RequestMappingHandlerMapping是SpringMvc中一个比较核心的类，查看下它的类结构图：
InitializingBean是个很神奇的接口，在Spring每个容器的bean构造方法、属性设置之后，会先调用InitializingBean接口的afterPropertiesSet方法；
RequestMappingHandlerMapping的afterPropertiesSet方法： 初始化了config对象，以及调用父类AbstractHandlerMethodMapping的afterPropertiesSet，父类方法afterPropertiesSet 逻辑是 initHandlerMethods，这也是SpringMvc初始化寻找Controller以及映射加载的核心逻辑；
@Override public void afterPropertiesSet() { this.config = new RequestMappingInfo.BuilderConfiguration(); this.config.setPathHelper(getUrlPathHelper()); this.config.setPathMatcher(getPathMatcher()); this.config.setSuffixPatternMatch(this.useSuffixPatternMatch); this.config.setTrailingSlashMatch(this.useTrailingSlashMatch); this.config.setRegisteredSuffixPatternMatch(this.useRegisteredSuffixPatternMatch); this.config.setContentNegotiationManager(getContentNegotiationManager()); // 初始化config对象，主要属性就是pathMatcher； 以及调用父类 afterPropertiesSet 方法，这是SpringMvc映射关系加载的核心； super.afterPropertiesSet(); }
AbstractHandlerMethodMapping 的 initHandlerMethods代码：
protected void initHandlerMethods() { if (logger.isDebugEnabled()) { logger.debug("Looking for request mappings in application context: " + getApplicationContext()); } String[] beanNames = (this.detectHandlerMethodsInAncestorContexts ? BeanFactoryUtils.beanNamesForTypeIncludingAncestors(getApplicationContext(), Object.class) : getApplicationContext().getBeanNamesForType(Object.class));
// detectHandlerMethodsInAncestorContexts 默认为false,代表不会检测SpringMvc父容器中的bean的映射关系 for (String beanName : beanNames) {
//遍历容器中的beanName, 代理的对象跳过，获取当前bean的类型，调用isHandler判断是否是处理器(handler\controller) if (!beanName.startsWith(SCOPED_TARGET_NAME_PREFIX)) { Class<?> beanType = null; try { beanType = getApplicationContext().getType(beanName); } catch (Throwable ex) { // An unresolvable bean type, probably from a lazy bean - let's ignore it.
if (logger.isDebugEnabled()) { logger.debug("Could not resolve target class for bean with name '" + beanName + "'", ex); } } if (beanType != null && isHandler(beanType)) { //isHandler方法判断是否是controller,判断逻辑下面有； detectHandlerMethods(beanName); //加载Controller和请求映射关系 } } } handlerMethodsInitialized(getHandlerMethods()); // 该方法是个空实现 }
isHandler方法： 判断当前bean的class属性，标注了Controller或者RequestMapping注解，就会去加载Controller和请求映射关系，如果不是handler，迭代下一个bean对象；
protected boolean isHandler(Class<?> beanType) { return (AnnotatedElementUtils.hasAnnotation(beanType, Controller.class) || AnnotatedElementUtils.hasAnnotation(beanType, RequestMapping.class)); }
detectHandlerMethods方法：
protected void detectHandlerMethods(final Object handler) { Class<?> handlerType = (handler instanceof String ? getApplicationContext().getType((String) handler) : handler.getClass());//之前传入handler为string类型，此处去容器获取handler的class final Class<?> userType = ClassUtils.getUserClass(handlerType); //处理class为CGLIB生成class，如果是CGLIB的获取父类class Map<Method, T> methods = MethodIntrospector.selectMethods(userType, new MethodIntrospector.MetadataLookup<T>() { @Override public T inspect(Method method) { return getMappingForMethod(method, userType); } }); if (logger.isDebugEnabled()) { logger.debug(methods.size() + " request handler methods found on " + userType + ": " + methods); } for (Map.Entry<Method, T> entry : methods.entrySet()) { Method invocableMethod = AopUtils.selectInvocableMethod(entry.getKey(), userType); T mapping = entry.getValue(); registerHandlerMethod(handler, invocableMethod, mapping); } }
MethodIntrospector.selectMethods(..)方法是个很全面的解析方法：注释写得很详细，☆方法处，metadataLookup.inspect方法，往上看，调用的就是getMappingForMethod方法获取RequestMappingInfo对象；
public static <T> Map<Method, T> selectMethods(Class<?> targetType, final MetadataLookup<T> metadataLookup) { final Map<Method, T> methodMap = new LinkedHashMap<Method, T>(); Set<Class<?>> handlerTypes = new LinkedHashSet<Class<?>>(); Class<?> specificHandlerType = null; if (!Proxy.isProxyClass(targetType)) {
//handler class不是JDK代理生成的,加入到handlerTypes集合，specificHandlerType为当前handler class handlerTypes.add(targetType); specificHandlerType = targetType; } handlerTypes.addAll(Arrays.asList(targetType.getInterfaces())); /
/handler class实现的接口加入到handlerTypes for (Class<?> currentHandlerType : handlerTypes) { final Class<?> targetClass = (specificHandlerType != null ? specificHandlerType : currentHandlerType); ReflectionUtils.doWithMethods(currentHandlerType, new ReflectionUtils.MethodCallback() {
//该工具类方法，遍历了该currentHandlerType本类中所有的方法 // 调用的是 getDeclaredMethods(),然后遍历method数组，调用doWith回调处理method方法 public void doWith(Method method) { Method specificMethod = ClassUtils.getMostSpecificMethod(method, targetClass); T result = metadataLookup.inspect(specificMethod);
// ☆ 核心！！！ 这里处理了方法以及类上的映射关系，并且返回泛型T，实际类型是RequesMappingInfo if (result != null) { Method bridgedMethod = BridgeMethodResolver.findBridgedMethod(specificMethod); if (bridgedMethod == specificMethod || metadataLookup.inspect(bridgedMethod) == null) { methodMap.put(specificMethod, result); } } } }, ReflectionUtils.USER_DECLARED_METHODS);
// ReflectionUtils.USER_DECLARED_METHODS是个methodFilter，作用是过滤方法是用户定义、且非桥接类型的方法； } return methodMap; }
RequestMappingHandlerMapping 的 getMappingForMethod 方法：先分析方法上的映射关系，再分析类所在方法上的映射关系，然后结合处理；
下面一点点记录我查看这个方法的发现；
protected RequestMappingInfo getMappingForMethod(Method method, Class<?> handlerType) { RequestMappingInfo info = createRequestMappingInfo(method);
// 解析类上RequestMapping注解 if (info != null) { RequestMappingInfo typeInfo = createRequestMappingInfo(handlerType);// 解析方法上@RequestMapping注解 if (typeInfo != null) { info = typeInfo.combine(info); //方法上RequestMapping注解不为空，就需要结合分析 } } return info; }
createRequestMappingInfo 方法：
private RequestMappingInfo createRequestMappingInfo(AnnotatedElement element) {
//调用Spring注解工具类AnnotatedElementUtils获取方法上注解 RequestMapping requestMapping = AnnotatedElementUtils.findMergedAnnotation(element, RequestMapping.class); RequestCondition<?> condition = (element instanceof Class<?> ? getCustomTypeCondition((Class<?>) element) : getCustomMethodCondition((Method) element));
//RequestMappingHandlerMapping两个方法都是返回null，空实现 return (requestMapping != null ? createRequestMappingInfo(requestMapping, condition) : null); }
具体的RequestMappingInfo的构造采用建造者模式还是其他模式的？
protected RequestMappingInfo createRequestMappingInfo( RequestMapping requestMapping, RequestCondition<?> customCondition) { //customCondition一般都为null return RequestMappingInfo .paths(resolveEmbeddedValuesInPatterns(requestMapping.path()))// @RequestMapping(path={....}) 将path属性设置上去 .methods(requestMapping.method()) // @RequestMapping(method={....}) 将method属性设置上去 .params(requestMapping.params()) // @RequestMapping(method={....}) 将method属性设置上去 .headers(requestMapping.headers()) // @RequestMapping(headers={....}) 将headers属性设置上去 .consumes(requestMapping.consumes()) // @RequestMapping(consumes={....}) 将consumes属性设置上去 .produces(requestMapping.produces()) // @RequestMapping(produces={....}) 将produces属性设置上去 .mappingName(requestMapping.name()) // @RequestMapping(name={....}) 将name属性设置上去 .customCondition(customCondition) .options(this.config) .build(); }
这里只分析一个开头、一个结尾这样；
RequestMappingInfo 的 paths 方法：
public static Builder paths(String... paths) { // paths是@RequestMapping的path属性，字符串数组，这里用可变参数来接收，效果一样
return new DefaultBuilder(paths); }
Builder接口所有方法都返回Builder对象，DefaultBuilder持有一堆属性，可以看到都是@ReuqestMapping的属性；
paths方法就将注解的path属性注入到DefaultBuilder中，其他方法methods、params、headers、consumes、produces、mappingName、customCondition都是这个套路；
而 options注入的config属性 ，最开始 afterPropertiesSet 里 ，this.config = new RequestMappingInfo.BuilderConfiguration()；
就是将RequestMappingHandleMapping中的config作为DefaultBuilder的options注入； 最后就是build方法。
DefaultBuilder 的 build方法：
public RequestMappingInfo build() { ContentNegotiationManager manager = this.options.getContentNegotiationManager(); // PatternsRequestCondition构造的主要属性就是paths,代表了映射的路径，不以/开头会添加 / 这个开头 PatternsRequestCondition patternsCondition = new PatternsRequestCondition( this.paths, this.options.getUrlPathHelper(), this.options.getPathMatcher(), this.options.useSuffixPatternMatch(), this.options.useTrailingSlashMatch(), this.options.getFileExtensions()); return new RequestMappingInfo(this.mappingName, patternsCondition, new RequestMethodsRequestCondition(methods), new ParamsRequestCondition(this.params), new HeadersRequestCondition(this.headers), new ConsumesRequestCondition(this.consumes, this.headers), new ProducesRequestCondition(this.produces, this.headers, manager), this.customCondition); // customCondition通常为null }
build方法返回 RequestMappingInfo，其中构造入参都是XXXRequestCondition这种，他们都实现了RequestCondition接口；
private PatternsRequestCondition(Collection<String> patterns, UrlPathHelper urlPathHelper, PathMatcher pathMatcher, boolean useSuffixPatternMatch, boolean useTrailingSlashMatch, List<String> fileExtensions) { //这里就是 prependLeadingSlash 会判断 @RequestMapping注解的 path属性，不是以 /开头会添加 / this.patterns = Collections.unmodifiableSet(prependLeadingSlash(patterns));
this.pathHelper = (urlPathHelper != null ? urlPathHelper : new UrlPathHelper()); this.pathMatcher = (pathMatcher != null ? pathMatcher : new AntPathMatcher()); this.useSuffixPatternMatch = useSuffixPatternMatch; this.useTrailingSlashMatch = useTrailingSlashMatch; if (fileExtensions != null) { for (String fileExtension : fileExtensions) { if (fileExtension.charAt(0) != '.') { fileExtension = "." + fileExtension; } this.fileExtensions.add(fileExtension); } } }
private static Set<String> prependLeadingSlash(Collection<String> patterns) { if (patterns == null) { return Collections.emptySet(); } Set<String> result = new LinkedHashSet<String>(patterns.size()); for (String pattern : patterns) { if (StringUtils.hasLength(pattern) && !pattern.startsWith("/")) { //URL不以 /开头就会自动添加 / pattern = "/" + pattern; } result.add(pattern); } return result; }
回到RequestMappingInfo的构造方法，将@RequestMapping的所有属性都以 RequestCondition的实现类 形式保存到 RequestMappingInfo对象中；
接口RequestCondition定义了三个方法，1.combine：一般用来 方法级别@RequestMapping与类级别@RequestMapping结合，返回新的(通常是RequestMappingInfo)；
2.getMatchingCondition：检查request对象是否满足条件，返回一个新的满足条件的RequestMappingInfo实例（T泛型用都是RequestMappingInfo）；
3.compareTo 用来多个匹配的情况排序挑选最合适的
public interface RequestCondition<T> {
T combine(T other);
T getMatchingCondition(HttpServletRequest request);
int compareTo(T other, HttpServletRequest request);
}
至此 回到 RequestMappingHandlerMapping 的 getMappingForMethod方法 ，第一个方法级别的createRequestMappingInfo方法分析完毕，下面两行解析了标注在 类上的 注解，并且返回 RequestMappingInfo对象，
第188行就是类上标注了@RequestMapping注解，和方法上同样标注@RequestMapping结合处理的步骤：调用类上的RequestMappingInfo的combine方法
查看RequestMappingInfo对象的combine方法：
public RequestMappingInfo combine(RequestMappingInfo other) {
// RequestMapping的name属性的处理方法，一般name属性很少写，处理方式:两个都不为空就返回this.name#other.name;有一个为空 就返回另外一个name String name = combineNames(other);
//下面逻辑A分析 调用AntPathMatcher的combine方法,将类上URL和方法上URL组合并放入新PatternsRequestCondition PatternsRequestCondition patterns = this.patternsCondition.combine(other.patternsCondition);
//下面逻辑B分析，并且接下来的methods、params、headers等等实现方式大体一致 RequestMethodsRequestCondition methods = this.methodsCondition.combine(other.methodsCondition); ParamsRequestCondition params = this.paramsCondition.combine(other.paramsCondition); HeadersRequestCondition headers = this.headersCondition.combine(other.headersCondition);
//！！comsume和produce判断逻辑不是相加，方法上的该属性优先级高于类级别上的 ConsumesRequestCondition consumes = this.consumesCondition.combine(other.consumesCondition); ProducesRequestCondition produces = this.producesCondition.combine(other.producesCondition); RequestConditionHolder custom = this.customConditionHolder.combine(other.customConditionHolder); return new RequestMappingInfo(name, patterns,//返回一个新的RequestMappingInfo对象，其中所有RequestCondition都是新创建的对象 methods, params, headers, consumes, produces, custom.getCondition()); }
逻辑A： PatternsRequestCondition 之前介绍过，其属性patterns 就是@RequestMapping的path / value 属性的集合，且判断 path是否以 / 开头，如果不是会自动补全 / 开头；
其实现了RequestCondition接口，查看其combine方法
public PatternsRequestCondition combine(PatternsRequestCondition other) {
// result作为新的请求路径集合 Set<String> result = new LinkedHashSet<String>();
//类上注解@RequestMapping path不为空，方法上注解注解@RequestMapping path不为空
//此处的AntPathMatcher就是RequestMappingHandlerMapping对象里的antPathMatcher对象
//@RequestMapping path属性是集合类型的，这类似笛卡尔积形式 调用AntPathMatcher的combine方式，进行URL组合 加入到result
if (!this.patterns.isEmpty() && !other.patterns.isEmpty()) { for (String pattern1 : this.patterns) {
for (String pattern2 : other.patterns) { result.add(this.pathMatcher.combine(pattern1, pattern2)); } } }
//已经说明有一方为空了，只要判断另外一方是否为空，不为空直接加入Set<String> else if (!this.patterns.isEmpty()) { result.addAll(this.patterns); } else if (!other.patterns.isEmpty()) { result.addAll(other.patterns); } else { result.add(""); } /返回了一个新的PatternsRequestCondition对象，patterns属性就是当前方法的请求路径 return new PatternsRequestCondition(result, this.pathHelper, this.pathMatcher, this.useSuffixPatternMatch, this.useTrailingSlashMatch, this.fileExtensions); / }
逻辑A-1：AntPathMatcher对象如何对请求路径进行结合combine?
| 类上path | 方法上path | 结合后path |
| null | null | |
| /hotels | null | /hotels |
| null | /hotels | /hotels |
| /hotels | /bookings | /hotels/bookings |
| /hotels | bookings | /hotels/bookings |
| /hotels/* | /bookings | /hotels/bookings |
| /hotels/** | /bookings | /hotels/**/bookings |
| /hotels | {hotel} | /hotels/{hotel} |
| /hotels/* | {hotel} | /hotels/{hotel} |
| /hotels/** | {hotel} | /hotels/**/{hotel} |
| /*.html | hotels.html | /hotels.html |
| /*.html | /hotels | /hotels.html |
| /*.html | /*.txt | IllegalArgumentException |
逻辑B：RequestMethodsRequestCondition 的 combine 方法，方法上注解@RequestMapping的method加入到类上注解的method属性里，然后返回一个全新的RequestMethodsRequestCondition，持有新的method集合；
public RequestMethodsRequestCondition combine(RequestMethodsRequestCondition other) { Set<RequestMethod> set = new LinkedHashSet<RequestMethod>(this.methods); set.addAll(other.methods); return new RequestMethodsRequestCondition(set); }
getMappingForMethod方法调用结束，返回结合后的RequestMappingInfo对象； 回到MethodIntrospector.selectMethods方法，第19行就是调用的getMappingForMethod方法，返回RequestMappingInfo对象result，result不为空之后，
会筛选不是桥接方法，存入methodMap这个Map对象，key-type是Method，value-type是RequestMappingInfo类型；
该方法selectMethods将Controller / Handler中所有方法都进行判断加载请求映射，返回methodMap对象；
1 public static <T> Map<Method, T> selectMethods(Class<?> targetType, final MetadataLookup<T> metadataLookup) { 2 final Map<Method, T> methodMap = new LinkedHashMap<Method, T>(); 3 Set<Class<?>> handlerTypes = new LinkedHashSet<Class<?>>(); 4 Class<?> specificHandlerType = null; 5
6 if (!Proxy.isProxyClass(targetType)) { 7 handlerTypes.add(targetType); 8 specificHandlerType = targetType; 9 } 10 handlerTypes.addAll(Arrays.asList(targetType.getInterfaces())); 11
12 for (Class<?> currentHandlerType : handlerTypes) { 13 final Class<?> targetClass = (specificHandlerType != null ? specificHandlerType : currentHandlerType); 14
15 ReflectionUtils.doWithMethods(currentHandlerType, new ReflectionUtils.MethodCallback() { 16 @Override 17 public void doWith(Method method) { 18 Method specificMethod = ClassUtils.getMostSpecificMethod(method, targetClass); 19 T result = metadataLookup.inspect(specificMethod); 20 if (result != null) { 21 Method bridgedMethod = BridgeMethodResolver.findBridgedMethod(specificMethod); 22 if (bridgedMethod == specificMethod || metadataLookup.inspect(bridgedMethod) == null) { 23 methodMap.put(specificMethod, result); 24 } 25 } 26 } 27 }, ReflectionUtils.USER_DECLARED_METHODS); 28 } 29
30 return methodMap; 31 }
回到最开始的分析detectHandlerMethods方法：methods对象就是上面返回的methodMap，如果日志设置了DEBUG，每遍历一个controller都会输出日志；
1 protected void detectHandlerMethods(final Object handler) { 2 Class<?> handlerType = (handler instanceof String ?
3 getApplicationContext().getType((String) handler) : handler.getClass()); 4 final Class<?> userType = ClassUtils.getUserClass(handlerType); 5
6 Map<Method, T> methods = MethodIntrospector.selectMethods(userType, 7 new MethodIntrospector.MetadataLookup<T>() { 8 @Override 9 public T inspect(Method method) { 10 return getMappingForMethod(method, userType); 11 } 12 }); 13
14 if (logger.isDebugEnabled()) { 15 logger.debug(methods.size() + " request handler methods found on " + userType + ": " + methods); 16 } 17 for (Map.Entry<Method, T> entry : methods.entrySet()) {//遍历methods，并且调用registerHandlerMethod注册映射信息 18 Method invocableMethod = AopUtils.selectInvocableMethod(entry.getKey(), userType); 19 T mapping = entry.getValue(); 20 registerHandlerMethod(handler, invocableMethod, mapping); 21 } 22 }
registerHandlerMethod：总结 ：
|
RequestMappingHandlerMapping.mappingRegistry属性
|
key-type | value-type |
mappingLookup |
RequestMappingInfo | HandlerMethod对象 |
urlLookup |
请求路径URL | RequestMappingInfo |
| nameLookup | controller name中大写字母#方法名(如UC#test) | HandlerMethod对象 |
registry |
RequestMappingInfo | MappingRegistration对象(持有 |
protected void registerHandlerMethod(Object handler, Method method, T mapping) { this.mappingRegistry.register(mapping, handler, method); } // this对象指RequestMappingHandlerMapping，mapping是RequestMappingInfo对象，handler是controler的name,method是当前@RequestMapping方法 public void register(T mapping, Object handler, Method method) { this.readWriteLock.writeLock().lock(); //可重入锁 写锁上锁 这里不太明白为什么要上锁 try {
//创建新的HandlerMethod对象 下面逻辑C 介绍HandlerMethod 逻辑D 分析createHandlerMethod方法 HandlerMethod handlerMethod = createHandlerMethod(handler, method);
//校验唯一性，一个RequestMappingInfo对应一个Handlermethod，如果根据RequestMappingInfo找到不同的hm 抛出异常 assertUniqueMethodMapping(handlerMethod, mapping); //INFO级别日志 比如Mapped "{[/user/test]}" onto public java.lang.String demo2.UserController.test(javax.servlet.http.HttpServletRequest,javax.servlet.http.HttpServletResponse) if (logger.isInfoEnabled()) { logger.info("Mapped \"" + mapping + "\" onto " + handlerMethod); }
//this指RequestMappingHandlerMapping.MappingRegistry，mappingLookup保存着RequestMappingInfo--HandlerMethod对象 this.mappingLookup.put(mapping, handlerMethod); //获取 mapping 的PatternsRequestCondition的patterns，也就是拼接的URL路径，并且路径不包含* ？的就加入到集合返回 , List<String> directUrls = getDirectUrls(mapping);
for (String url : directUrls) {
//MappingRegistry的urlLookup保存着 url--RequestMappingInfo对象 this.urlLookup.add(url, mapping); } String name = null; if (getNamingStrategy() != null) {
//name属性感觉没用，如果@RequestMapping有name属性就是这个属性 如果没有就是 controller名字中的大写字母#方法名字，比如UC#test name = getNamingStrategy().getName(handlerMethod, mapping);
//MappingRegistry的nameLookup保存着 name--HandlerMethod集合 addMappingName(name, handlerMethod); } CorsConfiguration corsConfig = initCorsConfiguration(handler, method, mapping); if (corsConfig != null) { this.corsLookup.put(handlerMethod, corsConfig); }
//MappingRegistry的registry保存着RequestMappingInfo--MappingRegistration，MappingRegistration几乎有映射的所有信息 this.registry.put(mapping, new MappingRegistration<T>(mapping, handlerMethod, directUrls, name)); } finally { this.readWriteLock.writeLock().unlock(); //可重入锁 写锁 释放锁 } }
逻辑C：HandlerMethod对象 属性有bean，就是controller对象实例；beanFactory当前Spring容器；beanType就是controller的类型；method就是handler method；birdgeMethod是handler method的桥接方法；MethodParameter是handler method的方法参数，handlerMethod一般为null；
HandlerMethod，作用Spring给出了：一个handler method对象，包含了method以及controller对象，此外提供了便捷方式获取方法入参、返回值、注解等等；
逻辑D：createHandlerMethod方法只是调用了HandlerMethod的构造方法，构造方法中对方法入参进行了处理；
1 protected HandlerMethod createHandlerMethod(Object handler, Method method) { 2 HandlerMethod handlerMethod; 3 if (handler instanceof String) { 4 String beanName = (String) handler; 5 handlerMethod = new HandlerMethod(beanName, 6 getApplicationContext().getAutowireCapableBeanFactory(), method); 7 } 8 else { 9 handlerMethod = new HandlerMethod(handler, method); 10 } 11 return handlerMethod; 12 } 13
14 public HandlerMethod(String beanName, BeanFactory beanFactory, Method method) { 15 Assert.hasText(beanName, "Bean name is required"); 16 Assert.notNull(beanFactory, "BeanFactory is required"); 17 Assert.notNull(method, "Method is required"); 18 this.bean = beanName; //controller beanName 19 this.beanFactory = beanFactory; //当前controller所在Spring工厂 20 this.beanType = ClassUtils.getUserClass(beanFactory.getType(beanName)); //获取当前controller类型 21 this.method = method; //当前handler method 22 this.bridgedMethod = BridgeMethodResolver.findBridgedMethod(method); //查找method的桥接方法，没有桥接方法就是返回自身 23 this.parameters = initMethodParameters(); //初始化MethodParameter对象 设置了每个MethodParameter的method、parameterIndex属性 具体方法下图 24 this.resolvedFromHandlerMethod = null; 25 }
至此，registerHandlerMethod方法分析完毕，detectHandlerMethods方法分析完成，
Spring主要做了哪些工作：将所有请求映射关系保存到上面RequestMappingHandlerMapping的mappingRegistry的相关属性中，详情见上面表格。
分析过SpringMvc的请求流程 SpringMvc流程
篇幅太长，只分析如何找根据请求找到对应的handler？ 遍历HandlerMapping对象，调用其getHanlder方法查找controller / handler ， RequestMappingHandlerMapping对象的父类AbstractHandlerMapping实现了getHandler方法，方法最开始Object handler = getHandlerInternal(request)； 那么我们从AbstractHandlerMapping 的 getHandlerInternal开始记录.
1 protected HandlerMethod getHandlerInternal(HttpServletRequest request) throws Exception {
// 根据request请求路径以及servlet-mapping得到要请求URL 2 String lookupPath = getUrlPathHelper().getLookupPathForRequest(request); 3 if (logger.isDebugEnabled()) { 4 logger.debug("Looking up handler method for path " + lookupPath); 5 } 6 this.mappingRegistry.acquireReadLock(); //读锁 上锁 7 try {
// 这里就是MVC寻找controller匹配的方法！ 下面花大篇幅介绍下 8 HandlerMethod handlerMethod = lookupHandlerMethod(lookupPath, request);
9 if (logger.isDebugEnabled()) { 10 if (handlerMethod != null) { 11 logger.debug("Returning handler method [" + handlerMethod + "]"); 12 } 13 else { 14 logger.debug("Did not find handler method for [" + lookupPath + "]"); 15 } 16 }
//找到handlerMethod,但bean是controller beanName，用beanFactory getBean替换bean 17 return (handlerMethod != null ? handlerMethod.createWithResolvedBean() : null); 18 } 19 finally { 20 this.mappingRegistry.releaseReadLock(); 21 } 22 }
lookupHandlerMethod方法：
逻辑是这样的，先根据请求的URL 从 RequestMappingHandlerMapping的mappingRegistry的urlLookup中尝试寻找RequestMappingInfo；
寻找大致分为两种情况：一种请求URL清楚，不需要通配符比对，那肯定可以直接找到RequestMappingInfo集合，创建Match对象并且添加到集合里面，然后根据规则对Match集合排序选出最优解；
第二种情况URL带有通配符，那需要遍历映射关系再重复第一种情况。
1 protected HandlerMethod lookupHandlerMethod(String lookupPath, HttpServletRequest request) throws Exception { 2 List<Match> matches = new ArrayList<Match>();
//return this.urlLookup.get(urlPath);调用mappingRegistry的urlLookup根据URL寻找RequestMappingInfo 3 List<T> directPathMatches = this.mappingRegistry.getMappingsByUrl(lookupPath);
4 if (directPathMatches != null) {
//遍历找到的RequestMappingInfo集合， 然后寻找匹配的对象并处理添加到matches集合，见 逻辑E 分析 5 addMatchingMappings(directPathMatches, matches, request);
6 } 7 if (matches.isEmpty()) { //matches为空，有可能是因为通配符匹配的情况需要再次匹配 8 // No choice but to go through all mappings...
9 addMatchingMappings(this.mappingRegistry.getMappings().keySet(), matches, request); 10 } 11
12 if (!matches.isEmpty()) {
//返回一个MatchComparator对象
// 持有Comparator属性，并且compare方法是调用了RequestMappingInfo的compareTo 13 Comparator<Match> comparator = new MatchComparator(getMappingComparator(request));
//说到底排序还是调用了RequestMappingInfo的compareTo方法, 也存在优先级之分 URL路径>params>headers>comsume>produce>method 排序分析见文章最后 14 Collections.sort(matches, comparator); 15 if (logger.isTraceEnabled()) { 16 logger.trace("Found " + matches.size() + " matching mapping(s) for [" +
17 lookupPath + "] : " + matches); 18 } 19 Match bestMatch = matches.get(0); //找到最优匹配 20 if (matches.size() > 1) { 21 if (CorsUtils.isPreFlightRequest(request)) { 22 return PREFLIGHT_AMBIGUOUS_MATCH; 23 } 24 Match secondBestMatch = matches.get(1); 25 if (comparator.compare(bestMatch, secondBestMatch) == 0) { //存在两个匹配且相等 抛出异常 26 Method m1 = bestMatch.handlerMethod.getMethod(); 27 Method m2 = secondBestMatch.handlerMethod.getMethod(); 28 throw new IllegalStateException("Ambiguous handler methods mapped for HTTP path '" +
29 request.getRequestURL() + "': {" + m1 + ", " + m2 + "}"); 30 } 31 } 32 handleMatch(bestMatch.mapping, lookupPath, request); //解析URL变量，完成设置request属性等工作 33 return bestMatch.handlerMethod; //返回最优匹配的HandlerMethod对象 34 } 35 else { //没找到handlerMethod 就返回null 36 return handleNoMatch(this.mappingRegistry.getMappings().keySet(), lookupPath, request); 37 } 38 }
逻辑E：遍历找到的RequestMappingInfo集合，调用RequestMappingInfo的getMatchCondition进行匹配以获取匹配的RequestMappingInfo对象；寻找到合适的RequestMappingInfo对象之后，创建一个Match对象加入matches集合；
mappingRegistry 的 getMappings方法返回mappingLookup属性，上述表格mappingLookup 存放 RequestMappingInfo--HandlerMethod，根据RequestMappingInfo对象从map中取对象，(逻辑G分析 RequestMappingInfo重写了的hashCode以及equals方法)。 Match对象持有RequestMappingInfo以及HandlerMethod属性；此处方法调用结束matches可能包含多个Match结果；
RequestMappingInfo的getMatchingCondition方法
1 public RequestMappingInfo getMatchingCondition(HttpServletRequest request) {
//如果RequestMappingInfo没有指定methods属性，返回RequestMappingInfo本身，否则方法匹配 2 RequestMethodsRequestCondition methods = this.methodsCondition.getMatchingCondition(request);
//下面几个匹配逻辑是一样的，匹配了返回自身，没匹配返回null,具体参数作用、如何匹配看吧； 3 ParamsRequestCondition params = this.paramsCondition.getMatchingCondition(request); 4 HeadersRequestCondition headers = this.headersCondition.getMatchingCondition(request); 5 ConsumesRequestCondition consumes = this.consumesCondition.getMatchingCondition(request); 6 ProducesRequestCondition produces = this.producesCondition.getMatchingCondition(request); 7 //有一个条件匹配不上就直接返回null
8 if (methods == null || params == null || headers == null || consumes == null || produces == null) {
9 return null; 10 } 11 //其他匹配上了，最重要的匹配请求URL， 路径匹配作为 逻辑F 分析
12 PatternsRequestCondition patterns = this.patternsCondition.getMatchingCondition(request);
13 if (patterns == null) { 14 return null; 15 } 16
17 RequestConditionHolder custom = this.customConditionHolder.getMatchingCondition(request); 18 if (custom == null) { 19 return null; 20 } 21
22 return new RequestMappingInfo(this.name, patterns,
//寻找到匹配之后，构造一个新的RequestMappingInfo对象，持有上述匹配之后的结果返回 23 methods, params, headers, consumes, produces, custom.getCondition()); 24 }
逻辑F：PatternsRequestCondition匹配
调用PatternsRequestCondition 的 getMatchingPattern 方法进行URL匹配；遍历PatternsRequestCondition的 patterns属性，逐个getMatchingPattern进行比较，匹配上将pattern存入集合，并且使用AntPatternComparator进行排序，排序之后集合加入到一个新的PatternsRequestCondition对象中；
//pattern就是patterns属性当前迭代的元素，lookupPath就是servlet-mapping下请求URL
1 private String getMatchingPattern(String pattern, String lookupPath) { 2 if (pattern.equals(lookupPath)) { //两者相等 无疑义直接返回 这种是没有通配符 * ？这种都会很容易匹配到并且返回 3 return pattern; 4 } 5 if (this.useSuffixPatternMatch) { // useSuffixPatternMatch默认为true 6 if (!this.fileExtensions.isEmpty() && lookupPath.indexOf('.') != -1) { // fileExtensions默认为空 7 for (String extension : this.fileExtensions) { 8 if (this.pathMatcher.match(pattern + extension, lookupPath)) { 9 return pattern + extension; 10 } 11 } 12 } 13 else { 14 boolean hasSuffix = pattern.indexOf('.') != -1; //pattern字符串是否有 . 15 if (!hasSuffix && this.pathMatcher.match(pattern + ".*", lookupPath)) { //没有 . 就用AntPathMatcher的match匹配 pattern.* lookupPath 16 return pattern + ".*"; 17 } 18 } 19 } 20 if (this.pathMatcher.match(pattern, lookupPath)) {
// 用AntPathMatcher的match匹配 pattern lookupPath,匹配上就返回pattern 21 return pattern; 22 } 23 if (this.useTrailingSlashMatch) { 24 if (!pattern.endsWith("/") && this.pathMatcher.match(pattern + "/", lookupPath)) { 25 return pattern +"/"; 26 } 27 } 28 return null; 29 }
逻辑G：先介绍下为什么要看RequestMappingInfo的hashCode以及equals方法？RequestMappingInfo作为key存储在Map中，肯定需要重写HashCode以及equals方法；
RequestMappingInfo的hashCode以及equals方法： 比较的时候会先调用hashCode判断值是否相等，相等再比较equals方法，如果相等则认为是同一个对象；
先来看hashCode方法，将RequestMappingInfo的所有RequestCondition属性按公式求和，这些属性都是AbstractRequestCondition，equals和hashCode方法都调用了getContent方法，而AbstractRequestCondition的各种实现类的getContent方法，比如PatternsRequestCondition实现方式就是返回patterns（URL）集合；比如RequestMethodsRequestCondition实现就是返回methods集合；
RequestMappingInfo
1 public boolean equals(Object other) { 2 if (this == other) { 3 return true; 4 } 5 if (!(other instanceof RequestMappingInfo)) { 6 return false; 7 } 8 RequestMappingInfo otherInfo = (RequestMappingInfo) other; 9 return (this.patternsCondition.equals(otherInfo.patternsCondition) &&
10 this.methodsCondition.equals(otherInfo.methodsCondition) &&
11 this.paramsCondition.equals(otherInfo.paramsCondition) &&
12 this.headersCondition.equals(otherInfo.headersCondition) &&
13 this.consumesCondition.equals(otherInfo.consumesCondition) &&
14 this.producesCondition.equals(otherInfo.producesCondition) &&
15 this.customConditionHolder.equals(otherInfo.customConditionHolder)); 16 } 17
18 @Override 19 public int hashCode() { 20 return (this.patternsCondition.hashCode() * 31 + // primary differentiation
21 this.methodsCondition.hashCode() + this.paramsCondition.hashCode() +
22 this.headersCondition.hashCode() + this.consumesCondition.hashCode() +
23 this.producesCondition.hashCode() + this.customConditionHolder.hashCode()); 24 }
AbstractRequestCondition
1 public boolean equals(Object obj) { 2 if (this == obj) { 3 return true; 4 } 5 if (obj != null && getClass() == obj.getClass()) { 6 AbstractRequestCondition<?> other = (AbstractRequestCondition<?>) obj; 7 return getContent().equals(other.getContent()); 8 } 9 return false; 10 } 11
12 @Override 13 public int hashCode() { 14 return getContent().hashCode(); 15 }
分析到上面，getHandlerInternal已经找到了对应的HandlerMethod对象，调用getHandlerExecutionChain封装成HandlerExecutionChain；
1 protected HandlerExecutionChain getHandlerExecutionChain(Object handler, HttpServletRequest request) { 2 HandlerExecutionChain chain = (handler instanceof HandlerExecutionChain ?
3 (HandlerExecutionChain) handler : new HandlerExecutionChain(handler));
// //构造一个HandlerExecutionChain对象，持有handlerMethod 4
5 String lookupPath = this.urlPathHelper.getLookupPathForRequest(request); 6 for (HandlerInterceptor interceptor : this.adaptedInterceptors) {
//adaptedInterceptors在开启<mvc:annotation-drvien/>之后不为空，多了一个MappedInterceptor拦截器 7 if (interceptor instanceof MappedInterceptor) { 8 MappedInterceptor mappedInterceptor = (MappedInterceptor) interceptor; 9 if (mappedInterceptor.matches(lookupPath, this.pathMatcher)) { 10 chain.addInterceptor(mappedInterceptor.getInterceptor());
//将ConversionServiceExposingInterceptor添加到HandlerExecutionChain的interceptorList属性中 11 } 12 } 13 else { 14 chain.addInterceptor(interceptor); 15 } 16 } 17 return chain; //返回HandlerExecutionChain对象 18 }
Tip1：这个RequestMappingHandlerMapping的MappedInterceptor是从哪里注入的呢？
开启了<mvc:annotation-driven />之后 Spring向容器中注入了这样两个bean的定义，MappedInterceptor，该对象持有ConversionServiceExposingInterceptor对象；
容器中有了MappedInterceptor对象，什么时候给RequestMappingHandlerMapping设置的adaptedInterceptors呢？通过打断点分析到，RequestMappingHandlerMapping实现了ApplicationContextAware接口，Spring向其注入ApplicationContext的时候，调用了initApplicationContext方法，不断进入方法最后进入到父类AbstractHandlerMapping的initApplicationContext方法，
1 protected void initApplicationContext() throws BeansException { 2 extendInterceptors(this.interceptors); 3 detectMappedInterceptors(this.adaptedInterceptors); //此处添加了RequestMappingHandlerMapping的adaptedInterceptors 4 initInterceptors(); 5 } 6
7 protected void detectMappedInterceptors(List<HandlerInterceptor> mappedInterceptors) {
//这里将容器中的MappedInterceptor添加到了RequestMappingHandlerMapping的adaptedInterceptors 8 mappedInterceptors.addAll(
9 BeanFactoryUtils.beansOfTypeIncludingAncestors( 10 getApplicationContext(), MappedInterceptor.class, true, false).values()); 11 }
至此，如何找到HandlerMethod已经分析完毕；
总结
SpringMvc请求寻找规则 ： 如果一个请求同时匹配上多个方法，按照如下顺序选择执行哪个方法：
先URL匹配的方法 >>>>> params满足的方法 >>>>> headers 满足的方法 >>>>>>consume满足的方法 >>>> produce 满足的方法 >>>> method满足的方法
如果一个请求匹配上了多个RequestMappingInfo筛选：
之前介绍过排序是调用 RequestMappingInfo的compareTo进行排序
1 public int compareTo(RequestMappingInfo other, HttpServletRequest request) { 2 int result = this.patternsCondition.compareTo(other.getPatternsCondition(), request); //优先URL进行匹配 3 if (result != 0) { 4 return result; 5 } 6 result = this.paramsCondition.compareTo(other.getParamsCondition(), request); 7 if (result != 0) { 8 return result; 9 } 10 result = this.headersCondition.compareTo(other.getHeadersCondition(), request); 11 if (result != 0) { 12 return result; 13 } 14 result = this.consumesCondition.compareTo(other.getConsumesCondition(), request); 15 if (result != 0) { 16 return result; 17 } 18 result = this.producesCondition.compareTo(other.getProducesCondition(), request); 19 if (result != 0) { 20 return result; 21 } 22 result = this.methodsCondition.compareTo(other.getMethodsCondition(), request); 23 if (result != 0) { 24 return result; 25 } 26 result = this.customConditionHolder.compareTo(other.customConditionHolder, request); 27 if (result != 0) { 28 return result; 29 } 30 return 0; 31 }
介绍下URL如何排序吧，其他类似； 假设两个URL /get1 可以被匹配 /get* 以及 /get?
1 public int compareTo(PatternsRequestCondition other, HttpServletRequest request) { 2 String lookupPath = this.pathHelper.getLookupPathForRequest(request); 3 Comparator<String> patternComparator = this.pathMatcher.getPatternComparator(lookupPath); //获取AntPatternComparator比较器 4 Iterator<String> iterator = this.patterns.iterator(); 5 Iterator<String> iteratorOther = other.patterns.iterator(); 6 while (iterator.hasNext() && iteratorOther.hasNext()) { 7 int result = patternComparator.compare(iterator.next(), iteratorOther.next()); //URL比较规则在这里 8 if (result != 0) { 9 return result; 10 } 11 } 12 if (iterator.hasNext()) { 13 return -1; 14 } 15 else if (iteratorOther.hasNext()) { 16 return 1; 17 } 18 else { 19 return 0; 20 } 21 }
URL比较规则：按照请求URL通配符按一定权重计算排序顺序，{个数+*个数+ ** 个数 ；所以 get* 比get?排在前面；
1 public int compare(String pattern1, String pattern2) { //例子中pattern1为 /get* pattern2为/get? 2 PatternInfo info1 = new PatternInfo(pattern1); //具体查看下面构造方法 3 PatternInfo info2 = new PatternInfo(pattern2); 4
5 if (info1.isLeastSpecific() && info2.isLeastSpecific()) { 6 return 0; 7 } 8 else if (info1.isLeastSpecific()) { 9 return 1; 10 } 11 else if (info2.isLeastSpecific()) { //上面三种情况是 比较 /**的情况 12 return -1; 13 } 14
15 boolean pattern1EqualsPath = pattern1.equals(path); 16 boolean pattern2EqualsPath = pattern2.equals(path); 17 if (pattern1EqualsPath && pattern2EqualsPath) { 18 return 0; 19 } 20 else if (pattern1EqualsPath) { 21 return -1; 22 } 23 else if (pattern2EqualsPath) { //这三种情况是比较 pattern1 pattern2存在和请求URL完全匹配的情况 24 return 1; 25 } 26
27 if (info1.isPrefixPattern() && info2.getDoubleWildcards() == 0) { 28 return 1; 29 } 30 else if (info2.isPrefixPattern() && info1.getDoubleWildcards() == 0) { //哪个pattern的 /**多 哪个排在前面 31 return -1; 32 } 33
34 if (info1.getTotalCount() != info2.getTotalCount()) { 35 return info1.getTotalCount() - info2.getTotalCount(); //按照权重来排序了 {算1 *算1 **算2 哪个大哪个排前面 /get*权重为1排前面 36 } 37
38 if (info1.getLength() != info2.getLength()) { 39 return info2.getLength() - info1.getLength(); 40 } 41
42 if (info1.getSingleWildcards() < info2.getSingleWildcards()) { 43 return -1; 44 } 45 else if (info2.getSingleWildcards() < info1.getSingleWildcards()) { 46 return 1; 47 } 48
49 if (info1.getUriVars() < info2.getUriVars()) { 50 return -1; 51 } 52 else if (info2.getUriVars() < info1.getUriVars()) { 53 return 1; 54 } 55
56 return 0; 57 } 58
59 public PatternInfo(String pattern) { 60 this.pattern = pattern; 61 if (this.pattern != null) { 62 initCounters(); 63 this.catchAllPattern = this.pattern.equals("/**"); //代表匹配所有就是pattern为 /** 64 this.prefixPattern = !this.catchAllPattern && this.pattern.endsWith("/**"); 65 } 66 if (this.uriVars == 0) { 67 this.length = (this.pattern != null ? this.pattern.length() : 0); 68 } 69 } 70
71 protected void initCounters() { 72 int pos = 0; 73 while (pos < this.pattern.length()) { 74 if (this.pattern.charAt(pos) == '{') { //存在变量 则uriVars自增 75 this.uriVars++; 76 pos++; 77 } 78 else if (this.pattern.charAt(pos) == '*') { //解析到* 79 if (pos + 1 < this.pattern.length() && this.pattern.charAt(pos + 1) == '*') { 80 this.doubleWildcards++; // doubleWildcards代表有两个*的 81 pos += 2; 82 } 83 else if (pos > 0 && !this.pattern.substring(pos - 1).equals(".*")) { //最后一位是* 且倒数第二位不是 * 84 this.singleWildcards++; // singleWildcards代表有单个* 85 pos++; 86 } 87 else { 88 pos++; 89 } 90 } 91 else { 92 pos++; 93 } 94 } 95 }