Use ICU case conversion/transliterator for case conversion

When I18N is enabled, use ICU's case conversion API and transliteration
API [1] to implement String.prototype.to{Upper,Lower}Case and
String.prototype.toLocale{Upper,Lower}Case.

* ICU-based case conversion was implemented in runtime-i18n.cc/i18n.js
* The above 4 functions are overridden with those in i18n.js when
  --icu_case_mapping flag is turned on. To control the override by the flag,
  they're overriden in icu-case-mapping.js

Previously, toLocale{U,L}Case just called to{U,L}Case so that they didn't
support locale-sensitive case conversion for Turkic languages (az, tr),
Greek (el) and Lithuanian (lt).

Before ICU APIs for the most general case are called, a fast-path for Latin-1
is tried. It's taken from Blink and adopted as necessary. This fast path
is always tried for to{U,L}Case. For toLocale{U,L}Case, it's only taken
when a locale (explicitly specified or default) is not in {az, el, lt, tr}.

With these changes, a build with --icu_case_mapping=true passes a bunch
of tests in test262/intl402/Strings/* and intl/* that failed before.

Handling of pure ASCII strings (aligned at word boundary) are not as fast
as Unibrow's implementation that uses word-by-word case conversion. OTOH,
Latin-1 input handling is faster than Unibrow. General Unicode input
handling is slower but more accurate.

See https://docs.google.com/spreadsheets/d/1KJCJxKc1FxFXjwmYqABS0_2cNdPetvnd8gY8_HGSbrg/edit?usp=sharing for the benchmark.

This CL started with http://crrev.com/1544023002#ps200001 by littledan@,
but has changed significantly since.

[1] See why transliteration API is needed for uppercasing in Greek.
    http://bugs.icu-project.org/trac/ticket/10582

R=yangguo
BUG=v8:4476,v8:4477
LOG=Y
TEST=test262/{built-ins,intl402}/Strings/*, webkit/fast/js/*, mjsunit/string-case,
     intl/general/case*

Review-Url: https://codereview.chromium.org/1812673005
Cr-Commit-Position: refs/heads/master@{#36187}

BUILD.gn

@@ -428,6 +428,10 @@ action("js2c_experimental") {
     "$target_gen_dir/experimental-libraries.cc",
   ]
 
+  if (v8_enable_i18n_support) {
+    sources += [ "src/js/icu-case-mapping.js" ]
+  }
+
   args = [
            rebase_path("$target_gen_dir/experimental-libraries.cc",
                        root_build_dir),

src/bootstrapper.cc

@@ -2478,6 +2478,9 @@ EMPTY_INITIALIZE_GLOBAL_FOR_FEATURE(harmony_instanceof)
 EMPTY_INITIALIZE_GLOBAL_FOR_FEATURE(harmony_restrictive_declarations)
 EMPTY_INITIALIZE_GLOBAL_FOR_FEATURE(harmony_exponentiation_operator)
 EMPTY_INITIALIZE_GLOBAL_FOR_FEATURE(harmony_string_padding)
+#ifdef V8_I18N_SUPPORT
+EMPTY_INITIALIZE_GLOBAL_FOR_FEATURE(icu_case_mapping)
+#endif
 
 void InstallPublicSymbol(Factory* factory, Handle<Context> native_context,
                          const char* name, Handle<Symbol> value) {
@@ -3046,6 +3049,10 @@ bool Genesis::InstallExperimentalNatives() {
   static const char* harmony_exponentiation_operator_natives[] = {nullptr};
   static const char* harmony_string_padding_natives[] = {
       "native harmony-string-padding.js", nullptr};
+#ifdef V8_I18N_SUPPORT
+  static const char* icu_case_mapping_natives[] = {"native icu-case-mapping.js",
+                                                   nullptr};
+#endif
 
   for (int i = ExperimentalNatives::GetDebuggerCount();
        i < ExperimentalNatives::GetBuiltinsCount(); i++) {

src/flag-definitions.h

@@ -193,12 +193,22 @@ DEFINE_IMPLICATION(es_staging, harmony_regexp_lookbehind)
 DEFINE_IMPLICATION(es_staging, move_object_start)
 
 // Features that are still work in progress (behind individual flags).
+#ifdef V8_I18N_SUPPORT
+#define HARMONY_INPROGRESS(V)                                           \
+  V(harmony_function_sent, "harmony function.sent")                     \
+  V(harmony_sharedarraybuffer, "harmony sharedarraybuffer")             \
+  V(harmony_simd, "harmony simd")                                       \
+  V(harmony_do_expressions, "harmony do-expressions")                   \
+  V(harmony_regexp_property, "harmony unicode regexp property classes") \
+  V(icu_case_mapping, "case mapping with ICU rather than Unibrow")
+#else
 #define HARMONY_INPROGRESS(V)                                  \
   V(harmony_function_sent, "harmony function.sent")            \
   V(harmony_sharedarraybuffer, "harmony sharedarraybuffer")    \
   V(harmony_simd, "harmony simd")                              \
   V(harmony_do_expressions, "harmony do-expressions")          \
   V(harmony_regexp_property, "harmony unicode regexp property classes")
+#endif
 
 // Features that are complete (but still behind --harmony/es-staging flag).
 #define HARMONY_STAGED(V)                                                    \

src/js/i18n.js

@@ -142,6 +142,13 @@ var AVAILABLE_LOCALES = {
  */
 var DEFAULT_ICU_LOCALE = UNDEFINED;
 
+function GetDefaultICULocaleJS() {
+  if (IS_UNDEFINED(DEFAULT_ICU_LOCALE)) {
+    DEFAULT_ICU_LOCALE = %GetDefaultICULocale();
+  }
+  return DEFAULT_ICU_LOCALE;
+}
+
 /**
  * Unicode extension regular expression.
  */
@@ -446,11 +453,7 @@ function lookupMatcher(service, requestedLocales) {
   }
 
   // Didn't find a match, return default.
-  if (IS_UNDEFINED(DEFAULT_ICU_LOCALE)) {
-    DEFAULT_ICU_LOCALE = %GetDefaultICULocale();
-  }
-
-  return {'locale': DEFAULT_ICU_LOCALE, 'extension': '', 'position': -1};
+  return {'locale': GetDefaultICULocaleJS(), 'extension': '', 'position': -1};
 }
 
 
@@ -722,21 +725,24 @@ function toTitleCaseTimezoneLocation(location) {
  */
 function canonicalizeLanguageTag(localeID) {
   // null is typeof 'object' so we have to do extra check.
-  if (typeof localeID !== 'string' && typeof localeID !== 'object' ||
+  if ((!IS_STRING(localeID) && !IS_RECEIVER(localeID)) ||
       IS_NULL(localeID)) {
     throw MakeTypeError(kLanguageID);
   }
 
+  // Optimize for the most common case; a language code alone in
+  // the canonical form/lowercase (e.g. "en", "fil").
+  if (IS_STRING(localeID) &&
+      !IS_NULL(InternalRegExpMatch(/^[a-z]{2,3}$/, localeID))) {
+    return localeID;
+  }
+
   var localeString = GlobalString(localeID);
 
   if (isValidLanguageTag(localeString) === false) {
     throw MakeRangeError(kInvalidLanguageTag, localeString);
   }
 
-  // This call will strip -kn but not -kn-true extensions.
-  // ICU bug filled - http://bugs.icu-project.org/trac/ticket/9265.
-  // TODO(cira): check if -u-kn-true-kc-true-kh-true still throws after
-  // upgrade to ICU 4.9.
   var tag = %CanonicalizeLanguageTag(localeString);
   if (tag === 'invalid-tag') {
     throw MakeRangeError(kInvalidLanguageTag, localeString);
@@ -1989,6 +1995,37 @@ function cachedOrNewService(service, locales, options, defaults) {
   return new savedObjects[service](locales, useOptions);
 }
 
+function LocaleConvertCase(s, locales, isToUpper) {
+  // ECMA 402 section 13.1.2 steps 1 through 12.
+  var language;
+  // Optimize for the most common two cases. initializeLocaleList() can handle
+  // them as well, but it's rather slow accounting for over 60% of
+  // toLocale{U,L}Case() and about 40% of toLocale{U,L}Case("<locale>").
+  if (IS_UNDEFINED(locales)) {
+    language = GetDefaultICULocaleJS();
+  } else if (IS_STRING(locales)) {
+    language = canonicalizeLanguageTag(locales);
+  } else {
+    var locales = initializeLocaleList(locales);
+    language = locales.length > 0 ? locales[0] : GetDefaultICULocaleJS();
+  }
+
+  // StringSplit is slower than this.
+  var pos = %_Call(StringIndexOf, language, '-');
+  if (pos != -1) {
+    language = %_Call(StringSubstring, language, 0, pos);
+  }
+
+  var CUSTOM_CASE_LANGUAGES = ['az', 'el', 'lt', 'tr'];
+  var langIndex = %_Call(ArrayIndexOf, CUSTOM_CASE_LANGUAGES, language);
+  if (langIndex == -1) {
+    // language-independent case conversion.
+    return isToUpper ? %StringToUpperCaseI18N(s) : %StringToLowerCaseI18N(s);
+  }
+  return %StringLocaleConvertCase(s, isToUpper,
+                                  CUSTOM_CASE_LANGUAGES[langIndex]);
+}
+
 /**
  * Compares this and that, and returns less than 0, 0 or greater than 0 value.
  * Overrides the built-in method.
@@ -2041,6 +2078,56 @@ OverrideFunction(GlobalString.prototype, 'normalize', function() {
   }
 );
 
+function ToLowerCaseI18N() {
+  if (!IS_UNDEFINED(new.target)) {
+    throw MakeTypeError(kOrdinaryFunctionCalledAsConstructor);
+  }
+  CHECK_OBJECT_COERCIBLE(this, "String.prototype.toLowerCase");
+  var s = TO_STRING(this);
+  return %StringToLowerCaseI18N(s);
+}
+
+function ToUpperCaseI18N() {
+  if (!IS_UNDEFINED(new.target)) {
+    throw MakeTypeError(kOrdinaryFunctionCalledAsConstructor);
+  }
+  CHECK_OBJECT_COERCIBLE(this, "String.prototype.toUpperCase");
+  var s = TO_STRING(this);
+  return %StringToUpperCaseI18N(s);
+}
+
+function ToLocaleLowerCaseI18N(locales) {
+  if (!IS_UNDEFINED(new.target)) {
+    throw MakeTypeError(kOrdinaryFunctionCalledAsConstructor);
+  }
+  CHECK_OBJECT_COERCIBLE(this, "String.prototype.toLocaleLowerCase");
+  return LocaleConvertCase(TO_STRING(this), locales, false);
+}
+
+%FunctionSetLength(ToLocaleLowerCaseI18N, 0);
+
+function ToLocaleUpperCaseI18N(locales) {
+  if (!IS_UNDEFINED(new.target)) {
+    throw MakeTypeError(kOrdinaryFunctionCalledAsConstructor);
+  }
+  CHECK_OBJECT_COERCIBLE(this, "String.prototype.toLocaleUpperCase");
+  return LocaleConvertCase(TO_STRING(this), locales, true);
+}
+
+%FunctionSetLength(ToLocaleUpperCaseI18N, 0);
+
+%FunctionRemovePrototype(ToLowerCaseI18N);
+%FunctionRemovePrototype(ToUpperCaseI18N);
+%FunctionRemovePrototype(ToLocaleLowerCaseI18N);
+%FunctionRemovePrototype(ToLocaleUpperCaseI18N);
+
+utils.Export(function(to) {
+  to.ToLowerCaseI18N = ToLowerCaseI18N;
+  to.ToUpperCaseI18N = ToUpperCaseI18N;
+  to.ToLocaleLowerCaseI18N = ToLocaleLowerCaseI18N;
+  to.ToLocaleUpperCaseI18N = ToLocaleUpperCaseI18N;
+});
+
 
 /**
  * Formats a Number object (this) using locale and options values.

src/js/icu-case-mapping.js

+// Copyright 2016 the V8 project authors. All rights reserved.
+// Use of this source code is governed by a BSD-style license that can be
+// found in the LICENSE file.
+
+(function(global, utils) {
+"use strict";
+
+%CheckIsBootstrapping();
+
+var GlobalString = global.String;
+var OverrideFunction = utils.OverrideFunction;
+var ToLowerCaseI18N = utils.ImportNow("ToLowerCaseI18N");
+var ToUpperCaseI18N = utils.ImportNow("ToUpperCaseI18N");
+var ToLocaleLowerCaseI18N = utils.ImportNow("ToLocaleLowerCaseI18N");
+var ToLocaleUpperCaseI18N = utils.ImportNow("ToLocaleUpperCaseI18N");
+
+OverrideFunction(GlobalString.prototype, 'toLowerCase', ToLowerCaseI18N, true);
+OverrideFunction(GlobalString.prototype, 'toUpperCase', ToUpperCaseI18N, true);
+OverrideFunction(GlobalString.prototype, 'toLocaleLowerCase',
+                 ToLocaleLowerCaseI18N, true);
+OverrideFunction(GlobalString.prototype, 'toLocaleUpperCase',
+                 ToLocaleUpperCaseI18N, true);
+
+})

src/js/prologue.js

@@ -208,7 +208,11 @@ function PostNatives(utils) {
     "SetIteratorNext",
     "SetValues",
     "SymbolToString",
+    "ToLocaleLowerCaseI18N",
+    "ToLocaleUpperCaseI18N",
+    "ToLowerCaseI18N",
     "ToPositiveInteger",
+    "ToUpperCaseI18N",
     // From runtime:
     "is_concat_spreadable_symbol",
     "iterator_symbol",

src/objects.h

@@ -8645,26 +8645,26 @@ class String: public Name {
   class FlatContent {
    public:
     // Returns true if the string is flat and this structure contains content.
-    bool IsFlat() { return state_ != NON_FLAT; }
+    bool IsFlat() const { return state_ != NON_FLAT; }
     // Returns true if the structure contains one-byte content.
-    bool IsOneByte() { return state_ == ONE_BYTE; }
+    bool IsOneByte() const { return state_ == ONE_BYTE; }
     // Returns true if the structure contains two-byte content.
-    bool IsTwoByte() { return state_ == TWO_BYTE; }
+    bool IsTwoByte() const { return state_ == TWO_BYTE; }
 
     // Return the one byte content of the string. Only use if IsOneByte()
     // returns true.
-    Vector<const uint8_t> ToOneByteVector() {
+    Vector<const uint8_t> ToOneByteVector() const {
       DCHECK_EQ(ONE_BYTE, state_);
       return Vector<const uint8_t>(onebyte_start, length_);
     }
     // Return the two-byte content of the string. Only use if IsTwoByte()
     // returns true.
-    Vector<const uc16> ToUC16Vector() {
+    Vector<const uc16> ToUC16Vector() const {
       DCHECK_EQ(TWO_BYTE, state_);
       return Vector<const uc16>(twobyte_start, length_);
     }
 
-    uc16 Get(int i) {
+    uc16 Get(int i) const {
       DCHECK(i < length_);
       DCHECK(state_ != NON_FLAT);
       if (state_ == ONE_BYTE) return onebyte_start[i];

src/runtime/runtime-strings.cc

@@ -1077,7 +1077,7 @@ MUST_USE_RESULT static Object* ConvertCase(
 
 RUNTIME_FUNCTION(Runtime_StringToLowerCase) {
   HandleScope scope(isolate);
-  DCHECK(args.length() == 1);
+  DCHECK_EQ(args.length(), 1);
   CONVERT_ARG_HANDLE_CHECKED(String, s, 0);
   return ConvertCase(s, isolate, isolate->runtime_state()->to_lower_mapping());
 }
@@ -1085,7 +1085,7 @@ RUNTIME_FUNCTION(Runtime_StringToLowerCase) {
 
 RUNTIME_FUNCTION(Runtime_StringToUpperCase) {
   HandleScope scope(isolate);
-  DCHECK(args.length() == 1);
+  DCHECK_EQ(args.length(), 1);
   CONVERT_ARG_HANDLE_CHECKED(String, s, 0);
   return ConvertCase(s, isolate, isolate->runtime_state()->to_upper_mapping());
 }

src/runtime/runtime.h

@@ -262,7 +262,10 @@ namespace internal {
   F(BreakIteratorFirst, 1, 1)                \
   F(BreakIteratorNext, 1, 1)                 \
   F(BreakIteratorCurrent, 1, 1)              \
-  F(BreakIteratorBreakType, 1, 1)
+  F(BreakIteratorBreakType, 1, 1)            \
+  F(StringToLowerCaseI18N, 1, 1)             \
+  F(StringToUpperCaseI18N, 1, 1)             \
+  F(StringLocaleConvertCase, 3, 1)
 #else
 #define FOR_EACH_INTRINSIC_I18N(F)
 #endif

src/v8.gyp

@@ -1988,17 +1988,6 @@
         }, {
           'toolsets': ['target'],
         }],
-        ['v8_enable_i18n_support==1', {
-          'variables': {
-            'i18n_library_files': [
-              'js/i18n.js',
-            ],
-          },
-        }, {
-          'variables': {
-            'i18n_library_files': [],
-          },
-        }],
       ],
       'variables': {
         'library_files': [
@@ -2048,6 +2037,12 @@
         'libraries_experimental_bin_file': '<(SHARED_INTERMEDIATE_DIR)/libraries-experimental.bin',
         'libraries_extras_bin_file': '<(SHARED_INTERMEDIATE_DIR)/libraries-extras.bin',
         'libraries_experimental_extras_bin_file': '<(SHARED_INTERMEDIATE_DIR)/libraries-experimental-extras.bin',
+        'conditions': [
+          ['v8_enable_i18n_support==1', {
+            'library_files': ['js/i18n.js'],
+            'experimental_library_files': ['js/icu-case-mapping.js'],
+          }],
+        ],
       },
       'actions': [
         {
@@ -2055,7 +2050,6 @@
           'inputs': [
             '../tools/js2c.py',
             '<@(library_files)',
-            '<@(i18n_library_files)'
           ],
           'outputs': ['<(SHARED_INTERMEDIATE_DIR)/libraries.cc'],
           'action': [
@@ -2064,7 +2058,6 @@
             '<(SHARED_INTERMEDIATE_DIR)/libraries.cc',
             'CORE',
             '<@(library_files)',
-            '<@(i18n_library_files)'
           ],
         },
         {
@@ -2072,7 +2065,6 @@
           'inputs': [
             '../tools/js2c.py',
             '<@(library_files)',
-            '<@(i18n_library_files)'
           ],
           'outputs': ['<@(libraries_bin_file)'],
           'action': [
@@ -2081,7 +2073,6 @@
             '<(SHARED_INTERMEDIATE_DIR)/libraries.cc',
             'CORE',
             '<@(library_files)',
-            '<@(i18n_library_files)',
             '--startup_blob', '<@(libraries_bin_file)',
             '--nojs',
           ],
@@ -2098,7 +2089,7 @@
             '../tools/js2c.py',
             '<(SHARED_INTERMEDIATE_DIR)/experimental-libraries.cc',
             'EXPERIMENTAL',
-            '<@(experimental_library_files)'
+            '<@(experimental_library_files)',
           ],
         },
         {

test/intl/general/case-mapping.js

+// Copyright 2016 the V8 project authors. All rights reserved.
+// Use of this source code is governed by a BSD-style license that can be
+// found in the LICENSE file.
+
+// Flags: --icu_case_mapping
+
+// Some edge cases that unibrow got wrong
+
+assertEquals("𐐘", "𐑀".toUpperCase());
+assertEquals("𐑀", "𐐘".toLowerCase());
+assertEquals("σ", "Σ".toLowerCase());
+
+// Some different paths in the ICU case conversion fastpath
+
+assertEquals("σς", "\u03A3\u03A3".toLowerCase());
+// Expand sharp s in latin1 fastpath
+assertEquals("ASSB", "A\u00DFB".toUpperCase());
+assertEquals("AB", "Ab".toUpperCase());
+// Find first upper case in fastpath
+assertEquals("ab", "aB".toLowerCase());
+assertEquals("AÜ", "aü".toUpperCase());
+assertEquals("AÜ", "AÜ".toUpperCase());
+assertEquals("aü", "aü".toLowerCase());
+assertEquals("aü", "AÜ".toLowerCase());
+assertEquals("aü", "AÜ".toLowerCase());
+
+// Starts with fastpath, but switches to full Unicode path
+// U+00FF is uppercased to U+0178.
+assertEquals("AŸ", "aÿ".toUpperCase());
+// U+00B5 (µ) is uppercased to U+039C (Μ)
+assertEquals("AΜ", "aµ".toUpperCase());
+
+// Buffer size increase
+assertEquals("CSSBẶ", "cßbặ".toUpperCase());
+assertEquals("FIFLFFIFFL", "\uFB01\uFB02\uFB03\uFB04".toUpperCase());
+// OneByte input with buffer size increase: non-fast path
+assertEquals("ABCSS", "abCß".toLocaleUpperCase("tr"));
+
+// More comprehensive tests for "tr", "az" and "lt" are in
+// test262/intl402/Strings/*
+
+// Buffer size decrease with a single locale or locale list.
+// In Turkic (tr, az), U+0307 preceeded by Capital Letter I is dropped.
+assertEquals("abci", "aBcI\u0307".toLocaleLowerCase("tr"));
+assertEquals("abci", "aBcI\u0307".toLocaleLowerCase("az"));
+assertEquals("abci", "aBcI\u0307".toLocaleLowerCase(["tr", "en"]));
+
+// Cons string
+assertEquals("abcijkl", ("aBcI" + "\u0307jkl").toLocaleLowerCase("tr"));
+assertEquals("abcijkl",
+             ("aB" + "cI" + "\u0307j" + "kl").toLocaleLowerCase("tr"));
+assertEquals("abci\u0307jkl", ("aBcI" + "\u0307jkl").toLocaleLowerCase("en"));
+assertEquals("abci\u0307jkl",
+             ("aB" + "cI" + "\u0307j" + "kl").toLocaleLowerCase("en"));
+assertEquals("abci\u0307jkl", ("aBcI" + "\u0307jkl").toLowerCase());
+assertEquals("abci\u0307jkl",
+             ("aB" + "cI" + "\u0307j" + "kl").toLowerCase());
+
+// "tr" and "az" should behave identically.
+assertEquals("aBcI\u0307".toLocaleLowerCase("tr"),
+             "aBcI\u0307".toLocaleLowerCase("az"));
+// What matters is the first locale in the locale list.
+assertEquals("aBcI\u0307".toLocaleLowerCase(["tr", "en", "fr"]),
+             "aBcI\u0307".toLocaleLowerCase("tr"));
+assertEquals("aBcI\u0307".toLocaleLowerCase(["en", "tr", "az"]),
+             "aBcI\u0307".toLocaleLowerCase("en"));
+assertEquals("aBcI\u0307".toLocaleLowerCase(["en", "tr", "az"]),
+             "aBcI\u0307".toLowerCase());
+
+// An empty locale list is the same as the default locale. Try these tests
+// under Turkish and Greek locale.
+assertEquals("aBcI\u0307".toLocaleLowerCase([]),
+             "aBcI\u0307".toLocaleLowerCase());
+assertEquals("aBcI\u0307".toLocaleLowerCase([]),
+             "aBcI\u0307".toLocaleLowerCase(Intl.GetDefaultLocale));
+assertEquals("άόύώ".toLocaleUpperCase([]), "άόύώ".toLocaleUpperCase());
+assertEquals("άόύώ".toLocaleUpperCase([]),
+             "άόύώ".toLocaleUpperCase(Intl.GetDefaultLocale));
+
+
+// English/root locale keeps U+0307 (combining dot above).
+assertEquals("abci\u0307", "aBcI\u0307".toLocaleLowerCase("en"));
+assertEquals("abci\u0307", "aBcI\u0307".toLocaleLowerCase(["en", "tr"]));
+assertEquals("abci\u0307", "aBcI\u0307".toLowerCase());
+
+// Greek uppercasing: not covered by intl402/String/*, yet. Tonos (U+0301) and
+// other diacritic marks are dropped. This rule is based on the current CLDR's
+// el-Upper transformation, but Greek uppercasing rules are more sophisticated
+// than this. See http://bugs.icu-project.org/trac/ticket/10582 and
+// http://unicode.org/cldr/trac/ticket/7905 .
+assertEquals("Α", "α\u0301".toLocaleUpperCase("el"));
+assertEquals("Α", "α\u0301".toLocaleUpperCase("el-GR"));
+assertEquals("Α", "α\u0301".toLocaleUpperCase("el-Grek"));
+assertEquals("Α", "α\u0301".toLocaleUpperCase("el-Grek-GR"));
+assertEquals("Α", "ά".toLocaleUpperCase("el"));
+assertEquals("ΑΟΥΩ", "άόύώ".toLocaleUpperCase("el"));
+assertEquals("ΑΟΥΩ", "α\u0301ο\u0301υ\u0301ω\u0301".toLocaleUpperCase("el"));
+assertEquals("ΑΟΥΩ", "άόύώ".toLocaleUpperCase("el"));
+assertEquals("ΟΕ", "Ό\u1f15".toLocaleUpperCase("el"));
+assertEquals("ΟΕ", "Ο\u0301ε\u0314\u0301".toLocaleUpperCase("el"));
+
+// Input and output are identical.
+assertEquals("αβγδε", "αβγδε".toLocaleLowerCase("el"));
+assertEquals("ΑΒΓΔΕ", "ΑΒΓΔΕ".toLocaleUpperCase("el"));
+assertEquals("ΑΒΓΔΕАБ𝐀𝐁", "ΑΒΓΔΕАБ𝐀𝐁".toLocaleUpperCase("el"));
+assertEquals("ABCDEÂÓḴ123", "ABCDEÂÓḴ123".toLocaleUpperCase("el"));
+// ASCII-only or Latin-1 only: 1-byte
+assertEquals("ABCDE123", "ABCDE123".toLocaleUpperCase("el"));
+assertEquals("ABCDEÂÓ123", "ABCDEÂÓ123".toLocaleUpperCase("el"));
+
+// To make sure that the input string is not overwritten in place.
+var strings = ["abCdef", "αβγδε", "άόύώ", "аб"];
+for (var s  of strings) {
+  var backupAsArray = s.split("");
+  var uppered = s.toLocaleUpperCase("el");
+  assertEquals(s, backupAsArray.join(""));
+}
+
+// In other locales, U+0301 is preserved.
+assertEquals("Α\u0301Ο\u0301Υ\u0301Ω\u0301",
+             "α\u0301ο\u0301υ\u0301ω\u0301".toLocaleUpperCase("en"));
+assertEquals("Α\u0301Ο\u0301Υ\u0301Ω\u0301",
+             "α\u0301ο\u0301υ\u0301ω\u0301".toUpperCase());
+
+// Plane 1; Deseret and Warang Citi Script.
+assertEquals("\u{10400}\u{118A0}", "\u{10428}\u{118C0}".toUpperCase());
+assertEquals("\u{10428}\u{118C0}", "\u{10400}\u{118A0}".toLowerCase());
+// Mathematical Bold {Capital, Small} Letter A do not change.
+assertEquals("\u{1D400}\u{1D41A}", "\u{1D400}\u{1D41A}".toUpperCase());
+assertEquals("\u{1D400}\u{1D41A}", "\u{1D400}\u{1D41A}".toLowerCase());
+// Plane 1; New characters in Unicode 8.0
+assertEquals("\u{10C80}", "\u{10CC0}".toUpperCase());
+assertEquals("\u{10CC0}", "\u{10C80}".toLowerCase());
+assertEquals("\u{10C80}", "\u{10CC0}".toLocaleUpperCase());
+assertEquals("\u{10CC0}", "\u{10C80}".toLocaleLowerCase());
+assertEquals("\u{10C80}", "\u{10CC0}".toLocaleUpperCase(["tr"]));
+assertEquals("\u{10C80}", "\u{10CC0}".toLocaleUpperCase(["tr"]));
+assertEquals("\u{10CC0}", "\u{10C80}".toLocaleLowerCase());

test/intl/testcfg.py

@@ -26,10 +26,12 @@
 # OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 
 import os
+import re
 
 from testrunner.local import testsuite
 from testrunner.objects import testcase
 
+FLAGS_PATTERN = re.compile(r"//\s+Flags:(.*)")
 
 class IntlTestSuite(testsuite.TestSuite):
 
@@ -55,7 +57,11 @@ class IntlTestSuite(testsuite.TestSuite):
     return tests
 
   def GetFlagsForTestCase(self, testcase, context):
+    source = self.GetSourceForTest(testcase)
     flags = ["--allow-natives-syntax"] + context.mode_flags
+    flags_match = re.findall(FLAGS_PATTERN, source)
+    for match in flags_match:
+      flags += match.strip().split()
 
     files = []
     files.append(os.path.join(self.root, "assert.js"))
@@ -71,6 +77,10 @@ class IntlTestSuite(testsuite.TestSuite):
 
     return testcase.flags + flags
 
+  def GetSourceForTest(self, testcase):
+    filename = os.path.join(self.root, testcase.path + self.suffix())
+    with open(filename) as f:
+      return f.read()
 
 def GetSuite(name, root):
   return IntlTestSuite(name, root)

test/test262/test262.status

@@ -139,14 +139,16 @@
   'intl402/NumberFormat/11.1.1_1': [FAIL],
 
   # https://code.google.com/p/v8/issues/detail?id=4476
-  'built-ins/String/prototype/toLocaleLowerCase/special_casing_conditional': [FAIL],
-  'built-ins/String/prototype/toLocaleLowerCase/supplementary_plane': [FAIL],
+  # The bug is fixed but behind a flag, --icu_case_mapping.
   'built-ins/String/prototype/toLowerCase/special_casing_conditional': [FAIL],
   'built-ins/String/prototype/toLowerCase/supplementary_plane': [FAIL],
-  'built-ins/String/prototype/toLocaleUpperCase/supplementary_plane': [FAIL],
   'built-ins/String/prototype/toUpperCase/supplementary_plane': [FAIL],
 
   # https://code.google.com/p/v8/issues/detail?id=4477
+  # The bug is fixed but behind a flag, --icu_case_mapping.
+  'built-ins/String/prototype/toLocaleUpperCase/supplementary_plane': [FAIL],
+  'built-ins/String/prototype/toLocaleLowerCase/supplementary_plane': [FAIL],
+  'built-ins/String/prototype/toLocaleLowerCase/special_casing_conditional': [FAIL],
   'intl402/String/prototype/toLocaleLowerCase/special_casing_Azeri': [FAIL],
   'intl402/String/prototype/toLocaleLowerCase/special_casing_Lithuanian': [FAIL],
   'intl402/String/prototype/toLocaleLowerCase/special_casing_Turkish': [FAIL],
@@ -423,6 +425,22 @@
   'built-ins/String/prototype/normalize/return-normalized-string': [SKIP],
   'built-ins/String/prototype/normalize/return-normalized-string-from-coerced-form': [SKIP],
   'built-ins/String/prototype/normalize/return-normalized-string-using-default-parameter': [SKIP],
+
+  # Case-conversion is not fully compliant to the Unicode spec with i18n off.
+  'built-ins/String/prototype/toLocaleLowerCase/special_casing_conditional': [FAIL],
+  'built-ins/String/prototype/toLocaleLowerCase/supplementary_plane': [FAIL],
+  'built-ins/String/prototype/toLowerCase/special_casing_conditional': [FAIL],
+  'built-ins/String/prototype/toLowerCase/supplementary_plane': [FAIL],
+  'built-ins/String/prototype/toLocaleUpperCase/supplementary_plane': [FAIL],
+  'built-ins/String/prototype/toUpperCase/supplementary_plane': [FAIL],
+
+  # Locale-sensitive case-conversion is not available with i18n off.
+  'intl402/String/prototype/toLocaleLowerCase/special_casing_Azeri': [FAIL],
+  'intl402/String/prototype/toLocaleLowerCase/special_casing_Lithuanian': [FAIL],
+  'intl402/String/prototype/toLocaleLowerCase/special_casing_Turkish': [FAIL],
+  'intl402/String/prototype/toLocaleUpperCase/special_casing_Azeri': [FAIL],
+  'intl402/String/prototype/toLocaleUpperCase/special_casing_Lithuanian': [FAIL],
+  'intl402/String/prototype/toLocaleUpperCase/special_casing_Turkish': [FAIL],
 }],  # no_i18n == True
 
 ['arch == arm or arch == mipsel or arch == mips or arch == arm64 or arch == mips64 or arch == mips64el', {