When a server fails to start it does not cleanup its WebSocketWorker threads

[vsch]

When server start fails because address is already in use, it calls stop which causes a null pointer exception because selector is not yet initialized.

This exception causes the server not to interrupt its web socket workers.

Expected Behavior

cleanup web workers if server start fails

Current Behavior

Leaves the workers running

Possible Solution

In WebSocket server stop(int) method

		synchronized ( this ) {
			if( selectorthread != null  ) {
				selector.wakeup();
				selectorthread.join( timeout );
			}
		}

change to:

		synchronized ( this ) {
			if( selectorthread != null  ) {
				if (selector != null) selector.wakeup();
				selectorthread.join( timeout );
			}
		}

Steps to Reproduce (for bugs)

1.Start a server
2. Start a second server with the same address
3. Once it fails the Worker threads are still alive
4.

Debug log (for bugs)

Context

Your Environment

• Version used: 1.3.7
• Java version:
• Operating System and version:
• Endpoint Name and version:
• Link to your project:

[marci4]

Hello @vsch,

thank you for your bug report.

I think this issue should already be solved on the master due to #661.

Could you please confirm this?

Greetings
marci4

[vsch]

@marci4, yes this is exactly the same issue.