Topology.restart() and Topology.clearMessages() do not clear delayed messages

[remikey]

#58 as added delayed messages support in the default MessageEngine.

Currently, when calling Topology.restart(), the Topology "manually" clears each Node's mailBox and sendQueue, using Topology.clearMessages().
Three points:

• I think that clearing messages should be under the responsibility of the MessageEngine

• The current implementation does not clear the delayMessages

  Because of these pending messages, in a simple delay-enabled broadcast example, a broadcast wave might oddly continue although the user asked for a restart.

• I don't think Topology.clearMessages() and Topology.restart() have the behavior a user would expect

I suggest creating a MessageEngine.reset() which would do the job, and would be called by Topology.clearMessages().

@acasteigts , do you agree?

[acasteigts]

Yes, this sounds very reasonable.

[remikey]

Applied and merged into develop.