[regexp] implement latest spec draft for property class.

See https://github.com/mathiasbynens/es-regexp-unicode-property-escapes

Changes:
- only allow General Category, binary properties, Script, and Script_Extensions.
- implement Script_Extensions.

R=littledan@chromium.org
BUG=v8:4743

Review-Url: https://codereview.chromium.org/2502933002
Cr-Commit-Position: refs/heads/master@{#41091}

src/regexp/regexp-parser.cc

@@ -1083,13 +1083,19 @@ bool IsExactPropertyValueAlias(const char* property_value_name,
 bool LookupPropertyValueName(UProperty property,
                              const char* property_value_name, bool negate,
                              ZoneList<CharacterRange>* result, Zone* zone) {
+  UProperty property_for_lookup = property;
+  if (property_for_lookup == UCHAR_SCRIPT_EXTENSIONS) {
+    // For the property Script_Extensions, we have to do the property value
+    // name lookup as if the property is Script.
+    property_for_lookup = UCHAR_SCRIPT;
+  }
   int32_t property_value =
-      u_getPropertyValueEnum(property, property_value_name);
+      u_getPropertyValueEnum(property_for_lookup, property_value_name);
   if (property_value == UCHAR_INVALID_CODE) return false;
 
   // We require the property name to match exactly to one of the property value
   // aliases. However, u_getPropertyValueEnum uses loose matching.
-  if (!IsExactPropertyValueAlias(property_value_name, property,
+  if (!IsExactPropertyValueAlias(property_value_name, property_for_lookup,
                                  property_value)) {
     return false;
   }
@@ -1197,9 +1203,14 @@ bool RegExpParser::ParsePropertyClass(ZoneList<CharacterRange>* result,
     const char* property_name = first_part.ToConstVector().start();
     const char* value_name = second_part.ToConstVector().start();
     UProperty property = u_getPropertyEnum(property_name);
-    if (property < UCHAR_INT_START) return false;
-    if (property >= UCHAR_INT_LIMIT) return false;
     if (!IsExactPropertyAlias(property_name, property)) return false;
+    if (property == UCHAR_GENERAL_CATEGORY) {
+      // We want to allow aggregate value names such as "Letter".
+      property = UCHAR_GENERAL_CATEGORY_MASK;
+    } else if (property != UCHAR_SCRIPT &&
+               property != UCHAR_SCRIPT_EXTENSIONS) {
+      return false;
+    }
     return LookupPropertyValueName(property, value_name, negate, result,
                                    zone());
   }

test/mjsunit/harmony/regexp-property-blocks.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: --harmony-regexp-property
-
-function t(re, s) { assertTrue(re.test(s)); }
-function f(re, s) { assertFalse(re.test(s)); }
-
-t(/\p{Block=ASCII}+/u, ".");
-t(/\p{Block=ASCII}+/u, "supercalifragilisticexpialidocious");
-t(/\p{Block=Basic_Latin}+/u, ".");
-t(/\p{Block=Basic_Latin}+/u, "supercalifragilisticexpialidocious");
-
-t(/\p{blk=CJK}+/u, "话说天下大势，分久必合，合久必分");
-t(/\p{blk=CJK_Unified_Ideographs}+/u, "吾庄后有一桃园，花开正盛");
-f(/\p{blk=CJK}+/u, "おはようございます");
-f(/\p{blk=CJK_Unified_Ideographs}+/u,
-  "Something is rotten in the state of Denmark");
-
-t(/\p{blk=Latin_1}+/u, "Wie froh bin ich, daß ich weg bin!");
-f(/\p{blk=Latin_1_Supplement}+/u, "奔腾千里荡尘埃，渡水登山紫雾开");
-f(/\p{blk=Latin_1_Sup}+/u, "いただきます");
-
-t(/\p{blk=Hiragana}/u, "いただきます");
-t(/\p{sc=Hiragana}/u, "\u{1b001}");   // This refers to the script "Hiragana".
-f(/\p{blk=Hiragana}/u, "\u{1b001}");  // This refers to the block "Hiragana".
-
-t(/\p{blk=Greek_And_Coptic}/u,
-  "ἄνδρα μοι ἔννεπε, μοῦσα, πολύτροπον, ὃς μάλα πολλὰ");
-t(/\p{blk=Greek}/u, "μῆνιν ἄειδε θεὰ Πηληϊάδεω Ἀχιλῆος");
-
-assertThrows("/\\p{In}/u");
-assertThrows("/\\pI/u");
-assertThrows("/\\p{I}/u");
-assertThrows("/\\p{CJK}/u");

test/mjsunit/harmony/regexp-property-enumerated.js

@@ -4,16 +4,13 @@
 
 // Flags: --harmony-regexp-property
 
-function t(re, s) { assertTrue(re.test(s)); }
-function f(re, s) { assertFalse(re.test(s)); }
+assertThrows("/\\p{Bidi_Class=L}+/u");
+assertThrows("/\\p{bc=Left_To_Right}+/u");
+assertThrows("/\\p{bc=AL}+/u");
+assertThrows("/\\p{bc=Arabic_Letter}+/u");
 
-t(/\p{Bidi_Class=L}+/u, "Is this the real life?");
-t(/\p{bc=Left_To_Right}+/u, "Is this just fantasy?");
-t(/\p{bc=AL}+/u, "السلام عليكم‎");
-t(/\p{bc=Arabic_Letter}+/u, "متشرف بمعرفتك‎");
-
-t(/\p{Line_Break=Glue}/u, "\u00A0");
-t(/\p{lb=AL}/u, "~");
+assertThrows("/\\p{Line_Break=Glue}/u");
+assertThrows("/\\p{lb=AL}/u");
 
 assertThrows("/\\p{Block=}/u");
 assertThrows("/\\p{=}/u");
@@ -22,7 +19,7 @@ assertThrows("/\\p{=Hiragana}/u");
 assertThrows("/\\p{Block=CJK=}/u");
 
 assertThrows("/\\p{Age=V8_0}/u");
-assertThrows("/\\p{General_Category=Letter}/u");
-assertThrows("/\\p{gc=L}/u");
+assertDoesNotThrow("/\\p{General_Category=Letter}/u");
+assertDoesNotThrow("/\\p{gc=L}/u");
 assertThrows("/\\p{General_Category_Mask=Letter}/u");
 assertThrows("/\\p{gcm=L}/u");

test/mjsunit/harmony/regexp-property-exact-match.js

@@ -36,7 +36,7 @@ assertThrows("/\\p{BidiClass=LeftToRight}/u");
 assertThrows("/\\p{BidiC=LeftToRight}/u");
 assertThrows("/\\p{bidi_c=Left_To_Right}/u");
 
-assertDoesNotThrow("/\\p{Block=CJK}/u");
+assertThrows("/\\p{Block=CJK}/u");
 assertThrows("/\\p{Block = CJK}/u");
 assertThrows("/\\p{Block=cjk}/u");
 assertThrows("/\\p{BLK=CJK}/u");

test/mjsunit/harmony/regexp-property-general-category.js

@@ -63,3 +63,7 @@ assertTrue(/\P{L}/u.test("\uA6EE"));
 
 assertTrue(/\p{Lowercase_Letter}/u.test("a"));
 assertTrue(/\p{Math_Symbol}/u.test("+"));
+
+assertTrue(/\p{gc=Ll}/u.test("a"));
+assertTrue(/\p{General_Category=Math_Symbol}/u.test("+"));
+assertTrue(/\p{General_Category=L}/u.test("X"));

test/mjsunit/harmony/regexp-property-invalid.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: --harmony-regexp-property
+
+assertThrows("/\p{Block=ASCII}+/u");
+assertThrows("/\p{Block=ASCII}+/u");
+assertThrows("/\p{Block=Basic_Latin}+/u");
+assertThrows("/\p{Block=Basic_Latin}+/u");
+
+assertThrows("/\p{blk=CJK}+/u");
+assertThrows("/\p{blk=CJK_Unified_Ideographs}+/u");
+assertThrows("/\p{blk=CJK}+/u");
+assertThrows("/\p{blk=CJK_Unified_Ideographs}+/u");
+
+assertThrows("/\p{Block=ASCII}+/u");
+assertThrows("/\p{Block=ASCII}+/u");
+assertThrows("/\p{Block=Basic_Latin}+/u");
+assertThrows("/\p{Block=Basic_Latin}+/u");
+
+assertThrows("/\p{NFKD_Quick_Check=Y}+/u");
+assertThrows("/\p{NFKD_QC=Yes}+/u");
+
+assertThrows("/\p{Numeric_Type=Decimal}+/u");
+assertThrows("/\p{nt=De}+/u");
+
+assertThrows("/\p{Bidi_Class=Arabic_Letter}+/u");
+assertThrows("/\p{Bidi_Class=AN}+/u");
+
+assertThrows("/\p{ccc=OV}+/u");
+
+assertThrows("/\p{Sentence_Break=Format}+/u");
+
+assertThrows("/\\p{In}/u");
+assertThrows("/\\pI/u");
+assertThrows("/\\p{I}/u");
+assertThrows("/\\p{CJK}/u");