Reland sticky regexps https://codereview.chromium.org/567313003/

src/regexp.js

 // Copyright 2012 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.
 
 // This file relies on the fact that the following declaration has been made
 // in runtime.js:
 // var $Object = global.Object;
 // var $Array = global.Array;
 
 var $RegExp = global.RegExp;
 
 // -------------------------------------------------------------------
 
 // A recursive descent parser for Patterns according to the grammar of
 // ECMA-262 15.10.1, with deviations noted below.
 function DoConstructRegExp(object, pattern, flags) {
   // RegExp : Called as constructor; see ECMA-262, section 15.10.4.
   if (IS_REGEXP(pattern)) {
     if (!IS_UNDEFINED(flags)) {
       throw MakeTypeError('regexp_flags', []);
     }
     flags = (pattern.global ? 'g' : '')
         + (pattern.ignoreCase ? 'i' : '')
         + (pattern.multiline ? 'm' : '');
+    if (harmony_regexps)
+        flags += (pattern.sticky ? 'y' : '');
     pattern = pattern.source;
   }
 
   pattern = IS_UNDEFINED(pattern) ? '' : ToString(pattern);
   flags = IS_UNDEFINED(flags) ? '' : ToString(flags);
 
   var global = false;
   var ignoreCase = false;
   var multiline = false;
+  var sticky = false;
   for (var i = 0; i < flags.length; i++) {
     var c = %_CallFunction(flags, i, StringCharAt);
     switch (c) {
       case 'g':
         if (global) {
           throw MakeSyntaxError("invalid_regexp_flags", [flags]);
         }
         global = true;
         break;
       case 'i':
         if (ignoreCase) {
           throw MakeSyntaxError("invalid_regexp_flags", [flags]);
         }
         ignoreCase = true;
         break;
       case 'm':
         if (multiline) {
           throw MakeSyntaxError("invalid_regexp_flags", [flags]);
         }
         multiline = true;
         break;
+      case 'y':
+        if (!harmony_regexps || sticky) {
+          throw MakeSyntaxError("invalid_regexp_flags", [flags]);
+        }
+        sticky = true;
+        break;
       default:
         throw MakeSyntaxError("invalid_regexp_flags", [flags]);
     }
   }
 
-  %RegExpInitializeObject(object, pattern, global, ignoreCase, multiline);
+  %RegExpInitializeObject(object, pattern, global, ignoreCase, multiline, sticky);
 
   // Call internal function to compile the pattern.
   %RegExpCompile(object, pattern, flags);
 }
 
 
 function RegExpConstructor(pattern, flags) {
   if (%_IsConstructCall()) {
     DoConstructRegExp(this, pattern, flags);
   } else {
@@ skipping 81 matching lines @@
                         ['RegExp.prototype.exec', this]);
   }
 
   string = TO_STRING_INLINE(string);
   var lastIndex = this.lastIndex;
 
   // Conversion is required by the ES5 specification (RegExp.prototype.exec
   // algorithm, step 5) even if the value is discarded for non-global RegExps.
   var i = TO_INTEGER(lastIndex);
 
-  var global = this.global;
-  if (global) {
+  var updateLastIndex = this.global || (harmony_regexps && this.sticky);
+  if (updateLastIndex) {
     if (i < 0 || i > string.length) {
       this.lastIndex = 0;
       return null;
     }
   } else {
     i = 0;
   }
 
   // matchIndices is either null or the lastMatchInfo array.
   var matchIndices = %_RegExpExec(this, string, i, lastMatchInfo);
 
   if (IS_NULL(matchIndices)) {
     this.lastIndex = 0;
     return null;
   }
 
   // Successful match.
   lastMatchInfoOverride = null;
-  if (global) {
+  if (updateLastIndex) {
     this.lastIndex = lastMatchInfo[CAPTURE1];
   }
   RETURN_NEW_RESULT_FROM_MATCH_INFO(matchIndices, string);
 }
 
 
 // One-element cache for the simplified test regexp.
 var regexp_key;
 var regexp_val;
 
 // Section 15.10.6.3 doesn't actually make sense, but the intention seems to be
 // that test is defined in terms of String.prototype.exec. However, it probably
 // means the original value of String.prototype.exec, which is what everybody
 // else implements.
 function RegExpTest(string) {
   if (!IS_REGEXP(this)) {
     throw MakeTypeError('incompatible_method_receiver',
                         ['RegExp.prototype.test', this]);
   }
   string = TO_STRING_INLINE(string);
 
   var lastIndex = this.lastIndex;
 
   // Conversion is required by the ES5 specification (RegExp.prototype.exec
   // algorithm, step 5) even if the value is discarded for non-global RegExps.
   var i = TO_INTEGER(lastIndex);
 
-  if (this.global) {
+  if (this.global || (harmony_regexps && this.sticky)) {
     if (i < 0 || i > string.length) {
       this.lastIndex = 0;
       return false;
     }
     // matchIndices is either null or the lastMatchInfo array.
     var matchIndices = %_RegExpExec(this, string, i, lastMatchInfo);
     if (IS_NULL(matchIndices)) {
       this.lastIndex = 0;
       return false;
     }
     lastMatchInfoOverride = null;
     this.lastIndex = lastMatchInfo[CAPTURE1];
     return true;
   } else {
-    // Non-global regexp.
-    // Remove irrelevant preceeding '.*' in a non-global test regexp.
-    // The expression checks whether this.source starts with '.*' and
-    // that the third char is not a '?'.
+    // Non-global, non-sticky regexp.
+    // Remove irrelevant preceeding '.*' in a test regexp.  The expression
+    // checks whether this.source starts with '.*' and that the third char is
+    // not a '?'.  But see https://code.google.com/p/v8/issues/detail?id=3560
     var regexp = this;
-    if (%_StringCharCodeAt(regexp.source, 0) == 46 &&  // '.'
+    if (regexp.source.length >= 3 &&
+        %_StringCharCodeAt(regexp.source, 0) == 46 &&  // '.'
         %_StringCharCodeAt(regexp.source, 1) == 42 &&  // '*'
         %_StringCharCodeAt(regexp.source, 2) != 63) {  // '?'
       regexp = TrimRegExp(regexp);
     }
     // matchIndices is either null or the lastMatchInfo array.
     var matchIndices = %_RegExpExec(regexp, string, 0, lastMatchInfo);
     if (IS_NULL(matchIndices)) {
       this.lastIndex = 0;
       return false;
     }
@@ skipping 16 matching lines @@
 
 function RegExpToString() {
   if (!IS_REGEXP(this)) {
     throw MakeTypeError('incompatible_method_receiver',
                         ['RegExp.prototype.toString', this]);
   }
   var result = '/' + this.source + '/';
   if (this.global) result += 'g';
   if (this.ignoreCase) result += 'i';
   if (this.multiline) result += 'm';
+  if (harmony_regexps && this.sticky) result += 'y';
   return 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
 // of the last successful match.
 function RegExpGetLastMatch() {
   if (lastMatchInfoOverride !== null) {
@@ skipping 178 matching lines @@
 
   for (var i = 1; i < 10; ++i) {
     %DefineAccessorPropertyUnchecked($RegExp, '$' + i,
                                      RegExpMakeCaptureGetter(i), NoOpSetter,
                                      DONT_DELETE);
   }
   %ToFastProperties($RegExp);
 }
 
 SetUpRegExp();