Message Passing

Sending and receiving messages is a fundamental concept in CAF. Actors communicate by sending messages to each other. A message is a tuple of values that can be delivered to local or remote actors (network transparency).

The class message holds a sequence of values of arbitrary types. Users can create messages directly by calling make_message(...) or by using a message_builder. The latter allows users to build messages incrementally.

However, users rarely interact with messages directly. Instead, they define message handlers that process incoming messages and create message implicitly when using the message passing API.

Copy on Write

A message is a copy-on-write (COW) type. This means that copying a message is cheap because it only copies the reference to the message content. The actual copying of the content only happens when one of the copies is modified.

This allows sending the same message to multiple receivers without copying overhead, as long as all receivers only read the content of the message.

Actors copy message contents whenever other actors hold references to it and if one or more arguments of a message handler take a mutable reference. Again, this is transparent to the user most of the time. However, it is still important to know about the COW semantics for understanding the performance characteristics of an actor system.

Sending Messages: The Mail API

Sending messages in CAF is starts by calling the member function mail or the free function anon_mail. The arguments of these functions are the contents of the message. Then, CAF returns a builder object that allows users to specify additional information about the message, such as the priority or whether the message shall be send after some delay:

• Calling urgent() on the builder object sets the priority of the message to message_priority::high. This causes the receiver to process the message before regular messages.

• Calling schedule(actor_clock::time_point timeout) on the builder object instructs CAF to send the message at a specific point in time.

• Calling delay(actor_clock::duration timeout) on the builder object instructs CAF to send the message after a relative timeout has passed.

In all cases, CAF returns a new builder object that allows users to chain multiple calls. The final call to the builder object is one of:

• send(Handle receiver) to send the message to a specific actor as an asynchronous message (fire-and-forget).

• delegate(Handle receiver) to send the message to a specific actor and transfer the responsibility for responding to the original sender (see Delegating Messages).

• request(Handle receiver, timespan timeout) to send the message to a specific actor as a request message. CAF will raise an error if the receiver does not respond within the specified timeout (passing infinite as timeout disables the timeout).

When sending a delayed or scheduled message, these member functions have two additional, optional parameters: a RefTag and a SelfRefTag. These tags configure what kind of reference CAF will hold onto for the actors while the message is pending. The RefTag specifies what kind of reference CAF will hold onto for the receiver of the message. By default, CAF uses strong_ref, which instructs CAF to store a strong reference to the receiver. Passing weak_ref instead will cause CAF to store a weak reference instead. Likewise, SelfRefTag specifies what kind of reference CAF will hold onto for the sender of the message. By default, CAF uses strong_self_ref, which instructs CAF to store a strong reference to the sender. Passing weak_self_ref will cause CAF to store a weak reference instead.

When using weak references, CAF will try to convert them to strong references before sending the message. If the conversion fails, CAF will drop the message.

When calling send or delegate, the result is either void (immediate send) or a disposable (delayed or scheduled send). The latter allows users to cancel a send operation while it is still pending.

When calling request, CAF will return a handle for the response message. This handle either offers then and await member functions when using an event-based actor or receive when using a blocking actor (see Requests)

Note: the builder object from anon_send only supports send.

Requirements for Message Types

Message types in CAF must meet the following requirements:

1. Inspectable (see Type Inspection)

2. Default constructible

3. Copy constructible

A type T is inspectable if it provides a free function inspect(Inspector&, T&) or specializes inspector_access. Requirement 2 is a consequence of requirement 1, because CAF needs to be able to create an object for T when deserializing incoming messages. Requirement 3 allows CAF to implement Copy on Write (see Copy on Write).

Requests

A main feature of CAF is its ability to couple input and output types via the type system. For example, a typed_actor<result<int32_t>(int32_t)> essentially behaves like a function. It receives a single int32_t as input and responds with another int32_t. CAF embraces this functional take on actors by simply creating response messages from the result of message handlers. This allows CAF to match request to response messages and to provide a convenient API for this style of communication.

Sending Requests and Handling Responses

Actors send request messages by calling mail(content...).request(receiver, timeout). This function returns an intermediate object that allows an actor to set a one-shot handler for the response message. Event-based actors can use either request(...).then or request(...).await. The former multiplexes the one-shot handler with the regular actor behavior and handles requests as they arrive. The latter suspends the regular actor behavior until all awaited responses arrive and handles requests in LIFO order. Blocking actors always use request(...).receive, which blocks until the one-shot handler was called. Actors receive a sec::request_timeout (see Default Error Codes) error message (see Error Handler) if a timeout occurs. Users can set the timeout to infinite for unbound operations. This is only recommended if the receiver is known to run locally.

In our following example, we use the simple cell actor shown below as communication endpoint.

struct cell_trait {
  using signatures
    = type_list<result<void>(put_atom, int32_t), // 'put' writes to the cell
                result<int32_t>(get_atom)>;      // 'get 'reads from the cell
};
using cell = typed_actor<cell_trait>;

struct cell_state {
  static constexpr inline const char* name = "cell";

  cell::pointer self;

  int32_t value;

  cell_state(cell::pointer ptr, int32_t val) : self(ptr), value(val) {
    // nop
  }

  cell::behavior_type make_behavior() {
    return {
      [this](put_atom, int32_t val) { value = val; },
      [this](get_atom) { return value; },
    };
  }
};

To showcase the slight differences in API and processing order, we implement three testee actors that all operate on a list of cell actors.

void waiting_testee(event_based_actor* self, vector<cell> cells) {
  for (auto& x : cells)
    self->mail(get_atom_v).request(x, 1s).await([self, x](int32_t y) {
      self->println("cell #{} -> {}", x.id(), y);
    });
}

void multiplexed_testee(event_based_actor* self, vector<cell> cells) {
  for (auto& x : cells)
    self->mail(get_atom_v).request(x, 1s).then([self, x](int32_t y) {
      self->println("cell #{} -> {}", x.id(), y);
    });
}

void blocking_testee(scoped_actor& self, vector<cell> cells) {
  for (auto& x : cells)
    self->mail(get_atom_v)
      .request(x, 1s)
      .receive([&](int32_t y) { self->println("cell #{} -> {}", x.id(), y); },
               [&](error& err) {
                 self->println("cell #{} -> {}", x.id(), err);
               });
}

Our caf_main for the examples spawns five cells and assign the initial values 0, 1, 4, 9, and 16. Then it spawns one instance for each of our testee implementations and we can observe the different outputs.

Our waiting_testee actor will always print:

cell #9 -> 16
cell #8 -> 9
cell #7 -> 4
cell #6 -> 1
cell #5 -> 0

This is because await puts the one-shots handlers onto a stack and enforces LIFO order by re-ordering incoming response messages as necessary.

The multiplexed_testee implementation does not print its results in a predicable order. Response messages arrive in arbitrary order and are handled immediately.

Finally, the blocking_testee has a deterministic output again. This is because it blocks on each request until receiving the result before making the next request.

cell #5 -> 0
cell #6 -> 1
cell #7 -> 4
cell #8 -> 9
cell #9 -> 16

Both event-based approaches send all requests, install a series of one-shot handlers, and then return from the implementing function. In contrast, the blocking function waits for a response before sending another request.

Error Handling in Requests

Requests allow CAF to unambiguously correlate request and response messages. This is also true if the response is an error message. Hence, CAF allows to add an error handler as optional second parameter to then and await (this parameter is mandatory for receive). If no such handler is defined, the default error handler (see Error Handler) is used as a fallback in scheduled actors.

As an example, we consider a simple divider that returns an error on a division by zero. This examples uses a custom error category (see Errors).

struct divider_trait {
  using signatures = type_list<result<double>(div_atom, double, double)>;
};

using divider = typed_actor<divider_trait>;

divider::behavior_type divider_impl() {
  return {
    [](div_atom, double x, double y) -> result<double> {
      if (y == 0.0)
        return math_error::division_by_zero;
      return x / y;
    },
  };
}

When sending requests to the divider, we can use a custom error handlers to report errors to the user like this:

  auto div = system.spawn(divider_impl);
  scoped_actor self{system};
  self->mail(div_atom_v, x, y)
    .request(div, 10s)
    .receive([&](double z) { self->println("{} / {} = {}", x, y, z); },
             [&](const error& err) {
               self->println("*** cannot compute {} / {} => {}", x, y, err);
             });

Delegating Messages

Actors can transfer responsibility for a request by using delegate. This enables the receiver of the delegated message to reply as usual—simply by returning a value from its message handler—and the original sender of the message will receive the response. The following diagram illustrates request delegation from actor B to actor C.

A                  B                  C
|                  |                  |
| ---(request)---> |                  |
|                  | ---(delegate)--> |
|                  X                  |---\
|                                     |   | compute
|                                     |   | result
|                                     |<--/
| <-------------(reply)-------------- |
|                                     X
X

Returning the result of delegate(...) from a message handler, as shown in the example below, suppresses the implicit response message and allows the compiler to check the result type when using statically typed actors.

struct adder_trait {
  using signatures = type_list<result<int32_t>(add_atom, int32_t, int32_t)>;
};
using adder_actor = typed_actor<adder_trait>;

adder_actor::behavior_type worker_impl() {
  return {
    [](add_atom, int32_t x, int32_t y) { return x + y; },
  };
}
adder_actor::behavior_type server_impl(adder_actor::pointer self,
                                       adder_actor worker) {
  return {
    [=](add_atom add, int32_t x, int32_t y) {
      return self->delegate(worker, add, x, y);
    },
  };
}

void client_impl(event_based_actor* self, adder_actor adder, int32_t x,
                 int32_t y) {
  self->mail(add_atom_v, x, y).request(adder, 10s).then([=](int32_t result) {
    self->println("{} + {} = {}", x, y, result);
  });
}

void caf_main(actor_system& sys) {
  auto worker = sys.spawn(worker_impl);
  auto server = sys.spawn(server_impl, sys.spawn(worker_impl));
  sys.spawn(client_impl, server, 1, 2);
}

Response Promises

Response promises allow an actor to send and receive other messages prior to replying to a particular request. Actors create a response promise using self->make_response_promise<Ts...>(), where Ts is a template parameter pack describing the promised return type. Dynamically typed actors simply call self->make_response_promise(). After retrieving a promise, an actor can fulfill it by calling the member function deliver(...), as shown in the following example.

struct adder_trait {
  using signatures
    = caf::type_list<result<int32_t>(add_atom, int32_t, int32_t)>;
};
using adder_actor = typed_actor<adder_trait>;

adder_actor::behavior_type worker_impl() {
  return {
    [](add_atom, int32_t x, int32_t y) { return x + y; },
  };
}
adder_actor::behavior_type server_impl(adder_actor::pointer self,
                                       adder_actor worker) {
  return {
    [=](add_atom, int32_t y, int32_t z) {
      auto rp = self->make_response_promise<int32_t>();
      self->mail(add_atom_v, y, z)
        .request(worker, infinite)
        .then([rp](int32_t result) mutable { rp.deliver(result); },
              [rp](error& err) mutable { rp.deliver(std::move(err)); });
      return rp;
    },
  };
}

void client_impl(event_based_actor* self, adder_actor adder, int32_t x,
                 int32_t y) {
  self->mail(add_atom_v, x, y).request(adder, 10s).then([=](int32_t result) {
    self->println("{} + {} = {}", x, y, result);
  });
}

void caf_main(actor_system& sys) {
  auto worker = sys.spawn(worker_impl);
  auto server = sys.spawn(server_impl, sys.spawn(worker_impl));
  sys.spawn(client_impl, server, 1, 2);
}

This example is almost identical to the example for delegating messages. However, there is a big difference in the flow of messages. In our first version, the worker (C) directly responded to the client (A). This time, the worker sends the result to the server (B), which then fulfills the promise and thereby sends the result to the client:

A                  B                  C
|                  |                  |
| ---(request)---> |                  |
|                  | ---(request)---> |
|                  |                  |---\
|                  |                  |   | compute
|                  |                  |   | result
|                  |                  |<--/
|                  | <----(reply)---- |
|                  |                  X
| <----(reply)---- |
|                  X
X

Special Message Handlers

CAF has a few system-level message types (exit_msg, and error) that all actors should handle regardless of their current state. Consequently, event-based actors handle such messages in special-purpose message handlers. Note that blocking actors have neither of those special-purpose handlers (see Blocking Actors) and should instead provide callbacks for these messages when calling receive.

Exit Handler

Bidirectional monitoring with a strong lifetime coupling is established by calling self->link_to(other). This will cause the runtime to send an exit_msg if either this or other dies. Per default, actors terminate after receiving an exit_msg unless the exit reason is exit_reason::normal. This mechanism propagates failure states in an actor system. Linked actors form a sub system in which an error causes all actors to fail collectively. Actors can override the default handler via set_exit_handler(f), where f is a function object with signature void (exit_message&) or void (scheduled_actor*, exit_message&).

Error Handler

Actors send error messages to others by returning an error (see Errors) from a message handler. Similar to exit messages, error messages usually cause the receiving actor to terminate, unless a custom handler was installed via set_error_handler(f), where f is a function object with signature void (error&) or void (scheduled_actor*, error&). Additionally, request accepts an error handler as second argument to handle errors for a particular request (see error-response). The default handler is used as fallback if request is used without error handler.