在上篇文章中 Spring Ioc 之 Bean的加载(一)
,我们分析了Spring Ioc中Bean的加载 doGetBean()
方法的 2.2从缓存中获取单例bean
和 2.3获取最终的bean实例对象
两个步骤,我们接着分析余下几个步骤。
直接上代码:
//真正实现向IOC容器获取Bean的功能,也是触发依赖注入功能的地方 protected <T> T doGetBean(final String name, @Nullable final Class<T> requiredType, @Nullable final Object[] args, boolean typeCheckOnly) throws BeansException { //根据指定的名称获取被管理Bean的名称,剥离指定名称中对容器的相关依赖 // 如果指定的是别名,将别名转换为规范的Bean名称 <1> final String beanName = transformedBeanName(name); Object bean; // Eagerly check singleton cache for manually registered singletons. // 从缓存中获取已被创建过的单例Bean <2> Object sharedInstance = getSingleton(beanName); //如果缓存中有 if (sharedInstance != null && args == null) { if (logger.isDebugEnabled()) { if (isSingletonCurrentlyInCreation(beanName)) { logger.debug("Returning eagerly cached instance of singleton bean '" + beanName + "' that is not fully initialized yet - a consequence of a circular reference"); } else { logger.debug("Returning cached instance of singleton bean '" + beanName + "'"); } } //注意:BeanFactory是管理容器中Bean的工厂 // FactoryBean是创建创建对象的工厂Bean,两者之间有区别 //获取给定Bean的实例对象,该对象要么是 bean 实例本身,要么就是 FactoryBean 创建的 Bean 对象 //(为什么要再次获取呢,因为上面获取的sharedInstance不一定是完整的) <3> bean = getObjectForBeanInstance(sharedInstance, name, beanName, null); } else { // Fail if we're already creating this bean instance: // We're assumably within a circular reference. // 因为 Spring 只解决单例模式下的循环依赖,在原型模式下如果存在循环依赖则会抛出异常。 <4> if (isPrototypeCurrentlyInCreation(beanName)) { throw new BeanCurrentlyInCreationException(beanName); } // Check if bean definition exists in this factory. //对IOC容器中是否存在指定名称的BeanDefinition进行检查,首先检查是否 //能在当前的BeanFactory中获取的所需要的Bean,如果不能则委托当前容器 //的父级容器去查找,如果还是找不到则沿着容器的继承体系向父级容器查找 BeanFactory parentBeanFactory = getParentBeanFactory(); //当前容器的父级容器存在,且当前容器中不存在指定名称的Bean if (parentBeanFactory != null && !containsBeanDefinition(beanName)) { // Not found -> check parent. //解析指定Bean名称的原始名称 String nameToLookup = originalBeanName(name); // 若为 AbstractBeanFactory 类型,委托父类处理 if (parentBeanFactory instanceof AbstractBeanFactory) { return ((AbstractBeanFactory) parentBeanFactory).doGetBean( nameToLookup, requiredType, args, typeCheckOnly); } else if (args != null) { // Delegation to parent with explicit args. //委派父级容器根据指定名称和显式的参数查找 return (T) parentBeanFactory.getBean(nameToLookup, args); } else { // No args -> delegate to standard getBean method. //委派父级容器根据指定名称和类型查找 return parentBeanFactory.getBean(nameToLookup, requiredType); } } // 创建的Bean是否需要进行类型验证,一般不需要 <5> if (!typeCheckOnly) { //向容器标记指定的Bean已经被创建 markBeanAsCreated(beanName); } try { //从容器中获取 beanName 相应的 GenericBeanDefinition 对象,并将其转换为 RootBeanDefinition 对象 // 主要解决Bean继承时子类合并父类公共属性问题 <6> final RootBeanDefinition mbd = getMergedLocalBeanDefinition(beanName); // 检查给定的合并的 BeanDefinition (是否为抽象类) checkMergedBeanDefinition(mbd, beanName, args); // Guarantee initialization of beans that the current bean depends on. // 处理所依赖的 bean @DependsOn() // 获取当前Bean所有依赖Bean的名称 <7> String[] dependsOn = mbd.getDependsOn(); //如果有依赖 if (dependsOn != null) { for (String dep : dependsOn) { //校验该依赖是否已经注册过给当前 Bean if (isDependent(beanName, dep)) { //已注册,抛出异常 throw new BeanCreationException(mbd.getResourceDescription(), beanName, "Circular depends-on relationship between '" + beanName + "' and '" + dep + "'"); } //没有,则先注册依赖的bean registerDependentBean(dep, beanName); //递归调用getBean(),先生成依赖的bean getBean(dep); } } // Create bean instance. //创建单例Bean <8> if (mbd.isSingleton()) { //这里使用了一个匿名内部类,创建Bean实例对象,并且注册给所依赖的对象 sharedInstance = getSingleton(beanName, () -> { try { //创建一个指定Bean实例对象,如果有父级继承,则合并子类和父类的定义 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. //显式地从容器单例模式Bean缓存中清除实例对象 destroySingleton(beanName); throw ex; } }); //获取给定Bean的实例对象 bean = getObjectForBeanInstance(sharedInstance, name, beanName, mbd); } //创建多例Bean else if (mbd.isPrototype()) { //原型模式(Prototype)是每次都会创建一个新的对象 Object prototypeInstance = null; try { //加载前置处理,默认的功能是注册当前创建的原型对象 beforePrototypeCreation(beanName); //创建指定Bean对象实例 prototypeInstance = createBean(beanName, mbd, args); } finally { //加载后置处理,默认的功能告诉IOC容器指定Bean的原型对象不再创建 afterPrototypeCreation(beanName); } //获取给定Bean的实例对象 bean = getObjectForBeanInstance(prototypeInstance, name, beanName, mbd); } //要创建的Bean既不是Singleton也不是Prototype //如:request、session、application等生命周期 else { String scopeName = mbd.getScope(); final Scope scope = this.scopes.get(scopeName); //Bean定义资源中没有配置生命周期范围,则Bean定义不合法 if (scope == null) { throw new IllegalStateException("No Scope registered for scope name '" + scopeName + "'"); } try { //这里又使用了一个匿名内部类,获取一个指定生命周期范围的实例 Object scopedInstance = scope.get(beanName, () -> { //前置处理 beforePrototypeCreation(beanName); try { return createBean(beanName, mbd, args); } finally { //后置处理 afterPrototypeCreation(beanName); } }); //获取给定Bean的实例对象 bean = getObjectForBeanInstance(scopedInstance, name, beanName, mbd); } catch (IllegalStateException ex) { throw new BeanCreationException(beanName, "Scope '" + scopeName + "' is not active for the current thread; consider " + "defining a scoped proxy for this bean if you intend to refer to it from a singleton", ex); } } } catch (BeansException ex) { cleanupAfterBeanCreationFailure(beanName); throw ex; } } // Check if required type matches the type of the actual bean instance. //对创建的Bean实例对象进行类型检查 <9> if (requiredType != null && !requiredType.isInstance(bean)) { try { T convertedBean = getTypeConverter().convertIfNecessary(bean, requiredType); if (convertedBean == null) { throw new BeanNotOfRequiredTypeException(name, requiredType, bean.getClass()); } return convertedBean; } catch (TypeMismatchException ex) { if (logger.isDebugEnabled()) { logger.debug("Failed to convert bean '" + name + "' to required type '" + ClassUtils.getQualifiedName(requiredType) + "'", ex); } throw new BeanNotOfRequiredTypeException(name, requiredType, bean.getClass()); } } return (T) bean; } 复制代码
代码很长,需要一些耐心,下面我们来逐步分析这段代码:
<1>处:具体分析,见 2.1获取原始beanName
<2>处: 具体分析,见 2.2从缓存中获取单例bean
<3>处: 具体分析,见 2.3获取最终的bean实例对象
<4>处: 具体分析,见 2.4原型模式依赖检查(Prototype)和从 parentBeanFactory 获取 Bean
<5>处: 具体分析,见 2.5标记bean为已创建或即将创建
<6>处: 具体分析,见 2.6获取BeanDefinition
<7>处: 具体分析,见 2.7bean依赖处理
<8>处: 具体分析,见 2.8不同作用域bean的实例化
<9>处: 具体分析,见 2.9类型转换
原型模式依赖检查,对应代码如下:
if (isPrototypeCurrentlyInCreation(beanName)) { throw new BeanCurrentlyInCreationException(beanName); } 复制代码
跟踪进去:
/** Names of beans that are currently in creation */ private final ThreadLocal<Object> prototypesCurrentlyInCreation = new NamedThreadLocal<>("Prototype beans currently in creation"); protected boolean isPrototypeCurrentlyInCreation(String beanName) { //从ThreadLocal中取出正在创建的prototype Object curVal = this.prototypesCurrentlyInCreation.get(); return (curVal != null && (curVal.equals(beanName) || (curVal instanceof Set && ((Set<?>) curVal).contains(beanName)))); } 复制代码
Spring 只处理单例模式下得循环依赖,对于原型模式的循环依赖直接抛出异常。
Spring会把正在创建的 原型模式
Bean存入 ThreadLoacl
,在这里通过ThreadLoacl来判断当前Bean是否已经创建。
从 parentBeanFactory 获取 Bean,对应代码如下:
// Check if bean definition exists in this factory. //对IOC容器中是否存在指定名称的BeanDefinition进行检查,首先检查是否 //能在当前的BeanFactory中获取的所需要的Bean,如果不能则委托当前容器 //的父级容器去查找,如果还是找不到则沿着容器的继承体系向父级容器查找 BeanFactory parentBeanFactory = getParentBeanFactory(); //当前容器的父级容器存在,且当前容器中不存在指定名称的Bean if (parentBeanFactory != null && !containsBeanDefinition(beanName)) { // Not found -> check parent. //解析指定Bean名称的原始名称 String nameToLookup = originalBeanName(name); // 若为 AbstractBeanFactory 类型,委托父类处理 if (parentBeanFactory instanceof AbstractBeanFactory) { return ((AbstractBeanFactory) parentBeanFactory).doGetBean( nameToLookup, requiredType, args, typeCheckOnly); } else if (args != null) { // Delegation to parent with explicit args. //委派父级容器根据指定名称和显式的参数查找 return (T) parentBeanFactory.getBean(nameToLookup, args); } else { // No args -> delegate to standard getBean method. //委派父级容器根据指定名称和类型查找 return parentBeanFactory.getBean(nameToLookup, requiredType); } } 复制代码
如果当前容器缓存中没有相对应的 BeanDefinition 对象,则会尝试从父类工厂(parentBeanFactory)中加载,然后再去递归调用 getBean(...) 方法
对应代码如下:
//创建的Bean是否需要进行类型验证,一般不需要 if (!typeCheckOnly) { //向容器标记指定的Bean已经被创建 markBeanAsCreated(beanName); } 复制代码
typeCheckOnly
是doGetBean(final String name, @Nullable final Class requiredType,@Nullable final Object[] args, boolean typeCheckOnly)方法中的一个参数,一般这个参数传的都是false
接着追踪 markBeanAsCreated()
方法:
protected void markBeanAsCreated(String beanName) { // 没有创建 if (!this.alreadyCreated.contains(beanName)) { synchronized (this.mergedBeanDefinitions) { // 再次检查一次:DCL 双重校验 if (!this.alreadyCreated.contains(beanName)) { clearMergedBeanDefinition(beanName); // 添加到已创建 bean 集合中 this.alreadyCreated.add(beanName); } } } } 复制代码
这里用到了单例模式中耳熟能详的双重校验
对应代码如下:
//从容器中获取 beanName 相应的 GenericBeanDefinition 对象,并将其转换为 RootBeanDefinition 对象 //主要解决Bean继承时子类合并父类公共属性问题 final RootBeanDefinition mbd = getMergedLocalBeanDefinition(beanName); // 检查给定的合并的 BeanDefinition (是否为抽象类) checkMergedBeanDefinition(mbd, beanName, args); 复制代码
这段代码注释很详细,就不多解释了。
对应代码如下:
// Guarantee initialization of beans that the current bean depends on. // 处理所依赖的 bean @DependsOn() //获取当前Bean所有依赖Bean的名称 <1> String[] dependsOn = mbd.getDependsOn(); //如果有依赖 if (dependsOn != null) { for (String dep : dependsOn) { //校验该依赖是否已经注册过给当前 Bean <2> if (isDependent(beanName, dep)) { //已注册,抛出异常 throw new BeanCreationException(mbd.getResourceDescription(), beanName, "Circular depends-on relationship between '" + beanName + "' and '" + dep + "'"); } //没有,则先注册依赖的bean <3> registerDependentBean(dep, beanName); //递归调用getBean(),先生成依赖的bean <4> getBean(dep); } } 复制代码
在spring中有一个 @DependsOn
注解,它的作用是依赖加载,比如A对象要在B对象加载之后才能加载,那么可以在A上面加 @DependsOn(value = "B")
注解,就可以达到我们的要求。
其实 @DependsOn
实现的原理就是上面这段代码。
<1>、通过我们前面从IoC容器中拿到的 BeanDefinition
,调用 mbd.getDependsOn()
方法,获取当前bean所有的依赖。
<2>、遍历这些依赖,判断此依赖是否已注册给当前的Bean
<3>、没有,则先注册依赖的Bean
<4>、递归调用getBean(),先生成依赖的bean
代码:
// 保存的是bean与其依赖的映射关系:B - > A private final Map<String, Set<String>> dependentBeanMap = new ConcurrentHashMap<>(64); //保存的是bean与其依赖的映射关系:A - > B private final Map<String, Set<String>> dependenciesForBeanMap = new ConcurrentHashMap<>(64); private boolean isDependent(String beanName, String dependentBeanName, @Nullable Set<String> alreadySeen) { if (alreadySeen != null && alreadySeen.contains(beanName)) { return false; } // 获取当前原始 beanName String canonicalName = canonicalName(beanName); // 获取该bean依赖的其他bean集合 Set<String> dependentBeans = this.dependentBeanMap.get(canonicalName); if (dependentBeans == null) { return false; } // 存在,则证明该依赖已经注册到bean中 if (dependentBeans.contains(dependentBeanName)) { return true; } // 递归检测依赖 for (String transitiveDependency : dependentBeans) { if (alreadySeen == null) { alreadySeen = new HashSet<>(); } alreadySeen.add(beanName); if (isDependent(transitiveDependency, dependentBeanName, alreadySeen)) { return true; } } return false; } 复制代码
这段代码很简单,主要就是通过 dependentBeanMap
获取当前bean对应的所有依赖 dependentBeans
,然后判断是否已注册,接着递归检查依赖的Bean有没有依赖,如果有,就递归调用 isDependent()
检查
如果没有注册依赖的Bean到该 Bean,则执行注册 registerDependentBean(dep, beanName)
:
// 保存的是bean与其依赖的映射关系:B - > A private final Map<String, Set<String>> dependentBeanMap = new ConcurrentHashMap<>(64); //保存的是bean与其依赖的映射关系:A - > B private final Map<String, Set<String>> dependenciesForBeanMap = new ConcurrentHashMap<>(64); //为指定的Bean注入依赖的Bean public void registerDependentBean(String beanName, String dependentBeanName) { // A quick check for an existing entry upfront, avoiding synchronization... //获取原始beanName String canonicalName = canonicalName(beanName); Set<String> dependentBeans = this.dependentBeanMap.get(canonicalName); if (dependentBeans != null && dependentBeans.contains(dependentBeanName)) { return; } // No entry yet -> fully synchronized manipulation of the dependentBeans Set //先从容器中:bean名称-->全部依赖Bean名称集合找查找给定名称Bean的依赖Bean synchronized (this.dependentBeanMap) { //获取给定名称Bean的所有依赖Bean名称 dependentBeans = this.dependentBeanMap.get(canonicalName); if (dependentBeans == null) { //为Bean设置依赖Bean信息 dependentBeans = new LinkedHashSet<>(8); this.dependentBeanMap.put(canonicalName, dependentBeans); } //把映射关系存入集合 dependentBeans.add(dependentBeanName); } //从容器中:bean名称-->指定名称Bean的依赖Bean集合找查找给定名称Bean的依赖Bean synchronized (this.dependenciesForBeanMap) { Set<String> dependenciesForBean = this.dependenciesForBeanMap.get(dependentBeanName); if (dependenciesForBean == null) { dependenciesForBean = new LinkedHashSet<>(8); this.dependenciesForBeanMap.put(dependentBeanName, dependenciesForBean); } //把映射关系存入集合 dependenciesForBean.add(canonicalName); } } 复制代码
套用上面的例子,如果 A
@DependsOn(value = "B")
,也就是说A依赖于B,那么该方法 registerDependentBean(dep, beanName)
中,参数 dep 就是B,beanName 就是A。
这段代码中其实就是把bean之间的依赖关系注册到两个map中。
dependentBeanMap 存入(B,A)
dependenciesForBeanMap 存入(A,B)
到了这一步,递归调用getBean(beanName)方法也就是doGetBean(beanName)重走当前流程,来先实例化依赖的Bean。等依赖的Bean实例化之后,当前bean再接着往下执行。
代码:
// Create bean instance. //创建单例Bean if (mbd.isSingleton()) { //这里使用了一个匿名内部类,创建Bean实例对象,并且注册给所依赖的对象 sharedInstance = getSingleton(beanName, () -> { try { //创建一个指定Bean实例对象,如果有父级继承,则合并子类和父类的定义 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. //显式地从容器单例模式Bean缓存中清除实例对象 destroySingleton(beanName); throw ex; } }); //获取给定Bean的实例对象 bean = getObjectForBeanInstance(sharedInstance, name, beanName, mbd); } //创建多例Bean else if (mbd.isPrototype()) { //原型模式(Prototype)是每次都会创建一个新的对象 Object prototypeInstance = null; try { //加载前置处理,默认的功能是注册当前创建的原型对象 beforePrototypeCreation(beanName); //创建指定Bean对象实例 prototypeInstance = createBean(beanName, mbd, args); } finally { //加载后置处理,默认的功能告诉IOC容器指定Bean的原型对象不再创建 afterPrototypeCreation(beanName); } //获取给定Bean的实例对象 bean = getObjectForBeanInstance(prototypeInstance, name, beanName, mbd); } //要创建的Bean既不是Singleton也不是Prototype //如:request、session、application等生命周期 else { String scopeName = mbd.getScope(); final Scope scope = this.scopes.get(scopeName); //Bean定义资源中没有配置生命周期范围,则Bean定义不合法 if (scope == null) { throw new IllegalStateException("No Scope registered for scope name '" + scopeName + "'"); } try { //这里又使用了一个匿名内部类,获取一个指定生命周期范围的实例 Object scopedInstance = scope.get(beanName, () -> { //前置处理 beforePrototypeCreation(beanName); try { return createBean(beanName, mbd, args); } finally { //后置处理 afterPrototypeCreation(beanName); } }); //获取给定Bean的实例对象 bean = getObjectForBeanInstance(scopedInstance, name, beanName, mbd); } catch (IllegalStateException ex) { throw new BeanCreationException(beanName, "Scope '" + scopeName + "' is not active for the current thread; consider " + "defining a scoped proxy for this bean if you intend to refer to it from a singleton", ex); } } 复制代码
这段代码很明显,分成了3个部分:
singleton Bean实例化
Prototype Bean实例化
其他类型 Bean 实例化(session,request等)
我们先来看singleton Bean实例化:
if (mbd.isSingleton()) { //这里使用了一个匿名内部类,创建Bean实例对象,并且注册给所依赖的对象 sharedInstance = getSingleton(beanName, () -> { try { //创建一个指定Bean实例对象,如果有父级继承,则合并子类和父类的定义 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. //显式地从容器单例模式Bean缓存中清除实例对象 destroySingleton(beanName); throw ex; } }); //获取给定Bean的实例对象 bean = getObjectForBeanInstance(sharedInstance, name,beanName, mbd); } 复制代码
Spring Bean 的作用域默认为 singleton 。还有其他作用域,如 prototype、request、session 等。
不同的作用域会有不同的初始化策略。
详见
spring各个scope的bean创建
。
代码:
// Check if required type matches the type of the actual bean instance. //对创建的Bean实例对象进行类型检查 if (requiredType != null && !requiredType.isInstance(bean)) { try { //执行转换 T convertedBean = getTypeConverter().convertIfNecessary(bean, requiredType); // 转换失败,抛异常 if (convertedBean == null) { throw new BeanNotOfRequiredTypeException(name, requiredType, bean.getClass()); } return convertedBean; } catch (TypeMismatchException ex) { if (logger.isDebugEnabled()) { logger.debug("Failed to convert bean '" + name + "' to required type '" + ClassUtils.getQualifiedName(requiredType) + "'", ex); } throw new BeanNotOfRequiredTypeException(name, requiredType, bean.getClass()); } } return (T) bean; 复制代码
requiredType
是 getBean()
方法可传入的一个参数,即可以根据指定的 beanName 和 requiredType 来获取Bean。
但是一般情况下是不需要类型检查的, requiredType
一般为null,如 getBean(beanName)
当 requiredType
不为null的时候走这段逻辑。
总结:
至此,spring加载Bean也就是 getBean() 我们大致分析完了,之后会再写几篇文章对其中有些步骤进行详细介绍。
参考:
芋道源码