runtime/RegExpConstructor.cpp

   1 /*
   2  *  Copyright (C) 1999-2000 Harri Porten (porten@kde.org)
   3  *  Copyright (C) 2003, 2007, 2008 Apple Inc. All Rights Reserved.
   4  *  Copyright (C) 2009 Torch Mobile, Inc.
   5  *
   6  *  This library is free software; you can redistribute it and/or
   7  *  modify it under the terms of the GNU Lesser General Public
   8  *  License as published by the Free Software Foundation; either
   9  *  version 2 of the License, or (at your option) any later version.
  10  *
  11  *  This library is distributed in the hope that it will be useful,
  12  *  but WITHOUT ANY WARRANTY; without even the implied warranty of
  13  *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  14  *  Lesser General Public License for more details.
  15  *
  16  *  You should have received a copy of the GNU Lesser General Public
  17  *  License along with this library; if not, write to the Free Software
  18  *  Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301  USA
  19  *
  20  */
  21
  22 #include "config.h"
  23 #include "RegExpConstructor.h"
  24
  25 #include "ArrayPrototype.h"
  26 #include "Error.h"
  27 #include "ExceptionHelpers.h"
  28 #include "JSArray.h"
  29 #include "JSFunction.h"
  30 #include "JSString.h"
  31 #include "Lookup.h"
  32 #include "ObjectPrototype.h"
  33 #include "RegExpMatchesArray.h"
  34 #include "RegExpObject.h"
  35 #include "RegExpPrototype.h"
  36 #include "RegExp.h"
  37 #include "RegExpCache.h"
  38 #include "UStringConcatenate.h"
  39 #include <wtf/PassOwnPtr.h>
  40
  41 #include <wtf/PassOwnPtr.h>
  42
  43 namespace JSC {
  44
  45 static JSValue regExpConstructorInput(ExecState*, JSValue, const Identifier&);
  46 static JSValue regExpConstructorMultiline(ExecState*, JSValue, const Identifier&);
  47 static JSValue regExpConstructorLastMatch(ExecState*, JSValue, const Identifier&);
  48 static JSValue regExpConstructorLastParen(ExecState*, JSValue, const Identifier&);
  49 static JSValue regExpConstructorLeftContext(ExecState*, JSValue, const Identifier&);
  50 static JSValue regExpConstructorRightContext(ExecState*, JSValue, const Identifier&);
  51 static JSValue regExpConstructorDollar1(ExecState*, JSValue, const Identifier&);
  52 static JSValue regExpConstructorDollar2(ExecState*, JSValue, const Identifier&);
  53 static JSValue regExpConstructorDollar3(ExecState*, JSValue, const Identifier&);
  54 static JSValue regExpConstructorDollar4(ExecState*, JSValue, const Identifier&);
  55 static JSValue regExpConstructorDollar5(ExecState*, JSValue, const Identifier&);
  56 static JSValue regExpConstructorDollar6(ExecState*, JSValue, const Identifier&);
  57 static JSValue regExpConstructorDollar7(ExecState*, JSValue, const Identifier&);
  58 static JSValue regExpConstructorDollar8(ExecState*, JSValue, const Identifier&);
  59 static JSValue regExpConstructorDollar9(ExecState*, JSValue, const Identifier&);
  60
  61 static void setRegExpConstructorInput(ExecState*, JSObject*, JSValue);
  62 static void setRegExpConstructorMultiline(ExecState*, JSObject*, JSValue);
  63
  64 } // namespace JSC
  65
  66 #include "RegExpConstructor.lut.h"
  67
  68 namespace JSC {
  69
  70 ASSERT_CLASS_FITS_IN_CELL(RegExpConstructor);
  71
  72 const ClassInfo RegExpConstructor::s_info = { "Function", &InternalFunction::s_info, 0, ExecState::regExpConstructorTable };
  73
  74 /* Source for RegExpConstructor.lut.h
  75 @begin regExpConstructorTable
  76     input           regExpConstructorInput          None
  77     $_              regExpConstructorInput          DontEnum
  78     multiline       regExpConstructorMultiline      None
  79     $*              regExpConstructorMultiline      DontEnum
  80     lastMatch       regExpConstructorLastMatch      DontDelete|ReadOnly
  81     $&              regExpConstructorLastMatch      DontDelete|ReadOnly|DontEnum
  82     lastParen       regExpConstructorLastParen      DontDelete|ReadOnly
  83     $+              regExpConstructorLastParen      DontDelete|ReadOnly|DontEnum
  84     leftContext     regExpConstructorLeftContext    DontDelete|ReadOnly
  85     $`              regExpConstructorLeftContext    DontDelete|ReadOnly|DontEnum
  86     rightContext    regExpConstructorRightContext   DontDelete|ReadOnly
  87     $'              regExpConstructorRightContext   DontDelete|ReadOnly|DontEnum
  88     $1              regExpConstructorDollar1        DontDelete|ReadOnly
  89     $2              regExpConstructorDollar2        DontDelete|ReadOnly
  90     $3              regExpConstructorDollar3        DontDelete|ReadOnly
  91     $4              regExpConstructorDollar4        DontDelete|ReadOnly
  92     $5              regExpConstructorDollar5        DontDelete|ReadOnly
  93     $6              regExpConstructorDollar6        DontDelete|ReadOnly
  94     $7              regExpConstructorDollar7        DontDelete|ReadOnly
  95     $8              regExpConstructorDollar8        DontDelete|ReadOnly
  96     $9              regExpConstructorDollar9        DontDelete|ReadOnly
  97 @end
  98 */
  99
 100 RegExpConstructor::RegExpConstructor(ExecState* exec, JSGlobalObject* globalObject, Structure* structure, RegExpPrototype* regExpPrototype)
 101     : InternalFunction(&exec->globalData(), globalObject, structure, Identifier(exec, "RegExp"))
 102     , d(adoptPtr(new RegExpConstructorPrivate))
 103 {
 104     ASSERT(inherits(&s_info));
 105
 106     // ECMA 15.10.5.1 RegExp.prototype
 107     putDirectWithoutTransition(exec->globalData(), exec->propertyNames().prototype, regExpPrototype, DontEnum | DontDelete | ReadOnly);
 108
 109     // no. of arguments for constructor
 110     putDirectWithoutTransition(exec->globalData(), exec->propertyNames().length, jsNumber(2), ReadOnly | DontDelete | DontEnum);
 111 }
 112
 113 RegExpMatchesArray::RegExpMatchesArray(ExecState* exec, RegExpConstructorPrivate* data)
 114     : JSArray(exec->globalData(), exec->lexicalGlobalObject()->regExpMatchesArrayStructure(), data->lastNumSubPatterns + 1, CreateInitialized)
 115 {
 116     RegExpConstructorPrivate* d = new RegExpConstructorPrivate;
 117     d->input = data->lastInput;
 118     d->lastInput = data->lastInput;
 119     d->lastNumSubPatterns = data->lastNumSubPatterns;
 120     unsigned offsetVectorSize = (data->lastNumSubPatterns + 1) * 2; // only copying the result part of the vector
 121     d->lastOvector().resize(offsetVectorSize);
 122     memcpy(d->lastOvector().data(), data->lastOvector().data(), offsetVectorSize * sizeof(int));
 123     // d->multiline is not needed, and remains uninitialized
 124
 125     setSubclassData(d);
 126 }
 127
 128 RegExpMatchesArray::~RegExpMatchesArray()
 129 {
 130     delete static_cast<RegExpConstructorPrivate*>(subclassData());
 131 }
 132
 133 void RegExpMatchesArray::fillArrayInstance(ExecState* exec)
 134 {
 135     RegExpConstructorPrivate* d = static_cast<RegExpConstructorPrivate*>(subclassData());
 136     ASSERT(d);
 137
 138     unsigned lastNumSubpatterns = d->lastNumSubPatterns;
 139
 140     for (unsigned i = 0; i <= lastNumSubpatterns; ++i) {
 141         int start = d->lastOvector()[2 * i];
 142         if (start >= 0)
 143             JSArray::put(exec, i, jsSubstring(exec, d->lastInput, start, d->lastOvector()[2 * i + 1] - start));
 144         else
 145             JSArray::put(exec, i, jsUndefined());
 146     }
 147
 148     PutPropertySlot slot;
 149     JSArray::put(exec, exec->propertyNames().index, jsNumber(d->lastOvector()[0]), slot);
 150     JSArray::put(exec, exec->propertyNames().input, jsString(exec, d->input), slot);
 151
 152     delete d;
 153     setSubclassData(0);
 154 }
 155
 156 JSObject* RegExpConstructor::arrayOfMatches(ExecState* exec) const
 157 {
 158     return new (exec) RegExpMatchesArray(exec, d.get());
 159 }
 160
 161 JSValue RegExpConstructor::getBackref(ExecState* exec, unsigned i) const
 162 {
 163     if (!d->lastOvector().isEmpty() && i <= d->lastNumSubPatterns) {
 164         int start = d->lastOvector()[2 * i];
 165         if (start >= 0)
 166             return jsSubstring(exec, d->lastInput, start, d->lastOvector()[2 * i + 1] - start);
 167     }
 168     return jsEmptyString(exec);
 169 }
 170
 171 JSValue RegExpConstructor::getLastParen(ExecState* exec) const
 172 {
 173     unsigned i = d->lastNumSubPatterns;
 174     if (i > 0) {
 175         ASSERT(!d->lastOvector().isEmpty());
 176         int start = d->lastOvector()[2 * i];
 177         if (start >= 0)
 178             return jsSubstring(exec, d->lastInput, start, d->lastOvector()[2 * i + 1] - start);
 179     }
 180     return jsEmptyString(exec);
 181 }
 182
 183 JSValue RegExpConstructor::getLeftContext(ExecState* exec) const
 184 {
 185     if (!d->lastOvector().isEmpty())
 186         return jsSubstring(exec, d->lastInput, 0, d->lastOvector()[0]);
 187     return jsEmptyString(exec);
 188 }
 189
 190 JSValue RegExpConstructor::getRightContext(ExecState* exec) const
 191 {
 192     if (!d->lastOvector().isEmpty())
 193         return jsSubstring(exec, d->lastInput, d->lastOvector()[1], d->lastInput.length() - d->lastOvector()[1]);
 194     return jsEmptyString(exec);
 195 }
 196
 197 bool RegExpConstructor::getOwnPropertySlot(ExecState* exec, const Identifier& propertyName, PropertySlot& slot)
 198 {
 199     return getStaticValueSlot<RegExpConstructor, InternalFunction>(exec, ExecState::regExpConstructorTable(exec), this, propertyName, slot);
 200 }
 201
 202 bool RegExpConstructor::getOwnPropertyDescriptor(ExecState* exec, const Identifier& propertyName, PropertyDescriptor& descriptor)
 203 {
 204     return getStaticValueDescriptor<RegExpConstructor, InternalFunction>(exec, ExecState::regExpConstructorTable(exec), this, propertyName, descriptor);
 205 }
 206
 207 JSValue regExpConstructorDollar1(ExecState* exec, JSValue slotBase, const Identifier&)
 208 {
 209     return asRegExpConstructor(slotBase)->getBackref(exec, 1);
 210 }
 211
 212 JSValue regExpConstructorDollar2(ExecState* exec, JSValue slotBase, const Identifier&)
 213 {
 214     return asRegExpConstructor(slotBase)->getBackref(exec, 2);
 215 }
 216
 217 JSValue regExpConstructorDollar3(ExecState* exec, JSValue slotBase, const Identifier&)
 218 {
 219     return asRegExpConstructor(slotBase)->getBackref(exec, 3);
 220 }
 221
 222 JSValue regExpConstructorDollar4(ExecState* exec, JSValue slotBase, const Identifier&)
 223 {
 224     return asRegExpConstructor(slotBase)->getBackref(exec, 4);
 225 }
 226
 227 JSValue regExpConstructorDollar5(ExecState* exec, JSValue slotBase, const Identifier&)
 228 {
 229     return asRegExpConstructor(slotBase)->getBackref(exec, 5);
 230 }
 231
 232 JSValue regExpConstructorDollar6(ExecState* exec, JSValue slotBase, const Identifier&)
 233 {
 234     return asRegExpConstructor(slotBase)->getBackref(exec, 6);
 235 }
 236
 237 JSValue regExpConstructorDollar7(ExecState* exec, JSValue slotBase, const Identifier&)
 238 {
 239     return asRegExpConstructor(slotBase)->getBackref(exec, 7);
 240 }
 241
 242 JSValue regExpConstructorDollar8(ExecState* exec, JSValue slotBase, const Identifier&)
 243 {
 244     return asRegExpConstructor(slotBase)->getBackref(exec, 8);
 245 }
 246
 247 JSValue regExpConstructorDollar9(ExecState* exec, JSValue slotBase, const Identifier&)
 248 {
 249     return asRegExpConstructor(slotBase)->getBackref(exec, 9);
 250 }
 251
 252 JSValue regExpConstructorInput(ExecState* exec, JSValue slotBase, const Identifier&)
 253 {
 254     return jsString(exec, asRegExpConstructor(slotBase)->input());
 255 }
 256
 257 JSValue regExpConstructorMultiline(ExecState*, JSValue slotBase, const Identifier&)
 258 {
 259     return jsBoolean(asRegExpConstructor(slotBase)->multiline());
 260 }
 261
 262 JSValue regExpConstructorLastMatch(ExecState* exec, JSValue slotBase, const Identifier&)
 263 {
 264     return asRegExpConstructor(slotBase)->getBackref(exec, 0);
 265 }
 266
 267 JSValue regExpConstructorLastParen(ExecState* exec, JSValue slotBase, const Identifier&)
 268 {
 269     return asRegExpConstructor(slotBase)->getLastParen(exec);
 270 }
 271
 272 JSValue regExpConstructorLeftContext(ExecState* exec, JSValue slotBase, const Identifier&)
 273 {
 274     return asRegExpConstructor(slotBase)->getLeftContext(exec);
 275 }
 276
 277 JSValue regExpConstructorRightContext(ExecState* exec, JSValue slotBase, const Identifier&)
 278 {
 279     return asRegExpConstructor(slotBase)->getRightContext(exec);
 280 }
 281
 282 void RegExpConstructor::put(ExecState* exec, const Identifier& propertyName, JSValue value, PutPropertySlot& slot)
 283 {
 284     lookupPut<RegExpConstructor, InternalFunction>(exec, propertyName, value, ExecState::regExpConstructorTable(exec), this, slot);
 285 }
 286
 287 void setRegExpConstructorInput(ExecState* exec, JSObject* baseObject, JSValue value)
 288 {
 289     asRegExpConstructor(baseObject)->setInput(value.toString(exec));
 290 }
 291
 292 void setRegExpConstructorMultiline(ExecState* exec, JSObject* baseObject, JSValue value)
 293 {
 294     asRegExpConstructor(baseObject)->setMultiline(value.toBoolean(exec));
 295 }
 296
 297 // ECMA 15.10.4
 298 JSObject* constructRegExp(ExecState* exec, JSGlobalObject* globalObject, const ArgList& args)
 299 {
 300     JSValue arg0 = args.at(0);
 301     JSValue arg1 = args.at(1);
 302
 303     if (arg0.inherits(&RegExpObject::s_info)) {
 304         if (!arg1.isUndefined())
 305             return throwError(exec, createTypeError(exec, "Cannot supply flags when constructing one RegExp from another."));
 306         return asObject(arg0);
 307     }
 308
 309     UString pattern = arg0.isUndefined() ? UString("") : arg0.toString(exec);
 310     if (exec->hadException())
 311         return 0;
 312
 313     RegExpFlags flags = NoFlags;
 314     if (!arg1.isUndefined()) {
 315         flags = regExpFlags(arg1.toString(exec));
 316         if (exec->hadException())
 317             return 0;
 318         if (flags == InvalidFlags)
 319             return throwError(exec, createSyntaxError(exec, "Invalid flags supplied to RegExp constructor."));
 320     }
 321
 322     RegExp* regExp = RegExp::create(&exec->globalData(), pattern, flags);
 323     if (!regExp->isValid())
 324         return throwError(exec, createSyntaxError(exec, regExp->errorMessage()));
 325     return new (exec) RegExpObject(exec->lexicalGlobalObject(), globalObject->regExpStructure(), regExp);
 326 }
 327
 328 static EncodedJSValue JSC_HOST_CALL constructWithRegExpConstructor(ExecState* exec)
 329 {
 330     ArgList args(exec);
 331     return JSValue::encode(constructRegExp(exec, asInternalFunction(exec->callee())->globalObject(), args));
 332 }
 333
 334 ConstructType RegExpConstructor::getConstructData(ConstructData& constructData)
 335 {
 336     constructData.native.function = constructWithRegExpConstructor;
 337     return ConstructTypeHost;
 338 }
 339
 340 // ECMA 15.10.3
 341 static EncodedJSValue JSC_HOST_CALL callRegExpConstructor(ExecState* exec)
 342 {
 343     ArgList args(exec);
 344     return JSValue::encode(constructRegExp(exec, asInternalFunction(exec->callee())->globalObject(), args));
 345 }
 346
 347 CallType RegExpConstructor::getCallData(CallData& callData)
 348 {
 349     callData.native.function = callRegExpConstructor;
 350     return CallTypeHost;
 351 }
 352
 353 void RegExpConstructor::setInput(const UString& input)
 354 {
 355     d->input = input;
 356 }
 357
 358 const UString& RegExpConstructor::input() const
 359 {
 360     // Can detect a distinct initial state that is invisible to JavaScript, by checking for null
 361     // state (since jsString turns null strings to empty strings).
 362     return d->input;
 363 }
 364
 365 void RegExpConstructor::setMultiline(bool multiline)
 366 {
 367     d->multiline = multiline;
 368 }
 369
 370 bool RegExpConstructor::multiline() const
 371 {
 372     return d->multiline;
 373 }
 374
 375 } // namespace JSC