Fix example code for Docker events (#70)

1.  The event_task must be cancelled in any case. Cancelling a task
    never raises an exception even if the task is already done.
2.  Check that the event_task has not been cancelled and get its result.
    If the task has been cancelled properly, nothing will happen. If the
    task had an exception, the exception will be raised and displayed
    properly to the terminal.
3.  Add another level of try...except because even if the Docker
    exception is raised by the task, we still need to close the loop.

Fixes #34

examples/events.py

@@ -34,6 +34,8 @@ async def demo(docker):
 
     while True:
         event = await subscriber.get()
+        if event is None:
+            break
         print(f"event: {event!r}")
 
         # Demonstrate simple event-driven container mgmt.
@@ -50,14 +52,18 @@ async def demo(docker):
 
 if __name__ == '__main__':
     loop = asyncio.get_event_loop()
-    docker = Docker()
     try:
+        docker = Docker()
         # start a monitoring task.
         event_task = loop.create_task(docker.events.run())
-        # do our stuffs.
-        loop.run_until_complete(demo(docker))
-        # explicitly stop monitoring.
-        event_task.cancel()
+        try:
+            # do our stuffs.
+            loop.run_until_complete(demo(docker))
+        finally:
+            # explicitly stop monitoring.
+            event_task.cancel()
+            loop.run_until_complete(docker.close())
+            if not event_task.cancelled():
+                event_task.result()  # NOTE: maybe raise an exception
     finally:
-        loop.run_until_complete(docker.close())
         loop.close()