removed a crawler overhead (terminated loop which searches greatest

stack that has zero-waiting urls). This should cause a slightly faster
crawl for crawl stacks with many different domains in the crawl queue.

source/de/anomic/crawler/Balancer.java

@@ -458,7 +458,7 @@ public class Balancer {
     	byte[] besturlhash = null;
     	String besthost = null;
     	Map<String, byte[]> zeroWaitingCandidates = new HashMap<String, byte[]>();
-    	while (i.hasNext()) {
+    	while (i.hasNext() && zeroWaitingCandidates.size() < 10) {
             entry = i.next();
 
             // clean up empty entries