refactor: refactor supervisor strategy

actors/actor_system.go

@@ -163,7 +163,7 @@ type actorSystem struct {
 	// The default value is 1s
 	actorInitTimeout time.Duration
 	// Specifies the supervisor strategy
-	supervisorStrategy StrategyDirective
+	supervisorDirective supervisorDirective
 	// Specifies the telemetry config
 	telemetry *telemetry.Telemetry
 	// Specifies whether remoting is enabled.
@@ -240,7 +240,7 @@ func NewActorSystem(name string, opts ...Option) (ActorSystem, error) {
 		expireActorAfter:       DefaultPassivationTimeout,
 		askTimeout:             DefaultAskTimeout,
 		actorInitMaxRetries:    DefaultInitMaxRetries,
-		supervisorStrategy:     DefaultSupervisoryStrategy,
+		supervisorDirective:    DefaultSupervisoryStrategy,
 		telemetry:              telemetry.New(),
 		locker:                 sync.Mutex{},
 		shutdownTimeout:        DefaultShutdownTimeout,
@@ -437,7 +437,7 @@ func (x *actorSystem) Spawn(ctx context.Context, name string, actor Actor) (PID,
 		withAskTimeout(x.askTimeout),
 		withCustomLogger(x.logger),
 		withActorSystem(x),
-		withSupervisorStrategy(x.supervisorStrategy),
+		withSupervisorDirective(x.supervisorDirective),
 		withMailboxSize(x.mailboxSize),
 		withMailbox(mailbox), // nil mailbox is taken care during initiliazation by the newPID
 		withStash(x.stashCapacity),
@@ -490,7 +490,7 @@ func (x *actorSystem) SpawnFromFunc(ctx context.Context, receiveFunc ReceiveFunc
 	}
 
 	actorID := uuid.NewString()
-	actor := newFnActor(actorID, receiveFunc, opts...)
+	actor := newFuncActor(actorID, receiveFunc, opts...)
 
 	actorPath := NewPath(actorID, NewAddress(x.name, "", -1))
 	if x.remotingEnabled.Load() {
@@ -511,7 +511,7 @@ func (x *actorSystem) SpawnFromFunc(ctx context.Context, receiveFunc ReceiveFunc
 		withAskTimeout(x.askTimeout),
 		withCustomLogger(x.logger),
 		withActorSystem(x),
-		withSupervisorStrategy(x.supervisorStrategy),
+		withSupervisorDirective(x.supervisorDirective),
 		withMailboxSize(x.mailboxSize),
 		withMailbox(mailbox), // nil mailbox is taken care during initiliazation by the newPID
 		withStash(x.stashCapacity),

actors/actor_test.go

@@ -147,22 +147,24 @@ func newSupervised() *supervised {
 	return &supervised{}
 }
 
-func (c *supervised) PreStart(context.Context) error {
+func (x *supervised) PreStart(context.Context) error {
 	return nil
 }
 
-func (c *supervised) Receive(ctx ReceiveContext) {
+func (x *supervised) Receive(ctx ReceiveContext) {
 	switch ctx.Message().(type) {
 	case *goaktpb.PostStart:
 	case *testspb.TestSend:
+	case *testspb.TestReply:
+		ctx.Response(new(testspb.Reply))
 	case *testspb.TestPanic:
 		panic("panicked")
 	default:
 		panic(ErrUnhandled)
 	}
 }
 
-func (c *supervised) PostStop(context.Context) error {
+func (x *supervised) PostStop(context.Context) error {
 	return nil
 }
 
@@ -491,3 +493,7 @@ func startClusterSystem(t *testing.T, nodeName, serverAddr string) (ActorSystem,
 	// return the cluster startNode
 	return system, provider
 }
+
+type unhandledSupervisorDirective struct{}
+
+func (x unhandledSupervisorDirective) isSupervisorDirective() {}

actors/func_actor.go

@@ -80,8 +80,8 @@ type funcActor struct {
 	postStop    PostStopFunc
 }
 
-// newFnActor creates an instance of funcActor
-func newFnActor(id string, receiveFunc ReceiveFunc, opts ...FuncOption) Actor {
+// newFuncActor creates an instance of funcActor
+func newFuncActor(id string, receiveFunc ReceiveFunc, opts ...FuncOption) Actor {
 	// create the actor instance
 	actor := &funcActor{
 		receiveFunc: receiveFunc,
@@ -98,7 +98,7 @@ func newFnActor(id string, receiveFunc ReceiveFunc, opts ...FuncOption) Actor {
 var _ Actor = (*funcActor)(nil)
 
 // PreStart pre-starts the actor.
-func (x funcActor) PreStart(ctx context.Context) error {
+func (x *funcActor) PreStart(ctx context.Context) error {
 	// check whether the pre-start hook is set and call it
 	preStart := x.preStart
 	if preStart != nil {
@@ -108,7 +108,7 @@ func (x funcActor) PreStart(ctx context.Context) error {
 }
 
 // Receive processes any message dropped into the actor mailbox.
-func (x funcActor) Receive(ctx ReceiveContext) {
+func (x *funcActor) Receive(ctx ReceiveContext) {
 	switch m := ctx.Message().(type) {
 	case *goaktpb.PostStart:
 		x.pid = ctx.Self()
@@ -121,7 +121,7 @@ func (x funcActor) Receive(ctx ReceiveContext) {
 }
 
 // PostStop is executed when the actor is shutting down.
-func (x funcActor) PostStop(ctx context.Context) error {
+func (x *funcActor) PostStop(ctx context.Context) error {
 	// check whether the pre-start hook is set and call it
 	postStop := x.postStop
 	if postStop != nil {

actors/option.go

@@ -86,10 +86,10 @@ func WithPassivationDisabled() Option {
 	})
 }
 
-// WithSupervisorStrategy sets the supervisor strategy
-func WithSupervisorStrategy(strategy StrategyDirective) Option {
+// WithSupervisorDirective sets the supervisor strategy directive
+func WithSupervisorDirective(directive supervisorDirective) Option {
 	return OptionFunc(func(a *actorSystem) {
-		a.supervisorStrategy = strategy
+		a.supervisorDirective = directive
 	})
 }
 

actors/option_test.go

@@ -39,6 +39,7 @@ import (
 func TestOption(t *testing.T) {
 	tel := telemetry.New()
 	mailbox := newReceiveContextBuffer(10)
+	resumeDirective := NewResumeDirective()
 	var atomicTrue atomic.Bool
 	atomicTrue.Store(true)
 	clusterConfig := NewClusterConfig()
@@ -74,9 +75,9 @@ func TestOption(t *testing.T) {
 			expected: actorSystem{expireActorAfter: -1},
 		},
 		{
-			name:     "WithSupervisorStrategy",
-			option:   WithSupervisorStrategy(RestartDirective),
-			expected: actorSystem{supervisorStrategy: RestartDirective},
+			name:     "WithSupervisorDirective",
+			option:   WithSupervisorDirective(resumeDirective),
+			expected: actorSystem{supervisorDirective: resumeDirective},
 		},
 		{
 			name:     "WithRemoting",

actors/pid.go

@@ -249,7 +249,7 @@ type pid struct {
 	processingTimeLocker *sync.Mutex
 
 	// supervisor strategy
-	supervisorStrategy StrategyDirective
+	supervisorDirective supervisorDirective
 
 	// observability settings
 	telemetry *telemetry.Telemetry
@@ -291,7 +291,7 @@ func newPID(ctx context.Context, actorPath *Path, actor Actor, opts ...pidOption
 		logger:               log.DefaultLogger,
 		mailboxSize:          DefaultMailboxSize,
 		children:             newPIDMap(10),
-		supervisorStrategy:   DefaultSupervisoryStrategy,
+		supervisorDirective:  DefaultSupervisoryStrategy,
 		watchersList:         slices.NewConcurrentSlice[*watcher](),
 		telemetry:            telemetry.New(),
 		actorPath:            actorPath,
@@ -558,7 +558,7 @@ func (x *pid) SpawnChild(ctx context.Context, name string, actor Actor) (PID, er
 		withAskTimeout(x.askTimeout.Load()),
 		withCustomLogger(x.logger),
 		withActorSystem(x.system),
-		withSupervisorStrategy(x.supervisorStrategy),
+		withSupervisorDirective(x.supervisorDirective),
 		withMailboxSize(x.mailboxSize),
 		withStash(x.stashCapacity.Load()),
 		withMailbox(x.mailbox.Clone()),
@@ -1326,45 +1326,6 @@ func (x *pid) handleReceived(received ReceiveContext) {
 	}
 }
 
-// supervise watches for child actor's failure and act based upon the supervisory strategy
-func (x *pid) supervise(cid PID, watcher *watcher) {
-	for {
-		select {
-		case <-watcher.Done:
-			x.logger.Debugf("stop watching cid=(%s)", cid.ActorPath().String())
-			return
-		case err := <-watcher.ErrChan:
-			x.logger.Errorf("child actor=(%s) is failing: Err=%v", cid.ActorPath().String(), err)
-			switch x.supervisorStrategy {
-			case StopDirective:
-				x.UnWatch(cid)
-				x.children.delete(cid.ActorPath())
-				if err := cid.Shutdown(context.Background()); err != nil {
-					// this can enter into some infinite loop if we panic
-					// since we are just shutting down the actor we can just log the error
-					// TODO: rethink properly about PostStop error handling
-					x.logger.Error(err)
-				}
-			case RestartDirective:
-				x.UnWatch(cid)
-				if err := cid.Restart(context.Background()); err != nil {
-					x.logger.Panic(err)
-				}
-				x.Watch(cid)
-			default:
-				x.UnWatch(cid)
-				x.children.delete(cid.ActorPath())
-				if err := cid.Shutdown(context.Background()); err != nil {
-					// this can enter into some infinite loop if we panic
-					// since we are just shutting down the actor we can just log the error
-					// TODO: rethink properly about PostStop error handling
-					x.logger.Error(err)
-				}
-			}
-		}
-	}
-}
-
 // passivationListener checks whether the actor is processing public or not.
 // when the actor is idle, it automatically shuts down to free resources
 func (x *pid) passivationListener() {

actors/pid_option.go

@@ -71,11 +71,11 @@ func withActorSystem(sys ActorSystem) pidOption {
 	}
 }
 
-// withSupervisorStrategy sets the supervisor strategy to used when dealing
+// withSupervisorDirective sets the supervisor strategy to used when dealing
 // with child actors
-func withSupervisorStrategy(strategy StrategyDirective) pidOption {
+func withSupervisorDirective(directive supervisorDirective) pidOption {
 	return func(pid *pid) {
-		pid.supervisorStrategy = strategy
+		pid.supervisorDirective = directive
 	}
 }
 

actors/pid_option_test.go

@@ -37,6 +37,7 @@ import (
 
 func TestPIDOptions(t *testing.T) {
 	mailbox := newReceiveContextBuffer(10)
+	resumeDirective := NewResumeDirective()
 	var (
 		atomicDuration   atomic.Duration
 		atomicInt        atomic.Int32
@@ -78,8 +79,8 @@ func TestPIDOptions(t *testing.T) {
 		},
 		{
 			name:     "WithSupervisorStrategy",
-			option:   withSupervisorStrategy(RestartDirective),
-			expected: &pid{supervisorStrategy: RestartDirective},
+			option:   withSupervisorDirective(resumeDirective),
+			expected: &pid{supervisorDirective: resumeDirective},
 		},
 		{
 			name:     "WithShutdownTimeout",