Spring源码-循环依赖源码解读
Spring源码-循环依赖源码解读
笔者最近看书还是从网上找资料对循环依赖的问题,大家都是解读了Spring解决循环依赖的想法(大都解释也不准确,在《Spring源码深度解析》作者也是看别人的博客说明了一下),没有从源码的角度分析是怎么解决循环依赖的。
Spring中对象可以配置成单例模式也可配置为原型模式(原型模式很值得一看)。
Spring中可以通过构造函数注入、setter注入的方式来解决对象与对象间的依赖。
对象间的循环依赖只能配置单例、setter注入的方式来解决,其他方法就会报错,下面我们通过源码分析一下。
一、单例、setter注入解决循环依赖
假如有TestA、TestB、TestC三个对象,其中TestA依赖TestB,TestB依赖TestC,TestC依赖TestA。
下面具体通过代码分析Spring是如何解决单例通过Setter注入的循环依赖。
在org.springframework.beans.factory.support.DefaultSingletonBeanRegistry类中有几个集合类型的成员变量,用来做缓存用的需要特别留意,源码如下:
public class DefaultSingletonBeanRegistry extends SimpleAliasRegistry implements SingletonBeanRegistry { ...... private final Map<String, Object> singletonObjects = new ConcurrentHashMap(256); private final Map<String, ObjectFactory<?>> singletonFactories = new HashMap(16); private final Map<String, Object> earlySingletonObjects = new HashMap(16); private final Set<String> singletonsCurrentlyInCreation = Collections.newSetFromMap(new ConcurrentHashMap(16)); ...... }
上面的代码中:
singletonsCurrentlyInCreation:保存对象的BeanName,在创建对象之前就会把对象的beanName保存起来。
singletonFactories:保存对象的BeanName和创建bean的工厂AbstractAutowireCapableBeanFactory(ObjectFactory),(对象的构造函数是在这一步完成的)
earlySingletonObjects:保存对象BeanName和ObjectFactory#getObject的对象(TestA),(TestA还没setter注入),此时可以作为对象填充依赖。
singletonObjects:保存BeanName和bean的实例(此时对象已经完成了setter注入)
通过Spring获取testA对象:
通过Spring获取testA对象 ApplicationContext factory=new ClassPathXmlApplicationContext("classpath:applicationContext-beans.xml"); TestA testA = (TestA) factory.getBean("testA"); System.out.println(testA);
factory.getBean最终调用的是org.springframework.beans.factory.support.AbstractBeanFactory#doGetBean
下面看doGetBean源码:
protected <T> T doGetBean(String name, Class<T> requiredType, final Object[] args, boolean typeCheckOnly) throws BeansException { ...... //重点1 Object sharedInstance = this.getSingleton(beanName); ...... if (mbd.isSingleton()) { //重点2 sharedInstance = getSingleton(beanName, new ObjectFactory<Object>() { @Override public Object getObject() throws BeansException { try { //重点3 return createBean(beanName, mbd, args); } catch (BeansException ex) { // Explicitly remove instance from singleton cache: It might have been put there // eagerly by the creation process, to allow for circular reference resolution. // Also remove any beans that received a temporary reference to the bean. destroySingleton(beanName); throw ex; } } }); bean = getObjectForBeanInstance(sharedInstance, name, beanName, mbd); } }
在上面源码中:
1、重点1:
//重点1 根据beanName试图从缓存中获取已经创建的对象,第一次进入是肯定返回null,这个函数放在后面解释。
2、重点2:
在//重点2中才是真正创建TestA对象的方法,下面是//重点2 org.springframework.beans.factory.support.DefaultSingletonBeanRegistry#getSingleton的源码:
//重点2 public Object getSingleton(String beanName, ObjectFactory<?> singletonFactory) { Object singletonObject = this.singletonObjects.get(beanName); if (singletonObject == null) { ...... //重点2-1 beforeSingletonCreation(beanName); ...... try { //重点2-2 singletonObject = singletonFactory.getObject(); newSingleton = true; }catch (IllegalStateException ex) { ...... }catch (BeanCreationException ex) { ...... } finally { //重点2-3 afterSingletonCreation(beanName); } } }
在//重点2-1中beforeSingletonCreation方法中只做了一件事,就是保存beanName到singletonsCurrentlyInCreation(注意),这个时候testA就保存在singletonsCurrentlyInCreation里面了,源码如下:
//重点2-1 protected void beforeSingletonCreation(String beanName) { //this.singletonsCurrentlyInCreation.add(beanName) if (!this.inCreationCheckExclusions.contains(beanName) && !this.singletonsCurrentlyInCreation.add(beanName)) { throw new BeanCurrentlyInCreationException(beanName); } }
在 //重点2-2 会调用 //重点3 。
3、重点3:
在//重点3中org.springframework.beans.factory.support.AbstractAutowireCapableBeanFactory#createBean调用doCreateBean,doCreateBean方法源码:
protected Object doCreateBean(final String beanName, final RootBeanDefinition mbd, final Object[] args) { ...... //重点3-1 boolean earlySingletonExposure = (mbd.isSingleton() && this.allowCircularReferences && isSingletonCurrentlyInCreation(beanName)); if (earlySingletonExposure) { if (logger.isDebugEnabled()) { logger.debug("Eagerly caching bean '" + beanName + "' to allow for resolving potential circular references"); } //重点3-2 addSingletonFactory(beanName, new ObjectFactory<Object>() { @Override public Object getObject() throws BeansException { //重点3-3 return getEarlyBeanReference(beanName, mbd, bean); } }); } ...... }
在//重点3-1判断testA必须是单例,并存在在singletonsCurrentlyInCreation中,此时才会调用//重点3-2的addSingletonFactory方法,//重点3-2的addSingletonFactory方法源码,
//重点3-2 protected void addSingletonFactory(String beanName, ObjectFactory<?> singletonFactory) { Assert.notNull(singletonFactory, "Singleton factory must not be null"); synchronized (this.singletonObjects) { if (!this.singletonObjects.containsKey(beanName)) { //重点3-2-1 this.singletonFactories.put(beanName, singletonFactory); this.earlySingletonObjects.remove(beanName); this.registeredSingletons.add(beanName); } } }
在上面的代码中有一个singletonFactory的参数,这个参数是//重点3-3调用getEarlyBeanReference得到的,getEarlyBeanReference返回的就是ObjectFactory对象,是完成构造方法的。
在//重点3-2-1 向singletonFactories添加 ObjectFactory(注意),这个时候,testA和testA的ObjectFactory对象保存在singletonFactories,并移除earlySingletonObjects(现在earlySingletonObjects里面并没有testA)。
执行完//重点3-2,发现testA依赖TestB对象,此时会递归调用getBean获取TestB,testB执行步骤和上面testA一样,然后testB依赖TestC,递归调用TestC,此时singletonFactories里面保存的数据如下:
testA -> ObjectFactory(TestA)
testB -> ObjectFactory(TestB)
testC -> ObjectFactory(TestC)
创建testC过程中执行完//重点3-2,发现依赖testA,此时会递归调用getBean获取TestA,这时候执行到//重点1,下面开始解析//重点1 Object sharedInstance = getSingleton(beanName);方法。
4、重点1:
下面是调用//重点1 getSingleton(beanName),调用getSingleton(beanName, true)的源码:
//重点1 protected Object getSingleton(String beanName, boolean allowEarlyReference) { Object singletonObject = this.singletonObjects.get(beanName); if (singletonObject == null && isSingletonCurrentlyInCreation(beanName)) { synchronized (this.singletonObjects) { singletonObject = this.earlySingletonObjects.get(beanName); if (singletonObject == null && allowEarlyReference) { //重点1-1 ObjectFactory<?> singletonFactory = this.singletonFactories.get(beanName); if (singletonFactory != null) { //重点1-2 singletonObject = singletonFactory.getObject(); this.earlySingletonObjects.put(beanName, singletonObject); this.singletonFactories.remove(beanName); } } } } return (singletonObject != NULL_OBJECT ? singletonObject : null); }
在上面代码中isSingletonCurrentlyInCreation就是判断对象(testA)是否被保存过(在//重点2-1的时候testA就被保存了)
在//重点1-1中,从singletonFactories缓存中获取到ObjectFactory(TestA),
在//重点1-2并通过模板模式获取TestA对象,保存在earlySingletonObjects缓存中,并移除singletonFactories中的testA,
此时testC获取到TestA的早期对象,可以注入TestA对象,自此TestC完成依赖注入,并把testC保存到singletonObjects中。
TestC创建完成,返回给testB,testB也算是完成了创建,然后返回给testA,自此循环依赖算是完成了。
总结:
单例的setter注入,会先把创建testA、testB、testC对象的ObjectFactory(对象工厂)保存在singletonFactories里面,然后testC依赖testA,那就从singletonFactories缓存中拿到testA的ObjectFactory,通过ObjectFactory的getObject获取TestA的对象,并保存在earlySingletonObjects缓存中,清除singletonFactories缓存中的testA,此时testC就可以获取earlySingletonObjects缓存中TestA的对象,完成注入TestA的过程。TestC对象创建完成就可以注入到TestB对象中,然后TestB注入到TestA中。
singletonsCurrentlyInCreation就是用来保存是否试图创建某个对象的beanName,不管有没有创建成功,为后来从singletonFactories缓存中或earlySingletonObjects缓存中取值做个标识。
二、单利、构造函数注入循环依赖
假如有TestA、TestB两个对象,TestA依赖TestB,TestB依赖TestA;
构造函数注入和setter注入的不同在于,构造函数注入无法先调用构造函数实例化对象,当TestA依赖TestB,会先把testA保存到singletonsCurrentlyInCreation中,然后getBean(“testB”),然后把testB保存到singletonsCurrentlyInCreation中,发现TestB依赖TestA,然后再getBean(“testA”),此时执行下面的代码(和模块一,重点2是同一块代码):
//重点2 org.springframework.beans.factory.support.DefaultSingletonBeanRegistry#getSingleton的源码: public Object getSingleton(String beanName, ObjectFactory<?> singletonFactory) { //重点2-0 Object singletonObject = this.singletonObjects.get(beanName); if (singletonObject == null) { ...... //重点2-1 beforeSingletonCreation(beanName); ...... try { //重点2-2 singletonObject = singletonFactory.getObject(); newSingleton = true; }catch (IllegalStateException ex) { ...... }catch (BeanCreationException ex) { ...... } finally { //重点2-3 afterSingletonCreation(beanName); } } }
在上面的代码中 //重点2-0 Object singletonObject = this.singletonObjects.get(beanName); 此时singletonObject为空会执行beforeSingletonCreation方法,源码如下:
//重点2-1 protected void beforeSingletonCreation(String beanName) { //this.singletonsCurrentlyInCreation.add(beanName) if (!this.inCreationCheckExclusions.contains(beanName) && !this.singletonsCurrentlyInCreation.add(beanName)) { throw new BeanCurrentlyInCreationException(beanName); } }
结果发现singletonsCurrentlyInCreation已经存在testA,抛出BeanCurrentlyInCreationException。
setter注入为什么不会执行这一步呢,因为setter注入中会实例化TestA、TestB保存在缓存中,所以在执行 //重点2-0 Object singletonObject = this.singletonObjects.get(beanName); 时 singletonObject 不为空,并不会执行beforeSingletonCreation所以不会保存。
三、原型模式下循环依赖
假如有TestA、TestB两个对象,TestA依赖TestB,TestB依赖TestA,当调用getBean(“testA”)时,会先把beanName(testA)保存到isPrototypeCurrentlyInCreation里面,发现TestA依赖TestB,就会去getBean(“testB”),然后把beanName(testB)也保存到isPrototypeCurrentlyInCreation里面,此时TestB发现依赖TestA,去getBean(“testA”)时,发现isPrototypeCurrentlyInCreation已经存在testA,就会抛出BeanCurrentlyInCreationException异常,具体代码如下
factory.getBean最终调用的是org.springframework.beans.factory.support.AbstractBeanFactory#doGetBean,
下面看doGetBean源码:
protected <T> T doGetBean(String name, Class<T> requiredType, final Object[] args, boolean typeCheckOnly) throws BeansException { ...... if (isPrototypeCurrentlyInCreation(beanName)) { throw new BeanCurrentlyInCreationException(beanName); } ...... } //isPrototypeCurrentlyInCreation源码 protected boolean isPrototypeCurrentlyInCreation(String beanName) { Object curVal = this.prototypesCurrentlyInCreation.get(); return (curVal != null && (curVal.equals(beanName) || (curVal instanceof Set && ((Set<?>) curVal).contains(beanName)))); }