wait_for_publish() deadlocks with qos=2 on server or connection shutdown

[philipp-schmidt]

I am using loop_start() and publish messages with qos level 2 and then wait for the message to be properly published via wait_for_publish(). Stopping the mqtt server sometimes results in a race condition where wait_for_publish() will lock up infinitely, probably waiting for the qos handshake to finish, but the connection is already dead. This completely deadlocks the thread. The automatic reconnect of loop_start() does not resolve this either, even if the server restarts.

topic = "test"
payload = "test"
publish_status = client.publish(topic, payload, qos=2)
# This will sometimes deadlock when shutting down the server externally or dropping connection
publish_status.wait_for_publish()

[swamper123]

I inspect the same behaviour with qos=1 and a Mosquitto broker.

The Mosquitto server timeouts the client, but the python program is still running.

[MattBrittan]

Apologies for the delay in someone looking into this.

I'd appreciate it if someone could confirm if this is still an issue with the current release and provide a reproducer and/or logs (understand that this might be difficult to reproduce). In theory the client should reconnect, and the publish process should conclude as normal (but this gets complicated and I have not traced in detail).