"stop" fails on Linux about crust HOT 4 CLOSED

If I run multiple instances of
cargo run --example crust_peer
then the stop command works fine on all but the last remaining instance.
If I only run a single instance then it also fails and I require a ctrl-c to regain the prompt.
Latest git versions of all libs and last nights rust on Ubuntu 15.04

from crust.

This might be related to shut-down of the beacon server. I have not yet looked into the code.
Its only the first node which acquires the beacon port (on a machine). I will update after investigating it.

from crust.

The code blocks on beacon.rs:L172/std::net::addr::SocketAddr::recv_from.

I'm studying the commit that introduced such feature to fix the issue.

from crust.

std::net::addr::SocketAddr::recv_from is used to wait for the other thread to finish cleanly. The problem is that the crust_peer example calls ConnectionManager::stop twice and BroadcastAcceptor::stop doesn't handle "double stop correctly". I'll write a patch soon.

from crust.