DistributedTargetInvocationDecoder

Protocol requirements

Similar to the DistributedTargetInvocationEncoder and its recordArgument and recordReturnType protocol requirements, the DistributedTargetInvocationDecoder contains a method which is not possible to express in source due to advanced use of generics combined with associated types. Specifically, the decodeNextArgument method is not expressed in source as protocol requirement, but will be treated by the compiler as-if it was.

In addition to the compiler offering compile errors if this witness is missing in an adopting type, we present its signature here for reference:

/// Ad-hoc protocol requirement
///
/// Attempt to decode the next argument from the underlying buffers into pre-allocated storage
/// pointed at by 'pointer'.
///
/// This method should throw if it has no more arguments available, if decoding the argument failed,
/// or, optionally, if the argument type we're trying to decode does not match the stored type.
///
/// The result of the decoding operation must be stored into the provided 'pointer' rather than
/// returning a value. This pattern allows the runtime to use a heavily optimized, pre-allocated
/// buffer for all the arguments and their expected types. The 'pointer' passed here is a pointer
/// to a "slot" in that pre-allocated buffer. That buffer will then be passed to a thunk that
/// performs the actual distributed (local) instance method invocation.
mutating func decodeNextArgument<Argument: SerializationRequirement>() throws -> Argument

Decoding DistributedActor arguments using Codable

When using an actor system where ActorID is Codable, every distributed actor using that system is also implicitly Codable (see DistributedActorSystem). Such distributed actors are encoded as their ActorID stored in an Encoder.singleValueContainer. When Codable is being used by such a system, the decodeNextArgument method will be using Decoder to decode the incoming values, which may themselves be distributed actors.

An actor system must be provided to the Decoder in order for a distributed actor’s Decodable.init(from:) to be able to return the instance of the actor. Specifically, the decodedActorID is passed to the actor system’s resolve(id:as:) method in order to return either a local instance identified by this ID, or creating a remote actor reference. Thus, you must set the actor system the decoding is performed for, on the decoder’s userInfo, as follows:

mutating func decodeNextArgument<Argument: SerializationRequirement>() throws -> Argument {
  let argumentData: Data = /// ...
  // ...
  decoder.userInfo[.actorSystemKey] = self.actorSystem
  return try Argument.decode(
}