Tweaks handling of U+30F[C-E]

components/url_formatter/url_formatter.cc

 // Copyright 2015 The Chromium Authors. All rights reserved.
 // Use of this source code is governed by a BSD-style license that can be
 // found in the LICENSE file.
 
 #include "components/url_formatter/url_formatter.h"
 
 #include <algorithm>
 #include <utility>
 
 #include "base/lazy_instance.h"
@@ skipping 306 matching lines @@
   deviation_characters_.freeze();
 
   // Latin letters outside ASCII. 'Script_Extensions=Latin' is not necessary
   // because additional characters pulled in with scx=Latn are not included in
   // the allowed set.
   non_ascii_latin_letters_ = icu::UnicodeSet(
       UNICODE_STRING_SIMPLE("[[:Latin:] - [a-zA-Z]]"), status);
   non_ascii_latin_letters_.freeze();
 
   // These letters are parts of |dangerous_patterns_|.
-  kana_letters_exceptions_ = icu::UnicodeSet(UNICODE_STRING_SIMPLE(
-      "[\\u3078-\\u307a\\u30d8-\\u30da\\u30fb\\u30fc]"), status);
+  kana_letters_exceptions_ = icu::UnicodeSet(
+      UNICODE_STRING_SIMPLE("[\\u3078-\\u307a\\u30d8-\\u30da\\u30fb-\\u30fe]"),
+      status);
   kana_letters_exceptions_.freeze();
 
   // These Cyrillic letters look like Latin. A domain label entirely made of
   // these letters is blocked as a simplified whole-script-spoofable.
   cyrillic_letters_latin_alike_ =
       icu::UnicodeSet(icu::UnicodeString("[асԁеһіјӏорԛѕԝхуъЬҽпгѵѡ]"), status);
   cyrillic_letters_latin_alike_.freeze();
 
   cyrillic_letters_ =
       icu::UnicodeSet(UNICODE_STRING_SIMPLE("[[:Cyrl:]]"), status);
@@ skipping 60 matching lines @@
     // non-Japanese script on either side is disallowed, legitimate cases like
     // '{vitamin in Katakana}b6' are blocked. Note that trying to block those
     // characters when used alone as a label is futile because those cases
     // would not reach here.
     // Also disallow what used to be blocked by mixed-script-confusable (MSC)
     // detection. ICU 58 does not detect MSC any more for a single input string.
     // See http://bugs.icu-project.org/trac/ticket/12823 .
     // TODO(jshin): adjust the pattern once the above ICU bug is fixed.
     // - Disallow U+30FB (Katakana Middle Dot) and U+30FC (Hiragana-Katakana
     //   Prolonged Sound) used out-of-context.
+    // - Dislallow U+30FD/E (Katakana iteration mark/voiced iteration mark)
+    //   unless they're preceded by a Katakana.
     // - Disallow three Hiragana letters (U+307[8-A]) or Katakana letters
     //   (U+30D[8-A]) that look exactly like each other when they're used in a
     //   label otherwise entirely in Katakna or Hiragana.
     // - Disallow U+0585 (Armenian Small Letter Oh) and U+0581 (Armenian Small
     //   Letter Co) to be next to Latin.
     // - Disallow Latin 'o' and 'g' next to Armenian.
     dangerous_pattern = new icu::RegexMatcher(
         icu::UnicodeString(
             "[^\\p{scx=kana}\\p{scx=hira}\\p{scx=hani}]"
             "[\\u30ce\\u30f3\\u30bd\\u30be]"
             "[^\\p{scx=kana}\\p{scx=hira}\\p{scx=hani}]|"
-            "[^\\p{scx=kana}\\p{scx=hira}]\\u30fc|"
-            "\\u30fc[^\\p{scx=kana}\\p{scx=hira}]|"
+            "[^\\p{scx=kana}\\p{scx=hira}]\\u30fc|^\\u30fc|"
+            "[^\\p{scx=kana}][\\u30fd\\u30fe]|^[\\u30fd\\u30fe]|"
             "^[\\p{scx=kana}]+[\\u3078-\\u307a][\\p{scx=kana}]+$|"
             "^[\\p{scx=hira}]+[\\u30d8-\\u30da][\\p{scx=hira}]+$|"
             "[a-z]\\u30fb|\\u30fb[a-z]|"
             "^[\\u0585\\u0581]+[a-z]|[a-z][\\u0585\\u0581]+$|"
             "[a-z][\\u0585\\u0581]+[a-z]|"
             "^[og]+[\\p{scx=armn}]|[\\p{scx=armn}][og]+$|"
-            "[\\p{scx=armn}][og]+[\\p{scx=armn}]", -1, US_INV),
+            "[\\p{scx=armn}][og]+[\\p{scx=armn}]",
+            -1, US_INV),
         0, status);
     tls_index.Set(dangerous_pattern);
   }
   dangerous_pattern->reset(label_string);
   return !dangerous_pattern->find();
 }
 
 bool IDNSpoofChecker::IsMadeOfLatinAlikeCyrillic(
     const icu::UnicodeString& label_string) {
   // Collect all the Cyrillic letters in |label_string| and see if they're
@@ skipping 407 matching lines @@
   return base::StartsWith(text, www, base::CompareCase::SENSITIVE)
       ? text.substr(www.length()) : text;
 }
 
 base::string16 StripWWWFromHost(const GURL& url) {
   DCHECK(url.is_valid());
   return StripWWW(base::ASCIIToUTF16(url.host_piece()));
 }
 
 }  // namespace url_formatter