Do all messages require TimeoutIds?

[o-alex]

It seems to be hardcoded that all message require a TimeoutId:
in DiectMsgNettyFactory - Request 0 line 88
this constructor is always called be decoding code which hardcodes the timeoutid boolean to true for both Request and Response, which means the decoding code always wants to read a timeoutId

However, the encoding code in DirectMsgNetty - writeHeader method, line 74 checks if the message has a timeout defined. If it does it writes it and everything works correct. If the message does not have a timeout it will not write anything, effectively jumping over it. The encoding succeds(with no timeout written) - but if we try to decode this it will fail as it will actually try to read a timeout.
Should the encoding fail if no timeout present or should we write a marker flag to mark the existence or absence of a timeout

[jimdowling]

For Request/Response, you have to have a TimeoutId. For a OneWay msg, there is no TimeoutId. Line 74 is for OneWay msgs.