fix: properly prevent deadlock on container cleanup

`ecs` was being locked once in find_entities, and another time when
`ContainerRef::close()` called by the garbage collector from table.set.

src/events.rs

@@ -14,8 +14,6 @@ use tokio::net::TcpListener;
 
 #[allow(clippy::too_many_lines)]
 pub async fn handle_event(client: Client, event: Event, state: State) -> anyhow::Result<()> {
-    state.lua.gc_stop();
-
     match event {
         Event::AddPlayer(player_info) => {
             call_listeners(&state, "add_player", Player::from(player_info)).await;