Handle BroadcastChannelUpdate in push_excess_b_events

When process_msg_events processes a single message for node B (via
OneMessage or OnePendingMessage mode), remaining events are passed to
push_excess_b_events for routing. Timer ticks can generate
BroadcastChannelUpdate events from marking channels disabled on
disconnect, which may be left over after the single processed message.
Since it is a broadcast not directed at a specific peer, it can be
safely skipped.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

Author: jkczyz

fuzz/src/chanmon_consistency.rs

@@ -1687,6 +1687,7 @@ pub fn do_test<Out: Output + MaybeSend + MaybeSync>(data: &[u8], out: Out) {
 						},
 						MessageSendEvent::SendChannelReady { .. } => continue,
 						MessageSendEvent::SendAnnouncementSignatures { .. } => continue,
+						MessageSendEvent::BroadcastChannelUpdate { .. } => continue,
 						MessageSendEvent::SendChannelUpdate { ref node_id, .. } => {
 							if Some(*node_id) == expect_drop_id { panic!("peer_disconnected should drop msgs bound for the disconnected peer"); }
 							*node_id == a_id