IDN Display: Add whole-script-confusable characters for Greek

This CL adds a list of Greek characters that look like Latin
characters. A domain consisting only of these characters and not
having a .gr ccTLD or other Greek TLDs such as .ελ will be
converted to punycode, as is done for other
whole-script-confusables such as Cyrillic and Hebrew.

Bug: 722167
Change-Id: Id0157503c605a6a73eb49147c26b48d5995d1ff2
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1986209
Commit-Queue: Mustafa Emre Acer <meacer@chromium.org>
Reviewed-by: Joe DeBlasio <jdeblasio@chromium.org>
Cr-Commit-Position: refs/heads/master@{#728697}

components/url_formatter/spoof_checks/idn_spoof_checker.cc

@@ -9,6 +9,7 @@
 #include "base/strings/string_piece.h"
 #include "base/strings/string_split.h"
 #include "base/strings/string_util.h"
+#include "base/strings/utf_string_conversions.h"
 #include "base/threading/thread_local_storage.h"
 #include "build/build_config.h"
 #include "net/base/lookup_string_in_fixed_set.h"
@@ -209,12 +210,20 @@ IDNSpoofChecker::IDNSpoofChecker() {
        // (Potential set: [ሀሁሃሠሡሰሱሲስበቡቢተቱቲታነከኩኪካኬክዐዑዕዖዘዙዚዛዝዞጠጡጢጣጦፐፒꬁꬂꬅ])
        "[[:Ethi:]]",
        "[ሀሠሰስበነተከዐዕዘጠፐꬅ]",
-       {"er", "et"}}};
+       {"er", "et"}},
+      {// Greek
+       "[[:Grek:]]",
+       // This ignores variants such as ά, έ, ή, ί.
+       "[αικνρυωηοτ]",
+       {"gr"}},
+  };
   for (const WholeScriptConfusableData& data : kWholeScriptConfusables) {
     auto all_letters = std::make_unique<icu::UnicodeSet>(
         icu::UnicodeString::fromUTF8(data.script_regex), status);
+    DCHECK(U_SUCCESS(status));
     auto latin_lookalikes = std::make_unique<icu::UnicodeSet>(
         icu::UnicodeString::fromUTF8(data.latin_lookalike_letters), status);
+    DCHECK(U_SUCCESS(status));
     auto script = std::make_unique<WholeScriptConfusable>(
         std::move(all_letters), std::move(latin_lookalikes), data.allowed_tlds);
     wholescriptconfusables_.push_back(std::move(script));

components/url_formatter/spoof_checks/idn_spoof_checker.h

@@ -100,7 +100,9 @@ class IDNSpoofChecker {
         const std::vector<std::string>& allowed_tlds);
     ~WholeScriptConfusable();
 
-    // Captures all letters belonging to this script.
+    // Captures all letters belonging to this script. See kScriptNameCodeList in
+    // blink/renderer/platform/text/locale_to_script_mapping.cc for script
+    // codes.
     std::unique_ptr<icu::UnicodeSet> all_letters;
     // The subset of all_letters that look like Latin ASCII letters. A domain
     // label entirely made of them is blocked as a simplified

components/url_formatter/spoof_checks/idn_spoof_checker_unittest.cc

@@ -1207,6 +1207,11 @@ const IDNTestCase kIdnCases[] = {
     // Whole-script-confusable in Ethiopic.
     {"xn--6xd66aa62c.com", L"ሠዐዐፐ.com", kUnsafe},
     {"xn--6xd66aa62c.et", L"ሠዐዐፐ.et", kSafe},
+
+    // Whole-script-confusable in Greek.
+    {"xn--mxapd.com", L"ικα.com", kUnsafe},
+    {"xn--mxapd.gr", L"ικα.gr", kSafe},
+    {"xn--mxapd.xn--qxam", L"ικα.ελ", kSafe},
 };
 
 namespace test {