IDL command replies currently construct their own generic Command Reply serialization context objects, but these do not reflect the flags retrieved during Command Request parsing.  Since Command Reply serialization highly depends on these flags, we need a way to construct new Command Reply serialization context objects, while ensuring they retain flags gained during deserialization.

As the goal of these changes are in pursuit of fixing tests that will fail once serializeForCommands and deserializeForCommands are plumbed in, no new tests are being added for this ticket.