reattempt connect if failed during init connect #729
Conversation
| PeerConnection::connect_brontozaur(local_node, remote_node) | ||
| .expect("Unable to connect to the remote peer") | ||
|
|
||
| let mut connection = PeerConnection::connect_brontozaur(local_node, remote_node); |
There was a problem hiding this comment.
This is a hard idea to realize. The way the patch is right now, the user will never understand what is going on when the connection is failing in the cli, or a future gui, and has to look at the logs for diagnostics. If you fail seconds later, the farcaster daemon will never report the failure either and expect the connection to come up eventually, because it expects the connection to be up within 1 second. If you move this to the farcaster call site, you add further dependence on the delay introduced for waiting on the connection.
I think the right thing to do here is refactor peerd to launch its internal zmq runtime no matter what and independently manage its remote session from within its runtime. Then it can send errors directly to farcasterd.
|
Closing this in favor of #792 - a failed to connect is now cleanly aborts taking a swap with an error message to the user. |
resolves #728 - reuses logic in https://github.com/farcaster-project/farcaster-node/blob/main/src/peerd/runtime.rs#L432-L449