let mut peer_poller = Arc::new(Mutex::new(PeerPoller::new()?));
let mut peer_poller_clone = peer_poller.clone();
thread::spawn(move || {
    loop {
        let p = rx.recv().unwrap();
        match peer_poller_clone.lock() {    <---- A
            Ok(mut g) => {
                g.add_peer(p);
            }
            Err(e) => {}
        }
    }
});
tx.send(peer)
match peer_poller.lock() {        <---- B
    Ok(mut g) => {
        g.run();
    }
    Err(e) => {}
}

The run() method of PeerPoller is blocking, so the peer_poller.lock() at site B will never be released. This cause the lock() at site A to never succeed. How can I refactor this code to be more idomatic in Rust?

I understand that one option is to refactor code so that add_peer() takes &self and performs modifications via RefCell (interior mutability). However, I’m curious if there is a better solution. Any suggestions?