Spring源码情操陶冶-ComponentScanBeanDefinitionParser文件扫描解析器

承接前文Spring源码情操陶冶-自定义节点的解析，本文讲述spring通过context:component-scan节点干了什么事

ComponentScanBeanDefinitionParser#私有属性

罗列下context:component-scan可填的基础属性

	private static final String BASE_PACKAGE_ATTRIBUTE = "base-package";

	private static final String RESOURCE_PATTERN_ATTRIBUTE = "resource-pattern";

	private static final String USE_DEFAULT_FILTERS_ATTRIBUTE = "use-default-filters";

	private static final String ANNOTATION_CONFIG_ATTRIBUTE = "annotation-config";

	private static final String NAME_GENERATOR_ATTRIBUTE = "name-generator";

	private static final String SCOPE_RESOLVER_ATTRIBUTE = "scope-resolver";

	private static final String SCOPED_PROXY_ATTRIBUTE = "scoped-proxy";

	private static final String EXCLUDE_FILTER_ELEMENT = "exclude-filter";

	private static final String INCLUDE_FILTER_ELEMENT = "include-filter";

	private static final String FILTER_TYPE_ATTRIBUTE = "type";

	private static final String FILTER_EXPRESSION_ATTRIBUTE = "expression";

ComponentScanBeanDefinitionParser#parse()-主方法

统一接口parse()方法，看下总体逻辑，代码如下

@Override
	public BeanDefinition parse(Element element, ParserContext parserContext) {
		//解析base-package属性值，扫描的包可以,;分隔
		String[] basePackages = StringUtils.tokenizeToStringArray(element.getAttribute(BASE_PACKAGE_ATTRIBUTE),			ConfigurableApplicationContext.CONFIG_LOCATION_DELIMITERS);

		// Actually scan for bean definitions and register them.
		ClassPathBeanDefinitionScanner scanner = configureScanner(parserContext, element);
		//通过ClassPathBeanDefinitionScanner扫描类来获取包名下的所有class并将他们注册到spring的bean工厂中
		Set<BeanDefinitionHolder> beanDefinitions = scanner.doScan(basePackages);
		//注册其他注解组件
		registerComponents(parserContext.getReaderContext(), beanDefinitions, element);

		return null;
	}

我们关注下ComponentScanBeanDefinitionParser#configureScanner()创建扫描器操作和ClassPathBeanDefinitionScanner#doScan()扫描包方法

ComponentScanBeanDefinitionParser#configureScanner()-创建扫描器

观察下如何创建扫描器，以及相关的初始操作，代码奉上

protected ClassPathBeanDefinitionScanner configureScanner(ParserContext parserContext, Element element) {
		XmlReaderContext readerContext = parserContext.getReaderContext();
		//默认使用spring自带的注解过滤
		boolean useDefaultFilters = true;
		//解析`use-default-filters`，类型为boolean
		if (element.hasAttribute(USE_DEFAULT_FILTERS_ATTRIBUTE)) {
			useDefaultFilters = Boolean.valueOf(element.getAttribute(USE_DEFAULT_FILTERS_ATTRIBUTE));
		}

		// Delegate bean definition registration to scanner class.
		//此处如果`use-default-filters`为true，则添加`@Component`、`@Service`、`@Controller`、`@Repository`、`@ManagedBean`、`@Named`添加到includeFilters的集合过滤
		ClassPathBeanDefinitionScanner scanner = createScanner(readerContext, useDefaultFilters);
		scanner.setResourceLoader(readerContext.getResourceLoader());
		scanner.setEnvironment(parserContext.getDelegate().getEnvironment());
		scanner.setBeanDefinitionDefaults(parserContext.getDelegate().getBeanDefinitionDefaults());
		scanner.setAutowireCandidatePatterns(parserContext.getDelegate().getAutowireCandidatePatterns());
		//设置`resource-pattern`属性，扫描资源的模式匹配，支持正则表达式
		if (element.hasAttribute(RESOURCE_PATTERN_ATTRIBUTE)) {
                    scanner.setResourcePattern(element.getAttribute(RESOURCE_PATTERN_ATTRIBUTE));
		}

		try {
			//解析name-generator属性 beanName生成器
			parseBeanNameGenerator(element, scanner);
		}
		catch (Exception ex) {
			readerContext.error(ex.getMessage(), readerContext.extractSource(element), ex.getCause());
		}

		try {
			//解析scope-resolver属性和scoped-proxy属性，但两者只可存在其一
			//后者值为targetClass：cglib代理、interfaces：JDK代理、no：不使用代理
			parseScope(element, scanner);
		}
		catch (Exception ex) {
			readerContext.error(ex.getMessage(), readerContext.extractSource(element), ex.getCause());
		}
		//解析子节点`context:include-filter`、`context:exclude-filter`主要用于对扫描class类的过滤
               //例如<context:include-filter type="annotation" expression="org.springframework.stereotype.Controller.RestController" />
		parseTypeFilters(element, scanner, readerContext, parserContext);

		return scanner;
	}

此处只简单的罗列了如何创建一个文件扫描器以及相关的初始操作，具体的读者可自行去阅读分析

ClassPathBeanDefinitionScanner#doScan()-扫描操作

真实扫描base-package指定的目录并返回注册的所有beanDefinition，具体的扫描简析如下

protected Set<BeanDefinitionHolder> doScan(String... basePackages) {
		//表明base-package属性是需要被指定的
		Assert.notEmpty(basePackages, "At least one base package must be specified");
		Set<BeanDefinitionHolder> beanDefinitions = new LinkedHashSet<BeanDefinitionHolder>();
		for (String basePackage : basePackages) {
			//对每个基础包都进行扫描寻找并且对基础包下的所有class都注册为BeanDefinition
                        /**
                        **
                        **并对得到的candidates集合进行过滤，此处便用到include-filters和exclude-filters
                        */
			Set<BeanDefinition> candidates = findCandidateComponents(basePackage);
			for (BeanDefinition candidate : candidates) {
				//解析一个bean的scope属性，代表作用范围
				//prototype->每次请求都创建新的对象 singleton->单例模式，处理多请求
				ScopeMetadata scopeMetadata = this.scopeMetadataResolver.resolveScopeMetadata(candidate);
				candidate.setScope(scopeMetadata.getScopeName());
				//使用beanName生成器生成
				String beanName = this.beanNameGenerator.generateBeanName(candidate, this.registry);
				/**
				**对注册的bean进行另外的赋值处理，比如默认属性的配置
				*返回的candidate类型为ScannedGenericBeanDefinition，下面两者
				*条件满足
				*/
				if (candidate instanceof AbstractBeanDefinition) {
					//设置lazy-init/autowire-code默认属性，从spring配置的<beans>节点属性读取
					postProcessBeanDefinition((AbstractBeanDefinition) candidate, beanName);
				}
				if (candidate instanceof AnnotatedBeanDefinition) {
					//读取bean上的注解，比如`@Lazy`、`@Dependson`的值设置相应的属性
					AnnotationConfigUtils.processCommonDefinitionAnnotations((AnnotatedBeanDefinition) candidate);
				}
				//查看是否已注册
				if (checkCandidate(beanName, candidate)) {
					BeanDefinitionHolder definitionHolder = new BeanDefinitionHolder(candidate, beanName);
					//默认采取cglib来做代理
					definitionHolder = AnnotationConfigUtils.applyScopedProxyMode(scopeMetadata, definitionHolder, this.registry);
					beanDefinitions.add(definitionHolder);
					//注册bean信息到工厂中
					registerBeanDefinition(definitionHolder, this.registry);
				}
			}
		}
		return beanDefinitions;
	}

在这里我们只简单的看下其父类ClassPathScanningCandidateComponentProvider#findCandidateComponents获取包下的所有class资源文件并实例化为BeanDefinition对象

ClassPathScanningCandidateComponentProvider#findCandidateComponents()-找寻符合条件的资源文件

扫描包下的所有class文件并对其进行过滤，过滤的条件为includeFilters和excludeFilters集合。代码简析如下

public Set<BeanDefinition> findCandidateComponents(String basePackage) {
		Set<BeanDefinition> candidates = new LinkedHashSet<BeanDefinition>();
		try {
			//值类似为classpath*:com/question/sky/**/*.class
			String packageSearchPath = ResourcePatternResolver.CLASSPATH_ALL_URL_PREFIX +resolveBasePackage(basePackage) + "/" + this.resourcePattern;
			//通过PathMatchingResourcePatternResolver来找寻资源
			//常用的Resource为FileSystemResource
			Resource[] resources = this.resourcePatternResolver.getResources(packageSearchPath);
			boolean traceEnabled = logger.isTraceEnabled();
			boolean debugEnabled = logger.isDebugEnabled();
			for (Resource resource : resources) {

				if (resource.isReadable()) {
					try {
						//生成MetadataReader对象->SimpleMetadataReader，内部包含AnnotationMetadataReadingVisitor注解访问处理类
						MetadataReader metadataReader = this.metadataReaderFactory.getMetadataReader(resource);
						//判断class是否不属于excludeFilters集合内但至少符合一个includeFilters集合
						if (isCandidateComponent(metadataReader)) {
							//包装为ScannedGenericBeanDefinition对象
							ScannedGenericBeanDefinition sbd = new ScannedGenericBeanDefinition(metadataReader);
							//保存文件资源
							sbd.setResource(resource);
							sbd.setSource(resource);
							//判断class文件是否不为接口或者抽象类并且是独立的
							if (isCandidateComponent(sbd)) {
								//完成验证加入集合中
								candidates.add(sbd);
							}
						}
					}
					catch (Throwable ex) {
						throw new BeanDefinitionStoreException(
								"Failed to read candidate component class: " + resource, ex);
					}
				}
			}
		}
		catch (IOException ex) {
			throw new BeanDefinitionStoreException("I/O failure during classpath scanning", ex);
		}
		return candidates;
	}

对上面的代码解释作下补充，主要是验证beanDefinition的两个方法

• ClassPathScanningCandidateComponentProvider#isCandidateComponent(MetadataReader metadataReader)
  
  对class类进行filter集合过滤

protected boolean isCandidateComponent(MetadataReader metadataReader) throws IOException {
		for (TypeFilter tf : this.excludeFilters) {
			//满足excludeFilter集合中的一个便返回false，表示不对对应的beanDefinition注册
			if (tf.match(metadataReader, this.metadataReaderFactory)) {
				return false;
			}
		}
		for (TypeFilter tf : this.includeFilters) {
			//首先满足其中includeFilter集合中的一个
			if (tf.match(metadataReader, this.metadataReaderFactory)) {
				//判断对应的beanDifinition不存在@Conditional注解或者满足@Conditional中指定的条件，则返回true
				//@Conditional注解的使用可自行查看相关资料
				return isConditionMatch(metadataReader);
			}
		}
		return false;
	}

• ClassPathScanningCandidateComponentProvider#isCandidateComponent(AnnotatedBeanDefinition beanDefinition)
  
  验证beanDefinition class类是否为具体类

protected boolean isCandidateComponent(AnnotatedBeanDefinition beanDefinition) {
		//非抽象类、接口类并且有独立特性[它是一个顶级类还是一个嵌套类(静态内部类)，可以独立于封闭类构造。]
		return (beanDefinition.getMetadata().isConcrete() && beanDefinition.getMetadata().isIndependent());
	}

ComponentScanBeanDefinitionParser#registerComponents-注册其他组件

在扫描包内的class文件注册为beanDefinition之后，ComponentScanBeanDefinitionParser还需要注册其他的组件，具体是什么可简单看下相关的源码

protected void registerComponents(
			XmlReaderContext readerContext, Set<BeanDefinitionHolder> beanDefinitions, Element element) {

		Object source = readerContext.extractSource(element);
		//包装为CompositeComponentDefinition对象，内置多ComponentDefinition对象
		CompositeComponentDefinition compositeDef = new CompositeComponentDefinition(element.getTagName(), source);
		//将已注册的所有beanDefinitionHolder对象放到上述对象中
		for (BeanDefinitionHolder beanDefHolder : beanDefinitions) {
			compositeDef.addNestedComponent(new BeanComponentDefinition(beanDefHolder));
		}

		// Register annotation config processors, if necessary.
		boolean annotationConfig = true;
		//获取annotation-config的属性值，默认为true
		if (element.hasAttribute(ANNOTATION_CONFIG_ATTRIBUTE)) {
			annotationConfig = Boolean.valueOf(element.getAttribute(ANNOTATION_CONFIG_ATTRIBUTE));
		}
		if (annotationConfig) {
			//注册多个BeanPostProcessor接口，具体什么可自行查看，返回的是包含BeanPostProcessor接口的beanDefinitionHolder对象集合
			Set<BeanDefinitionHolder> processorDefinitions = AnnotationConfigUtils.registerAnnotationConfigProcessors(readerContext.getRegistry(), source);
			//继续装入CompositeComponentDefinition对象
			for (BeanDefinitionHolder processorDefinition : processorDefinitions) {
				compositeDef.addNestedComponent(new BeanComponentDefinition(processorDefinition));
			}
		}
		//此处为空
		readerContext.fireComponentRegistered(compositeDef);
	}

此处的目的主要是注册多个BeanPostProcessor接口实现类【供后续spring调用统一接口进行解析，比如>>>Spring源码情操陶冶-AbstractApplicationContext#invokeBeanFactoryPostProcessors可执行下述的@Configuration解析】具体的有

• ConfigurationClassPostProcessor解析@Configuration注解类
• AutowiredAnnotationBeanPostProcessor解析@Autowired/@Value注解
• RequiredAnnotationBeanPostProcessor解析@Required注解
• CommonAnnotationBeanPostProcessor解析@Resource注解
• PersistenceAnnotationBeanPostProcessor解析JPA注解，持久层

小结

1. context:component-scan节点的属性及其含义

• base-package 扫描的基础包名，必填项，也可指定多个包名，以,;分隔
• use-default-filters 默认为true，如果设置为false，则不启用@Component及其相关注解
• resource-pattern 自定义扫描的文件名，支持正则匹配，默认为**/*.class
• name-generator beanName生成器
• scope-resolver 指定bean的作用范围溶解器 与scope-proxy分开使用
• scope-proxy 与scope-resolver分开使用，targetClass：cglib代理、interfaces：JDK代理、no：不使用代理 ，默认使用cglib代理
• context:include-filter/context:exclude-filter 子节点，可有多个，表示可对beanDefinition上的注解过滤

1. 具体的通过扫描base-package指定的包名来得到所有的class文件，请看>>>Spring源码情操陶冶-PathMatchingResourcePatternResolver路径资源匹配溶解器。注意：扫描一个类的时候，如果其内部有静态内部类也是会被扫描注册的

2. 此context:component-scan的指定表明默认可将base-package指定的包下的所有注解class比如@Service等注册为bean到bean工厂，方便后续的业务调用

3. 注册BeanFactoryPostProcessor/BeanPostProcessor接口实现类用于后续的bean实例化，比如ConfigurationClassPostProcessor解析@Configuration注解类、AutowiredAnnotationBeanPostProcessor解析@Autowired/@Value注解、RequiredAnnotationBeanPostProcessor解析@Required注解、CommonAnnotationBeanPostProcessor解析@Resource注解、PersistenceAnnotationBeanPostProcessor解析JPA注解