looking for uppercase W as well and not accepting alphanumeric before www.

Author: MTDdk

src/main/java/org/nibor/autolink/LinkExtractor.java

@@ -50,6 +50,7 @@ private Scanner trigger(char c) {
             case '@':
                 return emailScanner;
             case 'w':
+            case 'W':
                 return wwwScanner;
         }
         return null;

src/main/java/org/nibor/autolink/internal/Scanners.java

@@ -18,7 +18,7 @@ public static boolean isNonAscii(char c) {
         return c >= 0x80;
     }
 
-    public static int findUrlEnd(CharSequence input, int beginIndex) {
+    public static final int findUrlEnd(CharSequence input, int beginIndex) {
         int round = 0;
         int square = 0;
         int curly = 0;

src/main/java/org/nibor/autolink/internal/WwwUrlScanner.java

@@ -12,12 +12,16 @@ public class WwwUrlScanner implements Scanner {
 
     @Override
     public LinkSpan scan(final CharSequence input, int triggerIndex, int rewindIndex) {
-        int afterDot = triggerIndex + 4;
-        if (afterDot >= input.length() || input.charAt(triggerIndex + 1) != 'w' || input.charAt(triggerIndex + 2) != 'w' || input.charAt(triggerIndex + 3) != '.') {
+        final int afterDot = triggerIndex + 4;
+        if (afterDot >= input.length() || !isWWW(input, triggerIndex)) {
             return null;
         }
 
-        int first = triggerIndex;
+        final int first = findFirst(input, triggerIndex, rewindIndex);
+        if (first == -1) {
+            return null;
+        }
+        
         int last = findLast(input, afterDot);
         if (last == -1) {
             return null;
@@ -26,8 +30,21 @@ public LinkSpan scan(final CharSequence input, int triggerIndex, int rewindIndex
         return new LinkSpanImpl(LinkType.WWW, first, last + 1);
     }
 
-    private int findLast(CharSequence input, int beginIndex) {
-        int last = Scanners.findUrlEnd(input, beginIndex);
+    private static final int findFirst(final CharSequence input, final int beginIndex, final int rewindIndex) {
+        if (beginIndex == rewindIndex) {
+            return beginIndex;
+        }
+        
+        // Is the character before www. allowed?
+        if (isAllowed(input.charAt(beginIndex - 1))) {
+            return beginIndex;
+        }
+        
+        return -1;
+    }
+    
+    private static final int findLast(final CharSequence input, final int beginIndex) {
+        final int last = Scanners.findUrlEnd(input, beginIndex);
 
         // Make sure there is at least one dot after the first dot,
         // so www.something is not allowed, but www.something.co.uk is
@@ -38,4 +55,15 @@ private int findLast(CharSequence input, int beginIndex) {
 
         return -1;
     }
+    
+    private static final boolean isAllowed(char c) {
+        return c != '.' && !Scanners.isAlnum(c);
+    }
+    
+    private static final boolean isWWW(final CharSequence input, final int triggerIndex) {
+        return 
+                (input.charAt(triggerIndex + 1) == 'w' || input.charAt(triggerIndex + 1) == 'W')
+             && (input.charAt(triggerIndex + 2) == 'w' || input.charAt(triggerIndex + 2) == 'W')
+             &&  input.charAt(triggerIndex + 3) == '.';
+    }
 }

src/test/java/org/nibor/autolink/AutolinkWwwUrlTest.java

@@ -34,12 +34,19 @@ public void notLinked() {
         assertNotLinked("w.bar.foo.co");
         assertNotLinked("www.something");
         assertNotLinked("www.go");
+        assertNotLinked("foo.www.fo.uk");
+        assertNotLinked("www..com");
+        assertNotLinked("wwww.toomany.com");
     }
 
     @Test
     public void linked() {
         assertLinked("www.s.com","|www.s.com|");
         assertLinked("www.fo.uk","|www.fo.uk|");
+        assertLinked("foo:www.fo.uk","foo:|www.fo.uk|");
+        assertLinked("foo-www.fo.uk","foo-|www.fo.uk|");
+        assertLinked("WWW.s.com","|WWW.s.com|");
+        assertLinked("Www.s.com","|Www.s.com|");
     }
 
     @Test