最近幾天跟同事聊起Spring的一些問題，對一些地方有些疑問，趁這兩天有點空，看看Spring的原始碼，瞭解下具體的實現細節。本文基於Spring 4.0.5版本。

首先Web專案使用Spring是通過在web.xml裡面配置
org.springframework.web.context.ContextLoaderListener初始化IOC容器的。

<listener>
    <listener-class>org.springframework.web.context.ContextLoaderListener</listener-class>
</listener>
 

那就以此為切入點順藤摸瓜。

public class ContextLoaderListener extends ContextLoader implements ServletContextListener 

ContextLoaderListener繼承了ContextLoader，並且實現ServletContextListener介面。當Server容器（一般指tomcat）啟動時，會收到事件初始化。

@Override
public void contextInitialized(ServletContextEvent event) {
    initWebApplicationContext(event.getServletContext());
}
 

initWebApplicationContext方法是在org.springframework.web.context.ContextLoader類裡面。方法太長，分段讀一下。

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!");
}
Log logger = LogFactory.getLog(ContextLoader.class);
servletContext.log("Initializing Spring root WebApplicationContext");
if (logger.isInfoEnabled()) {
    logger.info("Root WebApplicationContext: initialization started");
}
long startTime = System.currentTimeMillis();
 

首先是判斷servletContext中是否已經註冊了WebApplicationContext，如果有則丟擲異常，避免重複註冊。然後就是啟用log，啟動計時。本方法的關鍵就在於try程式碼塊裡的內容

try {
    // Store context in local instance variable, to guarantee that
    // it is available on ServletContext shutdown.
    if (this.context == null) {
        this.context = createWebApplicationContext(servletContext);
    }
    if (this.context instanceof ConfigurableWebApplicationContext) {
        ConfigurableWebApplicationContext cwac = (ConfigurableWebApplicationContext) this.context;
        if (!cwac.isActive()) {
            // The context has not yet been refreshed -> provide services such as
            // setting the parent context, setting the application context id, etc
            if (cwac.getParent() == null) {
                // The context instance was injected without an explicit parent ->
                // determine parent for root web application context, if any.
                ApplicationContext parent = loadParentContext(servletContext);
                cwac.setParent(parent);
            }
            configureAndRefreshWebApplicationContext(cwac, servletContext);
        }
    }
    servletContext.setAttribute(WebApplicationContext.ROOT_WEB_APPLICATION_CONTEXT_ATTRIBUTE, this.context);

    ClassLoader ccl = Thread.currentThread().getContextClassLoader();
    if (ccl == ContextLoader.class.getClassLoader()) {
        currentContext = this.context;
    }
    else if (ccl != null) {
        currentContextPerThread.put(ccl, this.context);
    }

    if (logger.isDebugEnabled()) {
        logger.debug("Published root WebApplicationContext as ServletContext attribute with name [" +
                WebApplicationContext.ROOT_WEB_APPLICATION_CONTEXT_ATTRIBUTE + "]");
    }
    if (logger.isInfoEnabled()) {
        long elapsedTime = System.currentTimeMillis() - startTime;
        logger.info("Root WebApplicationContext: initialization completed in " + elapsedTime + " ms");
    }

    return this.context;
}

這裡面有幾個關鍵的方法。首先看一下createWebApplicationContext()

protected WebApplicationContext createWebApplicationContext(ServletContext sc) {
    Class<?> contextClass = determineContextClass(sc);
    if (!ConfigurableWebApplicationContext.class.isAssignableFrom(contextClass)) {
        throw new ApplicationContextException("Custom context class [" + contextClass.getName() +
                "] is not of type [" + ConfigurableWebApplicationContext.class.getName() + "]");
    }
    return (ConfigurableWebApplicationContext) BeanUtils.instantiateClass(contextClass);
}

首先determineContextClass()方法查明具體的Context類，他會讀取servletContext的初始化引數contextClass，此引數我們一半不配置，所以Spring就會讀取跟org.springframework.web.context.WebApplicationContext同一個包下面的ContextLoader.properties檔案讀取預設設定，反射出org.springframework.web.context.support.XmlWebApplicationContext類來。接下來就是在configureAndRefreshWebApplicationContext()方法裡將新建立的XmlWebApplicationContext進行初始化。首先會設定一個預設ID，即org.springframework.web.context.WebApplicationContext:+你專案的ContextPath。

if (ObjectUtils.identityToString(wac).equals(wac.getId())) {
    // The application context id is still set to its original default
    // value
    // -> assign a more useful id based on available information
    String idParam = sc.getInitParameter(CONTEXT_ID_PARAM);
    if (idParam != null) {
        wac.setId(idParam);
    } else {
        // Generate default id...
        wac.setId(ConfigurableWebApplicationContext.APPLICATION_CONTEXT_ID_PREFIX + ObjectUtils.getDisplayString(sc.getContextPath()));
    }
}

緊接著就是將ServletContext設定成XmlWebApplicationContext的屬性，這樣Spring就能在上下文裡輕鬆拿到ServletContext了。

wac.setServletContext(sc);

接下來就是讀取web.xml檔案中的contextConfigLocation引數。如果沒有配置就會去讀WEB-INF下的applicationContext.xml檔案。

<context-param>
    <param-name>contextConfigLocation</param-name>
    <param-value>classpath:beans.xml</param-value>
</context-param>

並將值設定(就是我們的Spring配置檔案的路徑)進XmlWebApplicationContext中。然後就會在指定的路徑載入配置檔案。

String configLocationParam = sc.getInitParameter(CONFIG_LOCATION_PARAM);
if (configLocationParam != null) {
    wac.setConfigLocation(configLocationParam);
}

接下來就是customizeContext(sc, wac)方法，此方法會根據使用者配置的globalInitializerClasses引數來初始化一些使用者自定義的屬性，一般我們不配置，所以這裡什麼也不做。

最後登場的就是最核心的方法了，

wac.refresh();

在這個方法裡，會完成資原始檔的載入、配置檔案解析、Bean定義的註冊、元件的初始化等核心工作，我們一探究竟。

@Override
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;
        }
    }
}

次方法是同步的，避免重複重新整理，每個步驟都放在單獨的方法內，流程清晰，是值得學習的地方。這裡面有個重要的方法是finishBeanFactoryInitialization(beanFactory);，裡面的內容是Spring如何例項化bean，並注入依賴的，這個內容下一節講，本節只說明Spring是如何載入class檔案的。

首先就是prepareRefresh()方法。

protected void prepareRefresh() {
    this.startupDate = System.currentTimeMillis();

    synchronized (this.activeMonitor) {
        this.active = true;
    }

    if (logger.isInfoEnabled()) {
        logger.info("Refreshing " + this);
    }

    // Initialize any placeholder property sources in the context environment
    initPropertySources();

    // Validate that all properties marked as required are resolvable
    // see ConfigurablePropertyResolver#setRequiredProperties
    getEnvironment().validateRequiredProperties();
}

此方法做一些準備工作，如記錄開始時間，輸出日誌，initPropertySources();和getEnvironment().validateRequiredProperties();一般沒幹什麼事。

接下來就是初始化BeanFactory，是整個refresh()方法的核心，其中完成了配置檔案的載入、解析、註冊

ConfigurableListableBeanFactory beanFactory = obtainFreshBeanFactory();

看看它裡面都做了些什麼？

protected ConfigurableListableBeanFactory obtainFreshBeanFactory() {
    refreshBeanFactory();
    ConfigurableListableBeanFactory beanFactory = getBeanFactory();
    if (logger.isDebugEnabled()) {
        logger.debug("Bean factory for " + getDisplayName() + ": " + beanFactory);
    }
    return beanFactory;
}

首先refreshBeanFactory()：

protected final void refreshBeanFactory() throws BeansException {
    if (hasBeanFactory()) {
        destroyBeans();
        closeBeanFactory();
    }
    try {
        DefaultListableBeanFactory beanFactory = createBeanFactory();
        beanFactory.setSerializationId(getId());
        customizeBeanFactory(beanFactory);
        loadBeanDefinitions(beanFactory);
        synchronized (this.beanFactoryMonitor) {
            this.beanFactory = beanFactory;
        }
    }
    catch (IOException ex) {
        throw new ApplicationContextException("I/O error parsing bean definition source for " + getDisplayName(), ex);
    }
}

我們看到會建立一個DefaultListableBeanFactory例項

DefaultListableBeanFactory beanFactory = createBeanFactory();

再設定一個ID

beanFactory.setSerializationId(getId());

然後設定一些自定義引數：

customizeBeanFactory(beanFactory);

這裡面最重要的就是loadBeanDefinitions(beanFactory);方法了。

@Override
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);
}

此方法會通過XmlBeanDefinitionReader載入bean定義。具體的實現方法是在org.springframework.beans.factory.xml.XmlBeanDefinitionReader.loadBeanDefinitions方法中定義的。這裡設計了層層呼叫，有好多過載方法，主要就是載入Spring所有的配置檔案(可能會有多個)，以備後面解析，註冊之用。我一路追蹤到org.springframework.beans.factory.xml.DefaultBeanDefinitionDocumentReader.doRegisterBeanDefinitions(Element root)

protected void doRegisterBeanDefinitions(Element root) {
    String profileSpec = root.getAttribute(PROFILE_ATTRIBUTE);
    if (StringUtils.hasText(profileSpec)) {
        Assert.state(this.environment != null, "Environment must be set for evaluating profiles");
        String[] specifiedProfiles = StringUtils.tokenizeToStringArray(
                profileSpec, BeanDefinitionParserDelegate.MULTI_VALUE_ATTRIBUTE_DELIMITERS);
        if (!this.environment.acceptsProfiles(specifiedProfiles)) {
            return;
        }
    }
    BeanDefinitionParserDelegate parent = this.delegate;
    this.delegate = createDelegate(this.readerContext, root, parent);
    preProcessXml(root);
    parseBeanDefinitions(root, this.delegate);
    postProcessXml(root);
    this.delegate = parent;
}

這裡建立了一個BeanDefinitionParserDelegate示例，解析XML的過程就是委託它完成的，我們不關心它是怎樣解析XML的，我們只關心是怎麼載入類的，所以就要看parseBeanDefinitions(root, this.delegate)方法了。

protected void parseBeanDefinitions(Element root, BeanDefinitionParserDelegate delegate) {
    if (delegate.isDefaultNamespace(root)) {
        NodeList nl = root.getChildNodes();
        for (int i = 0; i < nl.getLength(); i++) {
            Node node = nl.item(i);
            if (node instanceof Element) {
                Element ele = (Element) node;
                if (delegate.isDefaultNamespace(ele)) {
                    parseDefaultElement(ele, delegate);
                }
                else {
                    delegate.parseCustomElement(ele);
                }
            }
        }
    }
    else {
        delegate.parseCustomElement(root);
    }
}

我們看到最終解析XML元素的是delegate.parseCustomElement(ele)方法，最終會走到一下方法.

public BeanDefinition parseCustomElement(Element ele, BeanDefinition containingBd) {
    String namespaceUri = getNamespaceURI(ele);
    NamespaceHandler handler = this.readerContext.getNamespaceHandlerResolver().resolve(namespaceUri);
    if (handler == null) {
        error("Unable to locate Spring NamespaceHandler for XML schema namespace [" + namespaceUri + "]", ele);
        return null;
    }
    return handler.parse(ele, new ParserContext(this.readerContext, this, containingBd));
}

這裡會根據不同的XML節點，會委託NamespaceHandlerSupport找出合適的BeanDefinitionParser，如果我們配置了

<context:component-scan
    base-package="com.geeekr.service,com.geeekr.dao" />

那麼對應BeanDefinitionParser就是org.springframework.context.annotation.ComponentScanBeanDefinitionParser，來看看它的parse方法。

@Override
public BeanDefinition parse(Element element, ParserContext parserContext) {
    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);
    Set<BeanDefinitionHolder> beanDefinitions = scanner.doScan(basePackages);
    registerComponents(parserContext.getReaderContext(), beanDefinitions, element);

    return null;
}

不難看出這裡定義了一個ClassPathBeanDefinitionScanner，通過它去掃描包中的類檔案，注意：這裡是類檔案而不是類，因為現在這些類還沒有被載入，只是ClassLoader能找到這些class的路徑而已。到目前為止，感覺真想距離我們越來越近了。順著繼續往下摸。進入doSacn方法裡，映入眼簾的又是一大坨程式碼，但是我們只關心觀點的部分。

protected Set<BeanDefinitionHolder> doScan(String... basePackages) {
    Assert.notEmpty(basePackages, "At least one base package must be specified");
    Set<BeanDefinitionHolder> beanDefinitions = new LinkedHashSet<BeanDefinitionHolder>();
    for (String basePackage : basePackages) {
        Set<BeanDefinition> candidates = findCandidateComponents(basePackage);
        for (BeanDefinition candidate : candidates) {
            ScopeMetadata scopeMetadata = this.scopeMetadataResolver.resolveScopeMetadata(candidate);
            candidate.setScope(scopeMetadata.getScopeName());
            String beanName = this.beanNameGenerator.generateBeanName(candidate, this.registry);
            if (candidate instanceof AbstractBeanDefinition) {
                postProcessBeanDefinition((AbstractBeanDefinition) candidate, beanName);
            }
            if (candidate instanceof AnnotatedBeanDefinition) {
                AnnotationConfigUtils.processCommonDefinitionAnnotations((AnnotatedBeanDefinition) candidate);
            }
            if (checkCandidate(beanName, candidate)) {
                BeanDefinitionHolder definitionHolder = new BeanDefinitionHolder(candidate, beanName);
                definitionHolder = AnnotationConfigUtils.applyScopedProxyMode(scopeMetadata, definitionHolder, this.registry);
                beanDefinitions.add(definitionHolder);
                registerBeanDefinition(definitionHolder, this.registry);
            }
        }
    }
    return beanDefinitions;
}

一眼就能看出是通過

Set<BeanDefinition> candidates = findCandidateComponents(basePackage);

有時候不得不佩服這些外國人起名字的功力，把掃描出來的類叫做candidates(候選人)；真是不服不行啊，這種名字真的很容易理解有不有？哈哈，貌似扯遠了。繼續往下看。這裡只列出方法的主題部分。

public Set<BeanDefinition> findCandidateComponents(String basePackage) {
    Set<BeanDefinition> candidates = new LinkedHashSet<BeanDefinition>();
    try {
        String packageSearchPath = ResourcePatternResolver.CLASSPATH_ALL_URL_PREFIX +
                resolveBasePackage(basePackage) + "/" + this.resourcePattern;
        Resource[] resources = this.resourcePatternResolver.getResources(packageSearchPath);
        boolean traceEnabled = logger.isTraceEnabled();
        boolean debugEnabled = logger.isDebugEnabled();
        for (Resource resource : resources) {
            if (traceEnabled) {
                logger.trace("Scanning " + resource);
            }
            if (resource.isReadable()) {
                try {
                    MetadataReader metadataReader = this.metadataReaderFactory.getMetadataReader(resource);
                    if (isCandidateComponent(metadataReader)) {
                        ScannedGenericBeanDefinition sbd = new ScannedGenericBeanDefinition(metadataReader);
                        sbd.setResource(resource);
                        sbd.setSource(resource);

先看這兩句：

String packageSearchPath = ResourcePatternResolver.CLASSPATH_ALL_URL_PREFIX + resolveBasePackage(basePackage) + "/" + this.resourcePattern;

假設我們配置的需要掃描的包名為com.geeekr.service，那麼packageSearchPath的值就是classpath*:com.geeekr.service/**/*.class，意思就是com.geeekr.service包(包括子包)下所有class檔案；如果配置的是*，那麼packageSearchPath的值就是classpath*:*/**/*.class。這裡的表示式是Spring自己定義的。Spring會根據這種表示式找出相關的class檔案。

Resource[] resources = this.resourcePatternResolver.getResources(packageSearchPath);

這一句就把相關class檔案加載出來了，那我們就要看看，Spring究竟是如何把class檔案找到的了。首先看看resourcePatternResolver的定義:

private ResourcePatternResolver resourcePatternResolver = new PathMatchingResourcePatternResolver();

進入getResources方法

@Override
public Resource[] getResources(String locationPattern) throws IOException {
    Assert.notNull(locationPattern, "Location pattern must not be null");
    if (locationPattern.startsWith(CLASSPATH_ALL_URL_PREFIX)) {
        // a class path resource (multiple resources for same name possible)
        if (getPathMatcher().isPattern(locationPattern.substring(CLASSPATH_ALL_URL_PREFIX.length()))) {
            // a class path resource pattern
            return findPathMatchingResources(locationPattern);
        }
        else {
            // all class path resources with the given name
            return findAllClassPathResources(locationPattern.substring(CLASSPATH_ALL_URL_PREFIX.length()));
        }
    }
    else {
        // Only look for a pattern after a prefix here
        // (to not get fooled by a pattern symbol in a strange prefix).
        int prefixEnd = locationPattern.indexOf(":") + 1;
        if (getPathMatcher().isPattern(locationPattern.substring(prefixEnd))) {
            // a file pattern
            return findPathMatchingResources(locationPattern);
        }
        else {
            // a single resource with the given name
            return new Resource[] {getResourceLoader().getResource(locationPattern)};
        }
    }
}

這裡會先判斷表示式是否以classpath*:開頭。前面我們看到Spring已經給我們添加了這個頭，這裡當然符合條件了。接著會進入findPathMatchingResources方法。在這裡又把**/*.class去掉了，然後在呼叫getResources方法，然後在進入findAllClassPathResources方法。這裡的引數只剩下包名了例如com/geeekr/service/。

protected Resource[] findAllClassPathResources(String location) throws IOException {
    String path = location;
    if (path.startsWith("/")) {
        path = path.substring(1);
    }
    ClassLoader cl = getClassLoader();
    Enumeration<URL> resourceUrls = (cl != null ? cl.getResources(path) : ClassLoader.getSystemResources(path));
    Set<Resource> result = new LinkedHashSet<Resource>(16);
    while (resourceUrls.hasMoreElements()) {
        URL url = resourceUrls.nextElement();
        result.add(convertClassLoaderURL(url));
    }
    return result.toArray(new Resource[result.size()]);
}

真相大白了，Spring也是用的ClassLoader載入的class檔案。一路追蹤，原始的ClassLoader是Thread.currentThread().getContextClassLoader();。到此為止，就拿到class檔案了。
Spring會將class資訊封裝成BeanDefinition，然後再放進DefaultListableBeanFactory的beanDefinitionMap中。

拿到了class檔案後，就要看看Spring是如何裝配bean的了，下一節，繼續看。