[client] Fix engine restart (#3435)

- Refactor the network monitoring to handle one event and it after return
- In the engine restart cancel the upper layer context and the responsibility of the engine stop will be the upper layer
- Before triggering a restart, the engine checks whether the state is already down. This helps avoid unnecessary delayed network restart events.

client/internal/connect.go

@@ -161,7 +161,7 @@ func (c *ConnectClient) run(mobileDependency MobileDependency, runningChan chan
 	defer c.statusRecorder.ClientStop()
 	operation := func() error {
 		// if context cancelled we not start new backoff cycle
-		if c.isContextCancelled() {
+		if c.ctx.Err() != nil {
 			return nil
 		}
 
@@ -379,15 +379,6 @@ func (c *ConnectClient) Stop() error {
 	return nil
 }
 
-func (c *ConnectClient) isContextCancelled() bool {
-	select {
-	case <-c.ctx.Done():
-		return true
-	default:
-		return false
-	}
-}
-
 // SetNetworkMapPersistence enables or disables network map persistence.
 // When enabled, the last received network map will be stored and can be retrieved
 // through the Engine's getLatestNetworkMap method. When disabled, any stored