Using Scopes

For a web application to use a bean that injects another bean class, the bean needs to be able to hold state over the duration of the user’s interaction with the application. The way to define this state is to give the bean a scope. You can give an object any of the scopes described in Table 25-1, depending on how you are using it.

Table 25-1 Scopes

Scope

Annotation

Duration

Request

@RequestScoped

A user’s interaction with a web application in a single HTTP request.

Session

@SessionScoped

A user’s interaction with a web application across multiple HTTP requests.

Application

@ApplicationScoped

Shared state across all users' interactions with a web application.

Dependent

@Dependent

The default scope if none is specified; it means that an object exists to serve exactly one client (bean) and has the same lifecycle as that client (bean).

Conversation

@ConversationScoped

A user’s interaction with a servlet, including JavaServer Faces applications. The conversation scope exists within developer-controlled boundaries that extend it across multiple requests for long-running conversations. All long-running conversations are scoped to a particular HTTP servlet session and may not cross session boundaries.

The first three scopes are defined by both JSR 365 and the JavaServer Faces specification. The last two are defined by JSR 365.

All predefined scopes except @Dependent are contextual scopes. CDI places beans of contextual scope in the context whose lifecycle is defined by the Java EE specifications. For example, a session context and its beans exist during the lifetime of an HTTP session. Injected references to the beans are contextually aware. The references always apply to the bean that is associated with the context for the thread that is making the reference. The CDI container ensures that the objects are created and injected at the correct time as determined by the scope that is specified for these objects.

You can also define and implement custom scopes, but that is an advanced topic. Custom scopes are likely to be used by those who implement and extend the CDI specification.

A scope gives an object a well-defined lifecycle context. A scoped object can be automatically created when it is needed and automatically destroyed when the context in which it was created ends. Moreover, its state is automatically shared by any clients that execute in the same context.

Java EE components, such as servlets and enterprise beans, and JavaBeans components do not by definition have a well-defined scope. These components are one of the following:

Singletons, such as Enterprise JavaBeans singleton beans, whose state is shared among all clients
Stateless objects, such as servlets and stateless session beans, which do not contain client-visible state
Objects that must be explicitly created and destroyed by their client, such as JavaBeans components and stateful session beans, whose state is shared by explicit reference passing between clients

However, if you create a Java EE component that is a managed bean, then it becomes a scoped object, which exists in a well-defined lifecycle context.

The web application for the Printer bean will use a simple request and response mechanism, so the managed bean can be annotated as follows:

import javax.enterprise.context.RequestScoped;
import javax.inject.Inject;

@RequestScoped
public class Printer {

    @Inject @Informal Greeting greeting;
    ...
}

Beans that use session, application, or conversation scope must be serializable, but beans that use request scope do not have to be serializable.