[es6] Implement @@split subclassing.

RegExp.prototye[@@split] is not yet implement to spec regarding creating
new RegExp object with the SpeciesConstructor.

R=littledan@chromium.org
BUG=v8:4345
LOG=N

Review URL: https://codereview.chromium.org/1427573005

Cr-Commit-Position: refs/heads/master@{#31911}

src/js/regexp.js

@@ -12,8 +12,10 @@
 var FLAG_harmony_tolength;
 var GlobalObject = global.Object;
 var GlobalRegExp = global.RegExp;
+var InternalArray = utils.InternalArray;
 var InternalPackedArray = utils.InternalPackedArray;
 var MakeTypeError;
+var splitSymbol = utils.ImportNow("split_symbol");
 
 utils.ImportFromExperimental(function(from) {
   FLAG_harmony_tolength = from.FLAG_harmony_tolength;
@@ -268,6 +270,76 @@ function RegExpToString() {
 }
 
 
+// ES6 21.2.5.11.
+function RegExpSplit(string, limit) {
+  // TODO(yangguo): allow non-regexp receivers.
+  if (!IS_REGEXP(this)) {
+    throw MakeTypeError(kIncompatibleMethodReceiver,
+                        "RegExp.prototype.@@split", this);
+  }
+  var separator = this;
+  var subject = TO_STRING(string);
+
+  limit = (IS_UNDEFINED(limit)) ? kMaxUint32 : TO_UINT32(limit);
+  var length = subject.length;
+
+  if (limit === 0) return [];
+
+  if (length === 0) {
+    if (DoRegExpExec(separator, subject, 0, 0) !== null) return [];
+    return [subject];
+  }
+
+  var currentIndex = 0;
+  var startIndex = 0;
+  var startMatch = 0;
+  var result = new InternalArray();
+
+  outer_loop:
+  while (true) {
+    if (startIndex === length) {
+      result[result.length] = %_SubString(subject, currentIndex, length);
+      break;
+    }
+
+    var matchInfo = DoRegExpExec(separator, subject, startIndex);
+    if (matchInfo === null || length === (startMatch = matchInfo[CAPTURE0])) {
+      result[result.length] = %_SubString(subject, currentIndex, length);
+      break;
+    }
+    var endIndex = matchInfo[CAPTURE1];
+
+    // We ignore a zero-length match at the currentIndex.
+    if (startIndex === endIndex && endIndex === currentIndex) {
+      startIndex++;
+      continue;
+    }
+
+    result[result.length] = %_SubString(subject, currentIndex, startMatch);
+
+    if (result.length === limit) break;
+
+    var matchinfo_len = NUMBER_OF_CAPTURES(matchInfo) + REGEXP_FIRST_CAPTURE;
+    for (var i = REGEXP_FIRST_CAPTURE + 2; i < matchinfo_len; ) {
+      var start = matchInfo[i++];
+      var end = matchInfo[i++];
+      if (end != -1) {
+        result[result.length] = %_SubString(subject, start, end);
+      } else {
+        result[result.length] = UNDEFINED;
+      }
+      if (result.length === limit) break outer_loop;
+    }
+
+    startIndex = currentIndex = endIndex;
+  }
+
+  var array_result = [];
+  %MoveArrayContents(result, array_result);
+  return array_result;
+}
+
+
 // Getters for the static properties lastMatch, lastParen, leftContext, and
 // rightContext of the RegExp constructor.  The properties are computed based
 // on the captures array of the last successful match and the subject string
@@ -384,7 +456,8 @@ utils.InstallFunctions(GlobalRegExp.prototype, DONT_ENUM, [
   "exec", RegExpExecJS,
   "test", RegExpTest,
   "toString", RegExpToString,
-  "compile", RegExpCompileJS
+  "compile", RegExpCompileJS,
+  splitSymbol, RegExpSplit,
 ]);
 
 utils.InstallGetter(GlobalRegExp.prototype, 'global', RegExpGetGlobal);

src/js/string.js

@@ -20,6 +20,7 @@ var MakeTypeError;
 var RegExpExec;
 var RegExpExecNoTests;
 var RegExpLastMatchInfo;
+var splitSymbol = utils.ImportNow("split_symbol");
 
 utils.Import(function(from) {
   ArrayIndexOf = from.ArrayIndexOf;
@@ -554,95 +555,38 @@ function StringSlice(start, end) {
 }
 
 
-// ECMA-262 section 15.5.4.14
+// ES6 21.1.3.17.
 function StringSplitJS(separator, limit) {
   CHECK_OBJECT_COERCIBLE(this, "String.prototype.split");
 
-  var subject = TO_STRING(this);
-  limit = (IS_UNDEFINED(limit)) ? 0xffffffff : TO_UINT32(limit);
-
-  var length = subject.length;
-  if (!IS_REGEXP(separator)) {
-    var separator_string = TO_STRING(separator);
-
-    if (limit === 0) return [];
-
-    // ECMA-262 says that if separator is undefined, the result should
-    // be an array of size 1 containing the entire string.
-    if (IS_UNDEFINED(separator)) return [subject];
-
-    var separator_length = separator_string.length;
-
-    // If the separator string is empty then return the elements in the subject.
-    if (separator_length === 0) return %StringToArray(subject, limit);
-
-    var result = %StringSplit(subject, separator_string, limit);
-
-    return result;
-  }
-
-  if (limit === 0) return [];
-
-  // Separator is a regular expression.
-  return StringSplitOnRegExp(subject, separator, limit, length);
-}
-
-
-function StringSplitOnRegExp(subject, separator, limit, length) {
-  if (length === 0) {
-    if (RegExpExec(separator, subject, 0, 0) != null) {
-      return [];
+  if (!IS_NULL_OR_UNDEFINED(separator)) {
+    var splitter = separator[splitSymbol];
+    if (!IS_UNDEFINED(splitter)) {
+      if (!IS_CALLABLE(splitter)) {
+        throw MakeTypeError(kCalledNonCallable, splitter);
+      }
+      return %_Call(splitter, separator, this, limit);
     }
-    return [subject];
   }
 
-  var currentIndex = 0;
-  var startIndex = 0;
-  var startMatch = 0;
-  var result = new InternalArray();
-
-  outer_loop:
-  while (true) {
-
-    if (startIndex === length) {
-      result[result.length] = %_SubString(subject, currentIndex, length);
-      break;
-    }
+  var subject = TO_STRING(this);
+  limit = (IS_UNDEFINED(limit)) ? kMaxUint32 : TO_UINT32(limit);
 
-    var matchInfo = RegExpExec(separator, subject, startIndex);
-    if (matchInfo == null || length === (startMatch = matchInfo[CAPTURE0])) {
-      result[result.length] = %_SubString(subject, currentIndex, length);
-      break;
-    }
-    var endIndex = matchInfo[CAPTURE1];
+  var length = subject.length;
+  var separator_string = TO_STRING(separator);
 
-    // We ignore a zero-length match at the currentIndex.
-    if (startIndex === endIndex && endIndex === currentIndex) {
-      startIndex++;
-      continue;
-    }
+  if (limit === 0) return [];
 
-    result[result.length] = %_SubString(subject, currentIndex, startMatch);
+  // ECMA-262 says that if separator is undefined, the result should
+  // be an array of size 1 containing the entire string.
+  if (IS_UNDEFINED(separator)) return [subject];
 
-    if (result.length === limit) break;
+  var separator_length = separator_string.length;
 
-    var matchinfo_len = NUMBER_OF_CAPTURES(matchInfo) + REGEXP_FIRST_CAPTURE;
-    for (var i = REGEXP_FIRST_CAPTURE + 2; i < matchinfo_len; ) {
-      var start = matchInfo[i++];
-      var end = matchInfo[i++];
-      if (end != -1) {
-        result[result.length] = %_SubString(subject, start, end);
-      } else {
-        result[result.length] = UNDEFINED;
-      }
-      if (result.length === limit) break outer_loop;
-    }
+  // If the separator string is empty then return the elements in the subject.
+  if (separator_length === 0) return %StringToArray(subject, limit);
 
-    startIndex = currentIndex = endIndex;
-  }
-  var array_result = [];
-  %MoveArrayContents(result, array_result);
-  return array_result;
+  return %StringSplit(subject, separator_string, limit);
 }
 
 

test/mjsunit/harmony/string-split.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.
+
+// Flags: --harmony-regexp-subclass
+
+var pattern = {};
+var limit = { value: 3 };
+pattern[Symbol.split] = function(string, limit) {
+  return string.length * limit.value;
+};
+// Check object coercible fails.
+assertThrows(() => String.prototype.split.call(null, pattern, limit),
+             TypeError);
+// Override is called.
+assertEquals(15, "abcde".split(pattern, limit));
+// Non-callable override.
+pattern[Symbol.split] = "dumdidum";
+assertThrows(() => "abcde".split(pattern, limit), TypeError);
+
+assertEquals("[Symbol.split]", RegExp.prototype[Symbol.split].name);