Use ChannelUnavailable for a peer disconnecting not MisuseError

This fixes a crash in the `full_stack_target` fuzz test (found by
Chaincode's generous fuzzing infrastructure!) but ultimately is a
better error code - a peer disconnecting before we can fund a
channel isn't a "misuse error" its an unavailable channel.

Author: TheBlueMatt

fuzz/src/full_stack.rs

@@ -632,7 +632,9 @@ pub fn do_test(data: &[u8], logger: &Arc<dyn Logger>) {
 						// It's possible the channel has been closed in the mean time, but any other
 						// failure may be a bug.
 						if let APIError::ChannelUnavailable { err } = e {
-							assert_eq!(err, "No such channel");
+							if !err.starts_with("Can't find a peer matching the passed counterparty node_id ") {
+								assert_eq!(err, "No such channel");
+							}
 						} else { panic!(); }
 					}
 					pending_funding_signatures.insert(funding_output, tx);

lightning/src/ln/channelmanager.rs

@@ -2552,7 +2552,7 @@ where
 		let per_peer_state = self.per_peer_state.read().unwrap();
 		let peer_state_mutex_opt = per_peer_state.get(counterparty_node_id);
 		if let None = peer_state_mutex_opt {
-			return Err(APIError::APIMisuseError { err: format!("Can't find a peer matching the passed counterparty node_id {}", counterparty_node_id) })
+			return Err(APIError::ChannelUnavailable { err: format!("Can't find a peer matching the passed counterparty node_id {}", counterparty_node_id) })
 		}
 
 		let mut peer_state_lock = peer_state_mutex_opt.unwrap().lock().unwrap();