Behavior of `Timeout` class is not entirely intuitive

[dallmair]

The stopped() method returns the information whether the timeout is still running or not, and it does not check for expiration. The expired() method checks for expiration, and also stops the timer, so it only returns true a single time. In a table:

Situation	stopped()	expired()
Stopped	true	false
Started, not expired yet	false	false
Started, expired	before expired() is called: false after expired() is called: true	true, but only once false on second and all following calls

The current implementation is good for use cases where some code needs to be run a single time after a timeout expires. However, it's not ideal for use cases where one actually needs "delay" semantics, i.e. do nothing while the timeout is running, but always run the code after expiration.

It's not clear at the moment what the best solution is. Ideas:

1. Change semantics of the expired() method to return true when the Timeout is stopped. Need to check all existing usages whether this is possible, and also whether it makes sense.
2. Add a method stoppedOrExpired(), which at least shows up in IntelliSense and should make developers think.
3. Add another class Delay that actually implements delay semantics. Could use Timeout internally.