Spring Framework Documentation - core - 1. The IoC Container - 1.4 依赖项[TODO]

原文链接：Core Technologies

spring version 5.3.9

上一篇：1.3Bean概述

1.4. 依赖 Dependencies

典型的企业级软件不止一个对象，或者说不止一个Bean。最简单的应用都会有几个对象共同工作。接下来的章节解释了如何定义若干个Bean，并让其协作实现应用目标。

A typical enterprise application does not consist of a single object (or bean in the Spring parlance). Even the simplest application has a few objects that work together to present what the end-user sees as a coherent application. This next section explains how you go from defining a number of bean definitions that stand alone to a fully realized application where objects collaborate to achieve a goal.

1.4.1. 依赖注入 Dependency Injection

依赖注入就是在把对象的依赖，只通过构造方法参数或者工程方法参数，或者构造方法实例化，工厂方法构建完后调用属性set方法进行设置。容器会在创建Bean的时候注入依赖项。这个流程和常规的Bean 自身控制依赖的实例化或者用服务定位器模式来控制其依赖项 的流程正好相反，所以也叫控制反转。

Dependency injection (DI) is a process whereby objects define their dependencies (that is, the other objects with which they work) only through constructor arguments, arguments to a factory method, or properties that are set on the object instance after it is constructed or returned from a factory method. The container then injects those dependencies when it creates the bean. This process is fundamentally the inverse (hence the name, Inversion of Control) of the bean itself controlling the instantiation or location of its dependencies on its own by using direct construction of classes or the Service Locator pattern.

使用DI原则代码会让依赖的过程更加清晰，解耦也更加高效。对象无需查找他们的依赖性项，也无需知道他们的依赖在哪，甚至不知道他们会依赖于那个类。这样的结果是，你的类可测试性更高，特别是依赖是基于接口或者抽象类，这样子类或者mock的实现在单元测试是很实用。

Code is cleaner with the DI principle, and decoupling is more effective when objects are provided with their dependencies. The object does not look up its dependencies and does not know the location or class of the dependencies. As a result, your classes become easier to test, particularly when the dependencies are on interfaces or abstract base classes, which allow for stub or mock implementations to be used in unit tests.

DI exists in two major variants: Constructor-based dependency injection and Setter-based dependency injection.

基于构造方法的依赖注入 Constructor-based Dependency Injection

基于构造函数的依赖注入是通过容器调用具有多个参数的构造函数来完成的，每个参数表示一个依赖项。这和调用带有特定参数的静态工厂方法来构造bean几乎是等效的，本章节讨论类似地处理构造函数和静态工厂方法的参数。以下示例显示了一个类，该类只能通过构造函数注入进行依赖项注入：

Constructor-based DI is accomplished by the container invoking a constructor with a number of arguments, each representing a dependency. Calling a static factory method with specific arguments to construct the bean is nearly equivalent, and this discussion treats arguments to a constructor and to a static factory method similarly. The following example shows a class that can only be dependency-injected with constructor injection:

public class SimpleMovieLister {

    // the SimpleMovieLister has a dependency on a MovieFinder
    private final MovieFinder movieFinder;

    // a constructor so that the Spring container can inject a MovieFinder
    public SimpleMovieLister(MovieFinder movieFinder) {
        this.movieFinder = movieFinder;
    }

    // business logic that actually uses the injected MovieFinder is omitted...
}

注意这个类没什么特殊的。这就是一个POJO，没有依赖容器的特定接口，基类或者注解。

Notice that there is nothing special about this class. It is a POJO that has no dependencies on container specific interfaces, base classes, or annotations.

构造器参数的解析处理 Constructor Argument Resolution

构造函数参数使用类型进行匹配注入。如果bean定义的构造函数参数中不存在歧义，那么在bean定义中定义构造函数参数的顺序就是在实例化bean时将这些参数提供给相应构造函数的顺序。可参考：

Constructor argument resolution matching occurs by using the argument’s type. If no potential ambiguity exists in the constructor arguments of a bean definition, the order in which the constructor arguments are defined in a bean definition is the order in which those arguments are supplied to the appropriate constructor when the bean is being instantiated. Consider the following class:

package x.y;

public class ThingOne {

    public ThingOne(ThingTwo thingTwo, ThingThree thingThree) {
        // ...
    }
}

假设ThingTwo和ThingTree类没有继承关系，则不存在歧义。因此，以下配置可以生效，您不需要在元素中显式指定构造函数参数索引或类型。

Assuming that the ThingTwo and ThingThree classes are not related by inheritance, no potential ambiguity exists. Thus, the following configuration works fine, and you do not need to specify the constructor argument indexes or types explicitly in the  element.

上面的Bean有引用的时候，且类型是确定的，类型匹配没问题。但是下面的Bean 有用一些基础类型，比如 true Spring就不能决定值得类型，所以就不能按照类型匹配

When another bean is referenced, the type is known, and matching can occur (as was the case with the preceding example). When a simple type is used, such as true, Spring cannot determine the type of the value, and so cannot match by type without help. Consider the following class:

package examples;

public class ExampleBean {

    // Number of years to calculate the Ultimate Answer
    private final int years;

    // The Answer to Life, the Universe, and Everything
    private final String ultimateAnswer;

    public ExampleBean(int years, String ultimateAnswer) {
        this.years = years;
        this.ultimateAnswer = ultimateAnswer;
    }
}

构造器参数匹配 Constructor argument type matching

在前面的例子中，如果你在构造器中用type 属性指定了类型，容器可以根据类型进行匹配：

In the preceding scenario, the container can use type matching with simple types if you explicitly specify the type of the constructor argument by using the type attribute, as the following example shows:

构造器参数索引 Constructor argument index

你可以使用index 属性指定传入构造参数的参数顺序。

You can use the index attribute to specify explicitly the index of constructor arguments, as the following example shows:

在处理构造函数有多相同类型的参数的时候，可以用这个来区分注入。

In addition to resolving the ambiguity of multiple simple values, specifying an index resolves ambiguity where a constructor has two arguments of the same type.

The index is 0-based.注意index 从0开始。

构造函数参数名称 Constructor argument name

你可以使用构造函数的参数名称进行匹配，如下

You can also use the constructor parameter name for value disambiguation, as the following example shows:

注意，想要成功，你的代码必须debug=enable ，以保证Spring 可以从构造方法中查找到参数名称。如果你不能或者不想debug=enable，你可以使用 @ConstructorProperties 的JDK 注解，声明构造器函数，如下：

Keep in mind that, to make this work out of the box, your code must be compiled with the debug flag enabled so that Spring can look up the parameter name from the constructor. If you cannot or do not want to compile your code with the debug flag, you can use the @ConstructorProperties JDK annotation to explicitly name your constructor arguments. The sample class would then have to look as follows:

package examples;

public class ExampleBean {

    // Fields omitted

    @ConstructorProperties({"years", "ultimateAnswer"})
    public ExampleBean(int years, String ultimateAnswer) {
        this.years = years;
        this.ultimateAnswer = ultimateAnswer;
    }
}

基于Setter的依赖注入 Setter-based Dependency Injection

基于setter 的依赖注入，容器会在你的bean调用无参构造方法或者无参的静态工厂方法实例化你的Bean后，调用setter方法。

下面是例子，展示了一个类只用纯setter注入。这个类是传统的Java。它是一个POJO，不依赖于特定于容器的接口、基类或注释。

Setter-based DI is accomplished by the container calling setter methods on your beans after invoking a no-argument constructor or a no-argument static factory method to instantiate your bean.

The following example shows a class that can only be dependency-injected by using pure setter injection. This class is conventional Java. It is a POJO that has no dependencies on container specific interfaces, base classes, or annotations.
 

public class SimpleMovieLister {

    // the SimpleMovieLister has a dependency on the MovieFinder
    private MovieFinder movieFinder;

    // a setter method so that the Spring container can inject a MovieFinder
    public void setMovieFinder(MovieFinder movieFinder) {
        this.movieFinder = movieFinder;
    }

    // business logic that actually uses the injected MovieFinder is omitted...
}

ApplicationContext  容器支持基于构造方法和基于setter的依赖注入。并且还支持在构造方法注入后在用setter注入。你可以以BeanDefinition形式配置依赖，这种形式让你用PropertyEditor 实例完成属性格式的转换。不过，大多Spring用户都不会直接在代码里使用这些类，基本都是用XML或者注解（类上的@Component, @Controller， @Configuration类中方法上 @Bean 的等等）进行配置。这些代码可以不转换成 BeanDefinition  并用于加载Spring IoC容器的实例化。

The ApplicationContext supports constructor-based and setter-based DI for the beans it manages. It also supports setter-based DI after some dependencies have already been injected through the constructor approach. You configure the dependencies in the form of a BeanDefinition, which you use in conjunction with PropertyEditor instances to convert properties from one format to another. However, most Spring users do not work with these classes directly (that is, programmatically) but rather with XML bean definitions, annotated components (that is, classes annotated with @Component, @Controller, and so forth), or @Bean methods in Java-based @Configuration classes. These sources are then converted internally into instances of BeanDefinition and used to load an entire Spring IoC container instance.

怎么选择注入方式呢 Constructor-based or setter-based DI?

既然可以混着用构造器注入和setter注入，所以对于强依赖项使用构造函数，对于可选依赖项使用setter方法或配置方法是一个很好的经验法则。注意，在setter方法上使用@Required注释可以使属性成为必需的依赖项；但是，构造函数注入和参数的编程验证更可取。

Since you can mix constructor-based and setter-based DI, it is a good rule of thumb to use constructors for mandatory dependencies and setter methods or configuration methods for optional dependencies. Note that use of the @Required annotation on a setter method can be used to make the property be a required dependency; however, constructor injection with programmatic validation of arguments is preferable.

Spring 开发团队建议构造函数注入，则可以让你以不变的对象形式实现应用，并且能保证需要的依赖不为null。而且，构造器注入组件总能让你返回调用者一个完全初始化好状态的对象。不过，很多参数的构造方法是一种坏味道，解决办法饮食是依赖太多了，应该重构，让功能恰当分离减少依赖。

The Spring team generally advocates constructor injection, as it lets you implement application components as immutable objects and ensures that required dependencies are not null. Furthermore, constructor-injected components are always returned to the client (calling) code in a fully initialized state. As a side note, a large number of constructor arguments is a bad code smell, implying that the class likely has too many responsibilities and should be refactored to better address proper separation of concerns.

setter 注入的话，应该仅用于可选的依赖，这些依赖项可以在类中分配合理的默认值。否则，必须在代码使用依赖项的任何地方执行非空检查。setter注入的一个好处是setter方法使该类的对象易于重复配置或重新注入。通过JMX MBeans进行管理是setter注入的一个好的例子。

Setter injection should primarily only be used for optional dependencies that can be assigned reasonable default values within the class. Otherwise, not-null checks must be performed everywhere the code uses the dependency. One benefit of setter injection is that setter methods make objects of that class amenable to reconfiguration or re-injection later. Management through JMX MBeans is therefore a compelling use case for setter injection.

使用DI方式对特定的类具有意义。有时候，处理第三方的类，有时可能你没有源码，那选择就取决于你了。比如，如果第三方的类没有暴露setter方法，你就只能选择构造方法注入了。

Use the DI style that makes the most sense for a particular class. Sometimes, when dealing with third-party classes for which you do not have the source, the choice is made for you. For example, if a third-party class does not expose any setter methods, then constructor injection may be the only available form of DI.

依赖项的处理流程 Dependency Resolution Process

容器处理bean的流程如下

• ApplicationContext 根据配置创建，初始化所有的Bean。配置数据可以是XML，java代码，注解等等
• 对于每一个bean，他的依赖在属性和构造方法中表达，或者是静态工厂的参数，这些依赖会在bean创建的时候提供给bean 
• 所有的属性或者是构造器参数实际上被定义成一个set，或者是其他的被容器管理的bean
• 所有的属性和构造参数是从指定格式转换成真正的属性类型或者构造参数类型。默认情况下，spring 可以把string 转换成任何内置类型（基本类型），比如int, long, String, boolean等等

The container performs bean dependency resolution as follows:

• The ApplicationContext is created and initialized with configuration metadata that describes all the beans. Configuration metadata can be specified by XML, Java code, or annotations.

• For each bean, its dependencies are expressed in the form of properties, constructor arguments, or arguments to the static-factory method (if you use that instead of a normal constructor). These dependencies are provided to the bean, when the bean is actually created.

• Each property or constructor argument is an actual definition of the value to set, or a reference to another bean in the container.

• Each property or constructor argument that is a value is converted from its specified format to the actual type of that property or constructor argument. By default, Spring can convert a value supplied in string format to all built-in types, such as int, long, String, boolean, and so forth.

The Spring container validates the configuration of each bean as the container is created. However, the bean properties themselves are not set until the bean is actually created. Beans that are singleton-scoped and set to be pre-instantiated (the default) are created when the container is created. Scopes are defined in Bean Scopes. Otherwise, the bean is created only when it is requested. Creation of a bean potentially causes a graph of beans to be created, as the bean’s dependencies and its dependencies' dependencies (and so on) are created and assigned. Note that resolution mismatches among those dependencies may show up late — that is, on first creation of the affected bean.

循环依赖 Circular dependencies

If you use predominantly constructor injection, it is possible to create an unresolvable circular dependency scenario.

For example: Class A requires an instance of class B through constructor injection, and class B requires an instance of class A through constructor injection. If you configure beans for classes A and B to be injected into each other, the Spring IoC container detects this circular reference at runtime, and throws a BeanCurrentlyInCreationException.

One possible solution is to edit the source code of some classes to be configured by setters rather than constructors. Alternatively, avoid constructor injection and use setter injection only. In other words, although it is not recommended, you can configure circular dependencies with setter injection.

Unlike the typical case (with no circular dependencies), a circular dependency between bean A and bean B forces one of the beans to be injected into the other prior to being fully initialized itself (a classic chicken-and-egg scenario).

You can generally trust Spring to do the right thing. It detects configuration problems, such as references to non-existent beans and circular dependencies, at container load-time. Spring sets properties and resolves dependencies as late as possible, when the bean is actually created. This means that a Spring container that has loaded correctly can later generate an exception when you request an object if there is a problem creating that object or one of its dependencies — for example, the bean throws an exception as a result of a missing or invalid property. This potentially delayed visibility of some configuration issues is why ApplicationContext implementations by default pre-instantiate singleton beans. At the cost of some upfront time and memory to create these beans before they are actually needed, you discover configuration issues when the ApplicationContext is created, not later. You can still override this default behavior so that singleton beans initialize lazily, rather than being eagerly pre-instantiated.

If no circular dependencies exist, when one or more collaborating beans are being injected into a dependent bean, each collaborating bean is totally configured prior to being injected into the dependent bean. This means that, if bean A has a dependency on bean B, the Spring IoC container completely configures bean B prior to invoking the setter method on bean A. In other words, the bean is instantiated (if it is not a pre-instantiated singleton), its dependencies are set, and the relevant lifecycle methods (such as a configured init method or the InitializingBean callback method) are invoked.

依赖注入的例子 Examples of Dependency Injection

The following example uses XML-based configuration metadata for setter-based DI. A small part of a Spring XML configuration file specifies some bean definitions as follows:

The following example shows the corresponding ExampleBean class:

public class ExampleBean {

    private AnotherBean beanOne;

    private YetAnotherBean beanTwo;

    private int i;

    public void setBeanOne(AnotherBean beanOne) {
        this.beanOne = beanOne;
    }

    public void setBeanTwo(YetAnotherBean beanTwo) {
        this.beanTwo = beanTwo;
    }

    public void setIntegerProperty(int i) {
        this.i = i;
    }
}

In the preceding example, setters are declared to match against the properties specified in the XML file. The following example uses constructor-based DI:

The following example shows the corresponding ExampleBean class:

public class ExampleBean {

    private AnotherBean beanOne;

    private YetAnotherBean beanTwo;

    private int i;

    public ExampleBean(
        AnotherBean anotherBean, YetAnotherBean yetAnotherBean, int i) {
        this.beanOne = anotherBean;
        this.beanTwo = yetAnotherBean;
        this.i = i;
    }
}

The constructor arguments specified in the bean definition are used as arguments to the constructor of the ExampleBean.

Now consider a variant of this example, where, instead of using a constructor, Spring is told to call a static factory method to return an instance of the object:

The following example shows the corresponding ExampleBean class:

public class ExampleBean {

    // a private constructor
    private ExampleBean(...) {
        ...
    }

    // a static factory method; the arguments to this method can be
    // considered the dependencies of the bean that is returned,
    // regardless of how those arguments are actually used.
    public static ExampleBean createInstance (
        AnotherBean anotherBean, YetAnotherBean yetAnotherBean, int i) {

        ExampleBean eb = new ExampleBean (...);
        // some other operations...
        return eb;
    }
}

Arguments to the static factory method are supplied by  elements, exactly the same as if a constructor had actually been used. The type of the class being returned by the factory method does not have to be of the same type as the class that contains the static factory method (although, in this example, it is). An instance (non-static) factory method can be used in an essentially identical fashion (aside from the use of the factory-bean attribute instead of the class attribute), so we do not discuss those details here.

1.4.2. 依赖和配置的细节 Dependencies and Configuration in Detail

As mentioned in the previous section, you can define bean properties and constructor arguments as references to other managed beans (collaborators) or as values defined inline. Spring’s XML-based configuration metadata supports sub-element types within its  and  elements for this purpose.

Straight Values (Primitives, Strings, and so on)

The value attribute of the  element specifies a property or constructor argument as a human-readable string representation. Spring’s conversion service is used to convert these values from a String to the actual type of the property or argument. The following example shows various values being set:

The following example uses the p-namespace for even more succinct XML configuration:

The preceding XML is more succinct. However, typos are discovered at runtime rather than design time, unless you use an IDE (such as IntelliJ IDEA or the Spring Tools for Eclipse) that supports automatic property completion when you create bean definitions. Such IDE assistance is highly recommended.

You can also configure a java.util.Properties instance, as follows:

    
    
        
            jdbc.driver.className=com.mysql.jdbc.Driver
            jdbc.url=jdbc:mysql://localhost:3306/mydb
        
    

The Spring container converts the text inside the  element into a java.util.Properties instance by using the JavaBeans PropertyEditor mechanism. This is a nice shortcut, and is one of a few places where the Spring team do favor the use of the nested  element over the value attribute style.

idref 元素  The idref element

The idref element is simply an error-proof way to pass the id (a string value - not a reference) of another bean in the container to a  or  element. The following example shows how to use it:

The preceding bean definition snippet is exactly equivalent (at runtime) to the following snippet:

The first form is preferable to the second, because using the idref tag lets the container validate at deployment time that the referenced, named bean actually exists. In the second variation, no validation is performed on the value that is passed to the targetName property of the client bean. Typos are only discovered (with most likely fatal results) when the client bean is actually instantiated. If the client bean is a prototype bean, this typo and the resulting exception may only be discovered long after the container is deployed.

The local attribute on the idref element is no longer supported in the 4.0 beans XSD, since it does not provide value over a regular bean reference any more. Change your existing idref local references to idref bean when upgrading to the 4.0 schema.

A common place (at least in versions earlier than Spring 2.0) where the  element brings value is in the configuration of AOP interceptors in a ProxyFactoryBean bean definition. Using  elements when you specify the interceptor names prevents you from misspelling an interceptor ID.

引用其他的Bean References to Other Beans (Collaborators)

The ref element is the final element inside a  or  definition element. Here, you set the value of the specified property of a bean to be a reference to another bean (a collaborator) managed by the container. The referenced bean is a dependency of the bean whose property is to be set, and it is initialized on demand as needed before the property is set. (If the collaborator is a singleton bean, it may already be initialized by the container.) All references are ultimately a reference to another object. Scoping and validation depend on whether you specify the ID or name of the other object through the bean or parent attribute.

Specifying the target bean through the bean attribute of the  tag is the most general form and allows creation of a reference to any bean in the same container or parent container, regardless of whether it is in the same XML file. The value of the bean attribute may be the same as the id attribute of the target bean or be the same as one of the values in the name attribute of the target bean. The following example shows how to use a ref element:

Specifying the target bean through the parent attribute creates a reference to a bean that is in a parent container of the current container. The value of the parent attribute may be the same as either the id attribute of the target bean or one of the values in the name attribute of the target bean. The target bean must be in a parent container of the current one. You should use this bean reference variant mainly when you have a hierarchy of containers and you want to wrap an existing bean in a parent container with a proxy that has the same name as the parent bean. The following pair of listings shows how to use the parent attribute:

    class="org.springframework.aop.framework.ProxyFactoryBean">
    
         
    
    

The local attribute on the ref element is no longer supported in the 4.0 beans XSD, since it does not provide value over a regular bean reference any more. Change your existing ref local references to ref bean when upgrading to the 4.0 schema.

内部Bean Inner Beans

A  element inside the  or  elements defines an inner bean, as the following example shows:

An inner bean definition does not require a defined ID or name. If specified, the container does not use such a value as an identifier. The container also ignores the scope flag on creation, because inner beans are always anonymous and are always created with the outer bean. It is not possible to access inner beans independently or to inject them into collaborating beans other than into the enclosing bean.

As a corner case, it is possible to receive destruction callbacks from a custom scope — for example, for a request-scoped inner bean contained within a singleton bean. The creation of the inner bean instance is tied to its containing bean, but destruction callbacks let it participate in the request scope’s lifecycle. This is not a common scenario. Inner beans typically simply share their containing bean’s scope.

Bean集合 Collections

The , , , and  elements set the properties and arguments of the Java Collection types List, Set, Map, and Properties, respectively. The following example shows how to use them:

    
    
        
            [email protected]
            [email protected]
            [email protected]
        
    
    
    
        
            a list element followed by a reference
            
        
    
    
    
        
    
    
    
        
            just some string
            
        
    

The value of a map key or value, or a set value, can also be any of the following elements:

bean | ref | idref | list | set | map | props | value | null

集合融合 Collection Merging

The Spring container also supports merging collections. An application developer can define a parent , ,  or  element and have child , ,  or  elements inherit and override values from the parent collection. That is, the child collection’s values are the result of merging the elements of the parent and child collections, with the child’s collection elements overriding values specified in the parent collection.

This section on merging discusses the parent-child bean mechanism. Readers unfamiliar with parent and child bean definitions may wish to read the relevant section before continuing.

The following example demonstrates collection merging:

    
        
            
                [email protected]
                [email protected]
            
        
    
    
        
            
            
                [email protected]
                [email protected]
            
        
    

Notice the use of the merge=true attribute on the  element of the adminEmails property of the child bean definition. When the child bean is resolved and instantiated by the container, the resulting instance has an adminEmails Properties collection that contains the result of merging the child’s adminEmails collection with the parent’s adminEmails collection. The following listing shows the result:

[email protected]
[email protected]
[email protected]

The child Properties collection’s value set inherits all property elements from the parent , and the child’s value for the support value overrides the value in the parent collection.

This merging behavior applies similarly to the , , and  collection types. In the specific case of the  element, the semantics associated with the List collection type (that is, the notion of an ordered collection of values) is maintained. The parent’s values precede all of the child list’s values. In the case of the Map, Set, and Properties collection types, no ordering exists. Hence, no ordering semantics are in effect for the collection types that underlie the associated Map, Set, and Properties implementation types that the container uses internally.

集合融合的局限性 Limitations of Collection Merging

You cannot merge different collection types (such as a Map and a List). If you do attempt to do so, an appropriate Exception is thrown. The merge attribute must be specified on the lower, inherited, child definition. Specifying the merge attribute on a parent collection definition is redundant and does not result in the desired merging.

强类型的集合 Strongly-typed collection

With the introduction of generic types in Java 5, you can use strongly typed collections. That is, it is possible to declare a Collection type such that it can only contain (for example) String elements. If you use Spring to dependency-inject a strongly-typed Collection into a bean, you can take advantage of Spring’s type-conversion support such that the elements of your strongly-typed Collection instances are converted to the appropriate type prior to being added to the Collection. The following Java class and bean definition show how to do so:

public class SomeClass {

    private Map accounts;

    public void setAccounts(Map accounts) {
        this.accounts = accounts;
    }
}

When the accounts property of the something bean is prepared for injection, the generics information about the element type of the strongly-typed Map is available by reflection. Thus, Spring’s type conversion infrastructure recognizes the various value elements as being of type Float, and the string values (9.99, 2.75, and 3.99) are converted into an actual Float type.

Null 和空字符串 Null and Empty String Values

Spring treats empty arguments for properties and the like as empty Strings. The following XML-based configuration metadata snippet sets the email property to the empty String value ("").

The preceding example is equivalent to the following Java code:

exampleBean.setEmail("");

The  element handles null values. The following listing shows an example:

The preceding configuration is equivalent to the following Java code:

exampleBean.setEmail(null);

XML p-namespace简写  XML Shortcut with the p-namespace

The p-namespace lets you use the bean element’s attributes (instead of nested  elements) to describe your property values collaborating beans, or both.

Spring supports extensible configuration formats with namespaces, which are based on an XML Schema definition. The beans configuration format discussed in this chapter is defined in an XML Schema document. However, the p-namespace is not defined in an XSD file and exists only in the core of Spring.

The following example shows two XML snippets (the first uses standard XML format and the second uses the p-namespace) that resolve to the same result:

The example shows an attribute in the p-namespace called email in the bean definition. This tells Spring to include a property declaration. As previously mentioned, the p-namespace does not have a schema definition, so you can set the name of the attribute to the property name.

This next example includes two more bean definitions that both have a reference to another bean:

This example includes not only a property value using the p-namespace but also uses a special format to declare property references. Whereas the first bean definition uses  to create a reference from bean john to bean jane, the second bean definition uses p:spouse-ref="jane" as an attribute to do the exact same thing. In this case, spouse is the property name, whereas the -ref part indicates that this is not a straight value but rather a reference to another bean.

The p-namespace is not as flexible as the standard XML format. For example, the format for declaring property references clashes with properties that end in Ref, whereas the standard XML format does not. We recommend that you choose your approach carefully and communicate this to your team members to avoid producing XML documents that use all three approaches at the same time.

XML c-namespace简写 XML Shortcut with the c-namespace

Similar to the XML Shortcut with the p-namespace, the c-namespace, introduced in Spring 3.1, allows inlined attributes for configuring the constructor arguments rather then nested constructor-arg elements.

The following example uses the c: namespace to do the same thing as the from Constructor-based Dependency Injection:

The c: namespace uses the same conventions as the p: one (a trailing -ref for bean references) for setting the constructor arguments by their names. Similarly, it needs to be declared in the XML file even though it is not defined in an XSD schema (it exists inside the Spring core).

For the rare cases where the constructor argument names are not available (usually if the bytecode was compiled without debugging information), you can use fallback to the argument indexes, as follows:

Due to the XML grammar, the index notation requires the presence of the leading _, as XML attribute names cannot start with a number (even though some IDEs allow it). A corresponding index notation is also available for  elements but not commonly used since the plain order of declaration is usually sufficient there.

In practice, the constructor resolution mechanism is quite efficient in matching arguments, so unless you really need to, we recommend using the name notation through-out your configuration.

复合（层级）属性的名称 Compound Property Names

You can use compound or nested property names when you set bean properties, as long as all components of the path except the final property name are not null. Consider the following bean definition:

The something bean has a fred property, which has a bob property, which has a sammy property, and that final sammy property is being set to a value of 123. In order for this to work, the fred property of something and the bob property of fred must not be null after the bean is constructed. Otherwise, a NullPointerException is thrown.

1.4.3.  depends-on 的使用 Using depends-on

如果一个Bean是其他bean的依赖项，这意味着某个bean是其他bean 的一个集合属性。

If a bean is a dependency of another bean, that usually means that one bean is set as a property of another. Typically you accomplish this with the  element in XML-based configuration metadata. However, sometimes dependencies between beans are less direct. An example is when a static initializer in a class needs to be triggered, such as for database driver registration. The depends-on attribute can explicitly force one or more beans to be initialized before the bean using this element is initialized. The following example uses the depends-on attribute to express a dependency on a single bean:

To express a dependency on multiple beans, supply a list of bean names as the value of the depends-on attribute (commas, whitespace, and semicolons are valid delimiters):

The depends-on attribute can specify both an initialization-time dependency and, in the case of singleton beans only, a corresponding destruction-time dependency. Dependent beans that define a depends-on relationship with a given bean are destroyed first, prior to the given bean itself being destroyed. Thus, depends-on can also control shutdown order.

1.4.4. Bean 的懒加载 Lazy-initialized Beans

By default, ApplicationContext implementations eagerly create and configure all singleton beans as part of the initialization process. Generally, this pre-instantiation is desirable, because errors in the configuration or surrounding environment are discovered immediately, as opposed to hours or even days later. When this behavior is not desirable, you can prevent pre-instantiation of a singleton bean by marking the bean definition as being lazy-initialized. A lazy-initialized bean tells the IoC container to create a bean instance when it is first requested, rather than at startup.

In XML, this behavior is controlled by the lazy-init attribute on the  element, as the following example shows:

When the preceding configuration is consumed by an ApplicationContext, the lazy bean is not eagerly pre-instantiated when the ApplicationContext starts, whereas the not.lazy bean is eagerly pre-instantiated.

However, when a lazy-initialized bean is a dependency of a singleton bean that is not lazy-initialized, the ApplicationContext creates the lazy-initialized bean at startup, because it must satisfy the singleton’s dependencies. The lazy-initialized bean is injected into a singleton bean elsewhere that is not lazy-initialized.

You can also control lazy-initialization at the container level by using the default-lazy-init attribute on the  element, as the following example shows:

1.4.5. 自动注入的合作 Autowiring Collaborators

The Spring container can autowire relationships between collaborating beans. You can let Spring resolve collaborators (other beans) automatically for your bean by inspecting the contents of the ApplicationContext. Autowiring has the following advantages:

• Autowiring can significantly reduce the need to specify properties or constructor arguments. (Other mechanisms such as a bean template discussed elsewhere in this chapter are also valuable in this regard.)

• Autowiring can update a configuration as your objects evolve. For example, if you need to add a dependency to a class, that dependency can be satisfied automatically without you needing to modify the configuration. Thus autowiring can be especially useful during development, without negating the option of switching to explicit wiring when the code base becomes more stable.

When using XML-based configuration metadata (see Dependency Injection), you can specify the autowire mode for a bean definition with the autowire attribute of the  element. The autowiring functionality has four modes. You specify autowiring per bean and can thus choose which ones to autowire. The following table describes the four autowiring modes:

Table 2. Autowiring modes

Mode	Explanation
no	(Default) No autowiring. Bean references must be defined by ref elements. Changing the default setting is not recommended for larger deployments, because specifying collaborators explicitly gives greater control and clarity. To some extent, it documents the structure of a system.
byName	Autowiring by property name. Spring looks for a bean with the same name as the property that needs to be autowired. For example, if a bean definition is set to autowire by name and it contains a master property (that is, it has a setMaster(..) method), Spring looks for a bean definition named master and uses it to set the property.
byType	Lets a property be autowired if exactly one bean of the property type exists in the container. If more than one exists, a fatal exception is thrown, which indicates that you may not use byType autowiring for that bean. If there are no matching beans, nothing happens (the property is not set).
constructor	Analogous to byType but applies to constructor arguments. If there is not exactly one bean of the constructor argument type in the container, a fatal error is raised.

With byType or constructor autowiring mode, you can wire arrays and typed collections. In such cases, all autowire candidates within the container that match the expected type are provided to satisfy the dependency. You can autowire strongly-typed Map instances if the expected key type is String. An autowired Map instance’s values consist of all bean instances that match the expected type, and the Map instance’s keys contain the corresponding bean names.

Limitations and Disadvantages of Autowiring

Autowiring works best when it is used consistently across a project. If autowiring is not used in general, it might be confusing to developers to use it to wire only one or two bean definitions.

Consider the limitations and disadvantages of autowiring:

• Explicit dependencies in property and constructor-arg settings always override autowiring. You cannot autowire simple properties such as primitives, Strings, and Classes (and arrays of such simple properties). This limitation is by-design.

• Autowiring is less exact than explicit wiring. Although, as noted in the earlier table, Spring is careful to avoid guessing in case of ambiguity that might have unexpected results. The relationships between your Spring-managed objects are no longer documented explicitly.

• Wiring information may not be available to tools that may generate documentation from a Spring container.

• Multiple bean definitions within the container may match the type specified by the setter method or constructor argument to be autowired. For arrays, collections, or Map instances, this is not necessarily a problem. However, for dependencies that expect a single value, this ambiguity is not arbitrarily resolved. If no unique bean definition is available, an exception is thrown.

In the latter scenario, you have several options:

• Abandon autowiring in favor of explicit wiring.

• Avoid autowiring for a bean definition by setting its autowire-candidate attributes to false, as described in the next section.

• Designate a single bean definition as the primary candidate by setting the primary attribute of its  element to true.

• Implement the more fine-grained control available with annotation-based configuration, as described in Annotation-based Container Configuration.

排除Excluding a Bean from Autowiring

On a per-bean basis, you can exclude a bean from autowiring. In Spring’s XML format, set the autowire-candidate attribute of the  element to false. The container makes that specific bean definition unavailable to the autowiring infrastructure (including annotation style configurations such as @Autowired).

The autowire-candidate attribute is designed to only affect type-based autowiring. It does not affect explicit references by name, which get resolved even if the specified bean is not marked as an autowire candidate. As a consequence, autowiring by name nevertheless injects a bean if the name matches.

You can also limit autowire candidates based on pattern-matching against bean names. The top-level  element accepts one or more patterns within its default-autowire-candidates attribute. For example, to limit autowire candidate status to any bean whose name ends with Repository, provide a value of *Repository. To provide multiple patterns, define them in a comma-separated list. An explicit value of true or false for a bean definition’s autowire-candidate attribute always takes precedence. For such beans, the pattern matching rules do not apply.

These techniques are useful for beans that you never want to be injected into other beans by autowiring. It does not mean that an excluded bean cannot itself be configured by using autowiring. Rather, the bean itself is not a candidate for autowiring other beans.

1.4.6. Method Injection

In most application scenarios, most beans in the container are singletons. When a singleton bean needs to collaborate with another singleton bean or a non-singleton bean needs to collaborate with another non-singleton bean, you typically handle the dependency by defining one bean as a property of the other. A problem arises when the bean lifecycles are different. Suppose singleton bean A needs to use non-singleton (prototype) bean B, perhaps on each method invocation on A. The container creates the singleton bean A only once, and thus only gets one opportunity to set the properties. The container cannot provide bean A with a new instance of bean B every time one is needed.

A solution is to forego some inversion of control. You can make bean A aware of the container by implementing the ApplicationContextAware interface, and by making a getBean("B") call to the container ask for (a typically new) bean B instance every time bean A needs it. The following example shows this approach:

Java

// a class that uses a stateful Command-style class to perform some processing
package fiona.apple;

// Spring-API imports
import org.springframework.beans.BeansException;
import org.springframework.context.ApplicationContext;
import org.springframework.context.ApplicationContextAware;

public class CommandManager implements ApplicationContextAware {

    private ApplicationContext applicationContext;

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

    protected Command createCommand() {
        // notice the Spring API dependency!
        return this.applicationContext.getBean("command", Command.class);
    }

    public void setApplicationContext(
            ApplicationContext applicationContext) throws BeansException {
        this.applicationContext = applicationContext;
    }
}

The preceding is not desirable, because the business code is aware of and coupled to the Spring Framework. Method Injection, a somewhat advanced feature of the Spring IoC container, lets you handle this use case cleanly.

You can read more about the motivation for Method Injection in this blog entry.

Lookup Method Injection

Lookup method injection is the ability of the container to override methods on container-managed beans and return the lookup result for another named bean in the container. The lookup typically involves a prototype bean, as in the scenario described in the preceding section. The Spring Framework implements this method injection by using bytecode generation from the CGLIB library to dynamically generate a subclass that overrides the method.

For this dynamic subclassing to work, the class that the Spring bean container subclasses cannot be final, and the method to be overridden cannot be final, either. Unit-testing a class that has an abstract method requires you to subclass the class yourself and to supply a stub implementation of the abstract method. Concrete methods are also necessary for component scanning, which requires concrete classes to pick up. A further key limitation is that lookup methods do not work with factory methods and in particular not with @Bean methods in configuration classes, since, in that case, the container is not in charge of creating the instance and therefore cannot create a runtime-generated subclass on the fly.

In the case of the CommandManager class in the previous code snippet, the Spring container dynamically overrides the implementation of the createCommand() method. The CommandManager class does not have any Spring dependencies, as the reworked example shows:

package fiona.apple;

// no more Spring imports!

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();
}

In the client class that contains the method to be injected (the CommandManager in this case), the method to be injected requires a signature of the following form:

 [abstract]  theMethodName(no-arguments);

If the method is abstract, the dynamically-generated subclass implements the method. Otherwise, the dynamically-generated subclass overrides the concrete method defined in the original class. Consider the following example:

The bean identified as commandManager calls its own createCommand() method whenever it needs a new instance of the myCommand bean. You must be careful to deploy the myCommand bean as a prototype if that is actually what is needed. If it is a singleton, the same instance of the myCommand bean is returned each time.

Alternatively, within the annotation-based component model, you can declare a lookup method through the @Lookup annotation, as the following example shows:

public abstract class CommandManager {

    public Object process(Object commandState) {
        Command command = createCommand();
        command.setState(commandState);
        return command.execute();
    }

    @Lookup("myCommand")
    protected abstract Command createCommand();
}

Or, more idiomatically, you can rely on the target bean getting resolved against the declared return type of the lookup method:

public abstract class CommandManager {

    public Object process(Object commandState) {
        MyCommand command = createCommand();
        command.setState(commandState);
        return command.execute();
    }

    @Lookup
    protected abstract MyCommand createCommand();
}

Note that you should typically declare such annotated lookup methods with a concrete stub implementation, in order for them to be compatible with Spring’s component scanning rules where abstract classes get ignored by default. This limitation does not apply to explicitly registered or explicitly imported bean classes.

Another way of accessing differently scoped target beans is an ObjectFactory/ Provider injection point. See Scoped Beans as Dependencies. You may also find the ServiceLocatorFactoryBean (in the org.springframework.beans.factory.config package) to be useful.

Arbitrary Method Replacement

A less useful form of method injection than lookup method injection is the ability to replace arbitrary methods in a managed bean with another method implementation. You can safely skip the rest of this section until you actually need this functionality.

With XML-based configuration metadata, you can use the replaced-method element to replace an existing method implementation with another, for a deployed bean. Consider the following class, which has a method called computeValue that we want to override:

public class MyValueCalculator {

    public String computeValue(String input) {
        // some real code...
    }

    // some other methods...
}

A class that implements the org.springframework.beans.factory.support.MethodReplacer interface provides the new method definition, as the following example shows:

/**
 * meant to be used to override the existing computeValue(String)
 * implementation in MyValueCalculator
 */
public class ReplacementComputeValue implements MethodReplacer {

    public Object reimplement(Object o, Method m, Object[] args) throws Throwable {
        // get the input value, work with it, and return a computed result
        String input = (String) args[0];
        ...
        return ...;
    }
}

The bean definition to deploy the original class and specify the method override would resemble the following example:

    
    
        String
    

You can use one or more  elements within the  element to indicate the method signature of the method being overridden. The signature for the arguments is necessary only if the method is overloaded and multiple variants exist within the class. For convenience, the type string for an argument may be a substring of the fully qualified type name. For example, the following all match java.lang.String:

java.lang.String
String
Str

Because the number of arguments is often enough to distinguish between each possible choice, this shortcut can save a lot of typing, by letting you type only the shortest string that matches an argument type.

下一篇： 1.5Bean 作用域

Version 5.3.9
Last updated 2021-07-14 06:36:18 UTC