better handling of whitespace

git-svn-id: https://svn.berlios.de/svnroot/repos/yacy/trunk@2311 6c8d7289-2bf4-0310-a012-ef5d649a1542

source/de/anomic/htmlFilter/htmlFilterContentTransformer.java

@@ -143,7 +143,7 @@ public class htmlFilterContentTransformer extends htmlFilterAbstractTransformer
             serverByteBuffer[] sbbs = httpTemplate.splitQuotations(sbb);
             sbb = new serverByteBuffer();
             for (int i = 0; i < sbbs.length; i++) {
-                if (sbbs[i].isWhitespace()) {
+                if (sbbs[i].isWhitespace(true)) {
                     sbb.append(sbbs[i]);
                 } else if ((sbbs[i].byteAt(0) == httpTemplate.hash) ||
                            (sbbs[i].startsWith(httpTemplate.dpdpa))) {
@@ -151,8 +151,8 @@ public class htmlFilterContentTransformer extends htmlFilterAbstractTransformer
                     sbb.append(sbbs[i]);
                 } else {
                     // this is a text fragment, generate gettext quotation
-                    int ws = sbbs[i].whitespaceStart();
-                    int we = sbbs[i].whitespaceEnd();
+                    int ws = sbbs[i].whitespaceStart(true);
+                    int we = sbbs[i].whitespaceEnd(true);
                     sbb.append(sbbs[i].getBytes(0, ws));
                     sbb.append('_');
                     sbb.append('(');

source/de/anomic/server/serverByteBuffer.java

@@ -290,26 +290,44 @@ public final class serverByteBuffer extends OutputStream {
         return trim(l, r);
     }
 
-    public boolean isWhitespace() {
+    public boolean isWhitespace(boolean includeNonLetterBytes) {
         // returns true, if trim() would result in an empty serverByteBuffer
-        for (int i = 0; i < length; i++) {
-            if (buffer[offset + i] > 32) return false;
+        if (includeNonLetterBytes) {
+            byte b;
+            for (int i = 0; i < length; i++) {
+                b = buffer[offset + i];
+                if (((b >= '0') && (b <= '9')) || ((b >= 'A') && (b <= 'Z')) || ((b >= 'a') && (b <= 'z'))) return false;
+            }
+        } else {
+            for (int i = 0; i < length; i++) if (buffer[offset + i] > 32) return false;
         }
         return true;
     }
     
-    public int whitespaceStart() {
+    public int whitespaceStart(boolean includeNonLetterBytes) {
         // returns number of whitespace bytes at the beginning of text
-        for (int i = 0; i < length; i++) {
-            if (buffer[offset + i] > 32) return i;
+        if (includeNonLetterBytes) {
+            byte b;
+            for (int i = 0; i < length; i++) {
+                b = buffer[offset + i];
+                if (((b >= '0') && (b <= '9')) || ((b >= 'A') && (b <= 'Z')) || ((b >= 'a') && (b <= 'z'))) return i;
+            }
+        } else {
+            for (int i = 0; i < length; i++) if (buffer[offset + i] > 32) return i;
         }
         return length;
     }
     
-    public int whitespaceEnd() {
+    public int whitespaceEnd(boolean includeNonLetterBytes) {
         // returns position of whitespace at the end of text
-        for (int i = length - 1; i >= 0; i--) {
-            if (buffer[offset + i] > 32) return i + 1;
+        if (includeNonLetterBytes) {
+            byte b;
+            for (int i = length - 1; i >= 0; i--) {
+                b = buffer[offset + i];
+                if (((b >= '0') && (b <= '9')) || ((b >= 'A') && (b <= 'Z')) || ((b >= 'a') && (b <= 'z'))) return i + 1;
+            }
+        } else {
+            for (int i = length - 1; i >= 0; i--) if (buffer[offset + i] > 32) return i + 1;
         }
         return 0;
     }