1 Concurrency in JavaFX

The Worker Interface

The Worker interface defines an object that performs some work on one or more background threads. The state of the Worker object is observable and usable from the JavaFX Application thread.

The lifecycle of the Worker object is defined as follows. When created, the Worker object is in the READY state. Upon being scheduled for work, the Worker object transitions to the SCHEDULED state. After that, when the Worker object is performing the work, its state becomes RUNNING. Note that even when the Worker object is immediately started without being scheduled, it first transitions to the SCHEDULED state and then to the RUNNING state. The state of a Worker object that completes successfully is SUCCEEDED, and the value property is set to the result of this Worker object. Otherwise, if any exceptions are thrown during the execution of the Worker object, its state becomes FAILED and the exception property is set to the type of the exception that occurred. At any time before the end of the Worker object the developer can interrupt it by invoking the cancel method, which puts the Worker object into the CANCELLED state.

Distinctions in the lifecycle of a ScheduledService object can be found in the The ScheduledService Class section.

The progress of the work being done by the Worker object can be obtained through three different properties such as totalWork, workDone, and progress.

For more information on the range of the parameter values, see the API documentation.

The Task Class

Tasks are used to implement the logic of work that needs to be done on a background thread. First, you need to extend the Task class. Your implementation of the Task class must override the call method to do the background work and return the result.

The call method is invoked on the background thread, therefore this method can only manipulate states that are safe to read and write from a background thread. For example, manipulating an active scene graph from the call method throws runtime exceptions. On the other hand, the Task class is designed to be used with JavaFX GUI applications, and it ensures that any changes to public properties, change notifications for errors or cancellation, event handlers, and states occur on the JavaFX Application thread. Inside the call method, you can use the updateProgress, updateMessage, updateTitle methods, which update the values of the corresponding properties on the JavaFX Application thread. However, if the task was canceled, a return value from the call method is ignored.

Note that the Task class fits into the Java concurrency libraries because it inherits from the java.utils.concurrent.FutureTask class, which implements the Runnable interface. For this reason, a Task object can be used within the Java concurrency Executor API and also can be passed to a thread as a parameter. You can call the Task object directly by using the FutureTask.run() method, which enables calling this task from another background thread. Having a good understanding of the Java concurrency API will help you understand concurrency in JavaFX.

A task can be started in one of the following ways:

• By starting a thread with the given task as a parameter:

  Thread th = new Thread(task);

  th.setDaemon(true);

  th.start();

• By using the ExecutorService API:

  ExecutorService.submit(task);

The Task class defines a one-time object that cannot be reused. If you need a reusable Worker object, use the Service class.

import javafx.concurrent.Task;

Task<Integer> task = new Task<Integer>() {
    @Override protected Integer call() throws Exception {
        int iterations;
        for (iterations = 0; iterations < 100000; iterations++) {
            if (isCancelled()) {
               break;
            }
            System.out.println("Iteration " + iterations);
        }
        return iterations;
    }
};

import javafx.concurrent.Task;

Task<Integer> task = new Task<Integer>() {
    @Override protected Integer call() throws Exception {
        int iterations;
        for (iterations = 0; iterations < 1000; iterations++) {
            if (isCancelled()) {
                updateMessage("Cancelled");
                break;
            }
            updateMessage("Iteration " + iterations);
            updateProgress(iterations, 1000);
 
            //Block the thread for a short time, but be sure
            //to check the InterruptedException for cancellation
            try {
                Thread.sleep(100);
            } catch (InterruptedException interrupted) {
                if (isCancelled()) {
                    updateMessage("Cancelled");
                    break;
                }
            }
        }
        return iterations;
    }
};

import javafx.concurrent.Task;

Task task = new Task<Void>() {
    @Override public Void call() {
        static final int max = 1000000;
        for (int i=1; i<=max; i++) {
            if (isCancelled()) {
               break;
            }
            updateProgress(i, max);
        }
        return null;
    }
};
ProgressBar bar = new ProgressBar();
bar.progressProperty().bind(task.progressProperty());
new Thread(task).start();

The Service Class

The Service class is designed to execute a Task object on one or several background threads. The Service class methods and states must only be accessed on the JavaFX Application thread. The purpose of this class is to help the developer to implement the correct interaction between the background threads and the JavaFX Application thread.

You have the following control over the Service object: you can start, cancel and restart it as you need. To start the Service object, use the Service.start() method.

Using the Service class, you can observe the state of the background work and optionally cancel it. Later, you can reset the service and restart it. Thus, the service can be defined declaratively and restarted on demand.

For a service that needs to be automatically restarted, see The ScheduledService Class section.

When implementing the subclasses of the Service class, be sure to expose the input parameters to the Task object as properties of the subclass.

The service can be executed in one of the following ways:

• By an Executor object, if it is specified for the given service

• By a daemon thread, if no executor is specified

• By a custom executor such as a ThreadPoolExecutor

Example 1-4 shows an implementation of the Service class which reads the first line from any URL and returns it as a string.

import java.io.BufferedReader;
import java.io.IOException;
import java.io.InputStreamReader;
import java.net.MalformedURLException;
import java.net.URL;
import javafx.application.Application;
import javafx.beans.property.SimpleStringProperty;
import javafx.beans.property.StringProperty;
import javafx.concurrent.Service;
import javafx.concurrent.Task;
import javafx.concurrent.WorkerStateEvent;
import javafx.event.EventHandler;
import javafx.stage.Stage;

public class FirstLineServiceApp extends Application {

    @Override
    public void start(Stage stage) throws Exception {
        FirstLineService service = new FirstLineService();
        service.setUrl("http://google.com");
        service.setOnSucceeded(new EventHandler<WorkerStateEvent>() {

            @Override
            public void handle(WorkerStateEvent t) {
                System.out.println("done:" + t.getSource().getValue());
            }
        });
        service.start();
    }

    public static void main(String[] args) {
        launch();
    }

    private static class FirstLineService extends Service<String> {
        private StringProperty url = new SimpleStringProperty();

        public final void setUrl(String value) {
            url.set(value);
        }

        public final String getUrl() {
            return url.get();
        }

        public final StringProperty urlProperty() {
           return url;
        }


        @Override
        protected Task<String> createTask() {
            return new Task<String>() {
                @Override
                protected String call()
                        throws IOException, MalformedURLException {
                    try ( BufferedReader in = new BufferedReader(
                                new InputStreamReader(
                                    new URL(getUrl()).openStream;
                            in = new BufferedReader(
                                new InputStreamReader(u.openStream()))) {
                        return in.readLine();
                    }
                }
        };
    }
}

The WorkerStateEvent Class and State Transitions

Whenever the state of the Worker implementation changes, an appropriate event, defined by the WorkerStateEvent class, occurs. For example, when the Task object transitions to the SUCCEEDED state, the WORKER_STATE_SUCCEEDED event occurs, the onSucceeded event handler is called, after which the protected convenience method succeeded is invoked on the JavaFX Application thread.

There are several protected convenience methods such as cancelled, failed, running, scheduled, and succeeded, which are invoked when the Worker implementation transitions to the corresponding state. These methods can be overridden by subclasses of the Task and Service classes when the state is changed to implement the logic of your application. Example 1-5 shows a Task implementation that updates the status message on the task's success, cancellation, and failure.

import javafx.concurrent.Task;

Task<Integer> task = new Task<Integer>() {
    @Override protected Integer call() throws Exception {
        int iterations = 0;
        for (iterations = 0; iterations < 100000; iterations++) {
            if (isCancelled()) {
                break;
            }
            System.out.println("Iteration " + iterations);
        }
        return iterations;
    }

    @Override protected void succeeded() {
        super.succeeded();
        updateMessage("Done!");
    }

    @Override protected void cancelled() {
        super.cancelled();
        updateMessage("Cancelled!");
    }

@Override protected void failed() {
    super.failed();
    updateMessage("Failed!");
    }
};

The ScheduledService Class

Many use cases that involve polling require a service that restarts automatically. To meet those needs, the Service class was extended to produce the ScheduledService class. The ScheduledService class represents a service that automatically restarts after a successful execution and, under special conditions, upon its failure.

When created, the ScheduledService object is in the READY state.

After calling the ScheduledService.start() or ScheduledService.restart() method, the ScheduledService object transitions to the SCHEDULED state for the duration specified by the delay property.

In the RUNNING state, the ScheduledService object executes its task.

Task completes successfully

After the task completes, the ScheduledService object transitions to the SUCCEEDED state, then to the READY state, and then back to the SCHEDULED state. The duration of being in the SCHEDULED state depends on the time when the last transition to the RUNNING state occurred, the current time, and the value of the period property, which defines the minimum amount of time between two consequent runs. If the previous execution completed before the period expired, then the ScheduledService object stays in the SCHEDULED state until the period expires. Otherwise, if the previous execution took longer than the specified period, then the ScheduledService object instantly transitions to the RUNNING state.

Task fails

In the case when the task terminates in the FAILED state, the ScheduledService object either restarts or quits, depending on the values for the restartOnFailure, backoffStrategy, and maximumFailureCount properties.

If the restartOnFailure property is false, then the ScheduledService object transitions to the FAILED state and quits. In this case, you can restart the failed ScheduledService object manually.

If the restartOnFailure property is true, then the ScheduledService object transitions to the SCHEDULED state and remains in this state for the duration of cumulativePeriod property, which is obtained as a result of calling the backoffStrategy property. Using the cumulativePeriod property, you can force the failed ScheduledService object to wait longer before the next run. After the ScheduledService completes successfully, the cumulativePeriod property is reset to the value of the period property. When the amount of consequent failures reaches the value of the maximumFailureCount property, the ScheduledService object transitions to the FAILED state and quits.

Any changes that happen to the delay and period properties while the ScheduledService object is running will be taken into account on the next iteration. The default values for the delay and period properties are set to 0.