Inbox stop (#50)

* added process status stopped

* try recover when panic on process start

* add condition in for loop

* Atomically check inbox stopped status + test

---------

Co-authored-by: Terri Prifti <[email protected]>

actor/engine_test.go

@@ -48,33 +48,28 @@ func TestSendRepeat(t *testing.T) {
 	repeater.Stop()
 }
 
-func TestRestarts(t *testing.T) {
+func TestRestartsMaxRestarts(t *testing.T) {
 	e := NewEngine()
-	wg := sync.WaitGroup{}
+	restarts := 2
 	type payload struct {
 		data int
 	}
-
-	wg.Add(1)
 	pid := e.SpawnFunc(func(c *Context) {
 		switch msg := c.Message().(type) {
 		case Started:
 		case Stopped:
-			fmt.Println("stopped!")
 		case payload:
 			if msg.data != 10 {
 				panic("I failed to process this message")
 			} else {
 				fmt.Println("finally processed all my messsages after borking.", msg.data)
-				wg.Done()
 			}
 		}
-	}, "foo", WithRestartDelay(time.Millisecond*10))
+	}, "foo", WithMaxRestarts(restarts))
 
-	e.Send(pid, payload{1})
-	e.Send(pid, payload{2})
-	e.Send(pid, payload{10})
-	wg.Wait()
+	for i := 0; i < 11; i++ {
+		e.Send(pid, payload{i})
+	}
 }
 
 func TestProcessInitStartOrder(t *testing.T) {
@@ -103,6 +98,35 @@ func TestProcessInitStartOrder(t *testing.T) {
 	wg.Wait()
 }
 
+func TestRestarts(t *testing.T) {
+	e := NewEngine()
+	wg := sync.WaitGroup{}
+	type payload struct {
+		data int
+	}
+
+	wg.Add(1)
+	pid := e.SpawnFunc(func(c *Context) {
+		switch msg := c.Message().(type) {
+		case Started:
+		case Stopped:
+			fmt.Println("stopped!")
+		case payload:
+			if msg.data != 10 {
+				panic("I failed to process this message")
+			} else {
+				fmt.Println("finally processed all my messsages after borking.", msg.data)
+				wg.Done()
+			}
+		}
+	}, "foo", WithRestartDelay(time.Millisecond*10))
+
+	e.Send(pid, payload{1})
+	e.Send(pid, payload{2})
+	e.Send(pid, payload{10})
+	wg.Wait()
+}
+
 func TestSendWithSender(t *testing.T) {
 	var (
 		e      = NewEngine()

actor/inbox.go

@@ -12,6 +12,7 @@ const defaultThroughput = 300
 const (
 	idle int32 = iota
 	running
+	stopped
 )
 
 type Scheduler interface {
@@ -71,7 +72,7 @@ func (in *Inbox) process() {
 
 func (in *Inbox) run() {
 	i, t := 0, in.scheduler.Throughput()
-	for {
+	for atomic.LoadInt32(&in.procStatus) != stopped {
 		if i > t {
 			i = 0
 			runtime.Gosched()
@@ -91,5 +92,6 @@ func (in *Inbox) Start(proc Processer) {
 }
 
 func (in *Inbox) Stop() error {
+	atomic.StoreInt32(&in.procStatus, stopped)
 	return nil
 }

actor/process.go

@@ -73,9 +73,7 @@ func (p *process) Invoke(msgs []Envelope) {
 			for i := 0; i < nmsg-nproc; i++ {
 				p.mbuffer[i] = msgs[i+nproc]
 			}
-			if p.Opts.MaxRestarts > 0 {
-				p.tryRestart(v)
-			}
+			p.tryRestart(v)
 		}
 	}()
 	for i := 0; i < len(msgs); i++ {
@@ -99,6 +97,13 @@ func (p *process) Invoke(msgs []Envelope) {
 func (p *process) Start() {
 	recv := p.Producer()
 	p.context.receiver = recv
+	defer func() {
+		if v := recover(); v != nil {
+			p.context.message = Stopped{}
+			p.context.receiver.Receive(p.context)
+			p.tryRestart(v)
+		}
+	}()
 	p.context.message = Initialized{}
 	applyMiddleware(recv.Receive, p.Opts.Middleware...)(p.context)
 
@@ -118,7 +123,6 @@ func (p *process) Start() {
 }
 
 func (p *process) tryRestart(v any) {
-	p.restarts++
 	// InternalError does not take the maximum restarts into account.
 	// For now, InternalError is getting triggered when we are dialing
 	// a remote node. By doing this, we can keep dialing until it comes
@@ -144,6 +148,8 @@ func (p *process) tryRestart(v any) {
 		p.cleanup(nil)
 		return
 	}
+
+	p.restarts++
 	// Restart the process after its restartDelay
 	log.Errorw("[PROCESS] actor restarting", log.M{
 		"n":           p.restarts,