with either one - it does seem like that could totally be a shorter deadline (taking two weeks to build the funding tx at all is absurd), but also fine with going back to what it was and just doing it based on block height.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Type

    No type

    Projects

    No projects

      Milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions