For the latest stable version, please use Spring Framework 6.2.0!

Using the @Configuration annotation

@Configuration is a class-level annotation indicating that an object is a source of bean definitions. @Configuration classes declare beans through @Bean-annotated methods. Calls to @Bean methods on @Configuration classes can also be used to define inter-bean dependencies. See Basic Concepts: @Bean and @Configuration for a general introduction.

Injecting Inter-bean Dependencies

When beans have dependencies on one another, expressing that dependency is as simple as having one bean method call another, as the following example shows:

  • Java

  • Kotlin

@Configuration
public class AppConfig {

	@Bean
	public BeanOne beanOne() {
		return new BeanOne(beanTwo());
	}

	@Bean
	public BeanTwo beanTwo() {
		return new BeanTwo();
	}
}
@Configuration
class AppConfig {

	@Bean
	fun beanOne() = BeanOne(beanTwo())

	@Bean
	fun beanTwo() = BeanTwo()
}

In the preceding example, beanOne receives a reference to beanTwo through constructor injection.

This method of declaring inter-bean dependencies works only when the @Bean method is declared within a @Configuration class. You cannot declare inter-bean dependencies by using plain @Component classes.

Lookup Method Injection

As noted earlier, lookup method injection is an advanced feature that you should use rarely. It is useful in cases where a singleton-scoped bean has a dependency on a prototype-scoped bean. Using Java for this type of configuration provides a natural means for implementing this pattern. The following example shows how to use lookup method injection:

  • Java

  • Kotlin

public abstract class CommandManager {
	public Object process(Object commandState) {
		// grab a new instance of the appropriate Command interface
		Command command = createCommand();
		// set the state on the (hopefully brand new) Command instance
		command.setState(commandState);
		return command.execute();
	}

	// okay... but where is the implementation of this method?
	protected abstract Command createCommand();
}
abstract class CommandManager {
	fun process(commandState: Any): Any {
		// grab a new instance of the appropriate Command interface
		val command = createCommand()
		// set the state on the (hopefully brand new) Command instance
		command.setState(commandState)
		return command.execute()
	}

	// okay... but where is the implementation of this method?
	protected abstract fun createCommand(): Command
}

By using Java configuration, you can create a subclass of CommandManager where the abstract createCommand() method is overridden in such a way that it looks up a new (prototype) command object. The following example shows how to do so:

  • Java

  • Kotlin

@Bean
@Scope("prototype")
public AsyncCommand asyncCommand() {
	AsyncCommand command = new AsyncCommand();
	// inject dependencies here as required
	return command;
}

@Bean
public CommandManager commandManager() {
	// return new anonymous implementation of CommandManager with createCommand()
	// overridden to return a new prototype Command object
	return new CommandManager() {
		protected Command createCommand() {
			return asyncCommand();
		}
	}
}
@Bean
@Scope("prototype")
fun asyncCommand(): AsyncCommand {
	val command = AsyncCommand()
	// inject dependencies here as required
	return command
}

@Bean
fun commandManager(): CommandManager {
	// return new anonymous implementation of CommandManager with createCommand()
	// overridden to return a new prototype Command object
	return object : CommandManager() {
		override fun createCommand(): Command {
			return asyncCommand()
		}
	}
}

Further Information About How Java-based Configuration Works Internally

Consider the following example, which shows a @Bean annotated method being called twice:

  • Java

  • Kotlin

@Configuration
public class AppConfig {

	@Bean
	public ClientService clientService1() {
		ClientServiceImpl clientService = new ClientServiceImpl();
		clientService.setClientDao(clientDao());
		return clientService;
	}

	@Bean
	public ClientService clientService2() {
		ClientServiceImpl clientService = new ClientServiceImpl();
		clientService.setClientDao(clientDao());
		return clientService;
	}

	@Bean
	public ClientDao clientDao() {
		return new ClientDaoImpl();
	}
}
@Configuration
class AppConfig {

	@Bean
	fun clientService1(): ClientService {
		return ClientServiceImpl().apply {
			clientDao = clientDao()
		}
	}

	@Bean
	fun clientService2(): ClientService {
		return ClientServiceImpl().apply {
			clientDao = clientDao()
		}
	}

	@Bean
	fun clientDao(): ClientDao {
		return ClientDaoImpl()
	}
}

clientDao() has been called once in clientService1() and once in clientService2(). Since this method creates a new instance of ClientDaoImpl and returns it, you would normally expect to have two instances (one for each service). That definitely would be problematic: In Spring, instantiated beans have a singleton scope by default. This is where the magic comes in: All @Configuration classes are subclassed at startup-time with CGLIB. In the subclass, the child method checks the container first for any cached (scoped) beans before it calls the parent method and creates a new instance.

The behavior could be different according to the scope of your bean. We are talking about singletons here.

It is not necessary to add CGLIB to your classpath because CGLIB classes are repackaged under the org.springframework.cglib package and included directly within the spring-core JAR.

There are a few restrictions due to the fact that CGLIB dynamically adds features at startup-time. In particular, configuration classes must not be final. However, any constructors are allowed on configuration classes, including the use of @Autowired or a single non-default constructor declaration for default injection.

If you prefer to avoid any CGLIB-imposed limitations, consider declaring your @Bean methods on non-@Configuration classes (for example, on plain @Component classes instead) or by annotating your configuration class with @Configuration(proxyBeanMethods = false). Cross-method calls between @Bean methods are then not intercepted, so you have to exclusively rely on dependency injection at the constructor or method level there.