*) Fixing Problems with MalformedURLs during Word Selection

- removing (lurl.toString() == null) comparison because toString() is never null
   - adding (lurl.url() == null) condition because url() is null if we have selected a word entry with
     a malformed URL

git-svn-id: https://svn.berlios.de/svnroot/repos/yacy/trunk@1083 6c8d7289-2bf4-0310-a012-ef5d649a1542

source/de/anomic/plasma/plasmaWordIndexDistribution.java

@@ -327,21 +327,18 @@ public final class plasmaWordIndexDistribution {
                         while (urlIter.hasNext()) {
                             indexEntry = (plasmaWordIndexEntry) urlIter.next();                            
                             lurl = this.urlPool.loadedURL.getEntry(indexEntry.getUrlHash());
-                            if ((lurl == null) || (lurl.toString() == null)) {
-                                unknownURLEntries.add(indexEntry.getUrlHash());
-                            } else {
-                                if (lurl.toString() == null) {
-                                    this.urlPool.loadedURL.remove(indexEntry.getUrlHash());
+                            if ((lurl == null) || (lurl.url() == null)) {
                                 unknownURLEntries.add(indexEntry.getUrlHash());
                             } else {
                                 knownURLs.put(indexEntry.getUrlHash(), lurl);
                             }
                         }
-                        }
                         // now delete all entries that have no url entry
                         hashIter = unknownURLEntries.iterator();
                         while (hashIter.hasNext()) {
-                            indexEntity.removeEntry((String) hashIter.next(), false);
+                            String nextUrlHash = (String) hashIter.next();
+                            indexEntity.removeEntry(nextUrlHash, false);
+                            this.urlPool.loadedURL.remove(nextUrlHash);
                         }
                         
                         if (indexEntity.size() == 0) {
@@ -366,11 +363,7 @@ public final class plasmaWordIndexDistribution {
                         while ((urlIter.hasNext()) && (count > 0)) {
                             indexEntry = (plasmaWordIndexEntry) urlIter.next();
                             lurl = this.urlPool.loadedURL.getEntry(indexEntry.getUrlHash());
-                            if (lurl == null) {
-                                unknownURLEntries.add(indexEntry.getUrlHash());
-                            } else {
-                                if (lurl.toString() == null) {
-                                    this.urlPool.loadedURL.remove(indexEntry.getUrlHash());
+                            if ((lurl == null) || (lurl.url()==null)) {
                                 unknownURLEntries.add(indexEntry.getUrlHash());
                             } else {
                                 knownURLs.put(indexEntry.getUrlHash(), lurl);
@@ -378,11 +371,12 @@ public final class plasmaWordIndexDistribution {
                                 count--;
                             }
                         }
-                        }
                         // now delete all entries that have no url entry
                         hashIter = unknownURLEntries.iterator();
                         while (hashIter.hasNext()) {
-                            indexEntity.removeEntry((String) hashIter.next(), true);
+                            String nextUrlHash = (String) hashIter.next();
+                            indexEntity.removeEntry(nextUrlHash, true);
+                            this.urlPool.loadedURL.remove(nextUrlHash);
                         }
                         // use whats remaining
                         this.log.logFine("Selected partial index (" + tmpEntity.size() + " from " + indexEntity.size() +" URLs, " + unknownURLEntries.size() + " not bound) for word " + tmpEntity.wordHash());