Spring框架系列 - Spring AOP实现原理详解之AOP切面的实现
Posted realpdai
tags:
篇首语：本文由小常识网(cha138.com)小编为大家整理，主要介绍了Spring框架系列 - Spring AOP实现原理详解之AOP切面的实现相关的知识，希望对你有一定的参考价值。
前文，我们分析了Spring IOC的初始化过程和Bean的生命周期等，而Spring AOP也是基于IOC的Bean加载来实现的。本文主要介绍Spring AOP原理解析的切面实现过程（将切面类的所有切面方法根据使用的注解生成对应Advice，并将Advice连同切入点匹配器和切面类等信息一并封装到Advisor，为后续交给代理增强实现做准备的过程）。@pdai
引入
我们应该从哪里开始着手看Spring AOP的源码呢？和我们上文分析的IOC源码实现有什么关系呢？
- 前文中我们写了AOP的Demo，根据其XML配置我们不难发现AOP是基于IOC的Bean加载来实现的；这便使我们的主要入口
所以理解Spring AOP的初始化必须要先理解Spring IOC的初始化。
- 然后我们就能找到如下初始化的流程和aop对应的handler类
即parseCustomElement方法找到parse aop:aspectj-autoproxy的handler(org.springframework.aop.config.AopNamespaceHandler)
（PS：其实你会发现，最重要的是知识点的关联关系，而不是知识点本身，就后续代码而言不就是打个断点慢慢看的事了么。）
aop配置标签的解析
上文中，我们找到了AopNamespaceHandler，其实就是注册BeanDefinition的解析器BeanDefinitionParser，将
aop:xxxxxx配置标签交给指定的parser来处理。
/**
* Register the @link BeanDefinitionParser BeanDefinitionParsers for the
* @code config, @code spring-configured, @code aspectj-autoproxy
* and @code scoped-proxy tags.
*/
@Override
()
// In 2.0 XSD as well as in 2.5+ XSDs
// 注册解析<aop:config> 配置
registerBeanDefinitionParser("config", ConfigBeanDefinitionParser());
// 注册解析<aop:aspectj-autoproxy> 配置
registerBeanDefinitionParser("aspectj-autoproxy", AspectJAutoProxyBeanDefinitionParser());
registerBeanDefinitionDecorator("scoped-proxy", ScopedProxyBeanDefinitionDecorator());
// Only in 2.0 XSD: moved to context namespace in 2.5+
registerBeanDefinitionParser("spring-configured", SpringConfiguredBeanDefinitionParser());
config配置标签的解析
<aop:config/>由ConfigBeanDefinitionParser这个类处理，作为parser类最重要的就是parse方法
@Override
@Nullable
BeanDefinition (Element element, ParserContext parserContext)
CompositeComponentDefinition compositeDef =
CompositeComponentDefinition(element.getTagName(), parserContext.extractSource(element));
parserContext.pushContainingComponent(compositeDef);
configureAutoProxyCreator(parserContext, element);
List<Element> childElts = DomUtils.getChildElements(element);
(Element elt: childElts)
String localName = parserContext.getDelegate().getLocalName(elt);
(POINTCUT.equals(localName))
parsePointcut(elt, parserContext);
(ADVISOR.equals(localName))
parseAdvisor(elt, parserContext);
(ASPECT.equals(localName))
parseAspect(elt, parserContext);
parserContext.popAndRegisterContainingComponent();
;
打个断点看下
parseAspect的方法如下, 处理方法不难，我这里就不展开了
(Element aspectElement, ParserContext parserContext)
String aspectId = aspectElement.getAttribute(ID);
String aspectName = aspectElement.getAttribute(REF);
.parseState.push( AspectEntry(aspectId, aspectName));
List<BeanDefinition> beanDefinitions = ArrayList<>();
List<BeanReference> beanReferences = ArrayList<>();
List<Element> declareParents = DomUtils.getChildElementsByTagName(aspectElement, DECLARE_PARENTS);
( i = METHOD_INDEX; i < declareParents.size(); i++)
Element declareParentsElement = declareParents.get(i);
beanDefinitions.add(parseDeclareParents(declareParentsElement, parserContext));
// We have to parse "advice" and all the advice kinds in one loop, to get the
// ordering semantics right.
NodeList nodeList = aspectElement.getChildNodes();
adviceFoundAlready = ;
( i = 0; i < nodeList.getLength(); i++)
Node node = nodeList.item(i);
(isAdviceNode(node, parserContext))
(!adviceFoundAlready)
adviceFoundAlready = ;
(!StringUtils.hasText(aspectName))
parserContext.getReaderContext().error(
"<aspect> tag needs aspect bean reference via ref attribute when declaring advices.",
aspectElement, .parseState.snapshot());
;
beanReferences.add( RuntimeBeanReference(aspectName));
AbstractBeanDefinition advisorDefinition = parseAdvice(
aspectName, i, aspectElement, (Element) node, parserContext, beanDefinitions, beanReferences);
beanDefinitions.add(advisorDefinition);
AspectComponentDefinition aspectComponentDefinition = createAspectComponentDefinition(
aspectElement, aspectId, beanDefinitions, beanReferences, parserContext);
parserContext.pushContainingComponent(aspectComponentDefinition);
List<Element> pointcuts = DomUtils.getChildElementsByTagName(aspectElement, POINTCUT);
(Element pointcutElement : pointcuts)
parsePointcut(pointcutElement, parserContext);
parserContext.popAndRegisterContainingComponent();
.parseState.pop();
aspectj-autoproxy配置标签的解析
<aop:aspectj-autoproxy/>则由AspectJAutoProxyBeanDefinitionParser这个类处理的，我们看下parse 方法
@Override
@Nullable
BeanDefinition (Element element, ParserContext parserContext)
// 注册AspectJAnnotationAutoProxyCreator
AopNamespaceUtils.registerAspectJAnnotationAutoProxyCreatorIfNecessary(parserContext, element);
// 拓展BeanDefinition
extendBeanDefinition(element, parserContext);
;
AopNamespaceUtils.registerAspectJAnnotationAutoProxyCreatorIfNecessary方法对应如下
(
ParserContext parserContext, Element sourceElement)
BeanDefinition beanDefinition = AopConfigUtils.registerAspectJAnnotationAutoProxyCreatorIfNecessary(
parserContext.getRegistry(), parserContext.extractSource(sourceElement));
useClassProxyingIfNecessary(parserContext.getRegistry(), sourceElement);
registerComponentIfNecessary(beanDefinition, parserContext);
AopConfigUtils.registerAspectJAnnotationAutoProxyCreatorIfNecessary对应如下
@Nullable
BeanDefinition (
BeanDefinitionRegistry registry, @Nullable Object source)
registerOrEscalateApcAsRequired(AnnotationAwareAspectJAutoProxyCreator., , );
到这里，我们发现AOP的创建工作是交给AnnotationAwareAspectJAutoProxyCreator来完成的。
注解切面代理创建类(AnnotationAwareAspectJAutoProxyCreator)
AnnotationAwareAspectJAutoProxyCreator是如何工作的呢？这时候我们就要看AnnotationAwareAspectJAutoProxyCreator类结构关系了。
如下是类结构关系
它实现了两类接口：
- BeanFactoryAware属于Bean级生命周期接口方法
- InstantiationAwareBeanPostProcessor 和 BeanPostProcessor 这两个接口实现，一般称它们的实现类为“后处理器”，是容器级生命周期接口方法；
结合前文Spring Bean生命周期的流程
我们就可以定位到核心的初始化方法肯定在postProcessBeforeInstantiation和postProcessAfterInitialization中。
postProcessBeforeInstantiation
如下是上述类结构中postProcessBeforeInstantiation的方法，读者在自己看代码的时候建议打个断点看，可以方便理解
@Override
Object (Class<?> beanClass, String beanName)
Object cacheKey = getCacheKey(beanClass, beanName);
(!StringUtils.hasLength(beanName) || !.targetSourcedBeans.contains(beanName))
// 如果已经在缓存中，则忽略
(.advisedBeans.containsKey(cacheKey))
;
// 是否是aop基础类？是否跳过？
(isInfrastructureClass(beanClass) || shouldSkip(beanClass, beanName))
.advisedBeans.put(cacheKey, Boolean.FALSE);
;
// Create proxy here if we have a custom TargetSource.
// Suppresses unnecessary default instantiation of the target bean:
// The TargetSource will handle target instances in a custom fashion.
TargetSource targetSource = getCustomTargetSource(beanClass, beanName);
(targetSource != )
(StringUtils.hasLength(beanName))
.targetSourcedBeans.add(beanName);
Object[] specificInterceptors = getAdvicesAndAdvisorsForBean(beanClass, beanName, targetSource);
Object proxy = createProxy(beanClass, beanName, specificInterceptors, targetSource);
.proxyTypes.put(cacheKey, proxy.getClass());
proxy;
;
判断是否是aop基础类
是否是aop基础类的判断方法 isInfrastructureClass 如下
@Override
(Class<?> beanClass)
// Previously we setProxyTargetClass(true) in the constructor, but that has too
// broad an impact. Instead we now override isInfrastructureClass to avoid proxying
// aspects. Im not entirely happy with that as there is no good reason not
// to advise aspects, except that it causes advice invocation to go through a
// proxy, and if the aspect implements e.g the Ordered interface it will be
// proxied by that interface and fail at runtime as the advice method is not
// defined on the interface. We could potentially relax the restriction about
// not advising aspects in the future.
// 父类判断它是aop基础类 or 使用@Aspect注解
(.isInfrastructureClass(beanClass) ||
(.aspectJAdvisorFactory != && .aspectJAdvisorFactory.isAspect(beanClass)));
父类判断它是否是aop基础类的方法 super.isInfrastructureClass(beanClass), 本质上就是判断该类是否实现了Advice, Pointcut, Advisor或者AopInfrastructureBean接口。
(Class<?> beanClass)
// 该类是否实现了Advice, Pointcut, Advisor或者AopInfrastructureBean接口
retVal = Advice..() ||
..() ||
..() ||
..();
(retVal && logger.isTraceEnabled())
logger.trace("Did not attempt to auto-proxy infrastructure class [" + beanClass.getName() + "]");
retVal;
是否应该跳过shouldSkip
通过断点辅助，candidateAdvisors是就是xml配置的通知是对应的
@Override
(Class<?> beanClass, String beanName)
// TODO: Consider optimization by caching the list of the aspect names
List<Advisor> candidateAdvisors = findCandidateAdvisors();
(Advisor advisor : candidateAdvisors)
(advisor AspectJPointcutAdvisor &&
((AspectJPointcutAdvisor) advisor).getAspectName().equals(beanName))
;
.shouldSkip(beanClass, beanName);
切面方法转成Advisor
findCandidateAdvisors方法如下：
@Override
List<Advisor> ()
// 在父类中找到所有的advisor：基于xml配置的<aop:before/>生成的
List<Advisor> advisors = .findCandidateAdvisors();
// 为bean Factory中AspectJ切面构建advistor：通过AspectJ注解的方式生成Advisor类
(.aspectJAdvisorsBuilder != )
advisors.addAll(.aspectJAdvisorsBuilder.buildAspectJAdvisors());
advisors;
在当前的bean Factory中通过AspectJ注解的方式生成Advisor类，buildAspectJAdvisors方法如下
/**
* Look for AspectJ-annotated aspect beans in the current bean factory,
* and return to a list of Spring AOP Advisors representing them.
* <p>Creates a Spring Advisor for each AspectJ advice method.
* @return the list of @link org.springframework.aop.Advisor beans
* @see #isEligibleBean
*/
List<Advisor> ()
List<String> aspectNames = .aspectBeanNames;
(aspectNames == )
()
aspectNames = .aspectBeanNames;
(aspectNames == )
List<Advisor> advisors = ArrayList<>();
aspectNames = ArrayList<>();
String[] beanNames = BeanFactoryUtils.beanNamesForTypeIncludingAncestors(
.beanFactory, Object., , );
(String beanName : beanNames)
(!isEligibleBean(beanName))
;
// We must be careful not to instantiate beans eagerly as in this case they
// would be cached by the Spring container but would not have been weaved.
Class<?> beanType = .beanFactory.getType(beanName, );
(beanType == )
;
(.advisorFactory.isAspect(beanType))
aspectNames.add(beanName);
AspectMetadata amd = AspectMetadata(beanType, beanName);
(amd.getAjType().getPerClause().getKind() == PerClauseKind.SINGLETON)
MetadataAwareAspectInstanceFactory factory =
BeanFactoryAspectInstanceFactory(.beanFactory, beanName);
List<Advisor> classAdvisors = .advisorFactory.getAdvisors(factory);
// 单例加到advisorsCache, 非单例加到aspectFactoryCache
(.beanFactory.isSingleton(beanName))
.advisorsCache.put(beanName, classAdvisors);
.aspectFactoryCache.put(beanName, factory);
advisors.addAll(classAdvisors);
// Per target or per this.
(.beanFactory.isSingleton(beanName))
IllegalArgumentException("Bean with name " + beanName +
" is a singleton, but aspect instantiation model is not singleton");
MetadataAwareAspectInstanceFactory factory =
PrototypeAspectInstanceFactory(.beanFactory, beanName);
.aspectFactoryCache.put(beanName, factory);
// advisorFactory工厂获取advisors
advisors.addAll(.advisorFactory.getAdvisors(factory));
.aspectBeanNames = aspectNames;
advisors;
(aspectNames.isEmpty())
Collections.emptyList();
List<Advisor> advisors = ArrayList<>();
(String aspectName : aspectNames)
List<Advisor> cachedAdvisors = .advisorsCache.get(aspectName);
(cachedAdvisors != )
advisors.addAll(cachedAdvisors);
MetadataAwareAspectInstanceFactory factory = .aspectFactoryCache.get(aspectName);
advisors.addAll(.advisorFactory.getAdvisors(factory));
advisors;
上述方法本质上的思路是：用DCL双重锁的单例实现方式，拿到切面类里的切面方法，将其转换成advisor（并放入缓存中）。
转换的成advisor的方法是：this.advisorFactory.getAdvisors
@Override
List<Advisor> (MetadataAwareAspectInstanceFactory aspectInstanceFactory)
Class<?> aspectClass = aspectInstanceFactory.getAspectMetadata().getAspectClass();
String aspectName = aspectInstanceFactory.getAspectMetadata().getAspectName();
validate(aspectClass);
// We need to wrap the MetadataAwareAspectInstanceFactory with a decorator
// so that it will only instantiate once.
MetadataAwareAspectInstanceFactory lazySingletonAspectInstanceFactory =
LazySingletonAspectInstanceFactoryDecorator(aspectInstanceFactory);
List<Advisor> advisors = ArrayList<>();
(Method method : getAdvisorMethods(aspectClass))
// Prior to Spring Framework 5.2.7, advisors.size() was supplied as the declarationOrderInAspect
// to getAdvisor(...) to represent the "current position" in the declared methods list.
// However, since Java 7 the "current position" is not valid since the JDK no longer
// returns declared methods in the order in which they are declared in the source code.
// Thus, we now hard code the declarationOrderInAspect to 0 for all advice methods
// discovered via reflection in order to support reliable advice ordering across JVM launches.
// Specifically, a value of 0 aligns with the default value used in
// AspectJPrecedenceComparator.getAspectDeclarationOrder(Advisor).
Advisor advisor = getAdvisor(method, lazySingletonAspectInstanceFactory, 0, aspectName);
(advisor != )
advisors.add(advisor);
// If its a per target aspect, emit the dummy instantiating aspect.
(!advisors.isEmpty() && lazySingletonAspectInstanceFactory.getAspectMetadata().isLazilyInstantiated())
Advisor instantiationAdvisor = SyntheticInstantiationAdvisor(lazySingletonAspectInstanceFactory);
advisors.add(0, instantiationAdvisor);
// Find introduction fields.
(Field field : aspectClass.getDeclaredFields())
Advisor advisor = getDeclareParentsAdvisor(field);
(advisor != )
advisors.add(advisor);
advisors;
getAdvisor方法如下
@Override
@Nullable
Advisor (Method candidateAdviceMethod, MetadataAwareAspectInstanceFactory aspectInstanceFactory,
declarationOrderInAspect, String aspectName)
validate(aspectInstanceFactory.getAspectMetadata().getAspectClass());
AspectJExpressionPointcut expressionPointcut = getPointcut(
candidateAdviceMethod, aspectInstanceFactory.getAspectMetadata().getAspectClass());
(expressionPointcut == )
;
// 封装成advisor
InstantiationModelAwarePointcutAdvisorImpl(expressionPointcut, candidateAdviceMethod,
, aspectInstanceFactory, declarationOrderInAspect, aspectName);
获取表达式的切点
获取表达式的切点的方法getPointcut如下：
@Nullable
AspectJExpressionPointcut (Method candidateAdviceMethod, Class<?> candidateAspectClass)
AspectJAnnotation<?> aspectJAnnotation =
AbstractAspectJAdvisorFactory.findAspectJAnnotationOnMethod(candidateAdviceMethod);
(aspectJAnnotation == )
;
AspectJExpressionPointcut ajexp =
AspectJExpressionPointcut(candidateAspectClass, String[0], Class<?>[0]);
ajexp.setExpression(aspectJAnnotation.getPointcutExpression());
(.beanFactory != )
ajexp.setBeanFactory(.beanFactory);
ajexp;
AbstractAspectJAdvisorFactory.findAspectJAnnotationOnMethod的方法如下
Class<?>[] ASPECTJ_ANNOTATION_CLASSES = Class<?>[]
Pointcut., ., ., ., ., .;
/**
* Find and return the first AspectJ annotation on the given method
* (there <i>should</i> only be one anyway...).
*/
@SuppressWarnings("unchecked")
@Nullable
AspectJAnnotation<?> findAspectJAnnotationOnMethod(Method method)
(Class<?> clazz : ASPECTJ_ANNOTATION_CLASSES)
AspectJAnnotation<?> foundAnnotation = findAnnotation(method, (Class<Annotation>) clazz);
(foundAnnotation != )
foundAnnotation;
;
findAnnotation方法如下
@Nullable
<A extends Annotation> AspectJAnnotation<A> (Method method, Class<A> toLookFor)
A result = AnnotationUtils.findAnnotation(method, toLookFor);
(result != )
AspectJAnnotation<>(result);
;
AnnotationUtils.findAnnotation 获取注解方法如下
/**
* Find a single @link Annotation of @code annotationType on the supplied
* @link Method, traversing its super methods (i.e. from superclasses and
* interfaces) if the annotation is not <em>directly present</em> on the given
* method itself.
* <p>Correctly handles bridge @link Method Methods generated by the compiler.
* <p>Meta-annotations will be searched if the annotation is not
* <em>directly present</em> on the method.
* <p>Annotations on methods are not inherited by default, so we need to handle
* this explicitly.
* @param method the method to look for annotations on
* @param annotationType the annotation type to look for
* @return the first matching annotation, or @code null if not found
* @see #getAnnotation(Method, Class)
*/
@Nullable
<A extends Annotation> A (Method method, @Nullable Class<A> annotationType)
(annotationType == )
;
// Shortcut: directly present on the element, with no merging needed?
(AnnotationFilter.PLAIN.matches(annotationType) ||
AnnotationsScanner.hasPlainJavaAnnotationsOnly(method))
method.getDeclaredAnnotation(annotationType);
// Exhaustive retrieval of merged annotations...
MergedAnnotations.from(method, SearchStrategy.TYPE_HIERARCHY, RepeatableContainers.none())
.get(annotationType).withNonMergedAttributes()
.synthesize(MergedAnnotation::isPresent).orElse();
封装成Advisor
注：Advisor 是 advice的包装器，包含了advice及其它信息
由InstantiationModelAwarePointcutAdvisorImpl构造完成
(AspectJExpressionPointcut declaredPointcut,
Method aspectJAdviceMethod, AspectJAdvisorFactory aspectJAdvisorFactory,
MetadataAwareAspectInstanceFactory aspectInstanceFactory, declarationOrder, String aspectName)
.declaredPointcut = declaredPointcut;
.declaringClass = aspectJAdviceMethod.getDeclaringClass();
.methodName = aspectJAdviceMethod.getName();
.parameterTypes = aspectJAdviceMethod.getParameterTypes();
.aspectJAdviceMethod = aspectJAdviceMethod;
.aspectJAdvisorFactory = aspectJAdvisorFactory;
.aspectInstanceFactory = aspectInstanceFactory;
.declarationOrder = declarationOrder;
.aspectName = aspectName;
(aspectInstanceFactory.getAspectMetadata().isLazilyInstantiated())
// Static part of the pointcut is a lazy type.
Pointcut preInstantiationPointcut = Pointcuts.union(
aspectInstanceFactory.getAspectMetadata().getPerClausePointcut(), .declaredPointcut);
// Make it dynamic: must mutate from pre-instantiation to post-instantiation state.
// If its not a dynamic pointcut, it may be optimized out
// by the Spring AOP infrastructure after the first evaluation.
.pointcut = PerTargetInstantiationModelPointcut(
.declaredPointcut, preInstantiationPointcut, aspectInstanceFactory);
.lazy = ;
// A singleton aspect.
.pointcut = .declaredPointcut;
.lazy = ;
.instantiatedAdvice = instantiateAdvice(.declaredPointcut);
通过pointcut获取advice
Advice (AspectJExpressionPointcut pointcut)
Advice advice = .aspectJAdvisorFactory.getAdvice(.aspectJAdviceMethod, pointcut,
.aspectInstanceFactory, .declarationOrder, .aspectName);
(advice != ? advice : EMPTY_ADVICE);
交给aspectJAdvisorFactory获取
@Override
@Nullable
Advice (Method candidateAdviceMethod, AspectJExpressionPointcut expressionPointcut,
MetadataAwareAspectInstanceFactory aspectInstanceFactory, declarationOrder, String aspectName)
// 获取切面类
Class<?> candidateAspectClass = aspectInstanceFactory.getAspectMetadata().getAspectClass();
validate(candidateAspectClass);
// 获取切面注解
AspectJAnnotation<?> aspectJAnnotation =
AbstractAspectJAdvisorFactory.findAspectJAnnotationOnMethod(candidateAdviceMethod);
(aspectJAnnotation == )
;
// If we get here, we know we have an AspectJ method.
// Check that its an AspectJ-annotated class
(!isAspect(candidateAspectClass))
AopConfigException("Advice must be declared inside an aspect type: " +
"Offending method " + candidateAdviceMethod + " in class [" +
candidateAspectClass.getName() + "]");
(logger.isDebugEnabled())
logger.debug("Found AspectJ method: " + candidateAdviceMethod);
// 切面注解转换成advice
AbstractAspectJAdvice springAdvice;
(aspectJAnnotation.getAnnotationType())
AtPointcut: // AtPointcut忽略
(logger.isDebugEnabled())
logger.debug("Processing pointcut " + candidateAdviceMethod.getName() + "");
;
AtAround:
springAdvice = AspectJAroundAdvice(
candidateAdviceMethod, expressionPointcut, aspectInstanceFactory);
;
AtBefore:
springAdvice = AspectJMethodBeforeAdvice(
candidateAdviceMethod, expressionPointcut, aspectInstanceFactory);
;
AtAfter:
springAdvice = AspectJAfterAdvice(
candidateAdviceMethod, expressionPointcut, aspectInstanceFactory);
;
AtAfterReturning:
springAdvice = AspectJAfterReturningAdvice(
candidateAdviceMethod, expressionPointcut, aspectInstanceFactory);
AfterReturning afterReturningAnnotation = (AfterReturning) aspectJAnnotation.getAnnotation();
(StringUtils.hasText(afterReturningAnnotation.returning()))
springAdvice.setReturningName(afterReturningAnnotation.returning());
;
AtAfterThrowing:
springAdvice = AspectJAfterThrowingAdvice(
candidateAdviceMethod, expressionPointcut, aspectInstanceFactory);
AfterThrowing afterThrowingAnnotation = (AfterThrowing) aspectJAnnotation.getAnnotation();
(StringUtils.hasText(afterThrowingAnnotation.throwing()))
springAdvice.setThrowingName(afterThrowingAnnotation.throwing());
;
:
UnsupportedOperationException(
"Unsupported advice type on method: " + candidateAdviceMethod);
// 最后将其它切面信息配置到advice
springAdvice.setAspectName(aspectName);
springAdvice.setDeclarationOrder(declarationOrder);
String[] argNames = .parameterNameDiscoverer.getParameterNames(candidateAdviceMethod);
(argNames != )
springAdvice.setArgumentNamesFromStringArray(argNames);
springAdvice.calculateArgumentBindings();
springAdvice;
小结
回头看，主要是处理使用了@Aspect注解的切面类，然后将切面类的所有切面方法根据使用的注解生成对应Advice，并将Advice连同切入点匹配器和切面类等信息一并封装到Advisor的过程。
postProcessAfterInitialization
有了Adisor, 注入到合适的位置并交给代理（cglib和jdk)实现了。
/**
* Create a proxy with the configured interceptors if the bean is
* identified as one to proxy by the subclass.
* @see #getAdvicesAndAdvisorsForBean
*/
@Override
Object (@Nullable Object bean, String beanName)
(bean != )
Object cacheKey = getCacheKey(bean.getClass(), beanName);
(.earlyProxyReferences.remove(cacheKey) != bean)
wrapIfNecessary(bean, beanName, cacheKey);
bean;
后文中将分别介绍代理的创建和实现：
- Spring进阶 - Spring AOP实现原理详解之AOP代理的创建
- Spring进阶 - Spring AOP实现原理详解之Cglib代理实现
- Spring进阶 - Spring AOP实现原理详解之JDK代理实现
总结
通过上文的分析，我们做下小结：
- 由IOC Bean加载方法栈中找到parseCustomElement方法，找到parse
aop:aspectj-autoproxy的handler(org.springframework.aop.config.AopNamespaceHandler) - AopNamespaceHandler注册了
<aop:aspectj-autoproxy/>的解析类是AspectJAutoProxyBeanDefinitionParser - AspectJAutoProxyBeanDefinitionParser的parse 方法 通过AspectJAwareAdvisorAutoProxyCreator类去创建
- AspectJAwareAdvisorAutoProxyCreator实现了两类接口，BeanFactoryAware和BeanPostProcessor；根据Bean生命周期方法找到两个核心方法：postProcessBeforeInstantiation和postProcessAfterInitialization
- postProcessBeforeInstantiation：主要是处理使用了@Aspect注解的切面类，然后将切面类的所有切面方法根据使用的注解生成对应Advice，并将Advice连同切入点匹配器和切面类等信息一并封装到Advisor
- postProcessAfterInitialization：主要负责将Advisor注入到合适的位置，创建代理（cglib或jdk)，为后面给代理进行增强实现做准备。
更多文章
首先， 从Spring框架的整体架构和组成对整体框架有个认知。
- Spring是什么？它是怎么诞生的？有哪些主要的组件和核心功能呢? 本文通过这几个问题帮助你构筑Spring和Spring Framework的整体认知。
其次，通过案例引出Spring的核心（IoC和AOP），同时对IoC和AOP进行案例使用分析。
- 上文中我们简单介绍了Spring和Spring Framework的组件，那么这些Spring Framework组件是如何配合工作的呢？本文主要承接上文，向你展示Spring Framework组件的典型应用场景和基于这个场景设计出的简单案例，并以此引出Spring的核心要点，比如IOC和AOP等；在此基础上还引入了不同的配置方式， 如XML，Java配置和注解方式的差异。
- 在Spring基础 - Spring简单例子引入Spring的核心中向你展示了IoC的基础含义，同时以此发散了一些IoC相关知识点; 本节将在此基础上进一步解读IOC的含义以及IOC的使用方式
- 在Spring基础 - Spring简单例子引入Spring的核心中向你展示了AOP的基础含义，同时以此发散了一些AOP相关知识点; 本节将在此基础上进一步解读AOP的含义以及AOP的使用方式。
基于Spring框架和IOC，AOP的基础，为构建上层web应用，需要进一步学习SpringMVC。
- 前文我们介绍了Spring框架和Spring框架中最为重要的两个技术点（IOC和AOP），那我们如何更好的构建上层的应用呢（比如web 应用），这便是SpringMVC；Spring MVC是Spring在Spring Container Core和AOP等技术基础上，遵循上述Web MVC的规范推出的web开发框架，目的是为了简化Java栈的web开发。 本文主要介绍SpringMVC的请求流程和基础案例的编写和运行。
Spring进阶 - IoC，AOP以及SpringMVC的源码分析
- 在对IoC有了初步的认知后，我们开始对IOC的实现原理进行深入理解。本文将帮助你站在设计者的角度去看IOC最顶层的结构设计
- 上文，我们看了IOC设计要点和设计结构；紧接着这篇，我们可以看下源码的实现了：Spring如何实现将资源配置（以xml配置为例）通过加载，解析，生成BeanDefination并注册到IoC容器中的
- 上文，我们看了IOC设计要点和设计结构；以及Spring如何实现将资源配置（以xml配置为例）通过加载，解析，生成BeanDefination并注册到IoC容器中的；容器中存放的是Bean的定义即BeanDefinition放到beanDefinitionMap中，本质上是一个
ConcurrentHashMap<String, Object>；并且BeanDefinition接口中包含了这个类的Class信息以及是否是单例等。那么如何从BeanDefinition中实例化Bean对象呢，这是本文主要研究的内容？
- 前文，我们分析了Spring IOC的初始化过程和Bean的生命周期等，而Spring AOP也是基于IOC的Bean加载来实现的。本文主要介绍Spring AOP原理解析的切面实现过程(将切面类的所有切面方法根据使用的注解生成对应Advice，并将Advice连同切入点匹配器和切面类等信息一并封装到Advisor，为后续交给代理增强实现做准备的过程)。
- 上文我们介绍了Spring AOP原理解析的切面实现过程(将切面类的所有切面方法根据使用的注解生成对应Advice，并将Advice连同切入点匹配器和切面类等信息一并封装到Advisor)。本文在此基础上继续介绍，代理（cglib代理和JDK代理）的实现过程。
- 我们在前文中已经介绍了SpringAOP的切面实现和创建动态代理的过程，那么动态代理是如何工作的呢？本文主要介绍Cglib动态代理的案例和SpringAOP实现的原理。
- 上文我们学习了SpringAOP Cglib动态代理的实现，本文主要是SpringAOP JDK动态代理的案例和实现部分。
- 前文我们有了IOC的源码基础以及SpringMVC的基础，我们便可以进一步深入理解SpringMVC主要实现原理，包含DispatcherServlet的初始化过程和DispatcherServlet处理请求的过程的源码解析。本文是第一篇：DispatcherServlet的初始化过程的源码解析。
- 前文我们有了IOC的源码基础以及SpringMVC的基础，我们便可以进一步深入理解SpringMVC主要实现原理，包含DispatcherServlet的初始化过程和DispatcherServlet处理请求的过程的源码解析。本文是第二篇：DispatcherServlet处理请求的过程的源码解析。
以上是关于Spring框架系列 - Spring AOP实现原理详解之AOP切面的实现的主要内容，如果未能解决你的问题，请参考以下文章
Spring框架系列(11) - Spring AOP实现原理详解之Cglib代理实现