Block domain labels made of Cyrillic letters that look alike Latin

components/url_formatter/url_formatter_unittest.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 <stddef.h>
 #include <string.h>
 
 #include <vector>
@@ skipping 14 matching lines @@
 using base::ASCIIToUTF16;
 
 const size_t kNpos = base::string16::npos;
 
 struct IDNTestCase {
   const char* const input;
   const wchar_t* unicode_output;
   const bool unicode_allowed;
 };
 
+// TODO(jshin): Replace L"..." with "..." in UTF-8 when it's easier to read.
 const IDNTestCase idn_cases[] = {
   // No IDN
   {"www.google.com", L"www.google.com", true},
   {"www.google.com.", L"www.google.com.", true},
   {".", L".", true},
   {"", L"", true},
   // IDN
   // Hanzi (Traditional Chinese)
   {"xn--1lq90ic7f1rc.cn", L"\x5317\x4eac\x5927\x5b78.cn", true},
   // Hanzi ('video' in Simplified Chinese
@@ skipping 157 matching lines @@
   {"xn--weka801xo02a.com", L"\x30fc\x52d5\x753b\x30fc.com", false},
   // Han + U+30FC + Han
   {"xn--wekz60nb2ay85atj0b.jp", L"\x65e5\x672c\x30fc\x91ce\x7403.jp", false},
   // Latin + U+30FC + Latin
   {"xn--abcdef-r64e.jp", L"abc\x30fc" L"def.jp", false},
   // Latin + U+30FB + Latin
   {"xn--abcdef-k64e.jp", L"abc\x30fb" L"def.jp", false},
   // U+30FB + Latin
   {"xn--abc-os4b.jp", L"\x30fb" L"abc.jp", false},
 
+  // Cyrillic labels made of Latin-look-alike Cyrillic letters.
+  // ѕсоре.com with ѕсоре in Cyrillic
+  {"xn--e1argc3h.com", L"\x0455\x0441\x043e\x0440\x0435.com", false},
+  // ѕсоре123.com with ѕсоре in Cyrillic.
+  {"xn--123-qdd8bmf3n.com",
+      L"\x0455\x0441\x043e\x0440\x0435" L"123.com", false},
+  // ѕсоре-рау.com with ѕсоре and рау in Cyrillic.
+  {"xn----8sbn9akccw8m.com",
+      L"\x0455\x0441\x043e\x0440\x0435-\x0440\x0430\x0443.com", false},
+  // ѕсоре·рау.com with scope and pay in Cyrillic and U+00B7 between them.
+  {"xn--uba29ona9akccw8m.com",
+      L"\x0455\x0441\x043e\x0440\x0435\u00b7\x0440\x0430\x0443.com", false},
+
+  // The same as above three, but in IDN TLD.
+  {"xn--e1argc3h.xn--p1ai",
+      L"\x0455\x0441\x043e\x0440\x0435.\x0440\x0444", true},
+  {"xn--123-qdd8bmf3n.xn--p1ai",
+      L"\x0455\x0441\x043e\x0440\x0435" L"123.\x0440\x0444", true},
+  {"xn--uba29ona9akccw8m.xn--p1ai",
+      L"\x0455\x0441\x043e\x0440\x0435\u00b7\x0440\x0430\x0443.\x0440\x0444",
+      true},
+
+  // ѕсоре-рау.한국 with ѕсоре and рау in Cyrillic.
+  {"xn----8sbn9akccw8m.xn--3e0b707e",
+      L"\x0455\x0441\x043e\x0440\x0435-\x0440\x0430\x0443.\xd55c\xad6d", true},
+
+  // музей (museum in Russian) has characters without a Latin-look-alike.
+  {"xn--e1adhj9a.com", L"\x043c\x0443\x0437\x0435\x0439.com", true},
+
   // Mixed digits: the first two will also fail mixed script test
   // Latin + ASCII digit + Deva digit
   {"xn--asc1deva-j0q.co.in", L"asc1deva\x0967.co.in", false},
   // Latin + Deva digit + Beng digit
   {"xn--devabeng-f0qu3f.co.in", L"deva\x0967" L"beng\x09e7.co.in", false},
   // ASCII digit + Deva digit
   {"xn--79-v5f.co.in", L"7\x09ea" L"9.co.in", false},
   //  Deva digit + Beng digit
   {"xn--e4b0x.co.in", L"\x0967\x09e7.co.in", false},
   // U+4E00 (CJK Ideograph One) is not a digit
@@ skipping 699 matching lines @@
     0, kNpos, kNpos, kNpos, kNpos, kNpos, kNpos, 0, kNpos, kNpos, kNpos, kNpos,
     0, 1, 2, 3, 4, 5, 6, 7
   };
   CheckAdjustedOffsets("http://user@foo.com/", kFormatUrlOmitAll,
                        net::UnescapeRule::NORMAL, omit_all_offsets);
 }
 
 }  // namespace
 
 }  // namespace url_formatter