Merge pull request #411 from nicolo-ribaudo/regexp-es2015

Make RegExp#[Symbol.*] methods call exec

packages/core-js/internals/advance-string-index.js

@@ -0,0 +1,8 @@
+'use strict';
+var at = require('../internals/string-at')(true);
+
+// `AdvanceStringIndex` abstract operation
+// https://tc39.github.io/ecma262/#sec-advancestringindex
+module.exports = function (S, index, unicode) {
+  return index + (unicode ? at(S, index).length : 1);
+};

packages/core-js/internals/fix-regexp-well-known-symbol-logic.js

@@ -5,24 +5,43 @@ var fails = require('../internals/fails');
 var requireObjectCoercible = require('../internals/require-object-coercible');
 var wellKnownSymbol = require('../internals/well-known-symbol');
 
-module.exports = function (KEY, length, exec) {
+var SPECIES = wellKnownSymbol('species');
+
+module.exports = function (KEY, length, exec, sham) {
   var SYMBOL = wellKnownSymbol(KEY);
   var methods = exec(requireObjectCoercible, SYMBOL, ''[KEY]);
   var stringMethod = methods[0];
   var regexMethod = methods[1];
   if (fails(function () {
+    // String methods call symbol-named RegEp methods
     var O = {};
     O[SYMBOL] = function () { return 7; };
     return ''[KEY](O) != 7;
+  }) || fails(function () {
+    // Symbol-named RegExp methods call .exec
+    var execCalled = false;
+    var re = /a/;
+    re.exec = function () { execCalled = true; return null; };
+
+    if (KEY === 'split') {
+      // RegExp[@@split] doesn't call the regex's exec method, but first creates
+      // a new one. We need to return the patched regex when creating the new one.
+      re.constructor = {};
+      re.constructor[SPECIES] = function () { return re; };
+    }
+
+    re[SYMBOL]('');
+    return !execCalled;
   })) {
     redefine(String.prototype, KEY, stringMethod);
-    hide(RegExp.prototype, SYMBOL, length == 2
+    redefine(RegExp.prototype, SYMBOL, length == 2
       // 21.2.5.8 RegExp.prototype[@@replace](string, replaceValue)
       // 21.2.5.11 RegExp.prototype[@@split](string, limit)
       ? function (string, arg) { return regexMethod.call(string, this, arg); }
       // 21.2.5.6 RegExp.prototype[@@match](string)
       // 21.2.5.9 RegExp.prototype[@@search](string)
       : function (string) { return regexMethod.call(string, this); }
     );
+    if (sham) hide(RegExp.prototype[SYMBOL], 'sham', true);
   }
 };

packages/core-js/internals/regexp-exec.js

@@ -0,0 +1,22 @@
+var classof = require('../internals/classof-raw');
+var builtinExec = RegExp.prototype.exec;
+
+// `RegExpExec` abstract operation
+// https://tc39.github.io/ecma262/#sec-regexpexec
+module.exports = function (R, S) {
+  var exec = R.exec;
+  if (typeof exec === 'function') {
+    var result = exec.call(R, S);
+    if (typeof result !== 'object') {
+      throw new TypeError('RegExp exec method returned something other than an Object or null');
+    }
+    return result;
+  }
+
+  if (classof(R) !== 'RegExp') {
+    throw new TypeError('RegExp#exec called on incompatible receiver');
+  }
+
+  return builtinExec.call(R, S);
+};
+

packages/core-js/modules/es.string.match.js

@@ -1,11 +1,43 @@
 'use strict';
+
+var anObject = require('../internals/an-object');
+var toLength = require('../internals/to-length');
+var advanceStringIndex = require('../internals/advance-string-index');
+var regExpExec = require('../internals/regexp-exec');
+var nativeExec = RegExp.prototype.exec;
+
 // @@match logic
 require('../internals/fix-regexp-well-known-symbol-logic')('match', 1, function (defined, MATCH, nativeMatch) {
-  // `String.prototype.match` method
-  // https://tc39.github.io/ecma262/#sec-string.prototype.match
-  return [function match(regexp) {
-    var O = defined(this);
-    var matcher = regexp == undefined ? undefined : regexp[MATCH];
-    return matcher !== undefined ? matcher.call(regexp, O) : new RegExp(regexp)[MATCH](String(O));
-  }, nativeMatch];
+  return [
+    // `String.prototype.match` method
+    // https://tc39.github.io/ecma262/#sec-string.prototype.match
+    function match(regexp) {
+      var O = defined(this);
+      var matcher = regexp == undefined ? undefined : regexp[MATCH];
+      return matcher !== undefined ? matcher.call(regexp, O) : new RegExp(regexp)[MATCH](String(O));
+    },
+    // `RegExp.prototype[@@match]` method
+    // https://tc39.github.io/ecma262/#sec-regexp.prototype-@@match
+    function (regexp) {
+      if (regexp.exec === nativeExec) return nativeMatch.call(this, regexp);
+
+      var rx = anObject(regexp);
+      var S = String(this);
+
+      if (!rx.global) return regExpExec(rx, S);
+
+      var fullUnicode = rx.unicode;
+      rx.lastIndex = 0;
+      var A = [];
+      var n = 0;
+      var result;
+      while ((result = regExpExec(rx, S)) !== null) {
+        var matchStr = String(result[0]);
+        A[n] = matchStr;
+        if (matchStr === '') rx.lastIndex = advanceStringIndex(S, toLength(rx.lastIndex), fullUnicode);
+        n++;
+      }
+      return n === 0 ? null : A;
+    }
+  ];
 });

packages/core-js/modules/es.string.replace.js

@@ -1,13 +1,118 @@
 'use strict';
+
+var anObject = require('../internals/an-object');
+var toObject = require('../internals/to-object');
+var toLength = require('../internals/to-length');
+var toInteger = require('../internals/to-integer');
+var advanceStringIndex = require('../internals/advance-string-index');
+var regExpExec = require('../internals/regexp-exec');
+var nativeExec = RegExp.prototype.exec;
+var max = Math.max;
+var min = Math.min;
+var floor = Math.floor;
+var SUBSTITUTION_SYMBOLS = /\$([$&`']|\d\d?|<[^>]*>)/g;
+var SUBSTITUTION_SYMBOLS_NO_NAMED = /\$([$&`']|\d\d?)/g;
+
+var maybeToString = function (it) {
+  return it === undefined ? it : String(it);
+};
+
 // @@replace logic
 require('../internals/fix-regexp-well-known-symbol-logic')('replace', 2, function (defined, REPLACE, nativeReplace) {
-  // `String.prototype.replace` method
-  // https://tc39.github.io/ecma262/#sec-string.prototype.replace
-  return [function replace(searchValue, replaceValue) {
-    var O = defined(this);
-    var replacer = searchValue == undefined ? undefined : searchValue[REPLACE];
-    return replacer !== undefined
+  return [
+    // `String.prototype.replace` method
+    // https://tc39.github.io/ecma262/#sec-string.prototype.replace
+    function replace(searchValue, replaceValue) {
+      var O = defined(this);
+      var replacer = searchValue == undefined ? undefined : searchValue[REPLACE];
+      return replacer !== undefined
       ? replacer.call(searchValue, O, replaceValue)
       : nativeReplace.call(String(O), searchValue, replaceValue);
-  }, nativeReplace];
+    },
+    // `RegExp.prototype[@@replace]` method
+    // https://tc39.github.io/ecma262/#sec-regexp.prototype-@@replace
+    function (regexp, replaceValue) {
+      if (regexp.exec === nativeExec) return nativeReplace.call(this, regexp, replaceValue);
+
+      var rx = anObject(regexp);
+      var S = String(this);
+
+      var functionalReplace = typeof replaceValue === 'function';
+      if (!functionalReplace) replaceValue = String(replaceValue);
+
+      var global = rx.global;
+      if (global) {
+        var fullUnicode = rx.unicode;
+        rx.lastIndex = 0;
+      }
+      var results = [];
+      while (true) {
+        var result = regExpExec(rx, S);
+        if (result === null) break;
+
+        results.push(result);
+        if (!global) break;
+
+        var matchStr = String(result[0]);
+        if (matchStr === '') rx.lastIndex = advanceStringIndex(S, toLength(rx.lastIndex), fullUnicode);
+      }
+
+      var accumulatedResult = '';
+      var nextSourcePosition = 0;
+      for (var i = 0; i < results.length; i++) {
+        result = results[i];
+
+        var matched = String(result[0]);
+        var position = max(min(toInteger(result.index), S.length), 0);
+        var captures = result.slice(1).map(maybeToString);
+        var namedCaptures = result.groups;
+        if (functionalReplace) {
+          var replacerArgs = [matched].concat(captures, position, S);
+          if (namedCaptures !== undefined) replacerArgs.push(namedCaptures);
+          var replacement = String(replaceValue.apply(undefined, replacerArgs));
+        } else {
+          replacement = getSubstitution(matched, S, position, captures, namedCaptures, replaceValue);
+        }
+        if (position >= nextSourcePosition) {
+          accumulatedResult += S.slice(nextSourcePosition, position) + replacement;
+          nextSourcePosition = position + matched.length;
+        }
+      }
+      return accumulatedResult + S.slice(nextSourcePosition);
+    }
+  ];
+
+  // https://tc39.github.io/ecma262/#sec-getsubstitution
+  function getSubstitution(matched, str, position, captures, namedCaptures, replacement) {
+    var tailPos = position + matched.length;
+    var m = captures.length;
+    var symbols = SUBSTITUTION_SYMBOLS_NO_NAMED;
+    if (namedCaptures !== undefined) {
+      namedCaptures = toObject(namedCaptures);
+      symbols = SUBSTITUTION_SYMBOLS;
+    }
+    return nativeReplace.call(replacement, symbols, function (match, ch) {
+      var capture;
+      switch (ch[0]) {
+        case '$': return '$';
+        case '&': return matched;
+        case '`': return str.slice(0, position);
+        case "'": return str.slice(tailPos);
+        case '<':
+          capture = namedCaptures[ch.slice(1, -1)];
+          break;
+        default: // \d\d?
+          var n = +ch;
+          if (n === 0) return ch;
+          if (n > m) {
+            var f = floor(n / 10);
+            if (f === 0) return ch;
+            if (f <= m) return captures[f - 1] === undefined ? ch[1] : captures[f - 1] + ch[1];
+            return ch;
+          }
+          capture = captures[n - 1];
+      }
+      return capture === undefined ? '' : capture;
+    });
+  }
 });

packages/core-js/modules/es.string.search.js

@@ -1,11 +1,33 @@
 'use strict';
+
+var anObject = require('../internals/an-object');
+var sameValue = require('../internals/same-value');
+var regExpExec = require('../internals/regexp-exec');
+var nativeExec = RegExp.prototype.exec;
+
 // @@search logic
 require('../internals/fix-regexp-well-known-symbol-logic')('search', 1, function (defined, SEARCH, nativeSearch) {
-  // `String.prototype.search` method
-  // https://tc39.github.io/ecma262/#sec-string.prototype.search
-  return [function search(regexp) {
-    var O = defined(this);
-    var searcher = regexp == undefined ? undefined : regexp[SEARCH];
-    return searcher !== undefined ? searcher.call(regexp, O) : new RegExp(regexp)[SEARCH](String(O));
-  }, nativeSearch];
+  return [
+    // `String.prototype.search` method
+    // https://tc39.github.io/ecma262/#sec-string.prototype.search
+    function search(regexp) {
+      var O = defined(this);
+      var searcher = regexp == undefined ? undefined : regexp[SEARCH];
+      return searcher !== undefined ? searcher.call(regexp, O) : new RegExp(regexp)[SEARCH](String(O));
+    },
+    // `RegExp.prototype[@@search]` method
+    // https://tc39.github.io/ecma262/#sec-regexp.prototype-@@search
+    function (regexp) {
+      if (regexp.exec === nativeExec) return nativeSearch.call(this, regexp);
+
+      var rx = anObject(regexp);
+      var S = String(this);
+
+      var previousLastIndex = rx.lastIndex;
+      if (!sameValue(previousLastIndex, 0)) rx.lastIndex = 0;
+      var result = regExpExec(rx, S);
+      if (!sameValue(rx.lastIndex, previousLastIndex)) rx.lastIndex = previousLastIndex;
+      return result === null ? -1 : result.index;
+    }
+  ];
 });

packages/core-js/modules/es.string.split.js

@@ -1,10 +1,22 @@
 'use strict';
+
+var isRegExp = require('../internals/is-regexp');
+var anObject = require('../internals/an-object');
+var speciesConstructor = require('../internals/species-constructor');
+var advanceStringIndex = require('../internals/advance-string-index');
+var toLength = require('../internals/to-length');
+var regExpExec = require('../internals/regexp-exec');
+var nativeExec = RegExp.prototype.exec;
+var arrayPush = [].push;
+var min = Math.min;
+var LENGTH = 'length';
+
+// eslint-disable-next-line no-empty
+var SUPPORTS_Y = !!(function () { try { return new RegExp('x', 'y'); } catch (e) {} })();
+
 // @@split logic
 require('../internals/fix-regexp-well-known-symbol-logic')('split', 2, function (defined, SPLIT, nativeSplit) {
-  var isRegExp = require('../internals/is-regexp');
   var internalSplit = nativeSplit;
-  var arrayPush = [].push;
-  var LENGTH = 'length';
   if (
     'abbc'.split(/(b)*/)[1] == 'c' ||
     'test'.split(/(?:)/, -1)[LENGTH] != 4 ||
@@ -62,13 +74,68 @@ require('../internals/fix-regexp-well-known-symbol-logic')('split', 2, function
       return separator === undefined && limit === 0 ? [] : nativeSplit.call(this, separator, limit);
     };
   }
-  // `String.prototype.split` method
-  // https://tc39.github.io/ecma262/#sec-string.prototype.split
-  return [function split(separator, limit) {
-    var O = defined(this);
-    var splitter = separator == undefined ? undefined : separator[SPLIT];
-    return splitter !== undefined
+
+  return [
+    // `String.prototype.split` method
+    // https://tc39.github.io/ecma262/#sec-string.prototype.split
+    function split(separator, limit) {
+      var O = defined(this);
+      var splitter = separator == undefined ? undefined : separator[SPLIT];
+      return splitter !== undefined
       ? splitter.call(separator, O, limit)
       : internalSplit.call(String(O), separator, limit);
-  }, internalSplit];
-});
+    },
+    // `RegExp.prototype[@@split]` method
+    // https://tc39.github.io/ecma262/#sec-regexp.prototype-@@split
+    //
+    // NOTE: This cannot be properly polyfilled in engines that don't support
+    // the 'y' flag.
+    function (regexp, limit) {
+      // We can never use `internalSplit` if exec has been changed, because
+      // internalSplit contains workarounds for things which might have been
+      // purposely changed by the developer.
+      if (regexp.exec === nativeExec) return internalSplit.call(this, regexp, limit);
+
+      var rx = anObject(regexp);
+      var S = String(this);
+      var C = speciesConstructor(rx, RegExp);
+
+      var unicodeMatching = rx.unicode;
+      var flags = (rx.ignoreCase ? 'i' : '') +
+                    (rx.multiline ? 'm' : '') +
+                    (rx.unicode ? 'u' : '') +
+                    (SUPPORTS_Y ? 'y' : 'g');
+
+      // ^(? + rx + ) is needed, in combination with some S slicing, to
+      // simulate the 'y' flag.
+      var splitter = new C(SUPPORTS_Y ? rx : '^(?:' + rx.source + ')', flags);
+      var lim = limit === undefined ? 0xffffffff : limit >>> 0;
+      if (lim === 0) return [];
+      if (S.length === 0) return regExpExec(splitter, S) === null ? [S] : [];
+      var p = 0;
+      var q = 0;
+      var A = [];
+      while (q < S.length) {
+        splitter.lastIndex = SUPPORTS_Y ? q : 0;
+        var z = regExpExec(splitter, SUPPORTS_Y ? S : S.slice(q));
+        var e;
+        if (
+          z === null ||
+          (e = min(toLength(splitter.lastIndex + (SUPPORTS_Y ? 0 : q)), S.length)) === p
+        ) {
+          q = advanceStringIndex(S, q, unicodeMatching);
+        } else {
+          A.push(S.slice(p, q));
+          if (A.length === lim) return A;
+          for (var i = 1; i <= z.length - 1; i++) {
+            A.push(z[i]);
+            if (A.length === lim) return A;
+          }
+          q = p = e;
+        }
+      }
+      A.push(S.slice(p));
+      return A;
+    }
+  ];
+}, !SUPPORTS_Y);