[regexp] Port remaining JS functions in regexp.js

src/js/i18n.js

 // Copyright 2013 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.
 
 // ECMAScript 402 API implementation.
 
 /**
  * Intl object is a single object that has some named properties,
  * all of which are constructors.
  */
 (function(global, utils) {
 
 "use strict";
 
 %CheckIsBootstrapping();
 
 // -------------------------------------------------------------------
 // Imports
 
 var ArrayJoin;
 var ArrayPush;
 var GlobalDate = global.Date;
 var GlobalNumber = global.Number;
 var GlobalRegExp = global.RegExp;
 var GlobalString = global.String;
 var InstallFunctions = utils.InstallFunctions;
 var InstallGetter = utils.InstallGetter;
 var InternalArray = utils.InternalArray;
-var InternalRegExpMatch;
-var InternalRegExpReplace
 var ObjectHasOwnProperty = utils.ImportNow("ObjectHasOwnProperty");
 var OverrideFunction = utils.OverrideFunction;
 var patternSymbol = utils.ImportNow("intl_pattern_symbol");
 var resolvedSymbol = utils.ImportNow("intl_resolved_symbol");
 var SetFunctionName = utils.SetFunctionName;
 var StringSubstr = GlobalString.prototype.substr;
 var StringSubstring = GlobalString.prototype.substring;
 
 utils.Import(function(from) {
   ArrayJoin = from.ArrayJoin;
   ArrayPush = from.ArrayPush;
-  InternalRegExpMatch = from.InternalRegExpMatch;
-  InternalRegExpReplace = from.InternalRegExpReplace;
 });
 
 // Utilities for definitions
 
 function InstallFunction(object, name, func) {
   InstallFunctions(object, DONT_ENUM, [name, func]);
 }
 
 
 function InstallConstructor(object, name, func) {
@@ skipping 188 matching lines @@
   }
   return TIMEZONE_NAME_LOCATION_PART_RE;
 }
 
 
 /**
  * Returns an intersection of locales and service supported locales.
  * Parameter locales is treated as a priority list.
  */
 function supportedLocalesOf(service, locales, options) {
-  if (IS_NULL(InternalRegExpMatch(GetServiceRE(), service))) {
+  if (IS_NULL(%regexp_internal_match(GetServiceRE(), service))) {
     throw %make_error(kWrongServiceType, service);
   }
 
   // Provide defaults if matcher was not specified.
   if (IS_UNDEFINED(options)) {
     options = {};
   } else {
     options = TO_OBJECT(options);
   }
 
@@ skipping 27 matching lines @@
 
 /**
  * Returns the subset of the provided BCP 47 language priority list for which
  * this service has a matching locale when using the BCP 47 Lookup algorithm.
  * Locales appear in the same order in the returned list as in the input list.
  */
 function lookupSupportedLocalesOf(requestedLocales, availableLocales) {
   var matchedLocales = new InternalArray();
   for (var i = 0; i < requestedLocales.length; ++i) {
     // Remove -u- extension.
-    var locale = InternalRegExpReplace(
+    var locale = %regexp_internal_replace(
         GetUnicodeExtensionRE(), requestedLocales[i], '');
     do {
       if (!IS_UNDEFINED(availableLocales[locale])) {
         // Push requested locale not the resolved one.
         %_Call(ArrayPush, matchedLocales, requestedLocales[i]);
         break;
       }
       // Truncate locale if possible, if not break.
       var pos = %StringLastIndexOf(locale, '-');
       if (pos === -1) {
@@ skipping 89 matching lines @@
 
   return resolved;
 }
 
 
 /**
  * Returns best matched supported locale and extension info using basic
  * lookup algorithm.
  */
 function lookupMatcher(service, requestedLocales) {
-  if (IS_NULL(InternalRegExpMatch(GetServiceRE(), service))) {
+  if (IS_NULL(%regexp_internal_match(GetServiceRE(), service))) {
     throw %make_error(kWrongServiceType, service);
   }
 
   // Cache these, they don't ever change per service.
   if (IS_UNDEFINED(AVAILABLE_LOCALES[service])) {
     AVAILABLE_LOCALES[service] = getAvailableLocalesOf(service);
   }
 
   for (var i = 0; i < requestedLocales.length; ++i) {
     // Remove all extensions.
-    var locale = InternalRegExpReplace(
+    var locale = %regexp_internal_replace(
         GetAnyExtensionRE(), requestedLocales[i], '');
     do {
       if (!IS_UNDEFINED(AVAILABLE_LOCALES[service][locale])) {
         // Return the resolved locale and extension.
-        var extensionMatch = InternalRegExpMatch(
+        var extensionMatch = %regexp_internal_match(
             GetUnicodeExtensionRE(), requestedLocales[i]);
         var extension = IS_NULL(extensionMatch) ? '' : extensionMatch[0];
         return {'locale': locale, 'extension': extension, 'position': i};
       }
       // Truncate locale if possible.
       var pos = %StringLastIndexOf(locale, '-');
       if (pos === -1) {
         break;
       }
       locale = %_Call(StringSubstring, locale, 0, pos);
@@ skipping 177 matching lines @@
     return original;
   }
 
   var locales = %GetLanguageTagVariants([original, resolved]);
   if (locales[0].maximized !== locales[1].maximized) {
     return resolved;
   }
 
   // Preserve extensions of resolved locale, but swap base tags with original.
   var resolvedBase = new GlobalRegExp('^' + locales[1].base, 'g');
-  return InternalRegExpReplace(resolvedBase, resolved, locales[0].base);
+  return %regexp_internal_replace(resolvedBase, resolved, locales[0].base);
 }
 
 
 /**
  * Returns an Object that contains all of supported locales for a given
  * service.
  * In addition to the supported locales we add xx-ZZ locale for each xx-Yyyy-ZZ
  * that is supported. This is required by the spec.
  */
 function getAvailableLocalesOf(service) {
   var available = %AvailableLocalesOf(service);
 
   for (var i in available) {
     if (HAS_OWN_PROPERTY(available, i)) {
-      var parts = InternalRegExpMatch(
+      var parts = %regexp_internal_match(
           /^([a-z]{2,3})-([A-Z][a-z]{3})-([A-Z]{2})$/, i);
       if (!IS_NULL(parts)) {
         // Build xx-ZZ. We don't care about the actual value,
         // as long it's not undefined.
         available[parts[1] + '-' + parts[3]] = null;
       }
     }
   }
 
   return available;
@@ skipping 51 matching lines @@
          %StringToLowerCase(%_Call(StringSubstr, word, 1));
 }
 
 /**
  * Returns titlecased location, bueNos_airES -> Buenos_Aires
  * or ho_cHi_minH -> Ho_Chi_Minh. It is locale-agnostic and only
  * deals with ASCII only characters.
  * 'of', 'au' and 'es' are special-cased and lowercased.
  */
 function toTitleCaseTimezoneLocation(location) {
-  var match = InternalRegExpMatch(GetTimezoneNameLocationPartRE(), location)
+  var match = %regexp_internal_match(GetTimezoneNameLocationPartRE(), location)
   if (IS_NULL(match)) throw %make_range_error(kExpectedLocation, location);
 
   var result = toTitleCaseWord(match[1]);
   if (!IS_UNDEFINED(match[2]) && 2 < match.length) {
     // The first character is a separator, '_' or '-'.
     // None of IANA zone names has both '_' and '-'.
     var separator = %_Call(StringSubstring, match[2], 0, 1);
     var parts = %StringSplit(match[2], separator, kMaxUint32);
     for (var i = 1; i < parts.length; i++) {
       var part = parts[i]
@@ skipping 14 matching lines @@
 function canonicalizeLanguageTag(localeID) {
   // null is typeof 'object' so we have to do extra check.
   if ((!IS_STRING(localeID) && !IS_RECEIVER(localeID)) ||
       IS_NULL(localeID)) {
     throw %make_type_error(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}$/, localeID))) {
+      !IS_NULL(%regexp_internal_match(/^[a-z]{2,3}$/, localeID))) {
     return localeID;
   }
 
   var localeString = TO_STRING(localeID);
 
   if (isStructuallyValidLanguageTag(localeString) === false) {
     throw %make_range_error(kInvalidLanguageTag, localeString);
   }
 
   // ECMA 402 6.2.3
@@ skipping 57 matching lines @@
  * primary/extended language, script, region, variant are not checked
  * against the IANA language subtag registry.
  *
  * ICU is too permissible and lets invalid tags, like
  * hant-cmn-cn, through.
  *
  * Returns false if the language tag is invalid.
  */
 function isStructuallyValidLanguageTag(locale) {
   // Check if it's well-formed, including grandfadered tags.
-  if (IS_NULL(InternalRegExpMatch(GetLanguageTagRE(), locale))) {
+  if (IS_NULL(%regexp_internal_match(GetLanguageTagRE(), locale))) {
     return false;
   }
 
   // Just return if it's a x- form. It's all private.
   if (%StringIndexOf(locale, 'x-', 0) === 0) {
     return true;
   }
 
   // Check if there are any duplicate variants or singletons (extensions).
 
   // Remove private use section.
   locale = %StringSplit(locale, '-x-', kMaxUint32)[0];
 
   // Skip language since it can match variant regex, so we start from 1.
   // We are matching i-klingon here, but that's ok, since i-klingon-klingon
   // is not valid and would fail LANGUAGE_TAG_RE test.
   var variants = new InternalArray();
   var extensions = new InternalArray();
   var parts = %StringSplit(locale, '-', kMaxUint32);
   for (var i = 1; i < parts.length; i++) {
     var value = parts[i];
-    if (!IS_NULL(InternalRegExpMatch(GetLanguageVariantRE(), value)) &&
+    if (!IS_NULL(%regexp_internal_match(GetLanguageVariantRE(), value)) &&
         extensions.length === 0) {
       if (%ArrayIndexOf(variants, value, 0) === -1) {
         %_Call(ArrayPush, variants, value);
       } else {
         return false;
       }
     }
 
-    if (!IS_NULL(InternalRegExpMatch(GetLanguageSingletonRE(), value))) {
+    if (!IS_NULL(%regexp_internal_match(GetLanguageSingletonRE(), value))) {
       if (%ArrayIndexOf(extensions, value, 0) === -1) {
         %_Call(ArrayPush, extensions, value);
       } else {
         return false;
       }
     }
   }
 
   return true;
  }
@@ skipping 248 matching lines @@
 
 AddBoundMethod(Intl.Collator, 'compare', compare, 2, 'collator');
 
 /**
  * Verifies that the input is a well-formed ISO 4217 currency code.
  * Don't uppercase to test. It could convert invalid code into a valid one.
  * For example \u00DFP (Eszett+P) becomes SSP.
  */
 function isWellFormedCurrencyCode(currency) {
   return typeof currency == "string" && currency.length == 3 &&
-      IS_NULL(InternalRegExpMatch(/[^A-Za-z]/, currency));
+      IS_NULL(%regexp_internal_match(/[^A-Za-z]/, currency));
 }
 
 
 /**
  * Returns the valid digit count for a property, or throws RangeError on
  * a value out of the range.
  */
 function getNumberOption(options, property, min, max, fallback) {
   var value = options[property];
   if (!IS_UNDEFINED(value)) {
@@ skipping 297 matching lines @@
     return '';
   }
 }
 
 
 /**
  * Returns object that matches LDML representation of the date.
  */
 function fromLDMLString(ldmlString) {
   // First remove '' quoted text, so we lose 'Uhr' strings.
-  ldmlString = InternalRegExpReplace(GetQuotedStringRE(), ldmlString, '');
+  ldmlString = %regexp_internal_replace(GetQuotedStringRE(), ldmlString, '');
 
   var options = {};
-  var match = InternalRegExpMatch(/E{3,5}/, ldmlString);
+  var match = %regexp_internal_match(/E{3,5}/, ldmlString);
   options = appendToDateTimeObject(
       options, 'weekday', match, {EEEEE: 'narrow', EEE: 'short', EEEE: 'long'});
 
-  match = InternalRegExpMatch(/G{3,5}/, ldmlString);
+  match = %regexp_internal_match(/G{3,5}/, ldmlString);
   options = appendToDateTimeObject(
       options, 'era', match, {GGGGG: 'narrow', GGG: 'short', GGGG: 'long'});
 
-  match = InternalRegExpMatch(/y{1,2}/, ldmlString);
+  match = %regexp_internal_match(/y{1,2}/, ldmlString);
   options = appendToDateTimeObject(
       options, 'year', match, {y: 'numeric', yy: '2-digit'});
 
-  match = InternalRegExpMatch(/M{1,5}/, ldmlString);
+  match = %regexp_internal_match(/M{1,5}/, ldmlString);
   options = appendToDateTimeObject(options, 'month', match, {MM: '2-digit',
       M: 'numeric', MMMMM: 'narrow', MMM: 'short', MMMM: 'long'});
 
   // Sometimes we get L instead of M for month - standalone name.
-  match = InternalRegExpMatch(/L{1,5}/, ldmlString);
+  match = %regexp_internal_match(/L{1,5}/, ldmlString);
   options = appendToDateTimeObject(options, 'month', match, {LL: '2-digit',
       L: 'numeric', LLLLL: 'narrow', LLL: 'short', LLLL: 'long'});
 
-  match = InternalRegExpMatch(/d{1,2}/, ldmlString);
+  match = %regexp_internal_match(/d{1,2}/, ldmlString);
   options = appendToDateTimeObject(
       options, 'day', match, {d: 'numeric', dd: '2-digit'});
 
-  match = InternalRegExpMatch(/h{1,2}/, ldmlString);
+  match = %regexp_internal_match(/h{1,2}/, ldmlString);
   if (match !== null) {
     options['hour12'] = true;
   }
   options = appendToDateTimeObject(
       options, 'hour', match, {h: 'numeric', hh: '2-digit'});
 
-  match = InternalRegExpMatch(/H{1,2}/, ldmlString);
+  match = %regexp_internal_match(/H{1,2}/, ldmlString);
   if (match !== null) {
     options['hour12'] = false;
   }
   options = appendToDateTimeObject(
       options, 'hour', match, {H: 'numeric', HH: '2-digit'});
 
-  match = InternalRegExpMatch(/m{1,2}/, ldmlString);
+  match = %regexp_internal_match(/m{1,2}/, ldmlString);
   options = appendToDateTimeObject(
       options, 'minute', match, {m: 'numeric', mm: '2-digit'});
 
-  match = InternalRegExpMatch(/s{1,2}/, ldmlString);
+  match = %regexp_internal_match(/s{1,2}/, ldmlString);
   options = appendToDateTimeObject(
       options, 'second', match, {s: 'numeric', ss: '2-digit'});
 
-  match = InternalRegExpMatch(/z|zzzz/, ldmlString);
+  match = %regexp_internal_match(/z|zzzz/, ldmlString);
   options = appendToDateTimeObject(
       options, 'timeZoneName', match, {z: 'short', zzzz: 'long'});
 
   return options;
 }
 
 
 function appendToDateTimeObject(options, option, match, pairs) {
   if (IS_NULL(match)) {
     if (!HAS_OWN_PROPERTY(options, option)) {
@@ skipping 308 matching lines @@
   var upperID = %StringToUpperCase(tzID);
   if (upperID === 'UTC' || upperID === 'GMT' ||
       upperID === 'ETC/UTC' || upperID === 'ETC/GMT') {
     return 'UTC';
   }
 
   // TODO(jshin): Add support for Etc/GMT[+-]([1-9]|1[0-2])
 
   // We expect only _, '-' and / beside ASCII letters.
   // All inputs should conform to Area/Location(/Location)* from now on.
-  var match = InternalRegExpMatch(GetTimezoneNameCheckRE(), tzID);
+  var match = %regexp_internal_match(GetTimezoneNameCheckRE(), tzID);
   if (IS_NULL(match)) throw %make_range_error(kExpectedTimezoneID, tzID);
 
   var result = toTitleCaseTimezoneLocation(match[1]) + '/' +
                toTitleCaseTimezoneLocation(match[2]);
 
   if (!IS_UNDEFINED(match[3]) && 3 < match.length) {
     var locations = %StringSplit(match[3], '/', kMaxUint32);
     // The 1st element is empty. Starts with i=1.
     for (var i = 1; i < locations.length; i++) {
       result = result + '/' + toTitleCaseTimezoneLocation(locations[i]);
@@ skipping 437 matching lines @@
   }
 );
 
 %FunctionRemovePrototype(FormatDateToParts);
 
 utils.Export(function(to) {
   to.FormatDateToParts = FormatDateToParts;
 });
 
 })