The PubSubWorkerThread now uses a `threading.Event` to control its life cycle.

The `Redis` class and the `ConnectionPool` class now support the
"health_check_interval=N" option. By default N=0, which turns off health
checks. `N` should be an integer, and when greater than 0, ensures that
a health check is performed just before command execution anytime the
underlying connection has been idle for more than N seconds. A health
check is a full PING/PONG round trip to the Redis server.

If a health check encounters a ConnectionError or TimeoutError, the connection
is disconnected and reconnected and the health check is retried exactly once.
Any error during the retry is raised to the caller. Health check retries
are not governed by any other options such as `retry_on_timeout`. In systems
where idle times are common, these health checks are the intended way to
reconnect to the Redis server without harming any user data.

When this option is enabled for PubSub connections, calling `get_message()` or
`listen()` will send a health check anytime a message has not been read on
the PubSub connection for `health_check_interval` seconds. Users should
call `get_message()` or `listen()` at least every `health_check_interval`
seconds in order to keep the connection open.

This has multiple uses:
  * Single connection clients will not be considered threadsafe. This means
    certain settings could temporarily be adjusted. For example, a context
    manager could temporarily modify the encoding behavior for a set
    of commands.
  * We can introduce more thorough health checks that only happen when a
    connection is handed out from the connection pool.
  * Workloads that issue many commands to Redis should be slightly faster.
    Prior to this change, the client must retrieve a connection from the
    pool for each command.

Fixes #1191