SpringMVC源码总结(十一)mvc:interceptors拦截器介绍

本文章针对mvc:interceptors标签进行介绍,它的注册过程以及在访问时的拦截过程。 

首先说下接口HandlerInterceptor,它有如下三个方法:
 
1
2
3
4
5
6
7
8
9
10
boolean preHandle(HttpServletRequest request, HttpServletResponse response, Object handler)
        throws Exception;
 
void postHandle(
            HttpServletRequest request, HttpServletResponse response, Object handler, ModelAndView modelAndView)
            throws Exception;
 
void afterCompletion(
            HttpServletRequest request, HttpServletResponse response, Object handler, Exception ex)
            throws Exception;


正常情况下,对于preHandle就是在在处理函数之前先执行,然后再执行处理函数,接着执行postHandle,最后再执行afterCompletion。afterCompletion无论是否出错是肯定要执行的,而postHandle则不是,不一定会执行。之后看源代码就知道他们的执行情况。 

AsyncHandlerInterceptor接口则增添了afterConcurrentHandlingStarted方法,对于此还未研究,先不讨论。 

HandlerInterceptorAdapter则默认实现了上述的接口,所以当我们仅仅要实现某个方法时,只需继承HandlerInterceptorAdapter,然后覆盖相应的方法。 

然后我们就写一个类继承HandlerInterceptorAdapter来进行实验:LoginInterceptor如下:
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
@Override
    public boolean preHandle(HttpServletRequest request,
            HttpServletResponse response, Object handler) throws Exception {
        System.out.println("preHandle");
        return super.preHandle(request, response, handler);
    }
 
    @Override
    public void postHandle(HttpServletRequest request,
            HttpServletResponse response, Object handler,
            ModelAndView modelAndView) throws Exception {
        System.out.println("postHandle");
        super.postHandle(request, response, handler, modelAndView);
    }
 
    @Override
    public void afterCompletion(HttpServletRequest request,
            HttpServletResponse response, Object handler, Exception ex)
            throws Exception {
        System.out.println("afterCompletion");
        super.afterCompletion(request, response, handler, ex);
    }
 
    @Override
    public void afterConcurrentHandlingStarted(HttpServletRequest request,
            HttpServletResponse response, Object handler) throws Exception {
        System.out.println("afterConcurrentHandlingStarted");
        super.afterConcurrentHandlingStarted(request, response, handler);
    }

没有做具体的内容,仅仅是打印出一些信息,方便查看执行顺序。 
该接口的基本内容说完了,然后就看下它的配置说明:
 
1
2
3
4
5
6
7
8
<mvc:interceptors path-matcher="xxx">
        <mvc:interceptor>
            <mvc:mapping path="xxx"/>
            <mvc:exclude-mapping path="xxxx"/>
            <bean class="xxxx"></bean>
        </mvc:interceptor>
        <bean class="com.lg.mvc.interceptor.LoginInterceptor" />
    </mvc:interceptors>

其实在mvc:interceptors标签中,有两种类型的配置,一种直接配置一个bean(bean和ref归为一类),另一种还要配置上拦截的路径和排除的路径。直接配置的bean那就代表对所有的请求进行拦截,而对于mvc:interceptor则代表有着更精细的控制。 

而mvc:interceptors的属性path-matcher则表示配置一个自定义的PathMatcher,它主要用来处理路径的匹配规则,默认采用的PathMatcher为AntPathMatcher,具有ant风格的路径规则,如?表示任何单字符,*表示0个或多个字符,**表示0个或多个目录。 
对于本工程来说具体的配置如下:
 
1
2
3
<mvc:interceptors>
        <bean class="com.lg.mvc.interceptor.LoginInterceptor" />
    </mvc:interceptors>

然后就进行源代码分析: 
如何来处理xml文件中所配置的这些HandlerInterceptor的呢? 
对于mvc:interceptors的解析同样需要我们去看BeanDefinitionParser的实现类,最终会找到InterceptorsBeanDefinitionParser:
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
public BeanDefinition parse(Element element, ParserContext parserContext) {
        CompositeComponentDefinition compDefinition = new CompositeComponentDefinition(element.getTagName(), parserContext.extractSource(element));
        parserContext.pushContainingComponent(compDefinition);
//判断是否自定义了PathMatcher
        RuntimeBeanReference pathMatcherRef = null;
        if (element.hasAttribute("path-matcher")) {
            pathMatcherRef = new RuntimeBeanReference(element.getAttribute("path-matcher"));
        }
//获取所有的interceptor,在这里我们可以看到所有的interceptor最终都会构建成一个
//MappedInterceptor
        List<Element> interceptors = DomUtils.getChildElementsByTagName(element, "bean", "ref", "interceptor");
        for (Element interceptor : interceptors) {
            RootBeanDefinition mappedInterceptorDef = new RootBeanDefinition(MappedInterceptor.class);
            mappedInterceptorDef.setSource(parserContext.extractSource(interceptor));
            mappedInterceptorDef.setRole(BeanDefinition.ROLE_INFRASTRUCTURE);
 
            ManagedList<String> includePatterns = null;
            ManagedList<String> excludePatterns = null;
            Object interceptorBean;
            if ("interceptor".equals(interceptor.getLocalName())) {
                includePatterns = getIncludePatterns(interceptor, "mapping");
                excludePatterns = getIncludePatterns(interceptor, "exclude-mapping");
                Element beanElem = DomUtils.getChildElementsByTagName(interceptor, "bean", "ref").get(0);
                interceptorBean = parserContext.getDelegate().parsePropertySubElement(beanElem, null);
            }
            else {
                interceptorBean = parserContext.getDelegate().parsePropertySubElement(interceptor, null);
            }
            mappedInterceptorDef.getConstructorArgumentValues().addIndexedArgumentValue(0, includePatterns);
            mappedInterceptorDef.getConstructorArgumentValues().addIndexedArgumentValue(1, excludePatterns);
            mappedInterceptorDef.getConstructorArgumentValues().addIndexedArgumentValue(2, interceptorBean);
 
            if (pathMatcherRef != null) {
                mappedInterceptorDef.getPropertyValues().add("pathMatcher", pathMatcherRef);
            }
 
            String beanName = parserContext.getReaderContext().registerWithGeneratedName(mappedInterceptorDef);
            parserContext.registerComponent(new BeanComponentDefinition(mappedInterceptorDef, beanName));
        }
 
        parserContext.popAndRegisterContainingComponent();
        return null;
    }

这里就引出来MappedInterceptor的结构类型: 
1
2
3
4
5
6
7
private final String[] includePatterns;
 
    private final String[] excludePatterns;
 
    private final HandlerInterceptor interceptor;
 
    private PathMatcher pathMatcher;

到这里就很明白了,虽然在mvc:interceptors标签中,配置interceptor形式不一样,但是最终都将以MappedInterceptor形式存储,同时来看下MappedInterceptor的match的方法: 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
public boolean matches(String lookupPath, PathMatcher pathMatcher) {
        PathMatcher pathMatcherToUse = (this.pathMatcher != null) ? this.pathMatcher : pathMatcher;
        if (this.excludePatterns != null) {
            for (String pattern : this.excludePatterns) {
                if (pathMatcherToUse.match(pattern, lookupPath)) {
                    return false;
                }
            }
        }
        if (this.includePatterns == null) {
            return true;
        }
        else {
            for (String pattern : this.includePatterns) {
                if (pathMatcherToUse.match(pattern, lookupPath)) {
                    return true;
                }
            }
            return false;
        }
    }

这里便是PathMatcher对于excludePatterns、includePatterns 的使用规则,同时表明本身的PathMatcher若为空,则使用外部传来的PathMatcher。 
至此解析mvc:interceptors标签的过程就完成了。它们最终会注册到ApplicationContext的上下文环境中,等待被使用。 

谁会是他们的使用者呢?我们慢慢来看: 
对于每一个请求,HandlerMapping都会找到对应的handler,并最终封装成一个HandlerExecutionChain,这个HandlerExecutionChain包含有handler和它对应的interceptors,HandlerExecutionChain如下:
 
1
2
3
4
5
6
7
private final Object handler;
 
    private HandlerInterceptor[] interceptors;
 
    private List<HandlerInterceptor> interceptorList;
 
    private int interceptorIndex = -1;

既然是由HandlerMapping来产生的HandlerExecutionChain,则它需要为每一个它所管辖的handler来装配HandlerInterceptor。所以HandlerMapping必然是mvc:interceptors标签内容的使用者。 
使用者:AbstractHandlerMapping,它的属性有:
 
1
2
3
4
5
6
7
private PathMatcher pathMatcher = new AntPathMatcher();
 
    private final List<Object> interceptors = new ArrayList<Object>();
 
    private final List<HandlerInterceptor> adaptedInterceptors = new ArrayList<HandlerInterceptor>();
 
    private final List<MappedInterceptor> mappedInterceptors = new ArrayList<MappedInterceptor>();

这里便可以看到,它所使用的默认的PathMatcher为AntPathMatcher。接下来我们看下AbstractHandlerMapping的初始化方法: 
1
2
3
4
5
protected void initApplicationContext() throws BeansException {
        extendInterceptors(this.interceptors);
        detectMappedInterceptors(this.mappedInterceptors);
        initInterceptors();
    }

detectMappedInterceptors探测ApplicationContext中已经解析过的MappedInterceptor,如下: 
1
2
3
4
5
protected void detectMappedInterceptors(List<MappedInterceptor> mappedInterceptors) {
        mappedInterceptors.addAll(
                BeanFactoryUtils.beansOfTypeIncludingAncestors(
                        getApplicationContext(), MappedInterceptor.class, true, false).values());
    }

全部存放到AbstractHandlerMapping的mappedInterceptors属性上。 
然后我们继续看看在请求到来时的具体拦截过程:
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
protected void doDispatch(HttpServletRequest request, HttpServletResponse response) throws Exception {
        HttpServletRequest processedRequest = request;
        HandlerExecutionChain mappedHandler = null;
        boolean multipartRequestParsed = false;
 
        WebAsyncManager asyncManager = WebAsyncUtils.getAsyncManager(request);
 
        try {
            ModelAndView mv = null;
            Exception dispatchException = null;
 
            try {
                processedRequest = checkMultipart(request);
                multipartRequestParsed = (processedRequest != request);
 
                // Determine handler for the current request.
                mappedHandler = getHandler(processedRequest);
//略
}

对于每个请求先找到对应的HandlerMapping,然后由这个handlerMapping来找到对应请求的handler,然后由handlerMapping自身的interceptor和这个handler来构建一个HandlerExecutionChain。代码如下: 
1
2
3
4
5
6
7
8
9
10
11
12
13
protected HandlerExecutionChain getHandler(HttpServletRequest request) throws Exception {
        for (HandlerMapping hm : this.handlerMappings) {
            if (logger.isTraceEnabled()) {
                logger.trace(
                        "Testing handler map [" + hm + "] in DispatcherServlet with name '" + getServletName() + "'");
            }
            HandlerExecutionChain handler = hm.getHandler(request);
            if (handler != null) {
                return handler;
            }
        }
        return null;
    }

这里便是找到一个合适的HandlerMapping,继续看下hm.getHandler(request)这个方法。 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
public final HandlerExecutionChain getHandler(HttpServletRequest request) throws Exception {
        Object handler = getHandlerInternal(request);
        if (handler == null) {
            handler = getDefaultHandler();
        }
        if (handler == null) {
            return null;
        }
        // Bean name or resolved handler?
        if (handler instanceof String) {
            String handlerName = (String) handler;
            handler = getApplicationContext().getBean(handlerName);
        }
        return getHandlerExecutionChain(handler, request);
    }

这里便是找到对应请求的handler。getHandlerExecutionChain(handler, request)这里便是构建HandlerExecutionChain的地方: 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
protected HandlerExecutionChain getHandlerExecutionChain(Object handler, HttpServletRequest request) {
        HandlerExecutionChain chain = (handler instanceof HandlerExecutionChain ?
                (HandlerExecutionChain) handler : new HandlerExecutionChain(handler));
        chain.addInterceptors(getAdaptedInterceptors());
 
        String lookupPath = this.urlPathHelper.getLookupPathForRequest(request);
        for (MappedInterceptor mappedInterceptor : this.mappedInterceptors) {
            if (mappedInterceptor.matches(lookupPath, this.pathMatcher)) {
                chain.addInterceptor(mappedInterceptor.getInterceptor());
            }
        }
 
        return chain;
    }

对于我们关注的重点为它会遍历AbstarctHandlerMapping的mappedInterceptors属性,然后使用默认的pathMatcher,即AntPathMatcher来判断当前的请求是否符合拦截条件,若符合则将mappedInterceptor放进HandlerExecutionChain 中。 

至此一个HandlerExecutionChain便构建好了,包含一个handler和这个handler对应的interceptor。然后看下interceptor的执行过程:
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
protected void doDispatch(HttpServletRequest request, HttpServletResponse response) throws Exception {
        HttpServletRequest processedRequest = request;
        HandlerExecutionChain mappedHandler = null;
        boolean multipartRequestParsed = false;
 
        WebAsyncManager asyncManager = WebAsyncUtils.getAsyncManager(request);
 
        try {
            ModelAndView mv = null;
            Exception dispatchException = null;
 
            try {
                processedRequest = checkMultipart(request);
                multipartRequestParsed = (processedRequest != request);
 
                // Determine handler for the current request.
                mappedHandler = getHandler(processedRequest);
                if (mappedHandler == null || mappedHandler.getHandler() == null) {
                    noHandlerFound(processedRequest, response);
                    return;
                }
 
                // Determine handler adapter for the current request.
                HandlerAdapter ha = getHandlerAdapter(mappedHandler.getHandler());
 
                // Process last-modified header, if supported by the handler.
                String method = request.getMethod();
                boolean isGet = "GET".equals(method);
                if (isGet || "HEAD".equals(method)) {
                    long lastModified = ha.getLastModified(request, mappedHandler.getHandler());
                    if (logger.isDebugEnabled()) {
                        logger.debug("Last-Modified value for [" + getRequestUri(request) + "] is: " + lastModified);
                    }
                    if (new ServletWebRequest(request, response).checkNotModified(lastModified) && isGet) {
                        return;
                    }
                }
//重点1 这里执行interceptor的preHandle方法
                if (!mappedHandler.applyPreHandle(processedRequest, response)) {
                    return;
                }
//这里执行处理函数
                try {
                    // Actually invoke the handler.
                    mv = ha.handle(processedRequest, response, mappedHandler.getHandler());
                }
                finally {
                    if (asyncManager.isConcurrentHandlingStarted()) {
                        return;
                    }
                }
 
                applyDefaultViewName(request, mv);
//重点2:这里执行interceptor的postHandle方法
                mappedHandler.applyPostHandle(processedRequest, response, mv);
            }
            catch (Exception ex) {
                dispatchException = ex;
            }
//重点3:这里执行interceptor的afterCompletion方法
            processDispatchResult(processedRequest, response, mappedHandler, mv, dispatchException);
        }
//重点4:当出现异常时,仍然执行afterCompletion方法
        catch (Exception ex) {
            triggerAfterCompletion(processedRequest, response, mappedHandler, ex);
        }
        catch (Error err) {
            triggerAfterCompletionWithError(processedRequest, response, mappedHandler, err);
        }
        finally {
            if (asyncManager.isConcurrentHandlingStarted()) {
                // Instead of postHandle and afterCompletion
                mappedHandler.applyAfterConcurrentHandlingStarted(processedRequest, response);
                return;
            }
            // Clean up any resources used by a multipart request.
            if (multipartRequestParsed) {
                cleanupMultipart(processedRequest);
            }
        }
    }

先看重点1: 
1
2
3
4
5
6
7
8
9
10
11
12
13
boolean applyPreHandle(HttpServletRequest request, HttpServletResponse response) throws Exception {
        if (getInterceptors() != null) {
            for (int i = 0; i < getInterceptors().length; i++) {
                HandlerInterceptor interceptor = getInterceptors()[i];
                if (!interceptor.preHandle(request, response, this.handler)) {
                    triggerAfterCompletion(request, response, null);
                    return false;
                }
                this.interceptorIndex = i;
            }
        }
        return true;
    }

执行preHandle方法,一旦有一个preHandle返回false,则触发triggerAfterCompletion: 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
void triggerAfterCompletion(HttpServletRequest request, HttpServletResponse response, Exception ex)
            throws Exception {
 
        if (getInterceptors() == null) {
            return;
        }
        for (int i = this.interceptorIndex; i >= 0; i--) {
            HandlerInterceptor interceptor = getInterceptors()[i];
            try {
                interceptor.afterCompletion(request, response, this.handler, ex);
            }
            catch (Throwable ex2) {
                logger.error("HandlerInterceptor.afterCompletion threw exception", ex2);
            }
        }
    }

看下这里的for循环的条件,从interceptorIndex开始到0,逆序执行interceptor.afterCompletion。 
重点2 postHandle:
 
1
2
3
4
5
6
7
8
9
void applyPostHandle(HttpServletRequest request, HttpServletResponse response, ModelAndView mv) throws Exception {
        if (getInterceptors() == null) {
            return;
        }
        for (int i = getInterceptors().length - 1; i >= 0; i--) {
            HandlerInterceptor interceptor = getInterceptors()[i];
            interceptor.postHandle(request, response, this.handler, mv);
        }
    }

这个没有什么特殊,preHandle只有参数HttpServletRequest和HttpServletResponse,而postHandle则加入了返回结果ModelAndView,我们可以对ModelAndView进行进一步的修改,此时的view(若有)还没有经过渲染。 
重点3 :
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
private void processDispatchResult(HttpServletRequest request, HttpServletResponse response,
            HandlerExecutionChain mappedHandler, ModelAndView mv, Exception exception) throws Exception {
 
        boolean errorView = false;
 
        if (exception != null) {
            if (exception instanceof ModelAndViewDefiningException) {
                logger.debug("ModelAndViewDefiningException encountered", exception);
                mv = ((ModelAndViewDefiningException) exception).getModelAndView();
            }
            else {
                Object handler = (mappedHandler != null ? mappedHandler.getHandler() : null);
                mv = processHandlerException(request, response, handler, exception);
                errorView = (mv != null);
            }
        }
 
        // Did the handler return a view to render?
        if (mv != null && !mv.wasCleared()) {
            render(mv, request, response);
            if (errorView) {
                WebUtils.clearErrorRequestAttributes(request);
            }
        }
        else {
            if (logger.isDebugEnabled()) {
                logger.debug("Null ModelAndView returned to DispatcherServlet with name '" + getServletName() +
                        "': assuming HandlerAdapter completed request handling");
            }
        }
 
        if (WebAsyncUtils.getAsyncManager(request).isConcurrentHandlingStarted()) {
            // Concurrent handling started during a forward
            return;
        }
 
        if (mappedHandler != null) {
            mappedHandler.triggerAfterCompletion(request, response, null);
        }
    }

这里可以看到,如果有view,则渲染完成之后,才会执行triggerAfterCompletion,同时不再拥有对ModelAndView的处理(已经完成了渲染)。所以我们就可以看到当有view时,afterCompletion和postHandle的明显区别。 
重点4:当执行过程发生异常时,也会执行interceptor的afterCompletion方法。 
这里要做下说明,对于preHandler方法是获取不到处理函数的参数值的,如果想对处理函数的参数值进行拦截处理,则要使用Spring AOP。 

上一篇:java基础之:详解内部类(转载)


下一篇:Hibernate【映射】知识要点(三)