Add a few more confusable map entries

1. Map Malaylam U+0D1F to 's'.
2. Map 'small-cap-like' Cyrillic letters to "look-alike" Latin lowercase
letters.

The characters in new confusable map entries are replaced by their Latin
"look-alike" characters before the skeleton is calculated to compare with
top domain names.

Bug: 784761,773930
Test: components_unittests --gtest_filter=*IDNToUni*
Change-Id: Ib26664e21ac5eb290e4a2993b01cbf0edaade0ee
Reviewed-on: https://chromium-review.googlesource.com/805214
Reviewed-by: Peter Kasting <pkasting@chromium.org>
Commit-Queue: Jungshik Shin <jshin@chromium.org>
Cr-Commit-Position: refs/heads/master@{#521648}

components/url_formatter/idn_spoof_checker.cc

@@ -133,15 +133,31 @@ IDNSpoofChecker::IDNSpoofChecker() {
 
   // Used for diacritics-removal before the skeleton calculation. Add
   // "ł > l; ø > o; đ > d" that are not handled by "NFD; Nonspacing mark
-  // removal; NFC". On top of that, supplement the Unicode confusable list by
-  // replacing {U+043A (к), U+0138(ĸ), U+03BA(κ)}, U+04CF (ӏ) and U+043F(п) by
-  // 'k', 'l' and 'n', respectively.
+  // removal; NFC".
   // TODO(jshin): Revisit "ł > l; ø > o" mapping.
   UParseError parse_error;
-  transliterator_.reset(icu::Transliterator::createFromRules(
+  diacritic_remover_.reset(icu::Transliterator::createFromRules(
       UNICODE_STRING_SIMPLE("DropAcc"),
       icu::UnicodeString("::NFD; ::[:Nonspacing Mark:] Remove; ::NFC;"
-                         " ł > l; ø > o; đ > d; ӏ > l; [кĸκ] > k; п > n;"),
+                         " ł > l; ø > o; đ > d;"),
+      UTRANS_FORWARD, parse_error, status));
+
+  // Supplement the Unicode confusable list by the following mapping.
+  //   - U+04CF (ӏ) => l
+  //   - {U+043A (к), U+0138(ĸ), U+03BA(κ)} => k
+  //   - U+043F(п) => n
+  //   - {U+0185 (ƅ), U+044C (ь)} => b
+  //   - U+0432 (в) => b
+  //   - U+043C (м) => m
+  //   - U+043D (н) => h
+  //   - U+0442 (т) => t
+  //   - {U+0448 (ш), U+0449 (щ)} => w
+  //   - U+0D1F (ട) => s
+  extra_confusable_mapper_.reset(icu::Transliterator::createFromRules(
+      UNICODE_STRING_SIMPLE("ExtraConf"),
+      icu::UnicodeString(
+          "ӏ > l; [кĸκ] > k; п > n; [ƅь] > b; в > b; м > m; н > h; "
+          "т > t; [шщ] > w; ട > s;"),
       UTRANS_FORWARD, parse_error, status));
   DCHECK(U_SUCCESS(status))
       << "Spoofchecker initalization failed due to an error: "
@@ -270,7 +286,8 @@ bool IDNSpoofChecker::SimilarToTopDomains(base::StringPiece16 hostname) {
   // attached to non-LGC characters are already blocked.
   if (lgc_letters_n_ascii_.span(ustr_host, 0, USET_SPAN_CONTAINED) ==
       ustr_host.length())
-    transliterator_.get()->transliterate(ustr_host);
+    diacritic_remover_.get()->transliterate(ustr_host);
+  extra_confusable_mapper_.get()->transliterate(ustr_host);
 
   UErrorCode status = U_ZERO_ERROR;
   icu::UnicodeString ustr_skeleton;
@@ -279,8 +296,7 @@ bool IDNSpoofChecker::SimilarToTopDomains(base::StringPiece16 hostname) {
   if (U_FAILURE(status))
     return false;
   std::string skeleton;
-  ustr_skeleton.toUTF8String(skeleton);
-  return LookupMatchInTopDomains(skeleton);
+  return LookupMatchInTopDomains(ustr_skeleton.toUTF8String(skeleton));
 }
 
 bool IDNSpoofChecker::IsMadeOfLatinAlikeCyrillic(

components/url_formatter/idn_spoof_checker.h

@@ -67,7 +67,8 @@ class IDNSpoofChecker {
   icu::UnicodeSet cyrillic_letters_;
   icu::UnicodeSet cyrillic_letters_latin_alike_;
   icu::UnicodeSet lgc_letters_n_ascii_;
-  std::unique_ptr<icu::Transliterator> transliterator_;
+  std::unique_ptr<icu::Transliterator> diacritic_remover_;
+  std::unique_ptr<icu::Transliterator> extra_confusable_mapper_;
 
   IDNSpoofChecker(const IDNSpoofChecker&) = delete;
   void operator=(const IDNSpoofChecker&) = delete;

components/url_formatter/top_domains/alexa_domains.list

@@ -9176,3 +9176,6 @@ stripe.com
 digklmo68.com
 digklmo68.co.uk
 islkpx123.com
+os345.com
+woder.com
+wmhtb.com

components/url_formatter/top_domains/alexa_skeletons.gperf

@@ -9185,4 +9185,7 @@ stripe.corn, 1
 digklrno68.corn, 1
 digklrno68.co.uk, 1
 islkpxl23.corn, 1
+os345.corn, 1
+woder.corn, 1
+wrnhtb.corn, 1
 %%

components/url_formatter/top_domains/make_alexa_top_list.py

@@ -53,3 +53,4 @@
   # Add a few made-up domains for testing.
   outfile.write("# for testing\ndigklmo68.com\ndigklmo68.co.uk\n")
   outfile.write("islkpx123.com\n")
+  outfile.write("os345.com\nwoder.com\nwmhtb.com\n")

components/url_formatter/url_formatter_unittest.cc

@@ -406,6 +406,19 @@ const IDNTestCase idn_cases[] = {
      L"123.com",
      false},
 
+    // wmhtb.com
+    {"xn--l1acpvx.com", L"\x0448\x043c\x043d\x0442\x044c.com", false},
+    // щмнть.com
+    {"xn--l1acpzs.com", L"\x0449\x043c\x043d\x0442\x044c.com", false},
+    // шмнтв.com
+    {"xn--b1atdu1a.com", L"\x0448\x043c\x043d\x0442\x0432.com", false},
+    // ഠട345.com
+    {"xn--345-jtke.com", L"\x0d20\x0d1f" L"345.com", false},
+
+    // At one point the skeleton of 'w' was 'vv', ensure that
+    // that it's treated as 'w'.
+    {"xn--wder-qqa.com", L"w\x00f3" L"der.com", false},
+
     // 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},
@@ -696,9 +709,6 @@ const IDNTestCase idn_cases[] = {
      true},
     // Can start with a RTL and end with AN
     {"xn--mgbjq0r.eg", L"\x062c\x0627\x0631\x0662.eg", true},
-    // At one point the skeleton of 'w' was 'vv', ensure that
-    // that it's treated as 'w'.
-    {"xn--wnderlist-58a.com", L"w\x00fanderlist.com", false},
 
     // Extremely rare Latin letters
     // Latin Ext B - Pinyin: ǔnion.com