Fix a panic-while-panic'ing error where we fail drop checks

In `bolt2_open_channel_sending_node_checks_part1`, when we drop
`nodes[1]` after `create_channel` panics, we fail the
no-pending-messages test as it as an expeted `accept_channel` in
its outbound buffer. This avoids that by fetching the
`accept_channel` before we go into the panic'ing `create_channel`
call.

Author: TheBlueMatt

lightning/src/ln/functional_tests.rs

@@ -5927,6 +5927,7 @@ fn bolt2_open_channel_sending_node_checks_part1() { //This test needs to be on i
 	nodes[0].node.create_channel(nodes[1].node.get_our_node_id(), channel_value_satoshis, push_msat, 42, None).unwrap();
 	let node0_to_1_send_open_channel = get_event_msg!(nodes[0], MessageSendEvent::SendOpenChannel, nodes[1].node.get_our_node_id());
 	nodes[1].node.handle_open_channel(&nodes[0].node.get_our_node_id(), InitFeatures::known(), &node0_to_1_send_open_channel);
+	get_event_msg!(nodes[1], MessageSendEvent::SendAcceptChannel, nodes[0].node.get_our_node_id());
 
 	// Create a second channel with the same random values. This used to panic due to a colliding
 	// channel_id, but now panics due to a colliding outbound SCID alias.