Update dependencies and remove old matterclient lib (#2067)

vendor/github.com/lrstanley/girc/conn.go

@@ -12,6 +12,8 @@ import (
 	"net"
 	"sync"
 	"time"
+
+	"github.com/lrstanley/girc/internal/ctxgroup"
 )
 
 // Messages are delimited with CR and LF line endings, we're using the last
@@ -142,17 +144,44 @@ type ErrParseEvent struct {
 
 func (e ErrParseEvent) Error() string { return "unable to parse event: " + e.Line }
 
-func (c *ircConn) decode() (event *Event, err error) {
-	line, err := c.io.ReadString(delim)
-	if err != nil {
-		return nil, err
-	}
+type decodedEvent struct {
+	event *Event
+	err   error
+}
 
-	if event = ParseEvent(line); event == nil {
-		return nil, ErrParseEvent{line}
-	}
+func (c *ircConn) decode() <-chan decodedEvent {
+	ch := make(chan decodedEvent)
 
-	return event, nil
+	go func() {
+		defer close(ch)
+
+		line, err := c.io.ReadString(delim)
+		if err != nil {
+			select {
+			case ch <- decodedEvent{err: err}:
+			default:
+			}
+
+			return
+		}
+
+		event := ParseEvent(line)
+		if event == nil {
+			select {
+			case ch <- decodedEvent{err: ErrParseEvent{Line: line}}:
+			default:
+			}
+
+			return
+		}
+
+		select {
+		case ch <- decodedEvent{event: event}:
+		default:
+		}
+	}()
+
+	return ch
 }
 
 func (c *ircConn) encode(event *Event) error {
@@ -291,20 +320,17 @@ startConn:
 	} else {
 		c.conn = newMockConn(mock)
 	}
+	c.mu.Unlock()
 
 	var ctx context.Context
 	ctx, c.stop = context.WithCancel(context.Background())
-	c.mu.Unlock()
 
-	errs := make(chan error, 4)
-	var wg sync.WaitGroup
-	// 4 being the number of goroutines we need to finish when this function
-	// returns.
-	wg.Add(4)
-	go c.execLoop(ctx, errs, &wg)
-	go c.readLoop(ctx, errs, &wg)
-	go c.sendLoop(ctx, errs, &wg)
-	go c.pingLoop(ctx, errs, &wg)
+	group := ctxgroup.New(ctx)
+
+	group.Go(c.execLoop)
+	group.Go(c.readLoop)
+	group.Go(c.sendLoop)
+	group.Go(c.pingLoop)
 
 	// Passwords first.
 
@@ -338,16 +364,15 @@ startConn:
 	c.RunHandlers(&Event{Command: INITIALIZED, Params: []string{addr}})
 
 	// Wait for the first error.
-	var result error
-	select {
-	case <-ctx.Done():
+	err := group.Wait()
+	if err != nil {
+		c.debug.Printf("received error, beginning cleanup: %v", err)
+	} else {
 		if !c.state.sts.beginUpgrade {
 			c.debug.Print("received request to close, beginning clean up")
 		}
+
 		c.RunHandlers(&Event{Command: CLOSED, Params: []string{addr}})
-	case err := <-errs:
-		c.debug.Printf("received error, beginning cleanup: %v", err)
-		result = err
 	}
 
 	// Make sure that the connection is closed if not already.
@@ -363,20 +388,13 @@ startConn:
 
 	c.RunHandlers(&Event{Command: DISCONNECTED, Params: []string{addr}})
 
-	// Once we have our error/result, let all other functions know we're done.
-	c.debug.Print("waiting for all routines to finish")
-
-	// Wait for all goroutines to finish.
-	wg.Wait()
-	close(errs)
-
 	// This helps ensure that the end user isn't improperly using the client
 	// more than once. If they want to do this, they should be using multiple
 	// clients, not multiple instances of Connect().
 	c.mu.Lock()
 	c.conn = nil
 
-	if result == nil {
+	if err == nil {
 		if c.state.sts.beginUpgrade {
 			c.state.sts.beginUpgrade = false
 			c.mu.Unlock()
@@ -389,76 +407,85 @@ startConn:
 	}
 	c.mu.Unlock()
 
-	return result
+	return err
 }
 
 // readLoop sets a timeout of 300 seconds, and then attempts to read from the
 // IRC server. If there is an error, it calls Reconnect.
-func (c *Client) readLoop(ctx context.Context, errs chan error, wg *sync.WaitGroup) {
+func (c *Client) readLoop(ctx context.Context) error {
 	c.debug.Print("starting readLoop")
 	defer c.debug.Print("closing readLoop")
 
-	var event *Event
-	var err error
+	var de decodedEvent
 
 	for {
 		select {
 		case <-ctx.Done():
-			wg.Done()
-			return
+			return nil
 		default:
 			_ = c.conn.sock.SetReadDeadline(time.Now().Add(300 * time.Second))
-			event, err = c.conn.decode()
-			if err != nil {
-				errs <- err
-				wg.Done()
-				return
+
+			select {
+			case <-ctx.Done():
+				return nil
+			case de = <-c.conn.decode():
+			}
+
+			if de.err != nil {
+				return de.err
 			}
 
 			// Check if it's an echo-message.
 			if !c.Config.disableTracking {
-				event.Echo = (event.Command == PRIVMSG || event.Command == NOTICE) &&
-					event.Source != nil && event.Source.ID() == c.GetID()
+				de.event.Echo = (de.event.Command == PRIVMSG || de.event.Command == NOTICE) &&
+					de.event.Source != nil && de.event.Source.ID() == c.GetID()
 			}
 
-			c.rx <- event
+			c.receive(de.event)
 		}
 	}
 }
 
-// Send sends an event to the server. Use Client.RunHandlers() if you are
-// simply looking to trigger handlers with an event.
+// Send sends an event to the server. Send will split events if the event is longer
+// than what the server supports, and is an event that supports splitting. Use
+// Client.RunHandlers() if you are simply looking to trigger handlers with an event.
 func (c *Client) Send(event *Event) {
 	var delay time.Duration
 
-	if !c.Config.AllowFlood {
-		c.mu.RLock()
-
-		// Drop the event early as we're disconnected, this way we don't have to wait
-		// the (potentially long) rate limit delay before dropping.
-		if c.conn == nil {
-			c.debugLogEvent(event, true)
-			c.mu.RUnlock()
-			return
-		}
-
-		c.conn.mu.Lock()
-		delay = c.conn.rate(event.Len())
-		c.conn.mu.Unlock()
-		c.mu.RUnlock()
-	}
-
 	if c.Config.GlobalFormat && len(event.Params) > 0 && event.Params[len(event.Params)-1] != "" &&
 		(event.Command == PRIVMSG || event.Command == TOPIC || event.Command == NOTICE) {
 		event.Params[len(event.Params)-1] = Fmt(event.Params[len(event.Params)-1])
 	}
 
-	<-time.After(delay)
-	c.write(event)
+	var events []*Event
+	events = event.split(c.MaxEventLength())
+
+	for _, e := range events {
+		if !c.Config.AllowFlood {
+			c.mu.RLock()
+
+			// Drop the event early as we're disconnected, this way we don't have to wait
+			// the (potentially long) rate limit delay before dropping.
+			if c.conn == nil {
+				c.debugLogEvent(e, true)
+				c.mu.RUnlock()
+				return
+			}
+
+			c.conn.mu.Lock()
+			delay = c.conn.rate(e.Len())
+			c.conn.mu.Unlock()
+			c.mu.RUnlock()
+		}
+
+		<-time.After(delay)
+		c.write(e)
+	}
 }
 
 // write is the lower level function to write an event. It does not have a
-// write-delay when sending events.
+// write-delay when sending events. write will timeout after 30s if the event
+// can't be sent.
 func (c *Client) write(event *Event) {
 	c.mu.RLock()
 	defer c.mu.RUnlock()
@@ -468,7 +495,19 @@ func (c *Client) write(event *Event) {
 		c.debugLogEvent(event, true)
 		return
 	}
-	c.tx <- event
+
+	t := time.NewTimer(30 * time.Second)
+	defer func() {
+		if !t.Stop() {
+			<-t.C
+		}
+	}()
+
+	select {
+	case c.tx <- event:
+	case <-t.C:
+		c.debugLogEvent(event, true)
+	}
 }
 
 // rate allows limiting events based on how frequent the event is being sent,
@@ -487,7 +526,7 @@ func (c *ircConn) rate(chars int) time.Duration {
 	return 0
 }
 
-func (c *Client) sendLoop(ctx context.Context, errs chan error, wg *sync.WaitGroup) {
+func (c *Client) sendLoop(ctx context.Context) error {
 	c.debug.Print("starting sendLoop")
 	defer c.debug.Print("closing sendLoop")
 
@@ -537,18 +576,14 @@ func (c *Client) sendLoop(ctx context.Context, errs chan error, wg *sync.WaitGro
 
 			if event.Command == QUIT {
 				c.Close()
-				wg.Done()
-				return
+				return nil
 			}
 
 			if err != nil {
-				errs <- err
-				wg.Done()
-				return
+				return err
 			}
 		case <-ctx.Done():
-			wg.Done()
-			return
+			return nil
 		}
 	}
 }
@@ -568,11 +603,10 @@ type ErrTimedOut struct {
 
 func (ErrTimedOut) Error() string { return "timed out waiting for a requested PING response" }
 
-func (c *Client) pingLoop(ctx context.Context, errs chan error, wg *sync.WaitGroup) {
+func (c *Client) pingLoop(ctx context.Context) error {
 	// Don't run the pingLoop if they want to disable it.
 	if c.Config.PingDelay <= 0 {
-		wg.Done()
-		return
+		return nil
 	}
 
 	c.debug.Print("starting pingLoop")
@@ -604,9 +638,8 @@ func (c *Client) pingLoop(ctx context.Context, errs chan error, wg *sync.WaitGro
 			}
 
 			c.conn.mu.RLock()
-			if pingSent && time.Since(c.conn.lastPong) > c.Config.PingDelay+(60*time.Second) {
-				// It's 60 seconds over what out ping delay is, connection
-				// has probably dropped.
+			if pingSent && time.Since(c.conn.lastPong) > c.Config.PingDelay+c.Config.PingTimeout {
+				// PingTimeout exceeded, connection has probably dropped.
 				err := ErrTimedOut{
 					TimeSinceSuccess: time.Since(c.conn.lastPong),
 					LastPong:         c.conn.lastPong,
@@ -615,9 +648,7 @@ func (c *Client) pingLoop(ctx context.Context, errs chan error, wg *sync.WaitGro
 				}
 
 				c.conn.mu.RUnlock()
-				errs <- err
-				wg.Done()
-				return
+				return err
 			}
 			c.conn.mu.RUnlock()
 
@@ -628,8 +659,7 @@ func (c *Client) pingLoop(ctx context.Context, errs chan error, wg *sync.WaitGro
 			c.Cmd.Ping(fmt.Sprintf("%d", time.Now().UnixNano()))
 			pingSent = true
 		case <-ctx.Done():
-			wg.Done()
-			return
+			return nil
 		}
 	}
 }