Merge branch 'develop' of github.com:code4craft/webmagic into develop

webmagic-core/src/main/java/us/codecraft/webmagic/Spider.java

@@ -106,7 +106,7 @@ public class Spider implements Runnable, Task {
 
     private Date startTime;
 
-    private int emptySleepTime = 30000;
+    private long emptySleepTime = 30000;
 
     /**
      * create a spider with pageProcessor.
@@ -306,15 +306,36 @@ public class Spider implements Runnable, Task {
         checkRunningStat();
         initComponent();
         logger.info("Spider {} started!", getUUID());
+        // interrupt won't be necessarily detected
         while (!Thread.currentThread().isInterrupted() && stat.get() == STAT_RUNNING) {
-            final Request request = scheduler.poll(this);
+            Request poll = scheduler.poll(this);
-            if (request == null) {
+            if (poll == null) {
-                if (threadPool.getThreadAlive() == 0 && exitWhenComplete) {
+                if (threadPool.getThreadAlive() == 0) {
+                    //no alive thread anymore , try again
+                    poll = scheduler.poll(this);
+                    if (poll == null) {
+                        if (exitWhenComplete) {
+                            break;
+                        } else {
+                            // wait
+                            try {
+                                Thread.sleep(emptySleepTime);
+                                continue;
+                            } catch (InterruptedException e) {
                                 break;
                             }
-                // wait until new url added
+                        }
-                waitNewUrl();
+                    }
                 } else {
+                    // wait until new url added，
+                    if (waitNewUrl())
+                        //if interrupted
+                        break;
+                    continue;
+                }
+            }
+            final Request request = poll;
+            //this may swallow the interruption
             threadPool.execute(new Runnable() {
                 @Override
                 public void run() {
@@ -331,7 +352,6 @@ public class Spider implements Runnable, Task {
                 }
             });
         }
-        }
         stat.set(STAT_STOPPED);
         // release some resources
         if (destroyWhenExit) {
@@ -565,16 +585,24 @@ public class Spider implements Runnable, Task {
         return this;
     }
 
-    private void waitNewUrl() {
+    /**
+     *
+     * @return isInterrupted
+     */
+    private boolean waitNewUrl() {
+        // now there may not be any thread live
         newUrlLock.lock();
         try {
-            //double check
+            //double check，unnecessary, unless very fast concurrent
-            if (threadPool.getThreadAlive() == 0 && exitWhenComplete) {
+            if (threadPool.getThreadAlive() == 0) {
-                return;
+                return false;
             }
+            //wait for amount of time
             newUrlCondition.await(emptySleepTime, TimeUnit.MILLISECONDS);
+            return false;
         } catch (InterruptedException e) {
-            logger.warn("waitNewUrl - interrupted, error {}", e);
+            // logger.warn("waitNewUrl - interrupted, error {}", e);
+            return true;
         } finally {
             newUrlLock.unlock();
         }
@@ -772,7 +800,10 @@ public class Spider implements Runnable, Task {
      *
      * @param emptySleepTime In MILLISECONDS.
      */
-    public void setEmptySleepTime(int emptySleepTime) {
+    public void setEmptySleepTime(long emptySleepTime) {
+        if(emptySleepTime<=0){
+            throw new IllegalArgumentException("emptySleepTime should be more than zero!");
+        }
         this.emptySleepTime = emptySleepTime;
     }
 }