JmsServiceExporter.destroy() doesn't close the underlying JmsProducer.

See the attached unit test (which includes a demonstration of the 'fix"). After calling destroy()
on the JmsServiceExporter instance, a number of threads are left lying around:
Thread[Durable Queue Worker,6,main]
Thread[Dispatch Worker,6,main]
Thread[Thread-1,6,main]
Thread[ActiveMQ ShutdownHook,5,main]
Thread[TcpTransportServerChannel@tcp://A0277:61615,5,main]
Thread[TcpTransportChannel: Socket[addr=/127.0.0.1,port=4016,localport=61615],5,main]
Thread[TcpTransportChannel: Socket[addr=localhost/127.0.0.1,port=61615,localport=4016],7,main]
Thread[JmsSessionDispatcher: 1,10,main]

Calling close() on the JmsProducer cleans these up.