Update

Author: macchiati

unicodetools/src/main/java/org/unicode/utilities/UrlUtilities.java

@@ -176,18 +176,23 @@ public String unescape(String substring) {
     static final UnicodeSet validHost =
             new UnicodeSet(IUP.getSet("Idn_Status=" + Idn_Status_Values.valid))
                     .addAll(IUP.getSet("Idn_Status=" + Idn_Status_Values.mapped))
-                    .add('.')
+                    .removeAll(new UnicodeSet("[:ascii:]"))
+                    .addAll(new UnicodeSet("[a-zA-Z0-9.]"))
                     .freeze();
 
     public static final class StringRange {
-        final int start;
-        final int limit;
+        public final int start;
+        public final int limit;
 
         public StringRange(int start, int limit) {
             super();
             this.start = start;
             this.limit = limit;
         }
+
+		public String substring(String source) {
+			return source.substring(start, limit);
+		}
     }
 
     /**
@@ -207,12 +212,12 @@ public static StringRange parseURL(String source, int startCodePointOffset) {
         int start = findStartMatcher.start();
         // dumb search for end of host, doesn't handle .. or edge cases, but this does not have to
         // be production-quality
-        int current = findStartMatcher.end();
-        int current2 = validHost.span(source, start, SpanCondition.CONTAINED);
-        if (current == current2) {
+        int protocolLimit = findStartMatcher.end();
+        int hostLimit = validHost.span(source, protocolLimit, SpanCondition.CONTAINED);
+        if (protocolLimit == hostLimit) {
             return null;
         }
-        int limit = parsePathQueryFragment(source, current2);
+        int limit = parsePathQueryFragment(source, hostLimit);
         return new StringRange(start, limit);
     }
 

unicodetools/src/test/java/org/unicode/unittest/TestUrl.java

@@ -33,6 +33,7 @@
 import org.unicode.utilities.UrlUtilities;
 import org.unicode.utilities.UrlUtilities.LinkTermination;
 import org.unicode.utilities.UrlUtilities.Part;
+import org.unicode.utilities.UrlUtilities.StringRange;
 
 /** The following is very temporary, just during the spec development. */
 public class TestUrl extends TestFmwk {
@@ -439,6 +440,15 @@ public void testOverlap() {
         }
     }
 
+    public void testParseUrl() {
+    	String source = "See http://example.com/foobar and http://a.us/foobar!";
+		StringRange position = UrlUtilities.parseURL(source, 0);
+		assertEquals(source, "http://example.com/foobar", position.substring(source));
+		
+		position = UrlUtilities.parseURL(source, position.limit);
+		assertEquals(source, "http://a.us/foobar", position.substring(source));
+
+    }
     //    private static final String SPLIT1 = "\t"; // for debugging, "\n";
     //
     //    private static final boolean VERBOSE_ASSERT = false;