Port and integrate the irregexp engine from V8

tests/language/regex/non-bmp_test.dart

Index: tests/language/regex/non-bmp_test.dart
diff --git a/tests/language/regex/non-bmp_test.dart b/tests/language/regex/non-bmp_test.dart
new file mode 100644
index 0000000000000000000000000000000000000000..37f3464f628ac3f7ee07d459bd5a528eb8122a46
--- /dev/null
+++ b/tests/language/regex/non-bmp_test.dart
@@ -0,0 +1,26 @@
+// Copyright (c) 2014, the Dart project authors.  Please see the AUTHORS file
+// for details. All rights reserved. Use of this source code is governed by a
+// BSD-style license that can be found in the LICENSE file.
+
+import 'util.dart';
+import 'package:expect/expect.dart';
+
+void main() {
+  description(
+  'Tests that regular expressions treat non-BMP characters as two separate characters. '
+  + 'From a Unicode correctness point of view this is wrong, but it is what other browsers do. '
+  + 'And given that we store strings as UTF-16, it is also more efficient to implement. '
+  + 'Also test some other cases related to UTF-8 and UTF-16.'
+  );
+
+  var surrogatePair = new String.fromCharCode(0xD800) + new String.fromCharCode(0xDC00);
+
+  assertEquals(new RegExp(r".").firstMatch(surrogatePair).group(0).length, 1);
+  assertEquals(new RegExp(r"\D").firstMatch(surrogatePair).group(0).length, 1);
+  assertEquals(new RegExp(r"\S").firstMatch(surrogatePair).group(0).length, 1);
+  assertEquals(new RegExp(r"\W").firstMatch(surrogatePair).group(0).length, 1);
+  assertEquals(new RegExp(r"[^x]").firstMatch(surrogatePair).group(0).length, 1);
+
+  assertEquals(new RegExp(r".{1,2}").firstMatch("!!" + new String.fromCharCode(0xA1)).group(0).length, 2);
+  shouldBeNull(new RegExp(r".").firstMatch(""));
+}