首页 > 代码库 > spring源码读书笔记

spring源码读书笔记

如果我们在web项目里面使用spring的话,通常会在web.xml里面配置一个listener.

 

<listener>

<listener-class>

org.springframework.web.context.ContextLoaderListener

</listener-class>

</listener>

这个litener实现了ServletContextListener接口,并从ContextLoader继承。由于实现了ServletContextListener接口,所以在web容器启动的时候会调用contextInitialized方法。以下是这个方法的实现:

 

public void contextInitialized(ServletContextEvent event) {

       //createContextLoader是一个废弃了的方法,什么也没有做。返回null值

       this.contextLoader = createContextLoader();

       //所以this.contextLoader为null,这里把自身赋值给contextLoader

//因为ContextLoaderListerner继承了ContextLoader,所以可把自身赋值给

//ContextLoader(这里有点别扭)

       if (this.contextLoader == null) {

           this.contextLoader = this;

       }

       //接着就实例化webApplicationContext,由父类ContextLoader实现

    this.contextLoader.initWebApplicationContext(event.getServletContext());

}

我们现在来看一下ContextLoader的initWebApplicationContext方法,这个方法比较长,我们分段逐步跟进去看一下。(会省略一些不重要的代码)

以下这段代码主要判断是否重复实例化的问题,因为实例化webApplicationContext后,会把它放到servletContext的一个属性里,所以我们可以从servletContext的属性取出webApplicationContext,如果不为空,则已经实例化,接着就会抛出异常.

if(servletContext.getAttribute(WebApplicationContext.ROOT_WEB_APPLICATION_CONTEXT_ATTRIBUTE) != null) {

           throw new IllegalStateException(

"Cannot initialize context because there is already a root application context present - " +

"check whether you have multiple ContextLoader* definitions in your web.xml!");

    }

接着就会创建webApplicationContext

if (this.context == null) {

    this.context = createWebApplicationContext(servletContext);

}

我们跟进去createWebApplicationContext这个方法看一下

protected WebApplicationContext createWebApplicationContext(ServletContext sc)

{

    //获取相应的class,其实就是ConfigurableWebApplicationContext.class

    Class<?> contextClass = determineContextClass(sc);

    //判断是否为 ConfigurableWebApplicationContext.class或从它继承   if(!ConfigurableWebApplicationContext.class.isAssignableFrom(cont         extClass))

{

//若非ConfigurableWebApplicationContext.class或非其子类则抛出异常

throw new ApplicationContextException("Custom context class [" + contextClass.getName() +

"] is not of type [" + ConfigurableWebApplicationContext.class.getName() + "]");

    }

    //创建一个contextClass的实例对象返回

return(ConfigurableWebApplicationContext)BeanUtils.instantiateClass(contextClass);

}

 

我们看一下determineContextClass是怎么实现的

protected Class<?> determineContextClass(ServletContext servletContext) {

       //从servletContext读取contextClassName

       String contextClassName = servletContext.getInitParameter(CONTEXT_CLASS_PARAM);

 

      

       if (contextClassName != null) {

           //如果从servletContext读取到的contextClassName不为空,就返回对应

           //的class类

           try {

              //返回className对应的Class类

              return ClassUtils.forName(contextClassName, ClassUtils.getDefaultClassLoader());

           }//如果找不到该类名的类就抛出异常

           catch (ClassNotFoundException ex) {

              throw new ApplicationContextException(

                     "Failed to load custom context class [" + contextClassName + "]", ex);

           }

       }

    else {//如果从servletContext读取到得contextClassName为空就取默认的className

           contextClassName = defaultStrategies.getProperty(WebApplicationContext.class.getName());

           try {//返回className对应的Class类

              return ClassUtils.forName(contextClassName, ContextLoader.class.getClassLoader());

           }

           catch (ClassNotFoundException ex) {

              throw new ApplicationContextException(

                     "Failed to load default context class [" + contextClassName + "]", ex);

           }

       }

    }

 

至于BeanUtils.instantiateClass(contextClass);

则是通过反射创建对应class的实体对象。

 

 

//然后就是把context强制转换为configrableWebApplicationContext

configrableWebApplicationContext cwac = (configrableWebApplicationContext) this.context;

 

//接着就是核心方法configureAndRefreshWebApplicationContext

configureAndRefreshWebApplicationContext(cwac, servletContext);

 

//最后把它放到servletContext的属性里

servletContext.setAttribute(WebApplicationContext.ROOT_WEB_APPLICATION_CONTEXT_ATTRIBUTE, this.context);

 

这样webApplicationContext就算加载完成了。

我们现在来看一下核心方法configureAndRefreshWebApplicationContext(cwac, servletContext);(省略一些不重要的代码)

protected void configureAndRefreshWebApplicationContext(ConfigurableWebApplicationContext wac, ServletContext sc) {

       //把servletContext放到webApplicationContext中,以后可以直接取出来用

       wac.setServletContext(sc);

       //用户自己的一些设置

       customizeContext(sc, wac);

       //进行加载

       wac.refresh();

}

我们进入webApplicationContext的refresh方法看一下

public void refresh() throws BeansException, IllegalStateException {

       synchronized (this.startupShutdownMonitor) {

           // Prepare this context for refreshing.

           prepareRefresh();

 

           // Tell the subclass to refresh the internal bean factory.

           ConfigurableListableBeanFactory beanFactory = obtainFreshBeanFactory();

 

           // Prepare the bean factory for use in this context.

           prepareBeanFactory(beanFactory);

 

           try {

              // Allows post-processing of the bean factory in context subclasses.

              postProcessBeanFactory(beanFactory);

 

              // Invoke factory processors registered as beans in the context.

              invokeBeanFactoryPostProcessors(beanFactory);

 

              // Register bean processors that intercept bean creation.

              registerBeanPostProcessors(beanFactory);

 

              // Initialize message source for this context.

              initMessageSource();

 

              // Initialize event multicaster for this context.

              initApplicationEventMulticaster();

 

              // Initialize other special beans in specific context subclasses.

              onRefresh();

 

              // Check for listener beans and register them.

              registerListeners();

 

              // Instantiate all remaining (non-lazy-init) singletons.

              finishBeanFactoryInitialization(beanFactory);

 

              // Last step: publish corresponding event.

              finishRefresh();

           }

 

           catch (BeansException ex) {

              // Destroy already created singletons to avoid dangling resources.

              destroyBeans();

 

              // Reset ‘active‘ flag.

              cancelRefresh(ex);

 

              // Propagate exception to caller.

              throw ex;

           }

       }

    }

这是一个同步的方法,这里最核心的方法就是obtainFreshBeanFactory方法。其他的方法都是对webApplicationContext和beanFactory做一些前后的装饰和准备。

我们进入obtainFreshBeanFactoty方法看看

    protected ConfigurableListableBeanFactory obtainFreshBeanFactory() {

       refreshBeanFactory();

       ConfigurableListableBeanFactory beanFactory = getBeanFactory();

       if (logger.isDebugEnabled()) {

           logger.debug("Bean factory for " + getDisplayName() + ": " + beanFactory);

       }

       return beanFactory;

}

这里的核心方法就是refreshBeanFactory();它负责生成BeanFactory并加载bean

protected final void refreshBeanFactory() throws BeansException {

       //判断是否已经存在beanFactory如果有则销毁。

       if (hasBeanFactory()) {

           destroyBeans();

           closeBeanFactory();

       }

       try {

           DefaultListableBeanFactory beanFactory = createBeanFactory();//创建一个beanFactory

           beanFactory.setSerializationId(getId());//给它一个标识

           customizeBeanFactory(beanFactory);//用户自己做一些设置

           //这个方法很关键,负责加载所有的bean

           loadBeanDefinitions(beanFactory);

           synchronized (this.beanFactoryMonitor) {

              this.beanFactory = beanFactory;

           }

       }

       catch (IOException ex) {

           throw new ApplicationContextException("I/O error parsing bean definition source for " + getDisplayName(), ex);

       }

    }

然后我们进入loadBeanDefinitions(beanFactory);看一下

protected void loadBeanDefinitions(DefaultListableBeanFactory beanFactory) throws BeansException, IOException {

       // Create a new XmlBeanDefinitionReader for the given BeanFactory.

       XmlBeanDefinitionReader beanDefinitionReader = new XmlBeanDefinitionReader(beanFactory);

 

       // Configure the bean definition reader with this context‘s

       // resource loading environment.

       beanDefinitionReader.setEnvironment(this.getEnvironment());

       beanDefinitionReader.setResourceLoader(this);

       beanDefinitionReader.setEntityResolver(new ResourceEntityResolver(this));

 

       // Allow a subclass to provide custom initialization of the reader,

       // then proceed with actually loading the bean definitions.

       initBeanDefinitionReader(beanDefinitionReader);

       loadBeanDefinitions(beanDefinitionReader);

}

这里主要工作就是new一个XmlBeanDefinitionReader,给它设置environment,resourceLoader和entityResolver,注意一下由于webApplicationContext实现了ResouceLoader接口,所以它本身就是一个ResourceLoader.

我们可以看到它并不自己去实现lobeanDefinitions方法,而是委托给XmlBeanDefinitionReader去实现。

 

    protected void loadBeanDefinitions(XmlBeanDefinitionReader reader) throws IOException {

       String[] configLocations = getConfigLocations();

       if (configLocations != null) {

           for (String configLocation : configLocations) {

              reader.loadBeanDefinitions(configLocation);

           }

       }

    }

这里就对configLocations进行bean的加载,调用重载的方法(spring重载的方法好多啊)

public int loadBeanDefinitions(String location) throws BeanDefinitionStoreException {

       return loadBeanDefinitions(location, null);

}

 

 

public int loadBeanDefinitions(String location, Set<Resource> actualResources) throws BeanDefinitionStoreException {

       ResourceLoader resourceLoader = getResourceLoader();

       if (resourceLoader == null) {

           throw new BeanDefinitionStoreException(

                  "Cannot import bean definitions from location [" + location + "]: no ResourceLoader available");

       }

 

       if (resourceLoader instanceof ResourcePatternResolver) {

           // Resource pattern matching available.

           try {

              Resource[] resources = ((ResourcePatternResolver) resourceLoader).getResources(location);

              int loadCount = loadBeanDefinitions(resources);

              if (actualResources != null) {

                  for (Resource resource : resources) {

                     actualResources.add(resource);

                  }

              }

              if (logger.isDebugEnabled()) {

                  logger.debug("Loaded " + loadCount + " bean definitions from location pattern [" + location + "]");

              }

              return loadCount;

           }

           catch (IOException ex) {

              throw new BeanDefinitionStoreException(

                     "Could not resolve bean definition resource pattern [" + location + "]", ex);

           }

       }

       else {

           // Can only load single resources by absolute URL.

           Resource resource = resourceLoader.getResource(location);

           int loadCount = loadBeanDefinitions(resource);

           if (actualResources != null) {

              actualResources.add(resource);

           }

           if (logger.isDebugEnabled()) {

              logger.debug("Loaded " + loadCount + " bean definitions from location [" + location + "]");

           }

           return loadCount;

       }

    }

上面这段代码其实就是取得resourceLoader,通过location取得resouces,然后调用

loadBeanDefinitions(resource),这里又是一个重载的方法。

 

public int loadBeanDefinitions(Resource resource) throws BeanDefinitionStoreException {

       return loadBeanDefinitions(new EncodedResource(resource));

}

这个方法把resource进行一下编码,再调用一下重载的方法

 

public int loadBeanDefinitions(EncodedResource encodedResource) throws BeanDefinitionStoreException {

       Assert.notNull(encodedResource, "EncodedResource must not be null");

        if (logger.isInfoEnabled()) {

           logger.info("Loading XML bean definitions from " + encodedResource.getResource());

       }

 

       Set<EncodedResource> currentResources = this.resourcesCurrentlyBeingLoaded.get();

       if (currentResources == null) {

           currentResources = new HashSet<EncodedResource>(4);

           this.resourcesCurrentlyBeingLoaded.set(currentResources);

       }

       if (!currentResources.add(encodedResource)) {

           throw new BeanDefinitionStoreException(

                  "Detected cyclic loading of " + encodedResource + " - check your import definitions!");

       }

       try {

           InputStream inputStream = encodedResource.getResource().getInputStream();

           try {

              InputSource inputSource = new InputSource(inputStream);

              if (encodedResource.getEncoding() != null) {

                  inputSource.setEncoding(encodedResource.getEncoding());

              }

              return doLoadBeanDefinitions(inputSource, encodedResource.getResource());

           }

           finally {

              inputStream.close();

           }

       }

       catch (IOException ex) {

           throw new BeanDefinitionStoreException(

                  "IOException parsing XML document from " + encodedResource.getResource(), ex);

       }

       finally {

           currentResources.remove(encodedResource);

           if (currentResources.isEmpty()) {

              this.resourcesCurrentlyBeingLoaded.remove();

           }

       }

    }

通过resource取出inpustream,封装一个inputSource,调用doLoadBeanDefinitions(inputSource, encodedResource.getResource());

 

protected int doLoadBeanDefinitions(InputSource inputSource, Resource resource)

           throws BeanDefinitionStoreException {

      

           int validationMode = getValidationModeForResource(resource);

           Document doc = this.documentLoader.loadDocument(

                  inputSource, getEntityResolver(), this.errorHandler, validationMode, isNamespaceAware());

           return registerBeanDefinitions(doc, resource);

}

这个方法是从resource中读取一个doc对象,值得注意的是,这个doc是w3c的标准。然后进行bean的注册。

registerBeanDefinitions这个方法还没看完。 

 

 

spring源码读书笔记