在继续分析TimedCancelable类及QueryTraverser类之前有必要熟悉一下ThreadPool类和TaskHandle类， 这两者的作用是以线程池的方式执行连接器的功能

ThreadPool类源码如下：

/** * Pool for running {
    @link TimedCancelable}, time limited tasks. * * 
 * Users are provided a {
    @link TaskHandle} for each task. The {
    @link TaskHandle} * supports canceling the task and determining if the task is done running. * * 
 * The ThreadPool enforces a configurable maximum time interval for tasks. Each * task is guarded by a time out task that will cancel the primary task * if the primary task does not complete within the allowed interval. * * 
 * Task cancellation includes two actions that are visible for the task task's * {
    @link TimedCancelable} * 
     
       * 
      
Calling {
    @link Future#cancel(boolean)} to send the task an interrupt and * mark it as done. * 
      
Calling {
    @link TimedCancelable#cancel()} to send the task a second signal * that it is being canceled. This signal has the benefit that it does not * depend on the tasks interrupt handling policy. * 
     
 * Once a task has been canceled its {
    @link TaskHandle#isDone()} method will * immediately start returning true. * * 
 * {
    @link ThreadPool} performs the following processing when a task completes * 
     
       * 
      
Cancel the time out task for the completed task. * 
      
Log exceptions that indicate the task did not complete normally. * 
     
 */public class ThreadPool {  private static final Logger LOGGER =      Logger.getLogger(ThreadPool.class.getName());  /**   * The default amount of time in to wait for tasks to complete during during   * shutdown.   */  public static final int DEFAULT_SHUTDOWN_TIMEOUT_MILLIS = 10 * 1000;  /**   * Configured amount of time to let tasks run before automatic cancellation.   */  private final long maximumTaskLifeMillis;  /**   * ExecutorService for running submitted tasks. Tasks are only submitted   * through completionService.   */  private final ExecutorService executor =    Executors.newCachedThreadPool(        new ThreadNamingThreadFactory("ThreadPoolExecutor"));  /**   * CompletionService for running submitted tasks. All tasks are submitted   * through this CompletionService to provide blocking, queued access to   * completion information.   */  private final CompletionService
      completionService =      new ExecutorCompletionService

这里主要用到了java的多线程处理框架

TaskHandle类封装了线程执行的返回信息并提供对线程进行操作的方法，其源码如下：

/** * Handle for the management of a {
    @link TimedCancelable} primary task. */public class TaskHandle {  /**   * The primary {
    @link TimedCancelable} that is run by this task to   * perform some useful work.   */  final TimedCancelable cancelable;  /*   * The {@link future} for the primary task.   */  final Future
      taskFuture;  /*   * The time the task starts.   */  final long startTime;  /**   * Create a TaskHandle.   *   * @param cancelable {
    @link TimedCancelable} for the primary task.   * @param taskFuture {
    @link Future} for the primary task.   * @param startTime startTime for the primary task.   */  TaskHandle(TimedCancelable cancelable, Future
      taskFuture, long startTime) {    this.cancelable = cancelable;    this.taskFuture = taskFuture;    this.startTime = startTime;  }  /**   * Cancel the primary task and the time out task.   */  public void cancel() {    cancelable.cancel();    taskFuture.cancel(true);  }  /**   * Return true if the primary task has completed.   */  public boolean isDone() {    return taskFuture.isDone();  }}

---------------------------------------------------------------------------

本系列企业搜索引擎开发之连接器connector系本人原创

转载请注明出处 博客园 刺猬的温驯

本文链接http://www.cnblogs.com/chenying99/archive/2013/03/19/2968401.html