Google celebrates Women's History Month. See how.

Dart bindings


Given the library declaration:

library fuchsia.examples;

The bindings code for this library is generated into a fidl_fuchsia_examples_async dart library. The fidl_ prefix and _async suffix are hardcoded by the FIDL toolchain.

This code can then be imported using:

import 'package:fidl_fuchsia_examples/fidl_async.dart' as fidl_examples;


All constants are generated as a const. For example, the following constants:

const BOARD_SIZE uint8 = 9;
const NAME string = "Tic-Tac-Toe";

Are generated as:

const int BOARD_SIZE = 9;
const String NAME = "Tic-Tac-Toe";

The correspondence between FIDL primitive types and Dart types is outlined in built-in types.


This section describes how the FIDL toolchain converts FIDL types to native types in Dart. These types can appear as members in an aggregate type, as parameters to a protocol method, or as the type contained in an event or method response Future.

Nullable types do not have different generated types than their non-nullable counterparts in Dart.

Built-in types

The FIDL types are converted to Dart types based on the following table:

FIDL Type Dart Type
bool bool
int8, int16, int32, int64, uint8, uint16, uint32, uint64 int
float32, float64 double
array<int8, N>, vector<int8>:N Int8List
array<int16, N>, vector<int16>:N Int16List
array<int32, N>, vector<int32>:N Int32List
array<int64, N>, vector<int64>:N Int64List
array<uint8, N>, vector<uint8>:N Uint8List
array<uint16, N>, vector<uint16>:N Uint16List
array<uint32, N>, vector<uint32>:N Uint32List
array<uint64, N>, vector<uint64>:N Uint64List
array<float32, N>, vector<float32>:N Float32List
array<float64, N>, vector<float64>:N Float64List
array<T, N>, vector<T>:N List<T>
string String
server_end:P fidl.InterfaceRequest<P>
client_end:P fidl.InterfaceHandle<P>
zx.handle:CHANNEL zircon.Channel
zx.handle:EVENTPAIR zircon.EventPair
zx.handle:SOCKET zircon.Socket
zx.handle:VMO zircon.Vmo
zx.handle:S, zx.handle zircon.Handle

Response and event parameters

Method response and event types (see Protocols) are represented using Future<T>, where T is a type containing all of the response/event parameters. This section describes how the FIDL toolchain generates this inner type T.

  • Empty responses and events use void.
  • Responses and events with a single parameter T just use T as the response or event type.
  • Responses and events with multiple parameters use a generated wrapper class which follows the naming scheme [Protocol]$[Method]$Response. For example, an event OnOpponentMove for protocol TicTacToe that has multiple parameters would use generated class TicTacToe$OnOpponentMove$Response. This class provides a single method: the constructor, which has positional arguments corresponding to the response or event parameters.

Note that methods that do not have a response will have a response type of Future<void>, which is the same type used by methods with an empty response. In the former case, the Future can be expected to resolve immediately after sending the request, whereas in the latter case, the Future is only resolved after receiving the empty response from the server.

Type definitions


Given the bits definition:

type FileMode = strict bits : uint16 {
    READ = 0b001;
    WRITE = 0b010;
    EXECUTE = 0b100;

The FIDL toolchain generates a FileMode class with static const variables for each bits member, as well as for a FileMode with no flag set ($none) or every flag set ($mask):

  • static const FileMode read
  • static const FileMode write
  • static const FileMode execute
  • static const FileMode $none
  • static const FileMode $mask

FileMode provides the following methods:

  • int get $value: Getter for the underlying int value.
  • String toString(): Returns a readable representation of the FileMode.
  • FileMode operator |(FileMode other): Bitwise or operator.
  • FileMode operator &(FileMode other): Bitwise and operator.
  • bool operator==(Object other): Equality operator.
  • int getUnknownBits(): Returns only the set bits that are unknown. Always returns 0 for strict bits.
  • bool hasUnknownBits(): Returns whether this value contains any unknown bits. Always returns false for strict bits.

Example usage:

test('bits', () {
  expect(fidl_examples.FileMode.$none.$value, equals(0));
  expect($value, equals(1));
  final readWrite = | fidl_examples.FileMode.write;
  expect(readWrite.toString(), 'FileMode(3)');


Given the enum definition:

type LocationType = strict enum {
    MUSEUM = 1;
    AIRPORT = 2;

The FIDL toolchain generates a LocationType class with static const variables for each enum member:

  • static const LocationType museum
  • static const LocationType airport
  • static const LocationType restaurant

As well as the following variables:

  • static const Map<String, LocationType> $valuesMap: A mapping of the string representation of the member ('museum', 'airport', or 'restaurant') to its corresponding enum value (, LocationType.airport, or
  • static const List<LocationType> $values: A list of all of the enum values.

If LocationType is flexible, it will have an unknown placeholder member as well:

  • static const LocationType $unknown

If the enum has a member tagged with the [Unknown] attribute, the placeholder variable will have the same value as the tagged unknown member.

LocationType provides the following methods:

  • static LocationType $valueOf(String name): Look up a string name in the $valuesMap.
  • String toString(): Returns a readable representation of the LocationType.
  • bool isUnknown(): Returns whether this enum is unknown. Always returns false for strict enums.

Example usage:

test('enums', () {
  const museum =;
  expect(fidl_examples.LocationType(1), equals(museum));
  expect(fidl_examples.LocationType.$valueOf('museum'), equals(museum));
  expect(museum.toString(), 'LocationType(1)');


Given the struct declaration:

type Color = struct {
    id uint32;
    name string:MAX_STRING_LENGTH = "red";

The FIDL toolchain generates a Color class with the following methods:

  • const Color({@required id, name}): The constructor for Color takes named arguments corresponding to the struct's fields. Fields that are not nullable and do not have a default value specified are marked as @required.
  • int get id: Getter for the id field.
  • String get name: Getter for the name field.
  • Color.clone(Color, {int id, String name}): Clone constructor that will clone an existing Color, possibly overriding specific field values based on the provided named arguments.
  • List<Object> get $fields: Returns a list of fields in declaration order.
  • String toString(): Returns a readable string of the Color
  • bool operator==(Object other): Equality operator that performs a deep comparison when compared to another instance of a Color.

Example usage:

test('structs', () {
  final withDefaultName = fidl_examples.Color(id: 0);
  expect(, equals('red'));
  expect(withDefaultName.$fields, equals([0, 'red']));

  final blue = fidl_examples.Color(id: 1, name: 'blue');
  expect(blue == withDefaultName, equals(false));
  expect(blue.toString(), 'Color([1, blue])');

  final deepBlue = fidl_examples.Color.clone(blue, name: 'deep blue');
  expect(, equals(1));
  expect(blue == deepBlue, equals(false));


Given the union definition:

type JsonValue = strict union {
    1: reserved;
    2: int_value int32;
    3: string_value string:MAX_STRING_LENGTH;

FIDL generates an enum representing the tags of the union:

enum JsonValueTag {

As well as a JsonValue class with the following methods:

  • const JsonValue.withIntValue(int) and const JsonValue.withStringValue(String): Constructors for each variant.
  • JsonValueTag get $tag: Getter for the tag corresponding to this the variant of this union.
  • int? get intValue and String? get stringValue: Getter for the underlying value. If the instance's variant does not match the getter method, null is returned.
  • String toString(): Returns a readable string of the JsonValue.
  • int get $ordinal: Getter for the underlying ordinal value.
  • Object get $data: Getter for the underlying union data.
  • bool operator ==(Object other): Equality operator that performs deep comparison when compared to another JsonValue of the same variant.
  • fidl.UnknownRawData? get $unknownData: Returns the bytes and handles of the unknown data if this union contains an unknown variant, or null otherwise. Always returns null for strict unions.

If JsonValue is flexible, it will have the following additional methods:

  • const JsonValue.with$UnknownData(int ordinal, fidl.UnknownRawData data): Constructor for a value with an unknown variant set. This should only be used for testing, e.g. to check that code handles unknown unions correctly.

Example usage:

test('unions', () {
  final intVal = fidl_examples.JsonValue.withIntValue(1);
  expect(intVal.$tag, equals(fidl_examples.JsonValueTag.intValue));
  expect(intVal.intValue, equals(1));
      intVal == fidl_examples.JsonValue.withStringValue('1'), equals(false));

  expect(intVal.toString(), 'JsonValue(2: 1)');

Flexible unions and unknown variants

Flexible unions have an extra variant in the generated tag class:

enum JsonValueTag {

When a FIDL message containing a union with an unknown variant is decoded into JsonValue, JsonValue.$tag returns JsonValueTag.$unknown, and JsonValue.$ordinal returns the unknown ordinal.

Encoding a union with an unknown variant writes the unknown data and the original ordinal back onto the wire.

Strict unions fail when decoding an unknown variant. Flexible unions that are value types fail when decoding an unknown variant with handles.


Given the table definition:

type User = table {
    1: reserved;
    2: age uint8;
    3: name string:MAX_STRING_LENGTH;

The FIDL toolchain generates a User class that defines the following methods:

  • const User({$unknownData, age, name}): Constructor for User. Contains an optional parameter for each field as well as a map containing any unknown fields, as a Map<int, fidl.UnknownRawData>. Specifying a value for the unknown fields should only be done for testing, e.g. to test that a table with unknown fields is handled correctly.
  • int get age: Getter for the age field.
  • String get name: Getter for the name field.
  • Map<int, dynamic> get $fields: Returns a map of ordinals to field values.
  • Map<int, fidl.UnknownRawData>? get $unknownData: Returns a map of ordinals to unknown field values (i.e. bytes and handles). The list of handles is returned in traversal order, and is guaranteed to be empty if the table is a value type.
  • bool operator ==(Object other): Equality operator that performs deep comparison when compared to another User.

Example usage:

test('tables', () {
  final user = fidl_examples.User();
  expect(user.age, equals(null));
  expect(, equals(null));

  final fuchsia = fidl_examples.User(age: 100, name: 'fuchsia');
  expect(fuchsia.$fields, equals({2: 100, 3: 'fuchsia'}));
  expect(user == fuchsia, equals(false));

Inline layouts

The generated Dart code uses the the name reserved by fidlc for inline layouts.


Given the protocol:

protocol TicTacToe {
    StartGame(struct {
        start_first bool;
    MakeMove(struct {
        row uint8;
        col uint8;
    }) -> (struct {
        success bool;
        new_state box<GameState>;
    -> OnOpponentMove(struct {
        new_state GameState;

FIDL generates an abstract TicTacToe class, which defines the interface of the service used by clients to proxy calls to the server, and for the server for implementing the protocol.

TicTacToe contains a static const String $serviceName, which is defined depending on the presence of the @transitional attribute.

TicTacToe has the following abstract methods, representing the protocol methods:

  • async.Future<void> startGame(bool start_first): Abstract method for a fire and forget protocol method. It takes as arguments the request parameters and returns a future of void.
  • async.Future<TicTacToe$MakeMove$Response> makeMove(int row, int col): Abstract method for a two way protocol method. It takes as arguments the request parameters and returns a future of the response type.
  • async.Stream<GameState> get onOpponentMove: Getter for a Stream of onOpponentMove events.


The FIDL toolchain generates a TicTacToeProxy class that extends fidl.AsyncProxy<TicTacToe>, and provides an implementation for the abstract TicTacToe class that encodes and sends the request to the server end of the channel.

Example client code could thus look like the following:

final tictactoe = fidl_tictactoe.TicTacToeProxy();
// ...bind the proxy, omitted from this example
final state = await tictactoe.makeMove(0, 0);

Examples on how to set up and bind a proxy class to a channel are covered in the Dart tutorial.


Implementing a server for a FIDL protocol involves providing a concrete implementation of the appropriate interface. For a closed protocol, the TicTacToe abstract class is used directly as the interface implemented on the server. For an open or ajar protocol, an additional interface called TicTacToeServer is generated, which must be implemented on the server.

The bindings provide a TicTacToeBinding class that can bind to either a TicTacToe instance (if closed) or TicTacToeServer instance (if open or ajar), and a channel, and listens to incoming messages on the channel, dispatches them to the server implementation, and sends messages back through the channel. This class implements fidl.AsyncBinding<TicTacToe[Server]>.

Examples on how to set up and bind a server implementation are covered in the Dart tutorial.



The TicTacToeProxy class automatically implements the onOpponentMove getter. Clients obtain an async.Stream of onOpponentMove events sent from the server using this getter.


Servers send events by implementing the onOpponentMove getter on the abstract TicTacToe class. A TicTacToeBinding (see tutorial) that is bound to an instance of TicTacToe that has implemented the onOpponentMove getter will listen for events on the returned async.Stream, forwarding them to the client.


Given the method with an error type:

protocol TicTacToe {
    MakeMove(struct {
      row uint8;
      col uint8;
    }) -> (struct {
      new_state GameState;
    }) error MoveError;

The method signature for MakeMove on the generated abstract TicTacToe class is:

async.Future<GameState> makeMove(int row, int col)

The encapsulated Future corresponds to the generated response type for the success case, and the error case is represented by having the server implementation or the proxy class throw a fidl.MethodException.

Using this feature, an example implementation of MakeMove on the server side could look like:

async.Future<GameState> makeMove(int row, int col) {
  if (row > 9 || col > 9) {
    return async.Future.error(fidl.MethodException(MoveError.OutOfBounds));
  return async.Future.value(doSomething(row, col));

The TicTacToeBinding class will catch fidl.MethodExceptions and encode it as an error.

An example of using this on the client side would be:

myproxy.makeMove(1, 2).then((gameState) { ... })
                      .catchError((moveError) { ... });

Unknown interaction handling


When a protocol is declared as open or ajar, the backend will generate a TicTacToeServer class which inherits from the TicTacToe class. The server interface will add a single method to the base interface, called $unknownMethod, with this signature:

Future<void> $unknownMethod(fidl.UnknownMethodMetadata metadata);

This method will be called whenever the server receives a flexible unknown interaction which it can handle, that is a flexible one-way method in the case of an ajar protocol, or any flexible method in the case of an open protocol. The argument is a class that holds basic information about the unknown method that was received:

/// Metadata about an unknown flexible method that was received.
class UnknownMethodMetadata {
  UnknownMethodMetadata(this.ordinal, this.unknownMethodType);

  /// Ordinal of the method.
  final int ordinal;

  /// Type of the unknown method.
  /// For an ajar protocol, this will always be oneWay.
  final UnknownMethodType unknownMethodType;

UnknownMethodType is an enum with two variants, oneWay and twoWay. If the protocol is ajar, the unknownMethodType will always be oneWay, since two-way unknown methods cannot be handled.


There is no way for the client to tell if a flexible one-way method was known to the server or not. For flexible two-way methods, if the method is not known to the server, fidl.UnknownMethodException will be thrown, which is a subclass of FidlError which has the error code FidlErrorCode.fidlUnknownMethod.

Aside from the possibility of getting an UnknownMethodException, there are no API difference between strict and flexible methods on the client.

For open and ajar protocols the generated TicTacToeProxy class will have an additional field, called $unknownEvents:

Stream<UnknownEvent> get $unknownEvents;

This stream will emit an UnknownEvent whenever the client receives an unknown flexible event from the server. UnknownEvent is a class that holds information about the event that was received:

/// Event used when an unknown, flexible event is received.
class UnknownEvent {

  /// Ordinal of the event.
  final int ordinal;

Protocol composition

FIDL does not have a concept of inheritance, and generates full code as described above for all composed protocols. In other words, the code generated for:

protocol A {

protocol B {
    compose A;

Provides the same API as the code generated for:

protocol A {

protocol B {

The generated code is identical except for the method ordinals.

Protocol and method attributes


For protocol methods annotated with the @transitional attribute, the FIDL toolchain generates a default implementation on the abstract class so that server implementations will continue to compile without having to override the new method.


The generated class for a protocol annotated with the @discoverable attribute has a non-null $serviceName field.

Test scaffolding

The FIDL toolchain generates a fidl_test.dart file that contains convenience code for testing FIDL server implementations. This file contains a class for each protocol that provides stub implementations for each of the class’s methods, making it possible to implement only the methods that are used during testing.

Given the example protocol above, The FIDL toolchain generates a TicTacToe$TestBase class that extends the TicTacToe abstract class. All methods are implemented by returning async.Future.error(UnimplementedError()), and all events are implemented by returning a Stream with a single UnimplementedError event.