Don't unescape BiDi control characters in URL components

net/base/escape_unittest.cc

 // Copyright (c) 2012 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 <algorithm>
 #include <string>
 
 #include "net/base/escape.h"
 
 #include "base/basictypes.h"
@@ skipping 200 matching lines @@
   const UnescapeURLCase unescape_cases[] = {
     {L"", UnescapeRule::NORMAL, L""},
     {L"%2", UnescapeRule::NORMAL, L"%2"},
     {L"%%%%%%", UnescapeRule::NORMAL, L"%%%%%%"},
     {L"Don't escape anything", UnescapeRule::NORMAL, L"Don't escape anything"},
     {L"Invalid %escape %2", UnescapeRule::NORMAL, L"Invalid %escape %2"},
     {L"Some%20random text %25%2dOK", UnescapeRule::NONE,
      L"Some%20random text %25%2dOK"},
     {L"Some%20random text %25%2dOK", UnescapeRule::NORMAL,
      L"Some%20random text %25-OK"},
+    {L"Some%20random text %25%E2%80", UnescapeRule::NORMAL,
+     L"Some%20random text %25\xE2\x80"},
+    {L"Some%20random text %25%E2%80OK", UnescapeRule::NORMAL,
+     L"Some%20random text %25\xE2\x80OK"},
+    {L"Some%20random text %25%E2%80%84OK", UnescapeRule::NORMAL,
+     L"Some%20random text %25\xE2\x80\x84OK"},
+
+    // BiDi Control characters should not be unescaped.
+    {L"Some%20random text %25%D8%9COK", UnescapeRule::NORMAL,
+     L"Some%20random text %25%D8%9COK"},
+    {L"Some%20random text %25%E2%80%8EOK", UnescapeRule::NORMAL,
+     L"Some%20random text %25%E2%80%8EOK"},
+    {L"Some%20random text %25%E2%80%8FOK", UnescapeRule::NORMAL,
+     L"Some%20random text %25%E2%80%8FOK"},
+    {L"Some%20random text %25%E2%80%AAOK", UnescapeRule::NORMAL,
+     L"Some%20random text %25%E2%80%AAOK"},
+    {L"Some%20random text %25%E2%80%ABOK", UnescapeRule::NORMAL,
+     L"Some%20random text %25%E2%80%ABOK"},
+    {L"Some%20random text %25%E2%80%AEOK", UnescapeRule::NORMAL,
+     L"Some%20random text %25%E2%80%AEOK"},
+    {L"Some%20random text %25%E2%81%A6OK", UnescapeRule::NORMAL,
+     L"Some%20random text %25%E2%81%A6OK"},
+    {L"Some%20random text %25%E2%81%A9OK", UnescapeRule::NORMAL,
+     L"Some%20random text %25%E2%81%A9OK"},
+
     {L"Some%20random text %25%2dOK", UnescapeRule::SPACES,
      L"Some random text %25-OK"},
     {L"Some%20random text %25%2dOK", UnescapeRule::URL_SPECIAL_CHARS,
      L"Some%20random text %-OK"},
     {L"Some%20random text %25%2dOK",
      UnescapeRule::SPACES | UnescapeRule::URL_SPECIAL_CHARS,
      L"Some random text %-OK"},
     {L"%A0%B1%C2%D3%E4%F5", UnescapeRule::NORMAL, L"\xA0\xB1\xC2\xD3\xE4\xF5"},
     {L"%Aa%Bb%Cc%Dd%Ee%Ff", UnescapeRule::NORMAL, L"\xAa\xBb\xCc\xDd\xEe\xFf"},
     // Certain URL-sensitive characters should not be unescaped unless asked.
@@ skipping 192 matching lines @@
                 internal::AdjustEncodingOffset(adjustments));
   size_t expected_2[] = {0, kNpos, kNpos, 1, 2, 3, 4, kNpos, kNpos, 5, kNpos,
                          kNpos, 6, 7, 8, 9, kNpos, kNpos};
   EXPECT_EQ(offsets.size(), arraysize(expected_2));
   for (size_t i = 0; i < arraysize(expected_2); ++i)
     EXPECT_EQ(expected_2[i], offsets[i]);
 }
 
 }  // namespace
 }  // namespace net