Struts2攔截器原理

   //如果不是struts的請求則繼續由其它過濾器執行

//包裝request，對有文件上傳的特殊處理下

</blockquote>                 request = prepare.wrapRequest(request);

//查找對應的ActionMapping

</blockquote>                 ActionMapping mapping = prepare.findActionMapping(request, response, true);

//如果找不到ActionMapping則當作靜態資源來處理

</blockquote>                 if (mapping == null) {
                    boolean handled = execute.executeStaticResourceRequest(request, response);
                    if (!handled) {
                        chain.doFilter(request, response);
                    }
                } else {

  //使用ActionMapping來執行action

</blockquote>                     execute.executeAction(request, response, mapping);</strong>
                }
            }
        } finally {
            prepare.cleanupRequest(request);
        }
    }</span></td> </tr> </tbody> </table>   
跟蹤execute.executeAction()，則到了 org.apache.struts2.dispatcher.Dispatcher，如下：

public void serviceAction(HttpServletRequest request, HttpServletResponse response, ServletContext context,                               ActionMapping mapping) throws ServletException {         Map<String, Object> extraContext = createContextMap(request, response, mapping, context);         // If there was a previous value stack, then create a new copy and pass it in to be used by the new Action         ValueStack stack = (ValueStack) request.getAttribute(ServletActionContext.STRUTS_VALUESTACK_KEY);         boolean nullStack = stack == null;         if (nullStack) {             ActionContext ctx = ActionContext.getContext();             if (ctx != null) {                 stack = ctx.getValueStack();             }         }         if (stack != null) {             extraContext.put(ActionContext.VALUE_STACK, valueStackFactory.createValueStack(stack));         }         String timerKey = "Handling request from Dispatcher";         try {             UtilTimerStack.push(timerKey);             String namespace = mapping.getNamespace();             String name = mapping.getName();             String method = mapping.getMethod();             Configuration config = configurationManager.getConfiguration();    //使用StrutsActionProxyFactory(ActionProxyFactory的一個實現)創建action代理對象    //proxy實際上是org.apache.struts2.impl.StrutsActionProxy類型             ActionProxy proxy = config.getContainer().getInstance(ActionProxyFactory.class).createActionProxy(                     namespace, name, method, extraContext, true, false);</strong></span>             request.setAttribute(ServletActionContext.STRUTS_VALUESTACK_KEY, proxy.getInvocation().getStack());             // if the ActionMapping says to go straight to a result, do it!             if (mapping.getResult() != null) {                 Result result = mapping.getResult();                 result.execute(proxy.getInvocation());             } else { //執行action </blockquote>                 proxy.execute();             }             // If there was a previous value stack then set it back onto the request             if (!nullStack) {                 request.setAttribute(ServletActionContext.STRUTS_VALUESTACK_KEY, stack);             }         } catch (ConfigurationException e) {             // WW-2874 Only log error if in devMode             if(devMode) {                 String reqStr = request.getRequestURI();                 if (request.getQueryString() != null) {                     reqStr = reqStr + "?" + request.getQueryString();                 }                 LOG.error("Could not find action or result\n" + reqStr, e);             }             else {                 LOG.warn("Could not find action or result", e);             }             sendError(request, response, context, HttpServletResponse.SC_NOT_FOUND, e);         } catch (Exception e) {             sendError(request, response, context, HttpServletResponse.SC_INTERNAL_SERVER_ERROR, e);         } finally {             UtilTimerStack.pop(timerKey);         }     }</td> </tr> </tbody> </table> DefaultActionProxyFactory創建ActionProxy，在com.opensymphony.xwork2.DefaultActionProxyFactory： </tr> </tbody> </table>       接下來看看 org.apache.struts2.impl.StrutsActionProxy的execute()方法，如下： </span>      public ActionProxy createActionProxy(String namespace, String actionName, String methodName, Map<String, Object> extraContext, boolean executeResult, boolean cleanupContext) {                  ActionInvocation inv = new DefaultActionInvocation(extraContext, true);         container.inject(inv);         return createActionProxy(inv, namespace, actionName, methodName, executeResult, cleanupContext);     } </tr> </tbody> </table>    最關鍵的，com.opensymphony.xwork2.DefaultActionInvocation.invoke()方法，這個DefaultActionInvocation是ActionInvocation的一個實現類，如下：      public String execute() throws Exception {         ActionContext previous = ActionContext.getContext();         ActionContext.setContext(invocation.getInvocationContext());         try {        //這里就是調用攔截器的入口了             return invocation.invoke();         } finally {             if (cleanupContext)                 ActionContext.setContext(previous);         }     }   //保存了執行當前action方法時需要調用的攔截器棧，按照struts.xml中配制的攔截器順序，從前到后，依次加入到了這個Iterator里面   protected Iterator<InterceptorMapping> interceptors;</strong>   public String invoke() throws Exception {         String profileKey = "invoke: ";         try {             UtilTimerStack.push(profileKey);             if (executed) {                 throw new IllegalStateException("Action has already executed");             }            </span> //如果當前還有下一個，則繼續執行攔截器             if (interceptors.hasNext()</strong>) {                 final InterceptorMapping interceptor = (InterceptorMapping) interceptors.next();                 String interceptorMsg = "interceptor: " + interceptor.getName();                 UtilTimerStack.push(interceptorMsg);                 try { </span>      //執行攔截器的intercept()方法，并將當前ActionInvocation對象傳遞給這個方法      //這樣，當一個攔截器執行完自己的處理后，需要讓框架繼續執行下一個攔截器的時候，直接使用actionInvocation.invoke()方法，當前這個方法又會被調一次，這其實就是一個遞歸了，遞歸方法是ActionInvocation.invoke()，結束條件是 interceptors.hasNext() </blockquote> </blockquote>                                 resultCode = interceptor.getInterceptor().intercept(DefaultActionInvocation.this);                             }                 finally {                     UtilTimerStack.pop(interceptorMsg);                 }             } else { //攔截器全部都執行了，那么最后來執行action，跳出遞歸了</strong>                 resultCode = invokeActionOnly();             }             // this is needed because the result will be executed, then control will return to the Interceptor, which will             // return above and flow through again             if (!executed) {                 if (preResultListeners != null) {                     for (Object preResultListener : preResultListeners) {                         PreResultListener listener = (PreResultListener) preResultListener;                         String _profileKey = "preResultListener: ";                         try {                             UtilTimerStack.push(_profileKey);                             listener.beforeResult(this, resultCode);                         }                         finally {                             UtilTimerStack.pop(_profileKey);                         }                     }                 }                 // now execute the result, if we're supposed to                 if (proxy.getExecuteResult()) {                     executeResult();                 }                 executed = true;             }             return resultCode;         }         finally {             UtilTimerStack.pop(profileKey);         }     }</span></td> </tr> </tbody> </table>    基本原理到此為止，下面弄個小例子再說明一下： //攔截器，相當于struts2的攔截器  public interface Interceptor{     String intercept(InvocationContext context); } //很多攔截器的實現 public class ExceptionInterceptor implements Interceptor {     public String intercept(InvocationContext context) {         // 對異常的處理         System.out.println("\t\t\tExceptionInterceptor 處理異常");         return context.invoke();     } } public class FileUploadInterceptor implements Interceptor {     public String intercept(InvocationContext context) {         // 處理文件上傳相關         System.out.println("\t\t\tFileUploadInterceptor 處理文件上傳");         return context.invoke();     } } public class ParameterInterceptor implements Interceptor {     public String intercept(InvocationContext context) {         // 處理請求的參數         System.out.println("\t\t\tParameterInterceptor 處理請求參數");         return context.invoke();     } } //執行攔截器的invocation上下文，相當于struts2的ActionInvocation public class InvocationContext </strong>{     </span>// 這里存放當前執行當前action所需要執行的攔截器棧     private Iterator<Interceptor> interceptorIterator = null;     private String prefix = "";     public InvocationContext() {         // 模擬從配制文件中相應的規則取攔截器棧         ArrayList<Interceptor> list = new ArrayList<Interceptor>();         list.add(new ExceptionInterceptor());         list.add(new FileUploadInterceptor());         list.add(new ParameterInterceptor());         interceptorIterator = list.iterator();     }     public String invoke() {         // 是否還有攔截器需要執行         if (interceptorIterator.hasNext()) {             // 獲取下一個需要執行的攔截器             Interceptor interceptor = interceptorIterator.next();             String name = interceptor.getClass().getName();             name = prefix + name;             System.out.println(name + " intercept start...");             prefix += "\t";             // Interceptor的所有intercept方法實現里面，最后都調用了InvocationContext.invoke()方法             // 其實就是一個遞歸，只不過invoke()的下一個遞歸是在Interceptor.intercept()里面調用的             // 所以說為什么Interceptor.intercept()方法要加個InvocationContext的參數呢，作用就在于此             String result = interceptor.intercept(this);             System.out.println(name + " intercept end...");             return result;         } else {// 所有的攔截器都執行完了，那就來執行action對應的方法             return executeAction();         }     }     private String executeAction() {         System.out.println(prefix + "executeAction success.");         return "success";     } } //模擬請求進行測試 public class Test {     public static void main(String[] args) {         InvocationContext context = new InvocationContext();         System.out.println("請求開始了...");         context.invoke();         System.out.println("請求處理完了...");     } }  本文由用戶 jopen 自行上傳分享，僅供網友學習交流。所有權歸原作者，若您的權利被侵害，請聯系管理員。  轉載本站原創文章，請注明出處，并保留原始鏈接、圖片水印。  本站是一個以用戶分享為主的開源技術平臺，歡迎各類分享！  本文地址：http://www.baiduhome.net/lib/view/open1390530408367.html Struts2 Web框架 相關經驗 Struts2攔截器的原理與實現 Struts2攔截器原理 Struts2攔截器 Struts2 自定義攔截器(方法攔截器) Struts2基本原理 相關資訊 Android targetSdkVersion 原理 HashMap的工作原理 Google 圖片搜索的原理 Reddit排名算法工作原理 相似圖片搜索的原理 相關文檔  Struts2原理  Struts2 攔截器的使用  Struts2 攔截器  Struts2攔截器與sitemesh  02 Struts2攔截器課件  Struts2攔截器