Timezone name check fix

1. Location names with more than one underscores (e.g. Ho_Chi_Minh)
   didn't work because of the way capturing works with repeated patterns
   in RE. It's now supported by changing the RE to capture the whole string
   and splitting on '_' in the next step.

2. Adds support for location names with a hyphen

3. Adds support for timezone ids with three parts (e.g.
   American/Argentina/Buenos_Aires)

4. Adds special handling of 'au', 'es' and 'of' in zone ids. They need to be kept in lowercase. (see the full list at https://en.wikipedia.org/wiki/List_of_tz_database_time_zones )

5. Adds regression tests for all the above and make the existing tests
   more robust against future ICU changes. ICU canonicalizes zone names to
   deprecated names, but it may change. (
   http://bugs.icu-project.org/trac/ticket/12044 )

BUG=364374
LOG=Y

Review URL: https://codereview.chromium.org/1529363005

Cr-Commit-Position: refs/heads/master@{#33097}

src/js/i18n.js

@@ -176,12 +176,25 @@ var TIMEZONE_NAME_CHECK_RE = UNDEFINED;
 
 function GetTimezoneNameCheckRE() {
   if (IS_UNDEFINED(TIMEZONE_NAME_CHECK_RE)) {
-    TIMEZONE_NAME_CHECK_RE =
-        new GlobalRegExp('^([A-Za-z]+)/([A-Za-z]+)(?:_([A-Za-z]+))*$');
+    TIMEZONE_NAME_CHECK_RE = new GlobalRegExp(
+        '^([A-Za-z]+)/([A-Za-z_-]+)((?:\/[A-Za-z_-]+)+)*$');
   }
   return TIMEZONE_NAME_CHECK_RE;
 }
 
+/**
+ * Matches valid location parts of IANA time zone names.
+ */
+var TIMEZONE_NAME_LOCATION_PART_RE = UNDEFINED;
+
+function GetTimezoneNameLocationPartRE() {
+  if (IS_UNDEFINED(TIMEZONE_NAME_LOCATION_PART_RE)) {
+    TIMEZONE_NAME_LOCATION_PART_RE =
+        new GlobalRegExp('^([A-Za-z]+)((?:[_-][A-Za-z]+)+)*$');
+  }
+  return TIMEZONE_NAME_LOCATION_PART_RE;
+}
+
 /**
  * Adds bound method to the prototype of the given object.
  */
@@ -678,6 +691,34 @@ function toTitleCaseWord(word) {
          %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 = %_Call(StringMatch, location, GetTimezoneNameLocationPartRE());
+  if (IS_NULL(match)) throw MakeRangeError(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 = %_Call(StringSplit, match[2], separator);
+    for (var i = 1; i < parts.length; i++) {
+      var part = parts[i]
+      var lowercasedPart = %StringToLowerCase(part);
+      result = result + separator +
+          ((lowercasedPart !== 'es' &&
+            lowercasedPart !== 'of' && lowercasedPart !== 'au') ?
+          toTitleCaseWord(part) : lowercasedPart);
+    }
+  }
+  return result;
+}
+
 /**
  * Canonicalizes the language tag, or throws in case the tag is invalid.
  */
@@ -1735,8 +1776,8 @@ addBoundMethod(Intl.DateTimeFormat, 'v8Parse', parseDate, 1);
 
 
 /**
- * Returns canonical Area/Location name, or throws an exception if the zone
- * name is invalid IANA name.
+ * Returns canonical Area/Location(/Location) name, or throws an exception
+ * if the zone name is invalid IANA name.
  */
 function canonicalizeTimeZoneID(tzID) {
   // Skip undefined zones.
@@ -1751,16 +1792,22 @@ function canonicalizeTimeZoneID(tzID) {
     return 'UTC';
   }
 
-  // We expect only _ and / beside ASCII letters.
-  // All inputs should conform to Area/Location from now on.
+  // 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 = %_Call(StringMatch, tzID, GetTimezoneNameCheckRE());
-  if (IS_NULL(match)) throw MakeRangeError(kExpectedLocation, tzID);
+  if (IS_NULL(match)) throw MakeRangeError(kExpectedTimezoneID, tzID);
+
+  var result = toTitleCaseTimezoneLocation(match[1]) + '/' +
+               toTitleCaseTimezoneLocation(match[2]);
 
-  var result = toTitleCaseWord(match[1]) + '/' + toTitleCaseWord(match[2]);
-  var i = 3;
-  while (!IS_UNDEFINED(match[i]) && i < match.length) {
-    result = result + '_' + toTitleCaseWord(match[i]);
-    i++;
+  if (!IS_UNDEFINED(match[3]) && 3 < match.length) {
+    var locations = %_Call(StringSplit, match[3], '/');
+    // The 1st element is empty. Starts with i=1.
+    for (var i = 1; i < locations.length; i++) {
+      result = result + '/' + toTitleCaseTimezoneLocation(locations[i]);
+    }
   }
 
   return result;

src/messages.h

@@ -329,7 +329,11 @@ class CallSite {
   T(UnsupportedSuper, "Unsupported reference to 'super'")                      \
   /* RangeError */                                                             \
   T(DateRange, "Provided date is not in valid range.")                         \
-  T(ExpectedLocation, "Expected Area/Location for time zone, got %")           \
+  T(ExpectedTimezoneID,                                                        \
+    "Expected Area/Location(/Location)* for time zone, got %")                 \
+  T(ExpectedLocation,                                                          \
+    "Expected letters optionally connected with underscores or hyphens for "   \
+    "a location, got %")                                                       \
   T(InvalidArrayBufferLength, "Invalid array buffer length")                   \
   T(ArrayBufferAllocationFailed, "Array buffer allocation failed")             \
   T(InvalidArrayLength, "Invalid array length")                                \

test/mjsunit/regress/regress-487322.js

-// Copyright 2015 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.
-
-if (this.Intl) {
-  // Normalizes Kat{h,}mandu (chromium:487322)
-  df = new Intl.DateTimeFormat('en-US', {'timeZone': 'Asia/Katmandu'})
-  assertEquals('Asia/Katmandu', df.resolvedOptions().timeZone);
-
-  df = new Intl.DateTimeFormat('en-US', {'timeZone': 'Asia/Kathmandu'})
-  assertEquals('Asia/Katmandu', df.resolvedOptions().timeZone);
-
-  // Throws for unsupported time zones.
-  assertThrows(() => Intl.DateTimeFormat(undefined, {timeZone: 'Aurope/Paris'}));
-}

test/mjsunit/regress/regress-crbug-364374.js

+// Copyright 2015 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.
+
+if (this.Intl) {
+  // chromium:364374
+
+  // Locations with 2 underscores are accepted and normalized.
+  // 'of' and 'es' are always lowercased.
+  df = new Intl.DateTimeFormat('en-US', {'timeZone': 'eUrope/isLe_OF_man'})
+  assertEquals('Europe/Isle_of_Man', df.resolvedOptions().timeZone);
+
+  df = new Intl.DateTimeFormat('en-US', {'timeZone': 'africa/Dar_eS_salaam'})
+  assertEquals('Africa/Dar_es_Salaam', df.resolvedOptions().timeZone);
+
+  df = new Intl.DateTimeFormat('en-US', {'timeZone': 'America/port_of_spain'})
+  assertEquals('America/Port_of_Spain', df.resolvedOptions().timeZone);
+
+  // Zone ids with more than 2 parts are accepted and normalized.
+  df = new Intl.DateTimeFormat('en-US', {'timeZone': 'America/north_Dakota/new_salem'})
+  assertEquals('America/North_Dakota/New_Salem', df.resolvedOptions().timeZone);
+
+  // 3-part zone IDs are accepted and normalized.
+  // Two Buenose Aires aliases are identical.
+  df1 = new Intl.DateTimeFormat('en-US', {'timeZone': 'America/aRgentina/buenos_aIres'})
+  df2 = new Intl.DateTimeFormat('en-US', {'timeZone': 'America/Argentina/Buenos_Aires'})
+  assertEquals(df1.resolvedOptions().timeZone, df2.resolvedOptions().timeZone);
+
+  df2 = new Intl.DateTimeFormat('en-US', {'timeZone': 'America/Buenos_Aires'})
+  assertEquals(df1.resolvedOptions().timeZone, df2.resolvedOptions().timeZone);
+
+  df1 = new Intl.DateTimeFormat('en-US', {'timeZone': 'America/Indiana/Indianapolis'})
+  df2 = new Intl.DateTimeFormat('en-US', {'timeZone': 'America/Indianapolis'})
+  assertEquals(df1.resolvedOptions().timeZone, df2.resolvedOptions().timeZone);
+
+  // ICU does not recognize East-Indiana. Add later when it does.
+  // df2 = new Intl.DateTimeFormat('en-US', {'timeZone': 'America/East-Indiana'})
+  // assertEquals(df1.resolvedOptions().timeZone, df2.resolvedOptions().timeZone);
+
+
+  // Zone IDs with hyphens. 'au' has to be in lowercase.
+  df = new Intl.DateTimeFormat('en-US', {'timeZone': 'America/port-aU-pRince'})
+  assertEquals('America/Port-au-Prince', df.resolvedOptions().timeZone);
+
+  // Accepts Ho_Chi_Minh and treats it as identical to Saigon
+  df1 = new Intl.DateTimeFormat('en-US', {'timeZone': 'Asia/Ho_Chi_Minh'})
+  df2 = new Intl.DateTimeFormat('en-US', {'timeZone': 'Asia/Saigon'})
+  assertEquals(df1.resolvedOptions().timeZone, df2.resolvedOptions().timeZone);
+
+  // Throws for invalid timezone ids.
+  assertThrows(() => Intl.DateTimeFormat(undefined, {timeZone: 'Europe/_Paris'}));
+  assertThrows(() => Intl.DateTimeFormat(undefined, {timeZone: 'America/New__York'}));
+  assertThrows(() => Intl.DateTimeFormat(undefined, {timeZone: 'America//New_York'}));
+  assertThrows(() => Intl.DateTimeFormat(undefined, {timeZone: 'America/New_York_'}));
+  assertThrows(() => Intl.DateTimeFormat(undefined, {timeZone: 'America/New_Y0rk'}));
+}

test/mjsunit/regress/regress-crbug-487322.js

+// Copyright 2015 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.
+
+if (this.Intl) {
+  // Normalizes Kat{h,}mandu (chromium:487322)
+  // According to the IANA timezone db, Kathmandu is the current canonical
+  // name, but ICU got it backward. To make this test robust against a future
+  // ICU change ( http://bugs.icu-project.org/trac/ticket/12044 ),
+  // just check that Kat(h)mandu is resolved identically.
+  df1 = new Intl.DateTimeFormat('en-US', {'timeZone': 'Asia/Katmandu'})
+  df2 = new Intl.DateTimeFormat('en-US', {'timeZone': 'Asia/Kathmandu'})
+  assertEquals(df1.resolvedOptions().timeZone, df2.resolvedOptions().timeZone);
+
+  // Normalizes Ulan_Bator to Ulaanbaatar. Unlike Kat(h)mandu, ICU got this
+  // right so that we make sure that Ulan_Bator is resolved to Ulaanbaatar.
+  df = new Intl.DateTimeFormat('en-US', {'timeZone': 'Asia/Ulaanbaatar'})
+  assertEquals('Asia/Ulaanbaatar', df.resolvedOptions().timeZone);
+
+  df = new Intl.DateTimeFormat('en-US', {'timeZone': 'Asia/Ulan_Bator'})
+  assertEquals('Asia/Ulaanbaatar', df.resolvedOptions().timeZone);
+
+  // Throws for unsupported time zones.
+  assertThrows(() => Intl.DateTimeFormat(undefined, {timeZone: 'Aurope/Paris'}));
+}