3.2spring源码系列----循环依赖源码分析
首先,我们在3.1 spring5源码系列--循环依赖 之 手写代码模拟spring循环依赖 中手写了循环依赖的实现. 这个实现就是模拟的spring的循环依赖. 目的是为了更容易理解spring源码.
下面我们就进入正题, 看看spring的循环依赖源码.
一、getBean整体流程
目标很明确了, 就是要看看spring如何解决循环依赖的.
代码入口是refresh()#finishBeanFactoryInitialization(beanFactory);
二、拆解研究流程中的每一步
调用方法beanFactory.preInstantiateSingletons();实例化剩余的单例bean. 为什么是剩余的?很显然我们在上面已经实例化一部分了.比如配置类, postProcessor等.
2.1 入口
1 @Override 2 public void preInstantiateSingletons() throws BeansException { 3 if (logger.isTraceEnabled()) { 4 logger.trace("Pre-instantiating singletons in " + this); 5 } 6 7 8 // 获取容器中所有bean定义的名字 9 List<String> beanNames = new ArrayList<>(this.beanDefinitionNames); 10 11 // Trigger initialization of all non-lazy singleton beans... 12 /** 13 * 第一步: 循环bean定义的name 14 */ 15 for (String beanName : beanNames) { 16 // 获取bean定义 17 RootBeanDefinition bd = getMergedLocalBeanDefinition(beanName); 18 // 生产bean定义的条件: 不是抽象的, 是单例的, 不是懒加载的. 符合这个标准的, 最后才会调用getBean()生产bean 19 if (!bd.isAbstract() && bd.isSingleton() && !bd.isLazyInit()) { 20 // 这里判断是不是工厂bean, 这里和BeanFactory不是一个意思, 判断当前这个bean是否实现了beanFactory的接口 21 if (isFactoryBean(beanName)) { 22 Object bean = getBean(FACTORY_BEAN_PREFIX + beanName); 23 if (bean instanceof FactoryBean) { 24 final FactoryBean<?> factory = (FactoryBean<?>) bean; 25 boolean isEagerInit; 26 if (System.getSecurityManager() != null && factory instanceof SmartFactoryBean) { 27 isEagerInit = AccessController.doPrivileged((PrivilegedAction<Boolean>) 28 ((SmartFactoryBean<?>) factory)::isEagerInit, 29 getAccessControlContext()); 30 } 31 else { 32 isEagerInit = (factory instanceof SmartFactoryBean && 33 ((SmartFactoryBean<?>) factory).isEagerInit()); 34 } 35 if (isEagerInit) { 36 // 获取bean 37 getBean(beanName); 38 } 39 } 40 } 41 else { // 第二步: 调用bean定义 42 getBean(beanName); 43 } 44 } 45 } 46 47 // Trigger post-initialization callback for all applicable beans... 48 51 for (String beanName : beanNames) { 52 // 从缓存中得到实例instance 53 Object singletonInstance = getSingleton(beanName); 54 if (singletonInstance instanceof SmartInitializingSingleton) { 55 final SmartInitializingSingleton smartSingleton = (SmartInitializingSingleton) singletonInstance; 56 if (System.getSecurityManager() != null) { 57 AccessController.doPrivileged((PrivilegedAction<Object>) () -> { 58 smartSingleton.afterSingletonsInstantiated(); 59 return null; 60 }, getAccessControlContext()); 61 } 62 else { 63 smartSingleton.afterSingletonsInstantiated(); 64 } 65 } 66 } 67 }
首先, 循环bean定义, 这和我们模拟spring循环的第一步是一样的.
第二步: 判断从BeanDefinitionMap中取出来的这个bean是否满足生产bean的条件
我们注意代码注释中, 生产bean定义的条件: 不是抽象的, 是单例的, 不是懒加载的. 符合这个标准的, 最后才会调用getBean()生产bean
然后:调用getBean()
到目前为止,我们完成了上图源码图的第一部分:
2.2 创建bean前的准备工作
接下来看看getBean().doGetBean()方法
1 protected <T> T doGetBean(final String name, @Nullable final Class<T> requiredType, 2 @Nullable final Object[] args, boolean typeCheckOnly) throws BeansException { 3 4 // 第一步: 转换bean name. 在这里传入进来的name可能是别名, 也有可能是工厂bean的name, 所以在这里进行一个转换 5 final String beanName = transformedBeanName(name); 6 Object bean; 7 8 // Eagerly check singleton cache for manually registered singletons. 9 // 第二步: 尝试去缓存中获取对象, 如果没有获取到就创建bean 10 Object sharedInstance = getSingleton(beanName); 11 if (sharedInstance != null && args == null) { 12 if (logger.isTraceEnabled()) { 13 //判断当前类是否是正在创建中 14 if (isSingletonCurrentlyInCreation(beanName)) { 15 logger.trace("Returning eagerly cached instance of singleton bean '" + beanName + 16 "' that is not fully initialized yet - a consequence of a circular reference"); 17 } 18 else { 19 logger.trace("Returning cached instance of singleton bean '" + beanName + "'"); 20 } 21 } 22 bean = getObjectForBeanInstance(sharedInstance, name, beanName, null); 23 } 24 25 else { 26 // Fail if we're already creating this bean instance: 27 // We're assumably within a circular reference. 28 /** 29 * 判断当前的bean是不是多例, 如果是这抛出异常 30 * 31 * 判断当前这个bean是不是多例bean. 如果配置了@Scope("prototype") 就表示这是一个多例的bean 32 * spring 只能解决单例对象的setter注入的循环依赖, 不能解决构造器注入 33 * 34 * 如果是多例的bean, 当前正在创建bean, 也会抛出异常---这也是循环依赖的问题 35 */ 36 if (isPrototypeCurrentlyInCreation(beanName)) { 37 throw new BeanCurrentlyInCreationException(beanName); 38 } 39 40 /** 41 * 下面这段代码是关于子父容器的, 只有spring mvc继承自spring, 才会有子父容器的问题. 42 */ 43 // Check if bean definition exists in this factory. 44 BeanFactory parentBeanFactory = getParentBeanFactory(); 45 if (parentBeanFactory != null && !containsBeanDefinition(beanName)) { 46 // Not found -> check parent. 47 String nameToLookup = originalBeanName(name); 48 if (parentBeanFactory instanceof AbstractBeanFactory) { 49 return ((AbstractBeanFactory) parentBeanFactory).doGetBean( 50 nameToLookup, requiredType, args, typeCheckOnly); 51 } 52 else if (args != null) { 53 // Delegation to parent with explicit args. 54 return (T) parentBeanFactory.getBean(nameToLookup, args); 55 } 56 else if (requiredType != null) { 57 // No args -> delegate to standard getBean method. 58 return parentBeanFactory.getBean(nameToLookup, requiredType); 59 } 60 else { 61 return (T) parentBeanFactory.getBean(nameToLookup); 62 } 63 } 64 65 /** 66 * 方法参数typeCheckOnly是用来判断#getBean()方法时, 表示是否为仅仅进行类型检查, 67 * 如果不仅仅做类型检查, 而是创建bean对象, 则需要调用#markBeanAsCreated(String name) 68 * 69 */ 70 if (!typeCheckOnly) { 71 markBeanAsCreated(beanName); 72 } 73 74 try { 75 final RootBeanDefinition mbd = getMergedLocalBeanDefinition(beanName); 76 checkMergedBeanDefinition(mbd, beanName, args); 77 78 // Guarantee initialization of beans that the current bean depends on. 79 /** 80 * 现在有两个bean1, bean2 , 加载的时候调用的是bean1, bean2. 但如果我们想要bean2优先加载, 就使用@DependOn注解 81 * 用来解析带有dependOn注解的类 82 */ 83 String[] dependsOn = mbd.getDependsOn(); 84 if (dependsOn != null) { 85 for (String dep : dependsOn) { 86 if (isDependent(beanName, dep)) { 87 throw new BeanCreationException(mbd.getResourceDescription(), beanName, 88 "Circular depends-on relationship between '" + beanName + "' and '" + dep + "'"); 89 } 90 registerDependentBean(dep, beanName); 91 try { 92 getBean(dep); 93 } 94 catch (NoSuchBeanDefinitionException ex) { 95 throw new BeanCreationException(mbd.getResourceDescription(), beanName, 96 "'" + beanName + "' depends on missing bean '" + dep + "'", ex); 97 } 98 } 99 } 100 101 // Create bean instance. 102 /** 103 * 第三步: 创建单例bean实例 104 */ 105 if (mbd.isSingleton()) { // 处理单例bean 106 /** 107 * 这里getSingleton()和上面的getSigleton不一样, 上面的是从一级缓存中拿. 108 * 这个getSingleton()就办了一件事: 将bean设置为正在创建的状态. 这个状态很重要, 如果出现循环依赖, 发现bean正在创建, 就不会再创建了 109 */ 110 sharedInstance = getSingleton(beanName, () -> { 111 try { 112 return createBean(beanName, mbd, args); 113 } 114 catch (BeansException ex) { 115 // Explicitly remove instance from singleton cache: It might have been put there 116 // eagerly by the creation process, to allow for circular reference resolution. 117 // Also remove any beans that received a temporary reference to the bean. 118 destroySingleton(beanName); 119 throw ex; 120 } 121 }); 122 // 得到bean实例对象 123 bean = getObjectForBeanInstance(sharedInstance, name, beanName, mbd); 124 } 125 126 else if (mbd.isPrototype()) { // 处理多例bean 127 // It's a prototype -> create a new instance. 128 Object prototypeInstance = null; 129 try { 130 // 当前正在创建多例bean 131 beforePrototypeCreation(beanName); 132 // 执行创建bean 133 prototypeInstance = createBean(beanName, mbd, args); 134 } 135 finally { 136 afterPrototypeCreation(beanName); 137 } 138 // 获取bean实例对象 139 bean = getObjectForBeanInstance(prototypeInstance, name, beanName, mbd); 140 } 141 142 else { // 处理其他类型的bean 143 String scopeName = mbd.getScope(); 144 final Scope scope = this.scopes.get(scopeName); 145 if (scope == null) { 146 throw new IllegalStateException("No Scope registered for scope name '" + scopeName + "'"); 147 } 148 try { 149 Object scopedInstance = scope.get(beanName, () -> { 150 beforePrototypeCreation(beanName); 151 try { 152 return createBean(beanName, mbd, args); 153 } 154 finally { 155 afterPrototypeCreation(beanName); 156 } 157 }); 158 bean = getObjectForBeanInstance(scopedInstance, name, beanName, mbd); 159 } 160 catch (IllegalStateException ex) { 161 throw new BeanCreationException(beanName, 162 "Scope '" + scopeName + "' is not active for the current thread; consider " + 163 "defining a scoped proxy for this bean if you intend to refer to it from a singleton", 164 ex); 165 } 166 } 167 } 168 catch (BeansException ex) { 169 cleanupAfterBeanCreationFailure(beanName); 170 throw ex; 171 } 172 }
在这里, 首先从缓存中获取bean, 看缓存中是否已经存在了
Object sharedInstance = getSingleton(beanName);
然后, 如果缓存中已经存在了,那么久直接取出来. 代码如下:
if (sharedInstance != null && args == null) { if (logger.isTraceEnabled()) { //判断当前bean是否是正在创建中(单例bean) if (isSingletonCurrentlyInCreation(beanName)) { logger.trace("Returning eagerly cached instance of singleton bean '" + beanName + "' that is not fully initialized yet - a consequence of a circular reference"); } else { logger.trace("Returning cached instance of singleton bean '" + beanName + "'"); } } bean = getObjectForBeanInstance(sharedInstance, name, beanName, null); }
如果是空, 就说明是第一次创建, 执行else的部分
首先, 判断是否是正在创建的多例bean, 如果是正在创建的多例bean, 就抛出异常,
已经是正在创建了, 说明这至少是第二次了, 这里处理的是单例bean的循环依赖, 不处理多例bean的循环依赖, 所以抛出异常
对应的代码是这一句
// Fail if we're already creating this bean instance: 27 // We're assumably within a circular reference. 28 /** 29 * 判断当前的bean是不是多例, 如果是这抛出异常 30 * 31 * 判断当前这个bean是不是多例bean. 如果配置了@Scope("prototype") 就表示这是一个多例的bean 32 * spring 只能解决单例对象的setter注入的循环依赖, 不能解决构造器注入 33 * 34 * 如果是多例的bean, 当前正在创建bean, 也会抛出异常---这也是循环依赖的问题 35 */ 36 if (isPrototypeCurrentlyInCreation(beanName)) { 37 throw new BeanCurrentlyInCreationException(beanName); 38 }
那么, 接下来就是首次创建bean. 首次创建的bean有三种情况:
第一种, 这个bean是单例的.
第二种, 这个bean是多例的.
第三种. 其他类型
对应的代码就是这一块. 有行号, 可以和上面一一对应上
// Create bean instance. 102 /** 103 * 第三步: 创建单例bean实例 104 */ 105 if (mbd.isSingleton()) { // 处理单例bean 106 /** 107 * 这里getSingleton()和上面的getSigleton不一样, 上面的是从一级缓存中拿. 108 * 这个getSingleton()就办了一件事: 将bean设置为正在创建的状态. 这个状态很重要, 如果出现循环依赖, 发现bean正在创建, 就不会再创建了 109 */ 110 sharedInstance = getSingleton(beanName, () -> { 111 try { 112 return createBean(beanName, mbd, args); 113 } 114 catch (BeansException ex) { 115 // Explicitly remove instance from singleton cache: It might have been put there 116 // eagerly by the creation process, to allow for circular reference resolution. 117 // Also remove any beans that received a temporary reference to the bean. 118 destroySingleton(beanName); 119 throw ex; 120 } 121 }); 122 // 得到bean实例对象 123 bean = getObjectForBeanInstance(sharedInstance, name, beanName, mbd); 124 } 125 126 else if (mbd.isPrototype()) { // 处理多例bean 127 // It's a prototype -> create a new instance. 128 Object prototypeInstance = null; 129 try { 130 // 当前正在创建多例bean 131 beforePrototypeCreation(beanName); 132 // 执行创建bean 133 prototypeInstance = createBean(beanName, mbd, args); 134 } 135 finally { 136 afterPrototypeCreation(beanName); 137 } 138 // 获取bean实例对象 139 bean = getObjectForBeanInstance(prototypeInstance, name, beanName, mbd); 140 } 141 142 else { // 处理其他类型的bean 143 String scopeName = mbd.getScope(); 144 final Scope scope = this.scopes.get(scopeName); 145 if (scope == null) { 146 throw new IllegalStateException("No Scope registered for scope name '" + scopeName + "'"); 147 } 148 try { 149 Object scopedInstance = scope.get(beanName, () -> { 150 beforePrototypeCreation(beanName); 151 try { 152 return createBean(beanName, mbd, args); 153 } 154 finally { 155 afterPrototypeCreation(beanName); 156 } 157 }); 158 bean = getObjectForBeanInstance(scopedInstance, name, beanName, mbd); 159 } 160 catch (IllegalStateException ex) { 161 throw new BeanCreationException(beanName, 162 "Scope '" + scopeName + "' is not active for the current thread; consider " + 163 "defining a scoped proxy for this bean if you intend to refer to it from a singleton", 164 ex); 165 } 166 }
我们的重点研究对象是单例bean. 所以,重点看单例bean的实现
105 if (mbd.isSingleton()) { // 处理单例bean 106 /** 107 * 这里getSingleton()和上面的getSigleton不一样, 上面的是从一级缓存中拿. 108 * 这个getSingleton()就办了一件事: 将bean设置为正在创建的状态. 这个状态很重要, 如果出现循环依赖, 发现bean正在创建, 就不会再创建了 109 */ 110 sharedInstance = getSingleton(beanName, () -> { 111 try { 112 return createBean(beanName, mbd, args); 113 } 114 catch (BeansException ex) { 115 // Explicitly remove instance from singleton cache: It might have been put there 116 // eagerly by the creation process, to allow for circular reference resolution. 117 // Also remove any beans that received a temporary reference to the bean. 118 destroySingleton(beanName); 119 throw ex; 120 } 121 }); 122 // 得到bean实例对象 123 bean = getObjectForBeanInstance(sharedInstance, name, beanName, mbd); 124 }
这里的重点是调用了getSingleton(beanName, FactoryObject); FactoryObject是一个接口. 定义了一个钩子方法getObject().
这个接口在这里这是进行了定义, 并不会执行. 什么时候执行呢? 后面调用的时候执行.
下面来看看getSingleton()方法, 钩子方法也是在这里被调用的.
1 public Object getSingleton(String beanName, ObjectFactory<?> singletonFactory) { 2 Assert.notNull(beanName, "Bean name must not be null"); 3 synchronized (this.singletonObjects) { 4 // 第一步: 从一级缓存中获取单例对象 5 Object singletonObject = this.singletonObjects.get(beanName); 6 if (singletonObject == null) { 7 if (this.singletonsCurrentlyInDestruction) { 8 throw new BeanCreationNotAllowedException(beanName, 9 "Singleton bean creation not allowed while singletons of this factory are in destruction " + 10 "(Do not request a bean from a BeanFactory in a destroy method implementation!)"); 11 } 12 if (logger.isDebugEnabled()) { 13 logger.debug("Creating shared instance of singleton bean '" + beanName + "'"); 14 } 15 // 第二步: 将bean添加到singletonsCurrentlyInCreation中, 表示bean正在创建 16 beforeSingletonCreation(beanName); 17 boolean newSingleton = false; 18 boolean recordSuppressedExceptions = (this.suppressedExceptions == null); 19 if (recordSuppressedExceptions) { 20 this.suppressedExceptions = new LinkedHashSet<>(); 21 } 22 try { 23 // 第三步: 这里调用getObject()钩子方法, 就会回调匿名函数, 调用singletonFactory的createBean() 24 singletonObject = singletonFactory.getObject(); 25 newSingleton = true; 26 } 27 catch (IllegalStateException ex) { 28 // Has the singleton object implicitly appeared in the meantime -> 29 // if yes, proceed with it since the exception indicates that state. 30 singletonObject = this.singletonObjects.get(beanName); 31 if (singletonObject == null) { 32 throw ex; 33 } 34 } 35 catch (BeanCreationException ex) { 36 if (recordSuppressedExceptions) { 37 for (Exception suppressedException : this.suppressedExceptions) { 38 ex.addRelatedCause(suppressedException); 39 } 40 } 41 throw ex; 42 } 43 finally { 44 if (recordSuppressedExceptions) { 45 this.suppressedExceptions = null; 46 } 47 afterSingletonCreation(beanName); 48 } 49 if (newSingleton) { 50 addSingleton(beanName, singletonObject); 51 } 52 } 53 return singletonObject; 54 } 55 }
这里是调用getBean().
第一步: 去一级缓存中取成熟的单例bean. 如果拿到了, 就直接返回. 如果没拿到. 那么执行创建.
第二步: 在创建之前, 先把这个bean放入到正在创建的单例bean集合中. 标记这个bean正在创建中
第三步: 就是调用钩子方法getObject()了. 这个方法的方法体是在上面定义的. 其内容是去创建实例
sharedInstance = getSingleton(beanName, () -> { try { // 这里定义了一个钩子函数. 此时只是定义, 并不执行. 在真正需要创建bean的地方才会执行 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; } });
这里的代码逻辑是完成了创建之前的逻辑
2.3 创建bean
下面看看创建bean的过程
1 protected Object doCreateBean(final String beanName, final RootBeanDefinition mbd, final @Nullable Object[] args) 2 throws BeanCreationException { 3 4 // Instantiate the bean. 5 BeanWrapper instanceWrapper = null; 6 if (mbd.isSingleton()) { 7 instanceWrapper = this.factoryBeanInstanceCache.remove(beanName); 8 } 9 if (instanceWrapper == null) { 10 /** 11 * 第一步: 实例化 12 * 这里面的调用链非常深, 后面再看 13 * bean实例化有两种方式 14 * 1. 使用反射: 使用反射也有两种方式, 15 * a. 通过无参构造函数 (默认的方式) 16 * 从beanDefinition中可以得到beanClass, 17 * ClassName = BeanDefinition.beanClass 18 * Class clazz = Class.forName(ClassName); 19 * clazz.newInstance(); 20 * 这样就可以实例化bean了 21 * 22 * b. 通过有参函数. 23 * ClassName = BeanDefinition.beanClass 24 * Class clazz = Class.forName(ClassName); 25 * Constractor con = class.getConstractor(args....) 26 * con.newInstance(); 27 * 28 * 2. 使用工厂 29 * 我们使用@Bean的方式, 就是使用的工厂模式, 自己控制实例化过程 30 * 31 */ 32 instanceWrapper = createBeanInstance(beanName, mbd, args); 33 } 34 // 这里使用了装饰器的设计模式 35 final Object bean = instanceWrapper.getWrappedInstance(); 36 Class<?> beanType = instanceWrapper.getWrappedClass(); 37 if (beanType != NullBean.class) { 38 mbd.resolvedTargetType = beanType; 39 } 40 41 // Allow post-processors to modify the merged bean definition. 42 // 允许后置处理器修改已经合并的beanDefinition 43 synchronized (mbd.postProcessingLock) { 44 if (!mbd.postProcessed) { 45 try { 46 applyMergedBeanDefinitionPostProcessors(mbd, beanType, beanName); 47 } 48 catch (Throwable ex) { 49 throw new BeanCreationException(mbd.getResourceDescription(), beanName, 50 "Post-processing of merged bean definition failed", ex); 51 } 52 mbd.postProcessed = true; 53 } 54 } 55 56 /** 57 * 缓存单例bean到三级缓存中, 以防止循环依赖 58 * 判断是否是早期引用的bean, 如果是, 则允许提前暴露引用 59 * 60 * 判断是否能够早起暴露的条件 61 * 1. 是单例 62 * 2. 允许循环依赖 63 * 3. 正在创建的bean 64 */ 65 boolean earlySingletonExposure = (mbd.isSingleton() && this.allowCircularReferences && 66 isSingletonCurrentlyInCreation(beanName)); 67 if (earlySingletonExposure) { 68 if (logger.isTraceEnabled()) { 69 logger.trace("Eagerly caching bean '" + beanName + 70 "' to allow for resolving potential circular references"); 71 } 72 // 把我们的早期对象包装成一个singletonFactory对象, 该对象提供了getObject()方法, 把静态的bean放到三级缓存中去了. 73 addSingletonFactory(beanName, () -> getEarlyBeanReference(beanName, mbd, bean)); 74 } 75 76 // Initialize the bean instance. 77 Object exposedObject = bean; 78 try { 79 // 第二步:填充属性, 给属性赋值(调用set方法) 这里也是调用的后置处理器 80 populateBean(beanName, mbd, instanceWrapper); 81 // 第三步: 初始化. 82 exposedObject = initializeBean(beanName, exposedObject, mbd); 83 } 84 catch (Throwable ex) { 85 if (ex instanceof BeanCreationException && beanName.equals(((BeanCreationException) ex).getBeanName())) { 86 throw (BeanCreationException) ex; 87 } 88 else { 89 throw new BeanCreationException( 90 mbd.getResourceDescription(), beanName, "Initialization of bean failed", ex); 91 } 92 } 93 94 /** 95 * 初始化完成以后, 判断是否是早期的对象 96 * 是循环依赖. 才会走进这里来 97 */ 98 if (earlySingletonExposure) { 99 // 去缓存中获取到我们的对象 由于传递的allowEarlyReference是false, 要求只能在一级二级缓存中取 100 // 正常的普通的bean(不存在循环依赖的bean) 创建的过程中, 不会把三级缓存提升到二级缓存中. 101 Object earlySingletonReference = getSingleton(beanName, false); 102 if (earlySingletonReference != null) { 103 if (exposedObject == bean) { 104 exposedObject = earlySingletonReference; 105 } 106 else if (!this.allowRawInjectionDespiteWrapping && hasDependentBean(beanName)) { 107 String[] dependentBeans = getDependentBeans(beanName); 108 Set<String> actualDependentBeans = new LinkedHashSet<>(dependentBeans.length); 109 for (String dependentBean : dependentBeans) { 110 if (!removeSingletonIfCreatedForTypeCheckOnly(dependentBean)) { 111 actualDependentBeans.add(dependentBean); 112 } 113 } 114 if (!actualDependentBeans.isEmpty()) { 115 throw new BeanCurrentlyInCreationException(beanName, 116 "Bean with name '" + beanName + "' has been injected into other beans [" + 117 StringUtils.collectionToCommaDelimitedString(actualDependentBeans) + 118 "] in its raw version as part of a circular reference, but has eventually been " + 119 "wrapped. This means that said other beans do not use the final version of the " + 120 "bean. This is often the result of over-eager type matching - consider using " + 121 "'getBeanNamesForType' with the 'allowEagerInit' flag turned off, for example."); 122 } 123 } 124 } 125 } 126 127 // Register bean as disposable. 128 try { 129 registerDisposableBeanIfNecessary(beanName, bean, mbd); 130 } 131 catch (BeanDefinitionValidationException ex) { 132 throw new BeanCreationException( 133 mbd.getResourceDescription(), beanName, "Invalid destruction signature", ex); 134 } 135 136 return exposedObject; 137 }
首先, 实例化bean, 实例化的方式有两种. 一种是通过反射, 另一种是通过动态
1 /** 2 * 第一步: 实例化 3 * 这里面的调用链非常深, 后面再看 4 * bean实例化有两种方式 5 * 1. 使用反射: 使用反射也有两种方式, 6 * a. 通过无参构造函数 (默认的方式) 7 * 从beanDefinition中可以得到beanClass, 8 * ClassName = BeanDefinition.beanClass 9 * Class clazz = Class.forName(ClassName); 10 * clazz.newInstance(); 11 * 这样就可以实例化bean了 12 * 13 * b. 通过有参函数. 14 * ClassName = BeanDefinition.beanClass 15 * Class clazz = Class.forName(ClassName); 16 * Constractor con = class.getConstractor(args....) 17 * con.newInstance(); 18 * 19 * 2. 使用工厂 20 * 我们使用@Bean的方式, 就是使用的工厂模式, 自己控制实例化过程 21 * 22 */ 23 instanceWrapper = createBeanInstance(beanName, mbd, args);
判断是否是早期暴露的bean. 满足早期暴露的bean的三个条件是
1. 是单例的
2. 允许循环依赖
3. bean已经是处在正在创建中的行列了.
/* 判断是否能够早起暴露的条件 * 1. 是单例 * 2. 允许循环依赖 * 3. 正在创建的bean */ boolean earlySingletonExposure = (mbd.isSingleton() && this.allowCircularReferences && isSingletonCurrentlyInCreation(beanName));
创建bean的第二步: 属性赋值
// 第二步:填充属性, 给属性赋值(调用set方法) 这里也是调用的后置处理器 populateBean(beanName, mbd, instanceWrapper);
在这里会判断, 是否带有@Autowired的属性. 分为两种一种是Name,一种是Type
@SuppressWarnings("deprecation") // for postProcessPropertyValues protected void populateBean(String beanName, RootBeanDefinition mbd, @Nullable BeanWrapper bw) { if (bw == null) { if (mbd.hasPropertyValues()) { throw new BeanCreationException( mbd.getResourceDescription(), beanName, "Cannot apply property values to null instance"); } else { // Skip property population phase for null instance. return; } } // Give any InstantiationAwareBeanPostProcessors the opportunity to modify the // state of the bean before properties are set. This can be used, for example, // to support styles of field injection. if (!mbd.isSynthetic() && hasInstantiationAwareBeanPostProcessors()) { for (BeanPostProcessor bp : getBeanPostProcessors()) { if (bp instanceof InstantiationAwareBeanPostProcessor) { InstantiationAwareBeanPostProcessor ibp = (InstantiationAwareBeanPostProcessor) bp; if (!ibp.postProcessAfterInstantiation(bw.getWrappedInstance(), beanName)) { return; } } } } PropertyValues pvs = (mbd.hasPropertyValues() ? mbd.getPropertyValues() : null); // 判断属性是否有Autowired注解 int resolvedAutowireMode = mbd.getResolvedAutowireMode(); // Autowired是根据名字或者根据类型 if (resolvedAutowireMode == AUTOWIRE_BY_NAME || resolvedAutowireMode == AUTOWIRE_BY_TYPE) { MutablePropertyValues newPvs = new MutablePropertyValues(pvs); // Add property values based on autowire by name if applicable. if (resolvedAutowireMode == AUTOWIRE_BY_NAME) { autowireByName(beanName, mbd, bw, newPvs); } // Add property values based on autowire by type if applicable. if (resolvedAutowireMode == AUTOWIRE_BY_TYPE) { autowireByType(beanName, mbd, bw, newPvs); } pvs = newPvs; } ...... }
如果按照名字注入
1 protected void autowireByName( 2 String beanName, AbstractBeanDefinition mbd, BeanWrapper bw, MutablePropertyValues pvs) { 3 4 String[] propertyNames = unsatisfiedNonSimpleProperties(mbd, bw); 5 for (String propertyName : propertyNames) { 6 if (containsBean(propertyName)) { 7 // 调用getBean 8 Object bean = getBean(propertyName); 9 pvs.add(propertyName, bean); 10 registerDependentBean(propertyName, beanName); 11 if (logger.isTraceEnabled()) { 12 logger.trace("Added autowiring by name from bean name '" + beanName + 13 "' via property '" + propertyName + "' to bean named '" + propertyName + "'"); 14 } 15 } 16 else { 17 if (logger.isTraceEnabled()) { 18 logger.trace("Not autowiring property '" + propertyName + "' of bean '" + beanName + 19 "' by name: no matching bean found"); 20 } 21 } 22 } 23 }
会再次调用getBean方法. 构建bean. 这是就有可能出现循环依赖了.
按类型注入也是一样的.
只是解析bean的方式不同.
创建bean的第三步: 初始化
// 第三步: 初始化. exposedObject = initializeBean(beanName, exposedObject, mbd);
在初始化bean的时候, 会调用很多的aware. 还会调用init-method方法. 以及bean的后置处理器.
第四步:删除实例化和静态方法在缓存中的数据
/** * 初始化完成以后, 判断是否是早期的对象 * 是循环依赖. 才会走进这里来 */ if (earlySingletonExposure) { // 去缓存中获取到我们的对象 由于传递的allowEarlyReference是false, 要求只能在一级二级缓存中取 // 正常的普通的bean(不存在循环依赖的bean) 创建的过程中, 不会把三级缓存提升到二级缓存中. Object earlySingletonReference = getSingleton(beanName, false); if (earlySingletonReference != null) { if (exposedObject == bean) { exposedObject = earlySingletonReference; } else if (!this.allowRawInjectionDespiteWrapping && hasDependentBean(beanName)) { String[] dependentBeans = getDependentBeans(beanName); Set<String> actualDependentBeans = new LinkedHashSet<>(dependentBeans.length); for (String dependentBean : dependentBeans) { if (!removeSingletonIfCreatedForTypeCheckOnly(dependentBean)) { actualDependentBeans.add(dependentBean); } } if (!actualDependentBeans.isEmpty()) { throw new BeanCurrentlyInCreationException(beanName, "Bean with name '" + beanName + "' has been injected into other beans [" + StringUtils.collectionToCommaDelimitedString(actualDependentBeans) + "] in its raw version as part of a circular reference, but has eventually been " + "wrapped. This means that said other beans do not use the final version of the " + "bean. This is often the result of over-eager type matching - consider using " + "'getBeanNamesForType' with the 'allowEagerInit' flag turned off, for example."); } } } }
removeSingletonIfCreatedForTypeCheckOnly调用方法, 删除缓存.
这既是getBean()整个的过程. 中间还有很多细节, 没有往里面深入的看, 因为spring代码非常的深, 看的太深就忘了我们的目标了. 结合之前手写的spring循环依赖的思想看, 还是可以看得懂的.
三. 接下来有几个问题
问题1: 为什么需要二级缓存和三级缓存?
一级缓存: 用来存储完整的bean
二级缓存: 用来存储早期的,纯净的bean, 也就是没有注入属性的bean
三级缓存: 存的是函数的接口, 主要目的是为了解耦
二级缓存 和三级缓存 结合起来, 解决了循环依赖下的AOP动态代理的问题
问题2:有没有解决构造函数的循环依赖
答案是没有. 因为构造函数是在实例化的时候构建的. 这个时候bean都还没有创建, 所以没有办法处理循环依赖.如果出现构造函数的循环依赖, 是会直接报错的..
问题3:有没有解决多例下的循环依赖
也是没有的, 因为我们会判断, 如果是多例, 那么会抛出异常
1 /** 2 * 第二步: 判断当前bean是否是正在创建中的多例bean, 如果是就抛出异常 3 * 4 * 2. 判断当前这个bean是不是多例bean. 如果配置了@Scope("prototype") 就表示这是一个多例的bean 5 * spring 只能解决单例对象的setter注入的循环依赖, 不能解决构造器注入 6 * 7 * 如果是多例的bean, 当前正在创建bean, 也会抛出异常---这也是循环依赖的问题 8 */ 9 if (isPrototypeCurrentlyInCreation(beanName)) { 10 throw new BeanCurrentlyInCreationException(beanName); 11 }