In this tutorial you'll learn the difference bewteen the newUnreferencedActor and newReferencedActor method.
The following new concepts will be introduced:
- Actor::newUnreferencedActor()
- Actor::newReferencedActor()
Two methods are available to instantiate an Actor from within another Actor: newUnreferencedActor() (used in previous tutorials) and newReferencedActor().
As seen in previous tutorials, newUnreferencedActor() instantiates a new Actor on the same CPU core as the calling Actor. This method returns a unique ActorId corresponding to the newly created Actor instance.
Just as with newUnreferencedActor(), newReferencedActor() instantiates an Actor on the same CPU core as the caller Actor. The return type if different, however, here an ActorReference is returned. This object corresponds to an Actor reference.
While an Actor has an instance of a ActorReference object, the referenced Actor will not be able to terminate (the destructor will not be called). Hence, an ActorReference is always valid.
This tutorial makes use of the ActorReference to directly call a method from the referenced actor (PrinterActor). This bypasses having to send an Event through a pipe and the waiting for the EventLoop to dispatch said Event, which is obviously faster (when in the same CPU core).
ActorReference<PrinterActor> printerActorRef = newReferencedActor<PrinterActor>();
printerActorRef->printMessage("Hello using printerActorRef", getActorId());