12010-05-24 Gavin Barraclough <barraclough@apple.com> 2 3 Rubber Stamped by Sam Weinig. 4 5 Accidentally committed double write of codeblock in Interpreter. 6 7 * interpreter/Interpreter.cpp: 8 (JSC::Interpreter::privateExecute): 9 102010-05-24 Gavin Barraclough <barraclough@apple.com> 11 12 Reviewed by Sam Weinig. 13 14 https://bugs.webkit.org/show_bug.cgi?id=39583 15 Move creation of 'this' object from caller to callee in construction. 16 17 Presently the caller of a constructor is responsible for providing a this 18 object. Instead, move the object creation into a new op_create_this opcode, 19 planted in the head of the contructor bytecode for a function. Since the 20 prototype for the object is provided by performing a get_by_id on the callee, 21 also add a new get_callee opcode (this is used to get the callee JSFunction 22 into a register so that a normal get_by_id can be used). 23 24 Currently the caller is also responsible for detecting when op_construct is 25 performed on a JSFunction representing a host function, in which case an 26 exception is thrown – and this check currently takes place when constructing 27 the this object. Instead, mirroring the recent changes for non-host functions, 28 add a parallel code-path for native constructors to follow, with a thunk for 29 invoking native constructors provided by JITStubs, and a constructor-specific 30 NativeFunction on NativeExecutable. Provide an implementation of a host 31 constructor which will throw an exception. 32 33 * bytecode/CodeBlock.cpp: 34 (JSC::CodeBlock::dump): 35 (JSC::CodeBlock::getByIdExceptionInfoForBytecodeOffset): 36 * bytecode/CodeBlock.h: 37 * bytecode/Opcode.h: 38 * bytecompiler/BytecodeGenerator.cpp: 39 (JSC::BytecodeGenerator::BytecodeGenerator): 40 (JSC::BytecodeGenerator::emitConstruct): 41 * bytecompiler/BytecodeGenerator.h: 42 (JSC::BytecodeGenerator::emitGetByIdExceptionInfo): 43 * interpreter/Interpreter.cpp: 44 (JSC::Interpreter::privateExecute): 45 * jit/JIT.cpp: 46 (JSC::JIT::privateCompileMainPass): 47 * jit/JIT.h: 48 * jit/JITCall.cpp: 49 (JSC::JIT::compileOpCall): 50 (JSC::JIT::compileOpCallSlowCase): 51 * jit/JITCall32_64.cpp: 52 (JSC::JIT::compileOpCall): 53 (JSC::JIT::compileOpCallSlowCase): 54 * jit/JITOpcodes.cpp: 55 (JSC::JIT::privateCompileCTIMachineTrampolines): 56 (JSC::JIT::privateCompileCTINativeCall): 57 (JSC::JIT::emit_op_neq_null): 58 (JSC::JIT::emit_op_convert_this): 59 (JSC::JIT::emit_op_get_callee): 60 (JSC::JIT::emit_op_create_this): 61 * jit/JITOpcodes32_64.cpp: 62 (JSC::JIT::privateCompileCTIMachineTrampolines): 63 (JSC::JIT::privateCompileCTINativeCall): 64 (JSC::JIT::emit_op_get_callee): 65 (JSC::JIT::emit_op_create_this): 66 * jit/JITStubs.cpp: 67 (JSC::DEFINE_STUB_FUNCTION): 68 (JSC::JITThunks::hostFunctionStub): 69 * jit/JITStubs.h: 70 (JSC::JITThunks::ctiNativeConstruct): 71 (JSC::): 72 * runtime/ExceptionHelpers.cpp: 73 (JSC::createNotAnObjectError): 74 * runtime/Executable.h: 75 (JSC::NativeExecutable::create): 76 (JSC::NativeExecutable::NativeExecutable): 77 * runtime/JSFunction.cpp: 78 (JSC::callHostFunctionAsConstructor): 79 * runtime/JSFunction.h: 80 812010-05-23 Sam Weinig <sam@webkit.org> 82 83 Fix windows build. 84 85 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 86 872010-05-23 Sam Weinig <sam@webkit.org> 88 89 Reviewed by Oliver Hunt. 90 91 Fix for https://bugs.webkit.org/show_bug.cgi?id=39575 92 Make JS DOMObject inherit from JSObjectWithGlobalObject instead of JSObject 93 94 Expose the global object stored in JSObjectWithGlobalObject. 95 96 * JavaScriptCore.exp: 97 * JavaScriptCore.xcodeproj/project.pbxproj: 98 * runtime/JSObjectWithGlobalObject.cpp: 99 (JSC::JSObjectWithGlobalObject::JSObjectWithGlobalObject): 100 (JSC::JSObjectWithGlobalObject::globalObject): 101 * runtime/JSObjectWithGlobalObject.h: 102 1032010-05-21 Oliver Hunt <oliver@apple.com> 104 105 "asm volatile" isn't valid outside of functions. 106 107 Reviewed by Gavin Barraclough. 108 109 * jit/JITStubs.cpp: 110 1112010-05-21 Gavin Barraclough <barraclough@apple.com> 112 113 Unreviewed build fix. 114 115 Interpreter fix following r59974. 116 117 * interpreter/Interpreter.cpp: 118 (JSC::Interpreter::privateExecute): 119 * runtime/JSPropertyNameIterator.cpp: 120 (JSC::JSPropertyNameIterator::get): 121 * runtime/JSPropertyNameIterator.h: 122 1232010-05-21 Gavin Barraclough <barraclough@apple.com> 124 125 Rubber stamped by Oliver Hunt. 126 127 Interpreter fix following r59939. 128 129 * interpreter/Interpreter.cpp: 130 (JSC::Interpreter::privateExecute): 131 1322010-05-21 David Levin <levin@chromium.org> 133 134 Unreviewed build fix. 135 136 * wtf/SizeLimits.cpp: Removed a check while I figure out how to write it properly. 137 1382010-05-21 David Levin <levin@chromium.org> 139 140 Reviewed by Darin Adler. 141 142 Enforce size constraints on various data structures in JavaScriptCore/wtf. 143 https://bugs.webkit.org/show_bug.cgi?id=39327 144 145 I only modified the default build for OSX and Chromium's build file to include WTFCompileAsserts.cpp 146 as those should be sufficient to catch regressions on the size of the data structures. 147 148 * JavaScriptCore.gypi: Added the WTFCompileAsserts.cpp file. 149 * JavaScriptCore.xcodeproj/project.pbxproj: Added the WTFCompileAsserts.cpp file. 150 * runtime/UString.cpp: Added a compile assert for UString size. 151 * wtf/SizeLimits.cpp: Added compile asserts for data structures that didn't have cpp files. 152 * wtf/text/StringImpl.cpp: Added a compile assert for StringImpl size. 153 1542010-05-21 Gavin Barraclough <barraclough@apple.com> 155 156 Reviewed by Oliver Hunt. 157 158 Bug 39509 - Exceptions raised when JSString::value() is called not always properly thrown. 159 Missing some CHECK_FOR_EXCEPTION()s, cleanup NativeError construction (adding a property 160 via the prototype was inefficient, and whilst trivially true that the property is is not 161 a rope - and thus couldn't except - it would be hard to prove this). 162 163 * interpreter/Interpreter.cpp: 164 (JSC::Interpreter::callEval): 165 (JSC::Interpreter::privateExecute): 166 * jit/JITStubs.cpp: 167 (JSC::DEFINE_STUB_FUNCTION): 168 * runtime/InternalFunction.cpp: 169 (JSC::InternalFunction::name): 170 (JSC::InternalFunction::displayName): 171 * runtime/JSFunction.cpp: 172 (JSC::JSFunction::name): 173 (JSC::JSFunction::displayName): 174 * runtime/JSGlobalObject.cpp: 175 (JSC::JSGlobalObject::reset): 176 * runtime/JSPropertyNameIterator.cpp: 177 * runtime/JSPropertyNameIterator.h: 178 * runtime/JSString.h: 179 (JSC::RopeBuilder::tryGetValue): 180 (JSC::JSString::getIndex): 181 * runtime/NativeErrorConstructor.cpp: 182 (JSC::NativeErrorConstructor::NativeErrorConstructor): 183 * runtime/NativeErrorConstructor.h: 184 * runtime/NativeErrorPrototype.cpp: 185 (JSC::NativeErrorPrototype::NativeErrorPrototype): 186 * runtime/NativeErrorPrototype.h: 187 * runtime/StringPrototype.cpp: 188 (JSC::stringProtoFuncReplace): 189 1902010-05-21 Gavin Barraclough <barraclough@apple.com> 191 192 Reviewed by NOBODY (build fix). 193 194 Fix interpreter build. 195 196 * interpreter/Interpreter.cpp: 197 (JSC::Interpreter::privateExecute): 198 1992010-05-21 Geoffrey Garen <ggaren@apple.com> 200 201 Reviewed by Gavin Barraclough. 202 203 Split out JITCall32_64.cpp, and tightened up some #ifdefs in the hopes 204 of improving compile times. 205 https://bugs.webkit.org/show_bug.cgi?id=39507 206 207 * JavaScriptCore.xcodeproj/project.pbxproj: 208 * jit/JIT.cpp: 209 * jit/JITArithmetic.cpp: 210 * jit/JITArithmetic32_64.cpp: 211 * jit/JITCall.cpp: 212 * jit/JITCall32_64.cpp: Copied from jit/JITCall.cpp. 213 * jit/JITOpcodes.cpp: 214 * jit/JITOpcodes32_64.cpp: 215 * jit/JITPropertyAccess.cpp: 216 * jit/JITPropertyAccess32_64.cpp: 217 * jit/JITStubs.cpp: 218 2192010-05-21 Csaba Osztrogonác <ossy@webkit.org> 220 221 Unreviewed buildfix for Qt after r59941. 222 223 * JavaScriptCore.pro: missing runtime/JSObjectWithGlobalObject.cpp added to SOURCES. 224 2252010-05-21 Oliver Hunt <oliver@apple.com> 226 227 Windows build fix #3 228 229 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 230 2312010-05-21 Oliver Hunt <oliver@apple.com> 232 233 Windows build fix #2 234 235 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 236 2372010-05-21 Oliver Hunt <oliver@apple.com> 238 239 Windows build fix #1 240 241 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 242 2432010-05-21 Oliver Hunt <oliver@apple.com> 244 245 Reviewed by Geoffrey Garen. 246 247 All callable objects should have a global object reference 248 https://bugs.webkit.org/show_bug.cgi?id=39495 249 250 All objects that may ever return a value other CallTypeNone 251 or ConstructTypeNone now get a global object in their constructor 252 and store that in their first anonymous slot. We add a new type 253 JSObjectWithGlobalObject to allow us to share this logic as much 254 as possible, however some objects have specific inheritance 255 requirements so we can't just use it universally. 256 257 To enforce this requirement JSValue::getCallData and getConstructData 258 make use of a new "isValidCallee" function to assert that any object 259 that returns a value other than CallType/ConstructTypeNone has a 260 global object in anonymous slot 0. 261 262 In order to ensure that static function slots are converted into 263 function objects with the correct global object, all prototype objects 264 and other classes with static function slots also gain a global object 265 reference. Happily this fixes the long standing issue where host 266 function objects get a prototype from the lexical global object of the 267 first function that calls them, instead of the global object that they 268 are defined on. 269 270 * API/JSCallbackConstructor.cpp: 271 (JSC::JSCallbackConstructor::JSCallbackConstructor): 272 * API/JSCallbackConstructor.h: 273 * API/JSCallbackFunction.cpp: 274 (JSC::JSCallbackFunction::JSCallbackFunction): 275 * API/JSCallbackFunction.h: 276 * API/JSCallbackObject.cpp: 277 (JSC::): 278 * API/JSCallbackObject.h: 279 * API/JSCallbackObjectFunctions.h: 280 (JSC::::JSCallbackObject): 281 (JSC::::staticFunctionGetter): 282 * API/JSClassRef.cpp: 283 (OpaqueJSClass::prototype): 284 * API/JSContextRef.cpp: 285 * API/JSObjectRef.cpp: 286 (JSObjectMake): 287 (JSObjectMakeFunctionWithCallback): 288 (JSObjectMakeConstructor): 289 (JSObjectGetPrivate): 290 (JSObjectSetPrivate): 291 (JSObjectGetPrivateProperty): 292 (JSObjectSetPrivateProperty): 293 (JSObjectDeletePrivateProperty): 294 * API/JSValueRef.cpp: 295 (JSValueIsObjectOfClass): 296 * API/JSWeakObjectMapRefPrivate.cpp: 297 * CMakeLists.txt: 298 * GNUmakefile.am: 299 * JavaScriptCore.exp: 300 * JavaScriptCore.gypi: 301 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: 302 * JavaScriptCore.xcodeproj/project.pbxproj: 303 * interpreter/Interpreter.cpp: 304 (JSC::Interpreter::privateExecute): 305 * jit/JITStubs.cpp: 306 (JSC::DEFINE_STUB_FUNCTION): 307 * jsc.cpp: 308 (GlobalObject::GlobalObject): 309 * runtime/ArrayConstructor.cpp: 310 (JSC::ArrayConstructor::ArrayConstructor): 311 * runtime/ArrayConstructor.h: 312 * runtime/ArrayPrototype.cpp: 313 (JSC::ArrayPrototype::ArrayPrototype): 314 * runtime/ArrayPrototype.h: 315 (JSC::ArrayPrototype::createStructure): 316 * runtime/BooleanConstructor.cpp: 317 (JSC::BooleanConstructor::BooleanConstructor): 318 * runtime/BooleanConstructor.h: 319 * runtime/BooleanPrototype.cpp: 320 (JSC::BooleanPrototype::BooleanPrototype): 321 * runtime/BooleanPrototype.h: 322 * runtime/DateConstructor.cpp: 323 (JSC::DateConstructor::DateConstructor): 324 * runtime/DateConstructor.h: 325 * runtime/DatePrototype.cpp: 326 (JSC::DatePrototype::DatePrototype): 327 * runtime/DatePrototype.h: 328 * runtime/ErrorConstructor.cpp: 329 (JSC::ErrorConstructor::ErrorConstructor): 330 * runtime/ErrorConstructor.h: 331 * runtime/ErrorPrototype.cpp: 332 (JSC::ErrorPrototype::ErrorPrototype): 333 * runtime/ErrorPrototype.h: 334 * runtime/FunctionConstructor.cpp: 335 (JSC::FunctionConstructor::FunctionConstructor): 336 * runtime/FunctionConstructor.h: 337 * runtime/FunctionPrototype.cpp: 338 (JSC::FunctionPrototype::FunctionPrototype): 339 (JSC::FunctionPrototype::addFunctionProperties): 340 * runtime/FunctionPrototype.h: 341 * runtime/GlobalEvalFunction.cpp: 342 (JSC::GlobalEvalFunction::GlobalEvalFunction): 343 * runtime/GlobalEvalFunction.h: 344 * runtime/InternalFunction.cpp: 345 (JSC::InternalFunction::InternalFunction): 346 * runtime/InternalFunction.h: 347 * runtime/JSCell.h: 348 (JSC::JSValue::getCallData): 349 (JSC::JSValue::getConstructData): 350 * runtime/JSFunction.cpp: 351 (JSC::JSFunction::JSFunction): 352 * runtime/JSFunction.h: 353 * runtime/JSGlobalObject.cpp: 354 (JSC::JSGlobalObject::reset): 355 * runtime/JSGlobalObject.h: 356 (JSC::JSGlobalObject::JSGlobalObject): 357 * runtime/JSONObject.cpp: 358 (JSC::JSONObject::JSONObject): 359 * runtime/JSONObject.h: 360 * runtime/JSObject.h: 361 * runtime/JSObjectWithGlobalObject.cpp: Added. 362 (JSC::JSObjectWithGlobalObject::JSObjectWithGlobalObject): 363 * runtime/JSObjectWithGlobalObject.h: Added. 364 (JSC::JSObjectWithGlobalObject::createStructure): 365 (JSC::JSObjectWithGlobalObject::JSObjectWithGlobalObject): 366 * runtime/JSValue.cpp: 367 (JSC::JSValue::isValidCallee): 368 * runtime/JSValue.h: 369 * runtime/Lookup.cpp: 370 (JSC::setUpStaticFunctionSlot): 371 * runtime/MathObject.cpp: 372 (JSC::MathObject::MathObject): 373 * runtime/MathObject.h: 374 * runtime/NativeErrorConstructor.cpp: 375 (JSC::NativeErrorConstructor::NativeErrorConstructor): 376 * runtime/NativeErrorConstructor.h: 377 * runtime/NativeErrorPrototype.cpp: 378 (JSC::NativeErrorPrototype::NativeErrorPrototype): 379 * runtime/NativeErrorPrototype.h: 380 * runtime/NumberConstructor.cpp: 381 (JSC::NumberConstructor::NumberConstructor): 382 * runtime/NumberConstructor.h: 383 * runtime/NumberPrototype.cpp: 384 (JSC::NumberPrototype::NumberPrototype): 385 * runtime/NumberPrototype.h: 386 * runtime/ObjectConstructor.cpp: 387 (JSC::ObjectConstructor::ObjectConstructor): 388 * runtime/ObjectConstructor.h: 389 * runtime/ObjectPrototype.cpp: 390 (JSC::ObjectPrototype::ObjectPrototype): 391 * runtime/ObjectPrototype.h: 392 * runtime/PrototypeFunction.cpp: 393 (JSC::PrototypeFunction::PrototypeFunction): 394 * runtime/PrototypeFunction.h: 395 * runtime/RegExpConstructor.cpp: 396 (JSC::RegExpConstructor::RegExpConstructor): 397 (JSC::constructRegExp): 398 * runtime/RegExpConstructor.h: 399 * runtime/RegExpObject.cpp: 400 (JSC::RegExpObject::RegExpObject): 401 * runtime/RegExpObject.h: 402 * runtime/RegExpPrototype.cpp: 403 (JSC::RegExpPrototype::RegExpPrototype): 404 * runtime/RegExpPrototype.h: 405 * runtime/StringConstructor.cpp: 406 (JSC::StringConstructor::StringConstructor): 407 * runtime/StringConstructor.h: 408 * runtime/StringPrototype.cpp: 409 (JSC::StringPrototype::StringPrototype): 410 * runtime/StringPrototype.h: 411 4122010-05-21 Geoffrey Garen <ggaren@apple.com> 413 414 Reviewed by Oliver Hunt. 415 416 Removed the unused entry in the CallFrame 417 https://bugs.webkit.org/show_bug.cgi?id=39470 418 419 Removed the unused entry in the CallFrame, and reordered CallFrame 420 writes on the hot path to be in forwards order. 421 422 SunSpider says 0.3% faster. 423 424 * interpreter/RegisterFile.h: 425 (JSC::RegisterFile::): 426 * jit/JITCall.cpp: 427 (JSC::JIT::compileOpCall): 428 4292010-05-21 Jedrzej Nowacki <jedrzej.nowacki@nokia.com> 430 431 Reviewed by Kenneth Rohde Christiansen. 432 433 Split the QScriptValue autotest into smaller chunks. 434 435 File tst_qscriptvalue_generated.cpp was huge and it was difficult to compile 436 it with MSVS2008, because of the code optimization. 437 438 Tests were splited into a few files for better use of distributed 439 compilation. 440 Repeated calls to insert() and operator<<() where replaced by loops, 441 that should reduce time of code optimizing. 442 443 New files were generated by the same application, test logic wasn't changed. 444 445 [Qt] The QScriptValue autotest should be refactored. 446 https://bugs.webkit.org/show_bug.cgi?id=38987 447 448 * qt/tests/qscriptvalue/qscriptvalue.pro: 449 * qt/tests/qscriptvalue/tst_qscriptvalue.h: 450 * qt/tests/qscriptvalue/tst_qscriptvalue_generated.cpp: Removed. 451 * qt/tests/qscriptvalue/tst_qscriptvalue_generated_init.cpp: Added. 452 (tst_QScriptValue::initScriptValues): 453 * qt/tests/qscriptvalue/tst_qscriptvalue_generated_istype.cpp: Added. 454 (tst_QScriptValue::isValid_initData): 455 (tst_QScriptValue::isValid_makeData): 456 (tst_QScriptValue::isValid_test): 457 (tst_QScriptValue::isBool_initData): 458 (tst_QScriptValue::isBool_makeData): 459 (tst_QScriptValue::isBool_test): 460 (tst_QScriptValue::isBoolean_initData): 461 (tst_QScriptValue::isBoolean_makeData): 462 (tst_QScriptValue::isBoolean_test): 463 (tst_QScriptValue::isNumber_initData): 464 (tst_QScriptValue::isNumber_makeData): 465 (tst_QScriptValue::isNumber_test): 466 (tst_QScriptValue::isFunction_initData): 467 (tst_QScriptValue::isFunction_makeData): 468 (tst_QScriptValue::isFunction_test): 469 (tst_QScriptValue::isNull_initData): 470 (tst_QScriptValue::isNull_makeData): 471 (tst_QScriptValue::isNull_test): 472 (tst_QScriptValue::isString_initData): 473 (tst_QScriptValue::isString_makeData): 474 (tst_QScriptValue::isString_test): 475 (tst_QScriptValue::isUndefined_initData): 476 (tst_QScriptValue::isUndefined_makeData): 477 (tst_QScriptValue::isUndefined_test): 478 (tst_QScriptValue::isObject_initData): 479 (tst_QScriptValue::isObject_makeData): 480 (tst_QScriptValue::isObject_test): 481 * qt/tests/qscriptvalue/tst_qscriptvalue_generated_totype.cpp: Added. 482 (tst_QScriptValue::toString_initData): 483 (tst_QScriptValue::toString_makeData): 484 (tst_QScriptValue::toString_test): 485 (tst_QScriptValue::toNumber_initData): 486 (tst_QScriptValue::toNumber_makeData): 487 (tst_QScriptValue::toNumber_test): 488 (tst_QScriptValue::toBool_initData): 489 (tst_QScriptValue::toBool_makeData): 490 (tst_QScriptValue::toBool_test): 491 (tst_QScriptValue::toBoolean_initData): 492 (tst_QScriptValue::toBoolean_makeData): 493 (tst_QScriptValue::toBoolean_test): 494 (tst_QScriptValue::toInteger_initData): 495 (tst_QScriptValue::toInteger_makeData): 496 (tst_QScriptValue::toInteger_test): 497 (tst_QScriptValue::toInt32_initData): 498 (tst_QScriptValue::toInt32_makeData): 499 (tst_QScriptValue::toInt32_test): 500 (tst_QScriptValue::toUInt32_initData): 501 (tst_QScriptValue::toUInt32_makeData): 502 (tst_QScriptValue::toUInt32_test): 503 (tst_QScriptValue::toUInt16_initData): 504 (tst_QScriptValue::toUInt16_makeData): 505 (tst_QScriptValue::toUInt16_test): 506 5072010-05-21 Kwang Yul Seo <skyul@company100.net> 508 509 Reviewed by Eric Seidel. 510 511 Add COMPILER(MSVC) guard for strnstr 512 https://bugs.webkit.org/show_bug.cgi?id=39397 513 514 BREWMP uses MSVC for simulator build, but it is not OS(WINDOWS). 515 Add COMPILER(MSVC) guard for strnstr. 516 517 * wtf/StringExtras.h: 518 5192010-05-21 Kwang Yul Seo <skyul@company100.net> 520 521 Reviewed by Eric Seidel. 522 523 [BREWMP] Add !PLATFORM(BREWMP) guard for isMainThread 524 https://bugs.webkit.org/show_bug.cgi?id=39396 525 526 BREWMP sets ENABLE(SINGLE_THREADED), so it uses isMainThread from ThreadingNone.cpp. 527 Exclude isMainThread in MainThread.cpp. 528 529 * wtf/MainThread.cpp: 530 5312010-05-20 Kwang Yul Seo <skyul@company100.net> 532 533 Reviewed by Eric Seidel. 534 535 [BREWMP] Use skia as BREW MP's graphics backend. 536 https://bugs.webkit.org/show_bug.cgi?id=39407 537 538 Define WTF_PLATFORM_SKIA=1 for BREW MP. 539 540 * wtf/Platform.h: 541 5422010-05-20 Geoffrey Garen <ggaren@apple.com> 543 544 Reviewed by Beth Dakin. 545 546 Fixed naming mixup: it's op_call_put_result, not ret_result. 547 548 * bytecode/CodeBlock.cpp: 549 (JSC::CodeBlock::dump): 550 * interpreter/Interpreter.cpp: 551 (JSC::Interpreter::privateExecute): 552 5532010-05-20 Geoffrey Garen <ggaren@apple.com> 554 555 Reviewed by Oliver Hunt. 556 557 Whittling away at CodeBlock use: Removed a little wonkiness in closure lookup 558 https://bugs.webkit.org/show_bug.cgi?id=39444 559 560 Calculate all lookup depths at compile time. I'm not sure why we couldn't 561 do this before, but whatever the reason, it seems to be gone now. 562 563 * bytecompiler/BytecodeGenerator.cpp: 564 (JSC::BytecodeGenerator::findScopedProperty): 565 * bytecompiler/BytecodeGenerator.h: 566 * interpreter/Interpreter.cpp: 567 (JSC::Interpreter::resolveSkip): 568 (JSC::Interpreter::resolveGlobalDynamic): 569 (JSC::Interpreter::privateExecute): 570 * jit/JITOpcodes.cpp: 571 (JSC::JIT::emit_op_get_scoped_var): 572 (JSC::JIT::emit_op_put_scoped_var): 573 (JSC::JIT::emit_op_resolve_skip): 574 (JSC::JIT::emit_op_resolve_global_dynamic): 575 (JSC::JIT::emitSlow_op_resolve_global_dynamic): 576 * jit/JITOpcodes32_64.cpp: 577 (JSC::JIT::emit_op_get_scoped_var): 578 (JSC::JIT::emit_op_put_scoped_var): 579 (JSC::JIT::emit_op_resolve_skip): 580 5812010-05-20 Darin Adler <darin@apple.com> 582 583 Reviewed by Eric Seidel. 584 585 Fix warnings seen on the buildbots today 586 https://bugs.webkit.org/show_bug.cgi?id=39368 587 588 * profiler/ProfileNode.cpp: 589 (JSC::getCount): Removed unneeded initialization of global variable. 590 These are all initialized to zero anyway, so no need to try to initialize 591 it and get a warning. 592 593 * wtf/CurrentTime.cpp: 594 (WTF::currentTime): Removed unused global variable. 595 5962010-05-20 Geoffrey Garen <ggaren@apple.com> 597 598 Reviewed by Darin Adler. 599 600 Removed some reliance on callFame[CodeBlock] by storing CodeBlock in a 601 local variable in the Interpreter. 602 https://bugs.webkit.org/show_bug.cgi?id=39447 603 604 Small speedup on SunSpider in Interpreter mode. 605 606 * interpreter/Interpreter.cpp: 607 (JSC::Interpreter::privateExecute): 608 * jsc.cpp: 609 * wtf/Platform.h: 610 6112010-05-19 Gavin Barraclough <barraclough@apple.com> 612 613 Reviewed by Oliver Hunt. 614 615 Bug 39408 - Unify JIT/interpreter return behaviour. 616 617 Presently the JIT and the interpreter have differing implementations in how 618 they handle storing the result of a function call back to the register file. 619 In both cases the op_call ( / _eval / _varargs / op_construct) opcode has an 620 operand indicating the VM register in which the result should be stored. 621 The JIT plants code after the call to store the result, so the value will be 622 stored as soon as the callee has returned. In the interpreter the call 623 passes the return value register id into the callee via the callee callframe, 624 and the callee is responsible for writing the result back into its callers 625 register file after it has restored the parents callframe pointer, but before 626 returning. 627 628 Instead, move the task of writing the call result to the register file into a 629 new opcode (op_call_put_result), and after returning the callee should leave 630 the return value in a canonical location. In the case of the interpreter, 631 this canonical location is a local variable in privateExecute 632 (functionReturnValue), in the case of the JIT this is the normal return value 633 registers (regT0, or regT1:regT0 in JSVALUE32_64). op_call_put_result stores 634 the result from the canonical location to the registerfile. 635 636 In addition to unifying JIT & interpreter behaviour this change allows us to 637 remove a slot from the callframe, omit the write of the result where the 638 return value from the call is not used, and provides a 2% speedup on sunspider 639 in the interpreter. 640 641 * bytecode/CodeBlock.cpp: 642 (JSC::CodeBlock::dump): 643 * bytecode/Opcode.h: 644 * bytecompiler/BytecodeGenerator.cpp: 645 (JSC::BytecodeGenerator::emitCall): 646 (JSC::BytecodeGenerator::emitCallVarargs): 647 (JSC::BytecodeGenerator::emitConstruct): 648 * bytecompiler/BytecodeGenerator.h: 649 (JSC::BytecodeGenerator::finalDestinationOrIgnored): 650 * bytecompiler/NodesCodegen.cpp: 651 (JSC::NewExprNode::emitBytecode): 652 (JSC::FunctionCallValueNode::emitBytecode): 653 (JSC::FunctionCallResolveNode::emitBytecode): 654 (JSC::FunctionCallBracketNode::emitBytecode): 655 (JSC::FunctionCallDotNode::emitBytecode): 656 (JSC::CallFunctionCallDotNode::emitBytecode): 657 (JSC::ApplyFunctionCallDotNode::emitBytecode): 658 * interpreter/CallFrame.h: 659 (JSC::ExecState::init): 660 (JSC::ExecState::noCaller): 661 * interpreter/Interpreter.cpp: 662 (JSC::Interpreter::dumpRegisters): 663 (JSC::Interpreter::throwException): 664 (JSC::Interpreter::privateExecute): 665 * interpreter/RegisterFile.h: 666 (JSC::RegisterFile::): 667 * jit/JIT.cpp: 668 (JSC::JIT::privateCompileMainPass): 669 * jit/JIT.h: 670 * jit/JITCall.cpp: 671 (JSC::JIT::compileOpCallSetupArgs): 672 (JSC::JIT::compileOpConstructSetupArgs): 673 (JSC::JIT::emit_op_call_put_result): 674 (JSC::JIT::compileOpCallVarargs): 675 (JSC::JIT::compileOpCallVarargsSlowCase): 676 (JSC::JIT::compileOpCall): 677 (JSC::JIT::compileOpCallSlowCase): 678 (JSC::JIT::compileOpCallVarargsSetupArgs): 679 6802010-05-20 Steve Block <steveblock@google.com> 681 682 Reviewed by Jeremy Orlow. 683 684 Provide bindings for DeviceOrientation 685 https://bugs.webkit.org/show_bug.cgi?id=39210 686 687 Adds ENABLE_DEVICE_ORIENTATION to XCode project file, always disabled. 688 689 * Configurations/FeatureDefines.xcconfig: 690 6912010-05-19 Gavin Barraclough <barraclough@apple.com> 692 693 Reviewed by Geoff Garen. 694 695 Errk! rename op_constructor_ret as requested in review of Bug 39399. 696 697 * bytecode/CodeBlock.cpp: 698 (JSC::CodeBlock::dump): 699 * bytecode/Opcode.h: 700 * bytecompiler/BytecodeGenerator.cpp: 701 (JSC::BytecodeGenerator::emitReturn): 702 * interpreter/Interpreter.cpp: 703 (JSC::Interpreter::privateExecute): 704 * jit/JIT.cpp: 705 (JSC::JIT::privateCompileMainPass): 706 * jit/JIT.h: 707 * jit/JITCall.cpp: 708 (JSC::JIT::emit_op_ret_object_or_this): 709 * jit/JITOpcodes.cpp: 710 (JSC::JIT::emit_op_ret_object_or_this): 711 7122010-05-19 Gavin Barraclough <barraclough@apple.com> 713 714 Reviewed by Geoff Garen. 715 716 Bug 39399 - Move responsibility for verifying constructors return objects from the caller to the callee. 717 718 This is a necessary step to move object creation from caller to callee. 719 720 * bytecode/CodeBlock.cpp: 721 (JSC::CodeBlock::dump): 722 * bytecode/Opcode.h: 723 * bytecompiler/BytecodeGenerator.cpp: 724 (JSC::BytecodeGenerator::BytecodeGenerator): 725 (JSC::BytecodeGenerator::emitReturn): 726 (JSC::BytecodeGenerator::emitConstruct): 727 * bytecompiler/BytecodeGenerator.h: 728 (JSC::BytecodeGenerator::isConstructor): 729 * bytecompiler/NodesCodegen.cpp: 730 (JSC::FunctionBodyNode::emitBytecode): 731 * interpreter/Interpreter.cpp: 732 (JSC::Interpreter::privateExecute): 733 * jit/JIT.cpp: 734 (JSC::JIT::privateCompileMainPass): 735 (JSC::JIT::privateCompileSlowCases): 736 * jit/JIT.h: 737 * jit/JITCall.cpp: 738 (JSC::JIT::emit_op_constructor_ret): 739 * jit/JITOpcodes.cpp: 740 (JSC::JIT::emit_op_constructor_ret): 741 7422010-05-19 Gavin Barraclough <barraclough@apple.com> 743 744 Reviewed by NOBODY (build fix). 745 746 Fix windows build II. 747 748 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 749 7502010-05-19 Gavin Barraclough <barraclough@apple.com> 751 752 Reviewed by NOBODY (build fix). 753 754 Fix windows build. 755 756 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 757 7582010-05-19 Gavin Barraclough <barraclough@apple.com> 759 760 Reviewed by Oliver Hunt. 761 762 Bug 39393 - JSFunction need not be a subclass of InternalFunction. 763 764 re-landing r59800. 765 766 * JavaScriptCore.exp: 767 * interpreter/Interpreter.cpp: 768 (JSC::Interpreter::retrieveCaller): 769 (JSC::Interpreter::findFunctionCallFrame): 770 * interpreter/Interpreter.h: 771 * profiler/Profiler.cpp: 772 (JSC::Profiler::createCallIdentifier): 773 * runtime/FunctionPrototype.cpp: 774 (JSC::functionProtoFuncToString): 775 * runtime/JSFunction.cpp: 776 (JSC::): 777 (JSC::JSFunction::JSFunction): 778 (JSC::JSFunction::name): 779 (JSC::JSFunction::displayName): 780 (JSC::JSFunction::calculatedDisplayName): 781 * runtime/JSFunction.h: 782 * runtime/JSObject.cpp: 783 (JSC::JSObject::putDirectFunction): 784 (JSC::JSObject::putDirectFunctionWithoutTransition): 785 * runtime/JSObject.h: 786 * runtime/Lookup.cpp: 787 (JSC::setUpStaticFunctionSlot): 788 7892010-05-19 Gavin Barraclough <barraclough@apple.com> 790 791 Reviewed by NOBODY (build fix). 792 793 Reverting r59800, this b0rked stuff. 794 795 * JavaScriptCore.exp: 796 * interpreter/Interpreter.cpp: 797 (JSC::Interpreter::retrieveCaller): 798 (JSC::Interpreter::findFunctionCallFrame): 799 * interpreter/Interpreter.h: 800 * profiler/Profiler.cpp: 801 (JSC::Profiler::createCallIdentifier): 802 * runtime/FunctionPrototype.cpp: 803 (JSC::functionProtoFuncToString): 804 * runtime/JSFunction.cpp: 805 (JSC::): 806 (JSC::JSFunction::JSFunction): 807 * runtime/JSFunction.h: 808 * runtime/JSObject.cpp: 809 * runtime/JSObject.h: 810 * runtime/Lookup.cpp: 811 (JSC::setUpStaticFunctionSlot): 812 8132010-05-19 Gavin Barraclough <barraclough@apple.com> 814 815 Reviewed by Oliver Hunt. 816 817 Bug 39393 - JSFunction need not be a subclass of InternalFunction. 818 819 This may prevent us from introducing a more useful parent class to 820 JSFunction, e.g. a JSObject that holds an executable, which could 821 also reference an eval or program executable. 822 823 * JavaScriptCore.exp: 824 * interpreter/Interpreter.cpp: 825 (JSC::Interpreter::retrieveCaller): 826 (JSC::Interpreter::findFunctionCallFrame): 827 * interpreter/Interpreter.h: 828 * profiler/Profiler.cpp: 829 (JSC::Profiler::createCallIdentifier): 830 * runtime/FunctionPrototype.cpp: 831 (JSC::functionProtoFuncToString): 832 * runtime/JSFunction.cpp: 833 (JSC::): 834 (JSC::JSFunction::JSFunction): 835 (JSC::JSFunction::name): 836 (JSC::JSFunction::displayName): 837 (JSC::JSFunction::calculatedDisplayName): 838 * runtime/JSFunction.h: 839 * runtime/JSObject.cpp: 840 (JSC::JSObject::putDirectFunction): 841 (JSC::JSObject::putDirectFunctionWithoutTransition): 842 * runtime/JSObject.h: 843 * runtime/Lookup.cpp: 844 (JSC::setUpStaticFunctionSlot): 845 8462010-05-19 Oliver Hunt <oliver@apple.com> 847 848 Reviewed by Geoffrey Garen. 849 850 emitJumpIfNotJSCell should special case constant immediate values 851 https://bugs.webkit.org/show_bug.cgi?id=39392 852 <rdar://problem/8001324> 853 854 Make emitJumpSlowCaseIfNotJSCell special case constant immediate 855 values, in addition to the immediate JSCell optimisation. 856 857 Also add assertions to make sure no one else produces code that 858 attempts to load constants from the register file. 859 860 * jit/JITInlineMethods.h: 861 (JSC::JIT::emitJumpSlowCaseIfNotJSCell): 862 * jit/JSInterfaceJIT.h: 863 (JSC::JSInterfaceJIT::emitJumpIfNotJSCell): 864 (JSC::JSInterfaceJIT::emitLoadInt32): 865 (JSC::JSInterfaceJIT::tagFor): 866 (JSC::JSInterfaceJIT::payloadFor): 867 (JSC::JSInterfaceJIT::emitLoadDouble): 868 (JSC::JSInterfaceJIT::addressFor): 869 * jit/ThunkGenerators.cpp: 870 8712010-05-19 Geoffrey Garen <ggaren@apple.com> 872 873 Reviewed by Sam Weinig. 874 875 Slight refactoring to CodeBlock bytecode access 876 https://bugs.webkit.org/show_bug.cgi?id=39384 877 878 * bytecode/CodeBlock.h: 879 (JSC::CodeBlock::bytecodeOffset): 880 * interpreter/CallFrame.h: 881 (JSC::ExecState::returnPC): 882 * interpreter/Interpreter.cpp: 883 (JSC::Interpreter::unwindCallFrame): 884 (JSC::Interpreter::retrieveLastCaller): Moved bytecode access into a 885 CodeBlock helper function. Changed CallFrame to accurately represent 886 how the returnPC is stored in JIT vs Interpreter. 887 8882010-05-19 Geoffrey Garen <ggaren@apple.com> 889 890 Reviewed by Sam Weinig. 891 892 Don't build the Interpreter into JIT builds. 893 https://bugs.webkit.org/show_bug.cgi?id=39373 894 895 SunSpider says no change. 896 897 * interpreter/Interpreter.cpp: 898 (JSC::Interpreter::privateExecute): 899 * interpreter/Interpreter.h: 900 * wtf/Platform.h: 901 9022010-05-19 Geoffrey Garen <ggaren@apple.com> 903 904 Reviewed by Gavin Barraclough. 905 906 Standardized naming: "bytecode offset" wins; "bytecode index" loses. 907 908 Also renamed getBytecodeOffset to bytecodeOffset, to match the WebKit 909 naming style. 910 911 * bytecode/CodeBlock.h: 912 (JSC::CallReturnOffsetToBytecodeOffset::CallReturnOffsetToBytecodeOffset): 913 (JSC::getCallReturnOffset): 914 (JSC::CodeBlock::bytecodeOffset): 915 (JSC::CodeBlock::callReturnIndexVector): 916 * interpreter/Interpreter.cpp: 917 (JSC::bytecodeOffsetForPC): 918 * jit/JIT.cpp: 919 (JSC::JIT::JIT): 920 (JSC::JIT::privateCompileMainPass): 921 (JSC::JIT::privateCompileLinkPass): 922 (JSC::JIT::privateCompileSlowCases): 923 (JSC::JIT::privateCompile): 924 * jit/JIT.h: 925 (JSC::CallRecord::CallRecord): 926 (JSC::JumpTable::JumpTable): 927 (JSC::SwitchRecord::SwitchRecord): 928 * jit/JITCall.cpp: 929 (JSC::JIT::compileOpCallVarargsSlowCase): 930 (JSC::JIT::compileOpCall): 931 * jit/JITInlineMethods.h: 932 (JSC::JIT::emitNakedCall): 933 (JSC::JIT::addSlowCase): 934 (JSC::JIT::addJump): 935 (JSC::JIT::emitJumpSlowToHot): 936 (JSC::JIT::isLabeled): 937 (JSC::JIT::map): 938 (JSC::JIT::unmap): 939 (JSC::JIT::isMapped): 940 (JSC::JIT::getMappedPayload): 941 (JSC::JIT::getMappedTag): 942 (JSC::JIT::emitGetVirtualRegister): 943 * jit/JITOpcodes.cpp: 944 (JSC::JIT::emit_op_switch_imm): 945 (JSC::JIT::emit_op_switch_char): 946 (JSC::JIT::emit_op_switch_string): 947 (JSC::JIT::emit_op_new_error): 948 * jit/JITOpcodes32_64.cpp: 949 (JSC::JIT::emit_op_mov): 950 (JSC::JIT::emit_op_get_global_var): 951 (JSC::JIT::emit_op_put_global_var): 952 (JSC::JIT::emit_op_get_scoped_var): 953 (JSC::JIT::emit_op_put_scoped_var): 954 (JSC::JIT::emit_op_to_primitive): 955 (JSC::JIT::emit_op_resolve_global): 956 (JSC::JIT::emit_op_to_jsnumber): 957 (JSC::JIT::emit_op_catch): 958 (JSC::JIT::emit_op_switch_imm): 959 (JSC::JIT::emit_op_switch_char): 960 (JSC::JIT::emit_op_switch_string): 961 (JSC::JIT::emit_op_new_error): 962 (JSC::JIT::emit_op_convert_this): 963 * jit/JITPropertyAccess.cpp: 964 (JSC::JIT::emit_op_method_check): 965 (JSC::JIT::emitSlow_op_method_check): 966 * jit/JITPropertyAccess32_64.cpp: 967 (JSC::JIT::emit_op_method_check): 968 (JSC::JIT::emitSlow_op_method_check): 969 (JSC::JIT::emit_op_get_by_val): 970 (JSC::JIT::emit_op_get_by_id): 971 (JSC::JIT::emit_op_get_by_pname): 972 * jit/JITStubCall.h: 973 (JSC::JITStubCall::call): 974 * jit/JITStubs.cpp: 975 (JSC::DEFINE_STUB_FUNCTION): 976 9772010-05-19 Peter Varga <pvarga@inf.u-szeged.hu> 978 979 Reviewed by Darin Adler. 980 981 The backtrackParenthesesOnceEnd function should store the start position 982 of the subpattern's non-greedy match at the place of the subpattern's 983 start index in the output array instead of the place of the 984 subpattern's end index. 985 Add layout test for checking of non-greedy matching of subpattern in 986 regular expressions. 987 https://bugs.webkit.org/show_bug.cgi?id=39289 988 989 * yarr/RegexInterpreter.cpp: 990 (JSC::Yarr::Interpreter::backtrackParenthesesOnceEnd): 991 9922010-05-18 Gavin Barraclough <barraclough@apple.com> 993 994 Reviewed by Geoff Garen. 995 996 Bug 39343 - Dynamically generate a native call thunk per NativeFunction 997 998 https://bugs.webkit.org/show_bug.cgi?id=39252 regressed performance on i386, 999 by adding an extra indirection to making a native call. By introducing per- 1000 NativeFunction thunks we can hard code the function pointer into the thunk 1001 so that it need not be loaded from the callee. 1002 1003 * jit/JIT.h: 1004 (JSC::JIT::compileCTINativeCall): 1005 * jit/JITOpcodes.cpp: 1006 (JSC::JIT::privateCompileCTINativeCall): 1007 * jit/JITOpcodes32_64.cpp: 1008 (JSC::JIT::privateCompileCTINativeCall): 1009 * jit/JITPropertyAccess.cpp: 1010 (JSC::JIT::emitSlow_op_get_by_val): 1011 * jit/JITPropertyAccess32_64.cpp: 1012 (JSC::JIT::emitSlow_op_get_by_val): 1013 * jit/JITStubs.cpp: 1014 (JSC::JITThunks::ctiStub): 1015 (JSC::JITThunks::hostFunctionStub): 1016 * jit/JITStubs.h: 1017 * jit/SpecializedThunkJIT.h: 1018 (JSC::SpecializedThunkJIT::finalize): 1019 * jit/ThunkGenerators.cpp: 1020 (JSC::charCodeAtThunkGenerator): 1021 (JSC::charAtThunkGenerator): 1022 (JSC::fromCharCodeThunkGenerator): 1023 (JSC::sqrtThunkGenerator): 1024 (JSC::powThunkGenerator): 1025 * runtime/JSFunction.cpp: 1026 (JSC::JSFunction::JSFunction): 1027 * runtime/JSGlobalData.cpp: 1028 (JSC::JSGlobalData::getHostFunction): 1029 * runtime/JSGlobalData.h: 1030 (JSC::JSGlobalData::getCTIStub): 1031 * runtime/Lookup.cpp: 1032 (JSC::setUpStaticFunctionSlot): 1033 * runtime/StringConstructor.cpp: 1034 (JSC::StringConstructor::StringConstructor): 1035 * wtf/Platform.h: 1036 10372010-05-18 Geoffrey Garen <ggaren@apple.com> 1038 1039 Reviewed by Darin Adler. 1040 1041 Simplified handling of 'arguments' -- 1.2% SunSpider speedup 1042 https://bugs.webkit.org/show_bug.cgi?id=39200 1043 1044 Removed the reserved OptionalCalleeArguments slot from the CallFrame. 1045 Now, slots for 'arguments' are allocated and initialized only by 1046 functions that might need them. 1047 1048 * bytecode/CodeBlock.cpp: 1049 (JSC::CodeBlock::dump): Updated for new bytecode operands. 1050 1051 (JSC::CodeBlock::CodeBlock): 1052 * bytecode/CodeBlock.h: 1053 (JSC::unmodifiedArgumentsRegister): Added a helper function for mapping 1054 from the arguments register to its unmodified counterpart. 1055 1056 (JSC::CodeBlock::setArgumentsRegister): 1057 (JSC::CodeBlock::argumentsRegister): 1058 (JSC::CodeBlock::usesArguments): Changed from a "usesArguments" bool to 1059 an optional int index representing the arguments register. 1060 1061 * bytecode/Opcode.h: Updated for new bytecode operands. 1062 1063 * bytecompiler/BytecodeGenerator.cpp: 1064 (JSC::BytecodeGenerator::addVar): Factored out a helper function for 1065 allocating an anonymous var. 1066 1067 (JSC::BytecodeGenerator::BytecodeGenerator): Merged / simplified some 1068 arguments vs activation logic, and added code to allocate the arguments 1069 registers when needed. 1070 1071 (JSC::BytecodeGenerator::createArgumentsIfNecessary): Updated for new bytecode operands. 1072 1073 (JSC::BytecodeGenerator::emitCallEval): No need to create the arguments 1074 object before calling eval; the activation object will lazily create the 1075 arguments object if eval resolves it. 1076 1077 (JSC::BytecodeGenerator::emitReturn): Updated for new bytecode operands. 1078 1079 (JSC::BytecodeGenerator::emitPushScope): 1080 (JSC::BytecodeGenerator::emitPushNewScope): Ditto emitCallEval. 1081 1082 * bytecompiler/BytecodeGenerator.h: 1083 (JSC::BytecodeGenerator::addVar): Factored out a helper function for 1084 allocating an anonymous var. 1085 1086 (JSC::BytecodeGenerator::registerFor): No more need for special handling 1087 of the arguments registers; they're allocated just like normal registers 1088 now. 1089 1090 * interpreter/CallFrame.h: 1091 (JSC::ExecState::callerFrame): 1092 (JSC::ExecState::init): 1093 * interpreter/CallFrameClosure.h: 1094 (JSC::CallFrameClosure::resetCallFrame): Nixed optionalCalleeArguments. 1095 1096 * interpreter/Interpreter.cpp: 1097 (JSC::Interpreter::dumpRegisters): 1098 (JSC::Interpreter::unwindCallFrame): 1099 (JSC::Interpreter::privateExecute): 1100 (JSC::Interpreter::retrieveArguments): Opcodes accessing 'arguments' now 1101 take operands specifying registers, just like all other opcodes. 1102 JSActivation::copyRegisters is no longer responsible for tearing off the 1103 arguments object; instead, the VM is responsible for both. 1104 1105 Also, a behavior change: Each access to f.arguments creates a new object, 1106 unless f itself uses 'arguments'. This matches Chrome, and is necessary 1107 for the optimization. f.arguments is a nonstandard, deprecated feature, 1108 so high fidelity to a given implementation is not necessarily a goal. 1109 Also, as illustrated by the new test case, the identity of f.arguments 1110 has been broken since 2008, except in the case where f itself accesses 1111 f.arguments -- but nobody seemed to notice. So, hopefully this change won't 1112 break the web. 1113 1114 * interpreter/Register.h: Nixed the special arguments accessor. It's no 1115 longer needed. 1116 1117 * interpreter/RegisterFile.h: 1118 (JSC::RegisterFile::): 1119 * jit/JITCall.cpp: 1120 (JSC::JIT::compileOpCallInitializeCallFrame): 1121 (JSC::JIT::compileOpCall): 1122 * jit/JITOpcodes.cpp: 1123 (JSC::JIT::emit_op_tear_off_activation): 1124 (JSC::JIT::emit_op_tear_off_arguments): 1125 (JSC::JIT::emit_op_create_arguments): 1126 (JSC::JIT::emit_op_init_arguments): 1127 * jit/JITOpcodes32_64.cpp: 1128 (JSC::JIT::emit_op_tear_off_activation): 1129 (JSC::JIT::emit_op_tear_off_arguments): 1130 (JSC::JIT::emit_op_create_arguments): 1131 (JSC::JIT::emit_op_init_arguments): The actual optimization: Removed 1132 OptionalCalleeArguments from the callframe slot. Now, it doesn't need 1133 to be initialized for most calls. 1134 1135 * jit/JITStubs.cpp: 1136 (JSC::DEFINE_STUB_FUNCTION): 1137 * jit/JITStubs.h: 1138 (JSC::): Updated stubs to support arbitrary 'arguments' registers, 1139 instead of hard-coding something in the call frame. 1140 1141 * runtime/Arguments.h: 1142 (JSC::JSActivation::copyRegisters): Removed some obfuscatory abstraction. 1143 1144 * runtime/Executable.h: 1145 (JSC::FunctionExecutable::generatedByteCode): Added a helper for accessing 1146 the 'arguments' register. In a future patch, that kind of data should 1147 probably move out of CodeBlock and into Executable. 1148 1149 * runtime/JSActivation.cpp: 1150 (JSC::JSActivation::getOwnPropertySlot): 1151 (JSC::JSActivation::argumentsGetter): 1152 * runtime/JSActivation.h: Simplified / fixed access to 'arguments' via 1153 the activation object. It now implements the same behavior implemented 1154 by optimized variable access in the VM. This simplifies some other 1155 things, too -- like eval code generation. 1156 11572010-05-18 Sam Weinig <sam@webkit.org> 1158 1159 Try to fix the windows build. 1160 1161 * API/JSStringRefBSTR.cpp: 1162 11632010-05-18 Anders Carlsson <andersca@apple.com> 1164 1165 Reviewed by Sam Weinig. 1166 1167 Add an inlineCapacity template parameter to ListHashSet and use it to shrink the positioned object list hash set. 1168 https://bugs.webkit.org/show_bug.cgi?id=39304 1169 <rdar://problem/7998366> 1170 1171 Add an inlineCapacity template parameter to ListHashSet. 1172 1173 * wtf/ListHashSet.h: 1174 (WTF::::ListHashSet): 1175 (WTF::::operator): 1176 (WTF::::swap): 1177 (WTF::::~ListHashSet): 1178 (WTF::::size): 1179 (WTF::::capacity): 1180 (WTF::::isEmpty): 1181 (WTF::::begin): 1182 (WTF::::end): 1183 (WTF::::find): 1184 (WTF::::contains): 1185 (WTF::::add): 1186 (WTF::::insertBefore): 1187 (WTF::::remove): 1188 (WTF::::clear): 1189 (WTF::::unlinkAndDelete): 1190 (WTF::::appendNode): 1191 (WTF::::insertNodeBefore): 1192 (WTF::::deleteAllNodes): 1193 (WTF::::makeIterator): 1194 (WTF::::makeConstIterator): 1195 (WTF::deleteAllValues): 1196 11972010-05-18 Gabor Loki <loki@webkit.org> 1198 1199 Reviewed by Darin Adler. 1200 1201 Fix the interpreter after r59637 1202 https://bugs.webkit.org/show_bug.cgi?id=39287 1203 1204 * runtime/Executable.h: 1205 * runtime/JSFunction.cpp: 1206 (JSC::JSFunction::JSFunction): 1207 (JSC::JSFunction::getCallData): 1208 * runtime/JSGlobalData.cpp: 1209 * runtime/JSGlobalData.h: 1210 12112010-05-17 Gavin Barraclough <barraclough@apple.com> 1212 1213 Reviewed by Geoff Garen. 1214 1215 Oops, meant ebx not eax there. Fix Qt (and probably Win too). 1216 1217 * jit/JITOpcodes32_64.cpp: 1218 12192010-05-17 Gavin Barraclough <barraclough@apple.com> 1220 1221 Reviewed by Geoff Garen. 1222 1223 Bug 39252 - Move host/native JSFunction's NativeFunction onto NativeExecutable. 1224 1225 Currently host functions reuse JSFunction's ScopeChain as storage for their 1226 NativeFunction (the C function pointer to the host function implementation). 1227 Instead, move this onto NativeExecutable. This will allow host functions to 1228 have a scopechain (which will be implemented as a separate patch). 1229 1230 * JavaScriptCore.xcodeproj/project.pbxproj: 1231 * jit/JIT.h: 1232 * jit/JITCall.cpp: 1233 (JSC::JIT::compileOpCallInitializeCallFrame): 1234 (JSC::JIT::compileOpCall): 1235 * jit/JITOpcodes.cpp: 1236 (JSC::JIT::privateCompileCTIMachineTrampolines): 1237 * jit/JITOpcodes32_64.cpp: 1238 * jit/JITPropertyAccess.cpp: 1239 (JSC::JIT::stringGetByValStubGenerator): 1240 (JSC::JIT::emitSlow_op_get_by_val): 1241 * jit/JITPropertyAccess32_64.cpp: 1242 (JSC::JIT::stringGetByValStubGenerator): 1243 (JSC::JIT::emitSlow_op_get_by_val): 1244 * jit/JITStubs.cpp: 1245 (JSC::JITThunks::specializedThunk): 1246 * jit/JITStubs.h: 1247 (JSC::JITThunks::ctiNativeCall): 1248 * jit/SpecializedThunkJIT.h: 1249 (JSC::SpecializedThunkJIT::finalize): 1250 * jit/ThunkGenerators.cpp: 1251 (JSC::charCodeAtThunkGenerator): 1252 (JSC::charAtThunkGenerator): 1253 (JSC::fromCharCodeThunkGenerator): 1254 (JSC::sqrtThunkGenerator): 1255 (JSC::powThunkGenerator): 1256 * jit/ThunkGenerators.h: 1257 * runtime/Executable.h: 1258 (JSC::NativeExecutable::create): 1259 (JSC::NativeExecutable::function): 1260 (JSC::NativeExecutable::NativeExecutable): 1261 (JSC::JSFunction::nativeFunction): 1262 * runtime/JSFunction.cpp: 1263 (JSC::JSFunction::JSFunction): 1264 (JSC::JSFunction::~JSFunction): 1265 (JSC::JSFunction::markChildren): 1266 (JSC::JSFunction::getCallData): 1267 (JSC::JSFunction::call): 1268 (JSC::JSFunction::getOwnPropertySlot): 1269 (JSC::JSFunction::getConstructData): 1270 (JSC::JSFunction::construct): 1271 * runtime/JSFunction.h: 1272 (JSC::JSFunction::scope): 1273 * runtime/JSGlobalData.h: 1274 (JSC::JSGlobalData::getThunk): 1275 * runtime/Lookup.cpp: 1276 (JSC::setUpStaticFunctionSlot): 1277 * runtime/StringConstructor.cpp: 1278 (JSC::StringConstructor::StringConstructor): 1279 12802010-05-17 Patrick Gansterer <paroga@paroga.com> 1281 1282 Reviewed by Laszlo Gombos. 1283 1284 [Qt] Remove WinCE endian dedection. 1285 https://bugs.webkit.org/show_bug.cgi?id=38511 1286 1287 Windows CE supports little-endian format only. 1288 Correct dedection was added in r57804. 1289 1290 * wtf/Platform.h: 1291 12922010-05-16 Simon Fraser <simon.fraser@apple.com> 1293 1294 Reviewed by Darin Adler. 1295 1296 Reduce the size of FunctionExecutable 1297 https://bugs.webkit.org/show_bug.cgi?id=39180 1298 1299 Change m_numVariables from a size_t to 31 bits in a bitfield, 1300 packed with another bit for m_forceUsesArguments (which in turn 1301 get packed with the base class). 1302 1303 Reduces the size of FunctionExecutable from 160 to 152 bytes. 1304 1305 * runtime/Executable.h: 1306 (JSC::FunctionExecutable::variableCount): 1307 (JSC::FunctionExecutable::FunctionExecutable): 1308 13092010-05-15 Oliver Hunt <oliver@apple.com> 1310 1311 Reviewed by Maciej Stachowiak. 1312 1313 Incorrect codegen for slowcase of < in 64-bit 1314 https://bugs.webkit.org/show_bug.cgi?id=39151 1315 1316 Call the correct stud for the slowcases of the < operator. 1317 1318 * jit/JITArithmetic.cpp: 1319 (JSC::JIT::emitSlow_op_jnless): 1320 (JSC::JIT::emitSlow_op_jless): 1321 13222010-05-15 Leo Yang <leo.yang@torchmobile.com.cn> 1323 1324 Reviewed by Darin Adler. 1325 1326 Fix bug https://bugs.webkit.org/show_bug.cgi?id=38890 1327 1328 * runtime/DatePrototype.cpp: 1329 (JSC::formatLocaleDate): 1330 13312010-05-15 Leandro Pereira <leandro@profusion.mobi> 1332 1333 Reviewed by Adam Treat. 1334 1335 [EFL] Add build system for the EFL port. 1336 http://webkit.org/b/37945 1337 1338 * CMakeLists.txt: Added. 1339 * CMakeListsEfl.txt: Added. 1340 * config.h: Add conditional to look for cmakeconfig.h. 1341 * jsc/CMakeLists.txt: Added. 1342 * jsc/CMakeListsEfl.txt: Added. 1343 * wtf/CMakeLists.txt: Added. 1344 * wtf/CMakeListsEfl.txt: Added. 1345 13462010-05-15 Chao-ying Fu <fu@mips.com> 1347 1348 Reviewed by Oliver Hunt. 1349 1350 Update MIPS JIT for unsigned right shift, Math.sqrt, load16 1351 https://bugs.webkit.org/show_bug.cgi?id=38412 1352 1353 Fixed MIPS build failure. 1354 1355 * assembler/MIPSAssembler.h: 1356 (JSC::MIPSAssembler::srl): 1357 (JSC::MIPSAssembler::srlv): 1358 (JSC::MIPSAssembler::sqrtd): 1359 * assembler/MacroAssemblerMIPS.h: 1360 (JSC::MacroAssemblerMIPS::urshift32): 1361 (JSC::MacroAssemblerMIPS::sqrtDouble): 1362 (JSC::MacroAssemblerMIPS::load16): 1363 (JSC::MacroAssemblerMIPS::supportsFloatingPointSqrt): 1364 * jit/JSInterfaceJIT.h: 1365 13662010-05-15 Ilya Tikhonovsky <loislo@chromium.org> 1367 1368 Reviewed by Geoffrey Garen. 1369 1370 WebInspector: JSC Should provide heap size info for Timeline panel. 1371 https://bugs.webkit.org/show_bug.cgi?id=38420 1372 1373 * JavaScriptCore.exp: 1374 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 1375 * runtime/Collector.cpp: 1376 (JSC::Heap::size): 1377 * runtime/Collector.h: 1378 13792010-05-15 Dave Tapuska <dtapuska@rim.com> 1380 1381 Reviewed by Darin Adler. 1382 1383 Add a new define for we need strict arena alignment. ARMv5 with 1384 double word instructions set this value. 1385 1386 * wtf/Platform.h: 1387 13882010-05-14 Stephanie Lewis <slewis@apple.com> 1389 1390 Rubber-stamped by Mark Rowe. 1391 1392 Update order files. 1393 1394 * JavaScriptCore.order: 1395 13962010-05-14 Jedrzej Nowacki <jedrzej.nowacki@nokia.com> 1397 1398 Reviewed by Kenneth Rohde Christiansen. 1399 1400 Implementation of QtScript API that provide method to convert a value to an object. 1401 1402 [Qt] QtScript is missing toObject API 1403 https://bugs.webkit.org/show_bug.cgi?id=36594 1404 1405 * qt/api/qscriptengine.cpp: 1406 (QScriptEngine::toObject): 1407 * qt/api/qscriptengine.h: 1408 * qt/api/qscriptvalue.cpp: 1409 (QScriptValue::toObject): 1410 * qt/api/qscriptvalue.h: 1411 * qt/api/qscriptvalue_p.h: 1412 (QScriptValuePrivate::toObject): 1413 * qt/tests/qscriptengine/tst_qscriptengine.cpp: 1414 (tst_QScriptEngine::toObject): 1415 (tst_QScriptEngine::toObjectTwoEngines): 1416 * qt/tests/qscriptvalue/tst_qscriptvalue.cpp: 1417 (tst_QScriptValue::toObjectSimple): 1418 * qt/tests/qscriptvalue/tst_qscriptvalue.h: 1419 14202010-05-14 Chao-ying Fu <fu@mips.com> 1421 1422 Reviewed by Oliver Hunt. 1423 1424 String Indexing Failure on JSVALUE32 targets 1425 https://bugs.webkit.org/show_bug.cgi?id=39034 1426 1427 Remove zero-extend/shift-right code on regT1, because we already have 1428 it in emit_op_get_by_val(). 1429 1430 * jit/JITPropertyAccess.cpp: 1431 (JSC::JIT::stringGetByValStubGenerator): 1432 14332010-05-10 Jocelyn Turcotte <jocelyn.turcotte@nokia.com> 1434 1435 Reviewed by Alexey Proskuryakov. 1436 1437 Fix a VS2010 assert in std::copy 1438 https://bugs.webkit.org/show_bug.cgi?id=38630 1439 1440 The assert complains that the output iterator is null. 1441 1442 * wtf/Vector.h: 1443 (WTF::::operator): 1444 14452010-05-13 Maciej Stachowiak <mjs@apple.com> 1446 1447 No review, attempted build fix. 1448 1449 Try to fix Tiger build with some gratuitous initialization of 1450 seemingly uninitialized variables. 1451 1452 * wtf/text/StringImpl.h: 1453 (WebCore::StringImpl::tryCreateUninitialized): 1454 14552010-05-13 Maciej Stachowiak <mjs@apple.com> 1456 1457 Reviewed by Mark Rowe. 1458 1459 Various JavaScript string optimizations 1460 https://bugs.webkit.org/show_bug.cgi?id=39051 1461 1462 Approximately 1% SunSpider speedup. 1463 1464 * runtime/ArrayPrototype.cpp: 1465 (JSC::arrayProtoFuncJoin): Remove branches from the hot code path 1466 by moving the first pass outside the loop, and duplicating the hot loop 1467 to extract the loop-invariant branch. 1468 * runtime/RegExp.cpp: 1469 (JSC::RegExp::match): resize ovector to 0 instead of clearing to avoid 1470 thrash in case of large matches. 1471 * runtime/RegExpConstructor.h: 1472 (JSC::RegExpConstructor::performMatch): Mark ALWAYS_INLINE to make the 1473 compiler respect our authority. 1474 * runtime/StringPrototype.cpp: 1475 (JSC::jsSpliceSubstringsWithSeparators): Inline. 1476 (JSC::stringProtoFuncSubstring): Rewrite boundary condition checks to 1477 reduce the number of floating point comparisons and branches. 1478 14792010-05-12 Gavin Barraclough <barraclough@apple.com> 1480 1481 Reviewed by Oliver Hunt. 1482 1483 https://bugs.webkit.org/show_bug.cgi?id=39039 1484 Provide support for separate bytecode/JIT code translations for call/construct usage 1485 This will allow us to produce code generated specifically for use as a constructor, not for general function use. 1486 1487 * bytecode/CodeBlock.cpp: 1488 (JSC::CodeBlock::CodeBlock): 1489 (JSC::CodeBlock::unlinkCallers): 1490 * bytecode/CodeBlock.h: 1491 (JSC::CodeBlock::getBytecodeIndex): 1492 (JSC::CodeBlock::getJITCode): 1493 (JSC::CodeBlock::executablePool): 1494 (JSC::GlobalCodeBlock::GlobalCodeBlock): 1495 (JSC::FunctionCodeBlock::FunctionCodeBlock): 1496 * interpreter/Interpreter.cpp: 1497 (JSC::Interpreter::executeCall): 1498 (JSC::Interpreter::executeConstruct): 1499 (JSC::Interpreter::prepareForRepeatCall): 1500 (JSC::Interpreter::execute): 1501 (JSC::Interpreter::privateExecute): 1502 * interpreter/Interpreter.h: 1503 * jit/JIT.cpp: 1504 (JSC::JIT::unlinkCallOrConstruct): 1505 (JSC::JIT::linkConstruct): 1506 * jit/JIT.h: 1507 * jit/JITCall.cpp: 1508 (JSC::JIT::compileOpCall): 1509 (JSC::JIT::compileOpCallSlowCase): 1510 * jit/JITOpcodes.cpp: 1511 (JSC::JIT::privateCompileCTIMachineTrampolines): 1512 * jit/JITOpcodes32_64.cpp: 1513 (JSC::JIT::privateCompileCTIMachineTrampolines): 1514 * jit/JITPropertyAccess.cpp: 1515 (JSC::JIT::emitSlow_op_get_by_val): 1516 * jit/JITPropertyAccess32_64.cpp: 1517 (JSC::JIT::emitSlow_op_get_by_val): 1518 * jit/JITStubs.cpp: 1519 (JSC::DEFINE_STUB_FUNCTION): 1520 * jit/JITStubs.h: 1521 (JSC::JITThunks::ctiVirtualConstructLink): 1522 (JSC::JITThunks::ctiVirtualConstruct): 1523 (JSC::): 1524 * jit/SpecializedThunkJIT.h: 1525 (JSC::SpecializedThunkJIT::finalize): 1526 * runtime/Arguments.h: 1527 (JSC::JSActivation::copyRegisters): 1528 * runtime/ArrayPrototype.cpp: 1529 (JSC::isNumericCompareFunction): 1530 * runtime/Executable.cpp: 1531 (JSC::FunctionExecutable::~FunctionExecutable): 1532 (JSC::FunctionExecutable::compileForCall): 1533 (JSC::FunctionExecutable::compileForConstruct): 1534 (JSC::EvalExecutable::generateJITCode): 1535 (JSC::ProgramExecutable::generateJITCode): 1536 (JSC::FunctionExecutable::generateJITCodeForCall): 1537 (JSC::FunctionExecutable::generateJITCodeForConstruct): 1538 (JSC::FunctionExecutable::markAggregate): 1539 (JSC::FunctionExecutable::reparseExceptionInfo): 1540 (JSC::EvalExecutable::reparseExceptionInfo): 1541 (JSC::FunctionExecutable::recompile): 1542 * runtime/Executable.h: 1543 (JSC::ExecutableBase::ExecutableBase): 1544 (JSC::ExecutableBase::isHostFunction): 1545 (JSC::ExecutableBase::generatedJITCodeForCall): 1546 (JSC::ExecutableBase::generatedJITCodeForConstruct): 1547 (JSC::NativeExecutable::NativeExecutable): 1548 (JSC::EvalExecutable::jitCode): 1549 (JSC::ProgramExecutable::jitCode): 1550 (JSC::FunctionExecutable::bytecodeForCall): 1551 (JSC::FunctionExecutable::isGeneratedForCall): 1552 (JSC::FunctionExecutable::generatedBytecodeForCall): 1553 (JSC::FunctionExecutable::bytecodeForConstruct): 1554 (JSC::FunctionExecutable::isGeneratedForConstruct): 1555 (JSC::FunctionExecutable::generatedBytecodeForConstruct): 1556 (JSC::FunctionExecutable::symbolTable): 1557 (JSC::FunctionExecutable::FunctionExecutable): 1558 (JSC::FunctionExecutable::jitCodeForCall): 1559 (JSC::FunctionExecutable::jitCodeForConstruct): 1560 * runtime/JSActivation.h: 1561 (JSC::JSActivation::JSActivationData::JSActivationData): 1562 * runtime/JSFunction.cpp: 1563 (JSC::JSFunction::~JSFunction): 1564 (JSC::JSFunction::call): 1565 (JSC::JSFunction::construct): 1566 * runtime/JSGlobalData.cpp: 1567 (JSC::JSGlobalData::numericCompareFunction): 1568 15692010-05-12 Oliver Hunt <oliver@apple.com> 1570 1571 build fix 1572 1573 <rdar://problem/7977960> REGRESSION (r59282): Tiger PPC build broken 1574 1575 * runtime/UString.h: 1576 (JSC::tryMakeString): 1577 15782010-05-12 Oliver Hunt <oliver@apple.com> 1579 1580 Reviewed by Gavin Barraclough. 1581 1582 <rdar://7969718> Incorrect 32-bit codegen for op_jnlesseq 1583 https://bugs.webkit.org/show_bug.cgi?id=39038 1584 1585 Correct the use of incorrect comparison operators. 1586 1587 * jit/JITArithmetic32_64.cpp: 1588 (JSC::JIT::emit_op_jlesseq): 1589 (JSC::JIT::emitBinaryDoubleOp): 1590 15912010-05-12 Caio Marcelo de Oliveira Filho <caio.oliveira@openbossa.org> 1592 1593 Reviewed by Eric Seidel. 1594 1595 JavaScriptCore: fix build when some optimizations are disabled 1596 https://bugs.webkit.org/show_bug.cgi?id=38800 1597 1598 Remove outdated UNUSED_PARAMs when using JIT but with 1599 ENABLE_JIT_OPTIMIZE_CALL and ENABLE_JIT_OPTIMIZE_PROPERTY_ACCESS 1600 disabled. 1601 1602 * jit/JITOpcodes.cpp: 1603 * jit/JITOpcodes32_64.cpp: 1604 16052010-05-12 Oliver Hunt <oliver@apple.com> 1606 1607 Reviewed by Gavin Barraclough. 1608 1609 Work around GCC stupidity by either explicitly inlining or by 1610 using ALWAYS_INLINE 1611 1612 * runtime/StringPrototype.cpp: 1613 (JSC::stringProtoFuncReplace): 1614 * runtime/UString.h: 1615 (JSC::UString::~UString): 1616 * wtf/RefPtr.h: 1617 (WTF::RefPtr::RefPtr): 1618 (WTF::RefPtr::~RefPtr): 1619 * wtf/text/StringImpl.h: 1620 (WebCore::StringImpl::create): 1621 (WebCore::StringImpl::tryCreateUninitialized): 1622 16232010-05-12 Laszlo Gombos <laszlo.1.gombos@nokia.com> 1624 1625 Reviewed by Kenneth Rohde Christiansen. 1626 1627 [Qt] Detect debug mode consistently 1628 https://bugs.webkit.org/show_bug.cgi?id=38863 1629 1630 * JavaScriptCore.pri: 1631 * JavaScriptCore.pro: 1632 * jsc.pro: 1633 * qt/api/QtScript.pro: 1634 16352010-05-12 Peter Varga <pvarga@inf.u-szeged.hu> 1636 1637 Reviewed by Gavin Barraclough. 1638 1639 Remove a direct jump in generatePatternCharacterGreedy() and 1640 generatePatternCharacterClassGreedy() functions if quantityCount of the term 1641 is unlimited. 1642 https://bugs.webkit.org/show_bug.cgi?id=38898 1643 1644 * yarr/RegexJIT.cpp: 1645 (JSC::Yarr::RegexGenerator::generatePatternCharacterGreedy): 1646 (JSC::Yarr::RegexGenerator::generateCharacterClassGreedy): 1647 16482010-05-11 Geoffrey Garen <ggaren@apple.com> 1649 1650 Reviewed by Gavin Barraclough. 1651 1652 Slight simplification to the calling convention: read RegisterFile::end() 1653 from an absolute address, instead of from a pointer to the RegisterFile 1654 on the stack. 1655 1656 SunSpider reports no change. 1657 1658 This should free us to remove the RegisterFile argument from the stack. 1659 1660 * jit/JIT.cpp: 1661 (JSC::JIT::privateCompile): 1662 16632010-05-11 Brian Weinstein <bweinstein@apple.com> 1664 1665 Possible Gtk build fix - use ALWAYS_INLINE on functions that were inlined. 1666 1667 * wtf/text/AtomicString.h: 1668 (WebCore::AtomicString::add): Use ALWAYS_INLINE. 1669 * wtf/text/WTFString.h: 1670 (WebCore::String::length): Ditto. 1671 (WebCore::String::isEmpty): Ditto. 1672 16732010-05-11 Anders Carlsson <andersca@apple.com> 1674 1675 Reviewed by Sam Weinig. 1676 1677 Add a compile time assert that sizeof(String) == sizeof(AtomicString). 1678 1679 * wtf/text/AtomicString.cpp: 1680 16812010-05-11 Anders Carlsson <andersca@apple.com> 1682 1683 Reviewed by Mark Rowe. 1684 1685 https://bugs.webkit.org/show_bug.cgi?id=38961 1686 Move some member variables around to reduce class sizes. 1687 1688 Make AtomicString no longer inherit from FastAllocBase. 1689 1690 Since AtomicString's first (and only) member variable, a RefPtr, also inherits from FastAllocBase this 1691 was causing the size of AtomicString to contain sizeof(void*) bytes of padding. 1692 1693 * wtf/text/AtomicString.h: 1694 16952010-05-11 Brian Weinstein <bweinstein@apple.com> 1696 1697 Reviewed by Maciej Stachowiak. 1698 Patch by Gavin Barraclough. 1699 1700 Rest of REGRESSION (r57900-57919): 3% PLT Regression from moving strings into WTF. 1701 https://bugs.webkit.org/show_bug.cgi?id=38930 1702 <rdar://problem/7937188> 1703 1704 Inline string functions that are commonly used in WTFString, CString, and AtomicString. 1705 1706 * JavaScriptCore.exp: Remove exports of inline functions. 1707 * wtf/text/AtomicString.cpp: 1708 (WebCore::AtomicString::addSlowCase): Rename add to addSlowCase, and inline the fast path 1709 of add. 1710 * wtf/text/AtomicString.h: 1711 (WebCore::AtomicString::add): Inline the fast path of add. 1712 * wtf/text/CString.cpp: Inline implementation of data and length in the header. 1713 * wtf/text/CString.h: 1714 (WTF::CString::data): Inline implementation of data. 1715 (WTF::CString::length): Inline implementation of length. 1716 * wtf/text/WTFString.cpp: Remove implementations of functions to be inlined. 1717 * wtf/text/WTFString.h: 1718 (WebCore::String::String): Inline implementation. 1719 (WebCore::String::length): Ditto. 1720 (WebCore::String::characters): Ditto. 1721 (WebCore::String::operator[]): Ditto. 1722 (WebCore::String::isEmpty): Ditto. 1723 17242010-05-11 Brian Weinstein <bweinstein@apple.com> 1725 1726 Reviewed by Steve Falkenburg and Jon Honeycutt. 1727 Patch by Gavin Barraclough. 1728 1729 REGRESSION (r57900-57919): 3% PLT Regression from moving strings into WTF. 1730 https://bugs.webkit.org/show_bug.cgi?id=38930 1731 <rdar://problem/7937188> 1732 1733 When the String classes were moved from WebCore to WTF, it meant that on Windows, all operations 1734 on Strings in WebCore had to cross a DLL boundary (from WebKit.dll to JavaScript.dll). 1735 1736 We fix this by refactoring some of the WTF string code, so the code in AtomicString, StringImpl, and 1737 WTFString can be built by both WebCore and WTF, and we don't need to talk across a DLL to do operations 1738 on Strings. 1739 1740 * GNUmakefile.am: Add new file to build system. 1741 * JavaScriptCore.gypi: Ditto. 1742 * JavaScriptCore.pro: Ditto. 1743 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: Remove string exports, because these are now 1744 handled in WebCore. 1745 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: Remove the post-build step that was added 1746 here (the post build step is in JavaScriptCoreCommon.vsprops). 1747 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCoreCommon.vsprops: Copy the three cpp files that need to be compiled 1748 by WebCore into the WebKitOutputDir directory. 1749 * JavaScriptCore.vcproj/WTF/WTF.vcproj: Add the StringStatics file. 1750 * JavaScriptCore.vcproj/jsc/jsc.vcproj: Add the three WTF string cpp files to this project. 1751 * JavaScriptCore.vcproj/jsc/jscCommon.vsprops: Remove the need to link against WTF.lib (since jsc links against JavaScriptCore). 1752 * JavaScriptCore.xcodeproj/project.pbxproj: Add the StringStatics file. 1753 * wtf/text/AtomicString.cpp: Moved code to StringStatics. 1754 * wtf/text/StringImpl.cpp: Ditto. 1755 * wtf/text/StringStatics.cpp: Added. Move functions in WTF Strings that define static variables to here, so 1756 the rest of the files can be compiled in WebCore. 1757 (WebCore::StringImpl::empty): Moved from StringImpl.cpp to here. 1758 (WebCore::AtomicString::init): Moved from AtomicString.cpp to here. 1759 17602010-05-11 Alice Liu <alice.liu@apple.com> 1761 1762 Rubber-stamped by Gavin Barraclough. 1763 1764 Fix build error when enabling debugging block in WebKit win painting code 1765 1766 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 1767 17682010-05-11 Mark Rowe <mrowe@apple.com> 1769 1770 Fix the world. 1771 1772 In r59162 a change was made to WebCore's FeatureDefines.xcconfig that enabled FILE_READER and FILE_WRITER. 1773 The author and reviewer of that patch ignored the carefully-worded warning at the top of that file asking 1774 that changes to the file be kept in sync across JavaScriptCore, WebCore and WebKit, as well as being kept 1775 in sync with build-webkit. This led to WebCore and WebKit having different views of Document's vtable 1776 and results in crashes in Safari shortly after launch when virtual function calls resulted in the wrong 1777 function in WebCore being called. 1778 1779 We fix this by bringing the FeatureDefines.xcconfig files in to sync. Based on the ChangeLog message and 1780 other changes in r59162 it appears that enabling FILE_WRITER was unintentional so that particular change 1781 has been reverted. 1782 1783 * Configurations/FeatureDefines.xcconfig: 1784 17852010-05-11 Sheriff Bot <webkit.review.bot@gmail.com> 1786 1787 Unreviewed, rolling out r59171. 1788 http://trac.webkit.org/changeset/59171 1789 https://bugs.webkit.org/show_bug.cgi?id=38933 1790 1791 "Broke the world" (Requested by bweinstein on #webkit). 1792 1793 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 1794 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: 1795 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCoreCommon.vsprops: 1796 * JavaScriptCore.vcproj/WTF/WTF.vcproj: 1797 * JavaScriptCore.vcproj/jsc/jsc.vcproj: 1798 * JavaScriptCore.vcproj/jsc/jscCommon.vsprops: 1799 * JavaScriptCore.xcodeproj/project.pbxproj: 1800 * wtf/text/AtomicString.cpp: 1801 (WebCore::AtomicString::init): 1802 * wtf/text/StringImpl.cpp: 1803 (WebCore::StringImpl::empty): 1804 * wtf/text/StringStatics.cpp: Removed. 1805 18062010-05-11 Brian Weinstein <bweinstein@apple.com> 1807 1808 Reviewed by Steve Falkenburg. 1809 Patch by Gavin Barraclough. 1810 1811 REGRESSION (r57900-57919): 3% PLT Regression from moving strings into WTF. 1812 https://bugs.webkit.org/show_bug.cgi?id=38930 1813 <rdar://problem/7937188> 1814 1815 When the String classes were moved from WebCore to WTF, it meant that on Windows, all operations 1816 on Strings in WebCore had to cross a DLL boundary (from WebKit.dll to JavaScript.dll). 1817 1818 We fix this by refactoring some of the WTF string code, so the code in AtomicString, StringImpl, and 1819 WTFString can be built by both WebCore and WTF, and we don't need to talk across a DLL to do operations 1820 on Strings. 1821 1822 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: Remove string exports, because these are now 1823 handled in WebCore. 1824 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: Remove the post-build step that was added 1825 here (the post build step is in JavaScriptCoreCommon.vsprops). 1826 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCoreCommon.vsprops: Copy the three cpp files that need to be compiled 1827 by WebCore into the WebKitOutputDir directory. 1828 * JavaScriptCore.vcproj/WTF/WTF.vcproj: Add the StringStatics file. 1829 * JavaScriptCore.vcproj/jsc/jsc.vcproj: Add the three WTF string cpp files to this project. 1830 * JavaScriptCore.vcproj/jsc/jscCommon.vsprops: Remove the need to link against WTF.lib (since jsc links against JavaScriptCore). 1831 * JavaScriptCore.xcodeproj/project.pbxproj: Add the StringStatics file. 1832 * wtf/text/AtomicString.cpp: Moved code to StringStatics. 1833 * wtf/text/StringImpl.cpp: Ditto. 1834 * wtf/text/StringStatics.cpp: Added. Move functions in WTF Strings that define static variables to here, so 1835 the rest of the files can be compiled in WebCore. 1836 (WebCore::StringImpl::empty): Moved from StringImpl.cpp to here. 1837 (WebCore::AtomicString::init): Moved from AtomicString.cpp to here. 1838 18392010-05-11 Geoffrey Garen <ggaren@apple.com> 1840 1841 Fixed test failures seen on SnowLeopard buildbot. 1842 1843 * runtime/JSString.cpp: 1844 (JSC::JSString::replaceCharacter): Don't use size_t and wtf::notFound. 1845 Instead, use the individual types and notFound values of the string APIs 1846 we're using, since they're not necessarily the same in 64bit. 1847 18482010-05-11 Geoffrey Garen <ggaren@apple.com> 1849 1850 Reviewed by Oliver Hunt and Darin Adler. 1851 1852 Start using ropes in String.prototype.replace. 1853 1854 1%-1.5% speedup on SunSpider. 1855 1856 * runtime/JSString.cpp: 1857 (JSC::JSString::resolveRope): Updated for RopeImpl refactoring. 1858 1859 (JSC::JSString::replaceCharacter): Added a replaceCharacter function, which creates 1860 a rope for the resulting replacement. 1861 1862 * runtime/JSString.h: A few changes here: 1863 (JSC::): 1864 (JSC::RopeBuilder::RopeIterator::RopeIterator): 1865 (JSC::RopeBuilder::RopeIterator::operator++): 1866 (JSC::RopeBuilder::RopeIterator::operator*): 1867 (JSC::RopeBuilder::RopeIterator::operator!=): 1868 (JSC::RopeBuilder::RopeIterator::WorkItem::WorkItem): 1869 (JSC::RopeBuilder::RopeIterator::WorkItem::operator!=): 1870 (JSC::RopeBuilder::RopeIterator::skipRopes): Created a RopeIterator abstraction. 1871 We use this to do a substring find without having to resolve the rope. 1872 (We could use this iterator when resolving ropes, too, but resolving 1873 ropes backwards is usually more efficient.) 1874 1875 (JSC::RopeBuilder::JSString): Added constructors for 2 & 3 UStrings. 1876 1877 (JSC::RopeBuilder::appendValueInConstructAndIncrementLength): 1878 (JSC::RopeBuilder::size): Updated for RopeImpl refactoring. 1879 1880 * runtime/Operations.h: Updated for RopeImpl refactoring. 1881 (JSC::jsString): Added jsString functions for 2 & 3 UStrings. 1882 1883 * runtime/RopeImpl.cpp: 1884 (JSC::RopeImpl::derefFibersNonRecursive): 1885 * runtime/RopeImpl.h: 1886 (JSC::RopeImpl::initializeFiber): 1887 (JSC::RopeImpl::size): 1888 (JSC::RopeImpl::fibers): 1889 (JSC::RopeImpl::deref): 1890 (JSC::RopeImpl::RopeImpl): A little refactoring to make this patch easier: 1891 Moved statics to the top of the class; put multi-statement functions on 1892 multiple lines; renamed "fiberCount" to "size" to match other collections; 1893 changed the "fibers" accessor to return the fibers buffer, instead of an 1894 item in the buffer, to make iteration easier. 1895 1896 * runtime/StringPrototype.cpp: 1897 (JSC::stringProtoFuncReplace): Don't resolve a rope unless we need to. Do 1898 use our new replaceCharacter function if possible. Do use a rope to 1899 represent splicing three strings together. 1900 19012010-05-10 Laszlo Gombos <laszlo.1.gombos@nokia.com> 1902 1903 Reviewed by Darin Adler. 1904 1905 [Qt] Disable JIT support for mingw-w64 1906 https://bugs.webkit.org/show_bug.cgi?id=38747 1907 1908 Disale JIT for mingw-w64 as it is reportedly 1909 unstable. 1910 1911 Thanks for Vanboxem Rruben for the investigation. 1912 1913 * wtf/Platform.h: 1914 19152010-05-09 Fumitoshi Ukai <ukai@chromium.org> 1916 1917 Reviewed by Eric Seidel. 1918 1919 JavaScriptCore/wtf/MD5.h: checksum should take a reference to output. 1920 https://bugs.webkit.org/show_bug.cgi?id=38723 1921 1922 * JavaScriptCore.exp: 1923 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 1924 * wtf/MD5.cpp: 1925 (WTF::expectMD5): 1926 Fix for checksum change. 1927 (WTF::MD5::checksum): 1928 Take a reference to output, instead of returning the result by value, to reduce coping for performance. 1929 * wtf/MD5.h: 1930 19312010-05-09 Oliver Hunt <oliver@apple.com> 1932 1933 Build fix. 1934 1935 Fix accidental repeat addition of emit_op_new_regexp 1936 1937 * jit/JITOpcodes.cpp: 1938 19392010-05-09 Oliver Hunt <oliver@apple.com> 1940 1941 Reviewed by Gavin Barraclough. 1942 1943 REGRESSION(r57955): RegExp literals should not actually be cached, so r57955 should be rolled out. 1944 https://bugs.webkit.org/show_bug.cgi?id=38828 1945 <rdar://problem/7961634> 1946 1947 Rollout r57955 1948 1949 * bytecode/CodeBlock.cpp: 1950 (JSC::regexpToSourceString): 1951 (JSC::regexpName): 1952 (JSC::CodeBlock::dump): 1953 (JSC::CodeBlock::shrinkToFit): 1954 * bytecode/CodeBlock.h: 1955 (JSC::CodeBlock::addRegExp): 1956 (JSC::CodeBlock::regexp): 1957 * bytecode/Opcode.h: 1958 * bytecompiler/BytecodeGenerator.cpp: 1959 (JSC::BytecodeGenerator::addRegExp): 1960 (JSC::BytecodeGenerator::emitNewRegExp): 1961 * bytecompiler/BytecodeGenerator.h: 1962 * bytecompiler/NodesCodegen.cpp: 1963 (JSC::RegExpNode::emitBytecode): 1964 * interpreter/Interpreter.cpp: 1965 (JSC::Interpreter::privateExecute): 1966 * jit/JIT.cpp: 1967 (JSC::JIT::privateCompileMainPass): 1968 * jit/JIT.h: 1969 * jit/JITOpcodes.cpp: 1970 (JSC::JIT::emit_op_new_regexp): 1971 * jit/JITStubs.cpp: 1972 (JSC::DEFINE_STUB_FUNCTION): 1973 * jit/JITStubs.h: 1974 (JSC::): 1975 19762010-05-09 Geoffrey Garen <ggaren@apple.com> 1977 1978 Reviewed by Oliver Hunt. 1979 1980 Reserve a large-ish initial capacity for Lexer::m_buffer16. 1981 1982 SunSpider says 0.3% faster. 1983 1984 m_buffer16 is used when parsing complex strings -- for example, strings 1985 with escape sequences in them. These kinds of strings can be really long, 1986 and we want to avoid repeatedly copying as we grow m_buffer16. 1987 1988 The net memory cost is quite low, since it's proporitional to source 1989 code we already have in memory, and we throw away m_buffer16 right when 1990 we're done parsing. 1991 1992 * parser/Lexer.cpp: 1993 (JSC::Lexer::Lexer): No need to reserve initial capacity in our constructor, 1994 since setCode will be called before we're asked to lex anything. 1995 (JSC::Lexer::setCode): Reserve enough space to lex half the source code 1996 as a complex string without having to copy. 1997 (JSC::Lexer::clear): No need to reserve initial capacity here either, 1998 since setCode will be called before we're asked to lex anything. 1999 20002010-05-09 Laszlo Gombos <laszlo.1.gombos@nokia.com> 2001 2002 Reviewed by Eric Seidel. 2003 2004 [Qt] Remove YARR configuration rules from Qt make system 2005 https://bugs.webkit.org/show_bug.cgi?id=38819 2006 2007 Setting YARR based on JIT seeting is now in Platform.h 2008 for all ports. This patch essentially reverses r49238. 2009 2010 * JavaScriptCore.pri: 2011 20122010-05-09 Oliver Hunt <oliver@apple.com> 2013 2014 Reviewed by Maciej Stachowiak. 2015 2016 Improve string indexing performance 2017 https://bugs.webkit.org/show_bug.cgi?id=38814 2018 2019 Add an assembly stub to do indexed loads from strings much 2020 more cheaply than the current stub dispatch logic. We can 2021 do this because we are able to make guarantees about the 2022 register contents when entering the stub so the call overhead 2023 is negligible. 2024 2025 * jit/JIT.h: 2026 * jit/JITInlineMethods.h: 2027 * jit/JITOpcodes.cpp: 2028 * jit/JITPropertyAccess.cpp: 2029 (JSC::JIT::stringGetByValStubGenerator): 2030 (JSC::JIT::emitSlow_op_get_by_val): 2031 Moved from JITOpcodes.cpp to keep the slowcase next to 2032 the normal case codegen as we do for everything else. 2033 * jit/JITPropertyAccess32_64.cpp: 2034 (JSC::JIT::stringGetByValStubGenerator): 2035 (JSC::JIT::emitSlow_op_get_by_val): 2036 * jit/JSInterfaceJIT.h: 2037 (JSC::JSInterfaceJIT::emitFastArithImmToInt): 2038 20392010-05-09 Maciej Stachowiak <mjs@apple.com> 2040 2041 Fixed version of: "Optimized o[s] where o is a cell and s is a string" 2042 https://bugs.webkit.org/show_bug.cgi?id=38815 2043 2044 Fixed the previous patch for this from Geoff Garen. 2045 2046 The two problems were a missing exception check and a PropertySlot 2047 initialized improperly, leading to crashes and failures in the case 2048 of getters accessed with bracket syntax. 2049 2050 Previous patch: 2051 2052 Optimized o[s] where o is a cell and s is a string, removing some old 2053 code that wasn't really tuned for the JIT. 2054 2055 SunSpider says 0.8% faster. 2056 2057 * jit/JITStubs.cpp: 2058 (JSC::DEFINE_STUB_FUNCTION): 2059 * runtime/JSCell.h: 2060 20612010-05-08 Laszlo Gombos <laszlo.1.gombos@nokia.com> 2062 2063 Unreviewed, build fix. 2064 2065 Fix missing terminating ' character warning 2066 2067 The warning has been only seen on the Gtk buildbots. 2068 2069 * wtf/Platform.h: 2070 20712010-05-08 Laszlo Gombos <laszlo.1.gombos@nokia.com> 2072 2073 Reviewed by Gavin Barraclough. 2074 2075 Refactor YARR and YARR_JIT default rules in Platform.h 2076 https://bugs.webkit.org/show_bug.cgi?id=38727 2077 2078 Turn on YARR and YARR JIT by default of JIT is enabled. 2079 Ports can overrule this default decisions as they wish. 2080 2081 * wtf/Platform.h: 2082 20832010-05-08 Oliver Hunt <oliver@apple.com> 2084 2085 Reviewed by Gavin Barraclough. 2086 2087 Split JSVALUE32_64 code out of JITOpcodes.cpp and into JITOpcodes32_64.cpp 2088 https://bugs.webkit.org/show_bug.cgi?id=38808 2089 2090 * GNUmakefile.am: 2091 * JavaScriptCore.gypi: 2092 * JavaScriptCore.pro: 2093 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: 2094 * JavaScriptCore.xcodeproj/project.pbxproj: 2095 * jit/JITOpcodes.cpp: 2096 * jit/JITOpcodes32_64.cpp: Added. 2097 (JSC::JIT::privateCompileCTIMachineTrampolines): 2098 (JSC::JIT::emit_op_mov): 2099 (JSC::JIT::emit_op_end): 2100 (JSC::JIT::emit_op_jmp): 2101 (JSC::JIT::emit_op_loop_if_lesseq): 2102 (JSC::JIT::emitSlow_op_loop_if_lesseq): 2103 (JSC::JIT::emit_op_new_object): 2104 (JSC::JIT::emit_op_instanceof): 2105 (JSC::JIT::emitSlow_op_instanceof): 2106 (JSC::JIT::emit_op_new_func): 2107 (JSC::JIT::emit_op_get_global_var): 2108 (JSC::JIT::emit_op_put_global_var): 2109 (JSC::JIT::emit_op_get_scoped_var): 2110 (JSC::JIT::emit_op_put_scoped_var): 2111 (JSC::JIT::emit_op_tear_off_activation): 2112 (JSC::JIT::emit_op_tear_off_arguments): 2113 (JSC::JIT::emit_op_new_array): 2114 (JSC::JIT::emit_op_resolve): 2115 (JSC::JIT::emit_op_to_primitive): 2116 (JSC::JIT::emitSlow_op_to_primitive): 2117 (JSC::JIT::emit_op_strcat): 2118 (JSC::JIT::emit_op_resolve_base): 2119 (JSC::JIT::emit_op_resolve_skip): 2120 (JSC::JIT::emit_op_resolve_global): 2121 (JSC::JIT::emitSlow_op_resolve_global): 2122 (JSC::JIT::emit_op_not): 2123 (JSC::JIT::emitSlow_op_not): 2124 (JSC::JIT::emit_op_jfalse): 2125 (JSC::JIT::emitSlow_op_jfalse): 2126 (JSC::JIT::emit_op_jtrue): 2127 (JSC::JIT::emitSlow_op_jtrue): 2128 (JSC::JIT::emit_op_jeq_null): 2129 (JSC::JIT::emit_op_jneq_null): 2130 (JSC::JIT::emit_op_jneq_ptr): 2131 (JSC::JIT::emit_op_jsr): 2132 (JSC::JIT::emit_op_sret): 2133 (JSC::JIT::emit_op_eq): 2134 (JSC::JIT::emitSlow_op_eq): 2135 (JSC::JIT::emit_op_neq): 2136 (JSC::JIT::emitSlow_op_neq): 2137 (JSC::JIT::compileOpStrictEq): 2138 (JSC::JIT::emit_op_stricteq): 2139 (JSC::JIT::emitSlow_op_stricteq): 2140 (JSC::JIT::emit_op_nstricteq): 2141 (JSC::JIT::emitSlow_op_nstricteq): 2142 (JSC::JIT::emit_op_eq_null): 2143 (JSC::JIT::emit_op_neq_null): 2144 (JSC::JIT::emit_op_resolve_with_base): 2145 (JSC::JIT::emit_op_new_func_exp): 2146 (JSC::JIT::emit_op_throw): 2147 (JSC::JIT::emit_op_get_pnames): 2148 (JSC::JIT::emit_op_next_pname): 2149 (JSC::JIT::emit_op_push_scope): 2150 (JSC::JIT::emit_op_pop_scope): 2151 (JSC::JIT::emit_op_to_jsnumber): 2152 (JSC::JIT::emitSlow_op_to_jsnumber): 2153 (JSC::JIT::emit_op_push_new_scope): 2154 (JSC::JIT::emit_op_catch): 2155 (JSC::JIT::emit_op_jmp_scopes): 2156 (JSC::JIT::emit_op_switch_imm): 2157 (JSC::JIT::emit_op_switch_char): 2158 (JSC::JIT::emit_op_switch_string): 2159 (JSC::JIT::emit_op_new_error): 2160 (JSC::JIT::emit_op_debug): 2161 (JSC::JIT::emit_op_enter): 2162 (JSC::JIT::emit_op_enter_with_activation): 2163 (JSC::JIT::emit_op_create_arguments): 2164 (JSC::JIT::emit_op_init_arguments): 2165 (JSC::JIT::emit_op_convert_this): 2166 (JSC::JIT::emitSlow_op_convert_this): 2167 (JSC::JIT::emit_op_profile_will_call): 2168 (JSC::JIT::emit_op_profile_did_call): 2169 21702010-05-08 Gabor Loki <loki@webkit.org> 2171 2172 Reviewed by Gavin Barraclough. 2173 2174 Fix halfword loads on ARM 2175 https://bugs.webkit.org/show_bug.cgi?id=38741 2176 2177 The BaseIndex and ImplicitAddress are contain 32bit wide offset, but 2178 the load16 functions were working with 8 bit data (encoded in the 2179 instruction). If the offset cannot be encoded in an instruction, it 2180 should be stored in a temporary register. 2181 2182 * assembler/ARMAssembler.h: 2183 (JSC::ARMAssembler::getOffsetForHalfwordDataTransfer): 2184 * assembler/MacroAssemblerARM.h: 2185 (JSC::MacroAssemblerARM::load16): 2186 21872010-05-08 Gabor Loki <loki@webkit.org> 2188 2189 Reviewed by Gavin Barraclough. 2190 2191 Fix spanning branch instruction on Cortex-A8 with Thumb-2 JIT 2192 https://bugs.webkit.org/show_bug.cgi?id=38280 2193 2194 If the 32-bit Thumb-2 branch instruction spans two 4KiB regions and 2195 the target of the branch falls within the first region it is 2196 possible for the processor to incorrectly determine the branch 2197 instruction, and it is also possible in some cases for the processor 2198 to enter a deadlock state. 2199 2200 * assembler/ARMv7Assembler.h: 2201 (JSC::ARMv7Assembler::linkJumpAbsolute): 2202 22032010-05-08 Adam Barth <abarth@webkit.org> 2204 2205 Reviewed by Darin Fisher. 2206 2207 [WTFURL] Add a buffer for managing the output of the canonicalizer 2208 https://bugs.webkit.org/show_bug.cgi?id=38671 2209 2210 The canonicalizer produces a stream of canonicalized characters, which 2211 we will store in a URLBuffer. URLBuffer is quite similar to Vector, 2212 but serves as an abstraction to isolate us from the underlying 2213 character storage. In the case of WebKit, we'll probably implement a 2214 subclass of URLBuffer with Vector. In Chromium, the backing store is a 2215 std::string, which avoids an unnecessary memcpy that would be caused by 2216 using Vector instead. 2217 2218 * wtf/url/src/URLBuffer.h: Added. 2219 (WTF::URLBuffer::URLBuffer): 2220 (WTF::URLBuffer::~URLBuffer): 2221 (WTF::URLBuffer::at): 2222 (WTF::URLBuffer::set): 2223 (WTF::URLBuffer::capacity): 2224 (WTF::URLBuffer::length): 2225 (WTF::URLBuffer::data): 2226 (WTF::URLBuffer::setLength): 2227 (WTF::URLBuffer::append): 2228 (WTF::URLBuffer::grow): 2229 22302010-05-08 Eric Seidel <eric@webkit.org> 2231 2232 Unreviewed, just reverting commit. 2233 2234 REGRESSION(59000): r59000 contained all sorts of changes it should not have, needs revert. 2235 https://bugs.webkit.org/show_bug.cgi?id=38798 2236 2237 * bytecompiler/BytecodeGenerator.cpp: 2238 (JSC::BytecodeGenerator::emitResolve): 2239 * jit/JITOpcodes.cpp: 2240 (JSC::JIT::emit_op_resolve_global): 2241 22422010-05-04 Jer Noble <jer.noble@apple.com> 2243 2244 Reviewed by Maciej Stachowiak 2245 2246 Safari pegs CPU and drops tons of frames using HTML5 Vimeo player 2247 <https://bugs.webkit.org/show_bug.cgi?id=34005> 2248 <rdar://problem/7569713> 2249 2250 Added cancelCallOnMainThread. callOnMainThread should always now be paired 2251 with cancelCallOnMainThread in situations where the refcon passed to callOnMainThread 2252 may be dealloced before the main thread function can be dispatched. 2253 2254 * wtf/MainThread.cpp: 2255 (WTF::FunctionWithContext::operator == ): Supports the FunctionWithContextFinder predicate functor. 2256 (WTF::FunctionWithContextFinder::FunctionWithContextFinder): Predicate functor for use with Dequeue::findIf 2257 (WTF::FunctionWithContextFinder::operator()): 2258 (WTF::cancelCallOnMainThread): 2259 * wtf/MainThread.h: 2260 22612010-05-07 Oliver Hunt <oliver@apple.com> 2262 2263 Windows build fix. 2264 2265 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 2266 22672010-05-07 Eric Seidel <eric@webkit.org> 2268 2269 Unreviewed, rolling out r58990. 2270 http://trac.webkit.org/changeset/58990 2271 2272 Broke 4 tests on Snow Leopard and Qt. 2273 2274 * jit/JITStubs.cpp: 2275 (JSC::DEFINE_STUB_FUNCTION): 2276 * runtime/JSCell.h: 2277 22782010-05-07 Oliver Hunt <oliver@apple.com> 2279 2280 Reviewed by Gavin Barraclough. 2281 2282 Optimize access to the global object from a function that uses eval 2283 https://bugs.webkit.org/show_bug.cgi?id=38644 2284 2285 Fix bug where cross scope access to a global var (vs. property) would 2286 be allowed without checking for intervening dynamic scopes. 2287 2288 * bytecompiler/BytecodeGenerator.cpp: 2289 (JSC::BytecodeGenerator::emitResolve): 2290 22912010-05-07 Oliver Hunt <oliver@apple.com> 2292 2293 32-bit buildfix. 2294 2295 Macro expansion I stab at thee! 2296 2297 * jit/JITOpcodes.cpp: 2298 (JSC::JIT::emit_op_resolve_global): 2299 23002010-05-07 Geoffrey Garen <ggaren@apple.com> 2301 2302 Reviewed by Darin Adler. 2303 2304 Optimized o[s] where o is a cell and s is a string, removing some old 2305 code that wasn't really tuned for the JIT. 2306 2307 SunSpider says 0.8% faster. 2308 2309 * jit/JITStubs.cpp: 2310 (JSC::DEFINE_STUB_FUNCTION): 2311 * runtime/JSCell.h: 2312 23132010-05-07 Oliver Hunt <oliver@apple.com> 2314 2315 Windows build fix. 2316 2317 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 2318 23192010-05-07 Oliver Hunt <oliver@apple.com> 2320 2321 Reviewed by Geoffrey Garen. 2322 2323 Optimize access to the global object from a function that uses eval 2324 https://bugs.webkit.org/show_bug.cgi?id=38644 2325 2326 Add op_resolve_global_dynamic, a variant of op_resolve_global that 2327 checks each node in the scope chain for dynamically inserted properties 2328 and falls back to the normal resolve logic in that case. 2329 2330 * JavaScriptCore.exp: 2331 * bytecode/CodeBlock.cpp: 2332 (JSC::isGlobalResolve): 2333 (JSC::CodeBlock::printStructures): 2334 (JSC::CodeBlock::dump): 2335 (JSC::CodeBlock::derefStructures): 2336 * bytecode/Opcode.h: 2337 * bytecompiler/BytecodeGenerator.cpp: 2338 (JSC::BytecodeGenerator::findScopedProperty): 2339 Now take an additional reference parameter to used to indicate that 2340 there were nodes that may gain dynamic properties 2341 (JSC::BytecodeGenerator::emitResolve): 2342 (JSC::BytecodeGenerator::emitResolveBase): 2343 (JSC::BytecodeGenerator::emitResolveWithBase): 2344 deal with additional argument to findScopedProperty 2345 * bytecompiler/BytecodeGenerator.h: 2346 * bytecompiler/NodesCodegen.cpp: 2347 (JSC::FunctionCallResolveNode::emitBytecode): 2348 (JSC::PostfixResolveNode::emitBytecode): 2349 (JSC::PrefixResolveNode::emitBytecode): 2350 (JSC::ReadModifyResolveNode::emitBytecode): 2351 (JSC::AssignResolveNode::emitBytecode): 2352 These functions use findScopedProperty directly in order to 2353 optimise lookup. They cannot trivially handle any degree of 2354 dynamism in the lookup so we just give up in such case. 2355 * interpreter/Interpreter.cpp: 2356 (JSC::Interpreter::resolveGlobalDynamic): 2357 (JSC::Interpreter::execute): 2358 (JSC::Interpreter::privateExecute): 2359 * interpreter/Interpreter.h: 2360 * jit/JIT.cpp: 2361 (JSC::JIT::privateCompileMainPass): 2362 (JSC::JIT::privateCompileSlowCases): 2363 * jit/JIT.h: 2364 * jit/JITOpcodes.cpp: 2365 (JSC::JIT::emit_op_resolve_global): 2366 (JSC::JIT::emit_op_resolve_global_dynamic): 2367 (JSC::JIT::emitSlow_op_resolve_global): 2368 (JSC::JIT::emitSlow_op_resolve_global_dynamic): 2369 Happily resolve_global_dynamic can share the slow case! 2370 * jit/JITStubs.h: 2371 (JSC::): 2372 * runtime/JSActivation.cpp: 2373 (JSC::JSActivation::isDynamicScope): 2374 * runtime/JSActivation.h: 2375 * runtime/JSGlobalObject.cpp: 2376 (JSC::JSGlobalObject::isDynamicScope): 2377 * runtime/JSGlobalObject.h: 2378 * runtime/JSStaticScopeObject.cpp: 2379 (JSC::JSStaticScopeObject::isDynamicScope): 2380 * runtime/JSStaticScopeObject.h: 2381 * runtime/JSVariableObject.h: 2382 23832010-05-07 Geoffrey Garen <ggaren@apple.com> 2384 2385 Reviewed by Gavin Barraclough. 2386 2387 Fixed Sputnik failure seen on buildbot. 2388 2389 * runtime/NumberPrototype.cpp: 2390 (JSC::numberProtoFuncToString): Cast to unsigned before checking magnitude, 2391 to efficiently exclude negative numbers, which require two characters 2392 instead of one. 2393 23942010-05-07 Geoffrey Garen <ggaren@apple.com> 2395 2396 Reviewed by Alexey Proskuryakov. 2397 2398 Slightly more effective way to guarantee a compile-time constant, at 2399 least on gcc. 2400 2401 * runtime/NumberPrototype.cpp: 2402 (JSC::numberProtoFuncToString): Use pointer notation instead of array 2403 notation, so the compiler treats the string and its contents as constant. 2404 24052010-05-07 Adam Barth <abarth@webkit.org> 2406 2407 Reviewed by Alexey Proskuryakov. 2408 2409 [WTFURL] Add core URL parser 2410 https://bugs.webkit.org/show_bug.cgi?id=38572 2411 2412 This patch adds the core of the URL parser. The URL parser uses a 2413 templated notion of a code unit to support different string types. 2414 Later we'll add some non-templated APIs with concrete types that 2415 clients can use more conveniently. 2416 2417 The URLParser has a couple calls to ASSERT(), which I've commented out 2418 for now. I'm not 100% sure how to handle the dependency issues there 2419 yet, so I've punted on the issue for now. 2420 2421 Also, there are a number of methods that are declared public in this 2422 patch that are only used by the canonicalizer. My plan is to make 2423 those private or protected and make the canonicalizer a friend of the 2424 parser. The details will be in a later patch. 2425 2426 * wtf/url/src/URLComponent.h: 2427 (WTF::URLComponent::fromRange): 2428 (WTF::URLComponent::isNonEmpty): 2429 (WTF::URLComponent::isEmptyOrInvalid): 2430 * wtf/url/src/URLParser.h: Added. 2431 (WTF::URLParser::): 2432 (WTF::URLParser::isPossibleAuthorityTerminator): 2433 (WTF::URLParser::parseAuthority): 2434 (WTF::URLParser::extractScheme): 2435 (WTF::URLParser::parseAfterScheme): 2436 (WTF::URLParser::parseStandardURL): 2437 (WTF::URLParser::parsePath): 2438 (WTF::URLParser::parsePathURL): 2439 (WTF::URLParser::parseMailtoURL): 2440 (WTF::URLParser::parsePort): 2441 (WTF::URLParser::extractFileName): 2442 (WTF::URLParser::extractQueryKeyValue): 2443 (WTF::URLParser::isURLSlash): 2444 (WTF::URLParser::shouldTrimFromURL): 2445 (WTF::URLParser::trimURL): 2446 (WTF::URLParser::consecutiveSlashes): 2447 (WTF::URLParser::isPortDigit): 2448 (WTF::URLParser::nextAuthorityTerminator): 2449 (WTF::URLParser::parseUserInfo): 2450 (WTF::URLParser::parseServerInfo): 2451 24522010-05-07 Geoffrey Garen <ggaren@apple.com> 2453 2454 Reviewed by Oliver Hunt. 2455 2456 Added a fast path for number-to-character conversion via 2457 Number.prototype.toString base 36. 2458 2459 0.7% speedup on SunSpider. 2460 2461 * runtime/NumberPrototype.cpp: 2462 (JSC::numberProtoFuncToString): Made radix interpretation a little more 2463 efficient by keeping it in int space. Turned "const char" into 2464 "static const char" just in case. Added a fast path for base 36 2465 conversion of something that will turn into a character. 2466 24672010-05-07 Adam Barth <abarth@webkit.org> 2468 2469 Reviewed by Darin Fisher. 2470 2471 [WTFURL] Add a malloc-backed URLBuffer 2472 https://bugs.webkit.org/show_bug.cgi?id=38694 2473 2474 This patch adds a URLBuffer that uses new/delete to manage the backing 2475 store for the buffer. 2476 2477 * wtf/url/src/RawURLBuffer.h: Added. 2478 (WTF::RawURLBuffer::RawURLBuffer): 2479 (WTF::RawURLBuffer::~RawURLBuffer): 2480 (WTF::RawURLBuffer::resize): 2481 24822010-05-06 Fumitoshi Ukai <ukai@chromium.org> 2483 2484 Reviewed by Alexey Proskuryakov. 2485 2486 JavaScriptCore/wtf/RandomNumber.h should provide using WTF::* 2487 https://bugs.webkit.org/show_bug.cgi?id=38719 2488 2489 * wtf/RandomNumber.h: 2490 Add using directives. 2491 24922010-05-06 Laszlo Gombos <laszlo.1.gombos@nokia.com> 2493 2494 Reviewed by Kenneth Rohde Christiansen. 2495 2496 [Qt] Enable YARR_JIT for X86 Mac for QtWebKit 2497 https://bugs.webkit.org/show_bug.cgi?id=38668 2498 2499 * wtf/Platform.h: 2500 25012010-05-06 Gavin Barraclough <barraclough@apple.com> 2502 2503 Reviewed by Geoff Garen. 2504 2505 Bug 38714 - Add result caching for Math.sin 2506 2507 Math.sin is frequently called with the same operand, caching the result should 2508 improve performance. CachedTranscendentalFunction adds a generic mechanism to 2509 cache results for pure functions with the signature "double func(double)", and 2510 where NaN maps to NaN. 2511 2512 CachedTranscendentalFunction uses a very simple hash function designed to be 2513 fast to execute since we cannot cache the result, and trivial to JIT generate, 2514 should we wish to do so. 2515 2516 * JavaScriptCore.xcodeproj/project.pbxproj: 2517 * runtime/CachedTranscendentalFunction.h: Added. 2518 (JSC::CachedTranscendentalFunction::CachedTranscendentalFunction): 2519 (JSC::CachedTranscendentalFunction::~CachedTranscendentalFunction): 2520 (JSC::CachedTranscendentalFunction::operator()): 2521 (JSC::CachedTranscendentalFunction::initialize): 2522 (JSC::CachedTranscendentalFunction::hash): 2523 * runtime/JSGlobalData.h: 2524 * runtime/MathObject.cpp: 2525 (JSC::mathProtoFuncSin): 2526 * wtf/Platform.h: 2527 25282010-05-06 Geoffrey Garen <ggaren@apple.com> 2529 2530 Windows build fix. 2531 2532 * bytecode/SamplingTool.cpp: 2533 (JSC::SamplingFlags::sample): Use a cast, so Windows will be happy when 2534 this code is enabled. 2535 * wtf/Platform.h: Reverted last change to this file, which seems to have 2536 been accidental. 2537 25382010-05-06 Gavin Barraclough <barraclough@apple.com> 2539 2540 Reviewed by Oliver Hunt. 2541 2542 Add compile switch to make debugger keyword plant breakpoint instructions. 2543 2544 * bytecompiler/BytecodeGenerator.cpp: 2545 (JSC::BytecodeGenerator::emitDebugHook): 2546 * jit/JITOpcodes.cpp: 2547 (JSC::JIT::emit_op_debug): 2548 * wtf/Platform.h: 2549 25502010-05-06 Oliver Hunt <oliver@apple.com> 2551 2552 Build fix 2553 2554 Fix ARM builds 2555 2556 * assembler/MacroAssemblerARM.h: 2557 (JSC::MacroAssemblerARM::load16): 2558 * assembler/MacroAssemblerARMv7.h: 2559 (JSC::MacroAssemblerARMv7::load16): 2560 25612010-05-06 Oliver Hunt <oliver@apple.com> 2562 2563 Reviewed by Geoffrey Garen. 2564 2565 Improve performance of single character string compares 2566 https://bugs.webkit.org/show_bug.cgi?id=38659 2567 2568 Add logic to the jit to identify comparisons to single character string literals 2569 and then just perform the comparison inline, rather than ignoring the evidence 2570 and attempting to perform an integer comparison. 2571 2572 Multiple changes required -- add jnlesseq opcode, add helper function to identify 2573 single character string constants, add a helper to load single character strings. 2574 Then add the 32_64 and normal codepaths to the JIT. 2575 2576 * assembler/MacroAssemblerX86Common.h: 2577 (JSC::MacroAssemblerX86Common::load16): 2578 * bytecode/CodeBlock.cpp: 2579 (JSC::CodeBlock::dump): 2580 * bytecode/Opcode.h: 2581 * bytecompiler/BytecodeGenerator.cpp: 2582 (JSC::BytecodeGenerator::emitJumpIfTrue): 2583 * interpreter/Interpreter.cpp: 2584 (JSC::Interpreter::privateExecute): 2585 * jit/JIT.cpp: 2586 (JSC::JIT::privateCompileMainPass): 2587 (JSC::JIT::privateCompileSlowCases): 2588 * jit/JIT.h: 2589 * jit/JITArithmetic.cpp: 2590 (JSC::JIT::emit_op_jnless): 2591 (JSC::JIT::emitSlow_op_jnless): 2592 (JSC::JIT::emit_op_jless): 2593 (JSC::JIT::emitSlow_op_jless): 2594 (JSC::JIT::emit_op_jlesseq): 2595 (JSC::JIT::emit_op_jnlesseq): 2596 (JSC::JIT::emitSlow_op_jlesseq): 2597 (JSC::JIT::emitSlow_op_jnlesseq): 2598 * jit/JITArithmetic32_64.cpp: 2599 (JSC::JIT::emit_op_jnless): 2600 (JSC::JIT::emitSlow_op_jnless): 2601 (JSC::JIT::emit_op_jless): 2602 (JSC::JIT::emitSlow_op_jless): 2603 (JSC::JIT::emit_op_jlesseq): 2604 (JSC::JIT::emit_op_jnlesseq): 2605 (JSC::JIT::emitSlow_op_jlesseq): 2606 (JSC::JIT::emitSlow_op_jnlesseq): 2607 (JSC::JIT::emitBinaryDoubleOp): 2608 * jit/JITInlineMethods.h: 2609 (JSC::JIT::emitLoadCharacterString): 2610 (JSC::JIT::isOperandConstantImmediateChar): 2611 * jit/JSInterfaceJIT.h: 2612 (JSC::ThunkHelpers::stringImplDataOffset): 2613 (JSC::ThunkHelpers::jsStringLengthOffset): 2614 (JSC::ThunkHelpers::jsStringValueOffset): 2615 Moved from ThunkGenerators to make it possible to share. 2616 * jit/ThunkGenerators.cpp: 2617 26182010-05-06 Martin Robinson <mrobinson@webkit.org> 2619 2620 Reviewed by Gustavo Noronha Silva. 2621 2622 [GTK] Enable DOM clipboard and drag-and-drop access 2623 https://bugs.webkit.org/show_bug.cgi?id=30623 2624 2625 Convert dragging portion of drag-and-drop to use DataObjectGtk. 2626 2627 * wtf/gobject/GRefPtr.h: Add forward declarations for GObject functions. 2628 26292010-05-06 Steve Block <steveblock@google.com> 2630 2631 Reviewed by Eric Seidel. 2632 2633 MAC_JAVA_BRIDGE should be renamed JAVA_BRIDGE 2634 https://bugs.webkit.org/show_bug.cgi?id=38544 2635 2636 * wtf/Platform.h: 2637 26382010-05-06 Adam Barth <abarth@webkit.org> 2639 2640 Reviewed by Maciej Stachowiak. 2641 2642 [WTFURL] Add a class to represent the segments of a URL 2643 https://bugs.webkit.org/show_bug.cgi?id=38566 2644 2645 When paired with a string, this class represents a parsed URL. 2646 2647 * wtf/url/src/URLSegments.cpp: Added. 2648 (WTF::URLSegments::length): 2649 (WTF::URLSegments::charactersBefore): 2650 * wtf/url/src/URLSegments.h: Added. 2651 (WTF::URLSegments::): 2652 (WTF::URLSegments::URLSegments): 2653 * wtf/url/wtfurl.gyp: 2654 26552010-05-05 Tony Gentilcore <tonyg@chromium.org> 2656 2657 Reviewed by Adam Barth. 2658 2659 Adding comment with link to RefPtr documentation. 2660 https://bugs.webkit.org/show_bug.cgi?id=38601 2661 2662 * wtf/RefPtr.h: 2663 26642010-05-05 Gavin Barraclough <barraclough@apple.com> 2665 2666 Reviewed by Darin Adler. 2667 2668 Bug 38604 - workers-gc2 crashing on Qt 2669 2670 This appears to be due to a couple of issues. 2671 (1) When the atomic string table is deleted it does not clear the 'inTable' bit 2672 on AtomicStrings - it implicitly assumes that all AtomicStrings have already 2673 been deleted at this point (otherwise they will crash in their destructor when 2674 they try to remove themselves from the atomic string table). 2675 (2) We don't fix the ordering in which WTF::WTFThreadData and 2676 WebCore::ThreadGlobalData are destructed. 2677 2678 We should make sure that ThreadGlobalData is cleaned up before worker threads 2679 terminate and WTF::WTFThreadData is destroyed, and we should clear the inTable 2680 bit of members on atomic string table destruction. 2681 2682 WTF changes (fix issue 1, above) - ensure inTable property is cleared when the 2683 atomic string table is destroyed (also, rename InTable to IsAtomic, to make it 2684 clear which table we are refering to!) 2685 2686 * wtf/text/AtomicString.cpp: 2687 (WebCore::AtomicStringTable::destroy): 2688 (WebCore::CStringTranslator::translate): 2689 (WebCore::UCharBufferTranslator::translate): 2690 (WebCore::HashAndCharactersTranslator::translate): 2691 (WebCore::AtomicString::add): 2692 * wtf/text/StringImpl.cpp: 2693 (WebCore::StringImpl::~StringImpl): 2694 * wtf/text/StringImpl.h: 2695 (WebCore::StringImpl::isAtomic): 2696 (WebCore::StringImpl::setIsAtomic): 2697 * wtf/text/StringImplBase.h: 2698 26992010-05-05 Oliver Hunt <oliver@apple.com> 2700 2701 Reviewed by Geoffrey Garen. 2702 2703 Make Array.join specialised for actual arrays 2704 https://bugs.webkit.org/show_bug.cgi?id=38592 2705 2706 Bring Array.join in line with the other array methods by specialising for 2707 this === an array. 2708 2709 * runtime/ArrayPrototype.cpp: 2710 (JSC::arrayProtoFuncJoin): 2711 27122010-05-04 Yongjun Zhang <yongjun_zhang@apple.com> 2713 2714 Reviewed by Darin Adler. 2715 2716 WebKit crashes at DebuggerCallFrame::functionName() if m_callFrame is the top global callframe. 2717 https://bugs.webkit.org/show_bug.cgi?id=38535 2718 2719 Don't call asFunction if callee is not a FunctionType to prevent assertion failure 2720 in JSCell::isObject(). 2721 2722 * debugger/DebuggerCallFrame.cpp: 2723 (JSC::DebuggerCallFrame::functionName): 2724 (JSC::DebuggerCallFrame::calculatedFunctionName): 2725 27262010-05-04 Steve Falkenburg <sfalken@apple.com> 2727 2728 Remove redundant exports to fix Windows build warnings. 2729 2730 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 2731 27322010-05-04 Jedrzej Nowacki <jedrzej.nowacki@nokia.com> 2733 2734 Reviewed by Holger Freyther. 2735 2736 Fix a memory leak inside the QScriptValue. 2737 2738 QSciptValuePrivate::toString should release all temporary variables. 2739 2740 [Qt] QScriptValue::toString have a memory leak. 2741 https://bugs.webkit.org/show_bug.cgi?id=38112 2742 2743 * qt/api/qscriptvalue_p.h: 2744 (QScriptValuePrivate::toString): 2745 27462010-05-04 Gustavo Noronha Silva <gustavo.noronha@collabora.co.uk> 2747 2748 Reviewed by Darin Adler. 2749 2750 Does not work on SPARC because of alignment issues 2751 https://bugs.webkit.org/show_bug.cgi?id=38455 2752 2753 Use AllocAlignmentInteger instead of void* on pageheap_memory to 2754 avoid crashing at runtime in some architectures. Investigated by 2755 Mike Hommey <mh@glandium.org>, change suggested by Darin Adler. 2756 2757 * wtf/FastMalloc.cpp: 2758 27592010-05-04 Adam Barth <abarth@webkit.org> 2760 2761 Reviewed by Maciej Stachowiak. 2762 2763 Add first WTFURL file 2764 https://bugs.webkit.org/show_bug.cgi?id=38470 2765 2766 This class represents a component of a URL, such as the scheme or path. 2767 Essentially, the class is an index into a string buffer. 2768 2769 * wtf/url: Added. 2770 - There was some discussion about where to put the URL library. 2771 There are a number of reasonable options, but putting the library 2772 in the WTF directory seems like a reasonable alternative. This 2773 location will make more sense once WTF is moved out of 2774 JavaScriptCore and into a top-level directory. 2775 * wtf/url/src: Added. 2776 - I've added this extra directory to the usual structure because we 2777 eventually want a "public" folder that defines the public 2778 interface to this library that clients outside of svn.webkit.org 2779 can rely upon, much like the various subfolders of the WebKit 2780 directory often define a public API. 2781 * wtf/url/src/URLComponent.h: Added. 2782 - I've added this somewhat trivial class to get the ball rolling. 2783 (WTF::URLComponent::URLComponent): 2784 (WTF::URLComponent::isValid): 2785 (WTF::URLComponent::isNonempty): 2786 (WTF::URLComponent::reset): 2787 (WTF::URLComponent::operator==): 2788 (WTF::URLComponent::begin): 2789 (WTF::URLComponent::setBegin): 2790 (WTF::URLComponent::length): 2791 (WTF::URLComponent::setLength): 2792 (WTF::URLComponent::end): 2793 * wtf/url/wtfurl.gyp: Added. 2794 - WTFURL will be an independently buildable library, hence the 2795 beginnings of a wtfurl.gyp file. 2796 27972010-05-03 Stephanie Lewis <slewis@apple.com> 2798 2799 Reviewed by Geoff Garen. 2800 2801 https://bugs.webkit.org/show_bug.cgi?id=38368 2802 <rdar://problem/7834433> REGRESSSION: 1.5% PLT regression due to 56028 2803 (return memory quicker). 2804 Instead of returning everything but the smallest spans spread out 2805 the spans kept over several size lists. 2806 2807 * wtf/FastMalloc.cpp: 2808 (WTF::TCMalloc_PageHeap::scavenge): 2809 28102010-05-03 James Robinson <jamesr@chromium.org> 2811 2812 Reviewed by Eric Seidel. 2813 2814 Clean up a few compiler warnings 2815 https://bugs.webkit.org/show_bug.cgi?id=38073 2816 2817 * wtf/text/StringImpl.cpp: 2818 (WebCore::StringImpl::ascii): 2819 28202010-05-03 Gavin Barraclough <barraclough@apple.com> 2821 2822 Reviewed by NOBODY (reverting previous commit). 2823 2824 Rolling out r58114 - this introduced memory leaks of 2825 AtomicStrings then workers terminated. 2826 2827 * GNUmakefile.am: 2828 * JavaScriptCore.gypi: 2829 * JavaScriptCore.vcproj/WTF/WTF.vcproj: 2830 * JavaScriptCore.xcodeproj/project.pbxproj: 2831 * jit/ThunkGenerators.cpp: 2832 (JSC::ThunkHelpers::stringImplDataOffset): 2833 * runtime/Identifier.cpp: 2834 (JSC::IdentifierTable::~IdentifierTable): 2835 (JSC::IdentifierTable::add): 2836 (JSC::IdentifierCStringTranslator::hash): 2837 (JSC::IdentifierCStringTranslator::equal): 2838 (JSC::IdentifierCStringTranslator::translate): 2839 (JSC::Identifier::add): 2840 (JSC::IdentifierUCharBufferTranslator::hash): 2841 (JSC::IdentifierUCharBufferTranslator::equal): 2842 (JSC::IdentifierUCharBufferTranslator::translate): 2843 (JSC::Identifier::addSlowCase): 2844 * runtime/Identifier.h: 2845 * runtime/JSGlobalData.cpp: 2846 (JSC::JSGlobalData::JSGlobalData): 2847 (JSC::JSGlobalData::~JSGlobalData): 2848 * runtime/JSGlobalData.h: 2849 * wtf/WTFThreadData.cpp: 2850 (WTF::WTFThreadData::WTFThreadData): 2851 (WTF::WTFThreadData::~WTFThreadData): 2852 * wtf/WTFThreadData.h: 2853 (JSC::IdentifierTable::remove): 2854 (JSC::IdentifierTable::literalTable): 2855 (WTF::WTFThreadData::atomicStringTable): 2856 * wtf/text/AtomicString.cpp: 2857 (WebCore::AtomicStringTable::create): 2858 (WebCore::AtomicStringTable::table): 2859 (WebCore::AtomicStringTable::destroy): 2860 (WebCore::stringTable): 2861 (WebCore::CStringTranslator::hash): 2862 (WebCore::CStringTranslator::equal): 2863 (WebCore::CStringTranslator::translate): 2864 (WebCore::operator==): 2865 (WebCore::AtomicString::add): 2866 (WebCore::equal): 2867 (WebCore::UCharBufferTranslator::hash): 2868 (WebCore::UCharBufferTranslator::equal): 2869 (WebCore::UCharBufferTranslator::translate): 2870 (WebCore::HashAndCharactersTranslator::hash): 2871 (WebCore::HashAndCharactersTranslator::equal): 2872 (WebCore::HashAndCharactersTranslator::translate): 2873 (WebCore::AtomicString::find): 2874 (WebCore::AtomicString::remove): 2875 * wtf/text/AtomicStringTable.h: Removed. 2876 * wtf/text/StringImpl.cpp: 2877 (WebCore::StringImpl::~StringImpl): 2878 * wtf/text/StringImpl.h: 2879 (WebCore::StringImpl::inTable): 2880 (WebCore::StringImpl::setInTable): 2881 (WebCore::equal): 2882 * wtf/text/StringImplBase.h: 2883 (WTF::StringImplBase::StringImplBase): 2884 28852010-05-03 Kevin Watters <kevinwatters@gmail.com> 2886 2887 Reviewed by Kevin Ollivier. 2888 2889 [wx] Implement scheduleDispatchFunctionsOnMainThread for wx port. 2890 https://bugs.webkit.org/show_bug.cgi?id=38480 2891 2892 * wtf/wx/MainThreadWx.cpp: 2893 (wxCallAfter::wxCallAfter): 2894 (wxCallAfter::OnCallback): 2895 (WTF::scheduleDispatchFunctionsOnMainThread): 2896 28972010-05-03 Kevin Watters <kevinwatters@gmail.com> 2898 2899 Reviewed by Kevin Ollivier. 2900 2901 [wx] Build and use Mac's ComplexTextController to support complex text in wx. 2902 https://bugs.webkit.org/show_bug.cgi?id=38482 2903 2904 * wtf/Platform.h: 2905 29062010-05-03 Oliver Hunt <oliver@apple.com> 2907 2908 Reviewed by Maciej Stachowiak. 2909 2910 Interpreter crashes due to incorrect refcounting of cached structures. 2911 https://bugs.webkit.org/show_bug.cgi?id=38491 2912 rdar://problem/7926160 2913 2914 Make sure we ref/deref structures used for cached custom property getters 2915 2916 * bytecode/CodeBlock.cpp: 2917 (JSC::CodeBlock::derefStructures): 2918 (JSC::CodeBlock::refStructures): 2919 29202010-05-02 Laszlo Gombos <laszlo.1.gombos@nokia.com> 2921 2922 Reviewed by Eric Seidel. 2923 2924 [Qt] Enable JIT for QtWebKit on Symbian 2925 https://bugs.webkit.org/show_bug.cgi?id=38339 2926 2927 JIT on Symbian has been stable for quite some time, it 2928 is time to turn it on by default. 2929 2930 * wtf/Platform.h: 2931 29322010-04-29 Oliver Hunt <oliver@apple.com> 2933 2934 Reviewed by Gavin Barraclough. 2935 2936 Add codegen support for unsigned right shift 2937 https://bugs.webkit.org/show_bug.cgi?id=38375 2938 2939 Expose unsigned right shift in the macro assembler, and make use of it 2940 from the jit. Currently if the result is outside the range 0..2^31-1 2941 we simply fall back to the slow case, even in JSVALUE64 and JSVALUE32_64 2942 where technically we could still return an immediate value. 2943 2944 * assembler/MacroAssemblerARM.h: 2945 (JSC::MacroAssemblerARM::urshift32): 2946 * assembler/MacroAssemblerARMv7.h: 2947 (JSC::MacroAssemblerARMv7::urshift32): 2948 * assembler/MacroAssemblerX86Common.h: 2949 (JSC::MacroAssemblerX86Common::urshift32): 2950 * assembler/X86Assembler.h: 2951 (JSC::X86Assembler::): 2952 (JSC::X86Assembler::shrl_i8r): 2953 (JSC::X86Assembler::shrl_CLr): 2954 Add unsigned right shift to the x86 assembler 2955 * jit/JIT.cpp: 2956 (JSC::JIT::privateCompileMainPass): 2957 (JSC::JIT::privateCompileSlowCases): 2958 op_rshift no longer simply get thrown to a stub function 2959 * jit/JIT.h: 2960 * jit/JITArithmetic.cpp: 2961 (JSC::JIT::emit_op_urshift): 2962 (JSC::JIT::emitSlow_op_urshift): 2963 JSVALUE32 and JSVALUE64 implementation. Only supports 2964 double lhs in JSVALUE64. 2965 * jit/JITArithmetic32_64.cpp: 2966 (JSC::JIT::emit_op_rshift): 2967 (JSC::JIT::emitSlow_op_rshift): 2968 (JSC::JIT::emit_op_urshift): 2969 (JSC::JIT::emitSlow_op_urshift): 2970 Refactor right shift code to have shared implementation between signed 2971 and unsigned versions. 2972 29732010-04-29 Oliver Hunt <oliver@apple.com> 2974 2975 Reviewed by Gavin Barraclough. 2976 2977 Handle double on righthand side of a right shift 2978 https://bugs.webkit.org/show_bug.cgi?id=38363 2979 2980 In 64-bit we support right shift of a double, and there's no 2981 reason not to in 32-bit either. This is a 1.1% sunspider 2982 improvement in 32bit. 2983 2984 Given high performance javascript code frequently makes 2985 use of bit operators to perform double->int conversion i'll 2986 follow this patch up with similar patches for bitwise 'and' 2987 and 'or'. 2988 2989 * jit/JITArithmetic32_64.cpp: 2990 (JSC::JIT::emit_op_rshift): 2991 (JSC::JIT::emitSlow_op_rshift): 2992 29932010-04-29 Oliver Hunt <oliver@apple.com> 2994 2995 Reviewed by Geoff Garen. 2996 2997 Hoist JSVALUE32_64 arithmetic implementations into a separate file. 2998 2999 * GNUmakefile.am: 3000 * JavaScriptCore.gypi: 3001 * JavaScriptCore.pro: 3002 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: 3003 * JavaScriptCore.xcodeproj/project.pbxproj: 3004 * jit/JITArithmetic.cpp: 3005 (JSC::JIT::emitSlow_op_jnless): 3006 (JSC::JIT::emitSlow_op_jless): 3007 (JSC::JIT::emitSlow_op_jnlesseq): 3008 (JSC::JIT::emit_op_mod): 3009 (JSC::JIT::compileBinaryArithOpSlowCase): 3010 (JSC::JIT::emit_op_div): 3011 (JSC::JIT::compileBinaryArithOp): 3012 * jit/JITArithmetic32_64.cpp: Added. 3013 (JSC::JIT::emit_op_negate): 3014 (JSC::JIT::emitSlow_op_negate): 3015 (JSC::JIT::emit_op_jnless): 3016 (JSC::JIT::emitSlow_op_jnless): 3017 (JSC::JIT::emit_op_jless): 3018 (JSC::JIT::emitSlow_op_jless): 3019 (JSC::JIT::emit_op_jnlesseq): 3020 (JSC::JIT::emitSlow_op_jnlesseq): 3021 (JSC::JIT::emit_op_lshift): 3022 (JSC::JIT::emitSlow_op_lshift): 3023 (JSC::JIT::emit_op_rshift): 3024 (JSC::JIT::emitSlow_op_rshift): 3025 (JSC::JIT::emit_op_bitand): 3026 (JSC::JIT::emitSlow_op_bitand): 3027 (JSC::JIT::emit_op_bitor): 3028 (JSC::JIT::emitSlow_op_bitor): 3029 (JSC::JIT::emit_op_bitxor): 3030 (JSC::JIT::emitSlow_op_bitxor): 3031 (JSC::JIT::emit_op_bitnot): 3032 (JSC::JIT::emitSlow_op_bitnot): 3033 (JSC::JIT::emit_op_post_inc): 3034 (JSC::JIT::emitSlow_op_post_inc): 3035 (JSC::JIT::emit_op_post_dec): 3036 (JSC::JIT::emitSlow_op_post_dec): 3037 (JSC::JIT::emit_op_pre_inc): 3038 (JSC::JIT::emitSlow_op_pre_inc): 3039 (JSC::JIT::emit_op_pre_dec): 3040 (JSC::JIT::emitSlow_op_pre_dec): 3041 (JSC::JIT::emit_op_add): 3042 (JSC::JIT::emitAdd32Constant): 3043 (JSC::JIT::emitSlow_op_add): 3044 (JSC::JIT::emit_op_sub): 3045 (JSC::JIT::emitSub32Constant): 3046 (JSC::JIT::emitSlow_op_sub): 3047 (JSC::JIT::emitBinaryDoubleOp): 3048 (JSC::JIT::emit_op_mul): 3049 (JSC::JIT::emitSlow_op_mul): 3050 (JSC::JIT::emit_op_div): 3051 (JSC::JIT::emitSlow_op_div): 3052 (JSC::JIT::emit_op_mod): 3053 (JSC::JIT::emitSlow_op_mod): 3054 30552010-04-29 Zoltan Herczeg <zherczeg@webkit.org> 3056 3057 Reviewed by Oliver Hunt 3058 3059 Add sqrt() instruction support for ARM assembler (using VFP) 3060 https://bugs.webkit.org/show_bug.cgi?id=38312 3061 3062 * assembler/ARMAssembler.h: 3063 (JSC::ARMAssembler::): 3064 (JSC::ARMAssembler::fsqrtd_r): 3065 * assembler/MacroAssemblerARM.h: 3066 (JSC::MacroAssemblerARM::supportsFloatingPointSqrt): 3067 (JSC::MacroAssemblerARM::sqrtDouble): 3068 30692010-04-29 Laszlo Gombos <laszlo.1.gombos@nokia.com> 3070 3071 Reviewed by Simon Hausmann. 3072 3073 [Qt] Refactor qtscriptglobal.h 3074 https://bugs.webkit.org/show_bug.cgi?id=37953 3075 3076 Use the same pattern in qtscriptglobal.h as in 3077 qwebkitglobal.h without checking for specific OSs. 3078 3079 * qt/api/qtscriptglobal.h: 3080 30812010-04-29 Jedrzej Nowacki <jedrzej.nowacki@nokia.com> 3082 3083 Reviewed by Simon Hausmann. 3084 3085 Small cleanup. 3086 3087 QScriptValuePrivate constructor shouldn't take QScriptEngine pointer 3088 as a parameter, because it breaks a private / public separation rule. 3089 QScriptEnginePrivate::get(QScriptEngine*) should return 3090 a QScriptEnginePrivate pointer instead of a QScriptEnginePtr, it simplifies 3091 implementation. 3092 3093 [Qt] QScriptValuePrivate class needs some cleanup. 3094 https://bugs.webkit.org/show_bug.cgi?id=37729 3095 3096 * qt/api/qscriptengine_p.h: 3097 (QScriptEnginePrivate::get): 3098 * qt/api/qscriptvalue.cpp: 3099 (QScriptValue::QScriptValue): 3100 * qt/api/qscriptvalue_p.h: 3101 (QScriptValuePrivate::QScriptValuePrivate): 3102 31032010-04-29 Kwang Yul Seo <skyul@company100.net> 3104 3105 Reviewed by Simon Hausmann. 3106 3107 [WINCE] Export g_stackBase with JS_EXPORTDATA 3108 https://bugs.webkit.org/show_bug.cgi?id=37437 3109 3110 Export g_stackBase with JS_EXPORTDATA as it is used by SharedTimerWince.cpp. 3111 3112 * runtime/Collector.cpp: 3113 31142010-04-28 Oliver Hunt <oliver@apple.com> 3115 3116 Do an unordered compare so that we don't treat NaN incorrectly. 3117 3118 RS=Gavin Barraclough 3119 3120 * jit/ThunkGenerators.cpp: 3121 (JSC::powThunkGenerator): 3122 31232010-04-28 Oliver Hunt <oliver@apple.com> 3124 3125 Reviewed by Gavin Barraclough. 3126 3127 Add fast paths for Math.pow and Math.sqrt 3128 https://bugs.webkit.org/show_bug.cgi?id=38294 3129 3130 Add specialized thunks for Math.pow and Math.sqrt. 3131 This requires adding a sqrtDouble function to the MacroAssembler 3132 and sqrtsd to the x86 assembler. 3133 3134 Math.pow is slightly more complicated, in that we have 3135 to implement exponentiation ourselves rather than relying 3136 on hardware support. The inline exponentiation is restricted 3137 to positive integer exponents on a numeric base. Exponentiation 3138 is finally performed through the "Exponentiation by Squaring" 3139 algorithm. 3140 3141 * assembler/AbstractMacroAssembler.h: 3142 (JSC::AbstractMacroAssembler::ImmPtr::ImmPtr): 3143 * assembler/MacroAssemblerARM.h: 3144 (JSC::MacroAssemblerARM::supportsFloatingPointSqrt): 3145 (JSC::MacroAssemblerARM::loadDouble): 3146 (JSC::MacroAssemblerARM::sqrtDouble): 3147 * assembler/MacroAssemblerARMv7.h: 3148 (JSC::MacroAssemblerARMv7::supportsFloatingPointSqrt): 3149 (JSC::MacroAssemblerARMv7::sqrtDouble): 3150 * assembler/MacroAssemblerX86.h: 3151 (JSC::MacroAssemblerX86::loadDouble): 3152 (JSC::MacroAssemblerX86::supportsFloatingPointSqrt): 3153 * assembler/MacroAssemblerX86Common.h: 3154 (JSC::MacroAssemblerX86Common::sqrtDouble): 3155 * assembler/MacroAssemblerX86_64.h: 3156 (JSC::MacroAssemblerX86_64::loadDouble): 3157 (JSC::MacroAssemblerX86_64::supportsFloatingPointSqrt): 3158 * assembler/X86Assembler.h: 3159 (JSC::X86Assembler::): 3160 (JSC::X86Assembler::movsd_mr): 3161 (JSC::X86Assembler::sqrtsd_rr): 3162 (JSC::X86Assembler::X86InstructionFormatter::twoByteOp): 3163 (JSC::X86Assembler::X86InstructionFormatter::memoryModRM): 3164 * create_hash_table: 3165 * jit/JIT.h: 3166 * jit/JITInlineMethods.h: 3167 * jit/JITOpcodes.cpp: 3168 * jit/JITStubs.h: 3169 (JSC::JITThunks::ctiNativeCallThunk): 3170 * jit/JSInterfaceJIT.h: 3171 (JSC::JSInterfaceJIT::emitLoadDouble): 3172 (JSC::JSInterfaceJIT::emitJumpIfImmediateNumber): 3173 (JSC::JSInterfaceJIT::emitJumpIfNotImmediateNumber): 3174 (JSC::JSInterfaceJIT::emitLoadInt32): 3175 * jit/SpecializedThunkJIT.h: 3176 (JSC::SpecializedThunkJIT::loadDoubleArgument): 3177 (JSC::SpecializedThunkJIT::loadInt32Argument): 3178 (JSC::SpecializedThunkJIT::returnJSValue): 3179 (JSC::SpecializedThunkJIT::returnDouble): 3180 (JSC::SpecializedThunkJIT::finalize): 3181 * jit/ThunkGenerators.cpp: 3182 (JSC::sqrtThunkGenerator): 3183 (JSC::powThunkGenerator): 3184 * jit/ThunkGenerators.h: 3185 * runtime/Executable.h: 3186 (JSC::NativeExecutable::NativeExecutable): 3187 * runtime/JSFunction.cpp: 3188 (JSC::JSFunction::JSFunction): 3189 31902010-04-28 Mike Thole <mthole@apple.com> 3191 3192 Reviewed by Mark Rowe. 3193 3194 Reenable WTF_USE_PROTECTION_SPACE_AUTH_CALLBACK now that an 3195 exports file for it has been added. 3196 3197 * wtf/Platform.h: 3198 31992010-04-28 Mike Thole <mthole@apple.com> 3200 3201 Build fix, not reviewed. 3202 3203 * wtf/Platform.h: 3204 Disable WTF_USE_PROTECTION_SPACE_AUTH_CALLBACK until an exports file is created. 3205 32062010-04-28 Mike Thole <mthole@apple.com> 3207 3208 Reviewed by David Kilzer. 3209 3210 Add canAuthenticateAgainstProtectionSpace() to frame loader so that a protection space 3211 can be inspected before attempting to authenticate against it 3212 https://bugs.webkit.org/show_bug.cgi?id=38271 3213 3214 * wtf/Platform.h: Add WTF_USE_PROTECTION_SPACE_AUTH_CALLBACK define. 3215 32162010-04-28 Darin Adler <darin@apple.com> 3217 3218 Reviewed by Sam Weinig. 3219 3220 Don't use CFAbsoluteTimeGetCurrent on Mac OS X, since it's just a cover 3221 for gettimeofday anyway. 3222 3223 * wtf/CurrentTime.cpp: 3224 (WTF::currentTime): Remove the CF version of this. Change the "other POSIX" 3225 version to pass 0 for the time zone instead of ignoring a return value. 3226 Also get rid of unneeded type casts, since operations involving both 3227 integers and doubles are always evaluated as doubles in C without explicit 3228 type casting. 3229 32302010-04-28 Darin Fisher <darin@chromium.org> 3231 3232 Reviewed by David Levin. 3233 3234 [Chromium] callOnMainThread should equate to MessageLoop::PostTask 3235 https://bugs.webkit.org/show_bug.cgi?id=38276 3236 3237 * JavaScriptCore.gyp/JavaScriptCore.gyp: 3238 * wtf/MainThread.cpp: 3239 * wtf/chromium/ChromiumThreading.h: 3240 * wtf/chromium/MainThreadChromium.cpp: 3241 (WTF::initializeMainThread): 3242 (WTF::callOnMainThread): 3243 (WTF::callOnMainThreadAndWait): 3244 (WTF::setMainThreadCallbacksPaused): 3245 (WTF::isMainThread): 3246 32472010-04-28 Simon Hausmann <simon.hausmann@nokia.com>, Kent Hansen <kent.hansen@nokia.com> 3248 3249 Reviewed by Darin Adler. 3250 3251 JSC's currentThreadStackBase is not reentrant on some platforms 3252 https://bugs.webkit.org/show_bug.cgi?id=37195 3253 3254 This function needs to be reentrant to avoid memory corruption on platforms where 3255 the implementation uses global variables. 3256 3257 This patch adds a mutex lock where necessary and makes the Symbian implementation 3258 reentrant. 3259 3260 * runtime/Collector.cpp: 3261 (JSC::currentThreadStackBaseMutex): 3262 (JSC::currentThreadStackBase): 3263 32642010-04-28 Thiago Macieira <thiago.macieira@nokia.com> 3265 3266 Reviewed by Simon Hausmann. 3267 3268 Fix compilation with gcc 3.4. 3269 3270 AtomicStringImpl.h:29: error: base `WebCore::StringImpl' with only 3271 non-default constructor in class without a constructor 3272 3273 * wtf/text/AtomicStringImpl.h: 3274 (WebCore::AtomicStringImpl::AtomicStringImpl): Provide a default constructor. 3275 * wtf/text/StringImpl.h: Allow AtomicStringImpl to call a StringImpl constructor through a friend declaration. 3276 32772010-04-27 Stephanie Lewis <slewis@apple.com> 3278 3279 Fix Tiger build. 3280 3281 * wtf/FastMalloc.cpp: 3282 (WTF::TCMalloc_PageHeap::Carve): 3283 32842010-04-26 Stephanie Lewis <slewis@apple.com> 3285 3286 Reviewed by Geoff Garen. 3287 3288 https://bugs.webkit.org/show_bug.cgi?id=38154 FastMalloc calls madvise too often. 3289 <rdar://problem/7834433> REGRESSSION: 1.5% PLT regression due to 56028 (return memory quicker). 3290 To save on madvise calls when recommitting memory recommit the entire span and then carve it 3291 instead of carving the span up and only committing the part that will be used immediately. 3292 3293 * wtf/FastMalloc.cpp: 3294 (WTF::TCMalloc_PageHeap::New): 3295 (WTF::TCMalloc_PageHeap::AllocLarge): 3296 (WTF::TCMalloc_PageHeap::Carve): 3297 32982010-04-26 Oliver Hunt <oliver@apple.com> 3299 3300 Reviewed by Maciej Stachowiak. 3301 3302 Add specialized thunks for a char code related string functions 3303 https://bugs.webkit.org/show_bug.cgi?id=38177 3304 3305 This is a simple extension to my prior patch. Basically hoist the 3306 existing character loading logic from the charCodeAt thunk and into 3307 a helper function, then add a helper to load strings from the small 3308 string cache. This allows trivial implementations of fromCharCode 3309 and charAt. 3310 3311 * create_hash_table: 3312 * jit/SpecializedThunkJIT.h: 3313 (JSC::SpecializedThunkJIT::returnJSCell): 3314 (JSC::SpecializedThunkJIT::tagReturnAsJSCell): 3315 * jit/ThunkGenerators.cpp: 3316 (JSC::stringCharLoad): 3317 (JSC::charToString): 3318 (JSC::charCodeAtThunkGenerator): 3319 (JSC::charAtThunkGenerator): 3320 (JSC::fromCharCodeThunkGenerator): 3321 * jit/ThunkGenerators.h: 3322 * runtime/SmallStrings.h: 3323 (JSC::SmallStrings::singleCharacterStrings): 3324 * runtime/StringConstructor.cpp: 3325 (JSC::StringConstructor::StringConstructor): 3326 33272010-04-27 Thomas Zander <t.zander@nokia.com> 3328 3329 Reviewed by Simon Hausmann. 3330 3331 [Qt] Fix the build on Symbian on Linux using the Makefile based mkspec 3332 3333 * JavaScriptCore.pri: It is necessary to specify the library search path 3334 when linking against JavaScriptCore. 3335 33362010-04-21 Jocelyn Turcotte <jocelyn.turcotte@nokia.com> 3337 3338 Reviewed by Simon Hausmann. 3339 3340 Compile fix for Visual Studio 2010 3341 https://bugs.webkit.org/show_bug.cgi?id=37867 3342 3343 Patch by Prasanth Ullattil <prasanth.ullattil@nokia.com> 3344 3345 * runtime/Structure.cpp: 3346 (JSC::Structure::transitionTableAdd): 3347 * wtf/MathExtras.h: 3348 * wtf/Platform.h: 3349 33502010-04-27 Jocelyn Turcotte <jocelyn.turcotte@nokia.com> 3351 3352 Reviewed by Simon Hausmann. 3353 3354 Rename COMPILER(MSVC7) to COMPILER(MSVC7_OR_LOWER) 3355 3356 * config.h: 3357 (JSC::Structure::transitionTableAdd): 3358 * wtf/Assertions.h: 3359 * wtf/Atomics.h: 3360 * wtf/CurrentTime.h: 3361 (WTF::getLocalTime): 3362 * wtf/Platform.h: 3363 * wtf/StaticConstructors.h: 3364 * wtf/StringExtras.h: 3365 * wtf/Vector.h: 3366 (WTF::::append): 3367 33682010-04-15 Jocelyn Turcotte <jocelyn.turcotte@nokia.com> 3369 3370 Reviewed by Tor Arne Vestbø. 3371 3372 [Qt] Fix Mac build break. 3373 3374 https://bugs.webkit.org/show_bug.cgi?id=37867 3375 - Make sure JavaScriptCore's debug_and_release config follows that 3376 of WebCore. 3377 - Use different name for JavaScriptCore target in debug and release 3378 since they might be written in the same path at the same time on 3379 parallel builds 3380 - Consolidate the DESTDIR determination logic in JavaScriptCore.pri 3381 3382 * JavaScriptCore.pri: 3383 * JavaScriptCore.pro: 3384 33852010-04-21 Jocelyn Turcotte <jocelyn.turcotte@nokia.com> 3386 3387 Reviewed by Kenneth Rohde Christiansen. 3388 3389 [Qt] Make sure WebKit is not compiled using C++0x. 3390 https://bugs.webkit.org/show_bug.cgi?id=37867 3391 3392 As the rest of Qt compiles in the C++0x mode, people might start 3393 compiling it in this mode. WebKit don't support this yet. 3394 3395 Patch by Thiago Macieira <thiago.macieira@nokia.com> 3396 3397 * JavaScriptCore.pro: 3398 33992010-04-26 Oliver Hunt <oliver@apple.com> 3400 3401 Fix windows 3402 3403 * jit/SpecializedThunkJIT.h: 3404 (JSC::SpecializedThunkJIT::argumentToVirtualRegister): 3405 34062010-04-26 Oliver Hunt <oliver@apple.com> 3407 3408 Reviewed by Gavin Barraclough. 3409 3410 Need to support more efficient dispatch of lightweight builtins 3411 https://bugs.webkit.org/show_bug.cgi?id=38155 3412 3413 The problem with calling lightweight builtins is that marshalling 3414 the call from the nitro calling convention to the system calling 3415 convention is very expensive relative to the cost of the actually 3416 evaluating the function. To work around this problem this patch 3417 adds the ability to define a custom thunk for a builtin. 3418 3419 This allows us to use high performance custom implementations of 3420 the common and sensible versions of simple builtins. This patch 3421 includes a simple (use of which is currently hardcoded) thunk for 3422 charCodeAt. 3423 3424 This adds a JSInterfaceJIT subclass called SpecializedThunkJIT 3425 that has helper functions to remove (or at least reduce) the need 3426 to have separate thunk implementations for each JSValue encoding. 3427 3428 * GNUmakefile.am: 3429 * JavaScriptCore.pro: 3430 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: 3431 * JavaScriptCore.xcodeproj/project.pbxproj: 3432 Many project file updates 3433 3434 * create_hash_table: 3435 Add thunk generation callbacks to the hash tables, currently 3436 we hardcode the script to only support charCodeAt 3437 3438 * jit/JITStubCall.h: 3439 * jit/JITStubs.cpp: 3440 (JSC::JITThunks::~JITThunks): 3441 (JSC::JITThunks::specializedThunk): 3442 * jit/JITStubs.h: 3443 * jit/SpecializedThunkJIT.h: Added. 3444 (JSC::SpecializedThunkJIT::SpecializedThunkJIT): 3445 (JSC::SpecializedThunkJIT::loadCellArgument): 3446 (JSC::SpecializedThunkJIT::loadJSStringArgument): 3447 (JSC::SpecializedThunkJIT::loadInt32Argument): 3448 (JSC::SpecializedThunkJIT::appendFailure): 3449 (JSC::SpecializedThunkJIT::returnInt32): 3450 (JSC::SpecializedThunkJIT::finalize): 3451 (JSC::SpecializedThunkJIT::argumentToVirtualRegister): 3452 (JSC::SpecializedThunkJIT::tagReturnAsInt32): 3453 * jit/ThunkGenerators.cpp: Added. 3454 (JSC::ThunkHelpers::stringImplDataOffset): 3455 (JSC::ThunkHelpers::jsStringLengthOffset): 3456 (JSC::ThunkHelpers::jsStringValueOffset): 3457 (JSC::charCodeAtThunkGenerator): 3458 * jit/ThunkGenerators.h: Added. 3459 * runtime/Executable.h: 3460 (JSC::NativeExecutable::NativeExecutable): 3461 * runtime/JSFunction.cpp: 3462 (JSC::JSFunction::JSFunction): 3463 * runtime/JSFunction.h: 3464 * runtime/JSGlobalData.h: 3465 (JSC::JSGlobalData::getThunk): 3466 * runtime/JSString.h: 3467 (JSC::): 3468 * runtime/JSValue.h: 3469 Protect some of the JSVALUE32 and JSVALUE32_64 only constants 3470 and function behind appropriate USE() guards to make it harder 3471 to use the wrong flags for the target build. 3472 * runtime/Lookup.cpp: 3473 (JSC::HashTable::createTable): 3474 (JSC::setUpStaticFunctionSlot): 3475 * runtime/Lookup.h: 3476 (JSC::HashEntry::initialize): 3477 (JSC::HashEntry::generator): 3478 (JSC::HashEntry::): 3479 Make the lookup tables use a specialized thunkGenerator 3480 if present 3481 * wtf/text/StringImpl.h: 3482 34832010-04-26 Oliver Hunt <oliver@apple.com> 3484 3485 Build fix 3486 3487 * runtime/JSImmediate.h: 3488 * runtime/JSValue.h: 3489 34902010-04-26 Oliver Hunt <oliver@apple.com> 3491 3492 Reviewed by Gavin Barraclough. 3493 3494 Factor some basic JS interface logic out of JIT and into a superclass 3495 https://bugs.webkit.org/show_bug.cgi?id=38163 3496 3497 Create JSInterfaceJIT to hold some common low level functions for 3498 interacting with the JS RegisterFile. Doing this separately from 3499 https://bugs.webkit.org/show_bug.cgi?id=38155 to ease reviewing. 3500 3501 * JavaScriptCore.xcodeproj/project.pbxproj: 3502 * jit/JIT.h: 3503 * jit/JITInlineMethods.h: 3504 (JSC::JIT::emitJumpSlowCaseIfNotJSCell): 3505 * jit/JSInterfaceJIT.h: Added. 3506 (JSC::JSInterfaceJIT::emitLoadJSCell): 3507 (JSC::JSInterfaceJIT::emitJumpIfNotJSCell): 3508 (JSC::JSInterfaceJIT::emitLoadInt32): 3509 (JSC::JSInterfaceJIT::tagFor): 3510 (JSC::JSInterfaceJIT::payloadFor): 3511 (JSC::JSInterfaceJIT::addressFor): 3512 35132010-04-26 Sam Weinig <sam@webkit.org> 3514 3515 Fix Tiger build. 3516 3517 * wtf/mac/MainThreadMac.mm: 3518 (WTF::isMainThread): 3519 35202010-04-26 Geoffrey Garen <ggaren@apple.com> 3521 3522 Reviewed by Sam Weinig. 3523 3524 <rdar://problem/7766413> 3525 3526 Fixed a crash seen when using the JavaScriptCore API with WebKit. 3527 3528 No layout test because DumpRenderTree doesn't use the JavaScriptCore API 3529 in this way. 3530 3531 * interpreter/RegisterFile.cpp: 3532 (JSC::RegisterFile::setGlobalObject): 3533 (JSC::RegisterFile::clearGlobalObject): 3534 (JSC::RegisterFile::globalObject): 3535 * interpreter/RegisterFile.h: 3536 (JSC::RegisterFile::RegisterFile): Use WeakGCPtr for managing m_globalObject, 3537 since it's a weak pointer. (We never noticed this error before because, 3538 in WebKit, global objects always have a self-reference in a global variable, 3539 so marking the register file's global variables would keep m_globalObject 3540 alive. In the JavaScriptCore API, you can allocate a global object with 3541 no self-reference.) 3542 3543 * runtime/JSActivation.h: Removed unneeded #include. 3544 3545 * runtime/JSGlobalObject.cpp: 3546 (JSC::JSGlobalObject::~JSGlobalObject): Don't use ==, since a weak 3547 pointer is 0 when the object it points to runs its destructor. 3548 3549 * runtime/WeakGCPtr.h: 3550 (JSC::WeakGCPtr::clear): Changed to return a bool indicating whether the 3551 clear actually happened. 3552 (JSC::WeakGCPtr::assign): Changed to forbid assignment of 0 as a shorthand 3553 for calling clear(). A client should never clear by assigning 0, since 3554 clear() should be conditional on whether the object doing the clearing 3555 is still pointed to by the weak pointer. (Otherwise, a zombie object might 3556 clear a pointer to a new, valid object.) 3557 35582010-04-25 Sam Weinig <sam@webkit.org> 3559 3560 Reviewed by Maciej Stachowiak. 3561 3562 Fix for https://bugs.webkit.org/show_bug.cgi?id=38097 3563 Disentangle initializing the main thread from initializing threading 3564 3565 This is the first step in getting rid of the WEB_THREAD #define and 3566 paving the way to allowing mac both WebKit and WebKit2 to use the same 3567 WebCore. 3568 3569 * JavaScriptCore.exp: Export initializeMainThread and initializeMainThreadToProcessMainThread. 3570 * JavaScriptCore.xcodeproj/project.pbxproj: Sort files. 3571 3572 * wtf/MainThread.cpp: 3573 (WTF::mainThreadInitializationMutex): 3574 (WTF::mainThreadFunctionQueueMutex): 3575 (WTF::initializeMainThread): 3576 (WTF::deprecatedInitializeMainThread): 3577 * wtf/MainThread.h: 3578 Added Mac only initializeMainThreadToProcessMainThread which sets up callOnMainThread 3579 and isMainThead calls to assume that WebCore's main thread is also the 3580 process's main thread. Since it is possible that both initializeMainThread 3581 and initializeMainThreadToProcessMainThread could be called, the first one called 3582 wins (either will work). 3583 3584 * wtf/Threading.h: Moved to top of file. 3585 3586 * wtf/ThreadingPthreads.cpp: 3587 (WTF::initializeThreading): Remove call to initializeMainThread. 3588 * wtf/ThreadingWin.cpp: 3589 (WTF::initializeThreading): Ditto. 3590 * wtf/gtk/ThreadingGtk.cpp: 3591 (WTF::initializeThreading): Ditto. 3592 3593 * wtf/mac/MainThreadMac.mm: 3594 (WTF::initializeMainThreadPlatform): 3595 (WTF::initializeMainThreadToProcessMainThread): 3596 (WTF::scheduleDispatchFunctionsOnMainThread): 3597 (WTF::isMainThread): 3598 Add runtime switch between the old behavior of using the system's main 3599 thread and a stored pointer to the main thread. Tiger always uses the 3600 system's main thread. 3601 3602 * wtf/qt/ThreadingQt.cpp: 3603 (WTF::initializeThreading): Remove call to initializeMainThread. 3604 * wtf/win/MainThreadWin.cpp: 3605 (WTF::initializeMainThreadPlatform): Add call to initializeCurrentThreadInternal 3606 removed from initializeThreading. 3607 36082010-04-26 Jedrzej Nowacki <jedrzej.nowacki@nokia.com> 3609 3610 Reviewed by Kenneth Rohde Christiansen. 3611 3612 Fix a memory leak in the QScriptString. 3613 3614 QScriptStringPrivate's constructor shouldn't call JSStringRetain as 3615 QScriptConverter implicitly pass a JSStringRef ownership. 3616 3617 [Qt] QScriptStringPrivate c'tor problem 3618 https://bugs.webkit.org/show_bug.cgi?id=38110 3619 3620 * qt/api/qscriptstring_p.h: 3621 (QScriptStringPrivate::QScriptStringPrivate): 3622 36232010-04-24 Darin Adler <darin@apple.com> 3624 3625 Reviewed by Dan Bernstein. 3626 3627 REGRESSION (r56560): Crash in parseFloat if passed invalid UTF-16 data 3628 https://bugs.webkit.org/show_bug.cgi?id=38083 3629 rdar://problem/7901044 3630 3631 Tests: fast/js/ToNumber.html 3632 fast/js/parseFloat.html 3633 3634 * runtime/JSGlobalObjectFunctions.cpp: 3635 (JSC::parseInt): Added a FIXME comment about a problem I noticed. 3636 (JSC::parseFloat): Added a FIXME comment about a problem I noticed; 3637 covered by test cases in the test I added. 3638 * runtime/UString.cpp: 3639 (JSC::UString::toDouble): Added FIXME comments about two problems I 3640 noticed; covered by test cases in the tests I added. Added a return 3641 statement so we don't crash when illegal UTF-16 sequences are present. 3642 36432010-04-24 Anton Muhin <antonm@chromium.org> 3644 3645 Reviewed by Darin Adler. 3646 3647 Allow to construct HashTraits<WebCore::QualifiedName>::constructDeletedValue 3648 3649 Former implementation attempted to use AtomicString(HashTableDeletedValue) 3650 however those values cannot be used that way: one cannot construct 3651 QualifiedNameImpl out of such AtomicString as we'll try to lookup this string 3652 in the table, for example. 3653 https://bugs.webkit.org/show_bug.cgi?id=37722 3654 3655 * wtf/RefPtr.h: expose hash table deleted value 3656 36572010-04-23 Sam Weinig <sam@webkit.org> 3658 3659 Reviewed by David Levin. 3660 3661 Fix for https://bugs.webkit.org/show_bug.cgi?id=38060 3662 Split up Threading.h 3663 3664 Move bits for of Threading.h into 3 new files. 3665 - Atomics.h for atomic operations. 3666 - ThreadSafeShared.h for the ThreadSafeShared class. 3667 - ThreadingPrimitives.h for the primitives and platform types. 3668 3669 Basic threading operations (creation, etc.) remain in Threading.h. 3670 3671 * GNUmakefile.am: 3672 * JavaScriptCore.gypi: 3673 * JavaScriptCore.vcproj/WTF/WTF.vcproj: 3674 * JavaScriptCore.xcodeproj/project.pbxproj: 3675 * wtf/Atomics.h: Copied from wtf/Threading.h. 3676 * wtf/ThreadSafeShared.h: Copied from wtf/Threading.h. 3677 * wtf/Threading.h: 3678 * wtf/ThreadingPrimitives.h: Copied from wtf/Threading.h. 3679 36802010-04-23 Sam Weinig <sam@webkit.org> 3681 3682 Fix Qt build. 3683 3684 * wtf/qt/MainThreadQt.cpp: #include <QThread> 3685 36862010-04-22 Sam Weinig <sam@webkit.org> 3687 3688 Reviewed by Anders Carlsson. 3689 3690 Fix for https://bugs.webkit.org/show_bug.cgi?id=38022 3691 Move isMainThread predicate function to MainThread.h 3692 3693 * wtf/MainThread.cpp: 3694 (WTF::initializeMainThread): 3695 (WTF::isMainThread): 3696 * wtf/MainThread.h: 3697 * wtf/Threading.h: 3698 * wtf/ThreadingPthreads.cpp: 3699 (WTF::initializeThreading): 3700 * wtf/ThreadingWin.cpp: 3701 (WTF::initializeThreading): 3702 * wtf/gtk/ThreadingGtk.cpp: 3703 (WTF::initializeThreading): 3704 * wtf/mac/MainThreadMac.mm: 3705 (WTF::initializeMainThreadPlatform): 3706 (WTF::isMainThread): 3707 * wtf/qt/MainThreadQt.cpp: 3708 (WTF::isMainThread): 3709 * wtf/qt/ThreadingQt.cpp: 3710 (WTF::initializeThreading): 3711 * wtf/text/AtomicString.cpp: 3712 37132010-04-23 Gabor Rapcsanyi <rgabor@inf.u-szeged.hu> 3714 3715 Reviewed by Laszlo Gombos. 3716 3717 Add RVCT support for Linux 3718 3719 Developed in cooperation with Gabor Loki. 3720 3721 * API/JSStringRef.h: 3722 * jit/ExecutableAllocator.h: 3723 * jit/ExecutableAllocatorPosix.cpp: 3724 (JSC::ExecutableAllocator::cacheFlush): 3725 * jit/JITStubs.cpp: 3726 * wtf/MathExtras.h: 3727 * wtf/unicode/qt4/UnicodeQt4.h: 3728 37292010-04-23 Patrick Gansterer <paroga@paroga.com> 3730 3731 Reviewed by Laszlo Gombos. 3732 3733 [WIN] Move OwnPtrWin.cpp into platform specific folder. 3734 https://bugs.webkit.org/show_bug.cgi?id=38042 3735 3736 * JavaScriptCore.gypi: 3737 * JavaScriptCore.vcproj/WTF/WTF.vcproj: 3738 * wtf/OwnPtrWin.cpp: Removed. 3739 * wtf/win/OwnPtrWin.cpp: Copied from JavaScriptCore/wtf/OwnPtrWin.cpp. 3740 37412010-04-23 Patrick Gansterer <paroga@paroga.com> 3742 3743 Reviewed by Laszlo Gombos. 3744 3745 [BREWMP] Move OwnPtrBrew.cpp into platform specific folder. 3746 https://bugs.webkit.org/show_bug.cgi?id=38042 3747 3748 * wtf/OwnPtrBrew.cpp: Removed. 3749 * wtf/brew/OwnPtrBrew.cpp: Copied from JavaScriptCore/wtf/OwnPtrBrew.cpp. 3750 37512010-04-22 Steve Block <steveblock@google.com> 3752 3753 Reviewed by Adam Barth. 3754 3755 Update Android to use isfinite, isinf, isnan and signbit from namespace std. 3756 https://bugs.webkit.org/show_bug.cgi?id=37948 3757 3758 * wtf/MathExtras.h: 3759 37602010-04-22 Fumitoshi Ukai <ukai@chromium.org> 3761 3762 Unreviewed build fix. 3763 3764 * wtf/MD5.cpp: fix include path of CString.h 3765 37662010-04-22 Fumitoshi Ukai <ukai@chromium.org> 3767 3768 Reviewed by Adam Barth. 3769 3770 MD5 is required for WebSocket new protocol implementation 3771 https://bugs.webkit.org/show_bug.cgi?id=37913 3772 3773 * GNUmakefile.am: 3774 * JavaScriptCore.exp: 3775 * JavaScriptCore.gypi: 3776 * JavaScriptCore.pro: 3777 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 3778 * JavaScriptCore.vcproj/WTF/WTF.vcproj: 3779 * JavaScriptCore.xcodeproj/project.pbxproj: 3780 * wtf/MD5.cpp: Added. 3781 * wtf/MD5.h: Added. 3782 37832010-04-22 Gavin Barraclough <barraclough@apple.com> 3784 3785 Reviewed by NOBODY (build fix). 3786 Errk, accidentally committed commenting out two ASSERTs! reverting. 3787 3788 * runtime/Collector.cpp: 3789 (JSC::Heap::protect): 3790 (JSC::Heap::unprotect): 3791 37922010-04-22 Gavin Barraclough <barraclough@apple.com> 3793 3794 Reviewed by Geoff Garen. 3795 3796 https://bugs.webkit.org/show_bug.cgi?id=38006 3797 Change lifetime of JSC::IdentifierTables used by WebCores to match AtomicStringTable 3798 3799 Presently JSC's IdentifierTables are owned by the JSGlobalData. For 3800 JSGlobalData objects created via the API this should continue to be the case, 3801 but for the JSGlobalData objects used by WebCore (the main thread's common 3802 global data, and those for workers) use a IdentifierTable provided (and owned) 3803 by wtfThreadData. This allow the lifetime of these IdentifierTable to match 3804 those of the corresponding AtomicStringTables. 3805 3806 * API/APIShims.h: 3807 (JSC::APIEntryShim::APIEntryShim): 3808 * API/JSContextRef.cpp: 3809 (JSContextGroupCreate): 3810 * runtime/Collector.cpp: 3811 (JSC::Heap::protect): 3812 (JSC::Heap::unprotect): 3813 (JSC::Heap::markRoots): 3814 * runtime/JSGlobalData.cpp: 3815 (JSC::JSGlobalData::JSGlobalData): 3816 (JSC::JSGlobalData::~JSGlobalData): 3817 (JSC::JSGlobalData::createContextGroup): 3818 (JSC::JSGlobalData::create): 3819 (JSC::JSGlobalData::sharedInstance): 3820 * runtime/JSGlobalData.h: 3821 (JSC::JSGlobalData::): 3822 (JSC::JSGlobalData::isSharedInstance): 3823 * runtime/JSLock.cpp: 3824 (JSC::JSLock::JSLock): 3825 (JSC::JSLock::lock): 3826 (JSC::JSLock::unlock): 3827 (JSC::JSLock::DropAllLocks::DropAllLocks): 3828 * wtf/WTFThreadData.cpp: 3829 (WTF::WTFThreadData::WTFThreadData): 3830 (WTF::WTFThreadData::~WTFThreadData): 3831 38322010-04-22 Sheriff Bot <webkit.review.bot@gmail.com> 3833 3834 Unreviewed, rolling out r58110. 3835 http://trac.webkit.org/changeset/58110 3836 https://bugs.webkit.org/show_bug.cgi?id=38007 3837 3838 Caused an internal compiler error on Qt (Requested by abarth 3839 on #webkit). 3840 3841 * wtf/MathExtras.h: 3842 38432010-04-22 Gavin Barraclough <barraclough@apple.com> 3844 3845 Reviewed by NOBODY (windows build fix). 3846 3847 * API/JSWeakObjectMapRefPrivate.cpp: 3848 38492010-04-22 Gavin Barraclough <barraclough@apple.com> 3850 3851 Reviewed by NOBODY (windows build fix). 3852 3853 * API/JSBase.cpp: 3854 * API/JSCallbackObject.cpp: 3855 38562010-04-22 Gavin Barraclough <barraclough@apple.com> 3857 3858 Reviewed by Geoff Garen. 3859 3860 https://bugs.webkit.org/show_bug.cgi?id=37978 3861 Unify JSC::IdentifierTable and WebCore::AtomicStringTable implementations. 3862 3863 These two classes both implement a HashSet of uniqued StringImpls, with 3864 translator classes to avoid unnecessary object creation. The only difference 3865 between the classes is which flag (isIdentifier or inTable) is set. 3866 Combine the two classes using a template predicated on which flag to use. 3867 3868 New class AtomicStringTable created, containing all the goodness from 3869 IdentifierTable & AtomicStringTable, expect for Identifier's literalTable, 3870 which has been moved onto JSGlobalData. Removed duplicate string translator 3871 classes. Renamed StringImpl's inTable flag to more explicit 'isAtomic', 3872 and set this on the empty string (which matches Identifier behaviour, and 3873 removes a redundant check for zero-length). 3874 3875 * GNUmakefile.am: 3876 * JavaScriptCore.gypi: 3877 * JavaScriptCore.vcproj/WTF/WTF.vcproj: 3878 * JavaScriptCore.xcodeproj/project.pbxproj: 3879 * runtime/Identifier.cpp: 3880 (JSC::createLiteralTable): 3881 (JSC::deleteLiteralTable): 3882 (JSC::Identifier::add): 3883 (JSC::Identifier::addSlowCase): 3884 * runtime/Identifier.h: 3885 * runtime/JSGlobalData.cpp: 3886 (JSC::JSGlobalData::JSGlobalData): 3887 (JSC::JSGlobalData::~JSGlobalData): 3888 * runtime/JSGlobalData.h: 3889 * wtf/WTFThreadData.cpp: 3890 (WTF::WTFThreadData::WTFThreadData): 3891 (WTF::WTFThreadData::~WTFThreadData): 3892 * wtf/WTFThreadData.h: 3893 (WTF::WTFThreadData::atomicStringTable): 3894 * wtf/text/AtomicString.cpp: 3895 (WebCore::table): 3896 (WebCore::operator==): 3897 (WebCore::AtomicString::add): 3898 (WebCore::AtomicString::find): 3899 (WebCore::AtomicString::remove): 3900 * wtf/text/AtomicStringTable.h: Added. 3901 (WTF::CStringTranslator::hash): 3902 (WTF::CStringTranslator::equal): 3903 (WTF::CStringTranslator::translate): 3904 (WTF::UCharBufferTranslator::hash): 3905 (WTF::UCharBufferTranslator::equal): 3906 (WTF::UCharBufferTranslator::translate): 3907 (WTF::HashAndCharactersTranslator::hash): 3908 (WTF::HashAndCharactersTranslator::equal): 3909 (WTF::HashAndCharactersTranslator::translate): 3910 (WTF::IdentifierOrAtomicStringTable::remove): 3911 (WTF::::~IdentifierOrAtomicStringTable): 3912 (WTF::::add): 3913 (WTF::::find): 3914 * wtf/text/StringImpl.cpp: 3915 (WebCore::StringImpl::~StringImpl): 3916 * wtf/text/StringImpl.h: 3917 (WebCore::StringImpl::isAtomic): 3918 (WebCore::StringImpl::setIsAtomic): 3919 (WebCore::equal): 3920 * wtf/text/StringImplBase.h: 3921 (WTF::StringImplBase::StringImplBase): 3922 39232010-04-22 Steve Block <steveblock@google.com> 3924 3925 Reviewed by Adam Barth. 3926 3927 Update Android to use isfinite, isinf, isnan and signbit from namespace std. 3928 https://bugs.webkit.org/show_bug.cgi?id=37948 3929 3930 * wtf/MathExtras.h: 3931 39322010-04-22 Jocelyn Turcotte <jocelyn.turcotte@nokia.com> 3933 3934 Reviewed by Kenneth Rohde Christiansen. 3935 3936 [Qt] Remove dependency of JSC to QtGui 3937 https://bugs.webkit.org/show_bug.cgi?id=37867 3938 3939 The patch also make sure that hal.h is in the include path on Symbian. 3940 The dependency to QtGui took care of that before. 3941 3942 Patch by Thiago Macieira <thiago.macieira@nokia.com> 3943 and Rohan McGovern <rohan.mcgovern@nokia.com> 3944 3945 * JavaScriptCore.pri: 3946 * JavaScriptCore.pro: 3947 39482010-04-22 Jocelyn Turcotte <jocelyn.turcotte@nokia.com> 3949 3950 Reviewed by Tor Arne Vestbø. 3951 3952 [Qt] Corrects symbols visibility for JavaScriptCore. 3953 3954 https://bugs.webkit.org/show_bug.cgi?id=37867 3955 3956 * JavaScriptCore.pro: 3957 39582010-04-22 Gabor Loki <loki@webkit.org> 3959 3960 Reviewed by Gavin Barraclough. 3961 3962 Use BLX and BX to keep happy the return stack predictor above ARMv4 3963 https://bugs.webkit.org/show_bug.cgi?id=37862 3964 3965 Inspired by Jacob Bramley's patch from JaegerMonkey 3966 3967 * assembler/ARMAssembler.cpp: 3968 (JSC::ARMAssembler::executableCopy): 3969 * assembler/ARMAssembler.h: 3970 (JSC::ARMAssembler::): 3971 (JSC::ARMAssembler::bx): 3972 (JSC::ARMAssembler::blx): 3973 (JSC::ARMAssembler::loadBranchTarget): 3974 (JSC::ARMAssembler::jmp): 3975 (JSC::ARMAssembler::getLdrImmAddress): 3976 * assembler/MacroAssemblerARM.h: 3977 (JSC::MacroAssemblerARM::jump): 3978 (JSC::MacroAssemblerARM::nearCall): 3979 (JSC::MacroAssemblerARM::call): 3980 (JSC::MacroAssemblerARM::ret): 3981 (JSC::MacroAssemblerARM::prepareCall): 3982 (JSC::MacroAssemblerARM::call32): 3983 39842010-04-21 Andy Estes <aestes@apple.com> 3985 3986 Rubber stamped by Mark Rowe. 3987 3988 Export WTF::deleteOwnedPtr(HFONT). 3989 3990 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 3991 39922010-04-21 Gavin Barraclough <barraclough@apple.com> 3993 3994 Rubber Stamped by Oliver Hunt. 3995 3996 Make SmallStrings store an array of RefPtr<StringImpl>, 3997 instead of a direct array of StringImpls. This allows 3998 us to remove a friend (and a layering violation) from 3999 WTF::StringImpl, and makes it so that all StringImpls 4000 are individually heap allocated. 4001 4002 * runtime/SmallStrings.cpp: 4003 (JSC::SmallStringsStorage::rep): 4004 (JSC::SmallStringsStorage::SmallStringsStorage): 4005 * wtf/text/StringImpl.h: 4006 40072010-04-21 Sam Weinig <sam@webkit.org> 4008 4009 Reviewed by Geoffrey Garen. 4010 4011 Fix for https://bugs.webkit.org/show_bug.cgi?id=37937 4012 Wean JavaScriptCore off calls to isMainThread() 4013 4014 - Replace use of isMainThread() for interpreter reentry checks 4015 with a stored value on the JSGlobalData. 4016 - Replace use of isMainThread() for useMainThread only check in the 4017 collector with a stored exclusive thread. 4018 4019 * API/JSContextRef.cpp: 4020 (JSContextGroupCreate): 4021 Always default to a small stack type for uses of the JSC API. It is 4022 unlikely that the interpreter reentry required on the web will be as 4023 important for other uses of JavaScriptCore. 4024 4025 * JavaScriptCore.exp: 4026 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 4027 Update exports. 4028 4029 * interpreter/Interpreter.cpp: 4030 (JSC::Interpreter::execute): 4031 (JSC::Interpreter::prepareForRepeatCall): 4032 Use new stored JSGlobalData::maxReentryDepth instead of isMainThread(). 4033 4034 * interpreter/Interpreter.h: 4035 Rename MaxMainThreadReentryDepth to MaxLargeThreadReentryDepth and 4036 MaxSecondaryThreadReentryDepth to MaxSmallThreadReentryDepth. 4037 4038 * jsc.cpp: 4039 (main): Use the a large stack for jsc since it is always using the 4040 main thread. 4041 4042 * runtime/ArrayPrototype.cpp: 4043 (JSC::arrayProtoFuncToString): 4044 (JSC::arrayProtoFuncToLocaleString): 4045 (JSC::arrayProtoFuncJoin): 4046 Use new stored JSGlobalData::maxReentryDepth instead of isMainThread(). 4047 4048 * runtime/Collector.cpp: 4049 (JSC::Heap::registerThread): 4050 Use the concept of making JSC run on an exclusiveThread instead of 4051 forcing a mainThreadOnly assertion. 4052 4053 * runtime/JSGlobalData.cpp: 4054 (JSC::JSGlobalData::JSGlobalData): 4055 (JSC::JSGlobalData::createNonDefault): 4056 (JSC::JSGlobalData::create): 4057 (JSC::JSGlobalData::createLeaked): 4058 (JSC::JSGlobalData::sharedInstance): 4059 * runtime/JSGlobalData.h: 4060 Add ThreadStackType argument to JSGlobalData constructors and set 4061 maxReentryDepth based on it. 4062 40632010-04-21 Gavin Barraclough <barraclough@apple.com> 4064 4065 Reviewed by NOBODY (windows build fix pt. 3). 4066 4067 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 4068 40692010-04-21 Gavin Barraclough <barraclough@apple.com> 4070 4071 Reviewed by NOBODY (windows build fix pt. 2). 4072 4073 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 4074 40752010-04-21 Gavin Barraclough <barraclough@apple.com> 4076 4077 Reviewed by NOBODY (Qt build fix). 4078 4079 * JavaScriptCore.gypi: 4080 * JavaScriptCore.pro: 4081 * wtf/qt/StringQt.cpp: Copied from WebCore/platform/text/qt/StringQt.cpp. 4082 40832010-04-21 Gavin Barraclough <barraclough@apple.com> 4084 4085 Reviewed by NOBODY (windows build fix). 4086 4087 * API/JSValueRef.cpp: 4088 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: 4089 * runtime/Identifier.cpp: 4090 (JSC::IdentifierTable::~IdentifierTable): 4091 (JSC::IdentifierTable::add): 4092 * runtime/Identifier.h: 4093 * wtf/WTFThreadData.h: 4094 (JSC::IdentifierTable::remove): 4095 (JSC::IdentifierTable::literalTable): 4096 * wtf/text/StringImpl.cpp: 4097 (WebCore::StringImpl::~StringImpl): 4098 40992010-04-20 Gavin Barraclough <barraclough@apple.com> 4100 4101 Reviewed by Oliver Hunt, Darin Adler. 4102 4103 Bug 37906 - Remove JSC::UStringImpl; unify with StringImpl. 4104 4105 JSC::UStringImpl and WebCore::StringImpl (soon to be renamed to 4106 WTF::StringImpl) are almost identical. Remove duplication of code by unifying 4107 the two, move missing features from UStringImpl into StringImpl & delete the 4108 class UStringImpl. 4109 4110 * API/JSClassRef.cpp: 4111 * API/JSContextRef.cpp: 4112 * GNUmakefile.am: 4113 * JavaScriptCore.exp: 4114 * JavaScriptCore.pro: 4115 * JavaScriptCore.xcodeproj/project.pbxproj: 4116 * bytecode/EvalCodeCache.h: 4117 * bytecode/JumpTable.cpp: 4118 * profiler/ProfileNode.cpp: 4119 * runtime/Identifier.cpp: 4120 (JSC::Identifier::add): 4121 * runtime/Identifier.h: 4122 (JSC::Identifier::equal): 4123 * runtime/UString.cpp: 4124 * runtime/UString.h: 4125 (WTF::): 4126 * runtime/UStringImpl.cpp: Removed. 4127 * runtime/UStringImpl.h: 4128 * wtf/text/StringHash.h: 4129 (WebCore::StringHash::equal): 4130 (WebCore::CaseFoldingHash::equal): 4131 * wtf/text/StringImpl.cpp: 4132 (WebCore::StringImpl::~StringImpl): 4133 (WebCore::StringImpl::empty): 4134 (WebCore::StringImpl::sharedBuffer): 4135 (WebCore::equal): 4136 * wtf/text/StringImpl.h: 4137 (WebCore::StringImpl::StringImpl): 4138 (WebCore::StringImpl::create): 4139 (WebCore::StringImpl::tryCreateUninitialized): 4140 (WebCore::StringImpl::cost): 4141 (WebCore::StringImpl::isIdentifier): 4142 (WebCore::StringImpl::setIsIdentifier): 4143 (WebCore::StringImpl::computeHash): 4144 (WebCore::StringImpl::copyChars): 4145 (WebCore::StringImpl::): 4146 41472010-04-21 Patrick Gansterer <paroga@paroga.com> 4148 4149 Reviewed by Darin Adler. 4150 4151 Added missing #include "Lookup.h" in LUT source files. 4152 https://bugs.webkit.org/show_bug.cgi?id=37903 4153 4154 * runtime/ArrayPrototype.cpp: 4155 * runtime/DatePrototype.cpp: 4156 * runtime/JSONObject.cpp: 4157 * runtime/MathObject.cpp: 4158 * runtime/NumberConstructor.cpp: 4159 * runtime/RegExpConstructor.cpp: 4160 * runtime/RegExpObject.cpp: 4161 * runtime/StringPrototype.cpp: 4162 41632010-04-21 Gustavo Sverzut Barbieri <barbieri@profusion.mobi> 4164 4165 Reviewed by Nikolas Zimmermann. 4166 4167 Add missing EFL JavaScriptCore file. 4168 http://webkit.org/b/37854 4169 4170 * wtf/efl: Added. 4171 * wtf/efl/MainThreadEfl.cpp: Added. 4172 (WTF::initializeMainThreadPlatform): 4173 (WTF::timeoutFired): 4174 (WTF::scheduleDispatchFunctionsOnMainThread): 4175 41762010-04-20 Xan Lopez <xlopez@igalia.com> 4177 4178 Another attempt to fix the build. 4179 4180 * GNUmakefile.am: 4181 41822010-04-20 Oliver Hunt <oliver@apple.com> 4183 4184 Reviewed by Maciej Stachowiak. 4185 4186 [ES5] RegExp literals are constants that should be persistent across multiple function calls. 4187 https://bugs.webkit.org/show_bug.cgi?id=37908 4188 4189 Dump the separate RegExp constant pool, and just use the standard JS constant pool 4190 in codeblock. This allows us to drop op_new_regexp and all associated code as well. 4191 4192 * bytecode/CodeBlock.cpp: 4193 (JSC::CodeBlock::dump): 4194 (JSC::CodeBlock::shrinkToFit): 4195 * bytecode/CodeBlock.h: 4196 * bytecode/Opcode.h: 4197 * bytecompiler/BytecodeGenerator.cpp: 4198 (JSC::BytecodeGenerator::emitLoad): 4199 * bytecompiler/BytecodeGenerator.h: 4200 * bytecompiler/NodesCodegen.cpp: 4201 (JSC::RegExpNode::emitBytecode): 4202 * interpreter/Interpreter.cpp: 4203 (JSC::Interpreter::privateExecute): 4204 * jit/JIT.cpp: 4205 (JSC::JIT::privateCompileMainPass): 4206 * jit/JIT.h: 4207 * jit/JITOpcodes.cpp: 4208 * jit/JITStubs.cpp: 4209 * jit/JITStubs.h: 4210 (JSC::): 4211 42122010-04-20 Oliver Hunt <oliver@apple.com> 4213 4214 Fix license on create_regex_tables 4215 4216 * create_regex_tables: 4217 42182010-04-20 Oliver Hunt <oliver@apple.com> 4219 4220 Reviewed by NOBODY (Build fix). 4221 4222 Fix gtk 4223 4224 * GNUmakefile.am: 4225 * make-generated-sources.sh: 4226 42272010-04-20 Gavin Barraclough <barraclough@apple.com> 4228 4229 Reviewed by Oliver Hunt. 4230 4231 Bug 37895 - Share common code from UStringImplBase with StringImpl 4232 4233 The implementation of StringImpl & UStringImpl is very similar. Restructure 4234 StringImpl to match UStringImpl, moving the flags and length into a base class, 4235 so that this can be shared between both string types to increase code reuse. 4236 4237 * JavaScriptCore.xcodeproj/project.pbxproj: 4238 * runtime/RopeImpl.h: 4239 (JSC::RopeImpl::RopeImpl): 4240 * runtime/UStringImpl.h: 4241 (JSC::UStringImpl::UStringImpl): 4242 * wtf/text/StringImpl.h: 4243 (WebCore::StringImpl::StringImpl): 4244 (WebCore::StringImpl::characters): 4245 * wtf/text/StringImplBase.h: Copied from JavaScriptCore/runtime/UStringImpl.h. 4246 (WTF::StringImplBase::length): 4247 (WTF::StringImplBase::operator new): 4248 (WTF::StringImplBase::StringImplBase): 4249 42502010-04-20 Oliver Hunt <oliver@apple.com> 4251 4252 Reviewed by Gavin Barraclough. 4253 4254 Autogenerate yarr character tables 4255 https://bugs.webkit.org/show_bug.cgi?id=37877 4256 4257 Use a python script to automatically generate character tables 4258 for the builtin YARR character classes. This allows us to generate 4259 actual tables as well, by using these tables we can both increase 4260 performance of the check (for complex builtins) and reduce the actual 4261 code size. 4262 4263 4-8% win on string-unpack-code, but lots of noise on other tests so 4264 i'm only confident saying its a 1% win overall. 4265 4266 * DerivedSources.make: 4267 * JavaScriptCore.xcodeproj/project.pbxproj: 4268 * assembler/AbstractMacroAssembler.h: 4269 (JSC::AbstractMacroAssembler::ExtendedAddress::ExtendedAddress): 4270 * assembler/MacroAssembler.h: 4271 (JSC::MacroAssembler::branchTest8): 4272 * assembler/MacroAssemblerX86Common.h: 4273 (JSC::MacroAssemblerX86Common::branchTest8): 4274 * assembler/MacroAssemblerX86_64.h: 4275 (JSC::MacroAssemblerX86_64::branchTest8): 4276 * assembler/X86Assembler.h: 4277 (JSC::X86Assembler::cmpb_im): 4278 (JSC::X86Assembler::testb_im): 4279 * bytecode/SamplingTool.cpp: 4280 (JSC::SamplingTool::dump): 4281 * create_regex_tables: Added. 4282 * yarr/RegexCompiler.cpp: 4283 (JSC::Yarr::CharacterClassConstructor::charClass): 4284 * yarr/RegexJIT.cpp: 4285 (JSC::Yarr::RegexGenerator::matchCharacterClass): 4286 (JSC::Yarr::RegexGenerator::generatePatternCharacterGreedy): 4287 (JSC::Yarr::RegexGenerator::generatePatternCharacterNonGreedy): 4288 (JSC::Yarr::RegexGenerator::generateCharacterClassGreedy): 4289 * yarr/RegexPattern.h: 4290 (JSC::Yarr::CharacterClassTable::create): 4291 (JSC::Yarr::CharacterClassTable::CharacterClassTable): 4292 (JSC::Yarr::CharacterClass::CharacterClass): 4293 42942010-04-20 Gavin Barraclough <barraclough@apple.com> 4295 4296 Reviewed by NOBODY (speculative windows fix - missed a bit!). 4297 4298 * wtf/text/AtomicString.h: 4299 43002010-04-20 Gavin Barraclough <barraclough@apple.com> 4301 4302 Reviewed by NOBODY (speculative windows fix). 4303 4304 * wtf/text/AtomicString.h: 4305 43062010-04-20 Gavin Barraclough <barraclough@apple.com> 4307 4308 Reviewed by NOBODY (windows build fix). 4309 4310 Add missing .def file entries. 4311 4312 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 4313 43142010-04-20 Gavin Barraclough <barraclough@apple.com> 4315 4316 Reviewed by Geoff Garen. 4317 4318 Bug 37869 - Move URopeImpl to its own .h/.cpp 4319 4320 Currently Ropes are implemented by the class URopeImpl, which is defined in 4321 UStringImpl.h, and then typedefed to the name JSString::Rope. Remove the 4322 typedef, and rename all uses of URopeImpl and JSString::Rope to just RopeImpl. 4323 4324 Move RopeImpl to its own header, and remove all remaining references to ropes 4325 from UStringImpl (rename UStringOrRopeImpl to UStringImplBase, rename or move 4326 the isRope & deref methods from UStringOrRopeImpl). 4327 4328 * JavaScriptCore.xcodeproj/project.pbxproj: 4329 * runtime/JSString.cpp: 4330 (JSC::JSString::resolveRope): 4331 * runtime/JSString.h: 4332 (JSC::): 4333 (JSC::RopeBuilder::JSString): 4334 (JSC::RopeBuilder::~JSString): 4335 (JSC::RopeBuilder::appendStringInConstruct): 4336 (JSC::RopeBuilder::JSStringFinalizerStruct::): 4337 * runtime/RopeImpl.cpp: Copied from JavaScriptCore/runtime/UStringImpl.cpp. 4338 (JSC::RopeImpl::derefFibersNonRecursive): 4339 (JSC::RopeImpl::destructNonRecursive): 4340 * runtime/RopeImpl.h: Copied from JavaScriptCore/runtime/UStringImpl.h. 4341 (JSC::RopeImpl::tryCreateUninitialized): 4342 (JSC::RopeImpl::isRope): 4343 (JSC::RopeImpl::deref): 4344 (JSC::RopeImpl::RopeImpl): 4345 * runtime/UStringImpl.cpp: 4346 * runtime/UStringImpl.h: 4347 (JSC::UStringImplBase::isInvalid): 4348 (JSC::UStringImplBase::ref): 4349 (JSC::UStringImplBase::UStringImplBase): 4350 (JSC::UStringImplBase::): 4351 (JSC::UStringImpl::UStringImpl): 4352 43532010-04-20 Gavin Barraclough <barraclough@apple.com> 4354 4355 Reviewed by Geoff Garen. 4356 4357 Bug 37828 - Move WebCore's String classes to WTF 4358 4359 Move these classes up to WTF so they are available to all clients of WTF (in 4360 particular JSC). 4361 4362 As a first patch, making the most minimal change possible, since this patch 4363 could easily grow rather large since we'll have to change every class forward 4364 declaration ( e.g. every "namespace WebCore { class String; }" much change to 4365 "namespace WTF { class String; }"). 4366 4367 Moving the files, but leaving the classes logically in the WebCore namespace - 4368 which is technically a layering violation - I'll come back and fix this up in a 4369 subsequent patch. 4370 4371 * Android.mk: 4372 * Android.v8.wtf.mk: 4373 * GNUmakefile.am: 4374 * JavaScriptCore.exp: 4375 * JavaScriptCore.gypi: 4376 * JavaScriptCore.pro: 4377 * JavaScriptCore.vcproj/WTF/WTF.vcproj: 4378 * JavaScriptCore.xcodeproj/project.pbxproj: 4379 * config.h: 4380 * wtf/StaticConstructors.h: Copied from WebCore/platform/StaticConstructors.h. 4381 * wtf/text/AtomicString.cpp: Copied from WebCore/platform/text/AtomicString.cpp. 4382 * wtf/text/AtomicString.h: Copied from WebCore/platform/text/AtomicString.h. 4383 * wtf/text/AtomicStringImpl.h: Copied from WebCore/platform/text/AtomicStringImpl.h. 4384 * wtf/text/StringBuffer.h: Copied from WebCore/platform/text/StringBuffer.h. 4385 * wtf/text/StringHash.h: Copied from WebCore/platform/text/StringHash.h. 4386 * wtf/text/StringImpl.cpp: Copied from WebCore/platform/text/StringImpl.cpp. 4387 * wtf/text/StringImpl.h: Copied from WebCore/platform/text/StringImpl.h. 4388 * wtf/text/WTFString.cpp: Copied from WebCore/platform/text/String.cpp. 4389 (WebCore::charactersToFloat): 4390 * wtf/text/WTFString.h: Copied from WebCore/platform/text/PlatformString.h. 4391 43922010-04-20 Csaba Osztrogonác <ossy@webkit.org> 4393 4394 [Qt] Unreviewed speculative buildfix for WinCE after r57882 4395 https://bugs.webkit.org/show_bug.cgi?id=37701 4396 4397 * JavaScriptCore.pri: missing wince* case added. 4398 43992010-04-20 Gavin Barraclough <barraclough@apple.com> 4400 4401 Reviewed by NOBODY (build fix). 4402 Speculative Chromium/Win build fix, attempt #2. 4403 4404 * config.h: 4405 44062010-04-20 Gavin Barraclough <barraclough@apple.com> 4407 4408 Reviewed by NOBODY (build fix). 4409 Speculative Chromium/Win build fix. 4410 4411 * config.h: JS_EXPORTDATA should do nothing on !JSC builds. 4412 44132010-04-20 Csaba Osztrogonác <ossy@webkit.org> 4414 4415 Reviewed by Simon Hausmann. 4416 4417 [Qt] Target(WebCore,jsc,...) must depends on static library of JavaScriptCore 4418 https://bugs.webkit.org/show_bug.cgi?id=37701 4419 4420 * JavaScriptCore.pri: dependency added. 4421 44222010-04-20 Gavin Barraclough <barraclough@apple.com> 4423 4424 Rubber stamped by Maciej Stachowiak (relanding r57829). 4425 Added missing JS_EXPORTDATA 4426 4427 * API/APIShims.h: 4428 (JSC::APIEntryShimWithoutLock::APIEntryShimWithoutLock): 4429 (JSC::APIEntryShimWithoutLock::~APIEntryShimWithoutLock): 4430 (JSC::APICallbackShim::APICallbackShim): 4431 (JSC::APICallbackShim::~APICallbackShim): 4432 * API/JSContextRef.cpp: 4433 * Android.mk: 4434 * Android.v8.wtf.mk: 4435 * GNUmakefile.am: 4436 * JavaScriptCore.exp: 4437 * JavaScriptCore.gypi: 4438 * JavaScriptCore.pro: 4439 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 4440 * JavaScriptCore.vcproj/WTF/WTF.vcproj: 4441 * JavaScriptCore.xcodeproj/project.pbxproj: 4442 * runtime/Completion.cpp: 4443 (JSC::checkSyntax): 4444 (JSC::evaluate): 4445 * runtime/Identifier.cpp: 4446 (JSC::Identifier::remove): 4447 (JSC::Identifier::checkCurrentIdentifierTable): 4448 * runtime/Identifier.h: 4449 * runtime/InitializeThreading.cpp: 4450 (JSC::initializeThreadingOnce): 4451 * runtime/JSGlobalData.cpp: 4452 (JSC::JSGlobalData::create): 4453 * wtf/WTFThreadData.cpp: Copied from JavaScriptCore/wtf/WTFThreadData.cpp. 4454 * wtf/WTFThreadData.h: Copied from JavaScriptCore/wtf/WTFThreadData.h. 4455 44562010-04-19 Gavin Barraclough <barraclough@apple.com> 4457 4458 Reviewed by NOBODY (rolling out r57829). 4459 This broke windows. 4460 4461 * API/APIShims.h: 4462 (JSC::APIEntryShimWithoutLock::APIEntryShimWithoutLock): 4463 (JSC::APIEntryShimWithoutLock::~APIEntryShimWithoutLock): 4464 (JSC::APICallbackShim::APICallbackShim): 4465 (JSC::APICallbackShim::~APICallbackShim): 4466 * API/JSContextRef.cpp: 4467 * Android.mk: 4468 * Android.v8.wtf.mk: 4469 * GNUmakefile.am: 4470 * JavaScriptCore.exp: 4471 * JavaScriptCore.gypi: 4472 * JavaScriptCore.pro: 4473 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 4474 * JavaScriptCore.vcproj/WTF/WTF.vcproj: 4475 * JavaScriptCore.xcodeproj/project.pbxproj: 4476 * runtime/Completion.cpp: 4477 (JSC::checkSyntax): 4478 (JSC::evaluate): 4479 * runtime/Identifier.cpp: 4480 (JSC::Identifier::remove): 4481 (JSC::Identifier::checkCurrentIdentifierTable): 4482 (JSC::createIdentifierTableSpecificCallback): 4483 (JSC::createIdentifierTableSpecific): 4484 * runtime/Identifier.h: 4485 (JSC::ThreadIdentifierTableData::ThreadIdentifierTableData): 4486 (JSC::defaultIdentifierTable): 4487 (JSC::setDefaultIdentifierTable): 4488 (JSC::currentIdentifierTable): 4489 (JSC::setCurrentIdentifierTable): 4490 (JSC::resetCurrentIdentifierTable): 4491 * runtime/InitializeThreading.cpp: 4492 (JSC::initializeThreadingOnce): 4493 * runtime/JSGlobalData.cpp: 4494 (JSC::JSGlobalData::create): 4495 * wtf/WTFThreadData.cpp: Removed. 4496 * wtf/WTFThreadData.h: Removed. 4497 44982010-04-19 Douglas Gregor <dgregor@apple.com> 4499 4500 Reviewed and landed by Anders Carlsson. 4501 4502 * runtime/UStringImpl.h: 4503 Fix class/struct declaration mismatches. 4504 45052010-04-19 Geoffrey Garen <ggaren@apple.com> 4506 4507 Reviewed by Beth Dakin. 4508 4509 Checked in these tests I wrote becuase Balazs Kelemen wanted to use them. 4510 4511 * tests/perf: Added. 4512 * tests/perf/bench-allocate-nonretained.js: Added. 4513 * tests/perf/bench-allocate-retained.js: Added. 4514 45152010-04-19 Gavin Barraclough <barraclough@apple.com> 4516 4517 Reviewed by NOBODY (windows build fix). 4518 4519 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 4520 45212010-04-16 Gavin Barraclough <barraclough@apple.com> 4522 4523 Reviewed by Sam Weinig. 4524 4525 https://bugs.webkit.org/show_bug.cgi?id=37745 4526 Move string uniquing tables to (new) WTFThreadData class. 4527 4528 Remove AtomicString's dependency on ThreadGlobalData so that we can move 4529 WebCore's string classes up to WTF. 4530 4531 WTFThreadData.cpp/.h are based on ThreadGlobalData from WebCore. 4532 Moved JSC & WebCore's string uniquing tables to this class. 4533 4534 This patch introduces a temporary layering violation in providing forward 4535 declarations of classes from JSC and WTF; this will be resolved as we move 4536 more string code up to WTF. 4537 4538 * API/APIShims.h: 4539 (JSC::APIEntryShimWithoutLock::APIEntryShimWithoutLock): 4540 (JSC::APIEntryShimWithoutLock::~APIEntryShimWithoutLock): 4541 (JSC::APICallbackShim::APICallbackShim): 4542 (JSC::APICallbackShim::~APICallbackShim): 4543 * API/JSContextRef.cpp: 4544 * JavaScriptCore.exp: 4545 * JavaScriptCore.xcodeproj/project.pbxproj: 4546 * runtime/Completion.cpp: 4547 (JSC::checkSyntax): 4548 (JSC::evaluate): 4549 * runtime/Identifier.cpp: 4550 (JSC::Identifier::remove): 4551 (JSC::Identifier::checkCurrentIdentifierTable): 4552 * runtime/Identifier.h: 4553 * runtime/InitializeThreading.cpp: 4554 (JSC::initializeThreadingOnce): 4555 * runtime/JSGlobalData.cpp: 4556 (JSC::JSGlobalData::create): 4557 * wtf/WTFThreadData.cpp: Copied from WebCore/platform/ThreadGlobalData.cpp. 4558 (WTF::WTFThreadData::WTFThreadData): 4559 (WTF::WTFThreadData::~WTFThreadData): 4560 * wtf/WTFThreadData.h: Copied from WebCore/platform/ThreadGlobalData.h. 4561 (WTF::WTFThreadData::atomicStringTable): 4562 (WTF::WTFThreadData::initializeIdentifierTable): 4563 (WTF::WTFThreadData::currentIdentifierTable): 4564 (WTF::WTFThreadData::setCurrentIdentifierTable): 4565 (WTF::WTFThreadData::resetCurrentIdentifierTable): 4566 (WTF::wtfThreadData): 4567 45682010-04-19 Jocelyn Turcotte <jocelyn.turcotte@nokia.com> 4569 4570 Reviewed by Simon Hausmann. 4571 4572 [Qt] Build fix for WinCE. 4573 4574 Moved the include of the non-existing errno.h header file inside 4575 platform guard macros. 4576 4577 * jit/ExecutableAllocatorFixedVMPool.cpp: 4578 45792010-04-18 Kwang Yul Seo <skyul@company100.net> 4580 4581 Reviewed by Laszlo Gombos. 4582 4583 [WINCE] Don't define WTF_CPU_MIDDLE_ENDIAN=1 4584 https://bugs.webkit.org/show_bug.cgi?id=37434 4585 4586 Windows CE supports little-endian format only, so don't define 4587 WTF_CPU_MIDDLE_ENDIAN=1. 4588 4589 * wtf/Platform.h: 4590 45912010-04-18 Simon Hausmann <simon.hausmann@nokia.com> 4592 4593 Reviewed by Laszlo Gombos. 4594 4595 [Qt] Fix JavaScriptCore's include path for WinCE builds 4596 4597 https://bugs.webkit.org/show_bug.cgi?id=36751 4598 4599 * JavaScriptCore.pri: 4600 46012010-04-16 Gavin Barraclough <barraclough@apple.com> 4602 4603 Reviewed by nobody, build fix. 4604 46052010-04-16 Gavin Barraclough <barraclough@apple.com> 4606 4607 Reviewed by nobody, build fix. 4608 46092010-04-16 Gavin Barraclough <barraclough@apple.com> 4610 4611 Reviewed by Oliver Hunt. 4612 4613 Bug 37730 - Remove JSC::UString dependencies from WebCore::StringImpl 4614 (Following on from bug #37675). 4615 4616 Make the argument ordering for UStringImpl's constructor & create 4617 methods match, when passed a shared buffer. 4618 4619 * JavaScriptCore.exp: 4620 * runtime/UStringImpl.cpp: 4621 (JSC::UStringImpl::create): 4622 * runtime/UStringImpl.h: 4623 46242010-04-15 Jedrzej Nowacki <jedrzej.nowacki@nokia.com> 4625 4626 Reviewed by Kenneth Rohde Christiansen. 4627 4628 Fix memory leak in QScriptEngine::evaluate(). 4629 4630 QScriptEnginePrivate::evaluate should release temporary variables. 4631 4632 [Qt] QScriptEngine::evaluate has memory leak. 4633 https://bugs.webkit.org/show_bug.cgi?id=37596 4634 4635 * qt/api/qscriptengine_p.cpp: 4636 (QScriptEnginePrivate::evaluate): 4637 * qt/api/qscriptengine_p.h: 4638 46392010-04-14 Jedrzej Nowacki <jedrzej.nowacki@nokia.com> 4640 4641 Reviewed by Kenneth Rohde Christiansen. 4642 4643 Fix a memory leak in QScriptValue::inherits. 4644 4645 [Qt] QScriptValue::inherits has a memory leak. 4646 https://bugs.webkit.org/show_bug.cgi?id=37617 4647 4648 * qt/api/qscriptvalue_p.h: 4649 (QScriptValuePrivate::inherits): 4650 46512010-04-14 Jedrzej Nowacki <jedrzej.nowacki@nokia.com> 4652 4653 Reviewed by Kenneth Rohde Christiansen. 4654 4655 Fix a few memory leaks in QScriptEngine. 4656 4657 Syntax checking caused memory leak, not all temporary variables were released. 4658 4659 [Qt] Syntax checking in the QtScript cause a memory leak. 4660 https://bugs.webkit.org/show_bug.cgi?id=37610 4661 4662 * qt/api/qscriptengine_p.cpp: 4663 (QScriptEnginePrivate::checkSyntax): 4664 * qt/api/qscriptsyntaxcheckresult.cpp: 4665 (QScriptSyntaxCheckResultPrivate::errorMessage): 4666 (QScriptSyntaxCheckResultPrivate::errorLineNumber): 4667 46682010-04-14 Jedrzej Nowacki <jedrzej.nowacki@nokia.com> 4669 4670 Reviewed by Kenneth Rohde Christiansen. 4671 4672 Fix memory leak inside QScriptEnginePrivate::makeJSValue. 4673 4674 QScriptEnginePrivate::makeJSValue should release temporary JSStringRef variable. 4675 4676 [Qt] tst_QScriptValue::toString has a memory leak. 4677 https://bugs.webkit.org/show_bug.cgi?id=37598 4678 4679 * qt/api/qscriptengine_p.h: 4680 (QScriptEnginePrivate::makeJSValue): 4681 46822010-04-14 Peter Varga <pvarga@inf.u-szeged.hu> 4683 4684 Reviewed by Geoffrey Garen. 4685 4686 Move the YARR JIT fallback detection from RegexJIT.cpp to 4687 RegexCompiler.cpp. 4688 4689 https://bugs.webkit.org/show_bug.cgi?id=37571 4690 4691 * yarr/RegexCompiler.cpp: 4692 (JSC::Yarr::RegexPatternConstructor::atomBackReference): 4693 (JSC::Yarr::RegexPatternConstructor::quantifyAtom): 4694 * yarr/RegexJIT.cpp: 4695 (JSC::Yarr::RegexGenerator::generateTerm): 4696 (JSC::Yarr::RegexGenerator::RegexGenerator): 4697 (JSC::Yarr::jitCompileRegex): 4698 * yarr/RegexJIT.h: 4699 (JSC::Yarr::RegexCodeBlock::operator!): 4700 * yarr/RegexPattern.h: 4701 (JSC::Yarr::RegexPattern::RegexPattern): 4702 (JSC::Yarr::RegexPattern::reset): 4703 47042010-04-14 Kent Hansen <kent.hansen@nokia.com> 4705 4706 Reviewed by Maciej Stachowiak. 4707 4708 Mac OS X: Use deployment target to determine whether memory tagging should be enabled 4709 https://bugs.webkit.org/show_bug.cgi?id=34888 4710 4711 When building on (Snow) Leopard but targeting Tiger 4712 (TARGETING_TIGER defined, BUILDING_ON_TIGER not defined), 4713 WebKit would crash on Tiger because the tags passed to mmap 4714 caused those function calls to fail. 4715 4716 Conversely, when building on Tiger but targeting Leopard 4717 (BUILDING_ON_TIGER defined, TARGETING_LEOPARD defined), WebKit 4718 would crash on Leopard because the tags passed to vm_map and 4719 vm_allocate caused those function calls to fail. 4720 4721 Solution: Use TARGETING_TIGER rather than BUILDING_ON_TIGER to 4722 govern the tag definitions. Use the same tags for vm_map and 4723 vm_allocate regardless of target, since they work on 4724 both. Fall back to the mmap tags that work on Tiger (that is, 4725 "no tags") if targeting Tiger, since those tags also work on 4726 Leopard. 4727 4728 * wtf/VMTags.h: 4729 47302010-04-12 Jocelyn Turcotte <jocelyn.turcotte@nokia.com> 4731 4732 Reviewed by nobody, build fix. 4733 4734 [Qt] Build fix for Mac when building with build-webkit --qt 4735 4736 Specifying no configuration on Mac builds WebCore both in debug 4737 and release. JavaScriptCore has to follow this rule as well. 4738 4739 * JavaScriptCore.pro: 4740 47412010-04-11 Mark Rowe <mrowe@apple.com> 4742 4743 Reviewed by Sam Weinig. 4744 4745 <rdar://problem/7851332> Fix the build. 4746 4747 * wtf/FastMalloc.cpp: 4748 (WTF::TCMallocStats::): Initialize extra members of malloc_introspection_t to zero. 4749 47502010-04-09 Mikhail Naganov <mnaganov@chromium.org> 4751 4752 Reviewed by Pavel Feldman. 4753 4754 Make CallIdentifier constructor to handle null urls. 4755 4756 https://bugs.webkit.org/show_bug.cgi?id=37341 4757 4758 * profiler/CallIdentifier.h: 4759 (JSC::CallIdentifier::CallIdentifier): 4760 47612010-04-09 Jocelyn Turcotte <jocelyn.turcotte@nokia.com> 4762 4763 Reviewed by Simon Hausmann. 4764 4765 [Qt] Fix crashes with package builds in release 4766 4767 Add NDEBUG to the defines for package build in JavaScriptCore.pri, 4768 so that it's consistently used for JavaScriptCore, WebCore, jsc and all 4769 other tools using wtf, etc. data structures directly. Mixing NDEBUG with 4770 non-NDEBUG builds causes crashes due to differences in data structures when 4771 assertions/checks are enabled. 4772 4773 * JavaScriptCore.pri: 4774 47752010-04-09 Patrick Gansterer <paroga@paroga.com> 4776 4777 Reviewed by Darin Adler. 4778 4779 Implement NO_RETURN for COMPILER(MSVC). 4780 https://bugs.webkit.org/show_bug.cgi?id=33056 4781 4782 Added NO_RETURN_WITH_VALUE for functions with non-void return type. 4783 4784 * jsc.cpp: 4785 * wtf/AlwaysInline.h: 4786 * wtf/FastMalloc.cpp: 4787 47882010-04-08 Kwang Yul Seo <skyul@company100.net> 4789 4790 Reviewed by Simon Hausmann. 4791 4792 [WINCE] Check if ARM or _ARM_ is defined 4793 https://bugs.webkit.org/show_bug.cgi?id=37200 4794 4795 MSVC defines ARM and _ARM_ for Windows CE ARM. Define WTF_CPU_ARM=1 4796 when either ARM or _ARM_ is defined. 4797 4798 * wtf/Platform.h: 4799 48002010-04-08 Csaba Osztrogonác <ossy@webkit.org> 4801 4802 Reviewed Oliver Hunt. 4803 4804 [Qt]r57240 broke Qt build (might be a gcc bug) 4805 https://bugs.webkit.org/show_bug.cgi?id=37253 4806 4807 Workaround until fix. On PLATFORM(QT) use inline instead of ALWAYS_INLINE. 4808 4809 * wtf/PassRefPtr.h: Qt guards added. 4810 48112010-04-07 Oliver Hunt <oliver@apple.com> 4812 4813 Reviewed by Anders Carlsson. 4814 4815 Vector<UString> makes many needless calls to UString::UString and UString::~UString 4816 4817 Add a VectorTrait<UString> specialisation to allow vector to simply memset/memcpy 4818 data around. Only difference from the VectorTrait<RefPtr<T> > traits is the inability 4819 to use memset to initialize data. 4820 4821 * runtime/UString.h: 4822 (WTF::): 4823 48242010-04-07 Oliver Hunt <oliver@apple.com> 4825 4826 Reviewed by Geoff Garen. 4827 4828 Beat gcc with a clue bat -- force inlining of refIfNotNull and derefIfNotNull 4829 4830 * wtf/PassRefPtr.h: 4831 48322010-04-07 Kwang Yul Seo <skyul@company100.net> 4833 4834 Reviewed by Darin Adler. 4835 4836 Replace isprint with isASCIIPrintable 4837 https://bugs.webkit.org/show_bug.cgi?id=37223 4838 4839 WebKit does not use functions in <ctype.h> as they are dependent on the current 4840 locale. Use the equivalent functions in <wtf/ASCIICType.h>. isASCIIPrintable 4841 replaces isprint. 4842 4843 * pcre/pcre_exec.cpp: 4844 (pchars): 4845 48462010-04-07 Enrica Casucci <enrica@apple.com> 4847 4848 Reviewed by Darin Adler. 4849 4850 https://bugs.webkit.org/show_bug.cgi?id=37219 4851 4852 This change disables text caret for the iPhone platflorm. 4853 4854 * wtf/Platform.h: Disabled text caret for iPhone. 4855 48562010-04-06 Adam Barth <abarth@webkit.org> 4857 4858 Reviewed by Eric Seidel. 4859 4860 REGRESSION: Worker termination via JS timeout may cause worker tests like fast/workers/worker-terminate.html fail. 4861 https://bugs.webkit.org/show_bug.cgi?id=36646 4862 4863 Add a new exception type for forcibly terminating a JavaScript stack. 4864 The new exception functions similarly to the 4865 InterruptedExecutionException but is conceptually different because 4866 execution is terminated instead of just interrupted. 4867 4868 * GNUmakefile.am: 4869 - Added new Terminator.h file. 4870 * JavaScriptCore.gypi: 4871 - Added new Terminator.h file. 4872 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: 4873 - Added new Terminator.h file. 4874 * JavaScriptCore.xcodeproj/project.pbxproj: 4875 - Added new Terminator.h file. 4876 * interpreter/Interpreter.cpp: 4877 (JSC::Interpreter::throwException): 4878 - Fully unwind the stack for TerminatedExecutionException. 4879 (JSC::Interpreter::privateExecute): 4880 - Check if we've been terminated at the same time we check if we've 4881 timed out. 4882 * jit/JITStubs.cpp: 4883 (JSC::DEFINE_STUB_FUNCTION): 4884 - Check if we've been terminated at the same time we check if we've 4885 timed out. 4886 * runtime/Completion.cpp: 4887 - Some exceptions define special completion types so that calls can 4888 see why we terminated evaluation. 4889 (JSC::evaluate): 4890 * runtime/Completion.h: 4891 - Define a new completion type for termination. 4892 (JSC::): 4893 * runtime/ExceptionHelpers.cpp: 4894 - Define TerminatedExecutionException and refactor pseudo-RTTI 4895 virtual function to be more semantic. 4896 (JSC::InterruptedExecutionError::exceptionType): 4897 (JSC::TerminatedExecutionError::TerminatedExecutionError): 4898 (JSC::TerminatedExecutionError::exceptionType): 4899 (JSC::TerminatedExecutionError::toString): 4900 (JSC::createTerminatedExecutionException): 4901 * runtime/ExceptionHelpers.h: 4902 - Entry point for generating a TerminatedExecutionException. 4903 * runtime/JSGlobalData.cpp: 4904 (JSC::JSGlobalData::JSGlobalData): 4905 - Add a Terminator object that can be used to asynchronously 4906 terminate a JavaScript execution stack. 4907 * runtime/JSGlobalData.h: 4908 * runtime/JSObject.h: 4909 (JSC::JSObject::exceptionType): 4910 - Define that, by default, thrown objects have a normal exception 4911 type. 4912 * runtime/Terminator.h: Added. 4913 - Added a new controller object that can be used to terminate 4914 execution asynchronously. This object is more or less a 4915 glorified bool. 4916 (JSC::Terminator::Terminator): 4917 (JSC::Terminator::termianteSoon): 4918 (JSC::Terminator::shouldTerminate): 4919 49202010-04-05 Oliver Hunt <oliver@apple.com> 4921 4922 And another one. 4923 4924 * JavaScriptCore.vcproj/jsc/jscCommon.vsprops: 4925 49262010-04-05 Oliver Hunt <oliver@apple.com> 4927 4928 And another build fix. 4929 4930 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCoreGenerated.make: 4931 49322010-04-05 Oliver Hunt <oliver@apple.com> 4933 4934 Build fix 4935 4936 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: 4937 49382010-04-05 Oliver Hunt <oliver@apple.com> 4939 4940 Reviewed by Gavin Barraclough. 4941 4942 Support weak maps in JSC 4943 https://bugs.webkit.org/show_bug.cgi?id=37132 4944 4945 Expose an API to allow creation of a map for storing 4946 weak JS references. 4947 4948 * API/JSWeakObjectMapRefInternal.h: Added. 4949 (OpaqueJSWeakObjectMap::create): 4950 (OpaqueJSWeakObjectMap::map): 4951 (OpaqueJSWeakObjectMap::~OpaqueJSWeakObjectMap): 4952 (OpaqueJSWeakObjectMap::OpaqueJSWeakObjectMap): 4953 * API/JSWeakObjectMapRefPrivate.cpp: Added. 4954 * API/JSWeakObjectMapRefPrivate.h: Added. 4955 * JavaScriptCore.exp: 4956 * JavaScriptCore.xcodeproj/project.pbxproj: 4957 * runtime/JSGlobalObject.h: 4958 (JSC::JSGlobalObject::registerWeakMap): 4959 (JSC::JSGlobalObject::deregisterWeakMap): 4960 49612010-04-05 Laszlo Gombos <laszlo.1.gombos@nokia.com> 4962 4963 Reviewed by Kenneth Rohde Christiansen. 4964 4965 [Symbian] Consolidate Symbian WINSCW environment configuration 4966 https://bugs.webkit.org/show_bug.cgi?id=37100 4967 4968 Move the "undefinition" of WIN32 and _WIN32 from WebCore/config.h 4969 to JavaScriptCore/wtf/Platform.h as it is not specific to WebCore. 4970 4971 PLATFORM(WIN) and OS(WIN) no longer needs to be undefined as 4972 undefining WIN32 takes care of it. 4973 4974 * wtf/Platform.h: 4975 49762010-04-03 Gavin Barraclough <barraclough@apple.com> 4977 4978 Reviewed by Oliver Hunt. 4979 4980 https://bugs.webkit.org/show_bug.cgi?id=37068 4981 Change UString to use a 0 rep for null strings instead of a null object. 4982 4983 No performance impact. 4984 4985 * JavaScriptCore.exp: 4986 * runtime/InternalFunction.cpp: 4987 (JSC::InternalFunction::InternalFunction): 4988 * runtime/JSString.h: 4989 (JSC::RopeBuilder::JSString): 4990 * runtime/UString.cpp: 4991 (JSC::initializeUString): 4992 * runtime/UString.h: 4993 (JSC::UString::UString): 4994 (JSC::UString::data): 4995 (JSC::UString::size): 4996 (JSC::UString::isNull): 4997 (JSC::UString::isEmpty): 4998 (JSC::UString::cost): 4999 50002010-04-03 Balazs Kelemen <kb@inf.u-szeged.hu> 5001 5002 Reviewed by Oliver Hunt. 5003 5004 Fix uninitalised members in CallLinkInfo and BytecodeGenerator. 5005 5006 https://bugs.webkit.org/show_bug.cgi?id=36816 5007 5008 * bytecode/CodeBlock.cpp: 5009 (JSC::CodeBlock::CodeBlock): 5010 * bytecode/CodeBlock.h: 5011 (JSC::CallLinkInfo::CallLinkInfo): 5012 50132010-04-03 yael aharon <yael.aharon@nokia.com> 5014 5015 Reviewed by Darin Adler. 5016 5017 Enable HTMLProgressElement for Safari on OSX 5018 https://bugs.webkit.org/show_bug.cgi?id=36961 5019 5020 * Configurations/FeatureDefines.xcconfig: 5021 50222010-04-02 Ruben Van Boxem <vanboxem.ruben@gmail.com> 5023 5024 Reviewed by Eric Seidel. 5025 5026 Mingw-w64 fixes for JavaScriptCore 5027 https://bugs.webkit.org/show_bug.cgi?id=35607 5028 5029 * runtime/Collector.cpp: use the msvc code for mingw-w64 (but not mingw-w32) 5030 (JSC::Heap::allocateBlock): 5031 (JSC::Heap::freeBlockPtr): 5032 (JSC::currentThreadStackBase): 5033 (JSC::currentThreadStackBase): 5034 * wtf/Platform.h: added COMPILER(MINGW64) check to differentiate between mingw.org and mingw-w64 functions 5035 50362010-04-02 Geoffrey Garen <ggaren@apple.com> 5037 5038 Build fix: updated the .def file. 5039 5040 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 5041 50422010-04-02 Geoffrey Garen <ggaren@apple.com> 5043 5044 Reviewed by Oliver Hunt. 5045 5046 Inlined toThisString and toThisJSString to avoid virtual function call overhead 5047 https://bugs.webkit.org/show_bug.cgi?id=37039 5048 5049 Maybe a 1% speedup on iBench JS. 5050 5051 * JavaScriptCore.exp: New exports. 5052 5053 * runtime/JSCell.cpp: 5054 * runtime/JSCell.h: 5055 * runtime/JSNumberCell.cpp: 5056 * runtime/JSNumberCell.h: 5057 * runtime/JSString.cpp: 5058 * runtime/JSString.h: 5059 * runtime/JSValue.h: 5060 * runtime/JSZombie.h: 5061 (JSC::JSZombie::toThisObject): Nixed the old virtual-type implementation. 5062 5063 * runtime/JSObject.h: 5064 (JSC::JSValue::toThisString): 5065 (JSC::JSValue::toThisJSString): Added the inlined implementation. 5066 50672010-04-02 Jeremy Moskovich <jeremy@chromium.org> 5068 5069 Reviewed by Geoffrey Garen. 5070 5071 Beef up documentation for ASSERT* and CRASH macros a bit. 5072 5073 https://bugs.webkit.org/show_bug.cgi?id=36527 5074 5075 * wtf/Assertions.h: 5076 50772010-04-02 Laszlo Gombos <laszlo.1.gombos@nokia.com> 5078 5079 Unreviewed, minor build fix. 5080 5081 Change the order of the member initialisation list 5082 in constructor to match declaration order 5083 5084 * runtime/Collector.cpp: 5085 (JSC::Heap::Heap): 5086 50872010-04-01 Kinuko Yasuda <kinuko@chromium.org> 5088 5089 Reviewed by Dmitry Titov. 5090 5091 Add FileThread for async file operation support in FileReader and FileWriter 5092 https://bugs.webkit.org/show_bug.cgi?id=36896 5093 5094 Add ENABLE_FILE_READER and ENABLE_FILE_WRITER flags. 5095 5096 * Configurations/FeatureDefines.xcconfig: 5097 50982010-03-31 Gavin Barraclough <barraclough@apple.com> 5099 5100 Reviewed by NOBODY (windows build fix pt II). 5101 5102 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 5103 51042010-03-31 Gavin Barraclough <barraclough@apple.com> 5105 5106 Reviewed by NOBODY (windows build fix). 5107 5108 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 5109 51102010-03-31 Gavin Barraclough <barraclough@apple.com> 5111 5112 Reviewed by Geoff Garen. 5113 5114 Bug 36871 - Remove JSC::CString 5115 Use WTF::CString instead (which until recently was WebCore::CString). 5116 5117 * JavaScriptCore.exp: 5118 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 5119 * runtime/UString.cpp: 5120 * runtime/UString.h: 5121 51222010-03-31 Laszlo Gombos <laszlo.1.gombos@nokia.com> 5123 5124 Unreviewed, fix after r56842. 5125 5126 Add UNUSED_PARAM a to silence warning. 5127 5128 * jit/JITStubs.cpp: 5129 (JSC::DEFINE_STUB_FUNCTION): 5130 51312010-03-31 Laszlo Gombos <laszlo.1.gombos@nokia.com> 5132 5133 Unreviewed, Symbian build fix. 5134 5135 Refactor JITStubs.cpp so that the list of STUB_FUNCTIONs 5136 are not dependent on the JSVALUE32_64 guard. 5137 5138 * jit/JITStubs.cpp: Place the JSVALUE32_64 guard inside 5139 the body of cti_op_eq_strings. 5140 * jit/JITStubs.h: Remove JSVALUE32_64 guard from 5141 cti_op_eq_strings stub. 5142 51432010-03-30 Gavin Barraclough <barraclough@apple.com> 5144 5145 Reviewed by NOBODY (windows build fix). 5146 5147 Fixing b0rked version of JavaScriptCore.vcproj - added lines were truncated. 5148 5149 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: 5150 51512010-03-30 Gavin Barraclough <barraclough@apple.com> 5152 5153 Rubber stamped by Sam Weinig. 5154 5155 https://bugs.webkit.org/show_bug.cgi?id=36866 5156 Move CString to WTF 5157 5158 * Android.mk: 5159 * GNUmakefile.am: 5160 * JavaScriptCore.exp: 5161 * JavaScriptCore.gypi: 5162 * JavaScriptCore.pro: 5163 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 5164 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: 5165 * JavaScriptCore.vcproj/WTF/WTF.vcproj: 5166 * JavaScriptCore.xcodeproj/project.pbxproj: 5167 * wtf/text: Added. 5168 * wtf/text/CString.cpp: Copied from WebCore/platform/text/CString.cpp. 5169 * wtf/text/CString.h: Copied from WebCore/platform/text/CString.h. 5170 (WTF::CStringBuffer::data): 5171 (WTF::CStringBuffer::length): 5172 (WTF::CStringBuffer::create): 5173 (WTF::CStringBuffer::CStringBuffer): 5174 (WTF::CStringBuffer::mutableData): 5175 (WTF::CString::CString): 5176 (WTF::CString::isNull): 5177 (WTF::CString::buffer): 5178 (WTF::operator!=): 5179 51802010-03-30 Jocelyn Turcotte <jocelyn.turcotte@nokia.com> 5181 5182 Reviewed by nobody, build break. 5183 5184 [Qt] Fix build break on Qt Mac. 5185 5186 DESTDIR path on Mac do not include the configuration path by default 5187 like on Windows. Have to force it. 5188 5189 * JavaScriptCore.pro: 5190 51912010-03-29 Alice Liu <alice.liu@apple.com> 5192 5193 Reviewed by NOBODY (build fix). 5194 5195 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCoreGenerated.make: 5196 add JSObjectRefPrivate.h 5197 51982010-03-29 Chao-ying Fu <fu@mips.com> 5199 5200 Reviewed by Oliver Hunt. 5201 5202 MIPS JIT Supports 5203 https://bugs.webkit.org/show_bug.cgi?id=30144 5204 5205 The following changes enable MIPS JIT. 5206 5207 * assembler/MIPSAssembler.h: 5208 (JSC::MIPSAssembler::lbu): 5209 (JSC::MIPSAssembler::linkWithOffset): 5210 * assembler/MacroAssemblerMIPS.h: 5211 (JSC::MacroAssemblerMIPS::load8): 5212 (JSC::MacroAssemblerMIPS::branch8): 5213 (JSC::MacroAssemblerMIPS::branchTest8): 5214 (JSC::MacroAssemblerMIPS::setTest8): 5215 (JSC::MacroAssemblerMIPS::setTest32): 5216 * jit/JIT.h: 5217 * jit/JITInlineMethods.h: 5218 (JSC::JIT::preserveReturnAddressAfterCall): 5219 (JSC::JIT::restoreReturnAddressBeforeReturn): 5220 * jit/JITOpcodes.cpp: 5221 * jit/JITStubs.cpp: 5222 (JSC::JITThunks::JITThunks): 5223 * jit/JITStubs.h: 5224 (JSC::JITStackFrame::returnAddressSlot): 5225 * wtf/Platform.h: 5226 52272010-02-26 Kenneth Rohde Christiansen <kenneth@webkit.org> 5228 5229 Reviewed by Simon Fraser. 5230 5231 Add support for Widgets 1.0: View Mode Media Feature 5232 https://bugs.webkit.org/show_bug.cgi?id=35446 5233 5234 Add an enable flag for the Widgets (http://www.w3.org/TR/widgets-reqs/) 5235 and turn it on for Qt only. 5236 5237 * wtf/Platform.h: 5238 52392010-03-29 Patrick Gansterer <paroga@paroga.com> 5240 5241 Reviewed by Darin Adler. 5242 5243 Corrected name of (u)int64_t compile time assert. 5244 https://bugs.webkit.org/show_bug.cgi?id=36739 5245 5246 int64_t_is_four_bytes -> int64_t_is_eight_bytes 5247 5248 * os-win32/stdint.h: 5249 52502010-03-29 Jocelyn Turcotte <jocelyn.turcotte@nokia.com> 5251 5252 Reviewed by Simon Hausmann. 5253 5254 [Qt] Use the -l syntax for linking against JavaScriptCore on Windows. 5255 This allow qmake to extract dependencies correctly when generating VS 5256 solutions. 5257 5258 * JavaScriptCore.pri: 5259 52602010-03-29 Thomas Zander <t.zander@nokia.com> 5261 5262 Reviewed by Simon Hausmann. 5263 5264 https://bugs.webkit.org/show_bug.cgi?id=36742 5265 5266 gcc for Symbian doesn't support gcc extensions like atomicity.h - disable 5267 5268 * wtf/Threading.h: also detect os symbian 5269 52702010-03-28 Laszlo Gombos <laszlo.1.gombos@nokia.com> 5271 5272 Reviewed by Simon Hausmann. 5273 5274 [Qt] Remove the definition of WTF_CHANGES guards from the build system 5275 https://bugs.webkit.org/show_bug.cgi?id=31670 5276 5277 * JavaScriptCore.pro: Remove the definition of WTF_CHANGES 5278 as it is already defined in config.h 5279 52802010-03-28 Kent Hansen <kent.hansen@nokia.com> 5281 5282 Reviewed by Simon Hausmann. 5283 5284 [Qt] Add API for reporting additional memory cost of JavaScript objects 5285 https://bugs.webkit.org/show_bug.cgi?id=36650 5286 5287 * qt/api/qscriptengine.cpp: 5288 (QScriptEngine::reportAdditionalMemoryCost): 5289 * qt/api/qscriptengine.h: 5290 * qt/api/qscriptengine_p.h: 5291 (QScriptEnginePrivate::reportAdditionalMemoryCost): 5292 * qt/tests/qscriptengine/tst_qscriptengine.cpp: 5293 (tst_QScriptEngine::reportAdditionalMemoryCost): 5294 52952010-03-28 Jedrzej Nowacki <jedrzej.nowacki@nokia.com> 5296 5297 Reviewed by Simon Hausmann. 5298 5299 QScriptEngine API was enriched by globalObject() method 5300 which give an access to the global object. 5301 5302 [Qt] QScriptEngine doesn't give an access to global object 5303 https://bugs.webkit.org/show_bug.cgi?id=36603 5304 5305 * qt/api/qscriptengine.cpp: 5306 (QScriptEngine::globalObject): 5307 * qt/api/qscriptengine.h: 5308 * qt/api/qscriptengine_p.cpp: 5309 (QScriptEnginePrivate::globalObject): 5310 * qt/api/qscriptengine_p.h: 5311 * qt/tests/qscriptengine/tst_qscriptengine.cpp: 5312 (tst_QScriptEngine::globalObject): 5313 53142010-03-26 Jocelyn Turcotte <jocelyn.turcotte@nokia.com> 5315 5316 Reviewed by Simon Hausmann. 5317 5318 [Qt] Build JavaScriptCore as a static library. 5319 https://bugs.webkit.org/show_bug.cgi?id=36590 5320 5321 This patch takes what was left of the unused JavaScriptCore.pro 5322 and moved the compilation logic from JavaScriptCore.pri to 5323 JavaScriptCore.pro. 5324 5325 * JavaScriptCore.pri: 5326 * JavaScriptCore.pro: 5327 * jsc.pro: 5328 * qt/api/QtScript.pro: 5329 53302010-03-25 Gavin Barraclough <barraclough@apple.com> 5331 5332 Reviewed by NOBODY (build fix). 5333 5334 * profiler/ProfileGenerator.cpp: 5335 (JSC::ProfileGenerator::willExecute): 5336 (JSC::ProfileGenerator::didExecute): 5337 53382010-03-25 Gavin Barraclough <barraclough@apple.com> 5339 5340 Reviewed by Geoff Garen. 5341 5342 Bug 36611 - Cleanup JSC::CString 5343 Rename CString::c_str() -> CString::data(), CString::size() -> CString::length(), 5344 remove UString::getCString() (all uses are wrong, should use UString::UTF8String()). 5345 5346 * bytecode/CodeBlock.cpp: 5347 (JSC::CodeBlock::printUnaryOp): 5348 (JSC::CodeBlock::printBinaryOp): 5349 (JSC::CodeBlock::printConditionalJump): 5350 (JSC::CodeBlock::printGetByIdOp): 5351 (JSC::CodeBlock::printPutByIdOp): 5352 (JSC::printGlobalResolveInfo): 5353 (JSC::printStructureStubInfo): 5354 (JSC::CodeBlock::printStructure): 5355 (JSC::CodeBlock::printStructures): 5356 (JSC::CodeBlock::dump): 5357 * jsc.cpp: 5358 (functionPrint): 5359 (functionDebug): 5360 (runInteractive): 5361 (fillBufferWithContentsOfFile): 5362 * profiler/CallIdentifier.h: 5363 (JSC::CallIdentifier::c_str): 5364 * profiler/Profile.cpp: 5365 (JSC::Profile::debugPrintDataSampleStyle): 5366 * profiler/ProfileNode.cpp: 5367 (JSC::ProfileNode::debugPrintData): 5368 (JSC::ProfileNode::debugPrintDataSampleStyle): 5369 * runtime/DateConversion.cpp: 5370 (JSC::parseDate): 5371 * runtime/JSGlobalObjectFunctions.cpp: 5372 (JSC::encode): 5373 (JSC::globalFuncJSCPrint): 5374 * runtime/UString.cpp: 5375 (JSC::operator==): 5376 (JSC::UString::toDouble): 5377 * runtime/UString.h: 5378 (JSC::CString::length): 5379 (JSC::CString::data): 5380 53812010-03-25 Jocelyn Turcotte <jocelyn.turcotte@nokia.com> 5382 5383 Reviewed by nobody, build fix. 5384 5385 [Qt] Build fix on MSVC. Reverts r55633 for stdint.h 5386 5387 This file gets included in generated moc files which don't 5388 include the prefix header. 5389 5390 * os-win32/stdint.h: 5391 53922010-03-24 Gavin Barraclough <barraclough@apple.com> 5393 5394 Reviewed by NOBODY (windows build fix). 5395 53962010-03-24 Gavin Barraclough <barraclough@apple.com> 5397 5398 Reviewed by Sam Weinig. 5399 5400 Switch String::latin1, String::utf8, String::fromUTF8 to 5401 use WTF's Unicode conversion methods rather than TextEncoder. 5402 These methods only perform simple conversion, and don't need 5403 really require TextEncoder's full capability (to look up arbitrary 5404 encodings by name), switching to only be dependent on WTF will 5405 make it easier if we chose to move WebCore::String to WTF. 5406 5407 * JavaScriptCore.exp: 5408 54092010-03-24 Alexey Proskuryakov <ap@apple.com> 5410 5411 Reviewed by Geoff Garen. 5412 5413 * wtf/FastMalloc.h: Added a using directive for fastMallocSize, like we do for all public 5414 WTF symbols. Also sorted the list alphabetically. 5415 54162010-03-23 Gavin Barraclough <barraclough@apple.com> 5417 5418 Reviewed by NOBODY (speculative windows build fix part II). 5419 54202010-03-23 Gavin Barraclough <barraclough@apple.com> 5421 5422 Reviewed by NOBODY (speculative windows build fix). 5423 54242010-03-23 Gavin Barraclough <barraclough@apple.com> 5425 5426 Reviewed by Oliver Hunt. 5427 5428 Bug 36519 - JSGlobalContextRelease is unnecessarily slow 5429 5430 Since [ http://trac.webkit.org/changeset/35917 ], calling 5431 JSGlobalContextRelease always triggers a GC heap collection 5432 (if not a full destroy). As per 35917's changelog "This is 5433 only really necessary when the (JSGlobalObject's) last 5434 reference is released, but there is no way to determine that, 5435 and no harm in collecting slightly more often." 5436 5437 Well, we now know of cases of API clients who are harmed by 5438 the performance penalty of collecting too often, so it's time 5439 to add a way to determine whether a call to JSGlobalContextRelease 5440 is removing the last protect from it's global object. If further 5441 protects are retaining the global object (likely from other 5442 JSGlobalContextRefs), then don't trigger a GC collection. 5443 5444 * API/JSContextRef.cpp: 5445 * runtime/Collector.cpp: 5446 (JSC::Heap::unprotect): return a boolean indicating that the value is now unprotected. 5447 * runtime/Collector.h: 5448 * wtf/HashCountedSet.h: 5449 (WTF::::remove): return a boolean indicating whether the value was removed from the set. 5450 54512010-03-23 Mark Rowe <mrowe@apple.com> 5452 5453 Build fix. 5454 5455 * runtime/ArrayPrototype.cpp: 5456 (JSC::arrayProtoFuncSplice): Some versions of GCC emit a warning about the implicit 64- to 32-bit truncation 5457 that takes place here. An explicit cast is sufficient to silence it. 5458 54592010-03-23 Alexey Proskuryakov <ap@apple.com> 5460 5461 Build fix. 5462 5463 * runtime/ArrayPrototype.cpp: (JSC::arrayProtoFuncSplice): Fixed a typo - length doesn't 5464 need to be converted with toInteger(). 5465 54662010-03-23 Alexey Proskuryakov <ap@apple.com> 5467 5468 Reviewed by Geoff Garen. 5469 5470 https://bugs.webkit.org/show_bug.cgi?id=36511 5471 <rdar://problem/7753498> Safari freezes when using SPUTNIK JavaScript conformance check 5472 5473 Test: fast/js/sputnik-S15.4.4.12_A3_T3.html 5474 5475 * runtime/ArrayPrototype.cpp: (JSC::arrayProtoFuncSplice): We were incorrectly computing 5476 the start offset, and iterated over (almost) all integers. Note that this can be fixed 5477 without using doubles, but the code would be much more complicated, and there is no important 5478 reason to stick to integers here. 5479 54802010-03-23 Kent Hansen <kent.hansen@nokia.com> 5481 5482 Reviewed by Darin Adler. 5483 5484 Fix compilation on Itanium in 32-bit mode 5485 https://bugs.webkit.org/show_bug.cgi?id=36494 5486 5487 * wtf/Platform.h: Introduce CPU(IA64_32). Don't define 5488 WTF_USE_JSVALUE64 if the CPU is in 32-bit mode. 5489 54902010-03-23 Geoffrey Garen <ggaren@apple.com> 5491 5492 Reviewed by Mark Rowe. 5493 5494 Interpreter fix for <rdar://problem/7728196> REGRESSION (r46701): -(-2147483648) 5495 evaluates to -2147483648 on 32 bit (35842) 5496 5497 * interpreter/Interpreter.cpp: 5498 (JSC::Interpreter::privateExecute): Only take the fast negate path if 5499 a bit other than bit 31 is set. If none of bits 0-30 are set, then the 5500 value we're negating can only be 0 or -2147483648, and neither can be 5501 negated in int space. 5502 5503 * jit/JITArithmetic.cpp: 5504 (JSC::JIT::emit_op_negate): 5505 (JSC::JIT::emitSlow_op_negate): Updated the JIT implementation to match 5506 the interpreter, since it's slightly simpler. 5507 55082010-03-22 Siddharth Mathur <siddharth.mathur@nokia.com> 5509 5510 Reviewed by Laszlo Gombos. 5511 5512 [Symbian] More efficient aligned memory allocation for JSC Collector 5513 https://bugs.webkit.org/show_bug.cgi?id=34350 5514 5515 * JavaScriptCore.pri: Added 2 new Symbian source files and HAL linkage 5516 5517 * runtime/Collector.cpp: Reduced port-specific code and added private data member 5518 (JSC::Heap::Heap): 5519 (JSC::Heap::~Heap): 5520 (JSC::Heap::destroy): 5521 (JSC::Heap::allocateBlock): 5522 (JSC::Heap::freeBlockPtr): 5523 5524 * runtime/Collector.h: Added private data member 5525 5526 * wtf/symbian: Added. 5527 * wtf/symbian/BlockAllocatorSymbian.cpp: Added. 5528 (WTF::AlignedBlockAllocator::AlignedBlockAllocator): Helper class to allocate 5529 aligned blocks more efficiently as required by Collector 5530 (WTF::AlignedBlockAllocator::alloc): 5531 (WTF::AlignedBlockAllocator::free): 5532 (WTF::AlignedBlockAllocator::destroy): 5533 (WTF::AlignedBlockAllocator::~AlignedBlockAllocator): 5534 * wtf/symbian/BlockAllocatorSymbian.h: Added. 5535 55362010-03-22 Geoffrey Garen <ggaren@apple.com> 5537 5538 Reviewed by Sam Weinig. 5539 5540 Fixed <rdar://problem/7728196> REGRESSION (r46701): -(-2147483648) 5541 evaluates to -2147483648 on 32 bit (35842) 5542 5543 Two ways to fix the same bug: 5544 5545 1. Check for overflow when negating, since negating the largest negative 5546 int causes overflow. 5547 5548 2. Constant-fold even when negating a negative, since, like they say in 5549 high school, "math works." 5550 5551 * assembler/MacroAssemblerARM.h: 5552 (JSC::MacroAssemblerARM::branchNeg32): 5553 * assembler/MacroAssemblerX86Common.h: 5554 (JSC::MacroAssemblerX86Common::branchNeg32): Added a branching version 5555 of the negate operator. 5556 5557 * jit/JITArithmetic.cpp: 5558 (JSC::JIT::emit_op_negate): Use the branching version of the negate 5559 operator to check for overflow. 5560 5561 (JSC::JIT::emitSlow_op_negate): Link the check for overflow to a slow case. 5562 (We could emit inline code for this, since we know what the result would 5563 be, but that's probably just a waste of generated code.) 5564 5565 * parser/Grammar.y: Constant fold even when negating a negative. 5566 55672010-03-22 David Kilzer <ddkilzer@apple.com> 5568 5569 <http://webkit.org/b/36431> Clean up 'int' use in UString.cpp after r54789 5570 5571 Reviewed by Darin Adler. 5572 5573 * runtime/UString.cpp: 5574 (JSC::UString::from): Changed argument type from 'unsigned int' 5575 to 'unsigned' to match WebKit coding style. 5576 (JSC::UString::find): Changed static_cast<int>() to 5577 static_cast<unsigned>() now that this method returns unsigned. 5578 (JSC::UString::rfind): Ditto. 5579 * runtime/UString.h: 5580 (JSC::UString::from): Changed argument type from 'unsigned int' 5581 to 'unsigned' to match WebKit coding style. 5582 55832010-03-22 Jedrzej Nowacki <jedrzej.nowacki@nokia.com> 5584 5585 Reviewed by Kenneth Rohde Christiansen. 5586 5587 Add support for syntax checking in the QtScript API. 5588 5589 New class was created; the QScriptSyntaxCheckResult which main 5590 responsibility is to provide results of the ECMA Script code 5591 syntax check. The class is not fully functional as the JSC C API 5592 doesn't expose an error column number, but it is a good start point 5593 for a future development. 5594 5595 [Qt] QtScript functionality should be extended by syntax checking. 5596 https://bugs.webkit.org/show_bug.cgi?id=36123 5597 5598 * qt/api/QtScript.pro: 5599 * qt/api/qscriptengine.cpp: 5600 (QScriptEngine::checkSyntax): 5601 * qt/api/qscriptengine.h: 5602 * qt/api/qscriptengine_p.cpp: 5603 (QScriptEnginePrivate::checkSyntax): 5604 * qt/api/qscriptengine_p.h: 5605 * qt/api/qscriptsyntaxcheckresult.cpp: Added. 5606 (QScriptSyntaxCheckResult::QScriptSyntaxCheckResult): 5607 (QScriptSyntaxCheckResult::~QScriptSyntaxCheckResult): 5608 (QScriptSyntaxCheckResult::operator=): 5609 (QScriptSyntaxCheckResult::state): 5610 (QScriptSyntaxCheckResult::errorLineNumber): 5611 (QScriptSyntaxCheckResult::errorColumnNumber): 5612 (QScriptSyntaxCheckResult::errorMessage): 5613 * qt/api/qscriptsyntaxcheckresult.h: Added. 5614 * qt/api/qscriptsyntaxcheckresult_p.cpp: Added. 5615 (QScriptSyntaxCheckResultPrivate::~QScriptSyntaxCheckResultPrivate): 5616 (QScriptSyntaxCheckResultPrivate::errorMessage): 5617 (QScriptSyntaxCheckResultPrivate::errorLineNumber): 5618 * qt/api/qscriptsyntaxcheckresult_p.h: Added. 5619 (QScriptSyntaxCheckResultPrivate::get): 5620 (QScriptSyntaxCheckResultPrivate::QScriptSyntaxCheckResultPrivate): 5621 (QScriptSyntaxCheckResultPrivate::state): 5622 (QScriptSyntaxCheckResultPrivate::errorColumnNumber): 5623 * qt/tests/qscriptengine/tst_qscriptengine.cpp: 5624 (tst_QScriptEngine::checkSyntax_data): 5625 (tst_QScriptEngine::checkSyntax): 5626 56272010-03-21 Jedrzej Nowacki <jedrzej.nowacki@nokia.com> 5628 5629 Reviewed by Simon Hausmann. 5630 5631 New class; QScriptProgram. 5632 5633 The class should be used to evaluate the same script multiple times 5634 more efficiently. 5635 5636 [Qt] QtScript should have QScriptProgram class 5637 https://bugs.webkit.org/show_bug.cgi?id=36008 5638 5639 * qt/api/QtScript.pro: 5640 * qt/api/qscriptengine.cpp: 5641 (QScriptEngine::evaluate): 5642 * qt/api/qscriptengine.h: 5643 * qt/api/qscriptengine_p.cpp: 5644 (QScriptEnginePrivate::evaluate): 5645 * qt/api/qscriptengine_p.h: 5646 (QScriptEnginePrivate::evaluate): 5647 * qt/api/qscriptprogram.cpp: Added. 5648 (QScriptProgram::QScriptProgram): 5649 (QScriptProgram::~QScriptProgram): 5650 (QScriptProgram::operator=): 5651 (QScriptProgram::isNull): 5652 (QScriptProgram::sourceCode): 5653 (QScriptProgram::fileName): 5654 (QScriptProgram::firstLineNumber): 5655 (QScriptProgram::operator==): 5656 (QScriptProgram::operator!=): 5657 * qt/api/qscriptprogram.h: Added. 5658 * qt/api/qscriptprogram_p.h: Added. 5659 (QScriptProgramPrivate::get): 5660 (QScriptProgramPrivate::QScriptProgramPrivate): 5661 (QScriptProgramPrivate::~QScriptProgramPrivate): 5662 (QScriptProgramPrivate::isNull): 5663 (QScriptProgramPrivate::sourceCode): 5664 (QScriptProgramPrivate::fileName): 5665 (QScriptProgramPrivate::firstLineNumber): 5666 (QScriptProgramPrivate::operator==): 5667 (QScriptProgramPrivate::operator!=): 5668 (QScriptProgramPrivate::program): 5669 (QScriptProgramPrivate::file): 5670 (QScriptProgramPrivate::line): 5671 * qt/tests/qscriptengine/tst_qscriptengine.cpp: 5672 (tst_QScriptEngine::evaluateProgram): 5673 56742010-03-21 David Kilzer <ddkilzer@apple.com> 5675 5676 Blind attempt #2 to fix the Windows build after r56314 5677 5678 * API/tests/testapi.c: Include JSObjectRefPrivate.h for the new 5679 methods instead of declaring them locally (and non-extern). 5680 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 5681 Backed out previous change. 5682 56832010-03-21 David Kilzer <ddkilzer@apple.com> 5684 5685 Blind attempt to fix the Windows build after r56314 5686 5687 Try to fix the following errors on the Windows buildbot: 5688 5689 Linking... 5690 testapi.obj : error LNK2001: unresolved external symbol "bool __cdecl JSObjectSetPrivateProperty(struct OpaqueJSContext const *,struct OpaqueJSValue *,struct OpaqueJSString *,struct OpaqueJSValue const *)" (?JSObjectSetPrivateProperty@@YA_NPBUOpaqueJSContext@@PAUOpaqueJSValue@@PAUOpaqueJSString@@PBU2@@Z) 5691 testapi.obj : error LNK2001: unresolved external symbol "struct OpaqueJSValue const * __cdecl JSObjectGetPrivateProperty(struct OpaqueJSContext const *,struct OpaqueJSValue *,struct OpaqueJSString *)" (?JSObjectGetPrivateProperty@@YAPBUOpaqueJSValue@@PBUOpaqueJSContext@@PAU1@PAUOpaqueJSString@@@Z) 5692 C:\cygwin\home\buildbot\slave\win-release\build\WebKitBuild\bin\testapi.exe : fatal error LNK1120: 2 unresolved externals 5693 5694 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: Added 5695 missing symbols to be exported. 5696 56972010-03-21 Oliver Hunt <oliver@apple.com> 5698 5699 Reviewed by Maciej Stachowiak. 5700 5701 Documentation fix for previous patch. 5702 5703 * API/JSObjectRefPrivate.h: 5704 57052010-03-20 Oliver Hunt <oliver@apple.com> 5706 5707 Reviewed by Maciej Stachowiak. 5708 5709 JSC needs an API to allow custom objects to have aprivate GC-accessible properties 5710 https://bugs.webkit.org/show_bug.cgi?id=36420 5711 5712 Add new API methods to support "private" properties on custom 5713 objects. 5714 5715 * API/JSCallbackObject.h: 5716 (JSC::JSCallbackObjectData::JSCallbackObjectData): 5717 (JSC::JSCallbackObjectData::~JSCallbackObjectData): 5718 (JSC::JSCallbackObjectData::getPrivateProperty): 5719 (JSC::JSCallbackObjectData::setPrivateProperty): 5720 (JSC::JSCallbackObjectData::deletePrivateProperty): 5721 (JSC::JSCallbackObjectData::markChildren): 5722 (JSC::JSCallbackObjectData::JSPrivatePropertyMap::getPrivateProperty): 5723 (JSC::JSCallbackObjectData::JSPrivatePropertyMap::setPrivateProperty): 5724 (JSC::JSCallbackObjectData::JSPrivatePropertyMap::deletePrivateProperty): 5725 (JSC::JSCallbackObjectData::JSPrivatePropertyMap::markChildren): 5726 (JSC::JSCallbackObject::getPrivateProperty): 5727 (JSC::JSCallbackObject::setPrivateProperty): 5728 (JSC::JSCallbackObject::deletePrivateProperty): 5729 (JSC::JSCallbackObject::markChildren): 5730 * API/JSObjectRef.cpp: 5731 (JSObjectGetPrivateProperty): 5732 (JSObjectSetPrivateProperty): 5733 (JSObjectDeletePrivateProperty): 5734 * API/JSObjectRefPrivate.h: Added. 5735 * API/tests/testapi.c: 5736 (main): 5737 * JavaScriptCore.exp: 5738 * JavaScriptCore.xcodeproj/project.pbxproj: 5739 57402010-03-20 Kevin Ollivier <kevino@theolliviers.com> 5741 5742 [wx] Build fixes after introduction of Brew files. 5743 5744 * wscript: 5745 57462010-03-18 Tom Callaway <tcallawa@redhat.com> 5747 5748 Reviewed by Darin Adler. 5749 5750 Bug 35429: Fix compile on SPARC64 5751 https://bugs.webkit.org/show_bug.cgi?id=35429 5752 5753 * wtf/Platform.h: Set WTF_USE_JSVALUE64 for SPARC64 5754 57552010-03-18 Oliver Hunt <oliver@apple.com> 5756 5757 Reviewed by Sam Weinig. 5758 5759 Add API to directly expose JSON parsing 5760 https://bugs.webkit.org/show_bug.cgi?id=34887 5761 5762 Add API to expose JSON parsing directly, and add tests to testapi 5763 5764 * API/JSValueRef.cpp: 5765 (JSValueMakeFromJSONString): 5766 (JSValueCreateJSONString): 5767 * API/tests/testapi.c: 5768 (main): 5769 * JavaScriptCore.exp: 5770 * runtime/JSONObject.cpp: 5771 (JSC::JSONStringify): 5772 * runtime/JSONObject.h: 5773 57742010-03-16 Sam Weinig <sam@webkit.org> 5775 5776 Reviewed by Darin Adler and Mark Rowe. 5777 5778 Update WebKit availability macros for release after 4.0. 5779 5780 * API/WebKitAvailability.h: 5781 57822010-03-17 Oliver Hunt <oliver@apple.com> 5783 5784 Reviewed by Gavin Barraclough. 5785 5786 undefined, NaN, and Infinity should be ReadOnly 5787 https://bugs.webkit.org/show_bug.cgi?id=36263 5788 5789 Simply add the ReadOnly flag to these properties. 5790 5791 * runtime/JSGlobalObject.cpp: 5792 (JSC::JSGlobalObject::reset): 5793 57942010-03-17 Darin Adler <darin@apple.com> 5795 5796 Reviewed by Oliver Hunt. 5797 5798 Speed up Math.round a little by removing unneeded special case 5799 https://bugs.webkit.org/show_bug.cgi?id=36107 5800 5801 Test: fast/js/math.html 5802 5803 * runtime/MathObject.cpp: 5804 (JSC::mathProtoFuncRound): This function had a special case for numbers 5805 between -0.5 and -0.0 to return -0.0. But the algorithm in the function 5806 already yields -0.0 for those cases, so the extra checking and branching 5807 is unneeded. 5808 58092010-03-17 Mike Homey <glandium@debian.org> 5810 5811 Reviewed by Gustavo Noronha. 5812 5813 Build fix for SPARC. Fix missing macro value. 5814 5815 * wtf/Platform.h: 5816 58172010-03-16 Gavin Barraclough <barraclough@apple.com> 5818 5819 Reviewed by Oliver Hunt, Darin Adler. 5820 5821 Bug 36083 - REGRESSION (r55772-r55834): Crash in JavaScriptCore RegExp code on PowerPC 5822 5823 The problem is a bug in our port of PCRE - that a read may take place from the first character in an 5824 empty string. For the time being, revert to using a valid pointer in the data segment rather than 5825 an invalid non-null pointer into the zero-page for the empty string's data pointer. A better fix for 5826 this will be to remove PCRE. 5827 5828 * runtime/UStringImpl.cpp: 5829 (JSC::UStringImpl::empty): 5830 58312010-03-16 Darin Adler <darin@apple.com> 5832 5833 Rolled out r56081 since it broke the Windows build. 5834 58352010-03-16 Zoltan Horvath <zoltan@webkit.org> 5836 5837 Reviewed by Darin Adler. 5838 5839 Remove extra <new> include and add guards to operator new/delete definitions 5840 https://bugs.webkit.org/show_bug.cgi?id=35967 5841 5842 Remove extra <new> header include from FastAlloc.cpp since it is included in 5843 FastAlloc.h. Add ENABLE(GLOBAL_FASTMALLOC_NEW) macro guard to operator 5844 new/delete/new []/delete [] definitions. 5845 5846 * wtf/FastMalloc.cpp: 5847 58482010-03-15 Kwang Yul Seo <skyul@company100.net> 5849 5850 Reviewed by Eric Seidel. 5851 5852 [BREWMP] Add a function to create a BREW instance without local variable declarations. 5853 https://bugs.webkit.org/show_bug.cgi?id=34705 5854 5855 Add a template function to create a BREW instance in one line. 5856 5857 * wtf/brew/ShellBrew.h: Added. 5858 (WTF::createInstance): 5859 58602010-03-15 Geoffrey Garen <ggaren@apple.com> 5861 5862 Not reviewed. 5863 5864 Removed a now-incorrect comment I forgot to remove in my last check-in. 5865 5866 * wtf/FastMalloc.cpp: 5867 (WTF::TCMalloc_PageHeap::scavenge): 5868 58692010-03-15 Geoffrey Garen <ggaren@apple.com> 5870 5871 Reviewed by Sam Weinig. 5872 5873 Fixed a portion of: 5874 <rdar://problem/7165917> | https://bugs.webkit.org/show_bug.cgi?id=28676 5875 Safari 4 does not release memory back to the operating system fast enough (28676) 5876 5877 Every few seconds, release a percentage of the minimum unused page count 5878 during that time period. 5879 5880 SunSpider reports no change, command-line or in-browser, Mac or Windows. 5881 5882 * wtf/FastMalloc.cpp: 5883 (WTF::TCMalloc_PageHeap::init): 5884 (WTF::TCMalloc_PageHeap::signalScavenger): 5885 (WTF::TCMalloc_PageHeap::initializeScavenger): Renamed shouldContinueScavenging 5886 to shouldScavenge, since scavenging is no longer something that we interrupt. 5887 5888 (WTF::TCMalloc_PageHeap::scavenge): The new scavenging algorithm. Fixes 5889 a bug where the old code would release only one item from each size class 5890 per scavenge, potentially leaving large numbers of large-sized objects 5891 unreleased for a long time. 5892 5893 (WTF::TCMalloc_PageHeap::shouldScavenge): 5894 (WTF::TCMalloc_PageHeap::New): 5895 (WTF::TCMalloc_PageHeap::AllocLarge): 5896 (WTF::TCMalloc_PageHeap::Delete): 5897 (WTF::TCMalloc_PageHeap::GrowHeap): 5898 (WTF::TCMalloc_PageHeap::scavengerThread): 5899 (WTF::TCMalloc_PageHeap::periodicScavenge): Updated to track the minimum 5900 value of free_committed_pages_ during a given scavenge period. 5901 59022010-03-15 Gavin Barraclough <barraclough@apple.com> 5903 5904 Reviewed by Sam Weinig. 5905 5906 https://bugs.webkit.org/show_bug.cgi?id=35843 5907 Re-land reverted fix to JSString::getIndex() 5908 5909 Calling getIndex() on a JSString in rope form may result in a JSException being thrown 5910 if there is insuficient memory so value(exec) returns UString() with length zero, 5911 which will be passed to jsSingleCharacterSubstring. 5912 Add a slow case function to trap the error & return a safe null value, until the 5913 exception is handled. 5914 5915 * runtime/JSString.cpp: 5916 (JSC::JSString::getIndexSlowCase): 5917 (JSC::JSString::getStringPropertyDescriptor): 5918 * runtime/JSString.h: 5919 (JSC::jsSingleCharacterSubstring): 5920 (JSC::JSString::getIndex): 5921 (JSC::jsSingleCharacterString): 5922 (JSC::JSString::getStringPropertySlot): 5923 59242010-03-04 Kenneth Rohde Christiansen <kenneth@webkit.org> 5925 5926 Reviewed by Adam Roben. 5927 5928 Add a long long version of abs() for MSVC. 5929 5930 * wtf/MathExtras.h: 5931 (abs): 5932 59332010-03-15 Gabor Loki <loki@webkit.org> 5934 5935 Reviewed by Gavin Barraclough. 5936 5937 Combine ctiTrampolines on ARM and Thumb-2 5938 https://bugs.webkit.org/show_bug.cgi?id=36014 5939 5940 * jit/JITStubs.cpp: 5941 (JSC::JITThunks::JITThunks): 5942 59432010-03-12 Gavin Barraclough <barraclough@apple.com> 5944 5945 Reviewed by NOBODY (build fix). 5946 5947 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 5948 59492010-03-12 Gavin Barraclough <barraclough@apple.com> 5950 5951 Reviewed by NOBODY (build fix). 5952 5953 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 5954 59552010-03-11 Gavin Barraclough <barraclough@apple.com> 5956 5957 Reviewed by Oliver Hunt. 5958 5959 Bug 36075 - Clean up screwyness re static string impls & Identifiers. 5960 5961 * API/JSClassRef.cpp: 5962 (OpaqueJSClass::~OpaqueJSClass): Classname may be null/empty, and these are an identifer. This is okay, since the null/empty strings are shared across all threads. 5963 * JavaScriptCore.exp: 5964 * runtime/Identifier.cpp: 5965 (JSC::Identifier::add): No need to explicitly hash null reps, this is done in the ststic UStringImpl constructor. 5966 (JSC::Identifier::addSlowCase): UStringImpl::empty() handled & checkCurrentIdentifierTable now called in the header. 5967 (JSC::Identifier::checkCurrentIdentifierTable): Replaces checkSameIdentifierTable (this no longer checked the rep since the identifierTable pointer was removed from UString::Rep long ago). 5968 * runtime/Identifier.h: 5969 (JSC::Identifier::add): Replace call to checkSameIdentifierTable with call to checkCurrentIdentifierTable at head of function. 5970 * runtime/UStringImpl.cpp: 5971 (JSC::UStringImpl::~UStringImpl): Remove call to checkConsistency - this function no longer checks anything interesting. 5972 * runtime/UStringImpl.h: 5973 (JSC::UStringOrRopeImpl::UStringOrRopeImpl): Set s_refCountFlagIsIdentifier in static constructor. 5974 (JSC::UStringImpl::UStringImpl): remove calls to checkConsistency (see above), add new ASSERT to substring constructor. 5975 (JSC::UStringImpl::setHash): ASSERT not static (static strings set the hash in their constructor, should not reach this code path). 5976 (JSC::UStringImpl::create): Add missing ASSERT. 5977 (JSC::UStringImpl::setIsIdentifier): ASSERT !isStatic() (static strings hash set in constructor). 5978 59792010-03-12 Peter Varga <pvarga@inf.u-szeged.hu> 5980 5981 Reviewed by David Levin. 5982 5983 Two functions tryConsumeCharacter() and tryConsumeCharacterClass() are 5984 removed from yarr/RegexInterpreter.cpp because they are never called. 5985 5986 * yarr/RegexInterpreter.cpp: 5987 59882010-03-11 Jedrzej Nowacki <jedrzej.nowacki@nokia.com> 5989 5990 Reviewed by Simon Hausmann. 5991 5992 The JSNative state was renamed to JSPrimitive. The new name better 5993 coresponds to the ECMAScript standard. 5994 5995 Enum QScriptValuePrivate::States was renamed to State to obey Qt 5996 coding style rules ("States" name suggests that a state could 5997 mixed together with an other state using bitwise logic operators. 5998 5999 [Qt] QScriptValuePrivate::States has naming issues 6000 https://bugs.webkit.org/show_bug.cgi?id=35968 6001 6002 * qt/api/qscriptvalue_p.h: 6003 (QScriptValuePrivate::): 6004 (QScriptValuePrivate::QScriptValuePrivate): 6005 (QScriptValuePrivate::isBool): 6006 (QScriptValuePrivate::isNumber): 6007 (QScriptValuePrivate::isNull): 6008 (QScriptValuePrivate::isString): 6009 (QScriptValuePrivate::isUndefined): 6010 (QScriptValuePrivate::toString): 6011 (QScriptValuePrivate::toNumber): 6012 (QScriptValuePrivate::toBool): 6013 (QScriptValuePrivate::assignEngine): 6014 (QScriptValuePrivate::refinedJSValue): 6015 60162010-03-11 Gavin Barraclough <barraclough@apple.com> 6017 6018 Reviewed by NOBODY (Windows build fix). 6019 6020 Add export. 6021 6022 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 6023 60242010-03-11 Gavin Barraclough <barraclough@apple.com> 6025 6026 Reviewed by NOBODY (Windows build fix). 6027 6028 Add export. 6029 6030 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 6031 60322010-03-11 Gavin Barraclough <barraclough@apple.com> 6033 6034 Rubber stamped by Oliver Hunt. 6035 6036 Remove nonsense comments used in development & commited in error. 6037 6038 * runtime/UStringImpl.h: 6039 60402010-03-11 Gavin Barraclough <barraclough@apple.com> 6041 6042 Reviewed by NOBODY (Windows build fix). 6043 6044 Remove export. 6045 6046 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 6047 60482010-03-11 Gavin Barraclough <barraclough@apple.com> 6049 6050 Reviewed by Oliver Hunt. 6051 6052 https://bugs.webkit.org/show_bug.cgi?id=36041 6053 Remove unnecessary differences in common code between WebCore::StringImpl & JSC::UStringImpl 6054 6055 Much of the code in WebCore::StringImpl and JSC::UStringImpl is now very similar, 6056 but has trivial and unnecessary formatting differences, such as the exact wording 6057 of comments, missing ASSERTs, functions implemented in the .h vs .cpp etc. 6058 6059 * runtime/Identifier.cpp: 6060 (JSC::Identifier::add): UStringImpl::empty() now automatically hashes, uas per WebCore strings. 6061 (JSC::Identifier::addSlowCase): UStringImpl::empty() now automatically hashes, uas per WebCore strings. 6062 * runtime/UStringImpl.cpp: 6063 (JSC::UStringImpl::~UStringImpl): Only call bufferOwnership() once, add missing ASSERTs. 6064 (JSC::UStringImpl::createUninitialized): Move from .h, not commonly called, no need to inline. 6065 (JSC::UStringImpl::create): Move from .h, not commonly called, no need to inline. 6066 (JSC::UStringImpl::sharedBuffer): Rewritten to more closely match WebCore implementation, remove need for separate baseSharedBuffer() method. 6067 * runtime/UStringImpl.h: 6068 (JSC::UStringImpl::UStringImpl): Automatically hash static strings, ASSERT m_data & m_length are non-null/non-zero in non-static strings. 6069 (JSC::UStringImpl::setHash): Add missing ASSERT. 6070 (JSC::UStringImpl::create): Moved to .cpp / added missing check for empty string creation. 6071 (JSC::UStringImpl::adopt): Vector.size() returns size_t, not unsigned. 6072 (JSC::UStringImpl::cost): Renamed m_bufferSubstring -> m_substringBuffer 6073 (JSC::UStringImpl::hash): Reordered in file. 6074 (JSC::UStringImpl::existingHash): Reordered in file. 6075 (JSC::UStringImpl::computeHash): Reordered in file, renamed parameter. 6076 (JSC::UStringImpl::checkConsistency): rewrote ASSERT. 6077 (JSC::UStringImpl::bufferOwnership): Return type should be BufferOwnership. 6078 (JSC::UStringImpl::): Moved friends to head of class. 6079 60802010-03-11 Mark Rowe <mrowe@apple.com> 6081 6082 Reviewed by David Kilzer. 6083 6084 <rdar://problem/7745082> Make it possible to build WebKit for older Mac OS X versions from the current Mac OS X version 6085 6086 Default to using the appropriate SDK if the target Mac OS X version is not the current Mac OS X version. 6087 6088 * Configurations/Base.xcconfig: 6089 60902010-03-11 Mark Rowe <mrowe@apple.com> 6091 6092 Reviewed by Tim Hatcher. 6093 6094 <rdar://problem/7745082> Make it possible to build WebKit for older Mac OS X versions from the current Mac OS X version 6095 6096 Introduce TARGET_MAC_OS_X_VERSION_MAJOR to represent the Mac OS X version that is being targeted. It defaults to the 6097 current Mac OS X version unless otherwise specified. 6098 6099 Key off TARGET_MAC_OS_X_VERSION_MAJOR where we'd previously been keying off MAC_OS_X_VERSION_MAJOR. 6100 6101 Explicitly map from the target Mac OS X version to the preferred compiler since Xcode's default compiler choice 6102 may not be usable when targetting a different Mac OS X version. 6103 6104 Key off TARGET_GCC_VERSION rather than MAC_OS_X_VERSION_MAJOR in locations where we'd previously been keying off 6105 MAC_OS_X_VERSION_MAJOR but the decision is really related to the compiler version being used. 6106 6107 * Configurations/Base.xcconfig: 6108 * Configurations/DebugRelease.xcconfig: 6109 * Configurations/FeatureDefines.xcconfig: 6110 * Configurations/JavaScriptCore.xcconfig: 6111 * Configurations/Version.xcconfig: 6112 61132010-03-11 Simon Fraser <simon.fraser@apple.com> 6114 6115 Reviewed by Mark Rowe. 6116 6117 Sort the project file. 6118 6119 * JavaScriptCore.xcodeproj/project.pbxproj: 6120 61212010-03-11 Simon Fraser <simon.fraser@apple.com> 6122 6123 Reviewed by Mark Rowe. 6124 6125 Sort the project file . 6126 6127 * JavaScriptCore.xcodeproj/project.pbxproj: 6128 61292010-03-11 Gabor Loki <loki@webkit.org> 6130 6131 Reviewed by Gavin Barraclough. 6132 6133 Buildfix for Thumb-2 after r55684. Add branch8 and branchTest8 functions. 6134 https://bugs.webkit.org/show_bug.cgi?id=35892 6135 6136 * assembler/ARMv7Assembler.h: 6137 (JSC::ARMv7Assembler::): 6138 (JSC::ARMv7Assembler::ldrb): 6139 * assembler/MacroAssemblerARMv7.h: 6140 (JSC::MacroAssemblerARMv7::load8): 6141 (JSC::MacroAssemblerARMv7::branch8): 6142 (JSC::MacroAssemblerARMv7::branchTest8): 6143 (JSC::MacroAssemblerARMv7::setTest8): 6144 61452010-03-10 Gavin Barraclough <barraclough@apple.com> 6146 6147 Rubber stamped by Oliver Hunt. 6148 6149 Rename JSC::UStringImpl::data() to characters(), to match WebCore::StringImpl. 6150 6151 * API/JSClassRef.cpp: 6152 (OpaqueJSClassContextData::OpaqueJSClassContextData): 6153 * bytecompiler/BytecodeGenerator.cpp: 6154 (JSC::keyForCharacterSwitch): 6155 * bytecompiler/NodesCodegen.cpp: 6156 (JSC::processClauseList): 6157 * interpreter/Interpreter.cpp: 6158 (JSC::Interpreter::privateExecute): 6159 * jit/JITStubs.cpp: 6160 (JSC::DEFINE_STUB_FUNCTION): 6161 * runtime/ArrayPrototype.cpp: 6162 (JSC::arrayProtoFuncToString): 6163 * runtime/Identifier.cpp: 6164 (JSC::Identifier::equal): 6165 (JSC::Identifier::addSlowCase): 6166 * runtime/JSString.cpp: 6167 (JSC::JSString::resolveRope): 6168 * runtime/UString.cpp: 6169 (JSC::UString::toStrictUInt32): 6170 (JSC::equal): 6171 * runtime/UString.h: 6172 (JSC::UString::data): 6173 * runtime/UStringImpl.h: 6174 (JSC::UStringImpl::characters): 6175 (JSC::UStringImpl::hash): 6176 (JSC::UStringImpl::setHash): 6177 61782010-03-10 Gavin Barraclough <barraclough@apple.com> 6179 6180 Reviewed by Darin Adler, Geoffrey Garen, Maciej Stachowiak. 6181 6182 https://bugs.webkit.org/show_bug.cgi?id=35991 6183 Would be faster to not use a thread specific to implement StringImpl::empty() 6184 6185 Change JSC::UStringImpl's implementation of empty() match to match StringImpl's new implementation 6186 (use a static defined within the empty() method), and change the interface to match too (return 6187 a pointer not a reference). 6188 6189 ~0% performance impact (possible minor progression from moving empty() from .h to .cpp). 6190 6191 * JavaScriptCore.exp: 6192 * runtime/Identifier.cpp: 6193 (JSC::Identifier::add): 6194 (JSC::Identifier::addSlowCase): 6195 * runtime/PropertyNameArray.cpp: 6196 (JSC::PropertyNameArray::add): 6197 * runtime/UString.cpp: 6198 (JSC::initializeUString): 6199 (JSC::UString::UString): 6200 * runtime/UStringImpl.cpp: 6201 (JSC::UStringImpl::empty): 6202 (JSC::UStringImpl::create): 6203 * runtime/UStringImpl.h: 6204 (JSC::UStringImpl::adopt): 6205 (JSC::UStringImpl::createUninitialized): 6206 (JSC::UStringImpl::tryCreateUninitialized): 6207 62082010-03-10 Dmitry Titov <dimich@chromium.org> 6209 6210 Not reviewed, fixing Snow Leopard build. 6211 6212 * wtf/mac/MainThreadMac.mm: Forgot 'static' for a new local function. 6213 (WTF::postTimer): 6214 62152010-03-10 Dmitry Titov <dimich@chromium.org> 6216 6217 Reviewed by Darin Adler. 6218 6219 Make Document::postTask to use a single queue of tasks, to fire them in order 6220 https://bugs.webkit.org/show_bug.cgi?id=35943 6221 6222 The patch uses CFRunLoopTimer to schedule execution of tasks instead of performSelectorOnMainThread which apparently can starve other event sources. 6223 The timer is used when the schedule request is coming on the main thread itself. This happens when the task is posted on the main thread or 6224 when too many tasks are posted and the queue does 'stop and re-schedule' to make sure run loop has a chance to execute other events. 6225 6226 * wtf/mac/MainThreadMac.mm: 6227 (WTF::timerFired): 6228 (WTF::postTimer): 6229 (WTF::scheduleDispatchFunctionsOnMainThread): Use timer posted to the current RunLoop if scheduling the task execution while on the main thread. 6230 62312010-03-10 Geoffrey Garen <ggaren@apple.com> 6232 6233 Windows build fix: added new symbol. 6234 6235 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 6236 62372010-03-10 Geoffrey Garen <ggaren@apple.com> 6238 6239 Windows build fix: removed old symbol. 6240 6241 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 6242 62432010-03-09 Geoffrey Garen <ggaren@apple.com> 6244 6245 Reviewed by Alexey Proskuryakov, Darin Adler, and Sam Weinig. 6246 6247 Refactored fastCheckConsistency to match some review comments: 6248 - renamed fastCheckConsistency to fastMallocSize, and changed ValueCheck 6249 to ASSERT that a pointer's fastMallocSize is not 0. 6250 - implemented a version of fastMallocSize for tcmalloc. 6251 6252 Also moved some pre-existing code around to avoid a problem related to 6253 mismatched #define/#undef of malloc/free in this source file. 6254 6255 * JavaScriptCore.exp: 6256 * wtf/FastMalloc.cpp: 6257 (WTF::fastMallocSize): Renamed. Fixed indentation. 6258 6259 (WTF::TCMalloc_PageHeap::scavenge): Removed an incorrect ASSERT that 6260 got in the way of testing the tcmalloc implementation. (More information 6261 on why this ASSERT is incorrect is in <rdar://problem/7165917>.) 6262 6263 (WTF::TCMallocStats::fastMallocSize): Implemented for tcmalloc. 6264 6265 * wtf/FastMalloc.h: Updated for rename. 6266 6267 * wtf/ValueCheck.h: 6268 (WTF::): Moved the ASSERT that used to be in fastCheckConsistency here. 6269 62702010-03-10 Kevin Ollivier <kevino@theolliviers.com> 6271 6272 Reviewed by Eric Seidel. 6273 6274 Make global new/delete operators configurable for all ports and disable it 6275 for the wx port for now. 6276 6277 * wtf/FastMalloc.h: 6278 * wtf/Platform.h: 6279 62802010-03-09 Gavin Barraclough <barraclough@apple.com> 6281 6282 Reviewed by NOBODY (reverting r54510). 6283 6284 This caused a performance regression, by breaking the code 6285 generator's logic to calculate the skip level for resolving 6286 variables (traced by rdar:7683350) Reverting for now. 6287 6288 * parser/Grammar.y: 6289 * parser/NodeConstructors.h: 6290 (JSC::ContinueNode::ContinueNode): 6291 (JSC::BreakNode::BreakNode): 6292 (JSC::ForInNode::ForInNode): 6293 * runtime/CommonIdentifiers.cpp: 6294 (JSC::CommonIdentifiers::CommonIdentifiers): 6295 * runtime/CommonIdentifiers.h: 6296 * runtime/FunctionPrototype.cpp: 6297 (JSC::FunctionPrototype::FunctionPrototype): 6298 * runtime/Identifier.cpp: 6299 (JSC::Identifier::add): 6300 * runtime/PropertyNameArray.cpp: 6301 (JSC::PropertyNameArray::add): 6302 63032010-03-09 Geoffrey Garen <ggaren@apple.com> 6304 6305 Reviewed by Darin Adler. 6306 6307 Changed FastMalloc statistics reporting to be a bit clearer. We now 6308 report: 6309 - Reserved VM Bytes: the VM that has been mapped into the process. 6310 - Committed VM Bytes: the subset of Reserved VM Bytes actually in use. 6311 - Free List Bytes: the subset of Committed VM Bytes in a free list. 6312 6313 * wtf/FastMalloc.cpp: 6314 (WTF::fastMallocStatistics): 6315 (WTF::TCMallocStats::fastMallocStatistics): Updated to report the statistics 6316 above. Standardized use of "ifdef WTF_CHANGES". Added a SpinLockHolder 6317 around all statistics gathering, since it reads from the page heap. 6318 6319 * wtf/FastMalloc.h: Updated to report the statistics above. 6320 63212010-03-09 Gabor Loki <loki@webkit.org> 6322 6323 Rubber-stamped by Maciej Stachowiak. 6324 6325 Buildfix for ARM after r55684. Add branch8 and branchTest8 functions. 6326 https://bugs.webkit.org/show_bug.cgi?id=35892 6327 6328 * assembler/ARMAssembler.cpp: 6329 (JSC::ARMAssembler::dataTransfer32): 6330 * assembler/ARMAssembler.h: 6331 (JSC::ARMAssembler::): 6332 * assembler/MacroAssemblerARM.h: 6333 (JSC::MacroAssemblerARM::load8): 6334 (JSC::MacroAssemblerARM::branch8): 6335 (JSC::MacroAssemblerARM::branchTest8): 6336 63372010-03-08 Geoffrey Garen <ggaren@apple.com> 6338 6339 Windows build fix: 'P' is not a type. Luckily, 'void' is. 6340 6341 * wtf/FastMalloc.cpp: 6342 (WTF::fastCheckConsistency): 6343 63442010-03-08 Geoffrey Garen <ggaren@apple.com> 6345 6346 Windows build fix: export a new symbol. 6347 6348 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 6349 63502010-03-08 Geoffrey Garen <ggaren@apple.com> 6351 6352 Reviewed by Maciej Stachowiak. 6353 6354 Switching malloc implementations requires a world rebuild 6355 https://bugs.webkit.org/show_bug.cgi?id=35899 6356 6357 * wtf/FastMalloc.cpp: 6358 (WTF::fastCheckConsistency): 6359 (WTF::TCMallocStats::fastCheckConsistency): 6360 * wtf/FastMalloc.h: 6361 * wtf/ValueCheck.h: 6362 (WTF::): Moved pointer checking into a helper function in FastMalloc.cpp, 6363 so you can switch malloc implementations without rebuilding the world. 6364 63652010-03-07 Oliver Hunt <oliver@apple.com> 6366 6367 Reviewed by Darin Adler. 6368 6369 TypeInfo is unnecessarily large 6370 https://bugs.webkit.org/show_bug.cgi?id=35850 6371 6372 Reduce the size of the type and flags members to a single 6373 byte each, reducing the size of Structure by 8 bytes. 6374 6375 * assembler/MacroAssemblerX86Common.h: 6376 (JSC::MacroAssemblerX86Common::branch8): 6377 (JSC::MacroAssemblerX86Common::branchTest8): 6378 (JSC::MacroAssemblerX86Common::setTest8): 6379 Add single byte branches, and correct setTest8 to do a 6380 single byte read from memory, and actually store the result 6381 * assembler/X86Assembler.h: 6382 (JSC::X86Assembler::): 6383 (JSC::X86Assembler::cmpb_im): 6384 (JSC::X86Assembler::testb_im): 6385 * jit/JITCall.cpp: 6386 (JSC::JIT::emit_op_construct_verify): 6387 * jit/JITOpcodes.cpp: 6388 (JSC::JIT::emit_op_instanceof): 6389 (JSC::JIT::emit_op_jeq_null): 6390 (JSC::JIT::emit_op_jneq_null): 6391 (JSC::JIT::emit_op_get_pnames): 6392 (JSC::JIT::emit_op_convert_this): 6393 (JSC::JIT::emit_op_construct_verify): 6394 (JSC::JIT::emit_op_to_jsnumber): 6395 (JSC::JIT::emit_op_eq_null): 6396 (JSC::JIT::emit_op_neq_null): 6397 * runtime/JSTypeInfo.h: 6398 (JSC::TypeInfo::TypeInfo): 6399 (JSC::TypeInfo::type): 6400 64012010-03-08 Gavin Barraclough <barraclough@apple.com> 6402 6403 Reviewed by NOBODY (reverting regression). 6404 6405 Reverting 55035, this caused a regression. 6406 (https://bugs.webkit.org/show_bug.cgi?id=35843) 6407 6408 * runtime/JSString.cpp: 6409 (JSC::JSString::resolveRope): 6410 (JSC::JSString::getStringPropertyDescriptor): 6411 * runtime/JSString.h: 6412 (JSC::jsSingleCharacterSubstring): 6413 (JSC::JSString::getIndex): 6414 (JSC::JSString::getStringPropertySlot): 6415 * runtime/UStringImpl.cpp: 6416 * runtime/UStringImpl.h: 6417 64182010-03-08 Stuart Morgan <stuartmorgan@chromium.org> 6419 6420 Reviewed by Darin Adler. 6421 6422 Added a new USE definition for secure text mode on the Mac. 6423 https://bugs.webkit.org/show_bug.cgi?id=31265 6424 6425 * wtf/Platform.h: 6426 64272010-03-08 Jian Li <jianli@chromium.org> 6428 6429 Reviewed by Dmitry Titov. 6430 6431 Blob.slice support. 6432 https://bugs.webkit.org/show_bug.cgi?id=32993 6433 6434 Add ENABLE_BLOB_SLICE feature define. 6435 Also fix a problem that JSValue.toInteger is not exposed on Windows. 6436 6437 * Configurations/FeatureDefines.xcconfig: 6438 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 6439 64402010-03-07 Jedrzej Nowacki <jedrzej.nowacki@nokia.com> 6441 6442 Reviewed by Simon Hausmann. 6443 6444 Small performance fix in the QScriptConverter::toString(). 6445 6446 The QByteArray was replaced by the QVarLengthArray which doesn't 6447 have to allocate any memory on heap. 6448 6449 [Qt] QScriptConverter::toString() should use QVarLengthArray instead of QByteArray 6450 https://bugs.webkit.org/show_bug.cgi?id=35577 6451 6452 * qt/api/qscriptconverter_p.h: 6453 (QScriptConverter::toString): 6454 64552010-03-06 Mark Rowe <mrowe@apple.com> 6456 6457 Rubber-stamped by Sam Weinig. 6458 6459 Remove unnecessary includes of wtf/Platform.h. This is already pulled in by config.h. 6460 6461 * API/APICast.h: 6462 * API/JSCallbackFunction.cpp: 6463 * API/JSContextRef.cpp: 6464 * API/JSObjectRef.cpp: 6465 * API/JSValueRef.cpp: 6466 * assembler/ARMAssembler.h: 6467 * assembler/ARMv7Assembler.h: 6468 * assembler/AbstractMacroAssembler.h: 6469 * assembler/AssemblerBuffer.h: 6470 * assembler/AssemblerBufferWithConstantPool.h: 6471 * assembler/CodeLocation.h: 6472 * assembler/LinkBuffer.h: 6473 * assembler/MIPSAssembler.h: 6474 * assembler/MacroAssembler.h: 6475 * assembler/MacroAssemblerARM.h: 6476 * assembler/MacroAssemblerARMv7.h: 6477 * assembler/MacroAssemblerCodeRef.h: 6478 * assembler/MacroAssemblerMIPS.h: 6479 * assembler/MacroAssemblerX86.h: 6480 * assembler/MacroAssemblerX86Common.h: 6481 * assembler/MacroAssemblerX86_64.h: 6482 * assembler/RepatchBuffer.h: 6483 * assembler/X86Assembler.h: 6484 * jit/JIT.h: 6485 * jit/JITCode.h: 6486 * jit/JITInlineMethods.h: 6487 * jit/JITStubs.h: 6488 * os-win32/stdint.h: 6489 * runtime/JSAPIValueWrapper.h: 6490 * runtime/JSImmediate.h: 6491 * wtf/ASCIICType.h: 6492 * wtf/StdLibExtras.h: 6493 * wtf/VMTags.h: 6494 * yarr/RegexCompiler.h: 6495 * yarr/RegexInterpreter.h: 6496 * yarr/RegexJIT.h: 6497 * yarr/RegexParser.h: 6498 * yarr/RegexPattern.h: 6499 65002010-03-06 Kwang Yul Seo <skyul@company100.net> 6501 6502 Reviewed by Eric Seidel. 6503 6504 [BREWMP] Share OwnPtr. 6505 https://bugs.webkit.org/show_bug.cgi?id=35776 6506 6507 Share OwnPtr implementation with BREW MP and remove OwnPtrBrew. 6508 6509 * wtf/OwnPtrBrew.cpp: Added. 6510 (WTF::deleteOwnedPtr): 6511 * wtf/OwnPtrCommon.h: 6512 * wtf/brew/OwnPtrBrew.cpp: Removed. 6513 * wtf/brew/OwnPtrBrew.h: Removed. 6514 65152010-03-06 Patrick Gansterer <paroga@paroga.com> 6516 6517 Reviewed by Eric Seidel. 6518 6519 Implemented JIT_OPTIMIZE_NATIVE_CALL for WinCE 6520 https://bugs.webkit.org/show_bug.cgi?id=33426 6521 6522 * jit/JITOpcodes.cpp: 6523 (JSC::JIT::privateCompileCTIMachineTrampolines): 6524 65252010-03-05 Oliver Hunt <oliver@apple.com> 6526 6527 Reviewed by NOBODY (build fix). 6528 6529 Add enw exports to windows 6530 6531 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 6532 65332010-03-05 Oliver Hunt <oliver@apple.com> 6534 6535 Reviewed by Gavin Barraclough. 6536 6537 JSC should cache int to Identifier conversion as it does for ordinary strings 6538 https://bugs.webkit.org/show_bug.cgi?id=35814 6539 6540 Make the NumericStrings cache cache unsigned ints in addition to signed. 6541 We keep them separate from the int cache as it both simplifies code, and 6542 also because the unsigned path is exclusive to property access and therefore 6543 seems to have different usage patterns. 6544 6545 The primary trigger for the unsigned to Identifier propertyName conversion 6546 is the construction of array-like objects out of normal objects. Given these 6547 tend to be relative small numbers, and the array-like behaviour lends itself 6548 to sequential values this patch also adds a non-colliding cache for all small 6549 numbers. 6550 6551 * JavaScriptCore.exp: 6552 * runtime/Identifier.cpp: 6553 (JSC::Identifier::from): 6554 * runtime/Identifier.h: 6555 * runtime/NumericStrings.h: 6556 (JSC::NumericStrings::add): 6557 (JSC::NumericStrings::lookup): 6558 (JSC::NumericStrings::lookupSmallString): 6559 65602010-03-03 Oliver Hunt <oliver@apple.com> 6561 6562 Reviewed by Gavin Barraclough. 6563 6564 Allow static property getters to interact with JSCs caching 6565 https://bugs.webkit.org/show_bug.cgi?id=35716 6566 6567 Add new opcodes for handling cached lookup of static value getters. 6568 More or less the same as with JS getters, all that changes is that 6569 instead of calling through a JSFunction we always know that we have 6570 a C function to call. 6571 6572 For the patching routines in the JIT we now need to pass a few 6573 new parameters to allow us to pass enough information to the stub 6574 function to allow us to call the C function correctly. Logically 6575 this shouldn't actually be necessary as all of these functions ignore 6576 the identifier, but removing the ident parameter would require 6577 somewhat involved changes to the way we implement getOwnPropertySlot, 6578 etc. 6579 6580 * bytecode/CodeBlock.cpp: 6581 (JSC::CodeBlock::dump): 6582 (JSC::CodeBlock::derefStructures): 6583 (JSC::CodeBlock::refStructures): 6584 * bytecode/Instruction.h: 6585 (JSC::Instruction::Instruction): 6586 (JSC::Instruction::): 6587 * bytecode/Opcode.h: 6588 * interpreter/Interpreter.cpp: 6589 (JSC::Interpreter::tryCacheGetByID): 6590 (JSC::Interpreter::privateExecute): 6591 * jit/JIT.cpp: 6592 (JSC::JIT::privateCompileMainPass): 6593 * jit/JIT.h: 6594 (JSC::JIT::compileGetByIdProto): 6595 (JSC::JIT::compileGetByIdSelfList): 6596 (JSC::JIT::compileGetByIdProtoList): 6597 (JSC::JIT::compileGetByIdChainList): 6598 (JSC::JIT::compileGetByIdChain): 6599 * jit/JITPropertyAccess.cpp: 6600 (JSC::JIT::privateCompileGetByIdProto): 6601 (JSC::JIT::privateCompileGetByIdSelfList): 6602 (JSC::JIT::privateCompileGetByIdProtoList): 6603 (JSC::JIT::privateCompileGetByIdChainList): 6604 (JSC::JIT::privateCompileGetByIdChain): 6605 * jit/JITPropertyAccess32_64.cpp: 6606 (JSC::JIT::privateCompileGetByIdProto): 6607 (JSC::JIT::privateCompileGetByIdSelfList): 6608 (JSC::JIT::privateCompileGetByIdProtoList): 6609 (JSC::JIT::privateCompileGetByIdChainList): 6610 (JSC::JIT::privateCompileGetByIdChain): 6611 * jit/JITStubs.cpp: 6612 (JSC::JITThunks::tryCacheGetByID): 6613 (JSC::DEFINE_STUB_FUNCTION): 6614 * jit/JITStubs.h: 6615 (JSC::): 6616 * runtime/JSFunction.cpp: 6617 (JSC::JSFunction::getOwnPropertySlot): 6618 * runtime/Lookup.h: 6619 (JSC::getStaticPropertySlot): 6620 (JSC::getStaticValueSlot): 6621 * runtime/PropertySlot.h: 6622 (JSC::PropertySlot::): 6623 (JSC::PropertySlot::PropertySlot): 6624 (JSC::PropertySlot::cachedPropertyType): 6625 (JSC::PropertySlot::isCacheable): 6626 (JSC::PropertySlot::isCacheableValue): 6627 (JSC::PropertySlot::setValueSlot): 6628 (JSC::PropertySlot::setCacheableCustom): 6629 (JSC::PropertySlot::setGetterSlot): 6630 (JSC::PropertySlot::setCacheableGetterSlot): 6631 (JSC::PropertySlot::clearOffset): 6632 (JSC::PropertySlot::customGetter): 6633 66342010-03-04 Shinichiro Hamaji <hamaji@chromium.org> 6635 6636 Unreviewed. Remove a non-ASCII character introduced in the following bug. 6637 6638 put_by_id does will incorrectly cache writes where a specific value exists, where at the point of caching the same value is being written. 6639 https://bugs.webkit.org/show_bug.cgi?id=35537 6640 6641 * runtime/JSObject.h: 6642 (JSC::JSObject::putDirectInternal): 6643 66442010-03-04 Jocelyn Turcotte <jocelyn.turcotte@nokia.com> 6645 6646 Reviewed by Tor Arne Vestbø. 6647 6648 [Qt] Make the OUTPUT_DIR variable in qmake projects independent of build-webkit's logic. 6649 6650 This also allows shadow builds relying only on qmake to work properly. 6651 * jsc.pro: 6652 * qt/api/QtScript.pro: 6653 * qt/tests/qscriptengine/qscriptengine.pro: 6654 * qt/tests/qscriptvalue/qscriptvalue.pro: 6655 * qt/tests/tests.pri: 6656 66572010-03-03 Jedrzej Nowacki <jedrzej.nowacki@nokia.com> 6658 6659 Reviewed by Simon Hausmann. 6660 6661 QScriptValue::isObject fix. 6662 6663 Fix broken internal state evaluation from JSValue to JSNative / JSObject. 6664 New function was introduced which should take care about promoting 6665 JSValue state inside QScriptValuePrivate. It should be used instead of a 6666 direct JSC C API call. 6667 6668 The bug exposed a weakness in autotest suite, as the QScriptValuePrivate 6669 is based on state machine with lazy state evaluation, there is a possibility 6670 that serial sequencial calls to the same public const function could return 6671 different results. The patch fix the issue. 6672 6673 [Qt] Sometimes QScriptValue::isObject returns an incorrect value 6674 https://bugs.webkit.org/show_bug.cgi?id=35387 6675 6676 * qt/api/qscriptvalue_p.h: 6677 (QScriptValuePrivate::isBool): 6678 (QScriptValuePrivate::isNumber): 6679 (QScriptValuePrivate::isNull): 6680 (QScriptValuePrivate::isString): 6681 (QScriptValuePrivate::isUndefined): 6682 (QScriptValuePrivate::isError): 6683 (QScriptValuePrivate::isObject): 6684 (QScriptValuePrivate::isFunction): 6685 (QScriptValuePrivate::call): 6686 (QScriptValuePrivate::refineJSValue): 6687 * qt/tests/qscriptvalue/tst_qscriptvalue_generated.cpp: 6688 (tst_QScriptValue::initScriptValues): 6689 (tst_QScriptValue::isValid_makeData): 6690 (tst_QScriptValue::isValid_test): 6691 (tst_QScriptValue::isBool_makeData): 6692 (tst_QScriptValue::isBool_test): 6693 (tst_QScriptValue::isBoolean_makeData): 6694 (tst_QScriptValue::isBoolean_test): 6695 (tst_QScriptValue::isNumber_makeData): 6696 (tst_QScriptValue::isNumber_test): 6697 (tst_QScriptValue::isFunction_test): 6698 (tst_QScriptValue::isNull_makeData): 6699 (tst_QScriptValue::isNull_test): 6700 (tst_QScriptValue::isString_makeData): 6701 (tst_QScriptValue::isString_test): 6702 (tst_QScriptValue::isUndefined_makeData): 6703 (tst_QScriptValue::isUndefined_test): 6704 (tst_QScriptValue::isObject_makeData): 6705 (tst_QScriptValue::isObject_test): 6706 (tst_QScriptValue::toString_makeData): 6707 (tst_QScriptValue::toString_test): 6708 (tst_QScriptValue::toNumber_makeData): 6709 (tst_QScriptValue::toNumber_test): 6710 (tst_QScriptValue::toBool_makeData): 6711 (tst_QScriptValue::toBool_test): 6712 (tst_QScriptValue::toBoolean_makeData): 6713 (tst_QScriptValue::toBoolean_test): 6714 (tst_QScriptValue::toInteger_makeData): 6715 (tst_QScriptValue::toInteger_test): 6716 (tst_QScriptValue::toInt32_makeData): 6717 (tst_QScriptValue::toInt32_test): 6718 (tst_QScriptValue::toUInt32_makeData): 6719 (tst_QScriptValue::toUInt32_test): 6720 (tst_QScriptValue::toUInt16_makeData): 6721 (tst_QScriptValue::toUInt16_test): 6722 67232010-03-03 Chao-ying Fu <fu@mips.com> 6724 6725 Reviewed by Gavin Barraclough. 6726 6727 MIPS JIT Supports 6728 https://bugs.webkit.org/show_bug.cgi?id=30144 6729 6730 The following changes enable MIPS YARR and YARR_JIT. 6731 6732 * assembler/AbstractMacroAssembler.h: 6733 (JSC::AbstractMacroAssembler::Imm32::Imm32): 6734 * assembler/MIPSAssembler.h: Added. 6735 (JSC::MIPSRegisters::): 6736 (JSC::MIPSAssembler::MIPSAssembler): 6737 (JSC::MIPSAssembler::): 6738 (JSC::MIPSAssembler::JmpSrc::JmpSrc): 6739 (JSC::MIPSAssembler::JmpDst::JmpDst): 6740 (JSC::MIPSAssembler::JmpDst::isUsed): 6741 (JSC::MIPSAssembler::JmpDst::used): 6742 (JSC::MIPSAssembler::emitInst): 6743 (JSC::MIPSAssembler::nop): 6744 (JSC::MIPSAssembler::loadDelayNop): 6745 (JSC::MIPSAssembler::copDelayNop): 6746 (JSC::MIPSAssembler::move): 6747 (JSC::MIPSAssembler::li): 6748 (JSC::MIPSAssembler::lui): 6749 (JSC::MIPSAssembler::addiu): 6750 (JSC::MIPSAssembler::addu): 6751 (JSC::MIPSAssembler::subu): 6752 (JSC::MIPSAssembler::mult): 6753 (JSC::MIPSAssembler::mfhi): 6754 (JSC::MIPSAssembler::mflo): 6755 (JSC::MIPSAssembler::mul): 6756 (JSC::MIPSAssembler::andInsn): 6757 (JSC::MIPSAssembler::andi): 6758 (JSC::MIPSAssembler::nor): 6759 (JSC::MIPSAssembler::orInsn): 6760 (JSC::MIPSAssembler::ori): 6761 (JSC::MIPSAssembler::xorInsn): 6762 (JSC::MIPSAssembler::xori): 6763 (JSC::MIPSAssembler::slt): 6764 (JSC::MIPSAssembler::sltu): 6765 (JSC::MIPSAssembler::sltiu): 6766 (JSC::MIPSAssembler::sll): 6767 (JSC::MIPSAssembler::sllv): 6768 (JSC::MIPSAssembler::sra): 6769 (JSC::MIPSAssembler::srav): 6770 (JSC::MIPSAssembler::lw): 6771 (JSC::MIPSAssembler::lwl): 6772 (JSC::MIPSAssembler::lwr): 6773 (JSC::MIPSAssembler::lhu): 6774 (JSC::MIPSAssembler::sw): 6775 (JSC::MIPSAssembler::jr): 6776 (JSC::MIPSAssembler::jalr): 6777 (JSC::MIPSAssembler::jal): 6778 (JSC::MIPSAssembler::bkpt): 6779 (JSC::MIPSAssembler::bgez): 6780 (JSC::MIPSAssembler::bltz): 6781 (JSC::MIPSAssembler::beq): 6782 (JSC::MIPSAssembler::bne): 6783 (JSC::MIPSAssembler::bc1t): 6784 (JSC::MIPSAssembler::bc1f): 6785 (JSC::MIPSAssembler::newJmpSrc): 6786 (JSC::MIPSAssembler::appendJump): 6787 (JSC::MIPSAssembler::addd): 6788 (JSC::MIPSAssembler::subd): 6789 (JSC::MIPSAssembler::muld): 6790 (JSC::MIPSAssembler::lwc1): 6791 (JSC::MIPSAssembler::ldc1): 6792 (JSC::MIPSAssembler::swc1): 6793 (JSC::MIPSAssembler::sdc1): 6794 (JSC::MIPSAssembler::mtc1): 6795 (JSC::MIPSAssembler::mfc1): 6796 (JSC::MIPSAssembler::truncwd): 6797 (JSC::MIPSAssembler::cvtdw): 6798 (JSC::MIPSAssembler::ceqd): 6799 (JSC::MIPSAssembler::cngtd): 6800 (JSC::MIPSAssembler::cnged): 6801 (JSC::MIPSAssembler::cltd): 6802 (JSC::MIPSAssembler::cled): 6803 (JSC::MIPSAssembler::cueqd): 6804 (JSC::MIPSAssembler::coled): 6805 (JSC::MIPSAssembler::coltd): 6806 (JSC::MIPSAssembler::culed): 6807 (JSC::MIPSAssembler::cultd): 6808 (JSC::MIPSAssembler::label): 6809 (JSC::MIPSAssembler::align): 6810 (JSC::MIPSAssembler::getRelocatedAddress): 6811 (JSC::MIPSAssembler::getDifferenceBetweenLabels): 6812 (JSC::MIPSAssembler::size): 6813 (JSC::MIPSAssembler::executableCopy): 6814 (JSC::MIPSAssembler::getCallReturnOffset): 6815 (JSC::MIPSAssembler::linkJump): 6816 (JSC::MIPSAssembler::linkCall): 6817 (JSC::MIPSAssembler::linkPointer): 6818 (JSC::MIPSAssembler::relinkJump): 6819 (JSC::MIPSAssembler::relinkCall): 6820 (JSC::MIPSAssembler::repatchInt32): 6821 (JSC::MIPSAssembler::repatchPointer): 6822 (JSC::MIPSAssembler::repatchLoadPtrToLEA): 6823 (JSC::MIPSAssembler::relocateJumps): 6824 (JSC::MIPSAssembler::linkWithOffset): 6825 (JSC::MIPSAssembler::linkCallInternal): 6826 * assembler/MacroAssembler.h: 6827 * assembler/MacroAssemblerMIPS.h: Added. 6828 (JSC::MacroAssemblerMIPS::MacroAssemblerMIPS): 6829 (JSC::MacroAssemblerMIPS::): 6830 (JSC::MacroAssemblerMIPS::add32): 6831 (JSC::MacroAssemblerMIPS::and32): 6832 (JSC::MacroAssemblerMIPS::lshift32): 6833 (JSC::MacroAssemblerMIPS::mul32): 6834 (JSC::MacroAssemblerMIPS::not32): 6835 (JSC::MacroAssemblerMIPS::or32): 6836 (JSC::MacroAssemblerMIPS::rshift32): 6837 (JSC::MacroAssemblerMIPS::sub32): 6838 (JSC::MacroAssemblerMIPS::xor32): 6839 (JSC::MacroAssemblerMIPS::load32): 6840 (JSC::MacroAssemblerMIPS::load32WithUnalignedHalfWords): 6841 (JSC::MacroAssemblerMIPS::load32WithAddressOffsetPatch): 6842 (JSC::MacroAssemblerMIPS::loadPtrWithPatchToLEA): 6843 (JSC::MacroAssemblerMIPS::loadPtrWithAddressOffsetPatch): 6844 (JSC::MacroAssemblerMIPS::load16): 6845 (JSC::MacroAssemblerMIPS::store32WithAddressOffsetPatch): 6846 (JSC::MacroAssemblerMIPS::store32): 6847 (JSC::MacroAssemblerMIPS::supportsFloatingPoint): 6848 (JSC::MacroAssemblerMIPS::supportsFloatingPointTruncate): 6849 (JSC::MacroAssemblerMIPS::pop): 6850 (JSC::MacroAssemblerMIPS::push): 6851 (JSC::MacroAssemblerMIPS::move): 6852 (JSC::MacroAssemblerMIPS::swap): 6853 (JSC::MacroAssemblerMIPS::signExtend32ToPtr): 6854 (JSC::MacroAssemblerMIPS::zeroExtend32ToPtr): 6855 (JSC::MacroAssemblerMIPS::branch32): 6856 (JSC::MacroAssemblerMIPS::branch32WithUnalignedHalfWords): 6857 (JSC::MacroAssemblerMIPS::branch16): 6858 (JSC::MacroAssemblerMIPS::branchTest32): 6859 (JSC::MacroAssemblerMIPS::jump): 6860 (JSC::MacroAssemblerMIPS::branchAdd32): 6861 (JSC::MacroAssemblerMIPS::branchMul32): 6862 (JSC::MacroAssemblerMIPS::branchSub32): 6863 (JSC::MacroAssemblerMIPS::breakpoint): 6864 (JSC::MacroAssemblerMIPS::nearCall): 6865 (JSC::MacroAssemblerMIPS::call): 6866 (JSC::MacroAssemblerMIPS::ret): 6867 (JSC::MacroAssemblerMIPS::set32): 6868 (JSC::MacroAssemblerMIPS::setTest32): 6869 (JSC::MacroAssemblerMIPS::moveWithPatch): 6870 (JSC::MacroAssemblerMIPS::branchPtrWithPatch): 6871 (JSC::MacroAssemblerMIPS::storePtrWithPatch): 6872 (JSC::MacroAssemblerMIPS::tailRecursiveCall): 6873 (JSC::MacroAssemblerMIPS::makeTailRecursiveCall): 6874 (JSC::MacroAssemblerMIPS::loadDouble): 6875 (JSC::MacroAssemblerMIPS::storeDouble): 6876 (JSC::MacroAssemblerMIPS::addDouble): 6877 (JSC::MacroAssemblerMIPS::subDouble): 6878 (JSC::MacroAssemblerMIPS::mulDouble): 6879 (JSC::MacroAssemblerMIPS::convertInt32ToDouble): 6880 (JSC::MacroAssemblerMIPS::insertRelaxationWords): 6881 (JSC::MacroAssemblerMIPS::branchTrue): 6882 (JSC::MacroAssemblerMIPS::branchFalse): 6883 (JSC::MacroAssemblerMIPS::branchEqual): 6884 (JSC::MacroAssemblerMIPS::branchNotEqual): 6885 (JSC::MacroAssemblerMIPS::branchDouble): 6886 (JSC::MacroAssemblerMIPS::branchTruncateDoubleToInt32): 6887 (JSC::MacroAssemblerMIPS::linkCall): 6888 (JSC::MacroAssemblerMIPS::repatchCall): 6889 * jit/ExecutableAllocator.h: 6890 (JSC::ExecutableAllocator::cacheFlush): 6891 * wtf/Platform.h: 6892 * yarr/RegexJIT.cpp: 6893 (JSC::Yarr::RegexGenerator::generateEnter): 6894 (JSC::Yarr::RegexGenerator::generateReturn): 6895 68962010-03-03 Steve Falkenburg <sfalken@apple.com> 6897 6898 Windows build fix. 6899 6900 * JavaScriptCore.vcproj/jsc/jsc.vcproj: 6901 * JavaScriptCore.vcproj/jsc/jscCommon.vsprops: 6902 69032010-03-03 Steve Falkenburg <sfalken@apple.com> 6904 6905 Windows build fix. 6906 6907 * JavaScriptCore.vcproj/jsc/jsc.vcproj: 6908 69092010-03-03 Mark Rowe <mrowe@apple.com> 6910 6911 Reviewed by Geoff Garen. 6912 6913 Add virtual memory tags for TCMalloc and WebCore's purgeable buffers. 6914 6915 * wtf/TCSystemAlloc.cpp: 6916 (TryMmap): Use the VM tag. 6917 * wtf/VMTags.h: Make use of VM_MEMORY_TCMALLOC and VM_MEMORY_WEBCORE_PURGEABLE_BUFFERS. 6918 69192010-03-03 Steve Falkenburg <sfalken@apple.com> 6920 6921 Rubber stamped by Adam Roben. 6922 6923 Fix bogus xcopy that was polluting source tree at build time. 6924 6925 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCoreCommon.vsprops: 6926 69272010-03-02 Fridrich Strba <fridrich.strba@bluewin.ch> 6928 6929 Reviewed by Oliver Hunt. 6930 6931 Allow building smoothly on win32 and win64 using GCC 6932 https://bugs.webkit.org/show_bug.cgi?id=35607 6933 6934 * jit/JITStubs.h: 6935 * runtime/Collector.cpp: 6936 (JSC::Heap::allocateBlock): 6937 (JSC::Heap::freeBlockPtr): 6938 (JSC::currentThreadStackBase): 6939 69402010-03-02 Jeremy Orlow <jorlow@chromium.org> 6941 6942 Reviewed by David Levin. 6943 6944 Revert database thread changes that are no longer required 6945 https://bugs.webkit.org/show_bug.cgi?id=35519 6946 6947 Jochen Eisinger created 55214 and 55247 to track which database 6948 owns which thread. Dmitry suggested that this could also 6949 be done via TLS, though. After exploring the options, Jochen 6950 chose to go the TLS route, so these patches are no longer needed. 6951 6952 * wtf/Threading.h: 6953 * wtf/ThreadingNone.cpp: 6954 (WTF::isMainThread): 6955 * wtf/ThreadingPthreads.cpp: 6956 (WTF::identifierByPthreadHandle): 6957 (WTF::establishIdentifierForPthreadHandle): 6958 (WTF::pthreadHandleForIdentifier): 6959 (WTF::createThreadInternal): 6960 (WTF::currentThread): 6961 * wtf/ThreadingWin.cpp: 6962 (WTF::threadMap): 6963 (WTF::storeThreadHandleByIdentifier): 6964 (WTF::threadHandleForIdentifier): 6965 (WTF::createThreadInternal): 6966 69672010-03-02 Jedrzej Nowacki <jedrzej.nowacki@nokia.com> 6968 6969 Reviewed by Simon Hausmann. 6970 6971 Fix QScriptValue::toString(). 6972 6973 More ECMA Script compliance, especially for values as NaN, Inifinite 6974 and really big/small numbers. 6975 6976 [Qt] QScriptValue::toString() returns incorrect values 6977 https://bugs.webkit.org/show_bug.cgi?id=34850 6978 6979 * qt/api/qscriptconverter_p.h: 6980 (QScriptConverter::toString): 6981 * qt/api/qscriptvalue_p.h: 6982 (QScriptValuePrivate::toString): 6983 * qt/tests/qscriptvalue/tst_qscriptvalue.cpp: 6984 * qt/tests/qscriptvalue/tst_qscriptvalue.h: 6985 * qt/tests/qscriptvalue/tst_qscriptvalue_generated.cpp: 6986 (tst_QScriptValue::toString_initData): 6987 (tst_QScriptValue::toString_makeData): 6988 (tst_QScriptValue::toString_test): 6989 69902010-03-02 Jedrzej Nowacki <jedrzej.nowacki@nokia.com> 6991 6992 Reviewed by Simon Hausmann. 6993 6994 Introduce a new class; QScriptString. 6995 6996 The QScriptString class should act as a handle to "interned" 6997 strings in a QScriptEngine. 6998 6999 [Qt] QtScript should provide QScriptString 7000 https://bugs.webkit.org/show_bug.cgi?id=34843 7001 7002 * qt/api/QtScript.pro: 7003 * qt/api/qscriptengine.cpp: 7004 (QScriptEngine::toStringHandle): 7005 * qt/api/qscriptengine.h: 7006 * qt/api/qscriptengine_p.h: 7007 (QScriptEnginePrivate::toStringHandle): 7008 * qt/api/qscriptstring.cpp: Added. 7009 (QScriptString::QScriptString): 7010 (QScriptString::~QScriptString): 7011 (QScriptString::operator=): 7012 (QScriptString::isValid): 7013 (QScriptString::operator==): 7014 (QScriptString::operator!=): 7015 (QScriptString::toArrayIndex): 7016 (QScriptString::toString): 7017 (QScriptString::operator QString): 7018 (qHash): 7019 * qt/api/qscriptstring.h: Added. 7020 * qt/api/qscriptstring_p.h: Added. 7021 (QScriptStringPrivate::QScriptStringPrivate): 7022 (QScriptStringPrivate::~QScriptStringPrivate): 7023 (QScriptStringPrivate::get): 7024 (QScriptStringPrivate::isValid): 7025 (QScriptStringPrivate::operator==): 7026 (QScriptStringPrivate::operator!=): 7027 (QScriptStringPrivate::toArrayIndex): 7028 (QScriptStringPrivate::toString): 7029 (QScriptStringPrivate::id): 7030 * qt/tests/qscriptstring/qscriptstring.pro: Added. 7031 * qt/tests/qscriptstring/tst_qscriptstring.cpp: Added. 7032 (tst_QScriptString::tst_QScriptString): 7033 (tst_QScriptString::~tst_QScriptString): 7034 (tst_QScriptString::test): 7035 (tst_QScriptString::hash): 7036 (tst_QScriptString::toArrayIndex_data): 7037 (tst_QScriptString::toArrayIndex): 7038 * qt/tests/tests.pro: 7039 70402010-03-02 Oliver Hunt <oliver@apple.com> 7041 7042 Reviewed by NOBODY (Build fix). 7043 7044 Export function on windows. 7045 7046 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 7047 70482010-03-01 Oliver Hunt <oliver@apple.com> 7049 7050 Reviewed by Maciej Stachowiak. 7051 7052 Refactor named getter function signature to be in line with indexing getter signature 7053 https://bugs.webkit.org/show_bug.cgi?id=35563 7054 7055 This removes the PropertySlot argument from getter functions, and makes them directly 7056 pass the slot base. This makes the semantics for the functions match that of the 7057 indexing getters. 7058 7059 On the down side, this means that we can no longer simply use a proxy function for 7060 JS getters, so we now add another marker value to indicate that a getter is present 7061 and branch accordingly. 7062 7063 Against all rationality sunspider reports this as a perf win, but i suspect it's just noise. 7064 7065 * API/JSCallbackObject.h: 7066 * API/JSCallbackObjectFunctions.h: 7067 (JSC::::staticValueGetter): 7068 (JSC::::staticFunctionGetter): 7069 (JSC::::callbackGetter): 7070 * JavaScriptCore.exp: 7071 * runtime/JSActivation.cpp: 7072 (JSC::JSActivation::argumentsGetter): 7073 * runtime/JSActivation.h: 7074 * runtime/JSFunction.cpp: 7075 (JSC::JSFunction::argumentsGetter): 7076 (JSC::JSFunction::callerGetter): 7077 (JSC::JSFunction::lengthGetter): 7078 * runtime/JSFunction.h: 7079 * runtime/NumberConstructor.cpp: 7080 (JSC::numberConstructorNaNValue): 7081 (JSC::numberConstructorNegInfinity): 7082 (JSC::numberConstructorPosInfinity): 7083 (JSC::numberConstructorMaxValue): 7084 (JSC::numberConstructorMinValue): 7085 * runtime/PropertySlot.cpp: 7086 (JSC::PropertySlot::functionGetter): 7087 * runtime/PropertySlot.h: 7088 (JSC::PropertySlot::getValue): 7089 (JSC::PropertySlot::setGetterSlot): 7090 (JSC::PropertySlot::setCacheableGetterSlot): 7091 * runtime/RegExpConstructor.cpp: 7092 (JSC::regExpConstructorDollar1): 7093 (JSC::regExpConstructorDollar2): 7094 (JSC::regExpConstructorDollar3): 7095 (JSC::regExpConstructorDollar4): 7096 (JSC::regExpConstructorDollar5): 7097 (JSC::regExpConstructorDollar6): 7098 (JSC::regExpConstructorDollar7): 7099 (JSC::regExpConstructorDollar8): 7100 (JSC::regExpConstructorDollar9): 7101 (JSC::regExpConstructorInput): 7102 (JSC::regExpConstructorMultiline): 7103 (JSC::regExpConstructorLastMatch): 7104 (JSC::regExpConstructorLastParen): 7105 (JSC::regExpConstructorLeftContext): 7106 (JSC::regExpConstructorRightContext): 7107 * runtime/RegExpObject.cpp: 7108 (JSC::regExpObjectGlobal): 7109 (JSC::regExpObjectIgnoreCase): 7110 (JSC::regExpObjectMultiline): 7111 (JSC::regExpObjectSource): 7112 (JSC::regExpObjectLastIndex): 7113 71142010-03-01 Oliver Hunt <oliver@apple.com> 7115 7116 Reviewed by Gavin Barraclough. 7117 7118 PropertySlot::getValue(ExecState, unsigned) unnecessarily converts index to an Identifier 7119 https://bugs.webkit.org/show_bug.cgi?id=35561 7120 7121 Fix this by defining a separate property getter function for index getters. This allows 7122 us to pass an unsigned number without the conversion to an Identifier. We then update 7123 setCustomIndex to take this new getter type. 7124 7125 * runtime/PropertySlot.h: 7126 (JSC::PropertySlot::getValue): 7127 (JSC::PropertySlot::setCustom): 7128 (JSC::PropertySlot::setCustomIndex): 7129 71302010-03-01 Gavin Barraclough <barraclough@apple.com> 7131 7132 Reviewed by Oliver Hunt. 7133 7134 Bug 35537 - put_by_id does will incorrectly cache writes where a specific value exists, 7135 where at the point of caching the same value is being written. 7136 7137 When performing a put_by_id that is replacing a property already present on the object, 7138 there are three interesting cases regarding the state of the specific value: 7139 7140 (1) No specific value set - nothing to do, leave the structure in it's current state, 7141 can cache. 7142 (2) A specific value was set, the new put is not of a specified value (i.e. function), 7143 or is of a different specific value - in these cases we need to perform a despecifying 7144 transition to clear the specific value in the structure, but having done so this is a 7145 normal property so as such we can again cache normally. 7146 (3) A specific value was set, and we are overwriting with the same value - in these cases 7147 leave the structure unchanged, but since a specific value is set we cannot cache this 7148 put (we would need the JIT to dynamically check the value being written matched). 7149 7150 Unfortunately, the current behaviour does not match this. the checks for a specific value 7151 being present & the value matching are combined in such a way that in case (2), above we 7152 will unnecessarily prevent the transition being cached, but in case (3) we will incorrectly 7153 fail to prevent caching. 7154 7155 The bug exposes itself if multiple puts of the same specific value are performed to a 7156 property, and erroneously the put is allowed to be cached by the JIT. Method checks may be 7157 generated caching calls of this structure. Subsequent puts performed from JIT code may 7158 write different values without triggering a despecify transition, and as such cached method 7159 checks will continue to pass, despite the value having changed. 7160 7161 * runtime/JSObject.h: 7162 (JSC::JSObject::putDirectInternal): 7163 71642010-03-01 Tor Arne Vestbø <tor.arne.vestbo@nokia.com> 7165 7166 Reviewed by Simon Hausmann. 7167 7168 Fix the Qt build on Mac OS X/Cocoa 64-bit 7169 7170 * JavaScriptCore.pri: Add missing implementation file to resolve JSC symbols 7171 71722010-02-26 Gavin Barraclough <barraclough@apple.com> 7173 7174 Rubber Stamped by Geoff Garen. 7175 7176 Remove wrec. All builds should have switched to yarr by now. 7177 7178 * Android.mk: 7179 * GNUmakefile.am: 7180 * JavaScriptCore.gypi: 7181 * JavaScriptCore.pri: 7182 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: 7183 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCoreCommon.vsprops: 7184 * JavaScriptCore.vcproj/jsc/jscCommon.vsprops: 7185 * JavaScriptCore.xcodeproj/project.pbxproj: 7186 * runtime/RegExp.cpp: 7187 (JSC::RegExp::match): 7188 * runtime/RegExp.h: 7189 * wrec: Removed. 7190 * wrec/CharacterClass.cpp: Removed. 7191 * wrec/CharacterClass.h: Removed. 7192 * wrec/CharacterClassConstructor.cpp: Removed. 7193 * wrec/CharacterClassConstructor.h: Removed. 7194 * wrec/Escapes.h: Removed. 7195 * wrec/Quantifier.h: Removed. 7196 * wrec/WREC.cpp: Removed. 7197 * wrec/WREC.h: Removed. 7198 * wrec/WRECFunctors.cpp: Removed. 7199 * wrec/WRECFunctors.h: Removed. 7200 * wrec/WRECGenerator.cpp: Removed. 7201 * wrec/WRECGenerator.h: Removed. 7202 * wrec/WRECParser.cpp: Removed. 7203 * wrec/WRECParser.h: Removed. 7204 * wscript: 7205 72062010-02-26 Oliver Hunt <oliver@apple.com> 7207 7208 Reviewed by Geoff Garen. 7209 7210 Make the lookup table generator include an explicit cast to expected 7211 type of the function. We do this because otherwise the blind intptr_t 7212 cast that is subsequently applied allows incorrectly typed functions 7213 to be inserted into the table, where they will only fail at runtime. 7214 This change makes such errors produce a compile time failure. 7215 7216 * create_hash_table: 7217 72182010-02-26 Janne Koskinen <janne.p.koskinen@digia.com> 7219 7220 Reviewed by Simon Hausmann. 7221 7222 [Qt] Symbian specific getCPUTime implemetation 7223 https://bugs.webkit.org/show_bug.cgi?id=34742 7224 7225 Default implementation doesn't work on Symbian devices. 7226 This change adds a proper implementation by 7227 asking thread execution time from the current thread. 7228 7229 * runtime/TimeoutChecker.cpp: 7230 (JSC::getCPUTime): 7231 72322010-02-25 Alexey Proskuryakov <ap@apple.com> 7233 7234 Reviewed by Anders Carlsson. 7235 7236 https://bugs.webkit.org/show_bug.cgi?id=35406 7237 <rdar://problem/6945502> Make generic array methods work with JavaArray 7238 7239 Renamed lazyCreationData to subclassData. This is extra data that can be used by JSArray 7240 subclasses (you can't add new data members, because it wouldn't fit in JSCell otherwise). 7241 7242 * JavaScriptCore.exp: 7243 * runtime/JSArray.cpp: 7244 (JSC::JSArray::JSArray): 7245 (JSC::JSArray::subclassData): 7246 (JSC::JSArray::setSubclassData): 7247 * runtime/JSArray.h: 7248 * runtime/RegExpConstructor.cpp: 7249 (JSC::RegExpMatchesArray::RegExpMatchesArray): 7250 (JSC::RegExpMatchesArray::~RegExpMatchesArray): 7251 (JSC::RegExpMatchesArray::fillArrayInstance): 7252 * runtime/RegExpMatchesArray.h: 7253 (JSC::RegExpMatchesArray::getOwnPropertySlot): 7254 (JSC::RegExpMatchesArray::getOwnPropertyDescriptor): 7255 (JSC::RegExpMatchesArray::put): 7256 (JSC::RegExpMatchesArray::deleteProperty): 7257 (JSC::RegExpMatchesArray::getOwnPropertyNames): 7258 72592010-02-25 Oliver Hunt <oliver@apple.com> 7260 7261 Reviewed by Geoff Garen. 7262 7263 JSC crashes like crazy in the JSPropertyNameIterator destructor 7264 7265 Add back null check of m_cachedStructure. Curse last minute changes. 7266 7267 * runtime/JSPropertyNameIterator.cpp: 7268 (JSC::JSPropertyNameIterator::~JSPropertyNameIterator): 7269 72702010-02-25 Oliver Hunt <oliver@apple.com> 7271 7272 Reviewed by Maciej Stachowiak. 7273 7274 Race condition in JSPropertyNameIterator and Structure destruction 7275 https://bugs.webkit.org/show_bug.cgi?id=35398 7276 7277 JSPropertyNameIterator and Structure have a cyclic dependency that they 7278 manage by clearing the appropriate reference in each other during their 7279 destruction. However if the Structure is destroyed while the 7280 JSPropertyNameIterator is dead but not yet finalized the Structures 7281 WeakGCPtr will return null, and so prevent Structure from clearing 7282 the m_cachedStructure pointer of the iterator. When the iterator is 7283 then finalised the m_cachedStructure is invalid, and the attempt to 7284 clear the structures back reference fails. 7285 7286 To fix this we simply make JSPropertyNameIterator keep the Structure 7287 alive, using the weak pointer to break the ref cycle. 7288 7289 * runtime/JSPropertyNameIterator.cpp: 7290 (JSC::JSPropertyNameIterator::~JSPropertyNameIterator): 7291 The iterator now keeps m_cachedStructure alive itself, so no longer needs 7292 to check for it being cleared 7293 * runtime/JSPropertyNameIterator.h: 7294 (JSC::JSPropertyNameIterator::setCachedStructure): 7295 Add an assertion to ensure correct usage 7296 (JSC::JSPropertyNameIterator::cachedStructure): 7297 Add .get() 7298 * runtime/Structure.cpp: 7299 (JSC::Structure::~Structure): 7300 Add an assertion that our iterator isn't already dead, and remove 7301 the now unnecessary attempt to clear the ref in the iterator 7302 * runtime/WeakGCPtr.h: 7303 (JSC::WeakGCPtr::hasDeadObject): 7304 An assert-only function to allow us to assert correct behaviour 7305 in the Structure destructor 7306 73072010-02-25 Jochen Eisinger <jochen@chromium.org> 7308 7309 Reviewed by Jeremy Orlow. 7310 7311 Make the context that was passed to the ThreadFunction accessible. 7312 https://bugs.webkit.org/show_bug.cgi?id=35379 7313 7314 When a database is opened, right now you 7315 don't have any context from where it is opened. The problem is that 7316 the actual calls that open a database go through the sqlite3 vfs 7317 layer, so there's no easy way to pass this function down to to 7318 platform/sql/chromium/SQLFileSystemChromium*.cpp 7319 7320 This patch will allow you to get from anywhere within webkit a pointer 7321 to the Thread object that actually created the thread you're currently 7322 on (in case of the database, this can be either a thread forked of 7323 from the main thread or from a worker thread), and query the object 7324 for context information. 7325 7326 * wtf/Threading.h: 7327 * wtf/ThreadingNone.cpp: 7328 (WTF::threadContext): 7329 * wtf/ThreadingPthreads.cpp: 7330 (WTF::): 7331 (WTF::identifierByPthreadHandle): 7332 (WTF::establishIdentifierForPthreadHandle): 7333 (WTF::pthreadHandleForIdentifier): 7334 (WTF::contextForIdentifier): 7335 (WTF::createThreadInternal): 7336 (WTF::currentThread): 7337 (WTF::threadContext): 7338 * wtf/ThreadingWin.cpp: 7339 (WTF::): 7340 (WTF::threadMap): 7341 (WTF::storeThreadHandleByIdentifier): 7342 (WTF::threadHandleForIdentifier): 7343 (WTF::contextForIdentifier): 7344 (WTF::createThreadInternal): 7345 (WTF::threadContext): 7346 73472010-02-25 Jeremy Orlow <jorlow@chromium.org> 7348 7349 Reverting to re-submit with better change log. 7350 7351 * wtf/Threading.h: 7352 * wtf/ThreadingNone.cpp: 7353 (WTF::isMainThread): 7354 * wtf/ThreadingPthreads.cpp: 7355 (WTF::identifierByPthreadHandle): 7356 (WTF::establishIdentifierForPthreadHandle): 7357 (WTF::pthreadHandleForIdentifier): 7358 (WTF::createThreadInternal): 7359 (WTF::currentThread): 7360 * wtf/ThreadingWin.cpp: 7361 (WTF::threadMap): 7362 (WTF::storeThreadHandleByIdentifier): 7363 (WTF::threadHandleForIdentifier): 7364 (WTF::createThreadInternal): 7365 73662010-02-25 Jochen Eisinger <jochen@chromium.org> 7367 7368 Reviewed by Jeremy Orlow. 7369 7370 Make the context that was passed to the ThreadFunction accessible. 7371 https://bugs.webkit.org/show_bug.cgi?id=35379 7372 7373 * wtf/Threading.h: 7374 * wtf/ThreadingNone.cpp: 7375 (WTF::threadContext): 7376 * wtf/ThreadingPthreads.cpp: 7377 (WTF::): 7378 (WTF::identifierByPthreadHandle): 7379 (WTF::establishIdentifierForPthreadHandle): 7380 (WTF::pthreadHandleForIdentifier): 7381 (WTF::contextForIdentifier): 7382 (WTF::createThreadInternal): 7383 (WTF::currentThread): 7384 (WTF::threadContext): 7385 * wtf/ThreadingWin.cpp: 7386 (WTF::): 7387 (WTF::threadMap): 7388 (WTF::storeThreadHandleByIdentifier): 7389 (WTF::threadHandleForIdentifier): 7390 (WTF::contextForIdentifier): 7391 (WTF::createThreadInternal): 7392 (WTF::threadContext): 7393 73942010-02-24 Oliver Hunt <oliver@apple.com> 7395 7396 Reviewed by Geoffrey Garen. 7397 7398 [REGRESSION in r55185] EXC_BAD_ACCESS on opening inspector. 7399 https://bugs.webkit.org/show_bug.cgi?id=35335 7400 7401 compileGetDirectOffset modifies the contents of the object register 7402 when the object is not using the inline storage array. As the object 7403 register contains our 'this' pointer we can't allow it to be clobbered. 7404 The fix is simply to copy the register into a separate scratch register 7405 when we're loading off an object that doesn't use inline storage. 7406 7407 * jit/JITPropertyAccess.cpp: 7408 (JSC::JIT::privateCompileGetByIdSelfList): 7409 * jit/JITPropertyAccess32_64.cpp: 7410 (JSC::JIT::privateCompileGetByIdSelfList): 7411 74122010-02-24 Oliver Hunt <oliver@apple.com> 7413 7414 Reviewed by Gavin Barraclough. 7415 7416 Speed up getter performance in the jit 7417 https://bugs.webkit.org/show_bug.cgi?id=35332 7418 7419 Implement getter lookup caching in the interpreter. 7420 The getter stubs are generated through basically the 7421 same code paths as the normal get_by_id caching. 7422 Instead of simply loading a property and returning, 7423 we load the getter slot, and pass the getter, base value 7424 and return address to a shared stub used for getter 7425 dispatch. 7426 7427 * jit/JIT.h: 7428 (JSC::JIT::compileGetByIdProto): 7429 (JSC::JIT::compileGetByIdSelfList): 7430 (JSC::JIT::compileGetByIdProtoList): 7431 (JSC::JIT::compileGetByIdChainList): 7432 (JSC::JIT::compileGetByIdChain): 7433 * jit/JITPropertyAccess.cpp: 7434 (JSC::JIT::privateCompileGetByIdProto): 7435 (JSC::JIT::privateCompileGetByIdSelfList): 7436 (JSC::JIT::privateCompileGetByIdProtoList): 7437 (JSC::JIT::privateCompileGetByIdChainList): 7438 (JSC::JIT::privateCompileGetByIdChain): 7439 * jit/JITPropertyAccess32_64.cpp: 7440 (JSC::JIT::privateCompileGetByIdProto): 7441 (JSC::JIT::privateCompileGetByIdSelfList): 7442 (JSC::JIT::privateCompileGetByIdProtoList): 7443 (JSC::JIT::privateCompileGetByIdChainList): 7444 (JSC::JIT::privateCompileGetByIdChain): 7445 * jit/JITStubs.cpp: 7446 (JSC::JITThunks::tryCacheGetByID): 7447 (JSC::DEFINE_STUB_FUNCTION): 7448 * jit/JITStubs.h: 7449 (JSC::): 7450 * runtime/GetterSetter.h: 7451 74522010-02-23 Oliver Hunt <oliver@apple.com> 7453 7454 Reviewed by Maciej Stachowiak. 7455 7456 Web Inspector: Regression: r55027+: Inspector broken 7457 https://bugs.webkit.org/show_bug.cgi?id=35253 7458 7459 op_get_by_id_getter_chain was not passing the correct this parameter. 7460 The bug was caused by incorrect use of baseCell instead of baseValue, 7461 baseValue contains the original object for the lookup (and hence the 7462 correct this object), baseCell is clobbered as part of walking the 7463 prototype chain. 7464 7465 * interpreter/Interpreter.cpp: 7466 (JSC::Interpreter::privateExecute): 7467 74682010-02-23 Gustavo Noronha Silva <gns@gnome.org> 7469 7470 Rubber-stamped by Dimitri Glazkov. 7471 7472 Chromium build fix. 7473 7474 * JavaScriptCore.gyp/JavaScriptCore.gyp: 7475 74762010-02-23 Leandro Pereira <leandro@profusion.mobi> 7477 7478 Reviewed by Gustavo Noronha Silva. 7479 7480 Changes references of GOwnPtr to reflect their new place. 7481 http://webkit.org/b/35084 7482 7483 * JavaScriptCore/JavaScriptCore.gypi: 7484 * JavaScriptCore/wtf/Threading.h: 7485 * JavaScriptCore/wtf/unicode/glib/UnicodeGLib.h: 7486 74872010-02-23 Leandro Pereira <leandro@profusion.mobi> 7488 7489 Reviewed by Kenneth Rohde Christiansen. 7490 7491 Adding the EFL implementation of JavaScriptCore. 7492 See https://bugs.webkit.org/show_bug.cgi?id=35084 for details. 7493 7494 * GNUmakefile.am: Updated to reflect the new location of GOwnPtr and 7495 GRefPtr. 7496 * wtf/efl/MainThreadEfl.cpp: Added. 7497 * wtf/gobject/GOwnPtr.cpp: Moved from wtf/gtk. 7498 * wtf/gobject/GOwnPtr.h: Moved from wtf/gtk. 7499 * wtf/gobject/GRefPtr.cpp: Moved from wtf/gtk. 7500 * wtf/gobject/GRefPtr.h: Moved from wtf/gtk. 7501 75022010-02-22 Julien Chaffraix <jchaffraix@webkit.org> 7503 7504 Reviewed by Darin Adler. 7505 7506 Remove auto_ptr usage in JavaScriptCore. 7507 https://bugs.webkit.org/show_bug.cgi?id=35221 7508 7509 * parser/Nodes.h: Removed now unneeded adopt method. 7510 * parser/Parser.cpp: Removed <memory> include as it is not required anymore. 7511 * wtf/OwnPtr.h: Removed the constructor from auto_ptr. 7512 * wtf/VectorTraits.h: Removed a template specialization for auto_ptr. 7513 * wtf/unicode/Collator.h: Made userDefault return a PassOwnPtr. 7514 * wtf/unicode/CollatorDefault.cpp: 7515 (WTF::Collator::userDefault): Changed the method to match the next signature. 7516 * wtf/unicode/icu/CollatorICU.cpp: 7517 (WTF::Collator::userDefault): Ditto. 7518 75192010-02-22 Huahui Wu <hwu@google.com> 7520 7521 Reviewed by Eric Seidel. 7522 7523 Add code that enables SquirrelFish Extreme (a.k.a JSCX, JSC JIT) 7524 in Android. It's disabled by default, but is enabled when the 7525 enveronment variable ENABLE_JSC_JIT is set to true. 7526 https://bugs.webkit.org/show_bug.cgi?id=34855 7527 7528 * Android.mk: 7529 * wtf/Platform.h: 7530 75312010-02-22 Gavin Barraclough <barraclough@apple.com> 7532 7533 Reviewed by Oliver Hunt. 7534 7535 JSStringBuilder should not CRASH if allocation fails, it should throw a JSException. 7536 7537 * runtime/JSGlobalObjectFunctions.cpp: 7538 * runtime/JSStringBuilder.h: 7539 (JSC::JSStringBuilder::JSStringBuilder): 7540 (JSC::JSStringBuilder::append): 7541 (JSC::JSStringBuilder::build): 7542 * runtime/StringBuilder.h: 7543 (JSC::StringBuilder::build): 7544 * wtf/Vector.h: 7545 (WTF::VectorBufferBase::tryAllocateBuffer): 7546 (WTF::): 7547 (WTF::VectorBuffer::tryAllocateBuffer): 7548 (WTF::::tryExpandCapacity): 7549 (WTF::::tryReserveCapacity): 7550 (WTF::::tryAppend): 7551 75522010-02-22 Kwang Yul Seo <skyul@company100.net> 7553 7554 Reviewed by Eric Seidel. 7555 7556 [BREWMP] Map FastMalloc to BREW memory allocator 7557 https://bugs.webkit.org/show_bug.cgi?id=33570 7558 7559 Use MALLOC macro instead of the standard malloc function. 7560 Although RVCT provides malloc, we can't use it in BREW 7561 because the loader does not initialize the base address properly. 7562 7563 * wtf/FastMalloc.cpp: 7564 * wtf/brew/SystemMallocBrew.h: Added. 7565 (mallocBrew): 7566 (callocBrew): 7567 (freeBrew): 7568 (reallocBrew): 7569 75702010-02-22 Gustavo Noronha Silva <gustavo.noronha@collabora.co.uk> 7571 7572 Build fix for make distcheck. 7573 7574 * GNUmakefile.am: 7575 75762010-02-22 Laszlo Gombos <laszlo.1.gombos@nokia.com> 7577 7578 Unreviewed build fix. 7579 7580 [Qt] Build fix for RVCT. 7581 7582 Fix after r55024. The "-i" option is for perl not for the 7583 script. 7584 7585 * DerivedSources.pro: 7586 75872010-02-21 Gavin Barraclough <barraclough@apple.com> 7588 7589 Reviewed by Oliver Hunt. 7590 7591 Make UString::m_data be const, and make the UChar owned/ref-counted by CrossThreadRefCounted be const too. 7592 7593 * runtime/UStringImpl.cpp: 7594 (JSC::UStringImpl::baseSharedBuffer): 7595 (JSC::UStringImpl::~UStringImpl): 7596 * runtime/UStringImpl.h: 7597 (JSC::UStringImpl::create): 7598 (JSC::UStringImpl::data): 7599 (JSC::UStringImpl::UStringImpl): 7600 * wtf/OwnFastMallocPtr.h: 7601 (WTF::OwnFastMallocPtr::~OwnFastMallocPtr): 7602 76032010-02-21 Yuta Kitamura <yutak@chromium.org> 7604 7605 Reviewed by Darin Adler. 7606 7607 HashMapTranslatorAdapter::translate() needs to set the mapped value. 7608 7609 HTTPHeaderMap::add(const char*, const String&) does not work 7610 https://bugs.webkit.org/show_bug.cgi?id=35227 7611 7612 * wtf/HashMap.h: 7613 (WTF::HashMapTranslatorAdapter::translate): 7614 76152010-02-19 Maciej Stachowiak <mjs@apple.com> 7616 7617 Reviewed by David Levin. 7618 7619 Add an ENABLE flag for sandboxed iframes to make it possible to disable it in releases 7620 https://bugs.webkit.org/show_bug.cgi?id=35147 7621 7622 * Configurations/FeatureDefines.xcconfig: 7623 76242010-02-19 Gavin Barraclough <barraclough@apple.com> 7625 7626 Reviewed by Oliver Hunt. 7627 7628 JSString::getIndex() calls value() to resolve the string value (is a rope) 7629 to a UString, then passes the result to jsSingleCharacterSubstring without 7630 checking for an exception. In case of out-of-memory the returned UString 7631 is null(), which may result in an out-of-buounds substring being created. 7632 This is bad. 7633 7634 Simple fix is to be able to get an index from a rope without resolving to 7635 UString. This may be a useful optimization in some test cases. 7636 7637 The same bug exists in some other methods is JSString, these can be fixed 7638 by changing them to call getIndex(). 7639 7640 * runtime/JSString.cpp: 7641 (JSC::JSString::resolveRope): 7642 (JSC::JSString::getStringPropertyDescriptor): 7643 * runtime/JSString.h: 7644 (JSC::jsSingleCharacterSubstring): 7645 (JSC::JSString::getIndex): 7646 (JSC::jsSingleCharacterString): 7647 (JSC::JSString::getStringPropertySlot): 7648 * runtime/UStringImpl.cpp: 7649 (JSC::singleCharacterSubstring): 7650 * runtime/UStringImpl.h: 7651 (JSC::UStringImpl::singleCharacterSubstring): 7652 76532010-02-19 Oliver Hunt <oliver@apple.com> 7654 7655 RS = Gavin Barraclough. 7656 7657 Split the 32/64 version of JITPropertyAccess into a separate file. 7658 7659 * GNUmakefile.am: 7660 * JavaScriptCore.gypi: 7661 * JavaScriptCore.pri: 7662 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: 7663 * JavaScriptCore.xcodeproj/project.pbxproj: 7664 * jit/JITPropertyAccess.cpp: 7665 * jit/JITPropertyAccess32_64.cpp: Added. 7666 (JSC::JIT::emit_op_put_by_index): 7667 (JSC::JIT::emit_op_put_getter): 7668 (JSC::JIT::emit_op_put_setter): 7669 (JSC::JIT::emit_op_del_by_id): 7670 (JSC::JIT::emit_op_method_check): 7671 (JSC::JIT::emitSlow_op_method_check): 7672 (JSC::JIT::emit_op_get_by_val): 7673 (JSC::JIT::emitSlow_op_get_by_val): 7674 (JSC::JIT::emit_op_put_by_val): 7675 (JSC::JIT::emitSlow_op_put_by_val): 7676 (JSC::JIT::emit_op_get_by_id): 7677 (JSC::JIT::emitSlow_op_get_by_id): 7678 (JSC::JIT::emit_op_put_by_id): 7679 (JSC::JIT::emitSlow_op_put_by_id): 7680 (JSC::JIT::compileGetByIdHotPath): 7681 (JSC::JIT::compileGetByIdSlowCase): 7682 (JSC::JIT::compilePutDirectOffset): 7683 (JSC::JIT::compileGetDirectOffset): 7684 (JSC::JIT::testPrototype): 7685 (JSC::JIT::privateCompilePutByIdTransition): 7686 (JSC::JIT::patchGetByIdSelf): 7687 (JSC::JIT::patchMethodCallProto): 7688 (JSC::JIT::patchPutByIdReplace): 7689 (JSC::JIT::privateCompilePatchGetArrayLength): 7690 (JSC::JIT::privateCompileGetByIdProto): 7691 (JSC::JIT::privateCompileGetByIdSelfList): 7692 (JSC::JIT::privateCompileGetByIdProtoList): 7693 (JSC::JIT::privateCompileGetByIdChainList): 7694 (JSC::JIT::privateCompileGetByIdChain): 7695 (JSC::JIT::emit_op_get_by_pname): 7696 (JSC::JIT::emitSlow_op_get_by_pname): 7697 76982010-02-19 Patrick Gansterer <paroga@paroga.com> 7699 7700 Reviewed by Laszlo Gombos. 7701 7702 Added additional parameter to create_rvct_stubs 7703 for setting the regularexpression prefix. 7704 Renamed it because it now works for other platforms too. 7705 https://bugs.webkit.org/show_bug.cgi?id=34951 7706 7707 * DerivedSources.pro: 7708 * create_jit_stubs: Copied from JavaScriptCore/create_rvct_stubs. 7709 * create_rvct_stubs: Removed. 7710 77112010-02-18 Oliver Hunt <oliver@apple.com> 7712 7713 Reviewed by Gavin Barraclough. 7714 7715 Improve interpreter getter performance 7716 https://bugs.webkit.org/show_bug.cgi?id=35138 7717 7718 Improve the performance of getter dispatch by making it possible 7719 for the interpreter to cache the GetterSetter object lookup. 7720 7721 To do this we simply need to make PropertySlot aware of getters 7722 as a potentially cacheable property, and record the base and this 7723 objects for a getter access. This allows us to use more-or-less 7724 identical code to that used by the normal get_by_id caching, with 7725 the dispatch being the only actual difference. 7726 7727 I'm holding off of implementing this in the JIT until I do some 7728 cleanup to try and making coding in the JIT not be as horrible 7729 as it is currently. 7730 7731 * bytecode/CodeBlock.cpp: 7732 (JSC::CodeBlock::dump): 7733 (JSC::CodeBlock::derefStructures): 7734 (JSC::CodeBlock::refStructures): 7735 * bytecode/Opcode.h: 7736 * interpreter/Interpreter.cpp: 7737 (JSC::Interpreter::resolveGlobal): 7738 (JSC::Interpreter::tryCacheGetByID): 7739 (JSC::Interpreter::privateExecute): 7740 * jit/JIT.cpp: 7741 (JSC::JIT::privateCompileMainPass): 7742 * jit/JITStubs.cpp: 7743 (JSC::JITThunks::tryCacheGetByID): 7744 (JSC::DEFINE_STUB_FUNCTION): 7745 * runtime/JSObject.cpp: 7746 (JSC::JSObject::fillGetterPropertySlot): 7747 * runtime/PropertySlot.cpp: 7748 (JSC::PropertySlot::functionGetter): 7749 * runtime/PropertySlot.h: 7750 (JSC::PropertySlot::isGetter): 7751 (JSC::PropertySlot::isCacheable): 7752 (JSC::PropertySlot::isCacheableValue): 7753 (JSC::PropertySlot::setValueSlot): 7754 (JSC::PropertySlot::setGetterSlot): 7755 (JSC::PropertySlot::setCacheableGetterSlot): 7756 (JSC::PropertySlot::clearOffset): 7757 (JSC::PropertySlot::thisValue): 7758 77592010-02-17 Geoffrey Garen <ggaren@apple.com> 7760 7761 Reviewed by Oliver Hunt. 7762 7763 Fixed a portion of: 7764 <rdar://problem/7165917> | https://bugs.webkit.org/show_bug.cgi?id=28676 7765 Safari 4 does not release memory back to the operating system fast enough (28676) 7766 7767 This patch fixes a surprisingly common edge case in which the page heap 7768 would have only one free span, but that span would be larger than the 7769 minimum free size, so we would decide not to free it, even though it 7770 could be as large as 100MB or more! 7771 7772 SunSpider reports no change on Mac or Windows. 7773 7774 * wtf/FastMalloc.cpp: 7775 (WTF::TCMalloc_PageHeap::scavenge): Call shouldContinueScavenging() instead 7776 of doing the math ourselves. Don't keep a local value for pagesDecommitted 7777 because that lets free_committed_pages_ be wrong temporarily. Instead, 7778 update free_committed_pages_ as we go. ASSERT that we aren't releasing 7779 a span that has already been released, because we think this is impossible. 7780 Finally, don't be afraid to release all free memory in the page heap when 7781 scavenging. We only scavenge after 5 seconds of the application's working 7782 set not growing, and we keep both thread caches and a central cache on 7783 top of the page heap, so the extra free pages in the page heap were just 7784 overkill. 7785 77862010-02-17 Gavin Barraclough <barraclough@apple.com> 7787 7788 Reviewed by Oliver Hunt. 7789 7790 https://bugs.webkit.org/show_bug.cgi?id=35070 7791 Addition of 2 strings of length 2^31 may result in a string of length 0. 7792 7793 Check for overflow when creating a new JSString as a result of an addition 7794 or concatenation, throw an out of memory exception. 7795 7796 * runtime/JSString.h: 7797 (JSC::): 7798 * runtime/Operations.h: 7799 (JSC::jsString): 7800 78012010-02-17 Xan Lopez <xlopez@igalia.com> 7802 7803 Reviewed by Gustavo Noronha. 7804 7805 [Linux] Webkit incompatible with Java plugins 7806 https://bugs.webkit.org/show_bug.cgi?id=24912 7807 7808 Add support for GFile to GOwnPtr. 7809 7810 Based on original work by Gustavo Noronha. 7811 7812 * wtf/gtk/GOwnPtr.cpp: 7813 (WTF::GFile): 7814 * wtf/gtk/GOwnPtr.h: 7815 78162010-02-16 Gavin Barraclough <barraclough@apple.com> 7817 7818 Reviewed by Mark Rowe. 7819 7820 Fix a handful of other leaks seen on the buildbot. 7821 7822 * runtime/UStringImpl.h: 7823 (JSC::UStringOrRopeImpl::deref): Delegate through to the subclass version of deref to ensure that 7824 the correct cleanup takes place. This function previously featured some code that attempted to 7825 skip deletion of static UStringImpl's. Closer inspection revealed that it was in fact equivalent 7826 to "if (false)", meaning that UStringImpl's which had their final deref performed via this function 7827 were leaked. 7828 78292010-02-16 Mark Rowe <mrowe@apple.com> 7830 7831 Reviewed by Gavin Barraclough. 7832 7833 Fix a handful of leaks seen on the buildbot. 7834 7835 * runtime/UStringImpl.h: 7836 (JSC::UStringOrRopeImpl::deref): Call URopeImpl::destructNonRecursive rather than delete 7837 to ensure that the rope's fibers are also destroyed. 7838 78392010-02-16 Gavin Barraclough <barraclough@apple.com> 7840 7841 Reviewed by Oliver Hunt. 7842 7843 https://bugs.webkit.org/show_bug.cgi?id=34964 7844 Leaks tool reports false memory leaks due to Rope implementation. 7845 7846 A rope is a recursive data structure where each node in the rope holds a set of 7847 pointers, each of which may reference either a string (in UStringImpl form) or 7848 another rope node. A low bit in each pointer is used to distinguish between 7849 rope & string elements, in a fashion similar to the recently-removed 7850 PtrAndFlags class (see https://bugs.webkit.org/show_bug.cgi?id=33731 ). Again, 7851 this causes a problem for Leaks - refactor to remove the magic pointer 7852 mangling. 7853 7854 Move Rope out from JSString.h and rename to URopeImpl, to match UStringImpl. 7855 Give UStringImpl and URopeImpl a common parent class, UStringOrRopeImpl. 7856 Repurpose an otherwise invalid permutation to flags (static & should report 7857 memory cost) to identify ropes. 7858 7859 This allows us to change the rope's fibers to interrogate the object rather 7860 than storing a bool within the low bits of the pointer (or in some cases the 7861 use of a common parent class removes the need to determine the type at all - 7862 there is a common interface to ref or get the length of either ropes or strings). 7863 7864 * API/JSClassRef.cpp: 7865 (OpaqueJSClass::OpaqueJSClass): 7866 (OpaqueJSClassContextData::OpaqueJSClassContextData): 7867 * bytecompiler/BytecodeGenerator.cpp: 7868 (JSC::keyForCharacterSwitch): 7869 * interpreter/Interpreter.cpp: 7870 (JSC::Interpreter::privateExecute): 7871 * jit/JITStubs.cpp: 7872 (JSC::DEFINE_STUB_FUNCTION): 7873 * runtime/ArrayPrototype.cpp: 7874 (JSC::arrayProtoFuncToString): 7875 * runtime/Identifier.cpp: 7876 (JSC::Identifier::equal): 7877 (JSC::Identifier::addSlowCase): 7878 * runtime/JSString.cpp: 7879 (JSC::JSString::resolveRope): 7880 * runtime/JSString.h: 7881 (JSC::): 7882 (JSC::RopeBuilder::JSString): 7883 (JSC::RopeBuilder::~JSString): 7884 (JSC::RopeBuilder::appendStringInConstruct): 7885 (JSC::RopeBuilder::appendValueInConstructAndIncrementLength): 7886 (JSC::RopeBuilder::JSStringFinalizerStruct::JSStringFinalizerStruct): 7887 (JSC::RopeBuilder::JSStringFinalizerStruct::): 7888 * runtime/UString.cpp: 7889 (JSC::UString::toStrictUInt32): 7890 (JSC::equal): 7891 * runtime/UString.h: 7892 (JSC::UString::isEmpty): 7893 (JSC::UString::size): 7894 * runtime/UStringImpl.cpp: 7895 (JSC::URopeImpl::derefFibersNonRecursive): 7896 (JSC::URopeImpl::destructNonRecursive): 7897 * runtime/UStringImpl.h: 7898 (JSC::UStringOrRopeImpl::isRope): 7899 (JSC::UStringOrRopeImpl::length): 7900 (JSC::UStringOrRopeImpl::ref): 7901 (JSC::UStringOrRopeImpl::): 7902 (JSC::UStringOrRopeImpl::operator new): 7903 (JSC::UStringOrRopeImpl::UStringOrRopeImpl): 7904 (JSC::UStringImpl::adopt): 7905 (JSC::UStringImpl::createUninitialized): 7906 (JSC::UStringImpl::tryCreateUninitialized): 7907 (JSC::UStringImpl::data): 7908 (JSC::UStringImpl::cost): 7909 (JSC::UStringImpl::deref): 7910 (JSC::UStringImpl::UStringImpl): 7911 (JSC::UStringImpl::): 7912 (JSC::URopeImpl::tryCreateUninitialized): 7913 (JSC::URopeImpl::initializeFiber): 7914 (JSC::URopeImpl::fiberCount): 7915 (JSC::URopeImpl::fibers): 7916 (JSC::URopeImpl::deref): 7917 (JSC::URopeImpl::URopeImpl): 7918 (JSC::URopeImpl::hasOneRef): 7919 (JSC::UStringOrRopeImpl::deref): 7920 79212010-02-15 Gabor Loki <loki@webkit.org> 7922 7923 Reviewed by Gavin Barraclough. 7924 7925 Fix the SP at ctiOpThrowNotCaught on Thumb2 (JSVALUE32) 7926 https://bugs.webkit.org/show_bug.cgi?id=34939 7927 7928 * jit/JITStubs.cpp: 7929 79302010-02-15 Gavin Barraclough <barraclough@apple.com> 7931 7932 Reviewed by NOBODY (Build Fix!). 7933 7934 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 7935 79362010-02-15 Gavin Barraclough <barraclough@apple.com> 7937 7938 Reviewed by Oliver Hunt. 7939 7940 Some general Rope related refactoring. 7941 7942 Rename Rope::m_ropeLength to m_fiberCount, to be more descriptive. 7943 Rename Rope::m_stringLength to simply m_length (since this is the 7944 more conventional name for the length of a string). Move append 7945 behaviour out into a new RopeBuilder class, so that Rope no longer 7946 needs any knowledge of the JSString or UString implementation. 7947 7948 Make Rope no longer be nested within JSString. 7949 (Rope now no-longer need reside within JSString.h, but leaving 7950 the change of moving this out to a different header as a separate 7951 change from these renames). 7952 7953 * JavaScriptCore.exp: 7954 * jit/JITOpcodes.cpp: 7955 (JSC::JIT::privateCompileCTIMachineTrampolines): 7956 * runtime/JSString.cpp: 7957 (JSC::Rope::destructNonRecursive): 7958 (JSC::Rope::~Rope): 7959 (JSC::JSString::resolveRope): 7960 (JSC::JSString::toBoolean): 7961 (JSC::JSString::getStringPropertyDescriptor): 7962 * runtime/JSString.h: 7963 (JSC::Rope::Fiber::Fiber): 7964 (JSC::Rope::Fiber::deref): 7965 (JSC::Rope::Fiber::ref): 7966 (JSC::Rope::Fiber::refAndGetLength): 7967 (JSC::Rope::Fiber::isRope): 7968 (JSC::Rope::Fiber::rope): 7969 (JSC::Rope::Fiber::isString): 7970 (JSC::Rope::Fiber::string): 7971 (JSC::Rope::Fiber::nonFiber): 7972 (JSC::Rope::tryCreateUninitialized): 7973 (JSC::Rope::append): 7974 (JSC::Rope::fiberCount): 7975 (JSC::Rope::length): 7976 (JSC::Rope::fibers): 7977 (JSC::Rope::Rope): 7978 (JSC::Rope::operator new): 7979 (JSC::): 7980 (JSC::RopeBuilder::JSString): 7981 (JSC::RopeBuilder::~JSString): 7982 (JSC::RopeBuilder::length): 7983 (JSC::RopeBuilder::canGetIndex): 7984 (JSC::RopeBuilder::appendStringInConstruct): 7985 (JSC::RopeBuilder::appendValueInConstructAndIncrementLength): 7986 (JSC::RopeBuilder::isRope): 7987 (JSC::RopeBuilder::fiberCount): 7988 (JSC::JSString::getStringPropertySlot): 7989 * runtime/Operations.h: 7990 (JSC::jsString): 7991 79922010-02-15 Gavin Barraclough <barraclough@apple.com> 7993 7994 Reviewed by NOBODY (Build fix). 7995 7996 Add missing cast for !YARR (PPC) builds. 7997 7998 * runtime/RegExp.cpp: 7999 (JSC::RegExp::match): 8000 80012010-02-14 Gavin Barraclough <barraclough@apple.com> 8002 8003 Reviewed by Darin Adler. 8004 8005 https://bugs.webkit.org/show_bug.cgi?id=33731 8006 Many false leaks in release builds due to PtrAndFlags 8007 8008 StructureTransitionTable was effectively a smart pointer type, 8009 one machine word in size and wholly contained as a member of 8010 of Structure. It either pointed to an actual table, or could 8011 be used to describe a single transtion entry without use of a 8012 table. 8013 8014 This, however, worked by using a PtrAndFlags, which is not 8015 compatible with the leaks tool. Since there is no clear way to 8016 obtain another bit for 'free' here, and since there are bits 8017 available up in Structure, merge this functionality back up into 8018 Structure. Having this in a separate class was quite clean 8019 from an enacapsulation perspective, but this solution doesn't 8020 seem to bad - all table access is now intermediated through the 8021 Structure::structureTransitionTableFoo methods, keeping the 8022 optimization fairly well contained. 8023 8024 This was the last use of PtrAndFlags, so removing the file too. 8025 8026 * JavaScriptCore.xcodeproj/project.pbxproj: 8027 * bytecode/CodeBlock.h: 8028 * runtime/Structure.cpp: 8029 (JSC::Structure::Structure): 8030 (JSC::Structure::~Structure): 8031 (JSC::Structure::addPropertyTransitionToExistingStructure): 8032 (JSC::Structure::addPropertyTransition): 8033 (JSC::Structure::hasTransition): 8034 * runtime/Structure.h: 8035 (JSC::Structure::): 8036 (JSC::Structure::structureTransitionTableContains): 8037 (JSC::Structure::structureTransitionTableGet): 8038 (JSC::Structure::structureTransitionTableHasTransition): 8039 (JSC::Structure::structureTransitionTableRemove): 8040 (JSC::Structure::structureTransitionTableAdd): 8041 (JSC::Structure::structureTransitionTable): 8042 (JSC::Structure::setStructureTransitionTable): 8043 (JSC::Structure::singleTransition): 8044 (JSC::Structure::setSingleTransition): 8045 * runtime/StructureTransitionTable.h: 8046 * wtf/PtrAndFlags.h: Removed. 8047 80482010-02-15 Gavin Barraclough <barraclough@apple.com> 8049 8050 Rubber Stamped by Geoff Garen. 8051 8052 Bug 34948 - tryMakeString should fail on error in length calculation 8053 8054 Ooops! - "bool overflow" argument should have been "bool& overflow". 8055 8056 * runtime/UString.h: 8057 (JSC::sumWithOverflow): 8058 (JSC::tryMakeString): 8059 80602010-02-15 Gavin Barraclough <barraclough@apple.com> 8061 8062 Reviewed by NOBODY (Build Fix (pt 2!)). 8063 8064 Some symbol names have changed, remove, will readd if required. 8065 8066 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 8067 80682010-02-15 Gavin Barraclough <barraclough@apple.com> 8069 8070 Reviewed by NOBODY (Build Fix (pt 1?)). 8071 8072 Some symbol names have changed, remove, will readd if required. 8073 8074 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 8075 80762010-02-15 Geoffrey Garen <ggaren@apple.com> 8077 8078 Reviewed by Oliver Hunt. 8079 8080 Removed some mistaken code added in http://trac.webkit.org/changeset/53860. 8081 8082 * API/APIShims.h: 8083 (JSC::APICallbackShim::APICallbackShim): 8084 (JSC::APICallbackShim::~APICallbackShim): No need to start/stop the 8085 timeout checker when calling out from the API to the client; we want to 8086 monitor the VM for timeouts, not the client. This mistake was harmless / 8087 undetectable, since it's totally redundant with the APIEntryShim, which 8088 also starts / stops the timeout checker. 8089 80902010-02-15 Gavin Barraclough <barraclough@apple.com> 8091 8092 Reviewed by Geoff Garen. 8093 8094 Bug 34952 - String lengths in UString should be unsigned. 8095 This matches WebCore::StringImpl, and better unifies behaviour throughout JSC. 8096 8097 * JavaScriptCore.exp: 8098 * bytecode/EvalCodeCache.h: 8099 * runtime/Identifier.cpp: 8100 (JSC::Identifier::equal): 8101 * runtime/Identifier.h: 8102 * runtime/JSGlobalObjectFunctions.cpp: 8103 (JSC::globalFuncEscape): 8104 * runtime/JSONObject.cpp: 8105 (JSC::gap): 8106 (JSC::Stringifier::indent): 8107 * runtime/NumberPrototype.cpp: 8108 (JSC::numberProtoFuncToFixed): 8109 (JSC::numberProtoFuncToPrecision): 8110 * runtime/RegExp.cpp: 8111 (JSC::RegExp::match): 8112 * runtime/StringPrototype.cpp: 8113 (JSC::substituteBackreferencesSlow): 8114 (JSC::stringProtoFuncReplace): 8115 (JSC::stringProtoFuncSplit): 8116 (JSC::trimString): 8117 * runtime/UString.cpp: 8118 (JSC::UString::UString): 8119 (JSC::UString::from): 8120 (JSC::UString::getCString): 8121 (JSC::UString::ascii): 8122 (JSC::UString::operator[]): 8123 (JSC::UString::toStrictUInt32): 8124 (JSC::UString::find): 8125 (JSC::UString::rfind): 8126 (JSC::UString::substr): 8127 (JSC::operator<): 8128 (JSC::operator>): 8129 (JSC::compare): 8130 (JSC::equal): 8131 (JSC::UString::UTF8String): 8132 * runtime/UString.h: 8133 (JSC::UString::size): 8134 (JSC::operator==): 8135 * runtime/UStringImpl.cpp: 8136 (JSC::UStringImpl::create): 8137 * runtime/UStringImpl.h: 8138 (JSC::UStringImpl::create): 8139 (JSC::UStringImpl::size): 8140 (JSC::UStringImpl::computeHash): 8141 (JSC::UStringImpl::UStringImpl): 8142 81432010-02-15 Gavin Barraclough <barraclough@apple.com> 8144 8145 Reviewed by Geoff Garen. 8146 8147 Bug 34948 - tryMakeString should fail on error in length calculation 8148 8149 The sum of the length of substrings could overflow. 8150 8151 * runtime/UString.h: 8152 (JSC::sumWithOverflow): 8153 (JSC::tryMakeString): 8154 81552010-02-15 Geoffrey Garen <ggaren@apple.com> 8156 8157 Reviewed by Oliver Hunt. 8158 8159 Fixed <rdar://problem/7628524> Crash beneath JSGlobalContextRelease when 8160 typing in Google search field with GuardMalloc/full page heap enabled 8161 8162 * API/JSContextRef.cpp: Don't use APIEntryShim, since that requires 8163 a JSGlobalData, which this function destroys. Do use setCurrentIdentifierTable 8164 and JSLock instead, since those are the two features of APIEntryShim we 8165 require. 8166 81672010-02-15 Patrick Gansterer <paroga@paroga.com> 8168 8169 Reviewed by Laszlo Gombos. 8170 8171 Added additional parameter to create_rvct_stubs 8172 for setting the offset of thunkReturnAddress. 8173 https://bugs.webkit.org/show_bug.cgi?id=34657 8174 8175 * create_rvct_stubs: 8176 * jit/JITStubs.cpp: 8177 81782010-02-15 Jedrzej Nowacki <jedrzej.nowacki@nokia.com> 8179 8180 Reviewed by Simon Hausmann. 8181 8182 Fix QScriptValue::toIntXX methods. 8183 8184 More ECMA Script compliance. 8185 8186 [Qt] QScriptValue::toIntXX returns incorrect values 8187 https://bugs.webkit.org/show_bug.cgi?id=34847 8188 8189 * qt/api/qscriptvalue_p.h: 8190 (QScriptValuePrivate::toInteger): 8191 (QScriptValuePrivate::toInt32): 8192 (QScriptValuePrivate::toUInt32): 8193 (QScriptValuePrivate::toUInt16): 8194 * qt/tests/qscriptvalue/tst_qscriptvalue.h: 8195 * qt/tests/qscriptvalue/tst_qscriptvalue_generated.cpp: 8196 (tst_QScriptValue::toInteger_initData): 8197 (tst_QScriptValue::toInteger_makeData): 8198 (tst_QScriptValue::toInteger_test): 8199 (tst_QScriptValue::toInt32_initData): 8200 (tst_QScriptValue::toInt32_makeData): 8201 (tst_QScriptValue::toInt32_test): 8202 (tst_QScriptValue::toUInt32_initData): 8203 (tst_QScriptValue::toUInt32_makeData): 8204 (tst_QScriptValue::toUInt32_test): 8205 (tst_QScriptValue::toUInt16_initData): 8206 (tst_QScriptValue::toUInt16_makeData): 8207 (tst_QScriptValue::toUInt16_test): 8208 82092010-02-14 Laszlo Gombos <laszlo.1.gombos@nokia.com> 8210 8211 Reviewed by Adam Barth. 8212 8213 Implement NEVER_INLINE and NO_RETURN for RVCT 8214 https://bugs.webkit.org/show_bug.cgi?id=34740 8215 8216 * wtf/AlwaysInline.h: 8217 82182010-02-12 Gavin Barraclough <barraclough@apple.com> 8219 8220 Reviewed by Oliver Hunt. 8221 8222 https://bugs.webkit.org/show_bug.cgi?id=33731 8223 Remove uses of PtrAndFlags from JIT data stuctures. 8224 8225 These break the OS X Leaks tool. Free up a bit in CallLinkInfo, and invalid 8226 permutation of pointer states in MethodCallLinkInfo to represent the removed bits. 8227 8228 * bytecode/CodeBlock.h: 8229 (JSC::CallLinkInfo::seenOnce): 8230 (JSC::CallLinkInfo::setSeen): 8231 (JSC::MethodCallLinkInfo::MethodCallLinkInfo): 8232 (JSC::MethodCallLinkInfo::seenOnce): 8233 (JSC::MethodCallLinkInfo::setSeen): 8234 * jit/JIT.cpp: 8235 (JSC::JIT::unlinkCall): 8236 * jit/JITPropertyAccess.cpp: 8237 (JSC::JIT::patchMethodCallProto): 8238 * runtime/UString.h: 8239 82402010-02-12 Gavin Barraclough <barraclough@apple.com> 8241 8242 Reviewed by Darin Adler. 8243 8244 https://bugs.webkit.org/show_bug.cgi?id=33731 8245 Many false leaks in release builds due to PtrAndFlags 8246 8247 Remove UntypedPtrAndBitfield (similar to PtrAndFlags) in UStringImpl, 8248 and steal bits from the refCount instead. 8249 8250 * runtime/UStringImpl.cpp: 8251 (JSC::UStringImpl::baseSharedBuffer): 8252 (JSC::UStringImpl::~UStringImpl): 8253 * runtime/UStringImpl.h: 8254 (JSC::UStringImpl::cost): 8255 (JSC::UStringImpl::isIdentifier): 8256 (JSC::UStringImpl::setIsIdentifier): 8257 (JSC::UStringImpl::ref): 8258 (JSC::UStringImpl::deref): 8259 (JSC::UStringImpl::UStringImpl): 8260 (JSC::UStringImpl::bufferOwnerString): 8261 (JSC::UStringImpl::bufferOwnership): 8262 (JSC::UStringImpl::isStatic): 8263 (JSC::UStringImpl::): 8264 82652010-02-12 Geoffrey Garen <ggaren@apple.com> 8266 8267 Reviewed by Darin Adler. 8268 8269 Removed an unnecessary data dependency from my last patch. 8270 8271 * runtime/SmallStrings.cpp: 8272 (JSC::SmallStrings::markChildren): Since isAnyStringMarked being false 8273 is a condition of entering the loop, we can just use '=' instead of '|='. 8274 82752010-02-12 Janne Koskinen <janne.p.koskinen@digia.com> 8276 8277 Reviewed by Tor Arne Vestbø. 8278 8279 Additional refptr/passrefptr workarounds for WINSCW compiler 8280 https://bugs.webkit.org/show_bug.cgi?id=28054 8281 8282 * wtf/PassRefPtr.h: 8283 (WTF::refIfNotNull): 8284 (WTF::PassRefPtr::PassRefPtr): 8285 (WTF::PassRefPtr::~PassRefPtr): 8286 (WTF::PassRefPtr::clear): 8287 (WTF::::operator): 8288 * wtf/RefPtr.h: 8289 (WTF::RefPtr::RefPtr): 8290 (WTF::::operator): 8291 82922010-02-12 Janne Koskinen <janne.p.koskinen@digia.com> 8293 8294 Reviewed by Tor Arne Vestbø. 8295 8296 Additional refptr/passrefptr workarounds for WINSCW compiler 8297 https://bugs.webkit.org/show_bug.cgi?id=28054 8298 8299 * wtf/PassRefPtr.h: 8300 (WTF::refIfNotNull): 8301 (WTF::PassRefPtr::PassRefPtr): 8302 (WTF::PassRefPtr::~PassRefPtr): 8303 (WTF::PassRefPtr::clear): 8304 (WTF::::operator): 8305 * wtf/RefPtr.h: 8306 (WTF::RefPtr::RefPtr): 8307 (WTF::::operator): 8308 83092010-02-12 Janne Koskinen <janne.p.koskinen@digia.com> 8310 8311 Reviewed by Tor Arne Vestbø. 8312 8313 Additional refptr/passrefptr workarounds for WINSCW compiler 8314 https://bugs.webkit.org/show_bug.cgi?id=28054 8315 8316 * wtf/PassRefPtr.h: 8317 (WTF::refIfNotNull): 8318 (WTF::PassRefPtr::PassRefPtr): 8319 (WTF::PassRefPtr::~PassRefPtr): 8320 (WTF::PassRefPtr::clear): 8321 (WTF::::operator): 8322 * wtf/RefPtr.h: 8323 (WTF::RefPtr::RefPtr): 8324 (WTF::::operator): 8325 83262010-02-12 Janne Koskinen <janne.p.koskinen@digia.com> 8327 8328 Reviewed by Simon Hausmann. 8329 8330 Don't import the cmath functions from std:: for WINSCW. 8331 8332 * wtf/MathExtras.h: 8333 83342010-02-12 Kwang Yul Seo <skyul@company100.net> 8335 8336 Reviewed by Adam Barth. 8337 8338 Typedef both JSChar and UChar to wchar_t in RVCT. 8339 https://bugs.webkit.org/show_bug.cgi?id=34560 8340 8341 Define both JSChar and UChar to wchar_t as the size 8342 of wchar_t is 2 bytes in RVCT. 8343 8344 * API/JSStringRef.h: 8345 * wtf/unicode/qt4/UnicodeQt4.h: 8346 83472010-02-11 Geoffrey Garen <ggaren@apple.com> 8348 8349 Reviewed by Oliver Hunt and Darin Adler. 8350 8351 The rest of the fix for 8352 https://bugs.webkit.org/show_bug.cgi?id=34864 | <rdar://problem/7594198> 8353 Many objects left uncollected after visiting mail.google.com and closing 8354 window 8355 8356 Don't unconditionally hang onto small strings. Instead, hang onto all 8357 small strings as long as any small string is still referenced. 8358 8359 SunSpider reports no change. 8360 8361 * runtime/Collector.cpp: 8362 (JSC::Heap::markRoots): Mark the small strings cache last, so it can 8363 check if anything else has kept any strings alive. 8364 8365 * runtime/SmallStrings.cpp: 8366 (JSC::isMarked): 8367 (JSC::SmallStrings::markChildren): Only keep our strings alive if some 8368 other reference to at least one of them exists, too. 8369 83702010-02-11 Geoffrey Garen <ggaren@apple.com> 8371 8372 Reviewed by Gavin Barraclough. 8373 8374 Some progress toward fixing 8375 https://bugs.webkit.org/show_bug.cgi?id=34864 | <rdar://problem/7594198> 8376 Many objects left uncollected after visiting mail.google.com and closing 8377 window 8378 8379 SunSpider reports no change. 8380 8381 Keep weak references, rather than protected references, to cached for-in 8382 property name enumerators. 8383 8384 One problem with protected references is that a chain like 8385 [ gc object 1 ] => [ non-gc object ] => [ gc object 2 ] 8386 takes two GC passes to break, since the first pass collects [ gc object 1 ], 8387 releasing [ non-gc object ] and unprotecting [ gc object 2 ], and only 8388 then can a second pass collect [ gc object 2 ]. 8389 8390 Another problem with protected references is that they can keep a bunch 8391 of strings alive long after they're useful. In SunSpider and a few popular 8392 websites, the size-speed tradeoff seems to favor weak references. 8393 8394 * runtime/JSPropertyNameIterator.cpp: 8395 (JSC::JSPropertyNameIterator::JSPropertyNameIterator): Moved this constructor 8396 into the .cpp file, since it's not used elsewhere. 8397 8398 (JSC::JSPropertyNameIterator::~JSPropertyNameIterator): Added a destructor 8399 to support our weak reference. 8400 8401 * runtime/JSPropertyNameIterator.h: 8402 (JSC::Structure::setEnumerationCache): 8403 (JSC::Structure::clearEnumerationCache): 8404 (JSC::Structure::enumerationCache): Added a function for clearing a 8405 Structure's enumeration cache, used by our new destructor. Also fixed 8406 indentation to match the rest of the file. 8407 8408 * runtime/Structure.h: Changed from protected pointer to weak pointer. 8409 84102010-02-11 Chris Rogers <crogers@google.com> 8411 8412 Reviewed by David Levin. 8413 8414 audio engine: add Complex number class 8415 https://bugs.webkit.org/show_bug.cgi?id=34538 8416 8417 * wtf/Complex.h: Added. 8418 (WebCore::complexFromMagnitudePhase): 8419 84202010-02-10 Geoffrey Garen <ggaren@apple.com> 8421 8422 Reviewed by Oliver Hunt. 8423 8424 Added an SPI for asking about all the different live objects on the heap. 8425 Useful for memory debugging. 8426 8427 * JavaScriptCore.exp: Export the new SPI. 8428 8429 * runtime/Collector.cpp: 8430 (JSC::typeName): Use a little capitalization. Don't crash in the case of 8431 a non-object cell, since it might just be an uninitialized cell. 8432 8433 (JSC::Heap::objectTypeCounts): The new SPI. 8434 8435 * runtime/Collector.h: 8436 * runtime/CollectorHeapIterator.h: 8437 (JSC::CollectorHeapIterator::advance): 8438 (JSC::LiveObjectIterator::operator++): 8439 (JSC::DeadObjectIterator::operator++): 8440 (JSC::ObjectIterator::operator++): Made 2 tweaks to these iterators: 8441 (1) Skip the last cell in the block, since it's a dummy sentinel, and 8442 we don't want it to confuse the object count; (2) Fixed a logic error 8443 in LiveObjectIterator that could cause it to iterate dead objects if 8444 m_block were equal to m_heap.nextBlock and m_cell were less than 8445 m_heap.nextCell. No test for this since I can't think of a way that this 8446 could make WebKit behave badly. 8447 84482010-02-11 Steve Block <steveblock@google.com> 8449 8450 Reviewed by Darin Adler. 8451 8452 Guard cmath using declarations in MathExtras.h on Android 8453 https://bugs.webkit.org/show_bug.cgi?id=34840 8454 8455 Android does not provide these functions. 8456 8457 * wtf/MathExtras.h: 8458 84592010-02-08 Maciej Stachowiak <mjs@apple.com> 8460 8461 Reviewed by Cameron Zwarich. 8462 8463 Restore ENABLE_RUBY flag so vendors can ship with Ruby disabled if they choose. 8464 https://bugs.webkit.org/show_bug.cgi?id=34698 8465 8466 * Configurations/FeatureDefines.xcconfig: 8467 84682010-02-10 Kevin Watters <kevinwatters@gmail.com> 8469 8470 Reviewed by Kevin Ollivier. 8471 8472 [wx] Add Windows complex text support and Mac support for containsCharacters. 8473 8474 https://bugs.webkit.org/show_bug.cgi?id=34759 8475 8476 * wscript: 8477 84782010-02-10 Alexey Proskuryakov <ap@apple.com> 8479 8480 Addressing issues found by style bot. 8481 8482 * wtf/ValueCheck.h: Renamed header guard to match final file name. 8483 8484 * wtf/Vector.h: (WTF::::checkConsistency): Remove braces around a one-line clause. 8485 84862010-02-09 Alexey Proskuryakov <ap@apple.com> 8487 8488 Reviewed by Geoffrey Garen. 8489 8490 https://bugs.webkit.org/show_bug.cgi?id=34490 8491 WebCore::ImageEventSender::dispatchPendingEvents() crashes in certain conditions 8492 8493 * GNUmakefile.am: 8494 * JavaScriptCore.gypi: 8495 * JavaScriptCore.vcproj/WTF/WTF.vcproj: 8496 * JavaScriptCore.xcodeproj/project.pbxproj: 8497 Added ValueCheck.h. 8498 8499 * wtf/ValueCheck.h: Added. Moved code out of HashTraits, since it would be awkward to 8500 include that from Vector.h. 8501 (WTF::ValueCheck::checkConsistency): Allow null pointers, those are pretty consistent. 8502 8503 * wtf/HashTraits.h: Moved value checking code out of here. 8504 8505 * wtf/HashTable.h: (WTF::::checkTableConsistencyExceptSize): Updated for the above changes. 8506 8507 * wtf/Vector.h: 8508 (WTF::::checkConsistency): Check all vector elements. 8509 (WTF::ValueCheck): Support checking a Vector as an element in other containers. Currently 8510 unused. 8511 85122010-02-10 Jedrzej Nowacki <jedrzej.nowacki@nokia.com> 8513 8514 Reviewed by Simon Hausmann. 8515 8516 Fix QScriptValue::toBool. 8517 8518 Fix ECMA compliance in the QScriptValue for values like 0, NaN and 8519 empty strings. 8520 8521 [Qt] QScriptValue::toBool problem 8522 https://bugs.webkit.org/show_bug.cgi?id=34793 8523 8524 * qt/api/qscriptvalue_p.h: 8525 (QScriptValuePrivate::toBool): 8526 * qt/tests/qscriptvalue/tst_qscriptvalue.h: 8527 * qt/tests/qscriptvalue/tst_qscriptvalue_generated.cpp: 8528 (tst_QScriptValue::toBool_initData): 8529 (tst_QScriptValue::toBool_makeData): 8530 (tst_QScriptValue::toBool_test): 8531 (tst_QScriptValue::toBoolean_initData): 8532 (tst_QScriptValue::toBoolean_makeData): 8533 (tst_QScriptValue::toBoolean_test): 8534 85352009-10-06 Yongjun Zhang <yongjun.zhang@nokia.com> 8536 8537 Reviewed by Simon Hausmann. 8538 8539 Use derefIfNotNull() to work around WINSCW compiler forward declaration bug 8540 8541 The compiler bug is reported at 8542 https://xdabug001.ext.nokia.com/bugzilla/show_bug.cgi?id=9812. 8543 8544 The change should be reverted when the above bug is fixed in WINSCW compiler. 8545 8546 https://bugs.webkit.org/show_bug.cgi?id=28054 8547 85482009-10-06 Yongjun Zhang <yongjun.zhang@nokia.com> 8549 8550 Reviewed by Simon Hausmann. 8551 8552 Get rid of WINSCW hack for UnSpecifiedBoolType 8553 8554 Add parenthesis around (RefPtr::*UnspecifiedBoolType) to make the WINSCW 8555 compiler work with the default UnSpecifiedBoolType() operator. 8556 8557 https://bugs.webkit.org/show_bug.cgi?id=28054 8558 8559 * wtf/RefPtr.h: 8560 85612010-02-09 Jedrzej Nowacki <jedrzej.nowacki@nokia.com> 8562 8563 Reviewed by Simon Hausmann. 8564 8565 New functions nullValue() and undefinedValue(). 8566 8567 [Qt] QScriptEngine should contain nullValue and undefinedValue methods 8568 https://bugs.webkit.org/show_bug.cgi?id=34749 8569 8570 * qt/api/qscriptengine.cpp: 8571 (QScriptEngine::nullValue): 8572 (QScriptEngine::undefinedValue): 8573 * qt/api/qscriptengine.h: 8574 * qt/tests/qscriptengine/tst_qscriptengine.cpp: 8575 (tst_QScriptEngine::nullValue): 8576 (tst_QScriptEngine::undefinedValue): 8577 85782010-02-09 Jedrzej Nowacki <jedrzej.nowacki@nokia.com> 8579 8580 Reviewed by Simon Hausmann. 8581 8582 Fixes for QScriptValue::toNumber(). 8583 8584 Fix ECMA compliance in QScriptValue for values unbound 8585 to a QScriptEngine. 8586 8587 [Qt] QScriptValue::toNumber() is broken 8588 https://bugs.webkit.org/show_bug.cgi?id=34592 8589 8590 * qt/api/qscriptvalue_p.h: 8591 (QScriptValuePrivate::toNumber): 8592 * qt/tests/qscriptvalue/tst_qscriptvalue.h: 8593 * qt/tests/qscriptvalue/tst_qscriptvalue_generated.cpp: 8594 (tst_QScriptValue::toNumber_initData): 8595 (tst_QScriptValue::toNumber_makeData): 8596 (tst_QScriptValue::toNumber_test): 8597 85982010-02-09 Jedrzej Nowacki <jedrzej.nowacki@nokia.com> 8599 8600 Reviewed by Simon Hausmann. 8601 8602 Fix QScriptValue::isNumber(). 8603 8604 The isNumber() should return 'true' if the value is in the CNumber 8605 state. 8606 8607 [Qt] QScriptValue::isNumber() returns an incorrect value 8608 https://bugs.webkit.org/show_bug.cgi?id=34575 8609 8610 * qt/api/qscriptvalue_p.h: 8611 (QScriptValuePrivate::isNumber): 8612 * qt/tests/qscriptvalue/tst_qscriptvalue.h: 8613 * qt/tests/qscriptvalue/tst_qscriptvalue_generated.cpp: 8614 (tst_QScriptValue::isNumber_initData): 8615 (tst_QScriptValue::isNumber_makeData): 8616 (tst_QScriptValue::isNumber_test): 8617 86182010-02-09 Geoffrey Garen <ggaren@apple.com> 8619 8620 Reviewed by Oliver Hunt. 8621 8622 Small refactoring to the small strings cache to allow it to be cleared 8623 dynamically. 8624 8625 * runtime/SmallStrings.cpp: 8626 (JSC::SmallStrings::SmallStrings): 8627 (JSC::SmallStrings::clear): 8628 * runtime/SmallStrings.h: Moved initialization code into a shared function, 8629 and changed the constructor to call it. 8630 86312010-02-09 Gavin Barraclough <barraclough@apple.com> 8632 8633 Rubber Stamped by Geoff Garen. 8634 8635 Rename StringBuilder::release && JSStringBuilder::releaseJSString 8636 to 'build()'. 8637 8638 * runtime/ArrayPrototype.cpp: 8639 (JSC::arrayProtoFuncToLocaleString): 8640 (JSC::arrayProtoFuncJoin): 8641 * runtime/Executable.cpp: 8642 (JSC::FunctionExecutable::paramString): 8643 * runtime/FunctionConstructor.cpp: 8644 (JSC::constructFunction): 8645 * runtime/JSGlobalObjectFunctions.cpp: 8646 (JSC::encode): 8647 (JSC::decode): 8648 (JSC::globalFuncEscape): 8649 (JSC::globalFuncUnescape): 8650 * runtime/JSONObject.cpp: 8651 (JSC::Stringifier::stringify): 8652 * runtime/JSStringBuilder.h: 8653 (JSC::JSStringBuilder::build): 8654 * runtime/LiteralParser.cpp: 8655 (JSC::LiteralParser::Lexer::lexString): 8656 * runtime/NumberPrototype.cpp: 8657 (JSC::integerPartNoExp): 8658 (JSC::numberProtoFuncToFixed): 8659 * runtime/StringBuilder.h: 8660 (JSC::StringBuilder::build): 8661 86622010-02-09 John Sullivan <sullivan@apple.com> 8663 8664 https://bugs.webkit.org/show_bug.cgi?id=34772 8665 Overzealous new assertion in URStringImpl::adopt() 8666 8667 Reviewed by Adam Barth. 8668 8669 * runtime/UStringImpl.h: 8670 (JSC::UStringImpl::adopt): 8671 Only assert that vector.data() is non-zero if vector.size() is non-zero. 8672 86732010-02-09 Nikolas Zimmermann <nzimmermann@rim.com> 8674 8675 Not reviewed. Try to fix build problem on SnowLeopard slaves to bring them back. 8676 8677 * API/JSClassRef.cpp: 8678 (tryCreateStringFromUTF8): Mark method as 'static inline' to suppress "warning: no previous prototype for ..." 8679 86802010-02-09 Gavin Barraclough <barraclough@apple.com> 8681 8682 Reviewed by Oliver Hunt. 8683 8684 Three small string fixes: 8685 (1) StringBuilder::release should CRASH if the buffer allocation failed. 8686 (2) Remove weird, dead code from JSString::tryGetValue, replace with an ASSERT. 8687 (3) Move UString::createFromUTF8 out to the API, as tryCreateStringFromUTF8. 8688 This is only used from the API, and (now) unlike other UString::create 8689 methods may return UString::null() to indicate failure cases. Better 8690 handle these in the API. 8691 8692 * API/JSClassRef.cpp: 8693 (tryCreateStringFromUTF8): 8694 (OpaqueJSClass::OpaqueJSClass): 8695 (OpaqueJSClassContextData::OpaqueJSClassContextData): 8696 * runtime/JSString.h: 8697 (JSC::Fiber::tryGetValue): 8698 * runtime/StringBuilder.h: 8699 (JSC::StringBuilder::release): 8700 * runtime/UString.cpp: 8701 (JSC::UString::UString): 8702 (JSC::UString::from): 8703 (JSC::UString::find): 8704 * runtime/UString.h: 8705 87062010-02-09 Janne Koskinen <janne.p.koskinen@digia.com> 8707 8708 Reviewed by Laszlo Gombos. 8709 8710 [Qt] use nanval() for Symbian as nonInlineNaN 8711 https://bugs.webkit.org/show_bug.cgi?id=34170 8712 8713 numeric_limits<double>::quiet_NaN is broken in Symbian 8714 causing NaN to be evaluated as a number. 8715 8716 * runtime/JSValue.cpp: 8717 (JSC::nonInlineNaN): 8718 87192010-02-09 Tamas Szirbucz <szirbucz@inf.u-szeged.hu> 8720 8721 Reviewed by Gavin Barraclough. 8722 8723 Add a soft modulo operation to ARM JIT using a trampoline function. 8724 The performance progression is about ~1.8% on ARMv7 8725 https://bugs.webkit.org/show_bug.cgi?id=34424 8726 8727 Developed in cooperation with Gabor Loki. 8728 8729 * jit/JIT.h: 8730 * jit/JITArithmetic.cpp: 8731 (JSC::JIT::emit_op_mod): 8732 (JSC::JIT::emitSlow_op_mod): 8733 * jit/JITOpcodes.cpp: 8734 (JSC::JIT::softModulo): 8735 * jit/JITStubs.h: 8736 (JSC::JITThunks::ctiSoftModulo): 8737 * wtf/Platform.h: 8738 87392010-02-08 Gavin Barraclough <barraclough@apple.com> 8740 8741 Reviewed by NOBODY (SL/win build fixes). 8742 8743 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 8744 * runtime/StringPrototype.cpp: 8745 87462010-02-08 Gavin Barraclough <barraclough@apple.com> 8747 8748 Reviewed by Oliver Hunt 8749 8750 Make String.replace throw an exception on out-of-memory, rather than 8751 returning a null (err, empty-ish) string. Move String::replaceRange 8752 and String::spliceSubstringsWithSeparators out to StringPrototype - 8753 these were fairly specific use anyway, and we can better integrate 8754 throwing the JS expcetion this way. 8755 8756 Also removes redundant assignment operator from UString. 8757 8758 * JavaScriptCore.exp: 8759 * runtime/StringPrototype.cpp: 8760 (JSC::StringRange::StringRange): 8761 (JSC::jsSpliceSubstringsWithSeparators): 8762 (JSC::jsReplaceRange): 8763 (JSC::stringProtoFuncReplace): 8764 * runtime/UString.cpp: 8765 * runtime/UString.h: 8766 87672010-02-08 Kwang Yul Seo <skyul@company100.net> 8768 8769 Reviewed by Eric Seidel. 8770 8771 [BREWMP] Undefine WTF_OS_WINDOWS and WTF_PLATFORM_WIN 8772 https://bugs.webkit.org/show_bug.cgi?id=34561 8773 8774 As the binary for simulator is built with MSVC 2005, 8775 WTF_OS_WINDOWS and WTF_PLATFORM_WIN are defined. 8776 Undefine them as we don't target Windows. 8777 8778 * wtf/Platform.h: 8779 87802010-02-08 Chris Rogers <crogers@google.com> 8781 8782 Reviewed by Darin Adler. 8783 8784 audio engine: add Vector3 class 8785 https://bugs.webkit.org/show_bug.cgi?id=34548 8786 8787 * wtf/Vector3.h: Added. 8788 (WebCore::Vector3::Vector3): 8789 (WebCore::Vector3::abs): 8790 (WebCore::Vector3::isZero): 8791 (WebCore::Vector3::normalize): 8792 (WebCore::Vector3::x): 8793 (WebCore::Vector3::y): 8794 (WebCore::Vector3::z): 8795 (WebCore::operator+): 8796 (WebCore::operator-): 8797 (WebCore::operator*): 8798 (WebCore::dot): 8799 (WebCore::cross): 8800 (WebCore::distance): 8801 88022010-02-08 Oliver Hunt <oliver@apple.com> 8803 8804 Reviewed by Gavin Barraclough. 8805 8806 Fix warning in clang++ 8807 8808 * runtime/Structure.h: 8809 (JSC::Structure::propertyStorageSize): 8810 88112010-02-08 Gavin Barraclough <barraclough@apple.com> 8812 8813 Reviewed by Geoff Garen. 8814 8815 Make makeString CRASH if we fail to allocate a string. 8816 8817 (tryMakeString or jsMakeNontrivialString can be used where we 8818 expect allocation may fail and want to handle the error). 8819 8820 * runtime/JSStringBuilder.h: 8821 (JSC::jsMakeNontrivialString): 8822 * runtime/UString.h: 8823 (JSC::tryMakeString): 8824 (JSC::makeString): 8825 88262010-02-08 Gavin Barraclough <barraclough@apple.com> 8827 8828 Rubber Stamped by Oliver Hunt. 8829 8830 Remove a couple of unnecesary C-style casts spotted by Darin. 8831 8832 * runtime/JSGlobalObjectFunctions.cpp: 8833 (JSC::encode): 8834 (JSC::globalFuncEscape): 8835 88362010-02-08 Gavin Barraclough <barraclough@apple.com> 8837 8838 Reviewed by Geoff Garen. 8839 8840 Switch some more StringBuilder/jsNontrivialString code to use 8841 JSStringBuilder/jsMakeNontrivialString - these methods will 8842 throw an exception if we hit out-of-memory, rather than just 8843 CRASHing. 8844 8845 * runtime/FunctionPrototype.cpp: 8846 (JSC::functionProtoFuncToString): 8847 * runtime/JSGlobalObjectFunctions.cpp: 8848 (JSC::encode): 8849 (JSC::decode): 8850 (JSC::globalFuncEscape): 8851 88522010-02-08 Gavin Barraclough <barraclough@apple.com> 8853 8854 Reviewed by Sam Weinig. 8855 8856 Use an empty identifier instead of a null identifier for parse 8857 tokens without an identifier. 8858 8859 This helps encapsulate the null UStringImpl within UString. 8860 8861 * parser/Grammar.y: 8862 * parser/NodeConstructors.h: 8863 (JSC::ContinueNode::ContinueNode): 8864 (JSC::BreakNode::BreakNode): 8865 (JSC::ForInNode::ForInNode): 8866 * runtime/CommonIdentifiers.cpp: 8867 (JSC::CommonIdentifiers::CommonIdentifiers): 8868 * runtime/CommonIdentifiers.h: 8869 * runtime/FunctionPrototype.cpp: 8870 (JSC::FunctionPrototype::FunctionPrototype): 8871 88722010-02-08 Gustavo Noronha Silva <gustavo.noronha@collabora.co.uk> 8873 8874 Build fix for make distcheck. 8875 8876 * GNUmakefile.am: 8877 88782010-02-08 Simon Hausmann <simon.hausmann@nokia.com> 8879 8880 Unreviewed RVCT build fix. 8881 8882 Similar to r54391, don't import the cmath functions from std:: for RVCT. 8883 8884 * wtf/MathExtras.h: 8885 88862010-02-05 Gavin Barraclough <barraclough@apple.com> 8887 8888 Reviewed by Geoff Garen. 8889 8890 Change UStringImpl::create to CRASH if the string cannot be allocated, 8891 rather than returning a null string (which will behave like a zero-length 8892 string if used). 8893 8894 Also move createRep function from UString to become new overloaded 8895 UStringImpl::create methods. In doing so, bring their behaviour closer to 8896 being in line with WebCore::StringImpl, in removing the behaviour that they 8897 can be used to produce null UStrings (ASSERT the char* provided is non-null). 8898 This behaviour of converting null C-strings to null UStrings is inefficient 8899 (cmompared to just using UString::null()), incompatible with WebCore::StringImpl's 8900 behaviour, and may generate unexpected behaviour, since in many cases a null 8901 UString can be used like an empty string. 8902 8903 With these changes UStringImpl need not have a concept of null impls, we can 8904 start transitioning this to become an implementation detail of UString, that 8905 internally it chooses to use a null-object rather than an actually zero impl 8906 pointer. 8907 8908 * JavaScriptCore.exp: 8909 * debugger/Debugger.cpp: 8910 (JSC::Debugger::recompileAllJSFunctions): 8911 * debugger/DebuggerCallFrame.cpp: 8912 (JSC::DebuggerCallFrame::calculatedFunctionName): 8913 * parser/Parser.cpp: 8914 (JSC::Parser::parse): 8915 * profiler/Profile.cpp: 8916 (JSC::Profile::Profile): 8917 * profiler/ProfileGenerator.cpp: 8918 (JSC::ProfileGenerator::stopProfiling): 8919 * runtime/Error.cpp: 8920 (JSC::Error::create): 8921 (JSC::throwError): 8922 * runtime/ExceptionHelpers.cpp: 8923 (JSC::createError): 8924 * runtime/Identifier.cpp: 8925 (JSC::Identifier::add): 8926 * runtime/PropertyNameArray.cpp: 8927 (JSC::PropertyNameArray::add): 8928 * runtime/UString.cpp: 8929 (JSC::initializeUString): 8930 (JSC::UString::UString): 8931 (JSC::UString::operator=): 8932 * runtime/UString.h: 8933 (JSC::UString::isNull): 8934 (JSC::UString::null): 8935 (JSC::UString::rep): 8936 (JSC::UString::UString): 8937 * runtime/UStringImpl.cpp: 8938 (JSC::UStringImpl::create): 8939 * runtime/UStringImpl.h: 8940 89412010-02-05 Kwang Yul Seo <skyul@company100.net> 8942 8943 Reviewed by Eric Seidel. 8944 8945 [BREWMP] Define SYSTEM_MALLOC 1 8946 https://bugs.webkit.org/show_bug.cgi?id=34640 8947 8948 Make BREWMP use system malloc because FastMalloc is not ported. 8949 8950 * wtf/Platform.h: 8951 89522010-02-05 Kwang Yul Seo <skyul@company100.net> 8953 8954 Reviewed by Alexey Proskuryakov. 8955 8956 Don't call CRASH() in fastMalloc and fastCalloc when the requested memory size is 0 8957 https://bugs.webkit.org/show_bug.cgi?id=34569 8958 8959 With USE_SYSTEM_MALLOC=1, fastMalloc and fastCalloc call CRASH() 8960 if the return value of malloc and calloc is 0. 8961 8962 However, these functions can return 0 when the request size is 0. 8963 Libc manual says, "If size is 0, then malloc() returns either NULL, 8964 or a unique pointer value that can later be successfully passed to free()." 8965 Though malloc returns a unique pointer in most systems, 8966 0 can be returned in some systems. For instance, BREW's MALLOC returns 0 8967 when size is 0. 8968 8969 If malloc or calloc returns 0 due to allocation size, increase the size 8970 to 1 and try again. 8971 8972 * wtf/FastMalloc.cpp: 8973 (WTF::fastMalloc): 8974 (WTF::fastCalloc): 8975 89762010-02-04 Mark Rowe <mrowe@apple.com> 8977 8978 Reviewed by Timothy Hatcher. 8979 8980 Build fix. Remove a symbol corresponding to an inline function from the linker export 8981 file to prevent a weak external failure. 8982 8983 * JavaScriptCore.xcodeproj/project.pbxproj: Accommodate rename of script. 8984 89852010-02-04 Daniel Bates <dbates@webkit.org> 8986 8987 [Qt] Unreviewed, build fix for Qt bot. 8988 8989 * runtime/JSStringBuilder.h: Changed #include <X.h> notation #include "X.h". 8990 89912010-02-04 Geoffrey Garen <ggaren@apple.com> 8992 8993 Reviewed by Oliver Hunt. 8994 8995 Clearing a WeakGCPtr is weird 8996 https://bugs.webkit.org/show_bug.cgi?id=34627 8997 8998 Added a WeakGCPtr::clear interface. 8999 9000 As discussed in https://bugs.webkit.org/show_bug.cgi?id=33383, the old 9001 interface made it pretty weird for a client to conditionally clear a 9002 WeakGCPtr, which is exactly what clients want to do when objects are 9003 finalized. 9004 9005 * API/JSClassRef.cpp: 9006 (clearReferenceToPrototype): Use the new WeakGCPtr::clear() interface. 9007 9008 * runtime/WeakGCPtr.h: 9009 (JSC::WeakGCPtr::clear): Added an interface for clearing a WeakGCPtr, 9010 iff its current value is the value passed in. It's cumbersome for the 9011 client to do this test, since WeakGCPtr sometimes pretends to be null. 9012 90132010-02-04 Geoffrey Garen <ggaren@apple.com> 9014 9015 Build fix: export a header. 9016 9017 * JavaScriptCore.xcodeproj/project.pbxproj: 9018 90192010-02-04 Gavin Barraclough <barraclough@apple.com> 9020 9021 Reviewed by Oliver Hunt. 9022 9023 Add a JSStringBuilder class (similar-to, and derived-from StringBuilder) to 9024 construct JSStrings, throwing a JS exception should we run out of memory whilst 9025 allocating storage for the string. 9026 9027 Similarly, add jsMakeNontrivialString methods to use in cases where previously 9028 we were calling makeString & passing the result to jsNontrivialString. Again, 9029 these new methods throw if we hit an out of memory condition. 9030 9031 Move throwOutOfMemoryError into ExceptionHelpers, to make it more widely available. 9032 9033 * JavaScriptCore.xcodeproj/project.pbxproj: 9034 * runtime/ArrayPrototype.cpp: 9035 (JSC::arrayProtoFuncToString): 9036 (JSC::arrayProtoFuncToLocaleString): 9037 (JSC::arrayProtoFuncJoin): 9038 * runtime/DateConstructor.cpp: 9039 (JSC::callDate): 9040 * runtime/DatePrototype.cpp: 9041 (JSC::dateProtoFuncToString): 9042 (JSC::dateProtoFuncToUTCString): 9043 (JSC::dateProtoFuncToGMTString): 9044 * runtime/ErrorPrototype.cpp: 9045 (JSC::errorProtoFuncToString): 9046 * runtime/ExceptionHelpers.cpp: 9047 (JSC::throwOutOfMemoryError): 9048 * runtime/ExceptionHelpers.h: 9049 * runtime/JSStringBuilder.h: Added. 9050 (JSC::JSStringBuilder::releaseJSString): 9051 (JSC::jsMakeNontrivialString): 9052 * runtime/NumberPrototype.cpp: 9053 (JSC::numberProtoFuncToPrecision): 9054 * runtime/ObjectPrototype.cpp: 9055 (JSC::objectProtoFuncToString): 9056 * runtime/Operations.cpp: 9057 * runtime/Operations.h: 9058 * runtime/RegExpPrototype.cpp: 9059 (JSC::regExpProtoFuncToString): 9060 * runtime/StringBuilder.h: 9061 (JSC::StringBuilder::append): 9062 * runtime/StringPrototype.cpp: 9063 (JSC::stringProtoFuncBig): 9064 (JSC::stringProtoFuncSmall): 9065 (JSC::stringProtoFuncBlink): 9066 (JSC::stringProtoFuncBold): 9067 (JSC::stringProtoFuncFixed): 9068 (JSC::stringProtoFuncItalics): 9069 (JSC::stringProtoFuncStrike): 9070 (JSC::stringProtoFuncSub): 9071 (JSC::stringProtoFuncSup): 9072 (JSC::stringProtoFuncFontcolor): 9073 (JSC::stringProtoFuncFontsize): 9074 (JSC::stringProtoFuncAnchor): 9075 90762010-02-04 Steve Falkenburg <sfalken@apple.com> 9077 9078 Windows build fix. 9079 9080 * wtf/MathExtras.h: 9081 90822010-02-04 Darin Adler <darin@apple.com> 9083 9084 Reviewed by David Levin. 9085 9086 Make MathExtras.h compatible with <cmath> 9087 https://bugs.webkit.org/show_bug.cgi?id=34618 9088 9089 * wtf/MathExtras.h: Include <cmath> instead of <math.h>. 9090 Use "using" as we do elsewhere in WTF for the four functions from <cmath> 9091 we want to use without the prefix. Later we could consider making the std 9092 explicit at call sites instead. 9093 90942010-02-04 Tamas Szirbucz <szirbucz@inf.u-szeged.hu> 9095 9096 Reviewed by Gavin Barraclough. 9097 9098 Use an easily appendable structure for trampolines instead of pointer parameters. 9099 https://bugs.webkit.org/show_bug.cgi?id=34424 9100 9101 * assembler/ARMAssembler.cpp: 9102 (JSC::ARMAssembler::executableCopy): 9103 * jit/JIT.h: 9104 (JSC::JIT::compileCTIMachineTrampolines): 9105 * jit/JITOpcodes.cpp: 9106 (JSC::JIT::privateCompileCTIMachineTrampolines): 9107 * jit/JITStubs.cpp: 9108 (JSC::JITThunks::JITThunks): 9109 * jit/JITStubs.h: 9110 (JSC::JITThunks::ctiStringLengthTrampoline): 9111 (JSC::JITThunks::ctiVirtualCallLink): 9112 (JSC::JITThunks::ctiVirtualCall): 9113 (JSC::JITThunks::ctiNativeCallThunk): 9114 91152010-02-04 Jedrzej Nowacki <jedrzej.nowacki@nokia.com> 9116 9117 Reviewed by Simon Hausmann. 9118 9119 Increase test coverage for the QScriptValue. 9120 9121 https://bugs.webkit.org/show_bug.cgi?id=34533 9122 9123 * qt/tests/qscriptvalue/qscriptvalue.pro: 9124 * qt/tests/qscriptvalue/tst_qscriptvalue.cpp: 9125 (tst_QScriptValue::tst_QScriptValue): 9126 (tst_QScriptValue::~tst_QScriptValue): 9127 (tst_QScriptValue::dataHelper): 9128 (tst_QScriptValue::newRow): 9129 (tst_QScriptValue::testHelper): 9130 (tst_QScriptValue::ctor): 9131 * qt/tests/qscriptvalue/tst_qscriptvalue.h: Added. 9132 * qt/tests/qscriptvalue/tst_qscriptvalue_generated.cpp: Added. 9133 (tst_QScriptValue::initScriptValues): 9134 (tst_QScriptValue::isValid_initData): 9135 (tst_QScriptValue::isValid_makeData): 9136 (tst_QScriptValue::isValid_test): 9137 (tst_QScriptValue::isBool_initData): 9138 (tst_QScriptValue::isBool_makeData): 9139 (tst_QScriptValue::isBool_test): 9140 (tst_QScriptValue::isBoolean_initData): 9141 (tst_QScriptValue::isBoolean_makeData): 9142 (tst_QScriptValue::isBoolean_test): 9143 (tst_QScriptValue::isFunction_initData): 9144 (tst_QScriptValue::isFunction_makeData): 9145 (tst_QScriptValue::isFunction_test): 9146 (tst_QScriptValue::isNull_initData): 9147 (tst_QScriptValue::isNull_makeData): 9148 (tst_QScriptValue::isNull_test): 9149 (tst_QScriptValue::isString_initData): 9150 (tst_QScriptValue::isString_makeData): 9151 (tst_QScriptValue::isString_test): 9152 (tst_QScriptValue::isUndefined_initData): 9153 (tst_QScriptValue::isUndefined_makeData): 9154 (tst_QScriptValue::isUndefined_test): 9155 (tst_QScriptValue::isObject_initData): 9156 (tst_QScriptValue::isObject_makeData): 9157 (tst_QScriptValue::isObject_test): 9158 91592010-02-03 Kwang Yul Seo <skyul@company100.net> 9160 9161 Reviewed by Eric Seidel. 9162 9163 [BREWMP] Define WTF_PLATFORM_BREWMP_SIMULATOR when AEE_SIMULATOR is defined 9164 https://bugs.webkit.org/show_bug.cgi?id=34514 9165 9166 PLATFORM(BREWMP_SIMULATOR) guard is needed to make distinction between BREWMP 9167 and BREWMP simulator. 9168 9169 * wtf/Platform.h: 9170 91712010-02-03 Kwang Yul Seo <skyul@company100.net> 9172 9173 Reviewed by Eric Seidel. 9174 9175 [BREWMP] Remove COMPILE_ASSERT conflict with the underlying PLATFORM 9176 https://bugs.webkit.org/show_bug.cgi?id=34190 9177 9178 COMPILE_ASSERT conflicts with the underlying PLATFORM because it is defined 9179 both in WTF's Assertions.h and BREWMP's AEEClassIDs.h. Include AEEClassIDs.h 9180 in Assertions.h and undef COMPILE_ASSERT to avoid redefining COMPILE_ASSERT. 9181 9182 * wtf/Assertions.h: 9183 91842010-02-03 Kwang Yul Seo <skyul@company100.net> 9185 9186 Reviewed by Eric Seidel. 9187 9188 [BREWMP] Implement OwnPtrBrew to make sure BREW instances are freed. 9189 https://bugs.webkit.org/show_bug.cgi?id=34518 9190 9191 Add OwnPtrBrew to release IFile, IFileMgr and IBitmap instances. 9192 9193 * wtf/brew/OwnPtrBrew.cpp: Added. 9194 (WTF::IFileMgr): 9195 (WTF::IFile): 9196 (WTF::IBitmap): 9197 (WTF::freeOwnedPtrBrew): 9198 * wtf/brew/OwnPtrBrew.h: Added. 9199 (WTF::OwnPtrBrew::OwnPtrBrew): 9200 (WTF::OwnPtrBrew::~OwnPtrBrew): 9201 (WTF::OwnPtrBrew::get): 9202 (WTF::OwnPtrBrew::release): 9203 (WTF::OwnPtrBrew::outPtr): 9204 (WTF::OwnPtrBrew::set): 9205 (WTF::OwnPtrBrew::clear): 9206 (WTF::OwnPtrBrew::operator*): 9207 (WTF::OwnPtrBrew::operator->): 9208 (WTF::OwnPtrBrew::operator!): 9209 (WTF::OwnPtrBrew::operator UnspecifiedBoolType): 9210 (WTF::OwnPtrBrew::swap): 9211 (WTF::swap): 9212 (WTF::operator==): 9213 (WTF::operator!=): 9214 (WTF::getPtr): 9215 92162010-02-03 Kwang Yul Seo <skyul@company100.net> 9217 9218 Reviewed by Darin Adler. 9219 9220 Export WTF::fastStrDup symbol 9221 https://bugs.webkit.org/show_bug.cgi?id=34526 9222 9223 * JavaScriptCore.exp: 9224 92252010-02-03 Kevin Watters <kevinwatters@gmail.com> 9226 9227 Reviewed by Kevin Ollivier. 9228 9229 [wx] Enable JIT compilation for wx. 9230 9231 https://bugs.webkit.org/show_bug.cgi?id=34536 9232 9233 * wtf/Platform.h: 9234 92352010-02-02 Oliver Hunt <oliver@apple.com> 9236 9237 Reviewed by Geoffrey Garen. 9238 9239 Crash in CollectorBitmap::get at nbcolympics.com 9240 https://bugs.webkit.org/show_bug.cgi?id=34504 9241 9242 This was caused by the use of m_offset to determine the offset of 9243 a new property into the property storage. This patch corrects 9244 the effected cases by incorporating the anonymous slot count. It 9245 also removes the duplicate copy of anonymous slot count from the 9246 property table as keeping this up to date merely increased the 9247 chance of a mismatch. Finally I've added a large number of 9248 assertions in an attempt to prevent such a bug from happening 9249 again. 9250 9251 With the new assertions in place the existing anonymous slot tests 9252 all fail without the m_offset fixes. 9253 9254 * runtime/PropertyMapHashTable.h: 9255 * runtime/Structure.cpp: 9256 (JSC::Structure::materializePropertyMap): 9257 (JSC::Structure::addPropertyTransitionToExistingStructure): 9258 (JSC::Structure::addPropertyTransition): 9259 (JSC::Structure::removePropertyTransition): 9260 (JSC::Structure::flattenDictionaryStructure): 9261 (JSC::Structure::addPropertyWithoutTransition): 9262 (JSC::Structure::removePropertyWithoutTransition): 9263 (JSC::Structure::copyPropertyTable): 9264 (JSC::Structure::get): 9265 (JSC::Structure::put): 9266 (JSC::Structure::remove): 9267 (JSC::Structure::insertIntoPropertyMapHashTable): 9268 (JSC::Structure::createPropertyMapHashTable): 9269 (JSC::Structure::rehashPropertyMapHashTable): 9270 (JSC::Structure::checkConsistency): 9271 92722010-02-02 Steve Falkenburg <sfalken@apple.com> 9273 9274 Reviewed by Darin Adler. 9275 9276 Copyright year updating for Windows version resources should be automatic 9277 https://bugs.webkit.org/show_bug.cgi?id=34503 9278 9279 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.rc: 9280 92812010-02-02 Kwang Yul Seo <skyul@company100.net> 9282 9283 Reviewed by Eric Seidel. 9284 9285 [BREWMP] Add dummy main thread functions 9286 https://bugs.webkit.org/show_bug.cgi?id=33569 9287 9288 Add dummy initializeMainThreadPlatform and 9289 scheduleDispatchFunctionsOnMainThread. 9290 9291 * wtf/brew/MainThreadBrew.cpp: Added. 9292 (WTF::initializeMainThreadPlatform): 9293 (WTF::scheduleDispatchFunctionsOnMainThread): 9294 92952010-02-02 Kwang Yul Seo <skyul@company100.net> 9296 9297 Reviewed by Darin Adler. 9298 9299 Add using WTF::getLocalTime to CurrentTime.h 9300 https://bugs.webkit.org/show_bug.cgi?id=34493 9301 9302 * wtf/CurrentTime.h: 9303 93042010-02-02 Kwang Yul Seo <skyul@company100.net> 9305 9306 Reviewed by Eric Seidel. 9307 9308 [BREWMP] Add HAVE_XXX definitions 9309 https://bugs.webkit.org/show_bug.cgi?id=34414 9310 9311 Add HAVE_ERRNO_H=1 9312 9313 * wtf/Platform.h: 9314 93152010-02-02 Kwang Yul Seo <skyul@company100.net> 9316 9317 Reviewed by Eric Seidel. 9318 9319 [BREWMP] Don't define HAVE_TM_GMTOFF, HAVE_TM_ZONE and HAVE_TIMEGM 9320 https://bugs.webkit.org/show_bug.cgi?id=34388 9321 9322 BREWMP does not have these features. 9323 9324 * wtf/Platform.h: 9325 93262010-02-02 Kwang Yul Seo <skyul@company100.net> 9327 9328 Reviewed by Eric Seidel. 9329 9330 [BREWMP] Define WTF_PLATFORM_BREWMP=1 when BUILDING_BREWMP is defined 9331 https://bugs.webkit.org/show_bug.cgi?id=34386 9332 9333 Define WTF_PLATFORM_BREWMP=1 so that PLATFORM(BREWMP) guard can be used. 9334 9335 * wtf/Platform.h: 9336 93372010-02-01 Kent Tamura <tkent@chromium.org> 9338 9339 Reviewed by Darin Adler. 9340 9341 Date.UTC() should apply TimeClip operation. 9342 https://bugs.webkit.org/show_bug.cgi?id=34461 9343 9344 ECMAScript 5 15.9.4.3: 9345 > 9 Return TimeClip(MakeDate(MakeDay(yr, m, dt), MakeTime(h, min, s, milli))). 9346 9347 * runtime/DateConstructor.cpp: 9348 (JSC::dateUTC): Calls WTF::timeClip(). 9349 93502010-02-01 Kent Tamura <tkent@chromium.org> 9351 9352 Reviewed by Darin Adler. 9353 9354 Fix a bug that Math.round() retunrs incorrect results for huge integers 9355 https://bugs.webkit.org/show_bug.cgi?id=34462 9356 9357 * runtime/MathObject.cpp: 9358 (JSC::mathProtoFuncRound): Avoid "arg + 0.5". 9359 93602010-02-01 Kwang Yul Seo <skyul@company100.net> 9361 9362 Reviewed by Eric Seidel. 9363 9364 [BREWMP] Port WTF's currentTime 9365 https://bugs.webkit.org/show_bug.cgi?id=33567 9366 9367 Combine GETUTCSECONDS and GETTIMEMS to calculate the number 9368 of milliseconds since 1970/01/01 00:00:00 UTC. 9369 9370 * wtf/CurrentTime.cpp: 9371 (WTF::currentTime): 9372 93732010-02-01 Patrick Gansterer <paroga@paroga.com> 9374 9375 Reviewed by Darin Adler. 9376 9377 [Qt] WinCE buildfix after r52729 and fix for Q_BIG_ENDIAN typo. 9378 https://bugs.webkit.org/show_bug.cgi?id=34378 9379 9380 * wtf/Platform.h: 9381 93822010-02-01 Oliver Hunt <oliver@apple.com> 9383 9384 Reviewed by Gavin Barraclough. 9385 9386 Structure not accounting for anonymous slots when computing property storage size 9387 https://bugs.webkit.org/show_bug.cgi?id=34441 9388 9389 Previously any Structure with anonymous storage would have a property map, so we 9390 were only including anonymous slot size if there was a property map. Given this 9391 is no longer the case we should always include the anonymous slot count in the 9392 property storage size. 9393 9394 * runtime/Structure.h: 9395 (JSC::Structure::propertyStorageSize): 9396 93972010-02-01 Oliver Hunt <oliver@apple.com> 9398 9399 Windows build fix, update exports file (again) 9400 9401 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 9402 94032010-02-01 Oliver Hunt <oliver@apple.com> 9404 9405 Windows build fix, update exports file 9406 9407 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 9408 94092010-01-31 Oliver Hunt <oliver@apple.com> 9410 9411 Reviewed by Maciej Stachowiak. 9412 9413 JSC is failing to propagate anonymous slot count on some transitions 9414 https://bugs.webkit.org/show_bug.cgi?id=34321 9415 9416 Remove secondary Structure constructor, and make Structure store a copy 9417 of the number of anonymous slots directly so saving an immediate allocation 9418 of a property map for all structures with anonymous storage, which also 9419 avoids the leaked property map on new property transition in the original 9420 version of this patch. 9421 9422 We need to propagate the the anonymous slot count otherwise we can end up 9423 with a structure recording incorrect information about the available and 9424 needed space for property storage, or alternatively incorrectly reusing 9425 some slots. 9426 9427 * JavaScriptCore.exp: 9428 * runtime/Structure.cpp: 9429 (JSC::Structure::Structure): 9430 (JSC::Structure::materializePropertyMap): 9431 (JSC::Structure::addPropertyTransition): 9432 (JSC::Structure::changePrototypeTransition): 9433 (JSC::Structure::despecifyFunctionTransition): 9434 (JSC::Structure::getterSetterTransition): 9435 (JSC::Structure::toDictionaryTransition): 9436 (JSC::Structure::flattenDictionaryStructure): 9437 (JSC::Structure::copyPropertyTable): 9438 (JSC::Structure::put): 9439 (JSC::Structure::remove): 9440 (JSC::Structure::insertIntoPropertyMapHashTable): 9441 (JSC::Structure::createPropertyMapHashTable): 9442 * runtime/Structure.h: 9443 (JSC::Structure::create): 9444 (JSC::Structure::hasAnonymousSlots): 9445 (JSC::Structure::anonymousSlotCount): 9446 94472010-01-31 Patrick Gansterer <paroga@paroga.com> 9448 9449 Reviewed by Darin Adler. 9450 9451 Buildfix for WinCE + style fixes (TLS_OUT_OF_INDEXES is not defined). 9452 https://bugs.webkit.org/show_bug.cgi?id=34380 9453 9454 * wtf/ThreadSpecific.h: 9455 94562010-01-31 Kent Tamura <tkent@chromium.org> 9457 9458 Reviewed by Darin Adler. 9459 9460 [Windows] Fix a bug of round() with huge integral numbers 9461 https://bugs.webkit.org/show_bug.cgi?id=34297 9462 9463 Fix a bug that round() for huge integral numbers returns incorrect 9464 results. For example, round(8639999913600001) returns 9465 8639999913600002 without this change though the double type can 9466 represent 8639999913600001 precisely. 9467 9468 Math.round() of JavaScript has a similar problem. But this change 9469 doesn't fix it because Math.round() doesn't use round() of 9470 MathExtra.h. 9471 9472 * wtf/MathExtras.h: 9473 (round): Avoid to do "num + 0.5" or "num - 0.5". 9474 (roundf): Fixed similarly. 9475 (llround): Calls round(). 9476 (llroundf): Calls roundf(). 9477 (lround): Calls round(). 9478 (lroundf): Calls roundf(). 9479 94802010-01-29 Mark Rowe <mrowe@apple.com> 9481 9482 Sort Xcode projects. 9483 9484 * JavaScriptCore.xcodeproj/project.pbxproj: 9485 94862010-01-29 Mark Rowe <mrowe@apple.com> 9487 9488 Fix the Mac build. 9489 9490 Disable ENABLE_INDEXED_DATABASE since it is "completely non-functional". 9491 9492 As the comment in FeatureDefines.xcconfig notes, the list of feature defines 9493 needs to be kept in sync across the various files. The default values also 9494 need to be kept in sync between these files and build-webkit. 9495 9496 * Configurations/FeatureDefines.xcconfig: 9497 94982010-01-29 Simon Hausmann <simon.hausmann@nokia.com> 9499 9500 Rubber-stamped by Maciej Stachowiak. 9501 9502 Fix the ARM build. 9503 9504 * runtime/JSNumberCell.h: 9505 (JSC::JSNumberCell::createStructure): Call the right Structure::create overload. 9506 95072010-01-28 Kevin Ollivier <kevino@theolliviers.com> 9508 9509 [wx] Build fix for MSW, use ThreadingWin.cpp as the Windows pthreads implementation 9510 implements pthread_t in a way that makes it impossible to check its validity, 9511 which is needed by ThreadingPthreads.cpp. 9512 9513 * wscript: 9514 95152010-01-28 Oliver Hunt <oliver@apple.com> 9516 9517 Reviewed by Gavin Barraclough. 9518 9519 DOM Objects shouldn't all require custom mark functions 9520 https://bugs.webkit.org/show_bug.cgi?id=34291 9521 9522 Make getAnonymousValue const-friendly 9523 9524 * runtime/JSObject.h: 9525 (JSC::JSObject::getAnonymousValue): 9526 95272010-01-28 Oliver Hunt <oliver@apple.com> 9528 9529 Reviewed by Gavin Barraclough. 9530 9531 Simplify anonymous slot implementation 9532 https://bugs.webkit.org/show_bug.cgi?id=34282 9533 9534 A class must now specify the number of slots it needs at construction time 9535 rather than later on with a transition. This makes many things simpler, 9536 we no longer need to need an additional transition on object creation to 9537 add the anonymous slots, and we remove the need for a number of transition 9538 type checks. 9539 9540 * API/JSCallbackConstructor.h: 9541 (JSC::JSCallbackConstructor::createStructure): 9542 * API/JSCallbackFunction.h: 9543 (JSC::JSCallbackFunction::createStructure): 9544 * API/JSCallbackObject.h: 9545 (JSC::JSCallbackObject::createStructure): 9546 * JavaScriptCore.exp: 9547 * debugger/DebuggerActivation.h: 9548 (JSC::DebuggerActivation::createStructure): 9549 * runtime/Arguments.h: 9550 (JSC::Arguments::createStructure): 9551 * runtime/BooleanObject.h: 9552 (JSC::BooleanObject::createStructure): 9553 * runtime/DateInstance.h: 9554 (JSC::DateInstance::createStructure): 9555 * runtime/DatePrototype.h: 9556 (JSC::DatePrototype::createStructure): 9557 * runtime/FunctionPrototype.h: 9558 (JSC::FunctionPrototype::createStructure): 9559 * runtime/GetterSetter.h: 9560 (JSC::GetterSetter::createStructure): 9561 * runtime/GlobalEvalFunction.h: 9562 (JSC::GlobalEvalFunction::createStructure): 9563 * runtime/InternalFunction.h: 9564 (JSC::InternalFunction::createStructure): 9565 * runtime/JSAPIValueWrapper.h: 9566 (JSC::JSAPIValueWrapper::createStructure): 9567 * runtime/JSActivation.h: 9568 (JSC::JSActivation::createStructure): 9569 * runtime/JSArray.h: 9570 (JSC::JSArray::createStructure): 9571 * runtime/JSByteArray.cpp: 9572 (JSC::JSByteArray::createStructure): 9573 * runtime/JSCell.h: 9574 (JSC::JSCell::createDummyStructure): 9575 * runtime/JSFunction.h: 9576 (JSC::JSFunction::createStructure): 9577 * runtime/JSGlobalObject.h: 9578 (JSC::JSGlobalObject::createStructure): 9579 * runtime/JSNotAnObject.h: 9580 (JSC::JSNotAnObject::createStructure): 9581 * runtime/JSONObject.h: 9582 (JSC::JSONObject::createStructure): 9583 * runtime/JSObject.h: 9584 (JSC::JSObject::createStructure): 9585 (JSC::JSObject::putAnonymousValue): 9586 (JSC::JSObject::getAnonymousValue): 9587 * runtime/JSPropertyNameIterator.h: 9588 (JSC::JSPropertyNameIterator::createStructure): 9589 * runtime/JSStaticScopeObject.h: 9590 (JSC::JSStaticScopeObject::createStructure): 9591 * runtime/JSString.h: 9592 (JSC::Fiber::createStructure): 9593 * runtime/JSVariableObject.h: 9594 (JSC::JSVariableObject::createStructure): 9595 * runtime/JSWrapperObject.h: 9596 (JSC::JSWrapperObject::createStructure): 9597 (JSC::JSWrapperObject::JSWrapperObject): 9598 * runtime/MathObject.h: 9599 (JSC::MathObject::createStructure): 9600 * runtime/NumberConstructor.h: 9601 (JSC::NumberConstructor::createStructure): 9602 * runtime/NumberObject.h: 9603 (JSC::NumberObject::createStructure): 9604 * runtime/RegExpConstructor.h: 9605 (JSC::RegExpConstructor::createStructure): 9606 * runtime/RegExpObject.h: 9607 (JSC::RegExpObject::createStructure): 9608 * runtime/StringObject.h: 9609 (JSC::StringObject::createStructure): 9610 * runtime/StringObjectThatMasqueradesAsUndefined.h: 9611 (JSC::StringObjectThatMasqueradesAsUndefined::createStructure): 9612 * runtime/Structure.cpp: 9613 (JSC::Structure::~Structure): 9614 (JSC::Structure::materializePropertyMap): 9615 * runtime/Structure.h: 9616 (JSC::Structure::create): 9617 (JSC::Structure::anonymousSlotCount): 9618 * runtime/StructureTransitionTable.h: 9619 96202010-01-27 Oliver Hunt <oliver@apple.com> 9621 9622 Windows build fix. 9623 9624 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 9625 96262010-01-27 Oliver Hunt <oliver@apple.com> 9627 9628 Reviewed by Maciej Stachowiak. 9629 9630 MessageEvent.data should deserialize in the context of the MessageEvent's global object 9631 https://bugs.webkit.org/show_bug.cgi?id=34227 9632 9633 Add logic to allow us to create an Object, Array, or Date instance 9634 so we can create them in the context of a specific global object, 9635 rather than just using the current lexical global object. 9636 9637 * JavaScriptCore.exp: 9638 * runtime/DateInstance.cpp: 9639 (JSC::DateInstance::DateInstance): 9640 * runtime/DateInstance.h: 9641 * runtime/JSGlobalObject.h: 9642 (JSC::constructEmptyObject): 9643 (JSC::constructEmptyArray): 9644 96452010-01-27 Alexey Proskuryakov <ap@apple.com> 9646 9647 Reviewed by Darin Adler. 9648 9649 https://bugs.webkit.org/show_bug.cgi?id=34150 9650 WebKit needs a mechanism to catch stale HashMap entries 9651 9652 It is very difficult to catch stale pointers that are HashMap keys - since a pointer's hash 9653 is just its value, it is very unlikely that any observable problem is reproducible. 9654 9655 This extends hash table consistency checks to check that pointers are referencing allocated 9656 memory blocks, and makes it possible to invoke the checks explicitly (it is not feasible 9657 to enable CHECK_HASHTABLE_CONSISTENCY by default, because that affects performance too much). 9658 9659 * wtf/HashMap.h: (WTF::::checkConsistency): Call through to HashTable implementation. We can 9660 add similar calls to HashSet and HashCountedSet, but I haven't seen hard to debug problems 9661 with those yet. 9662 9663 * wtf/HashSet.h: (WTF::::remove): The version of checkTableConsistency that's guarded by 9664 CHECK_HASHTABLE_CONSISTENCY is now called internalCheckTableConsistency(). 9665 9666 * wtf/HashTable.h: 9667 (WTF::HashTable::internalCheckTableConsistency): 9668 (WTF::HashTable::internalCheckTableConsistencyExceptSize): 9669 (WTF::HashTable::checkTableConsistencyExceptSize): 9670 Expose checkTableConsistency() even if CHECK_HASHTABLE_CONSISTENCY is off. 9671 (WTF::::add): Updated for checkTableConsistency renaming. 9672 (WTF::::addPassingHashCode): Ditto. 9673 (WTF::::removeAndInvalidate): Ditto. 9674 (WTF::::remove): Ditto. 9675 (WTF::::rehash): Ditto. 9676 (WTF::::checkTableConsistency): The assertion for !shouldExpand() was not correct - this 9677 function returns true for tables with m_table == 0. 9678 (WTF::::checkTableConsistencyExceptSize): Call checkValueConsistency for key. Potentially, 9679 we could do the same for values. 9680 9681 * wtf/HashTraits.h: 9682 (WTF::GenericHashTraits::checkValueConsistency): An empty function that can be overridden 9683 to add checks. Currently, the only override is for pointer hashes. 9684 9685 * wtf/RefPtrHashMap.h: (WTF::::remove): Updated for checkTableConsistency renaming. 9686 96872010-01-27 Anton Muhin <antonm@chromium.org> 9688 9689 Reviewed by Darin Adler. 9690 9691 Remove trailing \ from inline function code 9692 https://bugs.webkit.org/show_bug.cgi?id=34223 9693 9694 * assembler/ARMv7Assembler.h: 9695 (JSC::ARMThumbImmediate::countLeadingZerosPartial): 9696 96972010-01-27 Kwang Yul Seo <skyul@company100.net> 9698 9699 Reviewed by Eric Seidel. 9700 9701 [BREWMP] Port WTF's randomNumber 9702 https://bugs.webkit.org/show_bug.cgi?id=33566 9703 9704 Use GETRAND to generate 4 byte random byte sequence to implement 9705 weakRandomNumber. Create a secure random number generator with 9706 AEECLSID_RANDOM to implement randomNumber. 9707 9708 * wtf/RandomNumber.cpp: 9709 (WTF::weakRandomNumber): 9710 (WTF::randomNumber): 9711 97122010-01-27 Kwang Yul Seo <skyul@company100.net> 9713 9714 Reviewed by Eric Seidel. 9715 9716 [BREWMP] Port getCPUTime 9717 https://bugs.webkit.org/show_bug.cgi?id=33572 9718 9719 Use GETUPTIMEMS which returns a continuously and 9720 linearly increasing millisecond timer from the time the device 9721 was powered on. This function is enough to implement getCPUTime. 9722 9723 * runtime/TimeoutChecker.cpp: 9724 (JSC::getCPUTime): 9725 97262010-01-27 Kwang Yul Seo <skyul@company100.net> 9727 9728 Reviewed by Oliver Hunt. 9729 9730 [BREWMP] Add MarkStack fastMalloc implementation for platforms without VirtualAlloc or mmap. 9731 https://bugs.webkit.org/show_bug.cgi?id=33582 9732 9733 Use fastMalloc and fastFree to implement MarkStack::allocateStack and 9734 MarkStack::releaseStack for platforms without page level allocation. 9735 9736 * runtime/MarkStack.h: 9737 (JSC::MarkStack::MarkStackArray::shrinkAllocation): 9738 * runtime/MarkStackNone.cpp: Added. 9739 (JSC::MarkStack::initializePagesize): 9740 (JSC::MarkStack::allocateStack): 9741 (JSC::MarkStack::releaseStack): 9742 97432010-01-27 Kwang Yul Seo <skyul@company100.net> 9744 9745 Reviewed by Eric Seidel. 9746 9747 [BREWMP] Don't use time function 9748 https://bugs.webkit.org/show_bug.cgi?id=33577 9749 9750 Calling time(0) in BREW devices causes a crash because time 9751 is not properly ported in most devices. Cast currentTime() to 9752 time_t to get the same result as time(0). 9753 9754 * wtf/DateMath.cpp: 9755 (WTF::calculateUTCOffset): 9756 97572010-01-27 Alexey Proskuryakov <ap@apple.com> 9758 9759 Revert r53899 (HashMap<AtomicStringImpl*, Value> key checks) and subsequent build fixes, 9760 because they make SVG tests crash in release builds. 9761 9762 * wtf/HashMap.h: 9763 (WTF::::remove): 9764 * wtf/HashSet.h: 9765 (WTF::::remove): 9766 * wtf/HashTable.h: 9767 (WTF::::add): 9768 (WTF::::addPassingHashCode): 9769 (WTF::::removeAndInvalidate): 9770 (WTF::::remove): 9771 (WTF::::rehash): 9772 (WTF::::checkTableConsistency): 9773 (WTF::::checkTableConsistencyExceptSize): 9774 * wtf/HashTraits.h: 9775 (WTF::GenericHashTraits::emptyValue): 9776 (WTF::): 9777 * wtf/RefPtrHashMap.h: 9778 (WTF::::remove): 9779 97802010-01-26 Alexey Proskuryakov <ap@apple.com> 9781 9782 More Windows build fixing. 9783 9784 * wtf/HashTraits.h: _msize takes void*, remove const qualifier from type. 9785 97862010-01-26 Alexey Proskuryakov <ap@apple.com> 9787 9788 Windows build fix. 9789 9790 * wtf/HashTraits.h: Include malloc.h for _msize(). 9791 97922010-01-26 Alexey Proskuryakov <ap@apple.com> 9793 9794 Build fix. 9795 9796 * wtf/HashTable.h: (WTF::HashTable::checkTableConsistencyExceptSize): Remove const from a 9797 static (empty) version of this function. 9798 97992010-01-26 Alexey Proskuryakov <ap@apple.com> 9800 9801 Reviewed by Darin Adler. 9802 9803 https://bugs.webkit.org/show_bug.cgi?id=34150 9804 WebKit needs a mechanism to catch stale HashMap entries 9805 9806 It is very difficult to catch stale pointers that are HashMap keys - since a pointer's hash 9807 is just its value, it is very unlikely that any observable problem is reproducible. 9808 9809 This extends hash table consistency checks to check that pointers are referencing allocated 9810 memory blocks, and makes it possible to invoke the checks explicitly (it is not feasible 9811 to enable CHECK_HASHTABLE_CONSISTENCY by default, because that affects performance too much). 9812 9813 * wtf/HashMap.h: (WTF::::checkConsistency): Call through to HashTable implementation. We can 9814 add similar calls to HashSet and HashCountedSet, but I haven't seen hard to debug problems 9815 with those yet. 9816 9817 * wtf/HashSet.h: (WTF::::remove): The version of checkTableConsistency that's guarded by 9818 CHECK_HASHTABLE_CONSISTENCY is now called internalCheckTableConsistency(). 9819 9820 * wtf/HashTable.h: 9821 (WTF::HashTable::internalCheckTableConsistency): 9822 (WTF::HashTable::internalCheckTableConsistencyExceptSize): 9823 (WTF::HashTable::checkTableConsistencyExceptSize): 9824 Expose checkTableConsistency() even if CHECK_HASHTABLE_CONSISTENCY is off. 9825 (WTF::::add): Updated for checkTableConsistency renaming. 9826 (WTF::::addPassingHashCode): Ditto. 9827 (WTF::::removeAndInvalidate): Ditto. 9828 (WTF::::remove): Ditto. 9829 (WTF::::rehash): Ditto. 9830 (WTF::::checkTableConsistency): The assertion for !shouldExpand() was not correct - this 9831 function returns true for tables with m_table == 0. 9832 (WTF::::checkTableConsistencyExceptSize): Call checkValueConsistency for key. Potentially, 9833 we could do the same for values. 9834 9835 * wtf/HashTraits.h: 9836 (WTF::GenericHashTraits::checkValueConsistency): An empty function that can be overridden 9837 to add checks. Currently, the only override is for pointer hashes. 9838 9839 * wtf/RefPtrHashMap.h: (WTF::::remove): Updated for checkTableConsistency renaming. 9840 98412010-01-26 Lyon Chen <liachen@rim.com> 9842 9843 Reviewed by Maciej Stachowiak. 9844 9845 Opcode.h use const void* for Opcode cause error #1211 for RVCT compiler 9846 https://bugs.webkit.org/show_bug.cgi?id=33902 9847 9848 * bytecode/Opcode.h: 9849 98502010-01-26 Steve Falkenburg <sfalken@apple.com> 9851 9852 Reviewed by Oliver Hunt. 9853 9854 Windows build references non-existent include paths 9855 https://bugs.webkit.org/show_bug.cgi?id=34175 9856 9857 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: 9858 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCoreCommon.vsprops: 9859 * JavaScriptCore.vcproj/WTF/WTFCommon.vsprops: 9860 * JavaScriptCore.vcproj/jsc/jscCommon.vsprops: 9861 * JavaScriptCore.vcproj/testapi/testapi.vcproj: 9862 * JavaScriptCore.vcproj/testapi/testapiCommon.vsprops: 9863 98642010-01-26 Oliver Hunt <oliver@apple.com> 9865 9866 Reviewed by Geoffrey Garen. 9867 9868 Using JavaScriptCore API with a webkit vended context can result in slow script dialog 9869 https://bugs.webkit.org/show_bug.cgi?id=34172 9870 9871 Make the APIShim correctly increment and decrement the timeout 9872 entry counter. 9873 9874 * API/APIShims.h: 9875 (JSC::APIEntryShimWithoutLock::APIEntryShimWithoutLock): 9876 (JSC::APIEntryShimWithoutLock::~APIEntryShimWithoutLock): 9877 (JSC::APICallbackShim::APICallbackShim): 9878 (JSC::APICallbackShim::~APICallbackShim): 9879 98802010-01-26 Simon Hausmann <simon.hausmann@nokia.com> 9881 9882 [Qt] Fix compilation of QtScript with non-gcc compilers 9883 9884 Variable length stack arrays are a gcc extension. Use QVarLengthArray 9885 as a more portable solution that still tries to allocate on the stack 9886 first. 9887 9888 * qt/api/qscriptvalue_p.h: 9889 (QScriptValuePrivate::call): 9890 98912010-01-26 Simon Hausmann <simon.hausmann@nokia.com> 9892 9893 Reviewed by Tor Arne Vestbø. 9894 9895 [Qt] Fix the build on platforms without JIT support. 9896 9897 The JIT support should be determined at compile-time via wtf/Platform.h 9898 9899 * qt/api/QtScript.pro: 9900 99012010-01-26 Jedrzej Nowacki <jedrzej.nowacki@nokia.com> 9902 9903 Reviewed by Simon Hausmann. 9904 9905 First steps of the QtScript API. 9906 9907 Two new classes were created; QScriptEngine and QScriptValue. 9908 The first should encapsulate a javascript context and the second a script 9909 value. 9910 9911 This API is still in development, so it isn't compiled by default. 9912 To trigger compilation, pass --qmakearg="CONFIG+=build-qtscript" to 9913 build-webkit. 9914 9915 https://bugs.webkit.org/show_bug.cgi?id=32565 9916 9917 * qt/api/QtScript.pro: Added. 9918 * qt/api/qscriptconverter_p.h: Added. 9919 (QScriptConverter::toString): 9920 * qt/api/qscriptengine.cpp: Added. 9921 (QScriptEngine::QScriptEngine): 9922 (QScriptEngine::~QScriptEngine): 9923 (QScriptEngine::evaluate): 9924 (QScriptEngine::collectGarbage): 9925 * qt/api/qscriptengine.h: Added. 9926 * qt/api/qscriptengine_p.cpp: Added. 9927 (QScriptEnginePrivate::QScriptEnginePrivate): 9928 (QScriptEnginePrivate::~QScriptEnginePrivate): 9929 (QScriptEnginePrivate::evaluate): 9930 * qt/api/qscriptengine_p.h: Added. 9931 (QScriptEnginePrivate::get): 9932 (QScriptEnginePrivate::collectGarbage): 9933 (QScriptEnginePrivate::makeJSValue): 9934 (QScriptEnginePrivate::context): 9935 * qt/api/qscriptvalue.cpp: Added. 9936 (QScriptValue::QScriptValue): 9937 (QScriptValue::~QScriptValue): 9938 (QScriptValue::isValid): 9939 (QScriptValue::isBool): 9940 (QScriptValue::isBoolean): 9941 (QScriptValue::isNumber): 9942 (QScriptValue::isNull): 9943 (QScriptValue::isString): 9944 (QScriptValue::isUndefined): 9945 (QScriptValue::isError): 9946 (QScriptValue::isObject): 9947 (QScriptValue::isFunction): 9948 (QScriptValue::toString): 9949 (QScriptValue::toNumber): 9950 (QScriptValue::toBool): 9951 (QScriptValue::toBoolean): 9952 (QScriptValue::toInteger): 9953 (QScriptValue::toInt32): 9954 (QScriptValue::toUInt32): 9955 (QScriptValue::toUInt16): 9956 (QScriptValue::call): 9957 (QScriptValue::engine): 9958 (QScriptValue::operator=): 9959 (QScriptValue::equals): 9960 (QScriptValue::strictlyEquals): 9961 * qt/api/qscriptvalue.h: Added. 9962 (QScriptValue::): 9963 * qt/api/qscriptvalue_p.h: Added. 9964 (QScriptValuePrivate::): 9965 (QScriptValuePrivate::get): 9966 (QScriptValuePrivate::QScriptValuePrivate): 9967 (QScriptValuePrivate::isValid): 9968 (QScriptValuePrivate::isBool): 9969 (QScriptValuePrivate::isNumber): 9970 (QScriptValuePrivate::isNull): 9971 (QScriptValuePrivate::isString): 9972 (QScriptValuePrivate::isUndefined): 9973 (QScriptValuePrivate::isError): 9974 (QScriptValuePrivate::isObject): 9975 (QScriptValuePrivate::isFunction): 9976 (QScriptValuePrivate::toString): 9977 (QScriptValuePrivate::toNumber): 9978 (QScriptValuePrivate::toBool): 9979 (QScriptValuePrivate::toInteger): 9980 (QScriptValuePrivate::toInt32): 9981 (QScriptValuePrivate::toUInt32): 9982 (QScriptValuePrivate::toUInt16): 9983 (QScriptValuePrivate::equals): 9984 (QScriptValuePrivate::strictlyEquals): 9985 (QScriptValuePrivate::assignEngine): 9986 (QScriptValuePrivate::call): 9987 (QScriptValuePrivate::engine): 9988 (QScriptValuePrivate::context): 9989 (QScriptValuePrivate::value): 9990 (QScriptValuePrivate::object): 9991 (QScriptValuePrivate::inherits): 9992 (QScriptValuePrivate::isJSBased): 9993 (QScriptValuePrivate::isNumberBased): 9994 (QScriptValuePrivate::isStringBased): 9995 * qt/api/qtscriptglobal.h: Added. 9996 * qt/tests/qscriptengine/qscriptengine.pro: Added. 9997 * qt/tests/qscriptengine/tst_qscriptengine.cpp: Added. 9998 (tst_QScriptEngine::tst_QScriptEngine): 9999 (tst_QScriptEngine::~tst_QScriptEngine): 10000 (tst_QScriptEngine::init): 10001 (tst_QScriptEngine::cleanup): 10002 (tst_QScriptEngine::collectGarbage): 10003 (tst_QScriptEngine::evaluate): 10004 * qt/tests/qscriptvalue/qscriptvalue.pro: Added. 10005 * qt/tests/qscriptvalue/tst_qscriptvalue.cpp: Added. 10006 (tst_QScriptValue::tst_QScriptValue): 10007 (tst_QScriptValue::~tst_QScriptValue): 10008 (tst_QScriptValue::init): 10009 (tst_QScriptValue::cleanup): 10010 (tst_QScriptValue::ctor): 10011 (tst_QScriptValue::toString_data): 10012 (tst_QScriptValue::toString): 10013 (tst_QScriptValue::copyConstructor_data): 10014 (tst_QScriptValue::copyConstructor): 10015 (tst_QScriptValue::assignOperator_data): 10016 (tst_QScriptValue::assignOperator): 10017 (tst_QScriptValue::dataSharing): 10018 (tst_QScriptValue::constructors_data): 10019 (tst_QScriptValue::constructors): 10020 (tst_QScriptValue::call): 10021 * qt/tests/tests.pri: Added. 10022 * qt/tests/tests.pro: Added. 10023 100242010-01-25 Dmitry Titov <dimich@chromium.org> 10025 10026 Reviewed by David Levin. 10027 10028 Fix Chromium Linux tests: the pthread functions on Linux produce segfault if they receive 0 thread handle. 10029 After r53714, we can have 0 thread handles passed to pthread_join and pthread_detach if corresponding threads 10030 were already terminated and their threadMap entries cleared. 10031 Add a 0 check. 10032 10033 * wtf/ThreadingPthreads.cpp: 10034 (WTF::waitForThreadCompletion): 10035 (WTF::detachThread): 10036 100372010-01-24 Laszlo Gombos <laszlo.1.gombos@nokia.com> 10038 10039 Reviewed by Maciej Stachowiak. 10040 10041 Refactor JITStubs.cpp so that DEFINE_STUB_FUNCTION is only used once for each function 10042 https://bugs.webkit.org/show_bug.cgi?id=33866 10043 10044 Place the guard USE(JSVALUE32_64) inside the body of the DEFINE_STUB_FUNCTION 10045 macro for those functions that are always present. 10046 10047 * jit/JITStubs.cpp: 10048 (JSC::DEFINE_STUB_FUNCTION): 10049 100502010-01-22 Kevin Watters <kevinwatters@gmail.com> 10051 10052 Reviewed by Kevin Ollivier. 10053 10054 [wx] Remove the Bakefile build system, which is no longer being used. 10055 10056 https://bugs.webkit.org/show_bug.cgi?id=34022 10057 10058 * JavaScriptCoreSources.bkl: Removed. 10059 * jscore.bkl: Removed. 10060 100612010-01-22 Steve Falkenburg <sfalken@apple.com> 10062 10063 Reviewed by Darin Adler. 10064 10065 https://bugs.webkit.org/show_bug.cgi?id=34025 10066 Enable client-based Geolocation abstraction for Mac, Windows AppleWebKit targets. 10067 10068 * Configurations/FeatureDefines.xcconfig: 10069 100702010-01-22 Dmitry Titov <dimich@chromium.org> 10071 10072 Not reviewed, attempted Snow Leopard build fix. 10073 10074 * wtf/ThreadingPthreads.cpp: Add a forward declaration of a function which is not 'static'. 10075 100762009-01-22 Dmitry Titov <dimich@chromium.org> 10077 10078 Reviewed by Maciej Stachowiak. 10079 10080 Fix the leak of ThreadIdentifiers in threadMap across threads. 10081 https://bugs.webkit.org/show_bug.cgi?id=32689 10082 10083 Test is added to DumpRenderTree.mm. 10084 10085 * Android.mk: Added file ThreadIdentifierDataPthreads.(h|cpp) to build. 10086 * Android.v8.wtf.mk: Ditto. 10087 * GNUmakefile.am: Ditto. 10088 * JavaScriptCore.gyp/JavaScriptCore.gyp: Ditto. 10089 * JavaScriptCore.gypi: Ditto. 10090 * JavaScriptCore.xcodeproj/project.pbxproj: Ditto. 10091 10092 * wtf/ThreadIdentifierDataPthreads.cpp: Added. Contains custom implementation of thread-specific data that uses custom destructor. 10093 (WTF::ThreadIdentifierData::~ThreadIdentifierData): Removes the ThreadIdentifier from the threadMap. 10094 (WTF::ThreadIdentifierData::identifier): 10095 (WTF::ThreadIdentifierData::initialize): 10096 (WTF::ThreadIdentifierData::destruct): Custom thread-specific destructor. Resets the value for the key again to cause second invoke. 10097 (WTF::ThreadIdentifierData::initializeKeyOnceHelper): 10098 (WTF::ThreadIdentifierData::initializeKeyOnce): Need to use pthread_once since initialization may come on any thread(s). 10099 * wtf/ThreadIdentifierDataPthreads.h: Added. 10100 (WTF::ThreadIdentifierData::ThreadIdentifierData): 10101 10102 * wtf/Threading.cpp: 10103 (WTF::threadEntryPoint): Move initializeCurrentThreadInternal to after the lock to make 10104 sure it is invoked when ThreadIdentifier is already established. 10105 10106 * wtf/Threading.h: Rename setThreadNameInternal -> initializeCurrentThreadInternal since it does more then only set the name now. 10107 * wtf/ThreadingNone.cpp: 10108 (WTF::initializeCurrentThreadInternal): Ditto. 10109 * wtf/ThreadingWin.cpp: 10110 (WTF::initializeCurrentThreadInternal): Ditto. 10111 (WTF::initializeThreading): Ditto. 10112 * wtf/gtk/ThreadingGtk.cpp: 10113 (WTF::initializeCurrentThreadInternal): Ditto. 10114 * wtf/qt/ThreadingQt.cpp: 10115 (WTF::initializeCurrentThreadInternal): Ditto. 10116 10117 * wtf/ThreadingPthreads.cpp: 10118 (WTF::establishIdentifierForPthreadHandle): 10119 (WTF::clearPthreadHandleForIdentifier): Make it not 'static' so the ~ThreadIdentifierData() in another file can call it. 10120 (WTF::initializeCurrentThreadInternal): Set the thread-specific data. The ThreadIdentifier is already established by creating thread. 10121 (WTF::waitForThreadCompletion): Remove call to clearPthreadHandleForIdentifier(threadID) since it is now done in ~ThreadIdentifierData(). 10122 (WTF::detachThread): Ditto. 10123 (WTF::currentThread): Use the thread-specific data to get the ThreadIdentifier. It's many times faster then Mutex-protected iteration through the map. 10124 Also, set the thread-specific data if called first time on the thread. 10125 101262010-01-21 Kwang Yul Seo <skyul@company100.net> 10127 10128 Reviewed by Alexey Proskuryakov. 10129 10130 Add ThreadSpecific for ENABLE(SINGLE_THREADED) 10131 https://bugs.webkit.org/show_bug.cgi?id=33878 10132 10133 Implement ThreadSpecific with a simple getter/setter 10134 when ENABLE(SINGLE_THREADED) is true. 10135 10136 Due to the change in https://bugs.webkit.org/show_bug.cgi?id=33236, 10137 an implementation of ThreadSpecific must be available to build WebKit. 10138 This causes a build failure for platforms without a proper 10139 ThreadSpecific implementation. 10140 10141 * wtf/ThreadSpecific.h: 10142 (WTF::::ThreadSpecific): 10143 (WTF::::~ThreadSpecific): 10144 (WTF::::get): 10145 (WTF::::set): 10146 (WTF::::destroy): 10147 101482010-01-21 Kwang Yul Seo <skyul@company100.net> 10149 10150 Reviewed by Maciej Stachowiak. 10151 10152 Add fastStrDup to FastMalloc 10153 https://bugs.webkit.org/show_bug.cgi?id=33937 10154 10155 The new string returned by fastStrDup is obtained with fastMalloc, 10156 and can be freed with fastFree. This makes the memory management 10157 more consistent because we don't need to keep strdup allocated pointers 10158 and free them with free(). Instead we can use fastFree everywhere. 10159 10160 * wtf/FastMalloc.cpp: 10161 (WTF::fastStrDup): 10162 * wtf/FastMalloc.h: 10163 101642010-01-21 Brady Eidson <beidson@apple.com> 10165 10166 Reviewed by Maciej Stachowiak. 10167 10168 history.back() for same-document history traversals isn't synchronous as the specification states. 10169 <rdar://problem/7535011> and https://bugs.webkit.org/show_bug.cgi?id=33538 10170 10171 * wtf/Platform.h: Add a "HISTORY_ALWAYS_ASYNC" enable and turn it on for Chromium. 10172 101732010-01-21 Geoffrey Garen <ggaren@apple.com> 10174 10175 Reviewed by Oliver Hunt. 10176 10177 Always create a prototype for automatically managed classes. 10178 10179 This fixes some errors where prototype chains were not correctly hooked 10180 up, and also ensures that API classes work correctly with features like 10181 instanceof. 10182 10183 * API/JSClassRef.cpp: 10184 (OpaqueJSClass::create): Cleaned up some of this code. Also changed it 10185 to always create a prototype class. 10186 10187 * API/tests/testapi.c: 10188 (Derived2_class): 10189 (main): Fixed a null value crash in the exception checking code. 10190 * API/tests/testapi.js: Added some tests for the case where a prototype 10191 chain would not be hooked up correctly. 10192 101932010-01-21 Oliver Hunt <oliver@apple.com> 10194 10195 Reviewed by Geoff Garen. 10196 10197 Force JSC to create a prototype chain for API classes with a 10198 parent class but no static functions. 10199 10200 * API/JSClassRef.cpp: 10201 (OpaqueJSClass::create): 10202 102032010-01-21 Kent Hansen <kent.hansen@nokia.com> 10204 10205 Reviewed by Geoffrey Garen. 10206 10207 Object.getOwnPropertyDescriptor always returns undefined for JS API objects 10208 https://bugs.webkit.org/show_bug.cgi?id=33946 10209 10210 Ideally the getOwnPropertyDescriptor() reimplementation should return an 10211 access descriptor that wraps the property getter and setter callbacks, but 10212 that approach is much more involved than returning a value descriptor. 10213 Keep it simple for now. 10214 10215 * API/JSCallbackObject.h: 10216 * API/JSCallbackObjectFunctions.h: 10217 (JSC::::getOwnPropertyDescriptor): 10218 * API/tests/testapi.js: 10219 102202010-01-20 Mark Rowe <mrowe@apple.com> 10221 10222 Build fix. 10223 10224 * wtf/FastMalloc.cpp: 10225 (WTF::TCMalloc_PageHeap::initializeScavenger): Remove unnecessary function call. 10226 102272010-01-20 Mark Rowe <mrowe@apple.com> 10228 10229 Reviewed by Oliver Hunt. 10230 10231 Use the inline i386 assembly for x86_64 as well rather than falling back to using pthread mutexes. 10232 10233 * wtf/TCSpinLock.h: 10234 (TCMalloc_SpinLock::Lock): 10235 (TCMalloc_SpinLock::Unlock): 10236 (TCMalloc_SlowLock): 10237 102382010-01-20 Mark Rowe <mrowe@apple.com> 10239 10240 Reviewed by Oliver Hunt. 10241 10242 <rdar://problem/7215063> Use GCD instead of an extra thread for FastMalloc scavenging on platforms where it is supported 10243 10244 Abstract the background scavenging slightly so that an alternate implementation that uses GCD can be used on platforms 10245 where it is supported. 10246 10247 * wtf/FastMalloc.cpp: 10248 (WTF::TCMalloc_PageHeap::init): 10249 (WTF::TCMalloc_PageHeap::initializeScavenger): 10250 (WTF::TCMalloc_PageHeap::signalScavenger): 10251 (WTF::TCMalloc_PageHeap::shouldContinueScavenging): 10252 (WTF::TCMalloc_PageHeap::Delete): 10253 (WTF::TCMalloc_PageHeap::periodicScavenge): 10254 * wtf/Platform.h: 10255 102562010-01-20 Geoffrey Garen <ggaren@apple.com> 10257 10258 Reviewed by Oliver Hunt. 10259 10260 <rdar://problem/7562708> REGRESSION(53460): Heap::destroy may not run 10261 all destructors 10262 10263 * runtime/Collector.cpp: 10264 (JSC::Heap::freeBlocks): Instead of fully marking protected objects, 10265 just set their mark bits. This prevents protected objects from keeping 10266 unprotected objects alive. Destructor order is not guaranteed, so it's 10267 OK to destroy objects pointed to by protected objects before destroying 10268 protected objects. 10269 102702010-01-19 David Levin <levin@chromium.org> 10271 10272 Reviewed by Oliver Hunt. 10273 10274 CrossThreadCopier needs to support ThreadSafeShared better. 10275 https://bugs.webkit.org/show_bug.cgi?id=33698 10276 10277 * wtf/TypeTraits.cpp: Added tests for the new type traits. 10278 * wtf/TypeTraits.h: 10279 (WTF::IsSubclass): Determines if a class is a derived from another class. 10280 (WTF::IsSubclassOfTemplate): Determines if a class is a derived from a 10281 template class (with one parameter that is unknown). 10282 (WTF::RemoveTemplate): Reveals the type for a template parameter. 10283 102842010-01-20 Steve Falkenburg <sfalken@apple.com> 10285 10286 Reviewed by Darin Adler and Adam Roben. 10287 10288 Feature defines are difficult to maintain on Windows builds 10289 https://bugs.webkit.org/show_bug.cgi?id=33883 10290 10291 FeatureDefines.vsprops are now maintained in a way similar to 10292 Configurations/FeatureDefines.xcconfig, with the added advantage 10293 of having a single FeatureDefines file across all projects. 10294 10295 * Configurations/FeatureDefines.xcconfig: Add comments about keeping feature definitions in sync. 10296 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: Add FeatureDefines.vsprops inherited property sheet. 10297 * JavaScriptCore.vcproj/WTF/WTF.vcproj: Add FeatureDefines.vsprops inherited property sheet. 10298 102992010-01-20 Csaba Osztrogonác <ossy@webkit.org> 10300 10301 [Qt] Unreviewed buildfix for r53547. 10302 10303 * DerivedSources.pro: 10304 103052010-01-20 Tor Arne Vestbø <tor.arne.vestbo@nokia.com> 10306 10307 Reviewed by Simon Hausmann. 10308 10309 [Qt] Make extraCompilers for generated sources depend on their scripts 10310 10311 * DerivedSources.pro: 10312 103132010-01-19 Brian Weinstein <bweinstein@apple.com> 10314 10315 Reviewed by Tim Hatcher. 10316 10317 When JavaScriptCore calls Debugger::Exception, have it pass a 10318 hasHandler variable that represents if exception is being handled 10319 in the same function (not in a parent on the call stack). 10320 10321 This just adds a new parameter, no behavior is changed. 10322 10323 * debugger/Debugger.h: 10324 * interpreter/Interpreter.cpp: 10325 (JSC::Interpreter::throwException): 10326 103272010-01-18 Maciej Stachowiak <mjs@apple.com> 10328 10329 Reviewed by Adam Barth. 10330 10331 Inline functions that are hot in DOM manipulation 10332 https://bugs.webkit.org/show_bug.cgi?id=33820 10333 10334 (3% speedup on Dromaeo DOM Core tests) 10335 10336 * runtime/WeakGCMap.h: 10337 (JSC::::get): inline 10338 103392010-01-19 Laszlo Gombos <laszlo.1.gombos@nokia.com> 10340 10341 Unreviewed build fix for JIT with RVCT. 10342 10343 Remove IMPORT statement; cti_vm_throw is already defined in JITStubs.h. 10344 Remove extra ')'. 10345 10346 * jit/JITStubs.cpp: 10347 (JSC::ctiVMThrowTrampoline): 10348 103492010-01-19 Geoffrey Garen <ggaren@apple.com> 10350 10351 Reviewed by Oliver Hunt. 10352 10353 REGRESSION (52082): Crash on worker thread when reloading http://radnan.public.iastate.edu/procedural/ 10354 https://bugs.webkit.org/show_bug.cgi?id=33826 10355 10356 This bug was caused by a GC-protected object being destroyed early by 10357 Heap::destroy. Clients of the GC protect APIs (reasonably) expect pointers 10358 to GC-protected memory to be valid. 10359 10360 The solution is to do two passes of tear-down in Heap::destroy. The first 10361 pass tears down all unprotected objects. The second pass ASSERTs that all 10362 previously protected objects are now unprotected, and then tears down 10363 all perviously protected objects. These two passes simulate the two passes 10364 that would have been required to free a protected object during normal GC. 10365 10366 * API/JSContextRef.cpp: Removed some ASSERTs that have moved into Heap. 10367 10368 * runtime/Collector.cpp: 10369 (JSC::Heap::destroy): Moved ASSERTs to here. 10370 (JSC::Heap::freeBlock): Tidied up the use of didShrink by moving its 10371 setter to the function that does the shrinking. 10372 (JSC::Heap::freeBlocks): Implemented above algorithm. 10373 (JSC::Heap::shrinkBlocks): Tidied up the use of didShrink. 10374 103752010-01-19 Gavin Barraclough <barraclough@apple.com> 10376 10377 Reviewed by NOBODY (build fix). 10378 10379 Reverting r53455, breaks 2 javascriptcore tests. 10380 10381 * API/JSContextRef.cpp: 10382 * runtime/Collector.cpp: 10383 (JSC::Heap::destroy): 10384 (JSC::Heap::freeBlock): 10385 (JSC::Heap::freeBlocks): 10386 (JSC::Heap::shrinkBlocks): 10387 103882010-01-18 Gavin Barraclough <barraclough@apple.com> 10389 10390 Reviewed by NOBODY (build fix). 10391 10392 Revert r53454, since it causes much sadness in this world. 10393 10394 * runtime/UString.cpp: 10395 (JSC::UString::spliceSubstringsWithSeparators): 10396 (JSC::UString::replaceRange): 10397 * runtime/UStringImpl.cpp: 10398 (JSC::UStringImpl::baseSharedBuffer): 10399 (JSC::UStringImpl::sharedBuffer): 10400 (JSC::UStringImpl::~UStringImpl): 10401 * runtime/UStringImpl.h: 10402 (JSC::UntypedPtrAndBitfield::UntypedPtrAndBitfield): 10403 (JSC::UntypedPtrAndBitfield::asPtr): 10404 (JSC::UntypedPtrAndBitfield::operator&=): 10405 (JSC::UntypedPtrAndBitfield::operator|=): 10406 (JSC::UntypedPtrAndBitfield::operator&): 10407 (JSC::UStringImpl::create): 10408 (JSC::UStringImpl::cost): 10409 (JSC::UStringImpl::isIdentifier): 10410 (JSC::UStringImpl::setIsIdentifier): 10411 (JSC::UStringImpl::ref): 10412 (JSC::UStringImpl::deref): 10413 (JSC::UStringImpl::checkConsistency): 10414 (JSC::UStringImpl::UStringImpl): 10415 (JSC::UStringImpl::bufferOwnerString): 10416 (JSC::UStringImpl::bufferOwnership): 10417 (JSC::UStringImpl::isStatic): 10418 * wtf/StringHashFunctions.h: 10419 (WTF::stringHash): 10420 104212010-01-18 Geoffrey Garen <ggaren@apple.com> 10422 10423 Reviewed by Oliver Hunt. 10424 10425 REGRESSION (52082): Crash on worker thread when reloading http://radnan.public.iastate.edu/procedural/ 10426 https://bugs.webkit.org/show_bug.cgi?id=33826 10427 10428 This bug was caused by a GC-protected object being destroyed early by 10429 Heap::destroy. Clients of the GC protect APIs (reasonably) expect pointers 10430 to GC-protected memory to be valid. 10431 10432 The solution is to do two passes of tear-down in Heap::destroy. The first 10433 pass tears down all unprotected objects. The second pass ASSERTs that all 10434 previously protected objects are now unprotected, and then tears down 10435 all perviously protected objects. These two passes simulate the two passes 10436 that would have been required to free a protected object during normal GC. 10437 10438 * API/JSContextRef.cpp: Removed some ASSERTs that have moved into Heap. 10439 10440 * runtime/Collector.cpp: 10441 (JSC::Heap::destroy): Moved ASSERTs to here. 10442 (JSC::Heap::freeBlock): Tidied up the use of didShrink by moving its 10443 setter to the function that does the shrinking. 10444 (JSC::Heap::freeBlocks): Implemented above algorithm. 10445 (JSC::Heap::shrinkBlocks): Tidied up the use of didShrink. 10446 104472010-01-18 Gavin Barraclough <barraclough@apple.com> 10448 10449 Reviewed by Oliver Hunt. 10450 10451 https://bugs.webkit.org/show_bug.cgi?id=33731 10452 Remove UntypedPtrAndBitfield from UStringImpl (akin to PtrAndFlags). 10453 10454 This break the OS X Leaks tool. Instead, free up some more bits from the refCount. 10455 10456 * runtime/UStringImpl.cpp: 10457 (JSC::UStringImpl::sharedBuffer): 10458 (JSC::UStringImpl::~UStringImpl): 10459 * runtime/UStringImpl.h: 10460 (JSC::UStringImpl::cost): 10461 (JSC::UStringImpl::checkConsistency): 10462 (JSC::UStringImpl::UStringImpl): 10463 (JSC::UStringImpl::bufferOwnerString): 10464 (JSC::UStringImpl::): 10465 * wtf/StringHashFunctions.h: 10466 (WTF::stringHash): 10467 104682010-01-18 Kent Tamura <tkent@chromium.org> 10469 10470 Reviewed by Darin Adler. 10471 10472 HTMLInputElement::valueAsDate setter support for type=month. 10473 https://bugs.webkit.org/show_bug.cgi?id=33021 10474 10475 Expose the following functions to be used by WebCore: 10476 - WTF::msToyear() 10477 - WTF::dayInYear() 10478 - WTF::monthFromDayInYear() 10479 - WTF::dayInMonthFromDayInYear() 10480 10481 * JavaScriptCore.exp: 10482 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 10483 * wtf/DateMath.cpp: 10484 (WTF::msToYear): Remove "static inline". 10485 (WTF::dayInYear): Remove "static inline". 10486 (WTF::monthFromDayInYear): Remove "static inline". 10487 (WTF::dayInMonthFromDayInYear): Remove "static inline". 10488 * wtf/DateMath.h: Declare the above functions. 10489 104902010-01-18 Darin Adler <darin@apple.com> 10491 10492 Fix build by reverting the previous change. 10493 10494 * runtime/UString.h: Rolled out the FastAllocBase base class. 10495 It was making UString larger, and therefore JSString larger, 10496 and too big for a garbage collection cell. 10497 10498 This raises the unpleasant possibility that many classes became 10499 larger because we added the FastAllocBase base class. I am 10500 worried about this, and it needs to be investigated. 10501 105022010-01-18 Zoltan Horvath <zoltan@webkit.org> 10503 10504 Reviewed by Darin Adler. 10505 10506 Allow custom memory allocation control for UString class 10507 https://bugs.webkit.org/show_bug.cgi?id=27831 10508 10509 Inherits the following class from FastAllocBase because it is 10510 instantiated by 'new' and no need to be copyable: 10511 10512 class name - instantiated at: 10513 classs UString - JavaScriptCore/runtime/UString.cpp:160 10514 10515 * runtime/UString.h: 10516 105172010-01-18 Evan Cheng <evan.cheng@apple.com> 10518 10519 Reviewed by Darin Adler. 10520 10521 Add some ALWAYS_INLINE for key functions not inlined by some versions of GCC. 10522 rdar://problem/7553780 10523 10524 * runtime/JSObject.h: 10525 (JSC::JSObject::getPropertySlot): ALWAYS_INLINE both overloads. 10526 * runtime/JSString.h: 10527 (JSC::JSString::JSString): ALWAYS_INLINE the version that takes a UString. 10528 * runtime/UString.h: 10529 (JSC::operator==): ALWAYS_INLINE the version that compares two UString objects. 10530 105312010-01-18 Csaba Osztrogonác <ossy@webkit.org> 10532 10533 Reviewed by Darin Adler. 10534 10535 Delete dftables-xxxxxxxx.in files automatically. 10536 https://bugs.webkit.org/show_bug.cgi?id=33796 10537 10538 * pcre/dftables: unlink unnecessary temporary file. 10539 105402010-01-18 Tor Arne Vestbø <tor.arne.vestbo@nokia.com> 10541 10542 Reviewed by Simon Hausmann. 10543 10544 [Qt] Force qmake to generate a single makefile for DerivedSources.pro 10545 10546 * DerivedSources.pro: 10547 105482010-01-18 Csaba Osztrogonác <ossy@webkit.org> 10549 10550 Rubber-stamped by Gustavo Noronha Silva. 10551 10552 Rolling out r53391 and r53392 because of random crashes on buildbots. 10553 https://bugs.webkit.org/show_bug.cgi?id=33731 10554 10555 * bytecode/CodeBlock.h: 10556 (JSC::CallLinkInfo::seenOnce): 10557 (JSC::CallLinkInfo::setSeen): 10558 (JSC::MethodCallLinkInfo::MethodCallLinkInfo): 10559 (JSC::MethodCallLinkInfo::seenOnce): 10560 (JSC::MethodCallLinkInfo::setSeen): 10561 * jit/JIT.cpp: 10562 (JSC::JIT::unlinkCall): 10563 * jit/JITPropertyAccess.cpp: 10564 (JSC::JIT::patchMethodCallProto): 10565 * runtime/UString.cpp: 10566 (JSC::UString::spliceSubstringsWithSeparators): 10567 (JSC::UString::replaceRange): 10568 * runtime/UString.h: 10569 * runtime/UStringImpl.cpp: 10570 (JSC::UStringImpl::baseSharedBuffer): 10571 (JSC::UStringImpl::sharedBuffer): 10572 (JSC::UStringImpl::~UStringImpl): 10573 * runtime/UStringImpl.h: 10574 (JSC::UntypedPtrAndBitfield::UntypedPtrAndBitfield): 10575 (JSC::UntypedPtrAndBitfield::asPtr): 10576 (JSC::UntypedPtrAndBitfield::operator&=): 10577 (JSC::UntypedPtrAndBitfield::operator|=): 10578 (JSC::UntypedPtrAndBitfield::operator&): 10579 (JSC::UStringImpl::create): 10580 (JSC::UStringImpl::cost): 10581 (JSC::UStringImpl::isIdentifier): 10582 (JSC::UStringImpl::setIsIdentifier): 10583 (JSC::UStringImpl::ref): 10584 (JSC::UStringImpl::deref): 10585 (JSC::UStringImpl::checkConsistency): 10586 (JSC::UStringImpl::UStringImpl): 10587 (JSC::UStringImpl::bufferOwnerString): 10588 (JSC::UStringImpl::bufferOwnership): 10589 (JSC::UStringImpl::isStatic): 10590 * wtf/StringHashFunctions.h: 10591 (WTF::stringHash): 10592 105932010-01-18 Simon Hausmann <simon.hausmann@nokia.com> 10594 10595 Reviewed by Kenneth Rohde Christiansen. 10596 10597 Fix the build with strict gcc and RVCT versions: It's not legal to cast a 10598 pointer to a function to a void* without an intermediate cast to a non-pointer 10599 type. A cast to a ptrdiff_t inbetween fixes it. 10600 10601 * runtime/JSString.h: 10602 (JSC::Fiber::JSString): 10603 106042010-01-15 Gavin Barraclough <barraclough@apple.com> 10605 10606 Reviewed by Oliver Hunt. 10607 10608 https://bugs.webkit.org/show_bug.cgi?id=33731 10609 Remove UntypedPtrAndBitfield from UStringImpl (akin to PtrAndFlags). 10610 10611 This break the OS X Leaks tool. Instead, free up some more bits from the refCount. 10612 10613 * runtime/UStringImpl.cpp: 10614 (JSC::UStringImpl::sharedBuffer): 10615 (JSC::UStringImpl::~UStringImpl): 10616 * runtime/UStringImpl.h: 10617 (JSC::UStringImpl::cost): 10618 (JSC::UStringImpl::checkConsistency): 10619 (JSC::UStringImpl::UStringImpl): 10620 (JSC::UStringImpl::bufferOwnerString): 10621 (JSC::UStringImpl::): 10622 * wtf/StringHashFunctions.h: 10623 (WTF::stringHash): 10624 106252010-01-15 Gavin Barraclough <barraclough@apple.com> 10626 10627 Reviewed by Oliver Hunt. 10628 10629 https://bugs.webkit.org/show_bug.cgi?id=33731 10630 Remove uses of PtrAndFlags from JIT data stuctures. 10631 10632 These break the OS X Leaks tool. Free up a bit in CallLinkInfo, and invalid 10633 permutation of pointer states in MethodCallLinkInfo to represent the removed bits. 10634 10635 * bytecode/CodeBlock.h: 10636 (JSC::CallLinkInfo::seenOnce): 10637 (JSC::CallLinkInfo::setSeen): 10638 (JSC::MethodCallLinkInfo::MethodCallLinkInfo): 10639 (JSC::MethodCallLinkInfo::seenOnce): 10640 (JSC::MethodCallLinkInfo::setSeen): 10641 * jit/JIT.cpp: 10642 (JSC::JIT::unlinkCall): 10643 * jit/JITPropertyAccess.cpp: 10644 (JSC::JIT::patchMethodCallProto): 10645 * runtime/UString.h: 10646 106472010-01-16 Maciej Stachowiak <mjs@apple.com> 10648 10649 Reviewed by Oliver Hunt. 10650 10651 Cache JS string values made from DOM strings (Dromaeo speedup) 10652 https://bugs.webkit.org/show_bug.cgi?id=33768 10653 <rdar://problem/7353576> 10654 10655 * runtime/JSString.h: 10656 (JSC::jsStringWithFinalizer): Added new mechanism for a string to have an optional 10657 finalizer callback, for the benefit of weak-referencing caches. 10658 (JSC::): 10659 (JSC::Fiber::JSString): 10660 (JSC::Fiber::~JSString): 10661 * runtime/JSString.cpp: 10662 (JSC::JSString::resolveRope): Clear fibers so this doesn't look like a string with a finalizer. 10663 * runtime/WeakGCMap.h: Include "Collector.h" to make this header includable by itself. 10664 106652010-01-15 Sam Weinig <sam@webkit.org> 10666 10667 Reviewed by Maciej Stachowiak. 10668 10669 Fix for <rdar://problem/7548432> 10670 Add ALWAYS_INLINE to jsLess for a 1% speedup on llvm-gcc. 10671 10672 * runtime/Operations.h: 10673 (JSC::jsLess): 10674 106752010-01-14 Geoffrey Garen <ggaren@apple.com> 10676 10677 Reviewed by Oliver Hunt. 10678 10679 REGRESISON: Google maps buttons not working properly 10680 https://bugs.webkit.org/show_bug.cgi?id=31871 10681 10682 REGRESSION(r52948): JavaScript exceptions thrown on Google Maps when 10683 getting directions for a second time 10684 https://bugs.webkit.org/show_bug.cgi?id=33446 10685 10686 SunSpider and v8 report no change. 10687 10688 * interpreter/Interpreter.cpp: 10689 (JSC::Interpreter::tryCacheGetByID): Update our cached offset in case 10690 flattening the dictionary changed any of its offsets. 10691 10692 * jit/JITStubs.cpp: 10693 (JSC::JITThunks::tryCacheGetByID): 10694 (JSC::DEFINE_STUB_FUNCTION): 10695 * runtime/Operations.h: 10696 (JSC::normalizePrototypeChain): ditto 10697 106982010-01-14 Gavin Barraclough <barraclough@apple.com> 10699 10700 Reviewed by Oliver Hunt. 10701 10702 https://bugs.webkit.org/show_bug.cgi?id=33705 10703 UStringImpl::create() should use internal storage 10704 10705 When creating a UStringImpl copying of a UChar*, we can use an internal buffer, 10706 by calling UStringImpl::tryCreateUninitialized(). 10707 10708 Also, remove duplicate of copyChars from JSString, call UStringImpl's version. 10709 10710 Small (max 0.5%) progression on Sunspidey. 10711 10712 * runtime/JSString.cpp: 10713 (JSC::JSString::resolveRope): 10714 * runtime/UStringImpl.h: 10715 (JSC::UStringImpl::create): 10716 107172010-01-14 Gavin Barraclough <barraclough@apple.com> 10718 10719 Reviewed by Sam Weinig. 10720 10721 Make naming & behaviour of UString[Impl] methods more consistent. 10722 https://bugs.webkit.org/show_bug.cgi?id=33702 10723 10724 UString::create() creates a copy of the UChar* passed, but UStringImpl::create() assumes 10725 that it should assume ownership of the provided buffer (with UString::createNonCopying() 10726 and UStringImpl::createCopying() providing the alternate behaviours). Unify on create() 10727 taking a copy of the provided buffer. For non-copying cases, use the name 'adopt', and 10728 make this method take a Vector<UChar>&. For cases where non-copying construction was being 10729 used, other than from a Vector<UChar>, change the code to allocate the storage along with 10730 the UStringImpl using UStringImpl::createUninitialized(). (The adopt() method also more 10731 closely matches that of WebCore::StringImpl). 10732 10733 Also, UString::createUninitialized() and UStringImpl::createUninitialized() have incompatible 10734 behaviours, in that the UString form sets the provided UChar* to a null or non-null value to 10735 indicate success or failure, but UStringImpl uses the returned PassRefPtr<UStringImpl> to 10736 indicate when allocation has failed (potentially leaving the output Char* uninitialized). 10737 This is also incompatible with WebCore::StringImpl's behaviour, in that 10738 StringImpl::createUninitialized() will CRASH() if unable to allocate. Some uses of 10739 createUninitialized() in JSC are unsafe, since they do not test the result for null. 10740 UStringImpl's indication is preferable, since we may want a successful call to set the result 10741 buffer to 0 (specifically, StringImpl returns 0 for the buffer where createUninitialized() 10742 returns the empty string, which seems reasonable to catch bugs early). UString's method 10743 cannot support UStringImpl's behaviour directly, since it returns an object rather than a 10744 pointer. 10745 - remove UString::createUninitialized(), replace with calls to UStringImpl::createUninitialized() 10746 - create a UStringImpl::tryCreateUninitialized() form UStringImpl::createUninitialized(), 10747 with current behaviour, make createUninitialized() crash on failure to allocate. 10748 - make cases in JSC that do not check the result call createUninitialized(), and cases that do 10749 check call tryCreateUninitialized(). 10750 10751 Rename computedHash() to existingHash(), to bring this in line wih WebCore::StringImpl. 10752 10753 * API/JSClassRef.cpp: 10754 (OpaqueJSClassContextData::OpaqueJSClassContextData): 10755 * JavaScriptCore.exp: 10756 * runtime/ArrayPrototype.cpp: 10757 (JSC::arrayProtoFuncToString): 10758 * runtime/Identifier.cpp: 10759 (JSC::CStringTranslator::translate): 10760 (JSC::UCharBufferTranslator::translate): 10761 * runtime/JSString.cpp: 10762 (JSC::JSString::resolveRope): 10763 * runtime/Lookup.cpp: 10764 (JSC::HashTable::createTable): 10765 * runtime/Lookup.h: 10766 (JSC::HashTable::entry): 10767 * runtime/StringBuilder.h: 10768 (JSC::StringBuilder::release): 10769 * runtime/StringConstructor.cpp: 10770 (JSC::stringFromCharCodeSlowCase): 10771 * runtime/StringPrototype.cpp: 10772 (JSC::substituteBackreferencesSlow): 10773 (JSC::stringProtoFuncToLowerCase): 10774 (JSC::stringProtoFuncToUpperCase): 10775 (JSC::stringProtoFuncFontsize): 10776 (JSC::stringProtoFuncLink): 10777 * runtime/Structure.cpp: 10778 (JSC::Structure::despecifyDictionaryFunction): 10779 (JSC::Structure::get): 10780 (JSC::Structure::despecifyFunction): 10781 (JSC::Structure::put): 10782 (JSC::Structure::remove): 10783 (JSC::Structure::insertIntoPropertyMapHashTable): 10784 (JSC::Structure::checkConsistency): 10785 * runtime/Structure.h: 10786 (JSC::Structure::get): 10787 * runtime/StructureTransitionTable.h: 10788 (JSC::StructureTransitionTableHash::hash): 10789 * runtime/UString.cpp: 10790 (JSC::createRep): 10791 (JSC::UString::UString): 10792 (JSC::UString::spliceSubstringsWithSeparators): 10793 (JSC::UString::replaceRange): 10794 (JSC::UString::operator=): 10795 * runtime/UString.h: 10796 (JSC::UString::adopt): 10797 (JSC::IdentifierRepHash::hash): 10798 (JSC::makeString): 10799 * runtime/UStringImpl.h: 10800 (JSC::UStringImpl::adopt): 10801 (JSC::UStringImpl::create): 10802 (JSC::UStringImpl::createUninitialized): 10803 (JSC::UStringImpl::tryCreateUninitialized): 10804 (JSC::UStringImpl::existingHash): 10805 108062010-01-13 Kent Hansen <kent.hansen@nokia.com> 10807 10808 Reviewed by Oliver Hunt. 10809 10810 JSON.stringify and JSON.parse needlessly process properties in the prototype chain 10811 https://bugs.webkit.org/show_bug.cgi?id=33053 10812 10813 * runtime/JSONObject.cpp: 10814 (JSC::Stringifier::Holder::appendNextProperty): 10815 (JSC::Walker::walk): 10816 108172010-01-13 Gavin Barraclough <barraclough@apple.com> 10818 10819 Reviewed by NOBODY (buildfix). 10820 10821 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 10822 108232010-01-13 Alexey Proskuryakov <ap@apple.com> 10824 10825 Reviewed by Darin Adler. 10826 10827 https://bugs.webkit.org/show_bug.cgi?id=33641 10828 Assertion failure in Lexer.cpp if input stream ends while in string escape 10829 10830 Test: fast/js/end-in-string-escape.html 10831 10832 * parser/Lexer.cpp: (JSC::Lexer::lex): Bail out quickly on end of stream, not giving the 10833 assertion a chance to fire. 10834 108352010-01-13 Gavin Barraclough <barraclough@apple.com> 10836 10837 Reviewed by NOBODY (buildfix). 10838 10839 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 10840 108412010-01-13 Gavin Barraclough <barraclough@apple.com> 10842 10843 Rubber stamped by Sam Weinig & Darin Adler. 10844 10845 Three quick fixes to UStringImpl. 10846 - The destroy() method can be switched back to a normal destructor; since we've switched 10847 the way we protect static strings to be using an odd ref-count the destroy() won't abort. 10848 - The cost() calculation logic was wrong. If you have multiple JSStrings wrapping substrings 10849 of a base string, they would each report the full cost of the base string to the heap. 10850 Instead we should only be reporting once for the base string. 10851 - Remove the overloaded new operator calling fastMalloc, replace this with a 'using' to pick 10852 up the implementation from the parent class. 10853 10854 * JavaScriptCore.exp: 10855 * runtime/UStringImpl.cpp: 10856 (JSC::UStringImpl::~UStringImpl): 10857 * runtime/UStringImpl.h: 10858 (JSC::UStringImpl::cost): 10859 (JSC::UStringImpl::deref): 10860 108612010-01-13 Jocelyn Turcotte <jocelyn.turcotte@nokia.com> 10862 10863 Reviewed by Simon Hausmann. 10864 10865 [Qt] Split the build process in two different .pro files. 10866 This allows qmake to be run once all source files are available. 10867 10868 * DerivedSources.pro: Added. 10869 * JavaScriptCore.pri: Moved source generation to DerivedSources.pro 10870 * pcre/pcre.pri: Moved source generation to DerivedSources.pro 10871 108722010-01-12 Kent Hansen <kent.hansen@nokia.com> 10873 10874 Reviewed by Geoffrey Garen. 10875 10876 [ES5] Implement Object.getOwnPropertyNames 10877 https://bugs.webkit.org/show_bug.cgi?id=32242 10878 10879 Add an extra argument to getPropertyNames() and getOwnPropertyNames() 10880 (and all reimplementations thereof) that indicates whether non-enumerable 10881 properties should be added. 10882 10883 * API/JSCallbackObject.h: 10884 * API/JSCallbackObjectFunctions.h: 10885 (JSC::::getOwnPropertyNames): 10886 * JavaScriptCore.exp: 10887 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 10888 * debugger/DebuggerActivation.cpp: 10889 (JSC::DebuggerActivation::getOwnPropertyNames): 10890 * debugger/DebuggerActivation.h: 10891 * runtime/Arguments.cpp: 10892 (JSC::Arguments::getOwnPropertyNames): 10893 * runtime/Arguments.h: 10894 * runtime/CommonIdentifiers.h: 10895 * runtime/JSArray.cpp: 10896 (JSC::JSArray::getOwnPropertyNames): 10897 * runtime/JSArray.h: 10898 * runtime/JSByteArray.cpp: 10899 (JSC::JSByteArray::getOwnPropertyNames): 10900 * runtime/JSByteArray.h: 10901 * runtime/JSFunction.cpp: 10902 (JSC::JSFunction::getOwnPropertyNames): 10903 * runtime/JSFunction.h: 10904 * runtime/JSNotAnObject.cpp: 10905 (JSC::JSNotAnObject::getOwnPropertyNames): 10906 * runtime/JSNotAnObject.h: 10907 * runtime/JSObject.cpp: 10908 (JSC::getClassPropertyNames): 10909 (JSC::JSObject::getPropertyNames): 10910 (JSC::JSObject::getOwnPropertyNames): 10911 * runtime/JSObject.h: 10912 * runtime/JSVariableObject.cpp: 10913 (JSC::JSVariableObject::getOwnPropertyNames): 10914 * runtime/JSVariableObject.h: 10915 * runtime/ObjectConstructor.cpp: 10916 (JSC::ObjectConstructor::ObjectConstructor): 10917 (JSC::objectConstructorGetOwnPropertyNames): 10918 * runtime/RegExpMatchesArray.h: 10919 (JSC::RegExpMatchesArray::getOwnPropertyNames): 10920 * runtime/StringObject.cpp: 10921 (JSC::StringObject::getOwnPropertyNames): 10922 * runtime/StringObject.h: 10923 * runtime/Structure.cpp: Rename getEnumerablePropertyNames() to getPropertyNames(), which takes an extra argument. 10924 (JSC::Structure::getPropertyNames): 10925 * runtime/Structure.h: 10926 (JSC::): 10927 109282010-01-12 Alexey Proskuryakov <ap@apple.com> 10929 10930 Reviewed by Darin Adler. 10931 10932 https://bugs.webkit.org/show_bug.cgi?id=33540 10933 Make it possible to build in debug mode with assertions disabled 10934 10935 * jit/JITStubs.cpp: (JSC::DEFINE_STUB_FUNCTION): 10936 * runtime/Identifier.cpp: (JSC::Identifier::checkSameIdentifierTable): 10937 * wtf/FastMalloc.cpp: 10938 * wtf/HashTable.h: (WTF::HashTableConstIterator::checkValidity): 10939 * yarr/RegexCompiler.cpp: (JSC::Yarr::compileRegex): 10940 109412009-11-23 Yong Li <yoli@rim.com> 10942 10943 Reviewed by Adam Treat. 10944 10945 Make GIF decoder support down-sampling 10946 https://bugs.webkit.org/show_bug.cgi?id=31806 10947 10948 * platform/image-decoders/ImageDecoder.cpp: 10949 (WebCore::ImageDecoder::upperBoundScaledY): 10950 (WebCore::ImageDecoder::lowerBoundScaledY): 10951 * platform/image-decoders/ImageDecoder.h: 10952 (WebCore::RGBA32Buffer::scaledRect): 10953 (WebCore::RGBA32Buffer::setScaledRect): 10954 (WebCore::ImageDecoder::scaledSize): 10955 * platform/image-decoders/gif/GIFImageDecoder.cpp: 10956 (WebCore::GIFImageDecoder::sizeNowAvailable): 10957 (WebCore::GIFImageDecoder::initFrameBuffer): 10958 (WebCore::copyOnePixel): 10959 (WebCore::GIFImageDecoder::haveDecodedRow): 10960 (WebCore::GIFImageDecoder::frameComplete): 10961 109622010-01-12 Adam Barth <abarth@webkit.org> 10963 10964 Reviewed by Eric Seidel. 10965 10966 ecma/Date/15.9.5.12-1.js fails every night at midnight 10967 https://bugs.webkit.org/show_bug.cgi?id=28041 10968 10969 Change the test to use a concrete time instead of "now". 10970 10971 * tests/mozilla/ecma/Date/15.9.5.10-1.js: 10972 * tests/mozilla/ecma/Date/15.9.5.12-1.js: 10973 109742010-01-11 Csaba Osztrogonác <ossy@webkit.org> 10975 10976 Reviewed by Ariya Hidayat. 10977 10978 [Qt] Enable JIT and YARR_JIT if (CPU(X86_64) && OS(LINUX) && GCC_VERSION >= 40100) 10979 10980 * wtf/Platform.h: 10981 109822010-01-11 Geoffrey Garen <ggaren@apple.com> 10983 10984 Reviewed by Alexey Proskuryakov. 10985 10986 https://bugs.webkit.org/show_bug.cgi?id=33481 10987 Uninitialized data members in ArrayStorage 10988 10989 SunSpider reports no change. 10990 10991 * runtime/JSArray.cpp: 10992 (JSC::JSArray::JSArray): Initialize missing data members in the two cases 10993 where we don't use fastZeroedMalloc, so it doesn't happen automatically. 10994 109952010-01-11 Steve Falkenburg <sfalken@apple.com> 10996 10997 Reviewed by Sam Weinig. 10998 10999 https://bugs.webkit.org/show_bug.cgi?id=33480 11000 11001 Improve debugging reliability for WTF on Windows. 11002 Store WTF static library's PDB file into a better location. 11003 11004 * JavaScriptCore.vcproj/WTF/WTF.vcproj: 11005 110062010-01-11 Steve Falkenburg <sfalken@apple.com> 11007 11008 Windows build fix. 11009 Remove extraneous entries from def file causing build warning. 11010 11011 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 11012 110132010-01-10 Kent Hansen <kent.hansen@nokia.com> 11014 11015 Reviewed by Darin Adler. 11016 11017 RegExp.prototype.toString returns "//" for empty regular expressions 11018 https://bugs.webkit.org/show_bug.cgi?id=33319 11019 11020 "//" starts a single-line comment, hence "/(?:)/" should be used, according to ECMA. 11021 11022 * runtime/RegExpPrototype.cpp: 11023 (JSC::regExpProtoFuncToString): 11024 11025 * tests/mozilla/ecma_2/RegExp/properties-001.js: 11026 (AddRegExpCases): 11027 * tests/mozilla/js1_2/regexp/toString.js: 11028 Update relevant Mozilla tests (Mozilla has had this behavior since November 2003). 11029 110302010-01-10 Darin Adler <darin@apple.com> 11031 11032 * tests/mozilla/ecma/Array/15.4.1.1.js: Added property allow-tabs. 11033 * tests/mozilla/ecma/Array/15.4.1.2.js: Added property allow-tabs. 11034 * tests/mozilla/ecma/Array/15.4.2.1-1.js: Added property allow-tabs. 11035 * tests/mozilla/ecma/Array/15.4.2.2-1.js: Added property allow-tabs. 11036 * tests/mozilla/ecma/Array/15.4.2.2-2.js: Added property allow-tabs. 11037 * tests/mozilla/ecma/Array/15.4.2.3.js: Added property allow-tabs. 11038 * tests/mozilla/ecma/Array/15.4.3.2.js: Added property allow-tabs. 11039 * tests/mozilla/ecma/Array/15.4.3.js: Added property allow-tabs. 11040 * tests/mozilla/ecma/Array/15.4.4.1.js: Added property allow-tabs. 11041 * tests/mozilla/ecma/Array/15.4.4.js: Added property allow-tabs. 11042 * tests/mozilla/ecma/LexicalConventions/7.7.4.js: Added property allow-tabs. 11043 * tests/mozilla/ecma/Math/15.8.2.13.js: Added property allow-tabs. 11044 * tests/mozilla/ecma/Math/15.8.2.16.js: Added property allow-tabs. 11045 * tests/mozilla/ecma/Math/15.8.2.18.js: Added property allow-tabs. 11046 * tests/mozilla/ecma/Math/15.8.2.2.js: Added property allow-tabs. 11047 * tests/mozilla/ecma/Math/15.8.2.4.js: Added property allow-tabs. 11048 * tests/mozilla/ecma/Math/15.8.2.5.js: Added property allow-tabs. 11049 * tests/mozilla/ecma/Math/15.8.2.7.js: Added property allow-tabs. 11050 * tests/mozilla/ecma/String/15.5.1.js: Added property allow-tabs. 11051 * tests/mozilla/ecma/String/15.5.2.js: Added property allow-tabs. 11052 * tests/mozilla/ecma/String/15.5.3.1-3.js: Added property allow-tabs. 11053 * tests/mozilla/ecma/String/15.5.3.1-4.js: Added property allow-tabs. 11054 * tests/mozilla/ecma/String/15.5.3.js: Added property allow-tabs. 11055 * tests/mozilla/ecma/TypeConversion/9.5-2.js: Added property allow-tabs. 11056 * tests/mozilla/ecma/jsref.js: Modified property allow-tabs. 11057 * tests/mozilla/ecma/shell.js: Modified property allow-tabs. 11058 * tests/mozilla/ecma_2/LexicalConventions/keywords-001.js: Added property allow-tabs. 11059 * tests/mozilla/ecma_2/RegExp/exec-001.js: Added property allow-tabs. 11060 * tests/mozilla/ecma_2/String/match-004.js: Added property allow-tabs. 11061 * tests/mozilla/ecma_2/String/replace-001.js: Added property allow-tabs. 11062 * tests/mozilla/ecma_2/String/split-002.js: Added property allow-tabs. 11063 * tests/mozilla/ecma_2/jsref.js: Modified property allow-tabs. 11064 * tests/mozilla/ecma_2/shell.js: Added property allow-tabs. 11065 * tests/mozilla/ecma_3/Date/shell.js: Modified property allow-tabs. 11066 * tests/mozilla/ecma_3/Exceptions/regress-181654.js: Added property allow-tabs. 11067 * tests/mozilla/ecma_3/RegExp/regress-209067.js: Added property allow-tabs. 11068 * tests/mozilla/ecma_3/RegExp/regress-85721.js: Added property allow-tabs. 11069 * tests/mozilla/importList.html: Added property allow-tabs. 11070 * tests/mozilla/js1_1/shell.js: Added property allow-tabs. 11071 * tests/mozilla/js1_2/Array/general1.js: Added property allow-tabs. 11072 * tests/mozilla/js1_2/Array/general2.js: Added property allow-tabs. 11073 * tests/mozilla/js1_2/Array/slice.js: Added property allow-tabs. 11074 * tests/mozilla/js1_2/Array/splice1.js: Added property allow-tabs. 11075 * tests/mozilla/js1_2/Array/splice2.js: Added property allow-tabs. 11076 * tests/mozilla/js1_2/Objects/toString-001.js: Added property allow-tabs. 11077 * tests/mozilla/js1_2/String/charCodeAt.js: Added property allow-tabs. 11078 * tests/mozilla/js1_2/String/concat.js: Modified property allow-tabs. 11079 * tests/mozilla/js1_2/String/match.js: Added property allow-tabs. 11080 * tests/mozilla/js1_2/String/slice.js: Added property allow-tabs. 11081 * tests/mozilla/js1_2/function/Function_object.js: Added property allow-tabs. 11082 * tests/mozilla/js1_2/function/Number.js: Modified property allow-tabs. 11083 * tests/mozilla/js1_2/function/String.js: Modified property allow-tabs. 11084 * tests/mozilla/js1_2/function/nesting.js: Added property allow-tabs. 11085 * tests/mozilla/js1_2/function/regexparg-1.js: Added property allow-tabs. 11086 * tests/mozilla/js1_2/function/regexparg-2-n.js: Added property allow-tabs. 11087 * tests/mozilla/js1_2/jsref.js: Added property allow-tabs. 11088 * tests/mozilla/js1_2/operator/equality.js: Added property allow-tabs. 11089 * tests/mozilla/js1_2/operator/strictEquality.js: Added property allow-tabs. 11090 * tests/mozilla/js1_2/regexp/RegExp_dollar_number.js: Added property allow-tabs. 11091 * tests/mozilla/js1_2/regexp/RegExp_input.js: Added property allow-tabs. 11092 * tests/mozilla/js1_2/regexp/RegExp_input_as_array.js: Added property allow-tabs. 11093 * tests/mozilla/js1_2/regexp/RegExp_lastIndex.js: Added property allow-tabs. 11094 * tests/mozilla/js1_2/regexp/RegExp_lastMatch.js: Added property allow-tabs. 11095 * tests/mozilla/js1_2/regexp/RegExp_lastMatch_as_array.js: Added property allow-tabs. 11096 * tests/mozilla/js1_2/regexp/RegExp_lastParen.js: Added property allow-tabs. 11097 * tests/mozilla/js1_2/regexp/RegExp_lastParen_as_array.js: Added property allow-tabs. 11098 * tests/mozilla/js1_2/regexp/RegExp_leftContext.js: Added property allow-tabs. 11099 * tests/mozilla/js1_2/regexp/RegExp_leftContext_as_array.js: Added property allow-tabs. 11100 * tests/mozilla/js1_2/regexp/RegExp_multiline.js: Added property allow-tabs. 11101 * tests/mozilla/js1_2/regexp/RegExp_multiline_as_array.js: Added property allow-tabs. 11102 * tests/mozilla/js1_2/regexp/RegExp_object.js: Added property allow-tabs. 11103 * tests/mozilla/js1_2/regexp/RegExp_rightContext.js: Added property allow-tabs. 11104 * tests/mozilla/js1_2/regexp/RegExp_rightContext_as_array.js: Added property allow-tabs. 11105 * tests/mozilla/js1_2/regexp/alphanumeric.js: Added property allow-tabs. 11106 * tests/mozilla/js1_2/regexp/asterisk.js: Added property allow-tabs. 11107 * tests/mozilla/js1_2/regexp/backslash.js: Added property allow-tabs. 11108 * tests/mozilla/js1_2/regexp/backspace.js: Added property allow-tabs. 11109 * tests/mozilla/js1_2/regexp/beginLine.js: Added property allow-tabs. 11110 * tests/mozilla/js1_2/regexp/character_class.js: Added property allow-tabs. 11111 * tests/mozilla/js1_2/regexp/compile.js: Added property allow-tabs. 11112 * tests/mozilla/js1_2/regexp/control_characters.js: Added property allow-tabs. 11113 * tests/mozilla/js1_2/regexp/digit.js: Added property allow-tabs. 11114 * tests/mozilla/js1_2/regexp/dot.js: Added property allow-tabs. 11115 * tests/mozilla/js1_2/regexp/endLine.js: Added property allow-tabs. 11116 * tests/mozilla/js1_2/regexp/everything.js: Added property allow-tabs. 11117 * tests/mozilla/js1_2/regexp/exec.js: Added property allow-tabs. 11118 * tests/mozilla/js1_2/regexp/flags.js: Added property allow-tabs. 11119 * tests/mozilla/js1_2/regexp/global.js: Added property allow-tabs. 11120 * tests/mozilla/js1_2/regexp/hexadecimal.js: Added property allow-tabs. 11121 * tests/mozilla/js1_2/regexp/ignoreCase.js: Added property allow-tabs. 11122 * tests/mozilla/js1_2/regexp/interval.js: Added property allow-tabs. 11123 * tests/mozilla/js1_2/regexp/octal.js: Added property allow-tabs. 11124 * tests/mozilla/js1_2/regexp/parentheses.js: Added property allow-tabs. 11125 * tests/mozilla/js1_2/regexp/plus.js: Added property allow-tabs. 11126 * tests/mozilla/js1_2/regexp/question_mark.js: Added property allow-tabs. 11127 * tests/mozilla/js1_2/regexp/simple_form.js: Added property allow-tabs. 11128 * tests/mozilla/js1_2/regexp/source.js: Added property allow-tabs. 11129 * tests/mozilla/js1_2/regexp/special_characters.js: Added property allow-tabs. 11130 * tests/mozilla/js1_2/regexp/string_replace.js: Added property allow-tabs. 11131 * tests/mozilla/js1_2/regexp/string_search.js: Added property allow-tabs. 11132 * tests/mozilla/js1_2/regexp/string_split.js: Added property allow-tabs. 11133 * tests/mozilla/js1_2/regexp/test.js: Added property allow-tabs. 11134 * tests/mozilla/js1_2/regexp/toString.js: Added property allow-tabs. 11135 * tests/mozilla/js1_2/regexp/vertical_bar.js: Added property allow-tabs. 11136 * tests/mozilla/js1_2/regexp/whitespace.js: Added property allow-tabs. 11137 * tests/mozilla/js1_2/regexp/word_boundary.js: Added property allow-tabs. 11138 * tests/mozilla/js1_2/shell.js: Added property allow-tabs. 11139 * tests/mozilla/js1_2/statements/break.js: Added property allow-tabs. 11140 * tests/mozilla/js1_2/statements/continue.js: Added property allow-tabs. 11141 * tests/mozilla/js1_2/statements/do_while.js: Added property allow-tabs. 11142 * tests/mozilla/js1_2/statements/switch.js: Added property allow-tabs. 11143 * tests/mozilla/js1_2/statements/switch2.js: Added property allow-tabs. 11144 * tests/mozilla/js1_3/shell.js: Added property allow-tabs. 11145 * tests/mozilla/js1_4/shell.js: Added property allow-tabs. 11146 * tests/mozilla/js1_5/Regress/regress-111557.js: Added property allow-tabs. 11147 * tests/mozilla/js1_5/Regress/regress-216320.js: Added property allow-tabs. 11148 * tests/mozilla/menuhead.html: Added property allow-tabs. 11149 * tests/mozilla/mklistpage.pl: Added property allow-tabs. 11150 * tests/mozilla/runtests.pl: Added property allow-tabs. 11151 111522010-01-08 Daniel Bates <dbates@webkit.org> 11153 11154 Reviewed by Adam Barth. 11155 11156 https://bugs.webkit.org/show_bug.cgi?id=33417 11157 11158 Cleans up style errors exposed by the patch for bug #33198. 11159 Moreover, fixes all "Weird number of spaces at line-start. Are you using a 4-space indent?" 11160 errors reported by check-webkit-style. 11161 11162 No functionality was changed. So, no new tests. 11163 11164 * wtf/Platform.h: 11165 111662010-01-08 Kent Hansen <kent.hansen@nokia.com> 11167 11168 Reviewed by Eric Seidel. 11169 11170 Don't store RegExp flags string representation 11171 https://bugs.webkit.org/show_bug.cgi?id=33321 11172 11173 It's unused; the string representation is reconstructed from flags. 11174 11175 * runtime/RegExp.cpp: 11176 (JSC::RegExp::RegExp): 11177 * runtime/RegExp.h: 11178 111792010-01-08 Geoffrey Garen <ggaren@apple.com> 11180 11181 Reviewed by Oliver Hunt. 11182 11183 Memory use grows grows possibly unbounded in this JavaScript Array test case 11184 https://bugs.webkit.org/show_bug.cgi?id=31675 11185 11186 This fixes one observed bug in this test case, which is that 11187 arrays don't report extra cost for the sparse value maps. 11188 11189 SunSpider reports a small speedup. 11190 11191 * runtime/JSArray.cpp: 11192 (JSC::JSArray::putSlowCase): Report extra memory cost for 11193 the sparse value map. 11194 * runtime/JSArray.h: 11195 111962010-01-08 Yong Li <yoli@rim.com> 11197 11198 Reviewed by Darin Adler. 11199 11200 Remove unnecessary #include from FastMalloc.cpp 11201 https://bugs.webkit.org/show_bug.cgi?id=33393 11202 11203 * wtf/FastMalloc.cpp: 11204 112052010-01-08 Eric Seidel <eric@webkit.org> 11206 11207 No review, rolling out r52983. 11208 http://trac.webkit.org/changeset/52983 11209 https://bugs.webkit.org/show_bug.cgi?id=33321 11210 11211 Broke 59 JavaScriptCore tests. I don't think Kent knew about 11212 run-javascriptcore-tests. Sadly neither does the commit-bot, 11213 yet. 11214 11215 * runtime/RegExp.cpp: 11216 (JSC::RegExp::RegExp): 11217 * runtime/RegExp.h: 11218 (JSC::RegExp::flags): 11219 112202010-01-08 Eric Seidel <eric@webkit.org> 11221 11222 No review, rolling out r52981. 11223 http://trac.webkit.org/changeset/52981 11224 https://bugs.webkit.org/show_bug.cgi?id=33319 11225 11226 Caused two JS tests to start failing: 11227 ecma_2/RegExp/properties-001.js and js1_2/regexp/toString.js 11228 11229 * runtime/RegExpPrototype.cpp: 11230 (JSC::regExpProtoFuncToString): 11231 112322010-01-08 Kent Hansen <kent.hansen@nokia.com> 11233 11234 Reviewed by Darin Adler. 11235 11236 Don't store RegExp flags string representation 11237 https://bugs.webkit.org/show_bug.cgi?id=33321 11238 11239 It's unused; the string representation is reconstructed from flags. 11240 11241 * runtime/RegExp.cpp: 11242 (JSC::RegExp::RegExp): 11243 * runtime/RegExp.h: 11244 112452010-01-08 Kent Hansen <kent.hansen@nokia.com> 11246 11247 Reviewed by Darin Adler. 11248 11249 RegExp.prototype.toString returns "//" for empty regular expressions 11250 https://bugs.webkit.org/show_bug.cgi?id=33319 11251 11252 "//" starts a single-line comment, hence "/(?:)/" should be used, according to ECMA. 11253 11254 * runtime/RegExpPrototype.cpp: 11255 (JSC::regExpProtoFuncToString): 11256 112572010-01-08 Norbert Leser <norbert.leser@nokia.com> 11258 11259 Reviewed by Darin Adler. 11260 11261 RVCT compiler with "-Otime -O3" optimization tries to optimize out 11262 inline new'ed pointers that are passed as arguments. 11263 Proposed patch assigns new'ed pointer explicitly outside function call. 11264 11265 https://bugs.webkit.org/show_bug.cgi?id=33084 11266 11267 * API/JSClassRef.cpp: 11268 (OpaqueJSClass::OpaqueJSClass): 11269 (OpaqueJSClassContextData::OpaqueJSClassContextData): 11270 112712010-01-08 Gabor Loki <loki@webkit.org> 11272 11273 Reviewed by Gavin Barraclough. 11274 11275 Remove an unnecessary cacheFlush from ARM_TRADITIONAL JIT 11276 https://bugs.webkit.org/show_bug.cgi?id=33203 11277 11278 * assembler/ARMAssembler.cpp: Remove obsolete linkBranch function. 11279 (JSC::ARMAssembler::executableCopy): Inline a clean linkBranch code. 11280 * assembler/ARMAssembler.h: 11281 (JSC::ARMAssembler::getLdrImmAddress): Use inline function. 11282 (JSC::ARMAssembler::getLdrImmAddressOnPool): Ditto. 11283 (JSC::ARMAssembler::patchPointerInternal): Remove an unnecessary cacheFlush. 11284 (JSC::ARMAssembler::linkJump): Use patchPointerInternal instead of linkBranch. 11285 (JSC::ARMAssembler::linkCall): Ditto. 11286 (JSC::ARMAssembler::relinkCall): Ditto. 11287 112882010-01-07 Gabor Loki <loki@webkit.org> 11289 11290 Reviewed by Gavin Barraclough. 11291 11292 Build fix for JSVALUE32 when ENABLE_JIT_OPTIMIZE* are disabled 11293 https://bugs.webkit.org/show_bug.cgi?id=33311 11294 11295 Move compileGetDirectOffset function to common part of JSVALUE32 11296 11297 * jit/JITPropertyAccess.cpp: 11298 (JSC::JIT::compileGetDirectOffset): 11299 113002010-01-07 Laszlo Gombos <laszlo.1.gombos@nokia.com> 11301 11302 Reviewed by Maciej Stachowiak. 11303 11304 Allow call sites to determine if ASSERT_* and LOG_* macros are operational 11305 https://bugs.webkit.org/show_bug.cgi?id=33020 11306 11307 * wtf/Assertions.h: Set ASSERT_MSG_DISABLED, FATAL_DISABLED, 11308 ERROR_DISABLED, LOG_DISABLED to 1 if the compiler does not support 11309 variadic macros. Refactor for better readibility. 11310 113112010-01-07 Daniel Bates <dbates@rim.com> 11312 11313 Reviewed by Eric Seidel. 11314 11315 https://bugs.webkit.org/show_bug.cgi?id=32987 11316 11317 Added ENABLE_XHTMLMP flag. Disabled by default. 11318 11319 * Configurations/FeatureDefines.xcconfig: 11320 113212010-01-07 Laszlo Gombos <laszlo.1.gombos@nokia.com> 11322 11323 Reviewed by Gavin Barraclough. 11324 11325 [Symbian] Port ARM traditional JIT Trampolines to RVCT 11326 https://bugs.webkit.org/show_bug.cgi?id=30552 11327 11328 Take the GCC implementation and mechanically convert 11329 it to RVCT syntax. 11330 11331 Use 'bx rX' instead of 'mov pc, rX' when it is available. 11332 11333 Developed in cooperation with Iain Campbell and Gabor Loki. 11334 11335 * JavaScriptCore.pri: Extra step to generate RVCT stubs. The 11336 script generation intentionally executed all the time not just 11337 for RVCT targets. 11338 11339 * create_rvct_stubs: Added. Perl script to expand precompiler macros 11340 for RVCT assembler - the template is defined in JITStubs.cpp. 11341 11342 * jit/JITStubs.cpp: 11343 (JSC::ctiTrampoline): 11344 (JSC::ctiVMThrowTrampoline): 11345 (JSC::ctiOpThrowNotCaught): 11346 113472010-01-07 Geoffrey Garen <ggaren@apple.com> 11348 11349 Reviewed by Sam Weinig. 11350 11351 Fix a crash seen on the buildbots. 11352 11353 * runtime/JSGlobalObject.cpp: 11354 (JSC::JSGlobalObject::init): Disable specific function tracking here, 11355 instead of in WebCore, to ensure that the disabling happens before a 11356 specific function can be registered. 11357 113582010-01-07 Alexey Proskuryakov <ap@apple.com> 11359 11360 Mac build fix. 11361 11362 * JavaScriptCore.exp: Export new JSGlobalData static data members. 11363 113642010-01-07 Alexey Proskuryakov <ap@apple.com> 11365 11366 Reviewed by Geoffrey Garen. 11367 11368 https://bugs.webkit.org/show_bug.cgi?id=33057 11369 REGRESSION(r49365): typeof(xhr.responseText) != "string" in Windows 11370 11371 <rdar://problem/7296920> REGRESSION: WebKit fails to start PeaceKeeper benchmark 11372 11373 Test: fast/js/webcore-string-comparison.html 11374 11375 In r49365, some code was moved from JSString.cpp to JSString.h, and as a result, WebCore 11376 got a way to directly instantiate JSStrings over DLL borders. Since vftable for JSString was 11377 not exported, objects created from WebCore got a different vptr, and JavaScriptCore 11378 optimizations that relied on vptr of all JSString objects being equal failed. 11379 11380 * config.h: Added a JS_EXPORTCLASS macro for exporting classes. It's currently the same as 11381 JS_EXPORTDATA, but it clearly needed a new name. 11382 11383 * runtime/InitializeThreading.cpp: 11384 (JSC::initializeThreadingOnce): 11385 * runtime/JSGlobalData.cpp: 11386 (JSC::JSGlobalData::storeVPtrs): 11387 (JSC::JSGlobalData::JSGlobalData): 11388 (JSC::JSGlobalData::createNonDefault): 11389 (JSC::JSGlobalData::create): 11390 (JSC::JSGlobalData::sharedInstance): 11391 * runtime/JSGlobalData.h: 11392 Store vptrs just once, no need to repeatedly pick and copy them. This makes it possible to 11393 assert vptr correctness in object destructors (which don't have access to JSGlobalData, 11394 and even Heap::heap(this) will fail for fake objects created from storeVPtrs()). 11395 11396 * runtime/JSArray.cpp: (JSC::JSArray::~JSArray): Assert that vptr is what we expect it to be. 11397 It's important to assert in destructor, because MSVC changes the vptr after constructor 11398 is invoked. 11399 * runtime/JSByteArray.cpp: (JSC::JSByteArray::~JSByteArray): Ditto. 11400 * runtime/JSByteArray.h: Ditto. 11401 * runtime/JSFunction.h: Ditto. 11402 * runtime/JSFunction.cpp: (JSC::JSFunction::~JSFunction): Ditto. 11403 11404 * runtime/JSCell.h: (JSC::JSCell::setVPtr): Added a method to substitute vptr for another 11405 one. 11406 11407 * runtime/JSString.h: Export JSString class together with its vftable, and tell other 11408 libraries tp import it. This is needed on platforms that have a separate JavaScriptCore 11409 dynamic library - and on Mac, we already did the export via JavaScriptCore.exp. 11410 (JSC::JSString::~JSString): Assert tha vptr is what we expect it to be. 11411 (JSC::fixupVPtr): Store a previously saved primary vftable pointer (do nothing if building 11412 JavaScriptCore itself). 11413 (JSC::jsSingleCharacterString): Call fixupVPtr in case this is call across DLL boundary. 11414 (JSC::jsSingleCharacterSubstring): Ditto. 11415 (JSC::jsNontrivialString): Ditto. 11416 (JSC::jsString): Ditto. 11417 (JSC::jsSubstring): Ditto. 11418 (JSC::jsOwnedString): Ditto. 11419 11420 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: Export the new static 11421 JSGlobalData members that are used in WebCore via inline functions. 11422 114232010-01-07 Geoffrey Garen <ggaren@apple.com> 11424 11425 Reviewed by Sam Weinig. 11426 11427 Safari memory usage skyrockets using new Google AdWords interface 11428 https://bugs.webkit.org/show_bug.cgi?id=33343 11429 11430 The memory use was caused by the global object creating too many structures 11431 as it thrashed between different specific functions. 11432 11433 * runtime/Structure.cpp: 11434 (JSC::Structure::Structure): 11435 (JSC::Structure::addPropertyTransition): 11436 (JSC::Structure::changePrototypeTransition): 11437 (JSC::Structure::despecifyFunctionTransition): 11438 (JSC::Structure::addAnonymousSlotsTransition): 11439 (JSC::Structure::getterSetterTransition): 11440 (JSC::Structure::toDictionaryTransition): 11441 (JSC::Structure::addPropertyWithoutTransition): 11442 (JSC::Structure::despecifyAllFunctions): 11443 * runtime/Structure.h: 11444 (JSC::Structure::disableSpecificFunctionTracking): Track a thrash count 11445 for specific functions. Disable specific function tracking once the 11446 thrash count has been hit. 11447 114482010-01-07 Csaba Osztrogonác <ossy@webkit.org> 11449 11450 Reviewed by Simon Hausmann. 11451 11452 [Qt] Enable JIT in debug mode on win32 after r51141 fixed the crashes. 11453 11454 * JavaScriptCore.pri: 11455 114562010-01-07 Zoltan Horvath <zoltan@webkit.org> 11457 11458 Reviewed by Holger Freyther. 11459 11460 [Mac] Build fix when FAST_MALLOC_MATCH_VALIDATION=1 11461 https://bugs.webkit.org/show_bug.cgi?id=33312 11462 11463 Using of operator += cause compile error on Mac, so it is changed to 11464 "= static_cast<AllocAlignmentInteger*>(old_ptr) + 1". 11465 11466 * wtf/FastMalloc.cpp: 11467 (WTF::TCMallocStats::realloc): 11468 114692010-01-07 Zoltan Horvath <zoltan@webkit.org> 11470 11471 Reviewed by Holger Freyther. 11472 11473 [Qt] Build fix when FAST_MALLOC_MATCH_VALIDATION=1 11474 https://bugs.webkit.org/show_bug.cgi?id=33312 11475 11476 Remove pByte (committed in r42344 from #20422), because pByte doesn't 11477 exist and it is unnecessary. 11478 11479 * wtf/FastMalloc.cpp: 11480 (WTF::TCMallocStats::realloc): 11481 114822010-01-06 Gavin Barraclough <barraclough@apple.com> 11483 11484 QT build fix. 11485 11486 * runtime/Identifier.cpp: 11487 (JSC::createIdentifierTableSpecific): 11488 114892010-01-06 Gavin Barraclough <barraclough@apple.com> 11490 11491 Windows build fix part I. 11492 11493 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 11494 114952010-01-06 Dan Bernstein <mitz@apple.com> 11496 11497 Build fix 11498 11499 * runtime/Identifier.cpp: 11500 (JSC::createIdentifierTableSpecificCallback): 11501 115022010-01-05 Gavin Barraclough <barraclough@apple.com> 11503 11504 Reviewed by Sam Weinig. 11505 11506 https://bugs.webkit.org/show_bug.cgi?id=33236 11507 Remove m_identifierTable pointer from UString 11508 11509 Currently every string holds a pointer so that during destruction, 11510 if a string has been used as an identifier, it can remove itself 11511 from the table. By instead accessing the identifierTable via a 11512 thread specific tracking the table associated with the current 11513 globaldata, we can save the memory cost of this pointer. 11514 11515 * API/APIShims.h: 11516 (JSC::APIEntryShimWithoutLock::APIEntryShimWithoutLock): 11517 (JSC::APIEntryShimWithoutLock::~APIEntryShimWithoutLock): 11518 (JSC::APICallbackShim::APICallbackShim): 11519 (JSC::APICallbackShim::~APICallbackShim): 11520 11521 - change the API shims to track the identifierTable of the current JSGlobalData. 11522 11523 * API/JSContextRef.cpp: 11524 (JSContextGroupCreate): 11525 11526 - update creation of JSGlobalData for API usage to use new create method. 11527 - fix shim instanciation bug in JSGlobalContextCreateInGroup. 11528 11529 * JavaScriptCore.exp: 11530 * runtime/Completion.cpp: 11531 (JSC::checkSyntax): 11532 (JSC::evaluate): 11533 11534 - add asserts to check the identifierTable is being tracked correctly. 11535 11536 * runtime/Identifier.cpp: 11537 (JSC::IdentifierTable::~IdentifierTable): 11538 (JSC::IdentifierTable::add): 11539 (JSC::Identifier::remove): 11540 (JSC::Identifier::checkSameIdentifierTable): 11541 (JSC::createIdentifierTableSpecificCallback): 11542 (JSC::createIdentifierTableSpecific): 11543 (JSC::createDefaultDataSpecific): 11544 11545 - Use currentIdentifierTable() instead of UStringImpl::m_identifierTable. 11546 - Define methods to access the thread specific identifier tables. 11547 11548 * runtime/Identifier.h: 11549 (JSC::ThreadIdentifierTableData::ThreadIdentifierTableData): 11550 (JSC::defaultIdentifierTable): 11551 (JSC::setDefaultIdentifierTable): 11552 (JSC::currentIdentifierTable): 11553 (JSC::setCurrentIdentifierTable): 11554 (JSC::resetCurrentIdentifierTable): 11555 11556 - Declare methods to access the thread specific identifier tables. 11557 11558 * runtime/JSGlobalData.cpp: 11559 (JSC::JSGlobalData::createNonDefault): 11560 (JSC::JSGlobalData::create): 11561 (JSC::JSGlobalData::sharedInstance): 11562 11563 - creation of JSGlobalData objects, other than for API usage, associate themselves with the current thread. 11564 11565 * runtime/JSGlobalData.h: 11566 * runtime/UStringImpl.cpp: 11567 (JSC::UStringImpl::destroy): 11568 11569 - destroy() method should be using isIdentifier(). 11570 11571 * runtime/UStringImpl.h: 11572 (JSC::UStringImpl::isIdentifier): 11573 (JSC::UStringImpl::setIsIdentifier): 11574 (JSC::UStringImpl::checkConsistency): 11575 (JSC::UStringImpl::UStringImpl): 11576 11577 - replace m_identifierTable with a single m_isIdentifier bit. 11578 11579 * wtf/StringHashFunctions.h: 11580 (WTF::stringHash): 11581 11582 - change string hash result from 32-bit to 31-bit, to free a bit in UStringImpl for m_isIdentifier. 11583 115842009-12-25 Patrick Gansterer <paroga@paroga.com> 11585 11586 Reviewed by Eric Seidel. 11587 11588 Buildfix for WinCE + style fixes. 11589 https://bugs.webkit.org/show_bug.cgi?id=32939 11590 11591 * jsc.cpp: 11592 (functionPrint): 11593 (functionQuit): 11594 (parseArguments): 11595 (fillBufferWithContentsOfFile): 11596 115972010-01-05 Patrick Gansterer <paroga@paroga.com> 11598 11599 Reviewed by Eric Seidel. 11600 11601 WinCE buildfix after r52791 (renamed PLATFORM(WINCE) to OS(WINCE)). 11602 https://bugs.webkit.org/show_bug.cgi?id=33205 11603 11604 * jit/ExecutableAllocator.h: 11605 116062010-01-05 Patrick Gansterer <paroga@paroga.com> 11607 11608 Reviewed by Darin Adler. 11609 11610 Added compiler error for unsupported platforms. 11611 https://bugs.webkit.org/show_bug.cgi?id=33112 11612 11613 * jit/JITStubs.cpp: 11614 116152010-01-05 Gabor Loki <loki@webkit.org> 11616 11617 Reviewed by Maciej Stachowiak. 11618 11619 Follow r52729 in ARMAssembler. 11620 https://bugs.webkit.org/show_bug.cgi?id=33208 11621 11622 Use WTF_ARM_ARCH_AT_LEAST instead of ARM_ARCH_VERSION 11623 11624 * assembler/ARMAssembler.cpp: 11625 (JSC::ARMAssembler::encodeComplexImm): Move tmp declaration to ARMv7 11626 * assembler/ARMAssembler.h: 11627 (JSC::ARMAssembler::): 11628 (JSC::ARMAssembler::bkpt): 11629 116302010-01-05 Maciej Stachowiak <mjs@apple.com> 11631 11632 Unreviewed build fix for Gtk+ 11633 11634 Don't use // comments in Platform.h, at least some of them seem to make the version of GCC 11635 used on the Gtk buildbot unhappy. 11636 11637 * wtf/Platform.h: 11638 116392010-01-04 Maciej Stachowiak <mjs@apple.com> 11640 11641 Reviewed by Darin Fisher. 11642 11643 Reorganize, document and rename OS() platform macros. 11644 https://bugs.webkit.org/show_bug.cgi?id=33198 11645 11646 * wtf/Platform.h: Rename, reorganize and document OS() macros. 11647 11648 Adapt to name changes. Also fixed a few incorrect OS checks. 11649 11650 * API/JSContextRef.cpp: 11651 * assembler/MacroAssemblerARM.cpp: 11652 (JSC::isVFPPresent): 11653 * assembler/MacroAssemblerX86Common.h: 11654 * bytecode/SamplingTool.cpp: 11655 * config.h: 11656 * interpreter/RegisterFile.cpp: 11657 (JSC::RegisterFile::~RegisterFile): 11658 * interpreter/RegisterFile.h: 11659 (JSC::RegisterFile::RegisterFile): 11660 (JSC::RegisterFile::grow): 11661 * jit/ExecutableAllocator.h: 11662 * jit/ExecutableAllocatorFixedVMPool.cpp: 11663 * jit/ExecutableAllocatorPosix.cpp: 11664 * jit/ExecutableAllocatorSymbian.cpp: 11665 * jit/ExecutableAllocatorWin.cpp: 11666 * jit/JITOpcodes.cpp: 11667 (JSC::JIT::privateCompileCTIMachineTrampolines): 11668 * jit/JITStubs.cpp: 11669 * jsc.cpp: 11670 (main): 11671 * parser/Grammar.y: 11672 * profiler/ProfileNode.cpp: 11673 (JSC::getCount): 11674 * runtime/Collector.cpp: 11675 (JSC::Heap::Heap): 11676 (JSC::Heap::allocateBlock): 11677 (JSC::Heap::freeBlockPtr): 11678 (JSC::currentThreadStackBase): 11679 (JSC::getCurrentPlatformThread): 11680 (JSC::suspendThread): 11681 (JSC::resumeThread): 11682 (JSC::getPlatformThreadRegisters): 11683 (JSC::otherThreadStackPointer): 11684 * runtime/Collector.h: 11685 * runtime/DateConstructor.cpp: 11686 * runtime/DatePrototype.cpp: 11687 (JSC::formatLocaleDate): 11688 * runtime/InitializeThreading.cpp: 11689 (JSC::initializeThreading): 11690 * runtime/MarkStack.h: 11691 (JSC::MarkStack::MarkStackArray::shrinkAllocation): 11692 * runtime/MarkStackPosix.cpp: 11693 * runtime/MarkStackSymbian.cpp: 11694 * runtime/MarkStackWin.cpp: 11695 * runtime/StringPrototype.cpp: 11696 (JSC::stringProtoFuncLastIndexOf): 11697 * runtime/TimeoutChecker.cpp: 11698 (JSC::getCPUTime): 11699 * runtime/UString.cpp: 11700 (JSC::UString::from): 11701 * wtf/Assertions.cpp: 11702 * wtf/Assertions.h: 11703 * wtf/CurrentTime.cpp: 11704 (WTF::lowResUTCTime): 11705 * wtf/CurrentTime.h: 11706 (WTF::getLocalTime): 11707 * wtf/DateMath.cpp: 11708 * wtf/FastMalloc.cpp: 11709 (WTF::TCMalloc_ThreadCache::InitModule): 11710 (WTF::TCMallocStats::): 11711 * wtf/FastMalloc.h: 11712 * wtf/MathExtras.h: 11713 * wtf/RandomNumber.cpp: 11714 (WTF::randomNumber): 11715 * wtf/RandomNumberSeed.h: 11716 (WTF::initializeRandomNumberGenerator): 11717 * wtf/StringExtras.h: 11718 * wtf/TCSpinLock.h: 11719 (TCMalloc_SpinLock::Unlock): 11720 (TCMalloc_SlowLock): 11721 * wtf/TCSystemAlloc.cpp: 11722 * wtf/ThreadSpecific.h: 11723 (WTF::::destroy): 11724 * wtf/Threading.h: 11725 * wtf/ThreadingPthreads.cpp: 11726 (WTF::initializeThreading): 11727 (WTF::isMainThread): 11728 * wtf/ThreadingWin.cpp: 11729 (WTF::wtfThreadEntryPoint): 11730 (WTF::createThreadInternal): 11731 * wtf/VMTags.h: 11732 * wtf/unicode/icu/CollatorICU.cpp: 11733 (WTF::Collator::userDefault): 11734 * wtf/win/MainThreadWin.cpp: 11735 (WTF::initializeMainThreadPlatform): 11736 117372010-01-04 Gustavo Noronha Silva <gns@gnome.org> 11738 11739 Add missing files to the build system - make distcheck build fix. 11740 11741 * GNUmakefile.am: 11742 117432010-01-04 Gavin Barraclough <barraclough@apple.com> 11744 11745 Reviewed by Sam Weinig, additional coding by Mark Rowe. 11746 11747 https://bugs.webkit.org/show_bug.cgi?id=33163 11748 Add string hashing functions to WTF. 11749 Use WTF's string hashing functions from UStringImpl. 11750 11751 * GNUmakefile.am: 11752 * JavaScriptCore.exp: 11753 * JavaScriptCore.gypi: 11754 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 11755 * JavaScriptCore.xcodeproj/project.pbxproj: 11756 * runtime/UStringImpl.cpp: 11757 * runtime/UStringImpl.h: 11758 (JSC::UStringImpl::computeHash): 11759 * wtf/HashFunctions.h: 11760 * wtf/StringHashFunctions.h: Added. 11761 (WTF::stringHash): 11762 117632010-01-04 Dmitry Titov <dimich@chromium.org> 11764 11765 Not reviewed, attempt to fix ARM bulid. 11766 11767 * wtf/Platform.h: 11768 117692010-01-04 Gavin Barraclough <barraclough@apple.com> 11770 11771 Rubber stamped by Geoff Garen. 11772 11773 Add an 'isIdentifier' to UStringImpl, use this where appropriate 11774 (where previously 'identifierTable' was being tested). 11775 11776 * API/JSClassRef.cpp: 11777 (OpaqueJSClass::~OpaqueJSClass): 11778 (OpaqueJSClassContextData::OpaqueJSClassContextData): 11779 * runtime/Identifier.cpp: 11780 (JSC::Identifier::addSlowCase): 11781 * runtime/Identifier.h: 11782 (JSC::Identifier::add): 11783 * runtime/PropertyNameArray.cpp: 11784 (JSC::PropertyNameArray::add): 11785 * runtime/UStringImpl.h: 11786 (JSC::UStringImpl::isIdentifier): 11787 117882010-01-04 Gavin Barraclough <barraclough@apple.com> 11789 11790 Reviewed by Sam "Shimmey Shimmey" Weinig. 11791 11792 https://bugs.webkit.org/show_bug.cgi?id=33158 11793 Refactor JSC API entry/exit to use RAII instead of copy/pasting code. 11794 Make it easier to change set of actions taken when passing across the API boundary. 11795 11796 * API/APIShims.h: Added. 11797 (JSC::APIEntryShimWithoutLock::APIEntryShimWithoutLock): 11798 (JSC::APIEntryShimWithoutLock::~APIEntryShimWithoutLock): 11799 (JSC::APIEntryShim::APIEntryShim): 11800 (JSC::APICallbackShim::APICallbackShim): 11801 (JSC::APICallbackShim::~APICallbackShim): 11802 * API/JSBase.cpp: 11803 (JSEvaluateScript): 11804 (JSCheckScriptSyntax): 11805 (JSGarbageCollect): 11806 (JSReportExtraMemoryCost): 11807 * API/JSCallbackConstructor.cpp: 11808 (JSC::constructJSCallback): 11809 * API/JSCallbackFunction.cpp: 11810 (JSC::JSCallbackFunction::call): 11811 * API/JSCallbackObjectFunctions.h: 11812 (JSC::::init): 11813 (JSC::::getOwnPropertySlot): 11814 (JSC::::put): 11815 (JSC::::deleteProperty): 11816 (JSC::::construct): 11817 (JSC::::hasInstance): 11818 (JSC::::call): 11819 (JSC::::getOwnPropertyNames): 11820 (JSC::::toNumber): 11821 (JSC::::toString): 11822 (JSC::::staticValueGetter): 11823 (JSC::::callbackGetter): 11824 * API/JSContextRef.cpp: 11825 * API/JSObjectRef.cpp: 11826 (JSObjectMake): 11827 (JSObjectMakeFunctionWithCallback): 11828 (JSObjectMakeConstructor): 11829 (JSObjectMakeFunction): 11830 (JSObjectMakeArray): 11831 (JSObjectMakeDate): 11832 (JSObjectMakeError): 11833 (JSObjectMakeRegExp): 11834 (JSObjectGetPrototype): 11835 (JSObjectSetPrototype): 11836 (JSObjectHasProperty): 11837 (JSObjectGetProperty): 11838 (JSObjectSetProperty): 11839 (JSObjectGetPropertyAtIndex): 11840 (JSObjectSetPropertyAtIndex): 11841 (JSObjectDeleteProperty): 11842 (JSObjectCallAsFunction): 11843 (JSObjectCallAsConstructor): 11844 (JSObjectCopyPropertyNames): 11845 (JSPropertyNameArrayRelease): 11846 (JSPropertyNameAccumulatorAddName): 11847 * API/JSValueRef.cpp: 11848 (JSValueGetType): 11849 (JSValueIsUndefined): 11850 (JSValueIsNull): 11851 (JSValueIsBoolean): 11852 (JSValueIsNumber): 11853 (JSValueIsString): 11854 (JSValueIsObject): 11855 (JSValueIsObjectOfClass): 11856 (JSValueIsEqual): 11857 (JSValueIsStrictEqual): 11858 (JSValueIsInstanceOfConstructor): 11859 (JSValueMakeUndefined): 11860 (JSValueMakeNull): 11861 (JSValueMakeBoolean): 11862 (JSValueMakeNumber): 11863 (JSValueMakeString): 11864 (JSValueToBoolean): 11865 (JSValueToNumber): 11866 (JSValueToStringCopy): 11867 (JSValueToObject): 11868 (JSValueProtect): 11869 (JSValueUnprotect): 11870 * JavaScriptCore.xcodeproj/project.pbxproj: 11871 118722010-01-04 Dan Bernstein <mitz@apple.com> 11873 11874 Reviewed by Ada Chan and Mark Rowe. 11875 11876 Updated copyright string 11877 11878 * Info.plist: 11879 * JavaScriptCore.vcproj/JavaScriptCore.resources/Info.plist: 11880 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.rc: 11881 118822010-01-04 Adam Roben <aroben@apple.com> 11883 11884 No review, rolling out r52741. 11885 http://trac.webkit.org/changeset/52741 11886 https://bugs.webkit.org/show_bug.cgi?id=33056 11887 11888 * wtf/AlwaysInline.h: 11889 118902010-01-04 Patrick Gansterer <paroga@paroga.com> 11891 11892 Reviewed by Darin Adler. 11893 11894 Add cacheFlush support for WinCE 11895 https://bugs.webkit.org/show_bug.cgi?id=33110 11896 11897 * jit/ExecutableAllocator.h: 11898 (JSC::ExecutableAllocator::cacheFlush): 11899 119002010-01-04 Patrick Gansterer <paroga@paroga.com> 11901 11902 Reviewed by Adam Roben. 11903 11904 Implement NO_RETURN for COMPILER(MSVC). 11905 https://bugs.webkit.org/show_bug.cgi?id=33056 11906 11907 * wtf/AlwaysInline.h: 11908 119092010-01-04 Maciej Stachowiak <mjs@apple.com> 11910 11911 Reviewed by Simon Hausmann. 11912 11913 Fix some PLATFORM(*_ENDIAN) uses to CPU() 11914 https://bugs.webkit.org/show_bug.cgi?id=33148 11915 11916 * runtime/JSCell.cpp: 11917 (JSC::): 11918 * runtime/JSValue.h: 11919 (JSC::JSValue::): 11920 119212010-01-04 Maciej Stachowiak <mjs@apple.com> 11922 11923 Reviewed by Adam Barth. 11924 11925 Document CPU() macros in comments. 11926 https://bugs.webkit.org/show_bug.cgi?id=33147 11927 11928 * wtf/Platform.h: 11929 119302010-01-04 Maciej Stachowiak <mjs@apple.com> 11931 11932 Reviewed by Adam Barth. 11933 11934 Reorganize, document and rename CPU() platform macros. 11935 https://bugs.webkit.org/show_bug.cgi?id=33145 11936 ExecutableAllocatorSymbian appears to have buggy ARM version check 11937 https://bugs.webkit.org/show_bug.cgi?id=33138 11938 11939 * wtf/Platform.h: 11940 Rename all macros related to detection of particular CPUs or 11941 classes of CPUs to CPU(), reorganize and document them. 11942 11943 All remaining changes are adapting to the renames, plus fixing the 11944 second bug cited above. 11945 11946 * assembler/ARMAssembler.cpp: 11947 * assembler/ARMAssembler.h: 11948 * assembler/ARMv7Assembler.h: 11949 * assembler/AbstractMacroAssembler.h: 11950 (JSC::AbstractMacroAssembler::Imm32::Imm32): 11951 * assembler/MacroAssembler.h: 11952 * assembler/MacroAssemblerARM.cpp: 11953 * assembler/MacroAssemblerARM.h: 11954 * assembler/MacroAssemblerCodeRef.h: 11955 (JSC::MacroAssemblerCodePtr::MacroAssemblerCodePtr): 11956 * assembler/MacroAssemblerX86.h: 11957 * assembler/MacroAssemblerX86Common.h: 11958 * assembler/MacroAssemblerX86_64.h: 11959 * assembler/X86Assembler.h: 11960 (JSC::X86Registers::): 11961 (JSC::X86Assembler::): 11962 (JSC::X86Assembler::movl_mEAX): 11963 (JSC::X86Assembler::movl_EAXm): 11964 (JSC::X86Assembler::repatchLoadPtrToLEA): 11965 (JSC::X86Assembler::X86InstructionFormatter::memoryModRM): 11966 * jit/ExecutableAllocator.h: 11967 * jit/ExecutableAllocatorFixedVMPool.cpp: 11968 * jit/ExecutableAllocatorPosix.cpp: 11969 * jit/ExecutableAllocatorSymbian.cpp: 11970 (JSC::ExecutableAllocator::intializePageSize): 11971 * jit/JIT.cpp: 11972 * jit/JIT.h: 11973 * jit/JITArithmetic.cpp: 11974 * jit/JITInlineMethods.h: 11975 (JSC::JIT::beginUninterruptedSequence): 11976 (JSC::JIT::restoreArgumentReferenceForTrampoline): 11977 (JSC::JIT::emitCount): 11978 * jit/JITOpcodes.cpp: 11979 (JSC::JIT::privateCompileCTIMachineTrampolines): 11980 * jit/JITPropertyAccess.cpp: 11981 (JSC::JIT::privateCompileGetByIdProto): 11982 (JSC::JIT::privateCompileGetByIdProtoList): 11983 (JSC::JIT::privateCompileGetByIdChainList): 11984 (JSC::JIT::privateCompileGetByIdChain): 11985 * jit/JITStubs.cpp: 11986 (JSC::JITThunks::JITThunks): 11987 * jit/JITStubs.h: 11988 * runtime/Collector.cpp: 11989 (JSC::currentThreadStackBase): 11990 (JSC::getPlatformThreadRegisters): 11991 (JSC::otherThreadStackPointer): 11992 * wrec/WREC.h: 11993 * wrec/WRECGenerator.cpp: 11994 (JSC::WREC::Generator::generateEnter): 11995 (JSC::WREC::Generator::generateReturnSuccess): 11996 (JSC::WREC::Generator::generateReturnFailure): 11997 * wrec/WRECGenerator.h: 11998 * wtf/FastMalloc.cpp: 11999 * wtf/TCSpinLock.h: 12000 (TCMalloc_SpinLock::Lock): 12001 (TCMalloc_SpinLock::Unlock): 12002 (TCMalloc_SlowLock): 12003 * wtf/Threading.h: 12004 * wtf/dtoa.cpp: 12005 * yarr/RegexJIT.cpp: 12006 (JSC::Yarr::RegexGenerator::generateEnter): 12007 (JSC::Yarr::RegexGenerator::generateReturn): 12008 * yarr/RegexJIT.h: 12009 120102010-01-04 Maciej Stachowiak <mjs@apple.com> 12011 12012 Reviewed by Adam Barth. 12013 12014 Clean up COMPILER macros and remove unused ones. 12015 https://bugs.webkit.org/show_bug.cgi?id=33132 12016 12017 Removed values are COMPILER(BORLAND) and COMPILER(CYGWIN) - they were 12018 not used anywhere. 12019 12020 * wtf/Platform.h: 12021 120222010-01-03 Maciej Stachowiak <mjs@apple.com> 12023 12024 Reviewed by Eric Seidel. 12025 12026 Update wtf/Platform.h to document the new system for porting macros. 12027 https://bugs.webkit.org/show_bug.cgi?id=33130 12028 12029 * wtf/Platform.h: 12030 120312009-12-29 Laszlo Gombos <laszlo.1.gombos@nokia.com> 12032 12033 Reviewed by Maciej Stachowiak. 12034 12035 PLATFORM(CAIRO) should be defined by WIN_CAIRO define 12036 https://bugs.webkit.org/show_bug.cgi?id=22250 12037 12038 * wtf/Platform.h: Define WTF_PLATFORM_CAIRO for GTK port only 12039 For the WinCairo port WTF_PLATFORM_CAIRO is already defined in config.h 12040 120412009-12-28 Shu Chang <Chang.Shu@nokia.com> 12042 12043 Reviewed by Laszlo Gombos. 12044 12045 [Qt] Delete ThreadPrivate instance after it is finished. 12046 https://bugs.webkit.org/show_bug.cgi?id=32614 12047 12048 * wtf/qt/ThreadingQt.cpp: 12049 (WTF::ThreadMonitor::instance): 12050 (WTF::ThreadMonitor::threadFinished): 12051 (WTF::createThreadInternal): 12052 (WTF::detachThread): 12053 120542009-12-28 Patrick Gansterer <paroga@paroga.com> 12055 12056 Reviewed by Maciej Stachowiak. 12057 12058 Cleanup of #define JS_EXPORT. 12059 12060 * API/JSBase.h: 12061 120622009-12-27 Patrick Gansterer <paroga@paroga.com> 12063 12064 Reviewed by Adam Barth. 12065 12066 WinCE buildfix (HWND_MESSAGE isn't supported there) 12067 12068 * wtf/win/MainThreadWin.cpp: 12069 (WTF::initializeMainThreadPlatform): 12070 120712009-12-27 Patrick Gansterer <paroga@paroga.com> 12072 12073 Reviewed by Adam Barth. 12074 12075 Added a file with WinMain function to link agains in WinCE. 12076 12077 * os-win32/WinMain.cpp: Added. 12078 (convertToUtf8): 12079 (WinMain): 12080 120812009-12-24 Laszlo Gombos <laszlo.1.gombos@nokia.com> 12082 12083 Unreviewed; revert of r52550. 12084 12085 The change regressed the following LayoutTests for QtWebKit. 12086 12087 fast/workers/worker-call.html -> crashed 12088 fast/workers/worker-close.html -> crashed 12089 12090 * wtf/qt/ThreadingQt.cpp: 12091 (WTF::waitForThreadCompletion): 12092 (WTF::detachThread): 12093 120942009-12-24 Shu Chang <Chang.Shu@nokia.com> 12095 12096 Reviewed by Laszlo Gombos. 12097 12098 [Qt] Fix memory leak by deleting instance of ThreadPrivate 12099 in function waitForThreadCompletion(), synchronously, or in 12100 detachThread(), asynchronously. 12101 https://bugs.webkit.org/show_bug.cgi?id=32614 12102 12103 * wtf/qt/ThreadingQt.cpp: 12104 (WTF::waitForThreadCompletion): 12105 (WTF::detachThread): 12106 121072009-12-23 Kwang Yul Seo <skyul@company100.net> 12108 12109 Reviewed by Laszlo Gombos. 12110 12111 Include stddef.h for ptrdiff_t 12112 https://bugs.webkit.org/show_bug.cgi?id=32891 12113 12114 ptrdiff_t is typedef-ed in stddef.h. 12115 Include stddef.h in jit/ExecutableAllocator.h. 12116 12117 * jit/ExecutableAllocator.h: 12118 121192009-12-23 Patrick Gansterer <paroga@paroga.com> 12120 12121 Reviewed by Eric Seidel. 12122 12123 Buildfix after r47092. 12124 12125 * wtf/wince/MemoryManager.cpp: 12126 (WTF::tryFastMalloc): 12127 (WTF::tryFastZeroedMalloc): 12128 (WTF::tryFastCalloc): 12129 (WTF::tryFastRealloc): 12130 121312009-12-23 Kent Tamura <tkent@chromium.org> 12132 12133 Reviewed by Darin Adler. 12134 12135 HTMLInputElement::valueAsDate getter support. 12136 https://bugs.webkit.org/show_bug.cgi?id=32876 12137 12138 Expose dateToDaysFrom1970(). 12139 12140 * JavaScriptCore.exp: 12141 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 12142 * wtf/DateMath.cpp: 12143 (WTF::dateToDaysFrom1970): 12144 * wtf/DateMath.h: 12145 121462009-12-22 Darin Adler <darin@apple.com> 12147 12148 Reviewed by Mark Rowe. 12149 12150 Turn off datagrid by default, at least for all platforms Apple ships. 12151 The datagrid implementation isn't ready for general web use yet. 12152 12153 * Configurations/FeatureDefines.xcconfig: Turn off datagrid by default. 12154 121552009-12-22 Steve Block <steveblock@google.com> 12156 12157 Reviewed by David Levin. 12158 12159 Updates Android's scheduleDispatchFunctionsOnMainThread() to use new 12160 AndroidThreading class, rather than using JavaSharedClient directly. 12161 This fixes the current layering violation. 12162 https://bugs.webkit.org/show_bug.cgi?id=32651 12163 12164 The pattern is copied from Chromium, which uses the ChromiumThreading 12165 class. This patch also fixes the style in ChromiumThreading.h. 12166 12167 * wtf/android/AndroidThreading.h: Added. Declares AndroidThreading. 12168 * wtf/android/MainThreadAndroid.cpp: Modified 12169 (WTF::scheduleDispatchFunctionsOnMainThread): Uses AndroidThreading. 12170 * wtf/chromium/ChromiumThreading.h: Modified. Fixes style. 12171 121722009-12-22 Gavin Barraclough <barraclough@apple.com> 12173 12174 Reviewed by Sam Weinig. 12175 12176 Fix a couple of problems with UntypedPtrAndBitfield. 12177 12178 Add a m_leaksPtr to reduce false positives from leaks in debug builds 12179 (this isn't perfect because we'd like a solution for release builds, 12180 but this is now at least as good as a PtrAndFlags would be). 12181 12182 Switch SmallStringsto use a regular string for the base, rather than 12183 a static one. UntypedPtrAndBitfield assumes all strings are at least 12184 8 byte aligned; this migt not be true of static strings. Shared buffers 12185 are heap allocated, as are all UStringImpls other than static strings. 12186 Static strings cannot end up being the owner string of substrings, 12187 since the only static strings are length 0. 12188 12189 * runtime/SmallStrings.cpp: 12190 (JSC::SmallStringsStorage::SmallStringsStorage): 12191 * runtime/UStringImpl.h: 12192 (JSC::UntypedPtrAndBitfield::UntypedPtrAndBitfield): 12193 (JSC::UStringImpl::UStringImpl): 12194 121952009-12-22 Kwang Yul Seo <skyul@company100.net> 12196 12197 Reviewed by Darin Adler. 12198 12199 RVCT (__ARMCC_VERSION < 400000) does not provide strcasecmp and strncasecmp 12200 https://bugs.webkit.org/show_bug.cgi?id=32857 12201 12202 Add implementation of strcasecmp and strncasecmp for RVCT < 4.0 12203 because earlier versions of RVCT 4.0 does not provide these functions. 12204 12205 * wtf/StringExtras.cpp: Added. 12206 (strcasecmp): 12207 (strncasecmp): 12208 * wtf/StringExtras.h: 12209 122102009-12-22 Kwang Yul Seo <skyul@company100.net> 12211 12212 Reviewed by Darin Adler. 12213 12214 Define ALWAYS_INLINE and WTF_PRIVATE_INLINE to __forceinline for RVCT 12215 https://bugs.webkit.org/show_bug.cgi?id=32853 12216 12217 Use __forceinline forces RVCT to compile a C or C++ function 12218 inline. The compiler attempts to inline the function, regardless of 12219 the characteristics of the function. 12220 12221 * wtf/AlwaysInline.h: 12222 * wtf/FastMalloc.h: 12223 122242009-12-21 Simon Hausmann <simon.hausmann@nokia.com> 12225 12226 Prospective GTK build fix: Add UStringImpl.cpp/h to the build. 12227 12228 * GNUmakefile.am: 12229 122302009-12-21 Simon Hausmann <simon.hausmann@nokia.com> 12231 12232 Fix the Qt build, add UStringImpl.cpp to the build. 12233 12234 * JavaScriptCore.pri: 12235 122362009-12-21 Gavin Barraclough <barraclough@apple.com> 12237 12238 Windows Build fix part 5. 12239 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: 12240 122412009-12-21 Gavin Barraclough <barraclough@apple.com> 12242 12243 Reviewed by NOBODY (build fix). 12244 Fix breakage of world introduced in build fix to r52463. 12245 12246 * runtime/UStringImpl.h: 12247 122482009-12-21 Gavin Barraclough <barraclough@apple.com> 12249 12250 Reviewed by Darin Adler. 12251 12252 https://bugs.webkit.org/show_bug.cgi?id=32831 12253 Replace UString::Rep implementation, following introduction of ropes to JSC. 12254 12255 * Remove redundant overcapacity mechanisms. 12256 * Reduce memory cost of Rep's. 12257 * Add an inline storage mechanism akin to that in WebCore's StringImpl. 12258 12259 ~1% Sunspider progression. 12260 12261 * JavaScriptCore.exp: 12262 * JavaScriptCore.xcodeproj/project.pbxproj: 12263 * runtime/JSString.cpp: 12264 (JSC::JSString::resolveRope): 12265 * runtime/SmallStrings.cpp: 12266 (JSC::SmallStringsStorage::SmallStringsStorage): 12267 * runtime/UString.cpp: 12268 (JSC::initializeUString): 12269 (JSC::createRep): 12270 (JSC::UString::createFromUTF8): 12271 (JSC::UString::createUninitialized): 12272 (JSC::UString::spliceSubstringsWithSeparators): 12273 (JSC::UString::replaceRange): 12274 (JSC::UString::ascii): 12275 (JSC::UString::operator=): 12276 (JSC::UString::toStrictUInt32): 12277 (JSC::equal): 12278 * runtime/UString.h: 12279 (JSC::UString::isEmpty): 12280 (JSC::UString::cost): 12281 (JSC::makeString): 12282 * runtime/UStringImpl.cpp: Added. 12283 (JSC::UStringImpl::baseSharedBuffer): 12284 (JSC::UStringImpl::sharedBuffer): 12285 (JSC::UStringImpl::destroy): 12286 (JSC::UStringImpl::computeHash): 12287 * runtime/UStringImpl.h: Added. 12288 (JSC::UntypedPtrAndBitfield::UntypedPtrAndBitfield): 12289 (JSC::UntypedPtrAndBitfield::asPtr): 12290 (JSC::UntypedPtrAndBitfield::operator&=): 12291 (JSC::UntypedPtrAndBitfield::operator|=): 12292 (JSC::UntypedPtrAndBitfield::operator&): 12293 (JSC::UStringImpl::create): 12294 (JSC::UStringImpl::createCopying): 12295 (JSC::UStringImpl::createUninitialized): 12296 (JSC::UStringImpl::data): 12297 (JSC::UStringImpl::size): 12298 (JSC::UStringImpl::cost): 12299 (JSC::UStringImpl::hash): 12300 (JSC::UStringImpl::computedHash): 12301 (JSC::UStringImpl::setHash): 12302 (JSC::UStringImpl::identifierTable): 12303 (JSC::UStringImpl::setIdentifierTable): 12304 (JSC::UStringImpl::ref): 12305 (JSC::UStringImpl::deref): 12306 (JSC::UStringImpl::allocChars): 12307 (JSC::UStringImpl::copyChars): 12308 (JSC::UStringImpl::computeHash): 12309 (JSC::UStringImpl::null): 12310 (JSC::UStringImpl::empty): 12311 (JSC::UStringImpl::checkConsistency): 12312 (JSC::UStringImpl::): 12313 (JSC::UStringImpl::UStringImpl): 12314 (JSC::UStringImpl::operator new): 12315 (JSC::UStringImpl::bufferOwnerString): 12316 (JSC::UStringImpl::bufferOwnership): 12317 (JSC::UStringImpl::isStatic): 12318 123192009-12-18 Laszlo Gombos <laszlo.1.gombos@nokia.com> 12320 12321 Reviewed by Kenneth Rohde Christiansen. 12322 12323 Move some build decisions from Qt build system into source files 12324 https://bugs.webkit.org/show_bug.cgi?id=31956 12325 12326 * JavaScriptCore.pri: Compile files unconditionally 12327 * jit/ExecutableAllocatorPosix.cpp: Guard with PLATFORM(UNIX) && !PLATFORM(SYMBIAN) 12328 * jit/ExecutableAllocatorWin.cpp: Guard with PLATFORM(WIN_OS) 12329 * runtime/MarkStackPosix.cpp: Guard with PLATFORM(UNIX) && !PLATFORM(SYMBIAN) 12330 * runtime/MarkStackSymbian.cpp: Guard with PLATFORM(SYMBIAN) 12331 * runtime/MarkStackWin.cpp: Guard with PLATFORM(WIN_OS) 12332 * wtf/Platform.h: Guard ENABLE_JSC_MULTIPLE_THREADS with ENABLE_SINGLE_THREADED for the Qt port 12333 * wtf/ThreadingNone.cpp: Guard with ENABLE(SINGLE_THREADED) 12334 * wtf/qt/ThreadingQt.cpp: Guard with !ENABLE(SINGLE_THREADED) 12335 123362009-12-18 Gavin Barraclough <barraclough@apple.com> 12337 12338 Reviewed by Sam Weinig. 12339 12340 Add createNonCopying method to UString to make replace constructor passed bool, 12341 to make behaviour more explicit. Add createFromUTF8 to UString (wrapping method 12342 on UString::Rep), since other cases of transliteration (e.g. from ascii) are 12343 performed in UString constructors. Add/use setHash & size() accessors on Rep, 12344 rather than accessing _hash/len directly. 12345 12346 * API/JSClassRef.cpp: 12347 (OpaqueJSClass::OpaqueJSClass): 12348 * API/OpaqueJSString.cpp: 12349 (OpaqueJSString::ustring): 12350 * JavaScriptCore.exp: 12351 * runtime/ArrayPrototype.cpp: 12352 (JSC::arrayProtoFuncToString): 12353 * runtime/Identifier.cpp: 12354 (JSC::Identifier::equal): 12355 (JSC::CStringTranslator::translate): 12356 (JSC::UCharBufferTranslator::translate): 12357 (JSC::Identifier::addSlowCase): 12358 * runtime/JSString.cpp: 12359 (JSC::JSString::resolveRope): 12360 * runtime/JSString.h: 12361 (JSC::JSString::Rope::Fiber::refAndGetLength): 12362 (JSC::JSString::Rope::append): 12363 * runtime/StringBuilder.h: 12364 (JSC::StringBuilder::release): 12365 * runtime/StringConstructor.cpp: 12366 (JSC::stringFromCharCodeSlowCase): 12367 * runtime/StringPrototype.cpp: 12368 (JSC::substituteBackreferencesSlow): 12369 (JSC::stringProtoFuncToLowerCase): 12370 (JSC::stringProtoFuncToUpperCase): 12371 (JSC::stringProtoFuncFontsize): 12372 (JSC::stringProtoFuncLink): 12373 * runtime/UString.cpp: 12374 (JSC::UString::UString): 12375 (JSC::UString::createNonCopying): 12376 (JSC::UString::createFromUTF8): 12377 * runtime/UString.h: 12378 (JSC::UString::Rep::setHash): 12379 (JSC::UString::~UString): 12380 (JSC::makeString): 12381 123822009-12-18 Geoffrey Garen <ggaren@apple.com> 12383 12384 Reviewed by Cameron Zwarich and Gavin Barraclough. 12385 12386 Changed Register constructors to assignment operators, to streamline 12387 moving values into registers. (In theory, there's no difference between 12388 the two, since the constructor should just inline away, but there seems 12389 to be a big difference in the addled mind of the GCC optimizer.) 12390 12391 In the interpreter, this is a 3.5% SunSpider speedup and a 1K-2K 12392 reduction in stack usage per privateExecute stack frame. 12393 12394 * interpreter/CallFrame.h: 12395 (JSC::ExecState::setCalleeArguments): 12396 (JSC::ExecState::setCallerFrame): 12397 (JSC::ExecState::setScopeChain): 12398 (JSC::ExecState::init): 12399 (JSC::ExecState::setArgumentCount): 12400 (JSC::ExecState::setCallee): 12401 (JSC::ExecState::setCodeBlock): Added a little bit of casting so these 12402 functions could use the new Register assignment operators. 12403 12404 * interpreter/Register.h: 12405 (JSC::Register::withInt): 12406 (JSC::Register::Register): 12407 (JSC::Register::operator=): Swapped in assignment operators for constructors. 12408 124092009-12-18 Yongjun Zhang <yongjun.zhang@nokia.com> 12410 12411 Reviewed by Simon Hausmann. 12412 12413 https://bugs.webkit.org/show_bug.cgi?id=32713 12414 [Qt] make wtf/Assertions.h compile in winscw compiler. 12415 12416 Add string arg before ellipsis to help winscw compiler resolve variadic 12417 macro definitions in wtf/Assertions.h. 12418 12419 * wtf/Assertions.h: 12420 124212009-12-18 Geoffrey Garen <ggaren@apple.com> 12422 12423 Reviewed by Adam Roben. 12424 12425 Fixed intermittent failure seen on Windows buildbot, and in other JSC 12426 API clients. 12427 12428 Added a WeakGCPtr class and changed OpaqueJSClass::cachedPrototype to 12429 use it, to avoid vending a stale object as a prototype. 12430 12431 * API/JSClassRef.cpp: 12432 (OpaqueJSClassContextData::OpaqueJSClassContextData): 12433 (OpaqueJSClass::prototype): 12434 * API/JSClassRef.h: Use WeakGCPtr. 12435 12436 * JavaScriptCore.xcodeproj/project.pbxproj: 12437 * runtime/WeakGCPtr.h: Added. 12438 (JSC::WeakGCPtr::WeakGCPtr): 12439 (JSC::WeakGCPtr::get): 12440 (JSC::WeakGCPtr::clear): 12441 (JSC::WeakGCPtr::operator*): 12442 (JSC::WeakGCPtr::operator->): 12443 (JSC::WeakGCPtr::operator!): 12444 (JSC::WeakGCPtr::operator bool): 12445 (JSC::WeakGCPtr::operator UnspecifiedBoolType): 12446 (JSC::WeakGCPtr::assign): 12447 (JSC::::operator): 12448 (JSC::operator==): 12449 (JSC::operator!=): 12450 (JSC::static_pointer_cast): 12451 (JSC::const_pointer_cast): 12452 (JSC::getPtr): Added WeakGCPtr to the project. 12453 124542009-12-18 Gavin Barraclough <barraclough@apple.com> 12455 12456 Reviewed by Sam Weinig. 12457 12458 https://bugs.webkit.org/show_bug.cgi?id=32720 12459 12460 * JavaScriptCore.exp: 12461 - Remove exports for UString::append 12462 * JavaScriptCore.xcodeproj/project.pbxproj: 12463 - Make StringBuilder a private header (was project). 12464 124652009-12-18 Martin Robinson <martin.james.robinson@gmail.com> 12466 12467 Reviewed by Gustavo Noronha Silva. 12468 12469 [GTK] GRefPtr does not take a reference when assigned a raw pointer 12470 https://bugs.webkit.org/show_bug.cgi?id=32709 12471 12472 Ensure that when assigning a raw pointer to a GRefPtr, the reference 12473 count is incremented. Also remove the GRefPtr conversion overload as 12474 GRefPtr types have necessarily incompatible reference counting. 12475 12476 * wtf/gtk/GRefPtr.h: 12477 (WTF::GRefPtr::operator=): 12478 124792009-12-18 Simon Hausmann <simon.hausmann@nokia.com> 12480 12481 Reviewed by Tor Arne Vestbø. 12482 12483 [Qt] Clean up the qmake build system to distinguish between trunk builds and package builds 12484 12485 https://bugs.webkit.org/show_bug.cgi?id=32716 12486 12487 * pcre/pcre.pri: Use standalone_package instead of QTDIR_build 12488 124892009-12-18 Martin Robinson <martin.james.robinson@gmail.com> 12490 12491 Reviewed by Gustavo Noronha Silva. 12492 12493 [GTK] Compile warning from line 29 of GRefPtr.cpp 12494 https://bugs.webkit.org/show_bug.cgi?id=32703 12495 12496 Fix memory leak and compiler warning in GRefPtr GHashTable template 12497 specialization. 12498 12499 * wtf/gtk/GRefPtr.cpp: 12500 (WTF::refGPtr): 12501 125022009-12-17 Sam Weinig <sam@webkit.org> 12503 12504 Reviewed by Mark Rowe. 12505 12506 Add BUILDING_ON_SNOW_LEOPARD and TARGETING_SNOW_LEOPARD #defines. 12507 12508 * wtf/Platform.h: 12509 125102009-12-17 Adam Roben <aroben@apple.com> 12511 12512 Sync JavaScriptCore.vcproj with JavaScriptCore.xcodeproj and the 12513 source tree 12514 12515 Fixes <http://webkit.org/b/32665>. 12516 12517 Reviewed by Ada Chan. 12518 12519 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: Moved 12520 around files and filters so that the structure matches 12521 JavaScriptCore.xcodeproj and the source tree. A few headers that were 12522 previously omitted have been added, as well as JSZombie.{cpp,h}. 12523 125242009-12-17 Adam Roben <aroben@apple.com> 12525 12526 Remove HeavyProfile and TreeProfile completely 12527 12528 These were mostly removed in r42808, but the empty files were left in 12529 place. 12530 12531 Fixes <http://webkit.org/b/32664>. 12532 12533 Reviewed by John Sullivan. 12534 12535 * Android.mk: 12536 * GNUmakefile.am: 12537 * JavaScriptCore.gypi: 12538 * JavaScriptCore.pri: 12539 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: 12540 * JavaScriptCoreSources.bkl: 12541 Removed HeavyProfile/TreeProfile source files. 12542 12543 * profiler/HeavyProfile.cpp: Removed. 12544 * profiler/HeavyProfile.h: Removed. 12545 * profiler/TreeProfile.cpp: Removed. 12546 * profiler/TreeProfile.h: Removed. 12547 125482009-12-17 Martin Robinson <martin.james.robinson@gmail.com> 12549 12550 Reviewed by Gustavo Noronha Silva. 12551 12552 [GTK] WebKit GTK needs a wrapper for ref counted glib/gobject structs 12553 https://bugs.webkit.org/show_bug.cgi?id=21599 12554 12555 Implement GRefPtr, a smart pointer for reference counted GObject types. 12556 12557 * GNUmakefile.am: 12558 * wtf/gtk/GOwnPtr.cpp: 12559 (WTF::GDir): 12560 * wtf/gtk/GRefPtr.h: Added. 12561 (WTF::): 12562 (WTF::GRefPtr::GRefPtr): 12563 (WTF::GRefPtr::~GRefPtr): 12564 (WTF::GRefPtr::clear): 12565 (WTF::GRefPtr::get): 12566 (WTF::GRefPtr::operator*): 12567 (WTF::GRefPtr::operator->): 12568 (WTF::GRefPtr::operator!): 12569 (WTF::GRefPtr::operator UnspecifiedBoolType): 12570 (WTF::GRefPtr::hashTableDeletedValue): 12571 (WTF::::operator): 12572 (WTF::::swap): 12573 (WTF::swap): 12574 (WTF::operator==): 12575 (WTF::operator!=): 12576 (WTF::static_pointer_cast): 12577 (WTF::const_pointer_cast): 12578 (WTF::getPtr): 12579 (WTF::adoptGRef): 12580 (WTF::refGPtr): 12581 (WTF::derefGPtr): 12582 125832009-12-17 Gustavo Noronha Silva <gustavo.noronha@collabora.co.uk> 12584 12585 Unreviewed. Build fixes for make distcheck. 12586 12587 * GNUmakefile.am: 12588 125892009-12-16 Geoffrey Garen <ggaren@apple.com> 12590 12591 Reviewed by Oliver Hunt. 12592 12593 Fixed <rdar://problem/7355025> Interpreter::privateExecute macro generates 12594 bloated code 12595 12596 This patch cuts Interpreter stack use by about a third. 12597 12598 * bytecode/Opcode.h: Changed Opcode to const void* to work with the 12599 const static initiliazation we want to do in Interpreter::privateExecute. 12600 12601 * interpreter/Interpreter.cpp: 12602 (JSC::Interpreter::Interpreter): Moved hashtable initialization here to 12603 avoid polluting Interpreter::privateExecute's stack, and changed it from a 12604 series of add() calls to one add() call in a loop, to cut down on code size. 12605 12606 (JSC::Interpreter::privateExecute): Changed a series of label computations 12607 to a copy of a compile-time constant array to cut down on code size. 12608 126092009-12-16 Mark Rowe <mrowe@apple.com> 12610 12611 Build fix. Disable debug variants of WebKit frameworks. 12612 12613 * JavaScriptCore.xcodeproj/project.pbxproj: 12614 126152009-12-15 Geoffrey Garen <ggaren@apple.com> 12616 12617 Reviewed by Sam "r=me" Weinig. 12618 12619 https://bugs.webkit.org/show_bug.cgi?id=32498 12620 <rdar://problem/7471495> 12621 REGRESSION(r51978-r52039): AJAX "Mark This Forum Read" function no longer 12622 works 12623 12624 Fixed a tyop. 12625 12626 * runtime/Operations.h: 12627 (JSC::jsAdd): Use the '&&' operator, not the ',' operator. 12628 126292009-12-15 Geoffrey Garen <ggaren@apple.com> 12630 12631 Try to fix the windows build: don't export this inlined function. 12632 12633 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 12634 126352009-12-15 Geoffrey Garen <ggaren@apple.com> 12636 12637 Reviewed by Beth Dakin. 12638 12639 Inlined JSCell's operator new. 12640 12641 3.7% speedup on bench-allocate-nonretained.js. 12642 12643 * JavaScriptCore.exp: 12644 * runtime/JSCell.cpp: 12645 * runtime/JSCell.h: 12646 (JSC::JSCell::operator new): 12647 126482009-12-15 Geoffrey Garen <ggaren@apple.com> 12649 12650 Reviewed by Oliver Hunt. 12651 12652 Removed the number heap, replacing it with a one-item free list for 12653 numbers, taking advantage of the fact that two number cells fit inside 12654 the space for one regular cell, and number cells don't require destruction. 12655 12656 SunSpider says 1.6% faster in JSVALUE32 mode (the only mode that 12657 heap-allocates numbers). 12658 12659 SunSpider says 1.1% faster in JSVALUE32_64 mode. v8 says 0.8% faster 12660 in JSVALUE32_64 mode. 10% speedup on bench-alloc-nonretained.js. 6% 12661 speedup on bench-alloc-retained.js. 12662 12663 There's a lot of formulaic change in this patch, but not much substance. 12664 12665 * JavaScriptCore.exp: 12666 * debugger/Debugger.cpp: 12667 (JSC::Debugger::recompileAllJSFunctions): 12668 * runtime/Collector.cpp: 12669 (JSC::Heap::Heap): 12670 (JSC::Heap::destroy): 12671 (JSC::Heap::allocateBlock): 12672 (JSC::Heap::freeBlock): 12673 (JSC::Heap::freeBlockPtr): 12674 (JSC::Heap::freeBlocks): 12675 (JSC::Heap::recordExtraCost): 12676 (JSC::Heap::allocate): 12677 (JSC::Heap::resizeBlocks): 12678 (JSC::Heap::growBlocks): 12679 (JSC::Heap::shrinkBlocks): 12680 (JSC::Heap::markConservatively): 12681 (JSC::Heap::clearMarkBits): 12682 (JSC::Heap::markedCells): 12683 (JSC::Heap::sweep): 12684 (JSC::Heap::markRoots): 12685 (JSC::Heap::objectCount): 12686 (JSC::Heap::addToStatistics): 12687 (JSC::Heap::statistics): 12688 (JSC::Heap::isBusy): 12689 (JSC::Heap::reset): 12690 (JSC::Heap::collectAllGarbage): 12691 (JSC::Heap::primaryHeapBegin): 12692 (JSC::Heap::primaryHeapEnd): 12693 * runtime/Collector.h: 12694 (JSC::): Removed all code pertaining to the number heap, and changed all 12695 heap template functions and classes to non-template functions and classes. 12696 12697 (JSC::Heap::allocateNumber): A new optimization to replace the number 12698 heap: allocate half-sized number cells in pairs, returning the first 12699 cell and caching the second cell for the next allocation. 12700 12701 * runtime/CollectorHeapIterator.h: 12702 (JSC::LiveObjectIterator::LiveObjectIterator): 12703 (JSC::LiveObjectIterator::operator++): 12704 (JSC::DeadObjectIterator::DeadObjectIterator): 12705 (JSC::DeadObjectIterator::operator++): 12706 (JSC::ObjectIterator::ObjectIterator): 12707 (JSC::ObjectIterator::operator++): 12708 * runtime/JSCell.h: 12709 (JSC::JSCell::isNumber): Removed all code pertaining to the number heap, 12710 and changed all heap template functions and classes to non-template functions 12711 and classes. 12712 127132009-12-15 Zoltan Horvath <zoltan@webkit.org> 12714 12715 Reviewed by Darin Adler. 12716 12717 Allow custom memory allocation control for WeakGCMap class 12718 https://bugs.webkit.org/show_bug.cgi?id=32547 12719 12720 Inherits WeakGCMap from FastAllocBase because it is instantiated by 12721 'new' at: WebCore/dom/Document.cpp:512. 12722 12723 * runtime/WeakGCMap.h: 12724 127252009-12-15 Zoltan Horvath <zoltan@webkit.org> 12726 12727 Reviewed by Darin Adler. 12728 12729 Allow custom memory allocation control for dtoa's P5Node struct 12730 https://bugs.webkit.org/show_bug.cgi?id=32544 12731 12732 Inherits P5Node struct from Noncopyable because it is instantiated by 12733 'new' at wtf/dtoa.cpp:588 and don't need to be copyable. 12734 12735 * wtf/dtoa.cpp: 12736 127372009-12-14 Geoffrey Garen <ggaren@apple.com> 12738 12739 Reviewed by Simon Fraser. 12740 12741 https://bugs.webkit.org/show_bug.cgi?id=32524 12742 REGRESSION(52084): fast/dom/prototypes.html failing two CSS tests 12743 12744 * wtf/StdLibExtras.h: 12745 (WTF::bitCount): The original patch put the parentheses in the wrong 12746 place, completely changing the calculation and making it almost always 12747 wrong. Moved the parentheses around the '+' operation, like the original 12748 compiler warning suggested. 12749 127502009-12-14 Gabor Loki <loki@inf.u-szeged.hu> 12751 12752 Unreviewed trivial buildfix. 12753 12754 Fix crosses initialization of usedPrimaryBlocks for JSValue32 12755 12756 * runtime/Collector.cpp: 12757 (JSC::Heap::markConservatively): 12758 127592009-12-14 Csaba Osztrogonác <ossy@webkit.org> 12760 12761 Reviewed by Simon Hausmann. 12762 12763 GCC 4.3.x warning fixed. Suggested parantheses added. 12764 warning: ../../../JavaScriptCore/wtf/StdLibExtras.h:77: warning: suggest parentheses around + or - in operand of & 12765 12766 * wtf/StdLibExtras.h: 12767 (WTF::bitCount): 12768 127692009-12-13 Geoffrey Garen <ggaren@apple.com> 12770 12771 Reviewed by Sam Weinig. 12772 12773 Changed GC from mark-sweep to mark-allocate. 12774 12775 Added WeakGCMap to keep WebCore blissfully ignorant about objects that 12776 have become garbage but haven't run their destructors yet. 12777 12778 1% SunSpider speedup. 12779 7.6% v8 speedup (37% splay speedup). 12780 17% speedup on bench-alloc-nonretained.js. 12781 18% speedup on bench-alloc-retained.js. 12782 12783 * API/JSBase.cpp: 12784 (JSGarbageCollect): 12785 * API/JSContextRef.cpp: 12786 * JavaScriptCore.exp: 12787 * JavaScriptCore.xcodeproj/project.pbxproj: Updated for renames and new 12788 files. 12789 12790 * debugger/Debugger.cpp: 12791 (JSC::Debugger::recompileAllJSFunctions): Updated to use the Collector 12792 iterator abstraction. 12793 12794 * jsc.cpp: 12795 (functionGC): Updated for rename. 12796 12797 * runtime/Collector.cpp: Slightly reduced the number of allocations per 12798 collection, so that small workloads only allocate on collector block, 12799 rather than two. 12800 12801 (JSC::Heap::Heap): Updated to use the new allocateBlock function. 12802 12803 (JSC::Heap::destroy): Updated to use the new freeBlocks function. 12804 12805 (JSC::Heap::allocateBlock): New function to initialize a block when 12806 allocating it. 12807 12808 (JSC::Heap::freeBlock): Consolidated the responsibility for running 12809 destructors into this function. 12810 12811 (JSC::Heap::freeBlocks): Updated to use freeBlock. 12812 12813 (JSC::Heap::recordExtraCost): Sweep the heap in this reporting function, 12814 so that allocation, which is more common, doesn't have to check extraCost. 12815 12816 (JSC::Heap::heapAllocate): Run destructors right before recycling a 12817 garbage cell. This has better cache utilization than a separate sweep phase. 12818 12819 (JSC::Heap::resizeBlocks): 12820 (JSC::Heap::growBlocks): 12821 (JSC::Heap::shrinkBlocks): New set of functions for managing the size of 12822 the heap, now that the heap doesn't maintain any information about its 12823 size. 12824 12825 (JSC::isPointerAligned): 12826 (JSC::isHalfCellAligned): 12827 (JSC::isPossibleCell): 12828 (JSC::isCellAligned): 12829 (JSC::Heap::markConservatively): Cleaned up this code a bit. 12830 12831 (JSC::Heap::clearMarkBits): 12832 (JSC::Heap::markedCells): Some helper functions for examining the the mark 12833 bitmap. 12834 12835 (JSC::Heap::sweep): Simplified this function by using a DeadObjectIterator. 12836 12837 (JSC::Heap::markRoots): Reordered some operations for clarity. 12838 12839 (JSC::Heap::objectCount): 12840 (JSC::Heap::addToStatistics): 12841 (JSC::Heap::statistics): Rewrote these functions to calculate an object 12842 count on demand, since the heap doesn't maintain this information by 12843 itself. 12844 12845 (JSC::Heap::reset): New function for resetting the heap once we've 12846 exhausted heap space. 12847 12848 (JSC::Heap::collectAllGarbage): This function matches the old collect() 12849 behavior, but it's now an uncommon function used only by API. 12850 12851 * runtime/Collector.h: 12852 (JSC::CollectorBitmap::count): 12853 (JSC::CollectorBitmap::isEmpty): Added some helper functions for managing 12854 the collector mark bitmap. 12855 12856 (JSC::Heap::reportExtraMemoryCost): Changed reporting from cell equivalents 12857 to bytes, so it's easier to understand. 12858 12859 * runtime/CollectorHeapIterator.h: 12860 (JSC::CollectorHeapIterator::CollectorHeapIterator): 12861 (JSC::CollectorHeapIterator::operator!=): 12862 (JSC::CollectorHeapIterator::operator*): 12863 (JSC::CollectorHeapIterator::advance): 12864 (JSC::::LiveObjectIterator): 12865 (JSC::::operator): 12866 (JSC::::DeadObjectIterator): 12867 (JSC::::ObjectIterator): New iterators for encapsulating details about 12868 heap layout, and what's live and dead on the heap. 12869 12870 * runtime/JSArray.cpp: 12871 (JSC::JSArray::putSlowCase): 12872 (JSC::JSArray::increaseVectorLength): Delay reporting extra cost until 12873 we're fully constructed, so the heap mark phase won't visit us in an 12874 invalid state. 12875 12876 * runtime/JSCell.h: 12877 (JSC::JSCell::): 12878 (JSC::JSCell::createDummyStructure): 12879 (JSC::JSCell::JSCell): 12880 * runtime/JSGlobalData.cpp: 12881 (JSC::JSGlobalData::JSGlobalData): 12882 * runtime/JSGlobalData.h: Added a dummy cell to simplify allocation logic. 12883 12884 * runtime/JSString.h: 12885 (JSC::jsSubstring): Don't report extra cost for substrings, since they 12886 share a buffer that's already reported extra cost. 12887 12888 * runtime/Tracing.d: 12889 * runtime/Tracing.h: Changed these dtrace hooks not to report object 12890 counts, since they're no longer cheap to compute. 12891 12892 * runtime/UString.h: Updated for renames. 12893 12894 * runtime/WeakGCMap.h: Added. 12895 (JSC::WeakGCMap::isEmpty): 12896 (JSC::WeakGCMap::uncheckedGet): 12897 (JSC::WeakGCMap::uncheckedBegin): 12898 (JSC::WeakGCMap::uncheckedEnd): 12899 (JSC::::get): 12900 (JSC::::take): 12901 (JSC::::set): 12902 (JSC::::uncheckedRemove): Mentioned above. 12903 12904 * wtf/StdLibExtras.h: 12905 (WTF::bitCount): Added a bit population count function, so the heap can 12906 count live objects to fulfill statistics questions. 12907 12908The very last cell in the block is not allocated -- should not be marked. 12909 129102009-12-13 Geoffrey Garen <ggaren@apple.com> 12911 12912 Windows build fix: Export some new symbols. 12913 12914 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 12915 129162009-12-13 Geoffrey Garen <ggaren@apple.com> 12917 12918 Windows build fix: Removed some old exports. 12919 12920 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 12921 129222009-12-13 Geoffrey Garen <ggaren@apple.com> 12923 12924 Windows build fix: Use unsigned instead of uint32_t to avoid dependencies. 12925 12926 * wtf/StdLibExtras.h: 12927 (WTF::bitCount): 12928 129292009-12-13 Gavin Barraclough <barraclough@apple.com> 12930 12931 Reviewed by NOBODY (speculative Windows build fix). 12932 12933 * runtime/JSGlobalObjectFunctions.cpp: 12934 129352009-12-13 Gavin Barraclough <barraclough@apple.com> 12936 12937 Reviewed by Sam Weinig. 12938 12939 https://bugs.webkit.org/show_bug.cgi?id=32496 12940 Switch remaining cases of string construction to use StringBuilder. 12941 Builds strings using a vector rather than using string append / addition. 12942 12943 * JavaScriptCore.exp: 12944 * JavaScriptCore.xcodeproj/project.pbxproj: 12945 * runtime/Executable.cpp: 12946 (JSC::FunctionExecutable::paramString): 12947 * runtime/FunctionConstructor.cpp: 12948 (JSC::constructFunction): 12949 * runtime/JSGlobalObjectFunctions.cpp: 12950 (JSC::encode): 12951 (JSC::decode): 12952 (JSC::globalFuncEscape): 12953 (JSC::globalFuncUnescape): 12954 * runtime/JSONObject.cpp: 12955 (JSC::Stringifier::stringify): 12956 (JSC::Stringifier::indent): 12957 * runtime/JSString.h: 12958 * runtime/LiteralParser.cpp: 12959 (JSC::LiteralParser::Lexer::lexString): 12960 * runtime/NumberPrototype.cpp: 12961 (JSC::integerPartNoExp): 12962 (JSC::numberProtoFuncToFixed): 12963 (JSC::numberProtoFuncToPrecision): 12964 * runtime/Operations.h: 12965 (JSC::jsString): 12966 * runtime/StringPrototype.cpp: 12967 (JSC::substituteBackreferencesSlow): 12968 (JSC::substituteBackreferences): 12969 (JSC::stringProtoFuncConcat): 12970 129712009-12-08 Jeremy Moskovich <jeremy@chromium.org> 12972 12973 Reviewed by Eric Seidel. 12974 12975 Add code to allow toggling ATSUI/Core Text rendering at runtime in ComplexTextController. 12976 https://bugs.webkit.org/show_bug.cgi?id=31802 12977 12978 The goal here is to allow for a zero runtime hit for ports that decide to select 12979 the API at compile time. 12980 When both USE(ATSUI) and USE(CORE_TEXT) are true, the API is toggled 12981 at runtime. Core Text is used for OS Versions >= 10.6. 12982 12983 * wtf/Platform.h: #define USE_CORE_TEXT and USE_ATSUI on Chrome/Mac. 12984 129852009-12-11 Maciej Stachowiak <mjs@apple.com> 12986 12987 Reviewed by Oliver Hunt. 12988 12989 Unify codegen for forward and backward variants of branches 12990 https://bugs.webkit.org/show_bug.cgi?id=32463 12991 12992 * jit/JIT.h: 12993 (JSC::JIT::emit_op_loop): Implemented in terms of forward variant. 12994 (JSC::JIT::emit_op_loop_if_true): ditto 12995 (JSC::JIT::emitSlow_op_loop_if_true): ditto 12996 (JSC::JIT::emit_op_loop_if_false): ditto 12997 (JSC::JIT::emitSlow_op_loop_if_false): ditto 12998 (JSC::JIT::emit_op_loop_if_less): ditto 12999 (JSC::JIT::emitSlow_op_loop_if_less): ditto 13000 * jit/JITOpcodes.cpp: 13001 130022009-12-11 Sam Weinig <sam@webkit.org> 13003 13004 Reviewed by Anders Carlsson. 13005 13006 Allow WTFs concept of the main thread to differ from pthreads when necessary. 13007 13008 * wtf/ThreadingPthreads.cpp: 13009 (WTF::initializeThreading): 13010 (WTF::isMainThread): 13011 * wtf/mac/MainThreadMac.mm: 13012 (WTF::initializeMainThreadPlatform): 13013 (WTF::scheduleDispatchFunctionsOnMainThread): 13014 130152009-12-11 Gavin Barraclough <barraclough@apple.com> 13016 13017 Reviewed by Oliver Hunt. 13018 13019 https://bugs.webkit.org/show_bug.cgi?id=32454 13020 Refactor construction of simple strings to avoid string concatenation. 13021 13022 Building strings through concatenation has a memory and performance cost - 13023 a memory cost since we must over-allocate the buffer to leave space to append 13024 into, and performance in that the string may still require reallocation (and 13025 thus copying during construction). Instead move the full construction to 13026 within a single function call (makeString), so that the arguments' lengths 13027 can be calculated and an appropriate sized buffer allocated before copying 13028 any characters. 13029 13030 ~No performance change (~2% progression on date tests). 13031 13032 * bytecode/CodeBlock.cpp: 13033 (JSC::escapeQuotes): 13034 (JSC::valueToSourceString): 13035 (JSC::constantName): 13036 (JSC::idName): 13037 (JSC::CodeBlock::registerName): 13038 (JSC::regexpToSourceString): 13039 (JSC::regexpName): 13040 * bytecompiler/NodesCodegen.cpp: 13041 (JSC::substitute): 13042 * profiler/Profiler.cpp: 13043 (JSC::Profiler::createCallIdentifier): 13044 * runtime/DateConstructor.cpp: 13045 (JSC::callDate): 13046 * runtime/DateConversion.cpp: 13047 (JSC::formatDate): 13048 (JSC::formatDateUTCVariant): 13049 (JSC::formatTime): 13050 (JSC::formatTimeUTC): 13051 * runtime/DateConversion.h: 13052 (JSC::): 13053 * runtime/DatePrototype.cpp: 13054 (JSC::dateProtoFuncToString): 13055 (JSC::dateProtoFuncToUTCString): 13056 (JSC::dateProtoFuncToDateString): 13057 (JSC::dateProtoFuncToTimeString): 13058 (JSC::dateProtoFuncToGMTString): 13059 * runtime/ErrorPrototype.cpp: 13060 (JSC::errorProtoFuncToString): 13061 * runtime/ExceptionHelpers.cpp: 13062 (JSC::createUndefinedVariableError): 13063 (JSC::createErrorMessage): 13064 (JSC::createInvalidParamError): 13065 * runtime/FunctionPrototype.cpp: 13066 (JSC::insertSemicolonIfNeeded): 13067 (JSC::functionProtoFuncToString): 13068 * runtime/ObjectPrototype.cpp: 13069 (JSC::objectProtoFuncToString): 13070 * runtime/RegExpConstructor.cpp: 13071 (JSC::constructRegExp): 13072 * runtime/RegExpObject.cpp: 13073 (JSC::RegExpObject::match): 13074 * runtime/RegExpPrototype.cpp: 13075 (JSC::regExpProtoFuncCompile): 13076 (JSC::regExpProtoFuncToString): 13077 * runtime/StringPrototype.cpp: 13078 (JSC::stringProtoFuncBig): 13079 (JSC::stringProtoFuncSmall): 13080 (JSC::stringProtoFuncBlink): 13081 (JSC::stringProtoFuncBold): 13082 (JSC::stringProtoFuncFixed): 13083 (JSC::stringProtoFuncItalics): 13084 (JSC::stringProtoFuncStrike): 13085 (JSC::stringProtoFuncSub): 13086 (JSC::stringProtoFuncSup): 13087 (JSC::stringProtoFuncFontcolor): 13088 (JSC::stringProtoFuncFontsize): 13089 (JSC::stringProtoFuncAnchor): 13090 * runtime/UString.h: 13091 (JSC::): 13092 (JSC::makeString): 13093 130942009-12-10 Gavin Barraclough <barraclough@apple.com> 13095 13096 Reviewed by Oliver Hunt. 13097 13098 https://bugs.webkit.org/show_bug.cgi?id=32400 13099 Switch remaining cases of string addition to use ropes. 13100 13101 Re-landing r51975 - added toPrimitiveString method, 13102 performs toPrimitive then subsequent toString operations. 13103 13104 ~1% progression on Sunspidey. 13105 13106 * jit/JITStubs.cpp: 13107 (JSC::DEFINE_STUB_FUNCTION): 13108 * runtime/JSString.h: 13109 (JSC::JSString::JSString): 13110 (JSC::JSString::appendStringInConstruct): 13111 * runtime/Operations.cpp: 13112 (JSC::jsAddSlowCase): 13113 * runtime/Operations.h: 13114 (JSC::jsString): 13115 (JSC::jsAdd): 13116 131172009-12-11 Adam Roben <aroben@apple.com> 13118 13119 Windows build fix 13120 13121 * JavaScriptCore.vcproj/jsc/jscCommon.vsprops: Added 13122 $(WebKitOutputDir)/include/private to the include path. 13123 131242009-12-11 Adam Roben <aroben@apple.com> 13125 13126 Move QuartzCorePresent.h to include/private 13127 13128 This fixes other projects that use wtf/Platform.h 13129 13130 Rubber-stamped by Steve Falkenburg. 13131 13132 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: Let VS do its thang. 13133 * JavaScriptCore.vcproj/JavaScriptCore/build-generated-files.sh: Write 13134 QuartzCorePresent.h to $(WebKitOutputDir)/include/private. 13135 13136 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCoreCommon.vsprops: 13137 * JavaScriptCore.vcproj/WTF/WTFCommon.vsprops: 13138 Added $(WebKitOutputDir)/include/private to the include path. 13139 131402009-12-11 Adam Roben <aroben@apple.com> 13141 13142 Fix clean builds and everything rebuilding on every build 13143 13144 Reviewed by Sam Weinig. 13145 13146 * JavaScriptCore.vcproj/JavaScriptCore/build-generated-files.sh: Don't 13147 write out QuartzCorePresent.h if it exists but is older than 13148 QuartzCore.h. Also, create the directory we write QuartzCorePresent.h 13149 into first. 13150 131512009-12-11 Adam Roben <aroben@apple.com> 13152 13153 Windows build fix for systems with spaces in their paths 13154 13155 * JavaScriptCore.vcproj/JavaScriptCore/build-generated-files.sh: Quote some paths. 13156 131572009-12-11 Chris Marrin <cmarrin@apple.com> 13158 13159 Reviewed by Adam Roben. 13160 13161 Add check for presence of QuartzCore headers 13162 https://bugs.webkit.org/show_bug.cgi?id=31856 13163 13164 The script now checks for the presence of QuartzCore.h. If present 13165 it will turn on ACCELERATED_COMPOSITING and 3D_RENDERING to enable 13166 HW compositing on Windows. The script writes QuartzCorePresent.h to 13167 the build directory which has a define telling whether QuartzCore is 13168 present. 13169 13170 * JavaScriptCore.vcproj/JavaScriptCore/build-generated-files.sh: 13171 * wtf/Platform.h: 13172 131732009-12-11 Kent Tamura <tkent@chromium.org> 13174 13175 Reviewed by Darin Adler. 13176 13177 Fix a problem that JSC::gregorianDateTimeToMS() returns a negative 13178 value for a huge year value. 13179 https://bugs.webkit.org/show_bug.cgi?id=32304 13180 13181 * wtf/DateMath.cpp: 13182 (WTF::dateToDaysFrom1970): Renamed from dateToDayInYear, and changed the return type to double. 13183 (WTF::calculateDSTOffset): Follow the dateToDaysFrom1970() change. 13184 (WTF::timeClip): Use maxECMAScriptTime. 13185 (JSC::gregorianDateTimeToMS): Follow the dateToDaysFrom1970() change. 13186 131872009-12-10 Adam Barth <abarth@webkit.org> 13188 13189 No review, rolling out r51975. 13190 http://trac.webkit.org/changeset/51975 13191 13192 * jit/JITStubs.cpp: 13193 (JSC::DEFINE_STUB_FUNCTION): 13194 * runtime/JSString.h: 13195 (JSC::JSString::JSString): 13196 (JSC::JSString::appendStringInConstruct): 13197 * runtime/Operations.cpp: 13198 (JSC::jsAddSlowCase): 13199 * runtime/Operations.h: 13200 (JSC::jsString): 13201 (JSC::jsAdd): 13202 132032009-12-10 Oliver Hunt <oliver@apple.com> 13204 13205 Reviewed by Gavin Barraclough. 13206 13207 Incorrect caching of prototype lookup with dictionary base 13208 https://bugs.webkit.org/show_bug.cgi?id=32402 13209 13210 Make sure we don't add cached prototype lookup to the proto_list 13211 lookup chain if the top level object is a dictionary. 13212 13213 * jit/JITStubs.cpp: 13214 (JSC::JITThunks::tryCacheGetByID): 13215 132162009-12-10 Gavin Barraclough <barraclough@apple.com> 13217 13218 Reviewed by Oliver Hunt. 13219 13220 https://bugs.webkit.org/show_bug.cgi?id=32400 13221 Switch remaining cases of string addition to use ropes. 13222 13223 ~1% progression on Sunspidey. 13224 13225 * jit/JITStubs.cpp: 13226 (JSC::DEFINE_STUB_FUNCTION): 13227 * runtime/JSString.h: 13228 (JSC::JSString::JSString): 13229 (JSC::JSString::appendStringInConstruct): 13230 * runtime/Operations.cpp: 13231 (JSC::jsAddSlowCase): 13232 * runtime/Operations.h: 13233 (JSC::jsString): 13234 (JSC::jsAdd): 13235 132362009-12-10 Kent Hansen <kent.hansen@nokia.com> 13237 13238 Reviewed by Geoffrey Garen. 13239 13240 Remove JSObject::getPropertyAttributes() and all usage of it. 13241 https://bugs.webkit.org/show_bug.cgi?id=31933 13242 13243 getOwnPropertyDescriptor() should be used instead. 13244 13245 * JavaScriptCore.exp: 13246 * JavaScriptCore.order: 13247 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 13248 * debugger/DebuggerActivation.cpp: 13249 (JSC::DebuggerActivation::getOwnPropertyDescriptor): 13250 * debugger/DebuggerActivation.h: 13251 * runtime/JSObject.cpp: 13252 (JSC::JSObject::propertyIsEnumerable): 13253 * runtime/JSObject.h: 13254 * runtime/JSVariableObject.cpp: 13255 * runtime/JSVariableObject.h: 13256 132572009-12-10 Gavin Barraclough <barraclough@apple.com> 13258 13259 Reviewed by Oliver Hunt & Mark Rowe. 13260 13261 https://bugs.webkit.org/show_bug.cgi?id=32367 13262 Add support for short Ropes (up to 3 entries) inline within JSString. 13263 (rather than externally allocating an object to hold the rope). 13264 Switch jsAdd of (JSString* + JSString*) to now make use of Ropes. 13265 13266 ~1% progression on Sunspidey. 13267 13268 * interpreter/Interpreter.cpp: 13269 (JSC::Interpreter::privateExecute): 13270 * jit/JITOpcodes.cpp: 13271 (JSC::JIT::privateCompileCTIMachineTrampolines): 13272 * jit/JITStubs.cpp: 13273 (JSC::DEFINE_STUB_FUNCTION): 13274 * runtime/JSString.cpp: 13275 (JSC::JSString::resolveRope): 13276 (JSC::JSString::toBoolean): 13277 (JSC::JSString::getStringPropertyDescriptor): 13278 * runtime/JSString.h: 13279 (JSC::JSString::Rope::Fiber::deref): 13280 (JSC::JSString::Rope::Fiber::ref): 13281 (JSC::JSString::Rope::Fiber::refAndGetLength): 13282 (JSC::JSString::Rope::append): 13283 (JSC::JSString::JSString): 13284 (JSC::JSString::~JSString): 13285 (JSC::JSString::value): 13286 (JSC::JSString::tryGetValue): 13287 (JSC::JSString::length): 13288 (JSC::JSString::canGetIndex): 13289 (JSC::JSString::appendStringInConstruct): 13290 (JSC::JSString::appendValueInConstructAndIncrementLength): 13291 (JSC::JSString::isRope): 13292 (JSC::JSString::string): 13293 (JSC::JSString::ropeLength): 13294 (JSC::JSString::getStringPropertySlot): 13295 * runtime/Operations.h: 13296 (JSC::jsString): 13297 (JSC::jsAdd): 13298 (JSC::resolveBase): 13299 133002009-12-09 Anders Carlsson <andersca@apple.com> 13301 13302 Reviewed by Geoffrey Garen. 13303 13304 Fix three more things found by compiling with clang++. 13305 13306 * runtime/Structure.h: 13307 (JSC::StructureTransitionTable::reifySingleTransition): 13308 Add the 'std' qualifier to the call to make_pair. 13309 13310 * wtf/DateMath.cpp: 13311 (WTF::initializeDates): 13312 Incrementing a bool is deprecated according to the C++ specification. 13313 13314 * wtf/PtrAndFlags.h: 13315 (WTF::PtrAndFlags::PtrAndFlags): 13316 Name lookup should not be done in dependent bases, so explicitly qualify the call to set. 13317 133182009-12-09 Maciej Stachowiak <mjs@apple.com> 13319 13320 Reviewed by Oliver Hunt. 13321 13322 Google reader gets stuck in the "Loading..." state and does not complete 13323 https://bugs.webkit.org/show_bug.cgi?id=32256 13324 <rdar://problem/7456388> 13325 13326 * jit/JITArithmetic.cpp: 13327 (JSC::JIT::emitSlow_op_jless): Fix some backward branches. 13328 133292009-12-09 Gavin Barraclough <barraclough@apple.com> 13330 13331 Reviewed by Oliver Hunt. 13332 13333 https://bugs.webkit.org/show_bug.cgi?id=32228 13334 Make destruction of ropes non-recursive to prevent stack exhaustion. 13335 Also, pass a UString& into initializeFiber rather than a Ustring::Rep*, 13336 since the Rep is not being ref counted this could result in usage of a 13337 Rep with refcount zero (where the Rep comes from a temporary UString 13338 returned from a function). 13339 13340 * runtime/JSString.cpp: 13341 (JSC::JSString::Rope::destructNonRecursive): 13342 (JSC::JSString::Rope::~Rope): 13343 * runtime/JSString.h: 13344 (JSC::JSString::Rope::initializeFiber): 13345 * runtime/Operations.h: 13346 (JSC::concatenateStrings): 13347 133482009-12-09 Zoltan Herczeg <zherczeg@inf.u-szeged.hu> 13349 13350 Reviewed by Eric Seidel. 13351 13352 https://bugs.webkit.org/show_bug.cgi?id=31930 13353 13354 Update to r51457. ASSERTs changed to COMPILE_ASSERTs. 13355 The speedup is 25%. 13356 13357 * runtime/JSGlobalData.cpp: 13358 (JSC::VPtrSet::VPtrSet): 13359 133602009-12-09 Steve Block <steveblock@google.com> 13361 13362 Reviewed by Adam Barth. 13363 13364 Updates Android Makefiles with latest additions. 13365 https://bugs.webkit.org/show_bug.cgi?id=32278 13366 13367 * Android.mk: Modified. 13368 * Android.v8.wtf.mk: Modified. 13369 133702009-12-09 Sam Weinig <sam@webkit.org> 13371 13372 Reviewed by Gavin Barraclough. 13373 13374 Fix a bug found while trying to compile JavaScriptCore with clang++. 13375 13376 * yarr/RegexPattern.h: 13377 (JSC::Yarr::PatternTerm::PatternTerm): Don't self assign here. Use false instead. 13378 133792009-12-09 Anders Carlsson <andersca@apple.com> 13380 13381 Reviewed by Sam Weinig. 13382 13383 Attempt to fix the Windows build. 13384 13385 * wtf/FastMalloc.h: 13386 133872009-12-09 Anders Carlsson <andersca@apple.com> 13388 13389 Reviewed by Sam Weinig. 13390 13391 Fix some things found while trying to compile JavaScriptCore with clang++. 13392 13393 * wtf/FastMalloc.h: 13394 Add correct exception specifications for the allocation/deallocation operators. 13395 13396 * wtf/Vector.h: 13397 * wtf/VectorTraits.h: 13398 Fix a bunch of struct/class mismatches. 13399 134002009-12-08 Maciej Stachowiak <mjs@apple.com> 13401 13402 Reviewed by Darin Adler. 13403 13404 move code generation portions of Nodes.cpp to bytecompiler directory 13405 https://bugs.webkit.org/show_bug.cgi?id=32284 13406 13407 * bytecompiler/NodesCodegen.cpp: Copied from parser/Nodes.cpp. Removed parts that 13408 are not about codegen. 13409 * parser/Nodes.cpp: Removed everything that is about codegen. 13410 13411 Update build systems: 13412 13413 * Android.mk: 13414 * GNUmakefile.am: 13415 * JavaScriptCore.gypi: 13416 * JavaScriptCore.pri: 13417 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: 13418 * JavaScriptCore.xcodeproj/project.pbxproj: 13419 * JavaScriptCoreSources.bkl: 13420 134212009-12-08 Kevin Watters <kevinwatters@gmail.com> 13422 13423 Reviewed by Kevin Ollivier. 13424 13425 [wx] Mac plugins support. 13426 13427 https://bugs.webkit.org/show_bug.cgi?id=32236 13428 13429 * wtf/Platform.h: 13430 134312009-12-08 Dmitry Titov <dimich@chromium.org> 13432 13433 Rubber-stamped by David Levin. 13434 13435 Revert and reopen "Add asserts to RefCounted to make sure ref/deref happens on the right thread." 13436 It may have caused massive increase of reported leaks on the bots. 13437 https://bugs.webkit.org/show_bug.cgi?id=31639 13438 13439 * GNUmakefile.am: 13440 * JavaScriptCore.gypi: 13441 * JavaScriptCore.vcproj/WTF/WTF.vcproj: 13442 * JavaScriptCore.xcodeproj/project.pbxproj: 13443 * runtime/Structure.cpp: 13444 (JSC::Structure::Structure): 13445 * wtf/RefCounted.h: 13446 (WTF::RefCountedBase::ref): 13447 (WTF::RefCountedBase::hasOneRef): 13448 (WTF::RefCountedBase::refCount): 13449 (WTF::RefCountedBase::derefBase): 13450 * wtf/ThreadVerifier.h: Removed. 13451 134522009-12-08 Gustavo Noronha Silva <gustavo.noronha@collabora.co.uk> 13453 13454 Reviewed by Darin Adler. 13455 13456 Make WebKit build correctly on FreeBSD, IA64, and Alpha. 13457 Based on work by Petr Salinger <Petr.Salinger@seznam.cz>, 13458 and Colin Watson <cjwatson@ubuntu.com>. 13459 13460 * wtf/Platform.h: 13461 134622009-12-08 Dmitry Titov <dimich@chromium.org> 13463 13464 Reviewed by Darin Adler. 13465 13466 Add asserts to RefCounted to make sure ref/deref happens on the right thread. 13467 https://bugs.webkit.org/show_bug.cgi?id=31639 13468 13469 * runtime/Structure.cpp: 13470 (JSC::Structure::Structure): Disable thread verification on this class since it uses addressOfCount(). 13471 * wtf/RefCounted.h: 13472 (WTF::RefCountedBase::ref): Add ASSERT. 13473 (WTF::RefCountedBase::hasOneRef): Ditto. 13474 (WTF::RefCountedBase::refCount): Ditto. 13475 (WTF::RefCountedBase::derefBase): Ditto. 13476 (WTF::RefCountedBase::disableThreadVerification): delegate to ThreadVerifier method. 13477 * wtf/ThreadVerifier.h: Added. 13478 (WTF::ThreadVerifier::ThreadVerifier): New Debug-only class to verify that ref/deref of RefCounted is done on the same thread. 13479 (WTF::ThreadVerifier::activate): Activates checks. Called when ref count becomes above 2. 13480 (WTF::ThreadVerifier::deactivate): Deactivates checks. Called when ref count drops below 2. 13481 (WTF::ThreadVerifier::disableThreadVerification): used on objects that should not be checked (StringImpl etc) 13482 (WTF::ThreadVerifier::verifyThread): 13483 * GNUmakefile.am: Add ThreadVerifier.h to the build file. 13484 * JavaScriptCore.gypi: Ditto. 13485 * JavaScriptCore.vcproj/WTF/WTF.vcproj: Ditto. 13486 * JavaScriptCore.xcodeproj/project.pbxproj: Ditto. 13487 134882009-12-08 Steve Block <steveblock@google.com> 13489 13490 Reviewed by Adam Barth. 13491 13492 [Android] Adds Makefiles for Android port. 13493 https://bugs.webkit.org/show_bug.cgi?id=31325 13494 13495 * Android.mk: Added. 13496 * Android.v8.wtf.mk: Added. 13497 134982009-12-07 Dmitry Titov <dimich@chromium.org> 13499 13500 Rubber-stamped by Darin Adler. 13501 13502 Remove ENABLE_SHARED_SCRIPT flags 13503 https://bugs.webkit.org/show_bug.cgi?id=32245 13504 This patch was obtained by "git revert" command and then un-reverting of ChangeLog files. 13505 13506 * Configurations/FeatureDefines.xcconfig: 13507 * wtf/Platform.h: 13508 135092009-12-07 Gavin Barraclough <barraclough@apple.com> 13510 13511 Reviewed by NOBODY (Windows build fixage part I). 13512 13513 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 13514 135152009-12-05 Gavin Barraclough <barraclough@apple.com> 13516 13517 Reviewed by Oliver Hunt. 13518 13519 https://bugs.webkit.org/show_bug.cgi?id=32184 13520 Handle out-of-memory conditions with JSC Ropes with a JS exception, rather than crashing. 13521 Switch from using fastMalloc to tryFastMalloc, pass an ExecState to record the exception on. 13522 13523 * API/JSCallbackObjectFunctions.h: 13524 (JSC::::toString): 13525 * API/JSValueRef.cpp: 13526 (JSValueIsStrictEqual): 13527 * JavaScriptCore.exp: 13528 * bytecompiler/BytecodeGenerator.cpp: 13529 (JSC::BytecodeGenerator::emitEqualityOp): 13530 * debugger/DebuggerCallFrame.cpp: 13531 (JSC::DebuggerCallFrame::functionName): 13532 (JSC::DebuggerCallFrame::calculatedFunctionName): 13533 * interpreter/Interpreter.cpp: 13534 (JSC::Interpreter::callEval): 13535 (JSC::Interpreter::privateExecute): 13536 * jit/JITStubs.cpp: 13537 (JSC::DEFINE_STUB_FUNCTION): 13538 * profiler/ProfileGenerator.cpp: 13539 (JSC::ProfileGenerator::addParentForConsoleStart): 13540 * profiler/Profiler.cpp: 13541 (JSC::Profiler::willExecute): 13542 (JSC::Profiler::didExecute): 13543 (JSC::Profiler::createCallIdentifier): 13544 (JSC::createCallIdentifierFromFunctionImp): 13545 * profiler/Profiler.h: 13546 * runtime/ArrayPrototype.cpp: 13547 (JSC::arrayProtoFuncIndexOf): 13548 (JSC::arrayProtoFuncLastIndexOf): 13549 * runtime/DateConstructor.cpp: 13550 (JSC::constructDate): 13551 * runtime/FunctionPrototype.cpp: 13552 (JSC::functionProtoFuncToString): 13553 * runtime/InternalFunction.cpp: 13554 (JSC::InternalFunction::name): 13555 (JSC::InternalFunction::displayName): 13556 (JSC::InternalFunction::calculatedDisplayName): 13557 * runtime/InternalFunction.h: 13558 * runtime/JSCell.cpp: 13559 (JSC::JSCell::getString): 13560 * runtime/JSCell.h: 13561 (JSC::JSValue::getString): 13562 * runtime/JSONObject.cpp: 13563 (JSC::gap): 13564 (JSC::Stringifier::Stringifier): 13565 (JSC::Stringifier::appendStringifiedValue): 13566 * runtime/JSObject.cpp: 13567 (JSC::JSObject::putDirectFunction): 13568 (JSC::JSObject::putDirectFunctionWithoutTransition): 13569 (JSC::JSObject::defineOwnProperty): 13570 * runtime/JSObject.h: 13571 * runtime/JSPropertyNameIterator.cpp: 13572 (JSC::JSPropertyNameIterator::get): 13573 * runtime/JSString.cpp: 13574 (JSC::JSString::Rope::~Rope): 13575 (JSC::JSString::resolveRope): 13576 (JSC::JSString::getPrimitiveNumber): 13577 (JSC::JSString::toNumber): 13578 (JSC::JSString::toString): 13579 (JSC::JSString::toThisString): 13580 (JSC::JSString::getStringPropertyDescriptor): 13581 * runtime/JSString.h: 13582 (JSC::JSString::Rope::createOrNull): 13583 (JSC::JSString::Rope::operator new): 13584 (JSC::JSString::value): 13585 (JSC::JSString::tryGetValue): 13586 (JSC::JSString::getIndex): 13587 (JSC::JSString::getStringPropertySlot): 13588 (JSC::JSValue::toString): 13589 * runtime/JSValue.h: 13590 * runtime/NativeErrorConstructor.cpp: 13591 (JSC::NativeErrorConstructor::NativeErrorConstructor): 13592 * runtime/Operations.cpp: 13593 (JSC::JSValue::strictEqualSlowCase): 13594 * runtime/Operations.h: 13595 (JSC::JSValue::equalSlowCaseInline): 13596 (JSC::JSValue::strictEqualSlowCaseInline): 13597 (JSC::JSValue::strictEqual): 13598 (JSC::jsLess): 13599 (JSC::jsLessEq): 13600 (JSC::jsAdd): 13601 (JSC::concatenateStrings): 13602 * runtime/PropertyDescriptor.cpp: 13603 (JSC::PropertyDescriptor::equalTo): 13604 * runtime/PropertyDescriptor.h: 13605 * runtime/StringPrototype.cpp: 13606 (JSC::stringProtoFuncReplace): 13607 (JSC::stringProtoFuncToLowerCase): 13608 (JSC::stringProtoFuncToUpperCase): 13609 136102009-12-07 Nikolas Zimmermann <nzimmermann@rim.com> 13611 13612 Reviewed by Holger Freyther. 13613 13614 Turn on (SVG) Filters support, by default. 13615 https://bugs.webkit.org/show_bug.cgi?id=32224 13616 13617 * Configurations/FeatureDefines.xcconfig: Enable FILTERS build flag. 13618 136192009-12-07 Steve Falkenburg <sfalken@apple.com> 13620 13621 Build fix. Be flexible about which version of ICU is used on Windows. 13622 13623 * JavaScriptCore.vcproj/jsc/jscCommon.vsprops: Add optional xcopy commands to copy ICU 4.2. 13624 136252009-12-07 Maciej Stachowiak <mjs@apple.com> 13626 13627 Reviewed by Oliver Hunt. 13628 13629 op_loop_if_less JIT codegen is broken for 64-bit 13630 https://bugs.webkit.org/show_bug.cgi?id=32221 13631 13632 * jit/JITOpcodes.cpp: 13633 (JSC::JIT::emit_op_loop_if_false): Fix codegen in this version - test was backwards. 13634 136352009-12-07 Oliver Hunt <oliver@apple.com> 13636 13637 Reviewed by Maciej Stachowiak. 13638 13639 Object.create fails if properties on the descriptor are getters 13640 https://bugs.webkit.org/show_bug.cgi?id=32219 13641 13642 Correctly initialise the PropertySlots with the descriptor object. 13643 13644 * runtime/ObjectConstructor.cpp: 13645 (JSC::toPropertyDescriptor): 13646 136472009-12-06 Maciej Stachowiak <mjs@apple.com> 13648 13649 Not reviewed, build fix. 13650 13651 Actually tested 64-bit *and* 32-bit build this time. 13652 13653 * jit/JITOpcodes.cpp: 13654 (JSC::JIT::emit_op_loop_if_false): 13655 136562009-12-06 Maciej Stachowiak <mjs@apple.com> 13657 13658 Not reviewed, build fix. 13659 13660 Really really fix 64-bit build for prior patch (actually tested this time). 13661 13662 * jit/JITOpcodes.cpp: 13663 (JSC::JIT::emit_op_loop_if_false): 13664 (JSC::JIT::emitSlow_op_loop_if_false): 13665 136662009-12-06 Maciej Stachowiak <mjs@apple.com> 13667 13668 Not reviewed, build fix. 13669 13670 Really fix 64-bit build for prior patch. 13671 13672 * jit/JITArithmetic.cpp: 13673 (JSC::JIT::emitSlow_op_jless): 13674 136752009-12-06 Maciej Stachowiak <mjs@apple.com> 13676 13677 Not reviewed, build fix. 13678 13679 Fix 64-bit build for prior patch. 13680 13681 * jit/JITOpcodes.cpp: 13682 (JSC::JIT::emitSlow_op_loop_if_less): 13683 136842009-12-05 Maciej Stachowiak <mjs@apple.com> 13685 13686 Reviewed by Oliver Hunt. 13687 13688 conway benchmark spends half it's time in op_less (jump fusion fails) 13689 https://bugs.webkit.org/show_bug.cgi?id=32190 13690 13691 <1% speedup on SunSpider and V8 13692 2x speedup on "conway" benchmark 13693 13694 Two optimizations: 13695 1) Improve codegen for logical operators &&, || and ! in a condition context 13696 13697 When generating code for combinations of &&, || and !, in a 13698 condition context (i.e. in an if statement or loop condition), we 13699 used to produce a value, and then separately jump based on its 13700 truthiness. Now we pass the false and true targets in, and let the 13701 logical operators generate jumps directly. This helps in four 13702 ways: 13703 13704 a) Individual clauses of a short-circuit logical operator can now 13705 jump directly to the then or else clause of an if statement (or to 13706 the top or exit of a loop) instead of jumping to a jump. 13707 13708 b) It used to be that jump fusion with the condition of the first 13709 clause of a logical operator was inhibited, because the register 13710 was ref'd to be used later, in the actual condition jump; this no 13711 longer happens since a jump straight to the final target is 13712 generated directly. 13713 13714 c) It used to be that jump fusion with the condition of the second 13715 clause of a logical operator was inhibited, because there was a 13716 jump target right after the second clause and before the actual 13717 condition jump. But now it's no longer necessary for the first 13718 clause to jump there so jump fusion is not blocked. 13719 13720 d) We avoid generating excess mov statements in some cases. 13721 13722 As a concrete example this source: 13723 13724 if (!((x < q && y < q) || (t < q && z < q))) { 13725 // ... 13726 } 13727 13728 Used to generate this bytecode: 13729 13730 [ 34] less r1, r-15, r-19 13731 [ 38] jfalse r1, 7(->45) 13732 [ 41] less r1, r-16, r-19 13733 [ 45] jtrue r1, 14(->59) 13734 [ 48] less r1, r-17, r-19 13735 [ 52] jfalse r1, 7(->59) 13736 [ 55] less r1, r-18, r-19 13737 [ 59] jtrue r1, 17(->76) 13738 13739 And now generates this bytecode (also taking advantage of the second optimization below): 13740 13741 [ 34] jnless r-15, r-19, 8(->42) 13742 [ 38] jless r-16, r-19, 26(->64) 13743 [ 42] jnless r-17, r-19, 8(->50) 13744 [ 46] jless r-18, r-19, 18(->64) 13745 13746 Note the jump fusion and the fact that there's less jump 13747 indirection - three of the four jumps go straight to the target 13748 clause instead of indirecting through another jump. 13749 13750 2) Implement jless opcode to take advantage of the above, since we'll now often generate 13751 a less followed by a jtrue where fusion is not forbidden. 13752 13753 * parser/Nodes.h: 13754 (JSC::ExpressionNode::hasConditionContextCodegen): Helper function to determine 13755 whether a node supports special conditional codegen. Return false as this is the default. 13756 (JSC::ExpressionNode::emitBytecodeInConditionContext): Assert not reached - only really 13757 defined for nodes that do have conditional codegen. 13758 (JSC::UnaryOpNode::expr): Add const version. 13759 (JSC::LogicalNotNode::hasConditionContextCodegen): Returne true only if subexpression 13760 supports it. 13761 (JSC::LogicalOpNode::hasConditionContextCodegen): Return true. 13762 * parser/Nodes.cpp: 13763 (JSC::LogicalNotNode::emitBytecodeInConditionContext): Implemented - just swap 13764 the true and false targets for the child node. 13765 (JSC::LogicalOpNode::emitBytecodeInConditionContext): Implemented - handle jumps 13766 directly, improving codegen quality. Also handles further nested conditional codegen. 13767 (JSC::ConditionalNode::emitBytecode): Use condition context codegen when available. 13768 (JSC::IfNode::emitBytecode): ditto 13769 (JSC::IfElseNode::emitBytecode): ditto 13770 (JSC::DoWhileNode::emitBytecode): ditto 13771 (JSC::WhileNode::emitBytecode): ditto 13772 (JSC::ForNode::emitBytecode): ditto 13773 13774 * bytecode/Opcode.h: 13775 - Added loop_if_false opcode - needed now that falsey jumps can be backwards. 13776 - Added jless opcode to take advantage of new fusion opportunities. 13777 * bytecode/CodeBlock.cpp: 13778 (JSC::CodeBlock::dump): Handle above. 13779 * bytecompiler/BytecodeGenerator.cpp: 13780 (JSC::BytecodeGenerator::emitJumpIfTrue): Add peephole for less + jtrue ==> jless. 13781 (JSC::BytecodeGenerator::emitJumpIfFalse): Add handling of backwrds falsey jumps. 13782 * bytecompiler/BytecodeGenerator.h: 13783 (JSC::BytecodeGenerator::emitNodeInConditionContext): Wrapper to handle tracking of 13784 overly deep expressions etc. 13785 * interpreter/Interpreter.cpp: 13786 (JSC::Interpreter::privateExecute): Implement the two new opcodes (loop_if_false, jless). 13787 * jit/JIT.cpp: 13788 (JSC::JIT::privateCompileMainPass): Implement JIT support for the two new opcodes. 13789 (JSC::JIT::privateCompileSlowCases): ditto 13790 * jit/JIT.h: 13791 * jit/JITArithmetic.cpp: 13792 (JSC::JIT::emit_op_jless): 13793 (JSC::JIT::emitSlow_op_jless): ditto 13794 (JSC::JIT::emitBinaryDoubleOp): ditto 13795 * jit/JITOpcodes.cpp: 13796 (JSC::JIT::emitSlow_op_loop_if_less): ditto 13797 (JSC::JIT::emit_op_loop_if_false): ditto 13798 (JSC::JIT::emitSlow_op_loop_if_false): ditto 13799 * jit/JITStubs.cpp: 13800 * jit/JITStubs.h: 13801 (JSC::): 13802 138032009-12-04 Kent Hansen <kent.hansen@nokia.com> 13804 13805 Reviewed by Darin Adler. 13806 13807 JavaScript delete operator should return false for string properties 13808 https://bugs.webkit.org/show_bug.cgi?id=32012 13809 13810 * runtime/StringObject.cpp: 13811 (JSC::StringObject::deleteProperty): 13812 138132009-12-03 Drew Wilson <atwilson@chromium.org> 13814 13815 Rolled back r51633 because it causes a perf regression in Chromium. 13816 13817 * wtf/Platform.h: 13818 138192009-12-03 Gavin Barraclough <barraclough@apple.com> 13820 13821 Try and fix the Windows build. 13822 13823 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: Export a symbol that should be exported. 13824 138252009-12-03 Mark Rowe <mrowe@apple.com> 13826 13827 Try and fix the Mac build. 13828 13829 * JavaScriptCore.exp: Export a symbol that should be exported. 13830 138312009-12-03 Oliver Hunt <oliver@apple.com> 13832 13833 Reviewed by Gavin Barraclough. 13834 13835 REGRESSION(4.0.3-48777): Crash in JSC::ExecState::propertyNames() (Debug-only?) 13836 https://bugs.webkit.org/show_bug.cgi?id=32133 13837 13838 Work around odd GCC-ism and correct the scopechain for use by 13839 calls made while a cachedcall is active on the callstack. 13840 13841 * interpreter/CachedCall.h: 13842 (JSC::CachedCall::newCallFrame): 13843 * runtime/JSArray.cpp: 13844 (JSC::AVLTreeAbstractorForArrayCompare::compare_key_key): 13845 * runtime/StringPrototype.cpp: 13846 (JSC::stringProtoFuncReplace): 13847 138482009-12-03 Gavin Barraclough <barraclough@apple.com> 13849 13850 Reviewed by Oliver "Brraaaaiiiinnnnnzzzzzzzz" Hunt. 13851 13852 https://bugs.webkit.org/show_bug.cgi?id=32136 13853 Add a rope representation to JSString. Presently JSString always holds its data in UString form. 13854 Instead, allow the result of a string concatenation to be represented in a tree form - with a 13855 variable sized, reference-counted rope node retaining a set of UString::Reps (or other rope nopes). 13856 13857 Strings must still currently be resolved down to a flat UString representation before being used, 13858 but by holding the string in a rope representation during construction we can avoid copying data 13859 until we know the final size of the string. 13860 13861 ~2% progression on SunSpider (~25% on date-format-xparb, ~20% on string-validate-input). 13862 13863 * JavaScriptCore.exp: 13864 13865 - Update exports. 13866 13867 * interpreter/Interpreter.cpp: 13868 (JSC::Interpreter::privateExecute): 13869 13870 - Make use of new JSString::length() method to avoid prematurely resolving ropes. 13871 13872 * jit/JITOpcodes.cpp: 13873 (JSC::JIT::privateCompileCTIMachineTrampolines): 13874 13875 - Switch the string length trampoline to read the length directly from JSString::m_length, 13876 rather than from the JSString's UString::Rep's 'len' property. 13877 13878 * jit/JITStubs.cpp: 13879 (JSC::DEFINE_STUB_FUNCTION): 13880 13881 - Modify op_add such that addition of two strings, where either or both strings are already 13882 in rope representation, produces a rope as a result. 13883 13884 * runtime/JSString.cpp: 13885 (JSC::JSString::Rope::~Rope): 13886 (JSC::copyChars): 13887 (JSC::JSString::resolveRope): 13888 (JSC::JSString::getPrimitiveNumber): 13889 (JSC::JSString::toBoolean): 13890 (JSC::JSString::toNumber): 13891 (JSC::JSString::toString): 13892 (JSC::JSString::toThisString): 13893 (JSC::JSString::getStringPropertyDescriptor): 13894 * runtime/JSString.h: 13895 (JSC::JSString::Rope::Fiber::Fiber): 13896 (JSC::JSString::Rope::Fiber::destroy): 13897 (JSC::JSString::Rope::Fiber::isRope): 13898 (JSC::JSString::Rope::Fiber::rope): 13899 (JSC::JSString::Rope::Fiber::string): 13900 (JSC::JSString::Rope::create): 13901 (JSC::JSString::Rope::initializeFiber): 13902 (JSC::JSString::Rope::ropeLength): 13903 (JSC::JSString::Rope::stringLength): 13904 (JSC::JSString::Rope::fibers): 13905 (JSC::JSString::Rope::Rope): 13906 (JSC::JSString::Rope::operator new): 13907 (JSC::JSString::JSString): 13908 (JSC::JSString::value): 13909 (JSC::JSString::length): 13910 (JSC::JSString::isRope): 13911 (JSC::JSString::rope): 13912 (JSC::JSString::string): 13913 (JSC::JSString::canGetIndex): 13914 (JSC::jsSingleCharacterSubstring): 13915 (JSC::JSString::getIndex): 13916 (JSC::jsSubstring): 13917 (JSC::JSString::getStringPropertySlot): 13918 13919 - Add rope form. 13920 13921 * runtime/Operations.h: 13922 (JSC::jsAdd): 13923 (JSC::concatenateStrings): 13924 13925 - Update string concatenation, and addition of ropes, to produce ropes. 13926 13927 * runtime/StringObject.cpp: 13928 (JSC::StringObject::getOwnPropertyNames): 13929 13930 - Make use of new JSString::length() method to avoid prematurely resolving ropes. 13931 139322009-11-23 Jeremy Moskovich <jeremy@chromium.org> 13933 13934 Reviewed by Eric Seidel. 13935 13936 Switch Chrome/Mac to use Core Text APIs rather than ATSUI APIs. 13937 https://bugs.webkit.org/show_bug.cgi?id=31802 13938 13939 No test since this is already covered by existing pixel tests. 13940 13941 * wtf/Platform.h: #define USE_CORE_TEXT for Chrome/Mac. 13942 139432009-12-02 Oliver Hunt <oliver@apple.com> 13944 13945 Reviewed by Gavin Barraclough. 13946 13947 Add files missed in prior patch. 13948 13949 * runtime/JSZombie.cpp: 13950 (JSC::): 13951 (JSC::JSZombie::leakedZombieStructure): 13952 * runtime/JSZombie.h: Added. 13953 (JSC::JSZombie::JSZombie): 13954 (JSC::JSZombie::isZombie): 13955 (JSC::JSZombie::classInfo): 13956 (JSC::JSZombie::isGetterSetter): 13957 (JSC::JSZombie::isAPIValueWrapper): 13958 (JSC::JSZombie::isPropertyNameIterator): 13959 (JSC::JSZombie::getCallData): 13960 (JSC::JSZombie::getConstructData): 13961 (JSC::JSZombie::getUInt32): 13962 (JSC::JSZombie::toPrimitive): 13963 (JSC::JSZombie::getPrimitiveNumber): 13964 (JSC::JSZombie::toBoolean): 13965 (JSC::JSZombie::toNumber): 13966 (JSC::JSZombie::toString): 13967 (JSC::JSZombie::toObject): 13968 (JSC::JSZombie::markChildren): 13969 (JSC::JSZombie::put): 13970 (JSC::JSZombie::deleteProperty): 13971 (JSC::JSZombie::toThisObject): 13972 (JSC::JSZombie::toThisString): 13973 (JSC::JSZombie::toThisJSString): 13974 (JSC::JSZombie::getJSNumber): 13975 (JSC::JSZombie::getOwnPropertySlot): 13976 139772009-12-02 Oliver Hunt <oliver@apple.com> 13978 13979 Reviewed by Gavin Barraclough. 13980 13981 Add zombies to JSC 13982 https://bugs.webkit.org/show_bug.cgi?id=32103 13983 13984 Add a compile time flag to make the JSC collector replace "unreachable" 13985 objects with zombie objects. The zombie object is a JSCell subclass that 13986 ASSERTs on any attempt to use the JSCell methods. In addition there are 13987 a number of additional assertions in bottleneck code to catch zombie usage 13988 as quickly as possible. 13989 13990 Grrr. Argh. Brains. 13991 13992 * JavaScriptCore.xcodeproj/project.pbxproj: 13993 * interpreter/Register.h: 13994 (JSC::Register::Register): 13995 * runtime/ArgList.h: 13996 (JSC::MarkedArgumentBuffer::append): 13997 (JSC::ArgList::ArgList): 13998 * runtime/Collector.cpp: 13999 (JSC::Heap::destroy): 14000 (JSC::Heap::sweep): 14001 * runtime/Collector.h: 14002 * runtime/JSCell.h: 14003 (JSC::JSCell::isZombie): 14004 (JSC::JSValue::isZombie): 14005 * runtime/JSValue.h: 14006 (JSC::JSValue::decode): 14007 (JSC::JSValue::JSValue): 14008 * wtf/Platform.h: 14009 140102009-12-01 Jens Alfke <snej@chromium.org> 14011 14012 Reviewed by Darin Adler. 14013 14014 Added variants of find/contains/add that allow a foreign key type to be used. 14015 This will allow AtomicString-keyed maps to be queried by C string without 14016 having to create a temporary AtomicString (see HTTPHeaderMap.) 14017 The code for this is adapted from the equivalent in HashSet.h. 14018 14019 * wtf/HashMap.h: 14020 (WTF::HashMap::find): 14021 (WTF::HashMap::contains): 14022 (WTF::HashMap::add): 14023 * wtf/HashSet.h: Changed "method" to "function member" in a comment. 14024 140252009-12-01 Gustavo Noronha Silva <gustavo.noronha@collabora.co.uk> 14026 14027 Revert 51551 because it broke GTK+. 14028 14029 * wtf/Platform.h: 14030 140312009-11-30 Gavin Barraclough <barraclough@apple.com> 14032 14033 Windows Build fix. Reviewed by NOBODY. 14034 14035 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 14036 140372009-11-24 Gavin Barraclough <barraclough@apple.com> 14038 14039 Reviewed by Geoff Garen. 14040 14041 Bug 31859 - Make world selection for JSC IsolatedWorlds automagical. 14042 14043 WebCore presently has to explicitly specify the world before entering into JSC, 14044 which is a little fragile (particularly since property access via a 14045 getter/setter might invoke execution). Instead derive the current world from 14046 the lexical global object. 14047 14048 Remove the temporary duct tape of willExecute/didExecute virtual hooks on the JSGlobalData::ClientData - these are no longer necessary. 14049 14050 * API/JSBase.cpp: 14051 (JSEvaluateScript): 14052 * API/JSObjectRef.cpp: 14053 (JSObjectCallAsFunction): 14054 * JavaScriptCore.exp: 14055 * runtime/JSGlobalData.cpp: 14056 * runtime/JSGlobalData.h: 14057 140582009-11-30 Laszlo Gombos <laszlo.1.gombos@nokia.com> 14059 14060 Reviewed by Kenneth Rohde Christiansen. 14061 14062 [Qt] Remove obsolete PLATFORM(KDE) code 14063 https://bugs.webkit.org/show_bug.cgi?id=31958 14064 14065 KDE is now using unpatched QtWebKit. 14066 14067 * parser/Lexer.cpp: Remove obsolete KDE_USE_FINAL guard 14068 * wtf/Platform.h: Remove PLATFORM(KDE) definition and code 14069 section that is guarded with it. 14070 140712009-11-30 Jan-Arve Sæther <jan-arve.saether@nokia.com> 14072 14073 Reviewed by Simon Hausmann. 14074 14075 [Qt] Fix compilation with win32-icc 14076 14077 The Intel compiler does not support the __has_trivial_constructor type 14078 trait. The Intel Compiler can report itself as _MSC_VER >= 1400. The 14079 reason for that is that the Intel Compiler depends on the Microsoft 14080 Platform SDK, and in order to try to be "fully" MS compatible it will 14081 "pretend" to be the same MS compiler as was shipped with the MS PSDK. 14082 (Thus, compiling with win32-icc with VC8 SDK will make the source code 14083 "think" the compiler at hand supports this type trait). 14084 14085 * wtf/TypeTraits.h: 14086 140872009-11-29 Laszlo Gombos <laszlo.1.gombos@nokia.com> 14088 14089 Reviewed by Eric Seidel. 14090 14091 [Qt] Mac build has JIT disabled 14092 https://bugs.webkit.org/show_bug.cgi?id=31828 14093 14094 * wtf/Platform.h: Enable JIT for Qt Mac builds 14095 140962009-11-28 Laszlo Gombos <laszlo.1.gombos@nokia.com> 14097 14098 Reviewed by Eric Seidel. 14099 14100 Apply workaround for the limitation of VirtualFree with MEM_RELEASE to all ports running on Windows 14101 https://bugs.webkit.org/show_bug.cgi?id=31943 14102 14103 * runtime/MarkStack.h: 14104 (JSC::MarkStack::MarkStackArray::shrinkAllocation): 14105 141062009-11-28 Zoltan Herczeg <zherczeg@inf.u-szeged.hu> 14107 14108 Reviewed by Gavin Barraclough. 14109 14110 https://bugs.webkit.org/show_bug.cgi?id=31930 14111 14112 Seems a typo. We don't need ~270k memory to determine the vptrs. 14113 14114 * runtime/JSGlobalData.cpp: 14115 (JSC::VPtrSet::VPtrSet): 14116 141172009-11-27 Shinichiro Hamaji <hamaji@chromium.org> 14118 14119 Unreviewed. 14120 14121 Move GOwnPtr* from wtf to wtf/gtk 14122 https://bugs.webkit.org/show_bug.cgi?id=31793 14123 14124 Build fix for chromium after r51423. 14125 Exclude gtk directory from chromium build. 14126 14127 * JavaScriptCore.gyp/JavaScriptCore.gyp: 14128 141292009-11-25 Oliver Hunt <oliver@apple.com> 14130 14131 Reviewed by Gavin Barraclough. 14132 14133 Incorrect behaviour of jneq_null in the interpreter 14134 https://bugs.webkit.org/show_bug.cgi?id=31901 14135 14136 Correct the logic of jneq_null. This is already covered by existing tests. 14137 14138 * interpreter/Interpreter.cpp: 14139 (JSC::Interpreter::privateExecute): 14140 141412009-11-26 Laszlo Gombos <laszlo.1.gombos@nokia.com> 14142 14143 Reviewed by Oliver Hunt. 14144 14145 Move GOwnPtr* from wtf to wtf/gtk 14146 https://bugs.webkit.org/show_bug.cgi?id=31793 14147 14148 * GNUmakefile.am: Change the path for GOwnPtr.*. 14149 * JavaScriptCore.gyp/JavaScriptCore.gyp: Remove 14150 GOwnPtr.cpp from the exclude list. 14151 * JavaScriptCore.gypi: Change the path for GOwnPtr.*. 14152 * wscript: Remove GOwnPtr.cpp from the exclude list. 14153 * wtf/GOwnPtr.cpp: Removed. 14154 * wtf/GOwnPtr.h: Removed. 14155 * wtf/Threading.h: Change the path for GOwnPtr.h. 14156 * wtf/gtk/GOwnPtr.cpp: Copied from JavaScriptCore/wtf/GOwnPtr.cpp. 14157 * wtf/gtk/GOwnPtr.h: Copied from JavaScriptCore/wtf/GOwnPtr.h. 14158 * wtf/unicode/glib/UnicodeGLib.h: Change the path for GOwnPtr.h. 14159 141602009-11-24 Dmitry Titov <dimich@chromium.org> 14161 14162 Reviewed by Eric Seidel. 14163 14164 Add ENABLE_SHARED_SCRIPT feature define and flag for build-webkit 14165 https://bugs.webkit.org/show_bug.cgi?id=31444 14166 14167 * Configurations/FeatureDefines.xcconfig: 14168 * wtf/Platform.h: 14169 141702009-11-24 Chris Marrin <cmarrin@apple.com> 14171 14172 Reviewed by Simon Fraser. 14173 14174 Add ability to enable ACCELERATED_COMPOSITING on Windows (currently disabled) 14175 https://bugs.webkit.org/show_bug.cgi?id=27314 14176 14177 * wtf/Platform.h: 14178 141792009-11-24 Jason Smith <dark.panda@gmail.com> 14180 14181 Reviewed by Alexey Proskuryakov. 14182 14183 RegExp#exec's returned Array-like object behaves differently from 14184 regular Arrays 14185 https://bugs.webkit.org/show_bug.cgi?id=31689 14186 14187 * JavaScriptCore/runtime/RegExpConstructor.cpp: ensure that undefined 14188 values are added to the returned RegExpMatchesArray 14189 141902009-11-24 Oliver Hunt <oliver@apple.com> 14191 14192 Reviewed by Alexey Proskuryakov. 14193 14194 JSON.stringify performance on undefined is very poor 14195 https://bugs.webkit.org/show_bug.cgi?id=31839 14196 14197 Switch from a UString to a Vector<UChar> when building 14198 the JSON string, allowing us to safely remove the substr-copy 14199 we otherwise did when unwinding an undefined property. 14200 14201 Also turns out to be a ~5% speedup on stringification. 14202 14203 * runtime/JSONObject.cpp: 14204 (JSC::Stringifier::StringBuilder::append): 14205 (JSC::Stringifier::stringify): 14206 (JSC::Stringifier::Holder::appendNextProperty): 14207 142082009-11-24 Mark Rowe <mrowe@apple.com> 14209 14210 Fix production builds where the source tree may be read-only. 14211 14212 * JavaScriptCore.xcodeproj/project.pbxproj: 14213 142142009-11-23 Laszlo Gombos <laszlo.1.gombos@nokia.com> 14215 14216 Reviewed by Kenneth Rohde Christiansen. 14217 14218 Include "config.h" to meet Coding Style Guidelines 14219 https://bugs.webkit.org/show_bug.cgi?id=31792 14220 14221 * wtf/unicode/UTF8.cpp: 14222 * wtf/unicode/glib/UnicodeGLib.cpp: 14223 * wtf/unicode/wince/UnicodeWince.cpp: 14224 142252009-11-23 Geoffrey Garen <ggaren@apple.com> 14226 14227 Reviewed by Oliver Hunt. 14228 14229 Streamlined some Math functions where we expect or know the result not 14230 to be representable as an int. 14231 14232 SunSpider says 0.6% faster. 14233 14234 * runtime/JSNumberCell.h: 14235 (JSC::JSValue::JSValue): 14236 * runtime/JSValue.h: 14237 (JSC::JSValue::): 14238 (JSC::jsDoubleNumber): 14239 (JSC::JSValue::JSValue): Added a function for making a numeric JSValue 14240 and skipping the "can I encode this as an int?" check, avoiding the 14241 overhead of int <-> double roundtripping and double <-> double comparison 14242 and branching. 14243 14244 * runtime/MathObject.cpp: 14245 (JSC::mathProtoFuncACos): 14246 (JSC::mathProtoFuncASin): 14247 (JSC::mathProtoFuncATan): 14248 (JSC::mathProtoFuncATan2): 14249 (JSC::mathProtoFuncCos): 14250 (JSC::mathProtoFuncExp): 14251 (JSC::mathProtoFuncLog): 14252 (JSC::mathProtoFuncRandom): 14253 (JSC::mathProtoFuncSin): 14254 (JSC::mathProtoFuncSqrt): 14255 (JSC::mathProtoFuncTan): For these functions, which we expect or know 14256 to produce results not representable as ints, call jsDoubleNumber instead 14257 of jsNumber. 14258 142592009-11-23 Mark Rowe <mrowe@apple.com> 14260 14261 Unreviewed. Unbreak the regression tests after r51329. 14262 14263 * API/JSBase.cpp: 14264 (JSEvaluateScript): Null-check clientData before dereferencing it. 14265 * API/JSObjectRef.cpp: 14266 (JSObjectCallAsFunction): Ditto. 14267 142682009-11-23 Gavin Barraclough <barraclough@apple.com> 14269 14270 Reviewed by Geoff Garen. 14271 14272 Part 1/3 of <rdar://problem/7377477> REGRESSION: Many web pages fail to render after interesting script runs in isolated world 14273 14274 Some clients of the JavaScriptCore API expect to be able to make callbacks over the JSC API, 14275 and for this to automagically cause execution to take place in the world associated with the 14276 global object associated with the ExecState (JSContextRef) passed. However this is not how 14277 things work - the world must be explicitly set within WebCore. 14278 14279 Making this work just for API calls to evaluate & call will be a far from perfect solution, 14280 since direct (non-API) use of JSC still relies on WebCore setting the current world correctly. 14281 A better solution would be to make this all work automagically all throughout WebCore, but this 14282 will require more refactoring. 14283 14284 Since the API is in JSC but worlds only exist in WebCore, add callbacks on the JSGlobalData::ClientData 14285 to allow it to update the current world on entry/exit via the JSC API. This is temporary duck 14286 tape, and should be removed once the current world no longer needs to be explicitly tracked. 14287 14288 * API/JSBase.cpp: 14289 (JSEvaluateScript): 14290 * API/JSObjectRef.cpp: 14291 (JSObjectCallAsFunction): 14292 * JavaScriptCore.exp: 14293 * runtime/JSGlobalData.cpp: 14294 (JSC::JSGlobalData::ClientData::beginningExecution): 14295 (JSC::JSGlobalData::ClientData::completedExecution): 14296 * runtime/JSGlobalData.h: 14297 142982009-11-23 Steve Block <steveblock@google.com> 14299 14300 Reviewed by Dmitry Titov. 14301 14302 Adds MainThreadAndroid.cpp with Android-specific WTF threading functions. 14303 https://bugs.webkit.org/show_bug.cgi?id=31807 14304 14305 * wtf/android: Added. 14306 * wtf/android/MainThreadAndroid.cpp: Added. 14307 (WTF::timeoutFired): 14308 (WTF::initializeMainThreadPlatform): 14309 (WTF::scheduleDispatchFunctionsOnMainThread): 14310 143112009-11-23 Alexey Proskuryakov <ap@apple.com> 14312 14313 Reviewed by Brady Eidson. 14314 14315 https://bugs.webkit.org/show_bug.cgi?id=31748 14316 Make WebSocketHandleCFNet respect proxy auto-configuration files via CFProxySupport 14317 14318 * JavaScriptCore.exp: Export callOnMainThreadAndWait. 14319 143202009-11-23 Laszlo Gombos <laszlo.1.gombos@nokia.com> 14321 14322 Reviewed by Kenneth Rohde Christiansen. 14323 14324 [Symbian] Fix lastIndexOf() for Symbian 14325 https://bugs.webkit.org/show_bug.cgi?id=31773 14326 14327 Symbian soft floating point library has problems with operators 14328 comparing NaN to numbers. Without a workaround lastIndexOf() 14329 function does not work. 14330 14331 Patch developed by David Leong. 14332 14333 * runtime/StringPrototype.cpp: 14334 (JSC::stringProtoFuncLastIndexOf):Add an extra test 14335 to check for NaN for Symbian. 14336 143372009-11-23 Steve Block <steveblock@google.com> 14338 14339 Reviewed by Eric Seidel. 14340 14341 Android port lacks implementation of atomicIncrement and atomicDecrement. 14342 https://bugs.webkit.org/show_bug.cgi?id=31715 14343 14344 * wtf/Threading.h: Modified. 14345 (WTF::atomicIncrement): Added Android implementation. 14346 (WTF::atomicDecrement): Added Android implementation. 14347 143482009-11-22 Laszlo Gombos <laszlo.1.gombos@nokia.com> 14349 14350 Unreviewed. 14351 14352 [Qt] Sort source lists and remove obsolete comments 14353 from the build system. 14354 14355 * JavaScriptCore.pri: 14356 143572009-11-21 Laszlo Gombos <laszlo.1.gombos@nokia.com> 14358 14359 Reviewed by Eric Seidel. 14360 14361 [Qt][Mac] Turn on multiple JavaScript threads for QtWebkit on Mac 14362 https://bugs.webkit.org/show_bug.cgi?id=31753 14363 14364 * wtf/Platform.h: 14365 143662009-11-19 Steve Block <steveblock@google.com> 14367 14368 Android port lacks configuration in Platform.h and config.h. 14369 https://bugs.webkit.org/show_bug.cgi?id=31671 14370 14371 * wtf/Platform.h: Modified. Added Android-specific configuration. 14372 143732009-11-19 Alexey Proskuryakov <ap@apple.com> 14374 14375 Reviewed by Darin Adler. 14376 14377 https://bugs.webkit.org/show_bug.cgi?id=31690 14378 Make SocketStreamHandleCFNet work on Windows 14379 14380 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 14381 * wtf/MainThread.cpp: 14382 (WTF::FunctionWithContext::FunctionWithContext): 14383 (WTF::dispatchFunctionsFromMainThread): 14384 (WTF::callOnMainThreadAndWait): 14385 * wtf/MainThread.h: 14386 Re-add callOnMainThreadAndWait(), which was removed in bug 23926. 14387 143882009-11-19 Dmitry Titov <dimich@chromium.org> 14389 14390 Reviewed by David Levin. 14391 14392 isMainThread() on Chromium (Mac and Linux) is so slow it timeouts LayoutTests.. 14393 https://bugs.webkit.org/show_bug.cgi?id=31693 14394 14395 * wtf/ThreadingPthreads.cpp: 14396 (WTF::initializeThreading): grab and use the pthread_t of the main thread instead of ThreadIdentifier. 14397 (WTF::isMainThread): Ditto. 14398 143992009-11-19 Laszlo Gombos <laszlo.1.gombos@nokia.com> 14400 14401 Reviewed by Darin Adler. 14402 14403 Remove HAVE(STRING_H) guard from JavaScriptCore 14404 https://bugs.webkit.org/show_bug.cgi?id=31668 14405 14406 * config.h: 14407 * runtime/UString.cpp: 14408 144092009-11-19 Dumitru Daniliuc <dumi@chromium.org> 14410 14411 Reviewed by Dmitry Titov. 14412 14413 Fixing a bug in MessageQueue::removeIf() that leads to an 14414 assertion failure. 14415 14416 https://bugs.webkit.org/show_bug.cgi?id=31657 14417 14418 * wtf/MessageQueue.h: 14419 (WTF::MessageQueue::removeIf): 14420 144212009-11-19 Laszlo Gombos <laszlo.1.gombos@nokia.com> 14422 14423 Reviewed by Darin Adler. 14424 14425 Remove HAVE(FLOAT_H) guard 14426 https://bugs.webkit.org/show_bug.cgi?id=31661 14427 14428 JavaScriptCore has a dependency on float.h, there is 14429 no need to guard float.h. 14430 14431 * runtime/DatePrototype.cpp: Remove include directive 14432 for float.h as it is included in MathExtras.h already. 14433 * runtime/Operations.cpp: Ditto. 14434 * runtime/UString.cpp: Ditto. 14435 * wtf/dtoa.cpp: Ditto. 14436 * wtf/MathExtras.h: Remove HAVE(FLOAT_H) guard. 14437 * wtf/Platform.h: Ditto. 14438 144392009-11-19 Thiago Macieira <thiago.macieira@nokia.com> 14440 14441 Reviewed by Simon Hausmann. 14442 14443 Build fix for 32-bit Sparc machines: these machines are big-endian. 14444 14445 * wtf/Platform.h: 14446 144472009-11-18 Laszlo Gombos <laszlo.1.gombos@nokia.com> 14448 14449 Reviewed by Kenneth Rohde Christiansen. 14450 14451 [Qt] Remove support for Qt v4.3 or older versions 14452 https://bugs.webkit.org/show_bug.cgi?id=29469 14453 14454 * JavaScriptCore.pro: 14455 * jsc.pro: 14456 * wtf/unicode/qt4/UnicodeQt4.h: 14457 144582009-11-18 Kent Tamura <tkent@chromium.org> 14459 14460 Reviewed by Darin Adler. 14461 14462 Move UString::from(double) implementation to new 14463 WTF::doubleToStringInJavaScriptFormat(), and expose it because WebCore 14464 code will use it. 14465 https://bugs.webkit.org/show_bug.cgi?id=31330 14466 14467 - Introduce new function createRep(const char*, unsigned) and 14468 UString::UString(const char*, unsigned) to reduce 2 calls to strlen(). 14469 - Fix a bug that dtoa() doesn't update *rve if the input value is NaN 14470 or Infinity. 14471 14472 No new tests because this doesn't change the behavior. 14473 14474 * JavaScriptCore.exp: 14475 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 14476 * runtime/UString.cpp: 14477 (JSC::createRep): 14478 (JSC::UString::UString): 14479 (JSC::UString::from): Move the code to doubleToStringInJavaScriptFormat(). 14480 * runtime/UString.h: 14481 * wtf/dtoa.cpp: 14482 (WTF::dtoa): Fix a bug about rve. 14483 (WTF::append): A helper for doubleToStringInJavaScriptFormat(). 14484 (WTF::doubleToStringInJavaScriptFormat): Move the code from UString::from(double). 14485 * wtf/dtoa.h: 14486 144872009-11-18 Laszlo Gombos <laszlo.1.gombos@nokia.com> 14488 14489 Reviewed by Kenneth Rohde Christiansen. 14490 14491 [Qt] Remove WTF_USE_JAVASCRIPTCORE_BINDINGS as it is no longer used 14492 https://bugs.webkit.org/show_bug.cgi?id=31643 14493 14494 * JavaScriptCore.pro: 14495 144962009-11-18 Nate Chapin <japhet@chromium.org> 14497 14498 Reviewed by Darin Fisher. 14499 14500 Remove Chromium's unnecessary dependency on wtf's tcmalloc files. 14501 14502 https://bugs.webkit.org/show_bug.cgi?id=31648 14503 14504 * JavaScriptCore.gyp/JavaScriptCore.gyp: 14505 145062009-11-18 Thiago Macieira <thiago.macieira@nokia.com> 14507 14508 Reviewed by Gavin Barraclough. 14509 14510 [Qt] Implement symbol hiding for JSC's JIT functions. 14511 14512 These functions are implemented directly in assembly, so they need the 14513 proper directives to enable/disable visibility. On ELF systems, it's 14514 .hidden, whereas on Mach-O systems (Mac) it's .private_extern. On 14515 Windows, it's not necessary since you have to explicitly export. I 14516 also implemented the AIX idiom, though it's unlikely anyone will 14517 implement AIX/POWER JIT. 14518 https://bugs.webkit.org/show_bug.cgi?id=30864 14519 14520 * jit/JITStubs.cpp: 14521 145222009-11-18 Oliver Hunt <oliver@apple.com> 14523 14524 Reviewed by Alexey Proskuryakov. 14525 14526 Interpreter may do an out of range access when throwing an exception in the profiler. 14527 https://bugs.webkit.org/show_bug.cgi?id=31635 14528 14529 Add bounds check. 14530 14531 * interpreter/Interpreter.cpp: 14532 (JSC::Interpreter::throwException): 14533 145342009-11-18 Gabor Loki <loki@inf.u-szeged.hu> 14535 14536 Reviewed by Darin Adler. 14537 14538 Fix the clobber list of cacheFlush for ARM and Thumb2 on Linux 14539 https://bugs.webkit.org/show_bug.cgi?id=31631 14540 14541 * jit/ExecutableAllocator.h: 14542 (JSC::ExecutableAllocator::cacheFlush): 14543 145442009-11-18 Harald Fernengel <harald.fernengel@nokia.com> 14545 14546 Reviewed by Simon Hausmann. 14547 14548 [Qt] Fix detection of linux-g++ 14549 14550 Never use "linux-g++*" to check for linux-g++, since this will break embedded 14551 builds which use linux-arm-g++ and friends. Use 'linux*-g++*' to check for any 14552 g++ on linux mkspec. 14553 14554 * JavaScriptCore.pri: 14555 145562009-11-17 Jon Honeycutt <jhoneycutt@apple.com> 14557 14558 Add JSContextRefPrivate.h to list of copied files. 14559 14560 Reviewed by Mark Rowe. 14561 14562 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCoreGenerated.make: 14563 145642009-11-17 Martin Robinson <martin.james.robinson@gmail.com> 14565 14566 Reviewed by Adam Barth. 14567 14568 [GTK] Style cleanup for GOwnPtr 14569 https://bugs.webkit.org/show_bug.cgi?id=31506 14570 14571 Remove forward declaration in GOwnPtr and do some style cleanup. 14572 14573 * wtf/GOwnPtr.cpp: 14574 * wtf/GOwnPtr.h: 14575 (WTF::GOwnPtr::GOwnPtr): 14576 (WTF::GOwnPtr::~GOwnPtr): 14577 (WTF::GOwnPtr::get): 14578 (WTF::GOwnPtr::release): 14579 (WTF::GOwnPtr::outPtr): 14580 (WTF::GOwnPtr::set): 14581 (WTF::GOwnPtr::clear): 14582 (WTF::GOwnPtr::operator*): 14583 (WTF::GOwnPtr::operator->): 14584 (WTF::GOwnPtr::operator!): 14585 (WTF::GOwnPtr::operator UnspecifiedBoolType): 14586 (WTF::GOwnPtr::swap): 14587 (WTF::swap): 14588 (WTF::operator==): 14589 (WTF::operator!=): 14590 (WTF::getPtr): 14591 (WTF::freeOwnedGPtr): 14592 145932009-11-17 Oliver Hunt <oliver@apple.com> 14594 14595 Reviewed by Maciej Stachowiak. 14596 14597 Incorrect use of JavaScriptCore API in DumpRenderTree 14598 https://bugs.webkit.org/show_bug.cgi?id=31577 14599 14600 Add assertions to the 'toJS' functions to catch mistakes like 14601 this early. Restructure existing code which blindly passed potentially 14602 null values to toJS when forwarding exceptions so that a null check is 14603 performed first. 14604 14605 * API/APICast.h: 14606 (toJS): 14607 (toJSForGC): 14608 * API/JSCallbackObjectFunctions.h: 14609 (JSC::::getOwnPropertySlot): 14610 (JSC::::put): 14611 (JSC::::deleteProperty): 14612 (JSC::::construct): 14613 (JSC::::hasInstance): 14614 (JSC::::call): 14615 (JSC::::toNumber): 14616 (JSC::::toString): 14617 (JSC::::staticValueGetter): 14618 (JSC::::callbackGetter): 14619 * API/tests/testapi.c: Fix errors in the API tester. 14620 (MyObject_getProperty): 14621 (MyObject_convertToType): 14622 (EvilExceptionObject_convertToType): 14623 146242009-11-16 Zoltan Herczeg <zherczeg@inf.u-szeged.hu> 14625 14626 Reviewed by Gavin Barraclough. 14627 14628 https://bugs.webkit.org/show_bug.cgi?id=31050 14629 14630 Minor fixes for JSVALUE32_64: branchConvertDoubleToInt32 14631 failed on a CortexA8 CPU, but not on a simulator; and 14632 JITCall.cpp modifications was somehow not committed to mainline. 14633 14634 * assembler/ARMAssembler.h: 14635 (JSC::ARMAssembler::fmrs_r): 14636 * assembler/MacroAssemblerARM.h: 14637 (JSC::MacroAssemblerARM::branchConvertDoubleToInt32): 14638 * jit/JITCall.cpp: 14639 (JSC::JIT::compileOpCall): 14640 146412009-11-16 Joerg Bornemann <joerg.bornemann@trolltech.com> 14642 14643 Reviewed by Simon Hausmann. 14644 14645 Fix Qt build on Windows CE 6. 14646 14647 * JavaScriptCore.pri: Add missing include path. 14648 * wtf/Platform.h: Include ce_time.h for Windows CE 6. 14649 146502009-11-13 Zoltan Herczeg <zherczeg@inf.u-szeged.hu> 14651 14652 Reviewed by Gavin Barraclough. 14653 14654 https://bugs.webkit.org/show_bug.cgi?id=31050 14655 14656 Adding optimization support for mode JSVALUE32_64 14657 on ARM systems. 14658 14659 * jit/JIT.h: 14660 * jit/JITCall.cpp: 14661 (JSC::JIT::compileOpCall): 14662 * jit/JITPropertyAccess.cpp: 14663 (JSC::JIT::emit_op_method_check): 14664 (JSC::JIT::compileGetByIdHotPath): 14665 (JSC::JIT::compileGetByIdSlowCase): 14666 (JSC::JIT::emit_op_put_by_id): 14667 146682009-11-14 Zoltan Herczeg <zherczeg@inf.u-szeged.hu> 14669 14670 Reviewed by Gavin Barraclough. 14671 14672 https://bugs.webkit.org/show_bug.cgi?id=31050 14673 14674 Adding JSVALUE32_64 support for ARM (but not turning it 14675 on by default). All optimizations must be disabled, since 14676 this patch is only the first of a series of patches. 14677 14678 During the work, a lot of x86 specific code revealed and 14679 made platform independent. 14680 See revisions: 50531 50541 50593 50594 50595 14681 14682 * assembler/ARMAssembler.h: 14683 (JSC::ARMAssembler::): 14684 (JSC::ARMAssembler::fdivd_r): 14685 * assembler/MacroAssemblerARM.h: 14686 (JSC::MacroAssemblerARM::lshift32): 14687 (JSC::MacroAssemblerARM::neg32): 14688 (JSC::MacroAssemblerARM::rshift32): 14689 (JSC::MacroAssemblerARM::branchOr32): 14690 (JSC::MacroAssemblerARM::set8): 14691 (JSC::MacroAssemblerARM::setTest8): 14692 (JSC::MacroAssemblerARM::loadDouble): 14693 (JSC::MacroAssemblerARM::divDouble): 14694 (JSC::MacroAssemblerARM::convertInt32ToDouble): 14695 (JSC::MacroAssemblerARM::zeroDouble): 14696 * jit/JIT.cpp: 14697 * jit/JIT.h: 14698 * jit/JITOpcodes.cpp: 14699 (JSC::JIT::privateCompileCTIMachineTrampolines): 14700 * jit/JITStubs.cpp: 14701 * wtf/StdLibExtras.h: 14702 147032009-11-13 Dominik Röttsches <dominik.roettsches@access-company.com> 14704 14705 Reviewed by Eric Seidel. 14706 14707 Unify TextBoundaries implementations by only relying on WTF Unicode abstractions 14708 https://bugs.webkit.org/show_bug.cgi?id=31468 14709 14710 Adding isAlphanumeric abstraction, required 14711 by TextBoundaries.cpp. 14712 14713 * wtf/unicode/glib/UnicodeGLib.h: 14714 (WTF::Unicode::isAlphanumeric): 14715 * wtf/unicode/icu/UnicodeIcu.h: 14716 (WTF::Unicode::isAlphanumeric): 14717 147182009-11-13 Norbert Leser <norbert.leser&nokia.com> 14719 14720 Reviewed by Eric Seidel. 14721 14722 Added macros for USERINCLUDE paths within symbian blocks 14723 to guarantee inclusion of respective header files from local path 14724 first (to avoid clashes with same names of header files in system include path). 14725 14726 * JavaScriptCore.pri: 14727 147282009-11-13 Oliver Hunt <oliver@apple.com> 14729 14730 Reviewed by Geoff Garen. 14731 14732 JSValueProtect and JSValueUnprotect don't protect API wrapper values 14733 https://bugs.webkit.org/show_bug.cgi?id=31485 14734 14735 Make JSValueProtect/Unprotect use a new 'toJS' function, 'toJSForGC' that 14736 does not attempt to to strip out API wrapper objects. 14737 14738 * API/APICast.h: 14739 (toJSForGC): 14740 * API/JSValueRef.cpp: 14741 (JSValueProtect): 14742 (JSValueUnprotect): 14743 * API/tests/testapi.c: 14744 (makeGlobalNumberValue): 14745 (main): 14746 147472009-11-13 İsmail Dönmez <ismail@namtrac.org> 14748 14749 Reviewed by Antti Koivisto. 14750 14751 Fix typo, ce_time.cpp should be ce_time.c 14752 14753 * JavaScriptCore.pri: 14754 147552009-11-12 Steve VanDeBogart <vandebo@chromium.org> 14756 14757 Reviewed by Adam Barth. 14758 14759 Calculate the time offset only if we were able to parse 14760 the date string. This saves an IPC in Chromium for 14761 invalid date strings. 14762 https://bugs.webkit.org/show_bug.cgi?id=31416 14763 14764 * wtf/DateMath.cpp: 14765 (WTF::parseDateFromNullTerminatedCharacters): 14766 (JSC::parseDateFromNullTerminatedCharacters): 14767 147682009-11-12 Oliver Hunt <oliver@apple.com> 14769 14770 Rollout r50896 until i can work out why it causes failures. 14771 14772 * bytecompiler/BytecodeGenerator.cpp: 14773 (JSC::BytecodeGenerator::emitReturn): 14774 * interpreter/Interpreter.cpp: 14775 (JSC::Interpreter::execute): 14776 * parser/Nodes.cpp: 14777 (JSC::EvalNode::emitBytecode): 14778 147792009-11-12 Steve Falkenburg <sfalken@apple.com> 14780 14781 Reviewed by Stephanie Lewis. 14782 14783 Remove LIBRARY directive from def file to fix Debug_All target. 14784 14785 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 14786 147872009-11-12 Gustavo Noronha Silva <gustavo.noronha@collabora.co.uk> 14788 14789 Rubber-stamped by Holger Freyther. 14790 14791 Revert r50204, since it makes DRT crash on 32 bits release builds 14792 for GTK+. 14793 14794 * wtf/FastMalloc.h: 14795 147962009-11-12 Oliver Hunt <oliver@apple.com> 14797 14798 Reviewed by Gavin Barraclough. 14799 14800 Start unifying entry logic for function and eval code. 14801 14802 Eval now uses a ret instruction to end execution, and sets up 14803 a callframe more in line with what we do for function entry. 14804 14805 * bytecompiler/BytecodeGenerator.cpp: 14806 (JSC::BytecodeGenerator::emitReturn): 14807 * interpreter/Interpreter.cpp: 14808 (JSC::Interpreter::execute): 14809 * parser/Nodes.cpp: 14810 (JSC::EvalNode::emitBytecode): 14811 148122009-11-12 Richard Moe Gustavsen <richard.gustavsen@nokia.com> 14813 14814 Reviewed by Kenneth Rohde Christiansen. 14815 14816 [Qt] Disable pthread_setname_np. 14817 14818 This allows Qt builds on Mac from 10.6 to run on earlier version 14819 where this symbol is not present. 14820 https://bugs.webkit.org/show_bug.cgi?id=31403 14821 14822 * wtf/Platform.h: 14823 148242009-11-12 Thiago Macieira <thiago.macieira@nokia.com> 14825 14826 Reviewed by Kenneth Rohde Christiansen. 14827 14828 [Qt] Fix linking on Linux 32-bit. 14829 14830 It was missing the ".text" directive at the top of the file, 14831 indicating that code would follow. Without it, the assembler created 14832 "NOTYPE" symbols, which would result in linker errors. 14833 https://bugs.webkit.org/show_bug.cgi?id=30863 14834 14835 * jit/JITStubs.cpp: 14836 148372009-11-11 Laszlo Gombos <laszlo.1.gombos@nokia.com> 14838 14839 Reviewed by Alexey Proskuryakov. 14840 14841 Refactor multiple JavaScriptCore threads 14842 https://bugs.webkit.org/show_bug.cgi?id=31328 14843 14844 Remove the id field from the PlatformThread structure 14845 as it is not used. 14846 14847 * runtime/Collector.cpp: 14848 (JSC::getCurrentPlatformThread): 14849 (JSC::suspendThread): 14850 (JSC::resumeThread): 14851 (JSC::getPlatformThreadRegisters): 14852 148532009-11-10 Geoffrey Garen <ggaren@apple.com> 14854 14855 Linux build fix: Added an #include for UINT_MAX. 14856 14857 * runtime/WeakRandom.h: 14858 148592009-11-10 Geoffrey Garen <ggaren@apple.com> 14860 14861 JavaScriptGlue build fix: Marked a file 'private' instead of 'project'. 14862 14863 * JavaScriptCore.xcodeproj/project.pbxproj: 14864 148652009-11-10 Geoffrey Garen <ggaren@apple.com> 14866 14867 Reviewed by Gavin "avGni arBalroguch" Barraclough. 14868 14869 Faster Math.random, based on GameRand. 14870 14871 SunSpider says 1.4% faster. 14872 14873 * GNUmakefile.am: 14874 * JavaScriptCore.gypi: 14875 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: 14876 * JavaScriptCore.xcodeproj/project.pbxproj: Added the header to the project. 14877 14878 * runtime/JSGlobalData.cpp: 14879 (JSC::JSGlobalData::JSGlobalData): 14880 * runtime/JSGlobalData.h: Use an object to track random number generation 14881 state, initialized to the current time. 14882 14883 * runtime/MathObject.cpp: 14884 (JSC::MathObject::MathObject): 14885 (JSC::mathProtoFuncRandom): Use the new hotness. 14886 14887 * runtime/WeakRandom.h: Added. 14888 (JSC::WeakRandom::WeakRandom): 14889 (JSC::WeakRandom::get): 14890 (JSC::WeakRandom::advance): The new hotness. 14891 148922009-11-09 Geoffrey Garen <ggaren@apple.com> 14893 14894 Reviewed by Oliver Hunt. 14895 14896 Imported the v8 DST cache. 14897 14898 SunSpider says 1.5% faster. 14899 14900 * runtime/JSGlobalData.cpp: 14901 (JSC::JSGlobalData::resetDateCache): Reset the DST cache when resetting 14902 other date data. 14903 14904 * runtime/JSGlobalData.h: 14905 (JSC::DSTOffsetCache::DSTOffsetCache): 14906 (JSC::DSTOffsetCache::reset): Added a struct for the DST cache. 14907 14908 * wtf/DateMath.cpp: 14909 (WTF::calculateDSTOffsetSimple): 14910 (WTF::calculateDSTOffset): 14911 (WTF::parseDateFromNullTerminatedCharacters): 14912 (JSC::getDSTOffset): 14913 (JSC::gregorianDateTimeToMS): 14914 (JSC::msToGregorianDateTime): 14915 (JSC::parseDateFromNullTerminatedCharacters): 14916 * wtf/DateMath.h: The imported code for probing and updating the cache. 14917 149182009-11-09 Geoffrey Garen <ggaren@apple.com> 14919 14920 Reviewed by Oliver Hunt. 14921 14922 Fixed an edge case that could cause the engine not to notice a timezone 14923 change. 14924 14925 No test because this case would require manual intervention to change 14926 the timezone during the test. 14927 14928 SunSpider reports no change. 14929 14930 * runtime/DateInstanceCache.h: 14931 (JSC::DateInstanceCache::DateInstanceCache): 14932 (JSC::DateInstanceCache::reset): Added a helper function for resetting 14933 this cache. Also, shrank the cache, since we'll be resetting it often. 14934 14935 * runtime/JSGlobalData.cpp: 14936 (JSC::JSGlobalData::resetDateCache): Include resetting the DateInstanceCache 14937 in resetting Date data. (Otherwise, a cache hit could bypass a necessary 14938 timezone update check.) 14939 149402009-11-09 Geoffrey Garen <ggaren@apple.com> 14941 14942 Reviewed by Sam Weinig. 14943 14944 Some manual inlining and constant propogation in Date code. 14945 14946 SunSpider reports a 0.4% speedup on date-*, no overall speedup. Shark 14947 says some previously evident stalls are now gone. 14948 14949 * runtime/DateConstructor.cpp: 14950 (JSC::callDate): 14951 * runtime/DateConversion.cpp: 14952 (JSC::formatTime): 14953 (JSC::formatTimeUTC): Split formatTime into UTC and non-UTC variants. 14954 14955 * runtime/DateConversion.h: 14956 * runtime/DateInstance.cpp: 14957 (JSC::DateInstance::calculateGregorianDateTime): 14958 (JSC::DateInstance::calculateGregorianDateTimeUTC): 14959 * runtime/DateInstance.h: 14960 (JSC::DateInstance::gregorianDateTime): 14961 (JSC::DateInstance::gregorianDateTimeUTC): Split gregorianDateTime into 14962 a UTC and non-UTC variant, and split each variant into a fast inline 14963 case and a slow out-of-line case. 14964 14965 * runtime/DatePrototype.cpp: 14966 (JSC::formatLocaleDate): 14967 (JSC::dateProtoFuncToString): 14968 (JSC::dateProtoFuncToUTCString): 14969 (JSC::dateProtoFuncToISOString): 14970 (JSC::dateProtoFuncToDateString): 14971 (JSC::dateProtoFuncToTimeString): 14972 (JSC::dateProtoFuncGetFullYear): 14973 (JSC::dateProtoFuncGetUTCFullYear): 14974 (JSC::dateProtoFuncToGMTString): 14975 (JSC::dateProtoFuncGetMonth): 14976 (JSC::dateProtoFuncGetUTCMonth): 14977 (JSC::dateProtoFuncGetDate): 14978 (JSC::dateProtoFuncGetUTCDate): 14979 (JSC::dateProtoFuncGetDay): 14980 (JSC::dateProtoFuncGetUTCDay): 14981 (JSC::dateProtoFuncGetHours): 14982 (JSC::dateProtoFuncGetUTCHours): 14983 (JSC::dateProtoFuncGetMinutes): 14984 (JSC::dateProtoFuncGetUTCMinutes): 14985 (JSC::dateProtoFuncGetSeconds): 14986 (JSC::dateProtoFuncGetUTCSeconds): 14987 (JSC::dateProtoFuncGetTimezoneOffset): 14988 (JSC::setNewValueFromTimeArgs): 14989 (JSC::setNewValueFromDateArgs): 14990 (JSC::dateProtoFuncSetYear): 14991 (JSC::dateProtoFuncGetYear): Updated for the gregorianDateTime change above. 14992 149932009-11-09 Geoffrey Garen <ggaren@apple.com> 14994 14995 Build fix: export a new symbol. 14996 14997 * JavaScriptCore.exp: 14998 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 14999 150002009-11-09 Geoffrey Garen <ggaren@apple.com> 15001 15002 Reviewed by Sam "Home Wrecker" Weinig. 15003 15004 Added a tiny cache for Date parsing. 15005 15006 SunSpider says 1.2% faster. 15007 15008 * runtime/DateConversion.cpp: 15009 (JSC::parseDate): Try to reuse the last parsed Date, if present. 15010 15011 * runtime/JSGlobalData.cpp: 15012 (JSC::JSGlobalData::resetDateCache): 15013 * runtime/JSGlobalData.h: Added storage for last parsed Date. Refactored 15014 this code to make resetting the date cache easier. 15015 15016 * runtime/JSGlobalObject.h: 15017 (JSC::DynamicGlobalObjectScope::DynamicGlobalObjectScope): Updated for 15018 refactoring. 15019 15020 * wtf/DateMath.cpp: 15021 (JSC::parseDateFromNullTerminatedCharacters): 15022 * wtf/DateMath.h: Changed ExecState to be first parameter, as is the JSC custom. 15023 150242009-11-09 Oliver Hunt <oliver@apple.com> 15025 15026 Reviewed by Gavin Barraclough. 15027 15028 Can cache prototype lookups on uncacheable dictionaries. 15029 https://bugs.webkit.org/show_bug.cgi?id=31198 15030 15031 Replace fromDictionaryTransition with flattenDictionaryObject and 15032 flattenDictionaryStructure. This change is necessary as we need to 15033 guarantee that our attempt to convert away from a dictionary structure 15034 will definitely succeed, and in some cases this requires mutating the 15035 object storage itself. 15036 15037 * interpreter/Interpreter.cpp: 15038 (JSC::Interpreter::tryCacheGetByID): 15039 * jit/JITStubs.cpp: 15040 (JSC::JITThunks::tryCacheGetByID): 15041 (JSC::DEFINE_STUB_FUNCTION): 15042 * runtime/BatchedTransitionOptimizer.h: 15043 (JSC::BatchedTransitionOptimizer::~BatchedTransitionOptimizer): 15044 * runtime/JSObject.h: 15045 (JSC::JSObject::flattenDictionaryObject): 15046 * runtime/Operations.h: 15047 (JSC::normalizePrototypeChain): 15048 * runtime/Structure.cpp: 15049 (JSC::Structure::flattenDictionaryStructure): 15050 (JSC::comparePropertyMapEntryIndices): 15051 * runtime/Structure.h: 15052 150532009-11-09 Laszlo Gombos <laszlo.1.gombos@nokia.com> 15054 15055 Not reviewed, build fix. 15056 15057 Remove extra character from r50701. 15058 15059 * JavaScriptCore.pri: 15060 150612009-11-09 Laszlo Gombos <laszlo.1.gombos@nokia.com> 15062 15063 Not reviewed, build fix. 15064 15065 Revert r50695 because it broke QtWebKit (clean builds). 15066 15067 * JavaScriptCore.pri: 15068 150692009-11-09 Norbert Leser <norbert.leser@nokia.com> 15070 15071 Reviewed by Kenneth Rohde Christiansen. 15072 15073 Prepended $$PWD to GENERATED_SOURCES_DIR to avoid potential ambiguities when included from WebCore.pro. 15074 Some preprocessors consider this GENERATED_SOURCES_DIR relative to current invoking dir (e.g., ./WebCore), 15075 and not the working dir of JavaCriptCore.pri (i.e., ../JavaScriptCore/). 15076 15077 * JavaScriptCore.pri: 15078 150792009-11-09 Laszlo Gombos <laszlo.1.gombos@nokia.com> 15080 15081 Reviewed by Kenneth Rohde Christiansen. 15082 15083 Use explicit parentheses to silence gcc 4.4 -Wparentheses warnings 15084 https://bugs.webkit.org/show_bug.cgi?id=31040 15085 15086 * interpreter/Interpreter.cpp: 15087 (JSC::Interpreter::privateExecute): 15088 150892009-11-08 David Levin <levin@chromium.org> 15090 15091 Reviewed by NOBODY (speculative snow leopard and windows build fixes). 15092 15093 * wtf/DateMath.cpp: 15094 (WTF::parseDateFromNullTerminatedCharacters): 15095 (JSC::gregorianDateTimeToMS): 15096 (JSC::msToGregorianDateTime): 15097 (JSC::parseDateFromNullTerminatedCharacters): 15098 * wtf/DateMath.h: 15099 (JSC::GregorianDateTime::GregorianDateTime): 15100 151012009-11-08 David Levin <levin@chromium.org> 15102 15103 Reviewed by NOBODY (chromium build fix). 15104 15105 Hopefully, the last build fix. 15106 15107 Create better separation in DateMath about the JSC 15108 and non-JSC portions. Also, only expose the non-JSC 15109 version in the exports. 15110 15111 * JavaScriptCore.exp: 15112 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 15113 * wtf/DateMath.cpp: 15114 (WTF::parseDateFromNullTerminatedCharacters): 15115 (JSC::getUTCOffset): 15116 (JSC::gregorianDateTimeToMS): 15117 (JSC::msToGregorianDateTime): 15118 (JSC::parseDateFromNullTerminatedCharacters): 15119 * wtf/DateMath.h: 15120 (JSC::gmtoffset): 15121 151222009-11-08 David Levin <levin@chromium.org> 15123 15124 Reviewed by NOBODY (chromium build fix). 15125 15126 For the change in DateMath. 15127 15128 * config.h: 15129 * wtf/DateMath.cpp: 15130 151312009-11-06 Geoffrey Garen <ggaren@apple.com> 15132 15133 Windows build fix: export some symbols. 15134 15135 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 15136 151372009-11-06 Geoffrey Garen <ggaren@apple.com> 15138 15139 Build fix: updated export file. 15140 15141 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 15142 151432009-11-06 Geoffrey Garen <ggaren@apple.com> 15144 15145 Build fix: added some #includes. 15146 15147 * wtf/CurrentTime.h: 15148 * wtf/DateMath.h: 15149 151502009-11-06 Geoffrey Garen <ggaren@apple.com> 15151 15152 Reviewed by Oliver Hunt. 15153 15154 https://bugs.webkit.org/show_bug.cgi?id=31197 15155 Implemented a timezone cache not based on Mac OS X's notify_check API. 15156 15157 If the VM calculates the local timezone offset from UTC, it caches the 15158 result until the end of the current VM invocation. (We don't want to cache 15159 forever, because the user's timezone may change over time.) 15160 15161 This removes notify_* overhead on Mac, and, more significantly, removes 15162 OS time and date call overhead on non-Mac platforms. 15163 15164 ~8% speedup on Date microbenchmark on Mac. SunSpider reports maybe a tiny 15165 speedup on Mac. (Speedup on non-Mac platforms should be even more noticeable.) 15166 15167 * JavaScriptCore.exp: 15168 15169 * interpreter/CachedCall.h: 15170 (JSC::CachedCall::CachedCall): 15171 * interpreter/Interpreter.cpp: 15172 (JSC::Interpreter::execute): 15173 * runtime/JSGlobalObject.h: 15174 (JSC::DynamicGlobalObjectScope::DynamicGlobalObjectScope): Made the 15175 DynamicGlobalObjectScope constructor responsible for checking whether a 15176 dynamicGlobalObject has already been set. This eliminated some duplicate 15177 client code, and allowed me to avoid adding even more duplicate client 15178 code. Made DynamicGlobalObjectScope responsible for resetting the 15179 local timezone cache upon first entry to the VM. 15180 15181 * runtime/DateConstructor.cpp: 15182 (JSC::constructDate): 15183 (JSC::callDate): 15184 (JSC::dateParse): 15185 (JSC::dateUTC): 15186 * runtime/DateConversion.cpp: 15187 (JSC::parseDate): 15188 * runtime/DateConversion.h: 15189 * runtime/DateInstance.cpp: 15190 (JSC::DateInstance::gregorianDateTime): 15191 * runtime/DateInstance.h: 15192 * runtime/DateInstanceCache.h: 15193 * runtime/DatePrototype.cpp: 15194 (JSC::setNewValueFromTimeArgs): 15195 (JSC::setNewValueFromDateArgs): 15196 (JSC::dateProtoFuncSetYear): 15197 * runtime/InitializeThreading.cpp: 15198 (JSC::initializeThreadingOnce): 15199 * runtime/JSGlobalData.cpp: 15200 (JSC::JSGlobalData::JSGlobalData): 15201 * runtime/JSGlobalData.h: 15202 * wtf/DateMath.cpp: 15203 (WTF::getCurrentUTCTime): 15204 (WTF::getCurrentUTCTimeWithMicroseconds): 15205 (WTF::getLocalTime): 15206 (JSC::getUTCOffset): Use the new cache. Also, see below. 15207 (JSC::gregorianDateTimeToMS): 15208 (JSC::msToGregorianDateTime): 15209 (JSC::initializeDates): 15210 (JSC::parseDateFromNullTerminatedCharacters): Simplified the way this function 15211 accounts for the local timezone offset, to accomodate our new caching API, 15212 and a (possibly misguided) caller in WebCore. Also, see below. 15213 * wtf/DateMath.h: 15214 (JSC::GregorianDateTime::GregorianDateTime): Moved most of the code in 15215 DateMath.* into the JSC namespace. The code needed to move so it could 15216 naturally interact with ExecState and JSGlobalData to support caching. 15217 Logically, it seemed right to move it, too, since this code is not really 15218 as low-level as the WTF namespace might imply -- it implements a set of 15219 date parsing and conversion quirks that are finely tuned to the JavaScript 15220 language. Also removed the Mac OS X notify_* infrastructure. 15221 15222 * wtf/CurrentTime.h: 15223 (WTF::currentTimeMS): 15224 (WTF::getLocalTime): Moved the rest of the DateMath code here, and renamed 15225 it to make it consistent with WTF's currentTime function. 15226 152272009-11-06 Gabor Loki <loki@inf.u-szeged.hu> 15228 15229 Unreviewed trivial buildfix after r50595. 15230 15231 Rename the remaining rshiftPtr calls to rshift32 15232 15233 * jit/JITArithmetic.cpp: 15234 (JSC::JIT::emit_op_rshift): 15235 * jit/JITInlineMethods.h: 15236 (JSC::JIT::emitFastArithImmToInt): 15237 152382009-11-06 Gavin Barraclough <barraclough@apple.com> 15239 15240 Reviewed by Oliver Hunt. 15241 15242 Tidy up the shift methods on the macro-assembler interface. 15243 15244 Currently behaviour of shifts of a magnitude > 0x1f is undefined. 15245 Instead defined that all shifts are masked to this range. This makes a lot of 15246 practical sense, both since having undefined behaviour is not particularly 15247 desirable, and because this behaviour is commonly required (particularly since 15248 it is required bt ECMA-262 for shifts). 15249 15250 Update the ARM assemblers to provide this behaviour. Remove (now) redundant 15251 masks from JITArithmetic, and remove rshiftPtr (this was used in case that 15252 could be rewritten in a simpler form using rshift32, only optimized JSVALUE32 15253 on x86-64, which uses JSVALUE64!) 15254 15255 * assembler/MacroAssembler.h: 15256 * assembler/MacroAssemblerARM.h: 15257 (JSC::MacroAssemblerARM::lshift32): 15258 (JSC::MacroAssemblerARM::rshift32): 15259 * assembler/MacroAssemblerARMv7.h: 15260 (JSC::MacroAssemblerARMv7::lshift32): 15261 (JSC::MacroAssemblerARMv7::rshift32): 15262 * assembler/MacroAssemblerX86_64.h: 15263 * jit/JITArithmetic.cpp: 15264 (JSC::JIT::emit_op_lshift): 15265 (JSC::JIT::emit_op_rshift): 15266 152672009-11-05 Gavin Barraclough <barraclough@apple.com> 15268 15269 Rubber Stamped by Oliver Hunt. 15270 15271 Remove a magic number (1) from the JIT, instead compute the value with OBJECT_OFFSET. 15272 15273 * jit/JITInlineMethods.h: 15274 (JSC::JIT::emitPutJITStubArg): 15275 (JSC::JIT::emitPutJITStubArgConstant): 15276 (JSC::JIT::emitGetJITStubArg): 15277 (JSC::JIT::emitPutJITStubArgFromVirtualRegister): 15278 * jit/JITStubCall.h: 15279 (JSC::JITStubCall::JITStubCall): 15280 (JSC::JITStubCall::getArgument): 15281 * jit/JITStubs.h: 15282 152832009-11-05 Zoltan Herczeg <zherczeg@inf.u-szeged.hu> 15284 15285 Reviewed by Gavin Barraclough. 15286 15287 https://bugs.webkit.org/show_bug.cgi?id=31159 15288 Fix branchDouble behaviour on ARM THUMB2 JIT. 15289 15290 The x86 branchDouble behaviour is reworked, and all JIT 15291 ports should follow the x86 port. See bug 31104 and 31151 15292 15293 This patch contains a fix for the traditional ARM port 15294 15295 * assembler/ARMAssembler.h: 15296 (JSC::ARMAssembler::): 15297 (JSC::ARMAssembler::fmrs_r): 15298 (JSC::ARMAssembler::ftosid_r): 15299 * assembler/MacroAssemblerARM.h: 15300 (JSC::MacroAssemblerARM::): 15301 (JSC::MacroAssemblerARM::branchDouble): 15302 (JSC::MacroAssemblerARM::branchConvertDoubleToInt32): 15303 153042009-11-05 Chris Jerdonek <chris.jerdonek@gmail.com> 15305 15306 Reviewed by Eric Seidel. 15307 15308 Removed the "this is part of the KDE project" comments from 15309 all *.h, *.cpp, *.idl, and *.pm files. 15310 15311 https://bugs.webkit.org/show_bug.cgi?id=31167 15312 15313 The maintenance and architecture page in the project wiki lists 15314 this as a task. 15315 15316 This change includes no changes or additions to test cases 15317 since the change affects only comments. 15318 15319 * wtf/wince/FastMallocWince.h: 15320 153212009-11-05 Gabor Loki <loki@inf.u-szeged.hu> 15322 15323 Reviewed by Gavin Barraclough. 15324 15325 Use ARMv7 specific encoding for immediate constants on ARMv7 target 15326 https://bugs.webkit.org/show_bug.cgi?id=31060 15327 15328 * assembler/ARMAssembler.cpp: 15329 (JSC::ARMAssembler::getOp2): Use INVALID_IMM 15330 (JSC::ARMAssembler::getImm): Use encodeComplexImm for complex immediate 15331 (JSC::ARMAssembler::moveImm): Ditto. 15332 (JSC::ARMAssembler::encodeComplexImm): Encode a constant by one or two 15333 instructions or a PC relative load. 15334 * assembler/ARMAssembler.h: Use INVALID_IMM if a constant cannot be 15335 encoded as an immediate constant. 15336 (JSC::ARMAssembler::): 15337 (JSC::ARMAssembler::movw_r): 16-bit immediate load 15338 (JSC::ARMAssembler::movt_r): High halfword 16-bit immediate load 15339 (JSC::ARMAssembler::getImm16Op2): Encode immediate constant for 15340 movw_r and mowt_r 15341 153422009-11-04 Mark Mentovai <mark@chromium.org> 15343 15344 Reviewed by Mark Rowe. 15345 15346 Provide TARGETING_TIGER and TARGETING_LEOPARD as analogues to 15347 BUILDING_ON_TIGER and BUILDING_ON_LEOPARD. The TARGETING_ macros 15348 consider the deployment target; the BUILDING_ON_ macros consider the 15349 headers being built against. 15350 15351 * wtf/Platform.h: 15352 153532009-11-04 Gavin Barraclough <barraclough@apple.com> 15354 15355 Reviewed by Oliver Hunt. 15356 15357 https://bugs.webkit.org/show_bug.cgi?id=31151 15358 Fix branchDouble behaviour on ARM THUMB2 JIT. 15359 15360 The ARMv7 JIT is currently using ARMv7Assembler::ConditionEQ to branch 15361 for DoubleEqualOrUnordered, however this is incorrect - ConditionEQ won't 15362 branch on unordered operands. Similarly, DoubleLessThanOrUnordered & 15363 DoubleLessThanOrEqualOrUnordered use ARMv7Assembler::ConditionLO & 15364 ARMv7Assembler::ConditionLS, whereas they should be using 15365 ARMv7Assembler::ConditionLT & ARMv7Assembler::ConditionLE. 15366 15367 Fix these, and fill out the missing DoubleConditions. 15368 15369 * assembler/MacroAssemblerARMv7.h: 15370 (JSC::MacroAssemblerARMv7::): 15371 (JSC::MacroAssemblerARMv7::branchDouble): 15372 153732009-11-04 Gavin Barraclough <barraclough@apple.com> 15374 15375 Rubber Stamped by Oliver Hunt. 15376 15377 Enable native call optimizations on ARMv7. (Existing ARM_TRADITIONAL 15378 implementation was generic, worked perfectly, just needed turning on). 15379 15380 * jit/JITOpcodes.cpp: 15381 * wtf/Platform.h: 15382 153832009-11-04 Gavin Barraclough <barraclough@apple.com> 15384 15385 Rubber Stamped by Mark Rowe, Oliver Hunt, and Sam Weinig. 15386 15387 Add a missing assert to the ARMv7 JIT. 15388 15389 * assembler/ARMv7Assembler.h: 15390 (JSC::ARMThumbImmediate::ARMThumbImmediate): 15391 153922009-11-04 Mark Rowe <mrowe@apple.com> 15393 15394 Rubber-stamped by Oliver Hunt. 15395 15396 Remove bogus op_ prefix on dumped version of three opcodes. 15397 15398 * bytecode/CodeBlock.cpp: 15399 (JSC::CodeBlock::dump): 15400 154012009-11-04 Mark Rowe <mrowe@apple.com> 15402 15403 Reviewed by Sam Weinig. 15404 15405 Fix dumping of constants in bytecode so that they aren't printed as large positive register numbers. 15406 15407 We do this by having the registerName function return information about the constant if the register 15408 number corresponds to a constant. This requires that registerName, and several functions that call it, 15409 be converted to member functions of CodeBlock so that the constant value can be retrieved. The 15410 ExecState also needs to be threaded down through these functions so that it can be passed on to 15411 constantName when needed. 15412 15413 * bytecode/CodeBlock.cpp: 15414 (JSC::constantName): 15415 (JSC::CodeBlock::registerName): 15416 (JSC::CodeBlock::printUnaryOp): 15417 (JSC::CodeBlock::printBinaryOp): 15418 (JSC::CodeBlock::printConditionalJump): 15419 (JSC::CodeBlock::printGetByIdOp): 15420 (JSC::CodeBlock::printPutByIdOp): 15421 (JSC::CodeBlock::dump): 15422 * bytecode/CodeBlock.h: 15423 (JSC::CodeBlock::isConstantRegisterIndex): 15424 154252009-11-04 Pavel Heimlich <tropikhajma@gmail.com> 15426 15427 Reviewed by Alexey Proskuryakov. 15428 15429 https://bugs.webkit.org/show_bug.cgi?id=30647 15430 Solaris build failure due to strnstr. 15431 15432 * wtf/StringExtras.h: Enable strnstr on Solaris, too. 15433 154342009-11-04 Gavin Barraclough <barraclough@apple.com> 15435 15436 Reviewed by Oliver Hunt. 15437 15438 https://bugs.webkit.org/show_bug.cgi?id=31104 15439 Refactor x86-specific behaviour out of the JIT. 15440 15441 - Add explicit double branch conditions for ordered and unordered comparisons (presently the brehaviour is a mix). 15442 - Refactor double to int conversion out into the MacroAssembler. 15443 - Remove broken double to int conversion for !JSVALUE32_64 builds - this code was broken and slowing us down, fixing it showed it not to be an improvement. 15444 - Remove exclusion of double to int conversion from (1 % X) cases in JSVALUE32_64 builds - if this was of benefit this is no longer the case; simplify. 15445 15446 * assembler/MacroAssemblerARM.h: 15447 (JSC::MacroAssemblerARM::): 15448 * assembler/MacroAssemblerARMv7.h: 15449 (JSC::MacroAssemblerARMv7::): 15450 * assembler/MacroAssemblerX86Common.h: 15451 (JSC::MacroAssemblerX86Common::): 15452 (JSC::MacroAssemblerX86Common::convertInt32ToDouble): 15453 (JSC::MacroAssemblerX86Common::branchDouble): 15454 (JSC::MacroAssemblerX86Common::branchConvertDoubleToInt32): 15455 * jit/JITArithmetic.cpp: 15456 (JSC::JIT::emitBinaryDoubleOp): 15457 (JSC::JIT::emit_op_div): 15458 (JSC::JIT::emitSlow_op_jnless): 15459 (JSC::JIT::emitSlow_op_jnlesseq): 15460 * jit/JITOpcodes.cpp: 15461 (JSC::JIT::emit_op_jfalse): 15462 154632009-11-04 Mark Mentovai <mark@chromium.org> 15464 15465 Reviewed by Eric Seidel. 15466 15467 Remove BUILDING_ON_LEOPARD from JavaScriptCore.gyp. This is supposed 15468 to be set as needed only in wtf/Platform.h. 15469 15470 * JavaScriptCore.gyp/JavaScriptCore.gyp: 15471 154722009-11-02 Oliver Hunt <oliver@apple.com> 15473 15474 Reviewed by Gavin Barraclough. 15475 15476 REGRESSION (r48573): JSC may incorrectly cache chain lookups with a dictionary at the head of the chain 15477 https://bugs.webkit.org/show_bug.cgi?id=31045 15478 15479 Add guards to prevent caching of prototype chain lookups with dictionaries at the 15480 head of the chain. Also add a few tighter assertions to cached prototype lookups 15481 to catch this in future. 15482 15483 * interpreter/Interpreter.cpp: 15484 (JSC::Interpreter::tryCacheGetByID): 15485 (JSC::Interpreter::privateExecute): 15486 * jit/JITStubs.cpp: 15487 (JSC::JITThunks::tryCacheGetByID): 15488 154892009-11-02 Laszlo Gombos <laszlo.1.gombos@nokia.com> 15490 15491 Reviewed by Darin Adler. 15492 15493 PLATFORM(CF) should be set when building for Qt on Darwin 15494 https://bugs.webkit.org/show_bug.cgi?id=23671 15495 15496 * wtf/Platform.h: Turn on CF support if both QT and DARWIN 15497 platforms are defined. 15498 154992009-11-02 Dmitry Titov <dimich@chromium.org> 15500 15501 Reviewed by David Levin. 15502 15503 Remove threadsafe refcounting from tasks used with WTF::MessageQueue. 15504 https://bugs.webkit.org/show_bug.cgi?id=30612 15505 15506 * wtf/MessageQueue.h: 15507 (WTF::MessageQueue::alwaysTruePredicate): 15508 (WTF::MessageQueue::~MessageQueue): 15509 (WTF::MessageQueue::append): 15510 (WTF::MessageQueue::appendAndCheckEmpty): 15511 (WTF::MessageQueue::prepend): 15512 (WTF::MessageQueue::waitForMessage): 15513 (WTF::MessageQueue::waitForMessageFilteredWithTimeout): 15514 (WTF::MessageQueue::tryGetMessage): 15515 (WTF::MessageQueue::removeIf): 15516 The MessageQueue is changed to act as a queue of OwnPtr<DataType>. It takes ownership 15517 of posted tasks and passes it to the new owner (in another thread) when the task is fetched. 15518 All methods have arguments of type PassOwnPtr<DataType> and return the same type. 15519 15520 * wtf/Threading.cpp: 15521 (WTF::createThread): 15522 Superficial change to trigger rebuild of JSC project on Windows, 15523 workaround for https://bugs.webkit.org/show_bug.cgi?id=30890 15524 155252009-10-30 Geoffrey Garen <ggaren@apple.com> 15526 15527 Reviewed by Oliver Hunt. 15528 15529 Fixed failing layout test: restore a special case I accidentally deleted. 15530 15531 * runtime/DatePrototype.cpp: 15532 (JSC::setNewValueFromDateArgs): In the case of applying a change to a date 15533 that is NaN, reset the date to 0 *and* then apply the change; don't just 15534 reset the date to 0. 15535 155362009-10-30 Geoffrey Garen <ggaren@apple.com> 15537 15538 Windows build fix: update for object-to-pointer change. 15539 15540 * runtime/DatePrototype.cpp: 15541 (JSC::formatLocaleDate): 15542 155432009-10-29 Geoffrey Garen <ggaren@apple.com> 15544 15545 Reviewed by Darin Adler. 15546 15547 https://bugs.webkit.org/show_bug.cgi?id=30942 15548 Use pointers instead of copies to pass GregorianDateTime objects around. 15549 15550 SunSpider reports a shocking 4.5% speedup on date-format-xparb, and 1.3% 15551 speedup on date-format-tofte. 15552 15553 * runtime/DateInstance.cpp: 15554 (JSC::DateInstance::gregorianDateTime): 15555 * runtime/DateInstance.h: 15556 * runtime/DatePrototype.cpp: 15557 (JSC::formatLocaleDate): 15558 (JSC::dateProtoFuncToString): 15559 (JSC::dateProtoFuncToUTCString): 15560 (JSC::dateProtoFuncToISOString): 15561 (JSC::dateProtoFuncToDateString): 15562 (JSC::dateProtoFuncToTimeString): 15563 (JSC::dateProtoFuncGetFullYear): 15564 (JSC::dateProtoFuncGetUTCFullYear): 15565 (JSC::dateProtoFuncToGMTString): 15566 (JSC::dateProtoFuncGetMonth): 15567 (JSC::dateProtoFuncGetUTCMonth): 15568 (JSC::dateProtoFuncGetDate): 15569 (JSC::dateProtoFuncGetUTCDate): 15570 (JSC::dateProtoFuncGetDay): 15571 (JSC::dateProtoFuncGetUTCDay): 15572 (JSC::dateProtoFuncGetHours): 15573 (JSC::dateProtoFuncGetUTCHours): 15574 (JSC::dateProtoFuncGetMinutes): 15575 (JSC::dateProtoFuncGetUTCMinutes): 15576 (JSC::dateProtoFuncGetSeconds): 15577 (JSC::dateProtoFuncGetUTCSeconds): 15578 (JSC::dateProtoFuncGetTimezoneOffset): 15579 (JSC::setNewValueFromTimeArgs): 15580 (JSC::setNewValueFromDateArgs): 15581 (JSC::dateProtoFuncSetYear): 15582 (JSC::dateProtoFuncGetYear): Renamed getGregorianDateTime to gregorianDateTime, 15583 since it no longer has an out parameter. Uses 0 to indicate invalid dates. 15584 155852009-10-30 Zoltan Horvath <zoltan@webkit.org> 15586 15587 Reviewed by Darin Adler. 15588 15589 Allow custom memory allocation control for JavaScriptCore's ListHashSet 15590 https://bugs.webkit.org/show_bug.cgi?id=30853 15591 15592 Inherits ListHashSet class from FastAllocBase because it is 15593 instantiated by 'new' in WebCore/rendering/RenderBlock.cpp:1813. 15594 15595 * wtf/ListHashSet.h: 15596 155972009-10-30 Oliver Hunt <oliver@apple.com> 15598 15599 Reviewed by Gavin Barraclough. 15600 15601 Regression: crash enumerating properties of an object with getters or setters 15602 https://bugs.webkit.org/show_bug.cgi?id=30948 15603 15604 Add a guard to prevent us trying to cache property enumeration on 15605 objects with getters or setters. 15606 15607 * runtime/JSPropertyNameIterator.cpp: 15608 (JSC::JSPropertyNameIterator::create): 15609 156102009-10-30 Roland Steiner <rolandsteiner@chromium.org> 15611 15612 Reviewed by Eric Seidel. 15613 15614 Remove ENABLE_RUBY guards as discussed with Dave Hyatt and Maciej Stachowiak. 15615 15616 Bug 28420 - Implement HTML5 <ruby> rendering 15617 (https://bugs.webkit.org/show_bug.cgi?id=28420) 15618 15619 No new tests (no functional change). 15620 15621 * Configurations/FeatureDefines.xcconfig: 15622 156232009-10-29 Oliver Hunt <oliver@apple.com> 15624 15625 Reviewed by Maciej Stachowiak. 15626 15627 REGRESSION (r50218-r50262): E*TRADE accounts page is missing content 15628 https://bugs.webkit.org/show_bug.cgi?id=30947 15629 <rdar://problem/7348833> 15630 15631 The logic for flagging that a structure has non-enumerable properties 15632 was in addPropertyWithoutTransition, rather than in the core Structure::put 15633 method. Despite this I was unable to produce a testcase that caused 15634 the failure that etrade was experiencing, but the new assertion in 15635 getEnumerablePropertyNames triggers on numerous layout tests without 15636 the fix, so in effecti all for..in enumeration in any test ends up 15637 doing the required consistency check. 15638 15639 * runtime/Structure.cpp: 15640 (JSC::Structure::addPropertyWithoutTransition): 15641 (JSC::Structure::put): 15642 (JSC::Structure::getEnumerablePropertyNames): 15643 (JSC::Structure::checkConsistency): 15644 156452009-10-29 Gabor Loki <loki@inf.u-szeged.hu> 15646 15647 Reviewed by Gavin Barraclough. 15648 15649 Add cacheFlush support for Thumb-2 on Linux 15650 https://bugs.webkit.org/show_bug.cgi?id=30865 15651 15652 * jit/ExecutableAllocator.h: 15653 (JSC::ExecutableAllocator::cacheFlush): 15654 156552009-10-28 Gavin Barraclough <barraclough@apple.com> 15656 15657 Reviewed by Oliver Hunt. 15658 15659 JSC JIT on ARMv7 cannot link jumps >16Mb range 15660 https://bugs.webkit.org/show_bug.cgi?id=30891 15661 15662 Start planing all relative jumps as move-32-bit-immediate-to-register-BX. 15663 In the cases where the jump would fall within a relative jump range, use a relative jump. 15664 15665 * JavaScriptCore.xcodeproj/project.pbxproj: 15666 * assembler/ARMv7Assembler.h: 15667 (JSC::ARMv7Assembler::~ARMv7Assembler): 15668 (JSC::ARMv7Assembler::LinkRecord::LinkRecord): 15669 (JSC::ARMv7Assembler::): 15670 (JSC::ARMv7Assembler::executableCopy): 15671 (JSC::ARMv7Assembler::linkJump): 15672 (JSC::ARMv7Assembler::relinkJump): 15673 (JSC::ARMv7Assembler::setInt32): 15674 (JSC::ARMv7Assembler::isB): 15675 (JSC::ARMv7Assembler::isBX): 15676 (JSC::ARMv7Assembler::isMOV_imm_T3): 15677 (JSC::ARMv7Assembler::isMOVT): 15678 (JSC::ARMv7Assembler::isNOP_T1): 15679 (JSC::ARMv7Assembler::isNOP_T2): 15680 (JSC::ARMv7Assembler::linkJumpAbsolute): 15681 (JSC::ARMv7Assembler::twoWordOp5i6Imm4Reg4EncodedImmFirst): 15682 (JSC::ARMv7Assembler::twoWordOp5i6Imm4Reg4EncodedImmSecond): 15683 (JSC::ARMv7Assembler::ARMInstructionFormatter::twoWordOp5i6Imm4Reg4EncodedImm): 15684 * assembler/MacroAssemblerARMv7.h: 15685 (JSC::MacroAssemblerARMv7::makeJump): 15686 (JSC::MacroAssemblerARMv7::makeBranch): 15687 * jit/JIT.h: 15688 * wtf/Platform.h: 15689 156902009-10-28 Oliver Hunt <oliver@apple.com> 15691 15692 Reviewed by Geoff Garen. 15693 15694 Improve for..in enumeration performance 15695 https://bugs.webkit.org/show_bug.cgi?id=30887 15696 15697 Improve indexing of an object with a for..in iterator by 15698 identifying cases where get_by_val is being used with a iterator 15699 as the subscript and replace it with a new get_by_pname 15700 bytecode. get_by_pname then optimizes lookups that directly access 15701 the base object. 15702 15703 * bytecode/CodeBlock.cpp: 15704 (JSC::CodeBlock::dump): 15705 * bytecode/Opcode.h: 15706 * bytecompiler/BytecodeGenerator.cpp: 15707 (JSC::BytecodeGenerator::emitGetByVal): 15708 * bytecompiler/BytecodeGenerator.h: 15709 (JSC::BytecodeGenerator::pushOptimisedForIn): 15710 (JSC::BytecodeGenerator::popOptimisedForIn): 15711 * interpreter/Interpreter.cpp: 15712 (JSC::Interpreter::privateExecute): 15713 * jit/JIT.cpp: 15714 (JSC::JIT::privateCompileMainPass): 15715 (JSC::JIT::privateCompileSlowCases): 15716 * jit/JIT.h: 15717 * jit/JITPropertyAccess.cpp: 15718 (JSC::JIT::compileGetDirectOffset): 15719 (JSC::JIT::emit_op_get_by_pname): 15720 (JSC::JIT::emitSlow_op_get_by_pname): 15721 * parser/Nodes.cpp: 15722 (JSC::ForInNode::emitBytecode): 15723 * runtime/JSObject.h: 15724 * runtime/JSPropertyNameIterator.cpp: 15725 (JSC::JSPropertyNameIterator::create): 15726 * runtime/JSPropertyNameIterator.h: 15727 (JSC::JSPropertyNameIterator::getOffset): 15728 (JSC::JSPropertyNameIterator::JSPropertyNameIterator): 15729 * runtime/JSValue.h: 15730 (JSC::JSValue::): 15731 * runtime/Structure.cpp: 15732 (JSC::Structure::addPropertyTransition): 15733 (JSC::Structure::changePrototypeTransition): 15734 (JSC::Structure::despecifyFunctionTransition): 15735 (JSC::Structure::addAnonymousSlotsTransition): 15736 (JSC::Structure::getterSetterTransition): 15737 (JSC::Structure::toDictionaryTransition): 15738 (JSC::Structure::addPropertyWithoutTransition): 15739 Track the existence (or not) of non-enumerable properties. 15740 * runtime/Structure.h: 15741 (JSC::Structure::propertyStorageCapacity): 15742 (JSC::Structure::propertyStorageSize): 15743 (JSC::Structure::hasNonEnumerableProperties): 15744 (JSC::Structure::hasAnonymousSlots): 15745 157462009-10-28 Dmitry Titov <dimich@chromium.org> 15747 15748 Not reviewed, attemp to fix Windows build. 15749 15750 Touch the cpp file to cause recompile. 15751 15752 * wtf/Threading.cpp: 15753 (WTF::threadEntryPoint): 15754 157552009-10-28 Dmitry Titov <dimich@chromium.org> 15756 15757 Reviewed by David Levin. 15758 15759 https://bugs.webkit.org/show_bug.cgi?id=30805 15760 Add MessageQueue::removeIf(Predicate&) to remove certain tasks without pulling them from the queue. 15761 Existing Database tests cover this since Database removes tasks when it is stopped. 15762 15763 * wtf/MessageQueue.h: 15764 (WTF::::removeIf): 15765 157662009-10-28 Afonso R. Costa Jr. <afonso.costa@openbossa.org> 15767 15768 Reviewed by Oliver Hunt. 15769 15770 [Qt] Enable YARR when YARR_JIT is enabled 15771 https://bugs.webkit.org/show_bug.cgi?id=30730 15772 15773 When enabling or disabling JIT using JAVASCRIPTCORE_JIT, the ENABLE_YARR should 15774 be toggled also. 15775 15776 * JavaScriptCore.pri: 15777 157782009-10-24 Martin Robinson <martin.james.robinson@gmail.com> 15779 15780 Reviewed by Oliver Hunt. 15781 15782 Fix strict aliasing warning by switching reinterpret_cast to bitwise_cast. 15783 15784 strict-aliasing warnings in JSFunction.h 15785 https://bugs.webkit.org/show_bug.cgi?id=27869 15786 15787 * runtime/JSFunction.h: 15788 (JSC::JSFunction::nativeFunction): 15789 (JSC::JSFunction::scopeChain): 15790 (JSC::JSFunction::setScopeChain): 15791 (JSC::JSFunction::setNativeFunction): 15792 157932009-10-28 Jan-Arve Sæther <jan-arve.saether@nokia.com> 15794 15795 Reviewed by Tor Arne Vestbø. 15796 15797 Build-fix for 64-bit Windows 15798 15799 * wtf/Platform.h: Make sure to use WTF_USE_JSVALUE64 15800 158012009-10-28 Gavin Barraclough <barraclough@apple.com> 15802 15803 Reviewed by NOBODY (build fix!). 15804 15805 * jit/JIT.h: 15806 158072009-10-26 Holger Hans Peter Freyther <zecke@selfish.org> 15808 15809 Rubber-stamped by Darin Adler. 15810 15811 Export fastMalloc, fastCalloc, fastRealloc and fastFree on GCC/Unix 15812 https://bugs.webkit.org/show_bug.cgi?id=30769 15813 15814 When using -fvisibility=hidden to hide all internal symbols by default 15815 the malloc symbols will be hidden as well. For memory instrumentation 15816 it is needed to provide an instrumented version of these symbols and 15817 override the normal routines and by changing the visibility back to 15818 default this becomes possible. 15819 15820 The only other solution would be to use system malloc instead of the 15821 TCmalloc implementation but this will not allow to analyze memory 15822 behavior with the default allocator. 15823 15824 * wtf/FastMalloc.h: Define WTF_FAST_MALLOC_EXPORT for GCC and !darwin 15825 158262009-10-27 Gavin Barraclough <barraclough@apple.com> 15827 15828 Rubber Stamped by Samuel Q. Weinig. 15829 15830 Make the asserts protecting the offsets in the JIT more descriptive. 15831 15832 * jit/JIT.h: 15833 * jit/JITCall.cpp: 15834 (JSC::JIT::compileOpCall): 15835 * jit/JITPropertyAccess.cpp: 15836 (JSC::JIT::emit_op_method_check): 15837 (JSC::JIT::compileGetByIdHotPath): 15838 (JSC::JIT::compileGetByIdSlowCase): 15839 (JSC::JIT::emit_op_put_by_id): 15840 158412009-10-27 Geoffrey Garen <ggaren@apple.com> 15842 15843 Reviewed by Sam Weinig. 15844 15845 A little bit of refactoring in the date code. 15846 15847 * JavaScriptCore.exp: Don't export this unused symbol. 15848 15849 * runtime/DateConstructor.cpp: 15850 (JSC::constructDate): 15851 15852 * runtime/DateInstance.cpp: 15853 (JSC::DateInstance::DateInstance): 15854 * runtime/DateInstance.h: Removed some unused functions. Changed the default 15855 constructor to ensure that a DateInstance is always initialized. 15856 15857 * runtime/DatePrototype.cpp: 15858 (JSC::DatePrototype::DatePrototype): Pass an initializer to our constructor, 15859 since it now requires one. 15860 15861 * wtf/DateMath.cpp: 15862 (WTF::msToGregorianDateTime): Only compute our offset from UTC if our 15863 output will require it. Otherwise, our offset is 0. 15864 158652009-10-27 Geoffrey Garen <ggaren@apple.com> 15866 15867 Build fix: Mark DateInstaceCache.h private, so other frameworks can see it. 15868 15869 * JavaScriptCore.xcodeproj/project.pbxproj: 15870 158712009-10-27 Geoffrey Garen <ggaren@apple.com> 15872 15873 Build fix: re-readded this file. 15874 15875 * runtime/DateInstanceCache.h: Added. 15876 (JSC::DateInstanceData::create): 15877 (JSC::DateInstanceData::DateInstanceData): 15878 (JSC::DateInstanceCache::DateInstanceCache): 15879 (JSC::DateInstanceCache::add): 15880 (JSC::DateInstanceCache::lookup): 15881 158822009-10-27 Geoffrey Garen <ggaren@apple.com> 15883 15884 Reviewed by Darin Adler and Oliver Hunt. 15885 15886 https://bugs.webkit.org/show_bug.cgi?id=30800 15887 Cache recently computed date data. 15888 15889 SunSpider reports a ~0.5% speedup, mostly from date-format-tofte.js. 15890 15891 * GNUmakefile.am: 15892 * JavaScriptCore.gypi: 15893 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: 15894 * JavaScriptCore.xcodeproj/project.pbxproj: Added new file. 15895 15896 * runtime/DateInstance.cpp: 15897 (JSC::DateInstance::DateInstance): 15898 (JSC::DateInstance::getGregorianDateTime): Use the shared cache. 15899 15900 * runtime/DateInstance.h: Renamed m_cache to m_data, to avoid the confusion 15901 of a "cache cache". 15902 15903 * runtime/DatePrototype.cpp: 15904 (JSC::formatLocaleDate): 15905 (JSC::dateProtoFuncToString): 15906 (JSC::dateProtoFuncToUTCString): 15907 (JSC::dateProtoFuncToISOString): 15908 (JSC::dateProtoFuncToDateString): 15909 (JSC::dateProtoFuncToTimeString): 15910 (JSC::dateProtoFuncGetFullYear): 15911 (JSC::dateProtoFuncGetUTCFullYear): 15912 (JSC::dateProtoFuncToGMTString): 15913 (JSC::dateProtoFuncGetMonth): 15914 (JSC::dateProtoFuncGetUTCMonth): 15915 (JSC::dateProtoFuncGetDate): 15916 (JSC::dateProtoFuncGetUTCDate): 15917 (JSC::dateProtoFuncGetDay): 15918 (JSC::dateProtoFuncGetUTCDay): 15919 (JSC::dateProtoFuncGetHours): 15920 (JSC::dateProtoFuncGetUTCHours): 15921 (JSC::dateProtoFuncGetMinutes): 15922 (JSC::dateProtoFuncGetUTCMinutes): 15923 (JSC::dateProtoFuncGetSeconds): 15924 (JSC::dateProtoFuncGetUTCSeconds): 15925 (JSC::dateProtoFuncGetTimezoneOffset): 15926 (JSC::setNewValueFromTimeArgs): 15927 (JSC::setNewValueFromDateArgs): 15928 (JSC::dateProtoFuncSetYear): 15929 (JSC::dateProtoFuncGetYear): Pass an ExecState to these functions, so they 15930 can access the DateInstanceCache. 15931 15932 * runtime/JSGlobalData.h: Keep a DateInstanceCache. 15933 159342009-10-27 James Robinson <jamesr@chromium.org> 15935 15936 Reviewed by Darin Fisher. 15937 15938 Ensures that JavaScriptCore/wtf/CurrentTime.cpp is not built in PLATFORM(CHROMIUM) builds. 15939 15940 Chromium uses a different method to calculate the current time than is used in 15941 JavaScriptCore/wtf/CurrentTime.cpp. This can lead to time skew when calls to currentTime() and Chromium's time 15942 function are mixed. In particular, timers can get scheduled in the past which leads to 100% CPU use. 15943 See http://code.google.com/p/chromium/issues/detail?id=25892 for an example. 15944 15945 https://bugs.webkit.org/show_bug.cgi?id=30833 15946 15947 * JavaScriptCore.gyp/JavaScriptCore.gyp: 15948 * wtf/CurrentTime.cpp: 15949 159502009-10-27 Peter Varga <pvarga@inf.u-szeged.hu> 15951 15952 Rubber-stamped by Tor Arne Vestbø. 15953 15954 Fix typo in RegexInterpreter.cpp and RegexJIT.cpp alterantive to 15955 alternative. 15956 15957 * yarr/RegexInterpreter.cpp: 15958 (JSC::Yarr::ByteCompiler::alternativeBodyDisjunction): 15959 (JSC::Yarr::ByteCompiler::alternativeDisjunction): 15960 (JSC::Yarr::ByteCompiler::emitDisjunction): 15961 * yarr/RegexJIT.cpp: 15962 (JSC::Yarr::RegexGenerator::generateDisjunction): 15963 159642009-10-26 Laszlo Gombos <laszlo.1.gombos@nokia.com> 15965 15966 Reviewed by Darin Adler. 15967 15968 Make .rc files compile on Windows without depending on MFC headers 15969 https://bugs.webkit.org/show_bug.cgi?id=30750 15970 15971 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.rc: Use 15972 winresrc.h because it exists even when MFC is not installed, and is 15973 all that's needed here. 15974 159752009-10-26 Gabor Loki <loki@inf.u-szeged.hu> 15976 15977 Reviewed by Gavin Barraclough. 15978 15979 The thunkReturnAddress is on JITStackFrame on ARM JIT as well 15980 https://bugs.webkit.org/show_bug.cgi?id=30782 15981 15982 Move the thunkReturnAddress from top of the stack into the JITStackFrame 15983 structure. This is a requirement for JSValue32_64 support on ARM. 15984 15985 * assembler/MacroAssemblerARM.h: 15986 (JSC::MacroAssemblerARM::ret): Return with link register 15987 (JSC::MacroAssemblerARM::prepareCall): Store the return address in link register 15988 * jit/JIT.h: Remove unused ctiReturnRegister 15989 * jit/JITInlineMethods.h: Same as ARMv7 15990 (JSC::JIT::restoreArgumentReference): Ditto. 15991 (JSC::JIT::restoreArgumentReferenceForTrampoline): Ditto. 15992 * jit/JITOpcodes.cpp: Remove ctiReturnRegister related instruction 15993 * jit/JITStubs.cpp: Store thunkReturnAddress on JITStackFrame. Use 15994 small trampoline functions which handle return addresses for each 15995 CTI_STUB_FUNCTION. 15996 * jit/JITStubs.h: Store thunkReturnAddress on JITStackFrame 15997 (JSC::JITStackFrame::returnAddressSlot): Return with the address of thunkReturnAddress 15998 * yarr/RegexJIT.cpp: 15999 (JSC::Yarr::RegexGenerator::generateEnter): Remove the unnecessary instruction 16000 160012009-10-26 Steve Block <steveblock@google.com> 16002 16003 Reviewed by Darin Adler. 16004 16005 Adds ability to disable ReadWriteLock on platforms (eg Android) that use pthreads but do not support pthread_rwlock. 16006 https://bugs.webkit.org/show_bug.cgi?id=30713 16007 16008 * wtf/Platform.h: Modified. Defines HAVE_PTHREAD_RWLOCK for all platforms currently using pthreads. 16009 * wtf/Threading.h: Modified. Use pthread_rwlock_t only when HAVE_PTHREAD_RWLOCK is defined. 16010 * wtf/ThreadingPthreads.cpp: Modified. Build ReadWriteLock methods only when HAVE_PTHREAD_RWLOCK is defined. 16011 160122009-10-24 Laszlo Gombos <laszlo.1.gombos@nokia.com> 16013 16014 Reviewed by Holger Freyther. 16015 16016 [Qt] [Symbian] Set the capability and memory required to run QtWebKit for Symbian 16017 https://bugs.webkit.org/show_bug.cgi?id=30476 16018 16019 Assign ReadUserData WriteUserData NetworkServices Symbian capabilities 16020 to jsc.exe. 16021 16022 * jsc.pro: 16023 160242009-10-23 Steve Block <steveblock@google.com> 16025 16026 Reviewed by Dmitry Titov. 16027 16028 Fixes a leak in createThreadInternal on Android. 16029 https://bugs.webkit.org/show_bug.cgi?id=30698 16030 16031 * wtf/ThreadingPthreads.cpp: Modified. 16032 (WTF::createThreadInternal): Avoid leaking a ThreadData object on failure. 16033 160342009-10-22 Geoffrey Garen <ggaren@apple.com> 16035 16036 Reviewed by Alexey Proskuryakov. 16037 16038 Fixed ASSERT when opening Safari's Caches window while the Web Inspector 16039 is open. 16040 16041 * runtime/Collector.cpp: 16042 (JSC::typeName): Added two new types to the type name list in the Collector. 16043 These types have been around for a while, but nobody remembered to consider them here. 16044 16045 * runtime/JSCell.h: 16046 (JSC::JSCell::isPropertyNameIterator): 16047 * runtime/JSPropertyNameIterator.h: 16048 (JSC::JSPropertyNameIterator::isPropertyNameIterator): Give the Collector 16049 a way to tell if a cell is a JSPropertyNameIterator. 16050 160512009-10-22 Steve Falkenburg <sfalken@apple.com> 16052 16053 Reviewed by Jon Honeycutt. 16054 16055 https://bugs.webkit.org/show_bug.cgi?id=30686 16056 Remove debug-specific def file. 16057 Only Debug_All target uses JavaScriptCore_debug.dll naming, and since 16058 that target is only used internally, maintaining two files just to 16059 suppress a single link warning isn't worthwhile. 16060 16061 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCoreCommon.vsprops: 16062 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: Removed. 16063 160642009-10-21 Jon Honeycutt <jhoneycutt@apple.com> 16065 16066 <rdar://problem/7270320> Screenshots of off-screen plug-ins are blank 16067 <rdar://problem/7270314> After halting a transparent PluginView on 16068 Windows, the transparency is applied twice 16069 16070 Reviewed by Dan Bernstein. 16071 16072 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 16073 Export WTF::deleteOwnedPtr(HDC). 16074 16075 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 16076 Ditto. 16077 160782009-10-20 Geoffrey Garen <ggaren@apple.com> 16079 16080 Windows build fix: updated variable name. 16081 16082 * runtime/DatePrototype.cpp: 16083 (JSC::formatLocaleDate): 16084 160852009-10-20 Geoffrey Garen <ggaren@apple.com> 16086 16087 Reviewed by Mark Rowe. 16088 16089 * jit/JITOpcodes.cpp: 16090 (JSC::JIT::emit_op_next_pname): Slightly tweaked this #ifdef to match the 16091 size of a JSValue because m_jsStrings is an array of JSValues. 16092 160932009-10-20 Geoffrey Garen <ggaren@apple.com> 16094 16095 Reviewed by Mark Rowe. 16096 16097 Fixed a 64-bit regression caused by the fix for 16098 https://bugs.webkit.org/show_bug.cgi?id=30570. 16099 16100 * jit/JITOpcodes.cpp: 16101 (JSC::JIT::emit_op_next_pname): Use TimesEight stepping on 64-bit, since 16102 64-bit pointers are eight bytes long. 16103 161042009-10-20 Geoffrey Garen <ggaren@apple.com> 16105 16106 Reviewed by Sam Weinig. 16107 16108 Refactored DateInstance::msToGregorianDateTime so that a DateInstance's 16109 caller doesn't need to supply the DateInstance's own internal value to 16110 the DateInstance. 16111 16112 * runtime/DateInstance.cpp: 16113 (JSC::DateInstance::getGregorianDateTime): Renamed from "msToGregorianDateTime". 16114 16115 * runtime/DateInstance.h: 16116 * runtime/DatePrototype.cpp: 16117 (JSC::formatLocaleDate): 16118 (JSC::dateProtoFuncToString): 16119 (JSC::dateProtoFuncToUTCString): 16120 (JSC::dateProtoFuncToISOString): 16121 (JSC::dateProtoFuncToDateString): 16122 (JSC::dateProtoFuncToTimeString): 16123 (JSC::dateProtoFuncToLocaleString): 16124 (JSC::dateProtoFuncToLocaleDateString): 16125 (JSC::dateProtoFuncToLocaleTimeString): 16126 (JSC::dateProtoFuncGetTime): 16127 (JSC::dateProtoFuncGetFullYear): 16128 (JSC::dateProtoFuncGetUTCFullYear): 16129 (JSC::dateProtoFuncToGMTString): 16130 (JSC::dateProtoFuncGetMonth): 16131 (JSC::dateProtoFuncGetUTCMonth): 16132 (JSC::dateProtoFuncGetDate): 16133 (JSC::dateProtoFuncGetUTCDate): 16134 (JSC::dateProtoFuncGetDay): 16135 (JSC::dateProtoFuncGetUTCDay): 16136 (JSC::dateProtoFuncGetHours): 16137 (JSC::dateProtoFuncGetUTCHours): 16138 (JSC::dateProtoFuncGetMinutes): 16139 (JSC::dateProtoFuncGetUTCMinutes): 16140 (JSC::dateProtoFuncGetSeconds): 16141 (JSC::dateProtoFuncGetUTCSeconds): 16142 (JSC::dateProtoFuncGetTimezoneOffset): 16143 (JSC::setNewValueFromTimeArgs): 16144 (JSC::setNewValueFromDateArgs): 16145 (JSC::dateProtoFuncSetYear): 16146 (JSC::dateProtoFuncGetYear): Also renamed "utc" to "outputIsUTC", for clarity. 16147 161482009-10-20 Gabor Loki <loki@inf.u-szeged.hu> 16149 16150 Reviewed by Geoffrey Garen. 16151 16152 The op_next_pname should use 4 bytes addressing mode in case of JSValue32 16153 https://bugs.webkit.org/show_bug.cgi?id=30570 16154 16155 * jit/JITOpcodes.cpp: 16156 (JSC::JIT::emit_op_next_pname): 16157 161582009-10-20 Gabor Loki <loki@inf.u-szeged.hu> 16159 16160 Reviewed by Oliver Hunt. 16161 16162 Move OverridesMarkChildren flag from DatePrototype to its parent class 16163 https://bugs.webkit.org/show_bug.cgi?id=30372 16164 16165 * runtime/DateInstance.h: 16166 (JSC::DateInstance::createStructure): 16167 * runtime/DatePrototype.h: 16168 161692009-10-19 Geoffrey Garen <ggaren@apple.com> 16170 16171 Reviewed by Oliver Hunt. 16172 16173 Tightened up some put_by_id_transition code generation. 16174 https://bugs.webkit.org/show_bug.cgi?id=30539 16175 16176 * jit/JIT.h: 16177 * jit/JITPropertyAccess.cpp: 16178 (JSC::JIT::testPrototype): 16179 (JSC::JIT::privateCompilePutByIdTransition): No need to do object type 16180 checks or read Structures and prototypes from objects: they're all known 16181 constants at compile time. 16182 161832009-10-19 Geoffrey Garen <ggaren@apple.com> 16184 16185 Reviewed by Sam Weinig. 16186 16187 Added a private API for getting a global context from a context, for 16188 clients who want to preserve a context for a later callback. 16189 16190 * API/APICast.h: 16191 (toGlobalRef): Added an ASSERT, since this function is used more often 16192 than before. 16193 16194 * API/JSContextRef.cpp: 16195 * API/JSContextRefPrivate.h: Added. The new API. 16196 16197 * API/tests/testapi.c: 16198 (print_callAsFunction): 16199 (main): Test the new API. 16200 16201 * JavaScriptCore.exp: 16202 * JavaScriptCore.xcodeproj/project.pbxproj: Build and export the new API. 16203 162042009-10-17 Geoffrey Garen <ggaren@apple.com> 16205 16206 Reviewed by Oliver Hunt. 16207 16208 Tightened up some instanceof code generation. 16209 https://bugs.webkit.org/show_bug.cgi?id=30488 16210 16211 * jit/JITOpcodes.cpp: 16212 (JSC::JIT::emit_op_instanceof): 16213 (JSC::JIT::emitSlow_op_instanceof): No need to do object type checks - 16214 cell type checks and ImplementsDefaultHasIntance checks implicitly 16215 supersede object type checks. 16216 162172009-10-18 Kwang Yul Seo <skyul@company100.net> 16218 16219 Reviewed by Darin Adler. 16220 16221 Use _stricmp and _strnicmp instead of deprecated stricmp and strnicmp. 16222 https://bugs.webkit.org/show_bug.cgi?id=30474 16223 16224 stricmp and strnicmp are deprecated beginning in Visual 16225 C++ 2005. Use _stricmp and _strnicmp instead in StringExtras.h. 16226 16227 * wtf/StringExtras.h: 16228 (strncasecmp): 16229 (strcasecmp): 16230 162312009-10-16 Geoffrey Garen <ggaren@apple.com> 16232 16233 Build fix: apparently we shouldn't export those symbols? 16234 16235 * JavaScriptCore.exp: 16236 162372009-10-16 Geoffrey Garen <ggaren@apple.com> 16238 16239 Build fix: export some symbols. 16240 16241 * JavaScriptCore.exp: 16242 162432009-10-16 Oliver Hunt <oliver@apple.com> 16244 16245 Reviewed by Gavin Barraclough. 16246 16247 structure typeinfo flags should be inherited. 16248 https://bugs.webkit.org/show_bug.cgi?id=30468 16249 16250 Add StructureFlag constant to the various JSC classes and use 16251 it for the TypeInfo construction. This allows us to simply 16252 accumulate flags by basing each classes StructureInfo on its parents. 16253 16254 * API/JSCallbackConstructor.h: 16255 (JSC::JSCallbackConstructor::createStructure): 16256 * API/JSCallbackFunction.h: 16257 (JSC::JSCallbackFunction::createStructure): 16258 * API/JSCallbackObject.h: 16259 (JSC::JSCallbackObject::createStructure): 16260 * debugger/DebuggerActivation.h: 16261 (JSC::DebuggerActivation::createStructure): 16262 * runtime/Arguments.h: 16263 (JSC::Arguments::createStructure): 16264 * runtime/BooleanObject.h: 16265 (JSC::BooleanObject::createStructure): 16266 * runtime/DatePrototype.h: 16267 (JSC::DatePrototype::createStructure): 16268 * runtime/FunctionPrototype.h: 16269 (JSC::FunctionPrototype::createStructure): 16270 * runtime/GlobalEvalFunction.h: 16271 (JSC::GlobalEvalFunction::createStructure): 16272 * runtime/InternalFunction.h: 16273 (JSC::InternalFunction::createStructure): 16274 * runtime/JSActivation.h: 16275 (JSC::JSActivation::createStructure): 16276 * runtime/JSArray.h: 16277 (JSC::JSArray::createStructure): 16278 * runtime/JSByteArray.cpp: 16279 (JSC::JSByteArray::createStructure): 16280 * runtime/JSByteArray.h: 16281 * runtime/JSFunction.h: 16282 (JSC::JSFunction::createStructure): 16283 * runtime/JSGlobalObject.h: 16284 (JSC::JSGlobalObject::createStructure): 16285 * runtime/JSNotAnObject.h: 16286 (JSC::JSNotAnObject::createStructure): 16287 * runtime/JSONObject.h: 16288 (JSC::JSONObject::createStructure): 16289 * runtime/JSObject.h: 16290 (JSC::JSObject::createStructure): 16291 * runtime/JSStaticScopeObject.h: 16292 (JSC::JSStaticScopeObject::createStructure): 16293 * runtime/JSVariableObject.h: 16294 (JSC::JSVariableObject::createStructure): 16295 * runtime/JSWrapperObject.h: 16296 (JSC::JSWrapperObject::createStructure): 16297 * runtime/MathObject.h: 16298 (JSC::MathObject::createStructure): 16299 * runtime/NumberConstructor.h: 16300 (JSC::NumberConstructor::createStructure): 16301 * runtime/NumberObject.h: 16302 (JSC::NumberObject::createStructure): 16303 * runtime/RegExpConstructor.h: 16304 (JSC::RegExpConstructor::createStructure): 16305 * runtime/RegExpObject.h: 16306 (JSC::RegExpObject::createStructure): 16307 * runtime/StringObject.h: 16308 (JSC::StringObject::createStructure): 16309 * runtime/StringObjectThatMasqueradesAsUndefined.h: 16310 (JSC::StringObjectThatMasqueradesAsUndefined::createStructure): 16311 163122009-10-16 Geoffrey Garen <ggaren@apple.com> 16313 16314 Reviewed by Oliver Hunt. 16315 16316 Fast for-in enumeration: Cache JSPropertyNameIterator; cache JSStrings 16317 in JSPropertyNameIterator; inline more code. 16318 16319 1.024x as fast on SunSpider (fasta: 1.43x as fast). 16320 16321 * bytecode/CodeBlock.cpp: 16322 (JSC::CodeBlock::dump): 16323 * bytecode/Opcode.h: 16324 * bytecompiler/BytecodeGenerator.cpp: 16325 (JSC::BytecodeGenerator::emitGetPropertyNames): 16326 (JSC::BytecodeGenerator::emitNextPropertyName): 16327 * bytecompiler/BytecodeGenerator.h: Added a few extra operands to 16328 op_get_pnames and op_next_pname so that we can track iteration state 16329 in the register file instead of in the JSPropertyNameIterator. (To be 16330 cacheable, the JSPropertyNameIterator must be stateless.) 16331 16332 * interpreter/Interpreter.cpp: 16333 (JSC::Interpreter::tryCachePutByID): 16334 (JSC::Interpreter::tryCacheGetByID): Updated for rename to 16335 "normalizePrototypeChain" and removal of "isCacheable". 16336 16337 (JSC::Interpreter::privateExecute): Updated for in-RegisterFile 16338 iteration state tracking. 16339 16340 * jit/JIT.cpp: 16341 (JSC::JIT::privateCompileMainPass): 16342 * jit/JIT.h: 16343 * jit/JITOpcodes.cpp: 16344 (JSC::JIT::emit_op_get_pnames): Updated for in-RegisterFile 16345 iteration state tracking. 16346 16347 (JSC::JIT::emit_op_next_pname): Inlined code generation for op_next_pname. 16348 16349 * jit/JITStubs.cpp: 16350 (JSC::JITThunks::tryCachePutByID): 16351 (JSC::JITThunks::tryCacheGetByID): Updated for rename to 16352 "normalizePrototypeChain" and removal of "isCacheable". 16353 16354 (JSC::DEFINE_STUB_FUNCTION): 16355 * jit/JITStubs.h: 16356 (JSC::): Added has_property and to_object stubs. Removed op_next_pname 16357 stub, since has_property is all we need anymore. 16358 16359 * parser/Nodes.cpp: 16360 (JSC::ForInNode::emitBytecode): Updated for in-RegisterFile 16361 iteration state tracking. 16362 16363 * runtime/JSCell.h: 16364 * runtime/JSObject.cpp: 16365 (JSC::JSObject::getPropertyNames): Don't do caching at this layer 16366 anymore, since we don't create a JSPropertyNameIterator at this layer. 16367 16368 * runtime/JSPropertyNameIterator.cpp: 16369 (JSC::JSPropertyNameIterator::create): Do do caching at this layer. 16370 (JSC::JSPropertyNameIterator::get): Updated for in-RegisterFile 16371 iteration state tracking. 16372 (JSC::JSPropertyNameIterator::markChildren): Mark our JSStrings. 16373 16374 * runtime/JSPropertyNameIterator.h: 16375 (JSC::JSPropertyNameIterator::size): 16376 (JSC::JSPropertyNameIterator::setCachedStructure): 16377 (JSC::JSPropertyNameIterator::cachedStructure): 16378 (JSC::JSPropertyNameIterator::setCachedPrototypeChain): 16379 (JSC::JSPropertyNameIterator::cachedPrototypeChain): 16380 (JSC::JSPropertyNameIterator::JSPropertyNameIterator): 16381 (JSC::Structure::setEnumerationCache): Don't store iteration state in 16382 a JSPropertyNameIterator. Do cache a JSPropertyNameIterator in a 16383 Structure. 16384 16385 * runtime/JSValue.h: 16386 (JSC::asCell): 16387 * runtime/MarkStack.h: Make those mischievous #include gods happy. 16388 16389 * runtime/ObjectConstructor.cpp: 16390 16391 * runtime/Operations.h: 16392 (JSC::normalizePrototypeChain): Renamed countPrototypeChainEntriesAndCheckForProxies 16393 to normalizePrototypeChain, since it changes dictionary prototypes to 16394 non-dictionary objects. 16395 16396 * runtime/PropertyNameArray.cpp: 16397 (JSC::PropertyNameArray::add): 16398 * runtime/PropertyNameArray.h: 16399 (JSC::PropertyNameArrayData::PropertyNameArrayData): 16400 (JSC::PropertyNameArray::data): 16401 (JSC::PropertyNameArray::size): 16402 (JSC::PropertyNameArray::begin): 16403 (JSC::PropertyNameArray::end): Simplified some code here to help with 16404 current and future refactoring. 16405 16406 * runtime/Protect.h: 16407 * runtime/Structure.cpp: 16408 (JSC::Structure::~Structure): 16409 (JSC::Structure::addPropertyWithoutTransition): 16410 (JSC::Structure::removePropertyWithoutTransition): No need to clear 16411 the enumeration cache with adding / removing properties without 16412 transition. It is an error to add / remove properties without transition 16413 once an object has been observed, and we can ASSERT to catch that. 16414 16415 * runtime/Structure.h: 16416 (JSC::Structure::enumerationCache): Changed the enumeration cache to 16417 hold a JSPropertyNameIterator. 16418 16419 * runtime/StructureChain.cpp: 16420 * runtime/StructureChain.h: 16421 (JSC::StructureChain::head): Removed StructureChain::isCacheable because 16422 it was wrong-headed in two ways: (1) It gave up when a prototype was a 16423 dictionary, but instead we want un-dictionary heavily accessed 16424 prototypes; (2) It folded a test for hasDefaultGetPropertyNames() into 16425 a generic test for "cacheable-ness", but hasDefaultGetPropertyNames() 16426 is only releavant to for-in caching. 16427 164282009-10-16 Steve Falkenburg <sfalken@apple.com> 16429 16430 Reviewed by Adam Roben. 16431 16432 Add a Debug_All configuration to build entire stack as debug. 16433 Change Debug_Internal to: 16434 - stop using _debug suffix for all WebKit/Safari binaries 16435 - not use _debug as a DLL naming suffix 16436 - use non-debug C runtime lib. 16437 16438 * JavaScriptCore.vcproj/JavaScriptCore.make: Debug build in makefile should build Debug_All. 16439 * JavaScriptCore.vcproj/JavaScriptCore.sln: Add Debug_All configuration. 16440 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: Add Debug_All configuration. 16441 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCoreGenerated.vcproj: Renamed single configuration from "Release" to "all". 16442 * JavaScriptCore.vcproj/JavaScriptCoreSubmit.sln: Add Debug_All configuration. 16443 * JavaScriptCore.vcproj/WTF/WTF.vcproj: Add Debug_All configuration. 16444 * JavaScriptCore.vcproj/jsc/jsc.vcproj: Add Debug_All configuration. 16445 * JavaScriptCore.vcproj/testapi/testapi.vcproj: Add Debug_All configuration. 16446 164472009-10-16 Oliver Hunt <oliver@apple.com> 16448 16449 Reviewed by Gavin Barraclough. 16450 16451 Make typeinfo flags default to false 16452 https://bugs.webkit.org/show_bug.cgi?id=30372 16453 16454 Last part -- replace HasDefaultGetPropertyNames with OverridesGetPropertyNames 16455 flag. 16456 16457 * API/JSCallbackConstructor.h: 16458 (JSC::JSCallbackConstructor::createStructure): 16459 * API/JSCallbackObject.h: 16460 (JSC::JSCallbackObject::createStructure): 16461 * debugger/DebuggerActivation.h: 16462 (JSC::DebuggerActivation::createStructure): 16463 * runtime/Arguments.h: 16464 (JSC::Arguments::createStructure): 16465 * runtime/BooleanObject.h: 16466 (JSC::BooleanObject::createStructure): 16467 * runtime/DatePrototype.h: 16468 (JSC::DatePrototype::createStructure): 16469 * runtime/FunctionPrototype.h: 16470 (JSC::FunctionPrototype::createStructure): 16471 * runtime/GlobalEvalFunction.h: 16472 (JSC::GlobalEvalFunction::createStructure): 16473 * runtime/JSAPIValueWrapper.h: 16474 (JSC::JSAPIValueWrapper::createStructure): 16475 * runtime/JSActivation.h: 16476 (JSC::JSActivation::createStructure): 16477 * runtime/JSArray.h: 16478 (JSC::JSArray::createStructure): 16479 * runtime/JSByteArray.cpp: 16480 (JSC::JSByteArray::createStructure): 16481 * runtime/JSFunction.h: 16482 (JSC::JSFunction::createStructure): 16483 * runtime/JSGlobalObject.h: 16484 (JSC::JSGlobalObject::createStructure): 16485 * runtime/JSNotAnObject.h: 16486 (JSC::JSNotAnObject::createStructure): 16487 * runtime/JSONObject.h: 16488 (JSC::JSONObject::createStructure): 16489 * runtime/JSObject.cpp: 16490 (JSC::JSObject::getPropertyNames): 16491 * runtime/JSObject.h: 16492 (JSC::JSObject::createStructure): 16493 * runtime/JSStaticScopeObject.h: 16494 (JSC::JSStaticScopeObject::createStructure): 16495 * runtime/JSTypeInfo.h: 16496 (JSC::TypeInfo::overridesGetPropertyNames): 16497 * runtime/JSVariableObject.h: 16498 (JSC::JSVariableObject::createStructure): 16499 * runtime/JSWrapperObject.h: 16500 (JSC::JSWrapperObject::createStructure): 16501 * runtime/MathObject.h: 16502 (JSC::MathObject::createStructure): 16503 * runtime/NumberConstructor.h: 16504 (JSC::NumberConstructor::createStructure): 16505 * runtime/NumberObject.h: 16506 (JSC::NumberObject::createStructure): 16507 * runtime/RegExpConstructor.h: 16508 (JSC::RegExpConstructor::createStructure): 16509 * runtime/RegExpObject.h: 16510 (JSC::RegExpObject::createStructure): 16511 * runtime/StringObject.h: 16512 (JSC::StringObject::createStructure): 16513 * runtime/StringObjectThatMasqueradesAsUndefined.h: 16514 (JSC::StringObjectThatMasqueradesAsUndefined::createStructure): 16515 * runtime/StructureChain.cpp: 16516 (JSC::StructureChain::isCacheable): 16517 165182009-10-16 Kevin Ollivier <kevino@theolliviers.com> 16519 16520 wxMSW build fix, we can't use the simple hash there because the PlatformModuleVersion 16521 structure differs. 16522 16523 * wtf/Platform.h: 16524 165252009-10-16 Laszlo Gombos <laszlo.1.gombos@nokia.com> 16526 16527 Reviewed by Simon Hausmann. 16528 16529 [Qt] Implement ExecutableAllocator for Symbian 16530 https://bugs.webkit.org/show_bug.cgi?id=29946 16531 16532 Tested with YARR JIT enabled for Symbian; 16533 This patch does not (yet) enable YARR JIT by default. 16534 16535 * JavaScriptCore.pri: 16536 * jit/ExecutableAllocator.h: 16537 * jit/ExecutableAllocatorSymbian.cpp: Added. 16538 (JSC::ExecutableAllocator::intializePageSize): 16539 (JSC::ExecutablePool::systemAlloc): 16540 (JSC::ExecutablePool::systemRelease): 16541 165422009-10-15 Oliver Hunt <oliver@apple.com> 16543 16544 Reviewed by Darin Adler. 16545 16546 Make typeinfo flags default to false 16547 https://bugs.webkit.org/show_bug.cgi?id=30372 16548 16549 Part 2 -- Reverse the TypeInfo HasDefaultMark flag to OverridesMarkChildren, etc 16550 16551 * API/JSCallbackConstructor.h: 16552 (JSC::JSCallbackConstructor::createStructure): 16553 * API/JSCallbackFunction.h: 16554 (JSC::JSCallbackFunction::createStructure): 16555 * API/JSCallbackObject.h: 16556 (JSC::JSCallbackObject::createStructure): 16557 * debugger/DebuggerActivation.h: 16558 (JSC::DebuggerActivation::createStructure): 16559 * runtime/Arguments.h: 16560 (JSC::Arguments::createStructure): 16561 * runtime/BooleanObject.h: 16562 (JSC::BooleanObject::createStructure): 16563 * runtime/DatePrototype.h: 16564 (JSC::DatePrototype::createStructure): 16565 * runtime/FunctionPrototype.h: 16566 (JSC::FunctionPrototype::createStructure): 16567 * runtime/GetterSetter.h: 16568 (JSC::GetterSetter::createStructure): 16569 * runtime/GlobalEvalFunction.h: 16570 (JSC::GlobalEvalFunction::createStructure): 16571 * runtime/InternalFunction.h: 16572 (JSC::InternalFunction::createStructure): 16573 * runtime/JSAPIValueWrapper.h: 16574 (JSC::JSAPIValueWrapper::createStructure): 16575 * runtime/JSActivation.h: 16576 (JSC::JSActivation::createStructure): 16577 * runtime/JSArray.h: 16578 (JSC::JSArray::createStructure): 16579 (JSC::MarkStack::markChildren): 16580 * runtime/JSByteArray.cpp: 16581 (JSC::JSByteArray::createStructure): 16582 * runtime/JSFunction.h: 16583 (JSC::JSFunction::createStructure): 16584 * runtime/JSGlobalObject.h: 16585 (JSC::JSGlobalObject::createStructure): 16586 * runtime/JSNotAnObject.h: 16587 (JSC::JSNotAnObject::createStructure): 16588 * runtime/JSNumberCell.h: 16589 (JSC::JSNumberCell::createStructure): 16590 * runtime/JSONObject.h: 16591 (JSC::JSONObject::createStructure): 16592 * runtime/JSObject.h: 16593 (JSC::JSObject::createStructure): 16594 * runtime/JSPropertyNameIterator.h: 16595 (JSC::JSPropertyNameIterator::createStructure): 16596 * runtime/JSStaticScopeObject.h: 16597 (JSC::JSStaticScopeObject::createStructure): 16598 * runtime/JSString.h: 16599 (JSC::JSString::createStructure): 16600 * runtime/JSTypeInfo.h: 16601 (JSC::TypeInfo::overridesMarkChildren): 16602 * runtime/JSVariableObject.h: 16603 (JSC::JSVariableObject::createStructure): 16604 * runtime/JSWrapperObject.h: 16605 (JSC::JSWrapperObject::createStructure): 16606 * runtime/MathObject.h: 16607 (JSC::MathObject::createStructure): 16608 * runtime/NumberConstructor.h: 16609 (JSC::NumberConstructor::createStructure): 16610 * runtime/NumberObject.h: 16611 (JSC::NumberObject::createStructure): 16612 * runtime/RegExpConstructor.h: 16613 (JSC::RegExpConstructor::createStructure): 16614 * runtime/RegExpObject.h: 16615 (JSC::RegExpObject::createStructure): 16616 * runtime/StringObject.h: 16617 (JSC::StringObject::createStructure): 16618 * runtime/StringObjectThatMasqueradesAsUndefined.h: 16619 (JSC::StringObjectThatMasqueradesAsUndefined::createStructure): 16620 166212009-10-14 Oliver Hunt <oliver@apple.com> 16622 16623 Reviewed by Geoff Garen. 16624 16625 Make typeinfo flags default to false 16626 https://bugs.webkit.org/show_bug.cgi?id=30372 16627 16628 Part 1. Reverse the HasStandardGetOwnPropertySlot flag. 16629 16630 * API/JSCallbackConstructor.h: 16631 (JSC::JSCallbackConstructor::createStructure): 16632 * API/JSCallbackFunction.h: 16633 (JSC::JSCallbackFunction::createStructure): 16634 * API/JSCallbackObject.h: 16635 (JSC::JSCallbackObject::createStructure): 16636 * debugger/DebuggerActivation.h: 16637 (JSC::DebuggerActivation::createStructure): 16638 * jit/JITStubs.cpp: 16639 (JSC::DEFINE_STUB_FUNCTION): 16640 * runtime/Arguments.h: 16641 (JSC::Arguments::createStructure): 16642 * runtime/BooleanObject.h: 16643 (JSC::BooleanObject::createStructure): 16644 * runtime/DatePrototype.h: 16645 (JSC::DatePrototype::createStructure): 16646 * runtime/FunctionPrototype.h: 16647 (JSC::FunctionPrototype::createStructure): 16648 * runtime/GlobalEvalFunction.h: 16649 (JSC::GlobalEvalFunction::createStructure): 16650 * runtime/InternalFunction.h: 16651 (JSC::InternalFunction::createStructure): 16652 * runtime/JSActivation.h: 16653 (JSC::JSActivation::createStructure): 16654 * runtime/JSArray.h: 16655 (JSC::JSArray::createStructure): 16656 * runtime/JSByteArray.cpp: 16657 (JSC::JSByteArray::createStructure): 16658 * runtime/JSFunction.h: 16659 (JSC::JSFunction::createStructure): 16660 * runtime/JSGlobalObject.h: 16661 (JSC::JSGlobalObject::createStructure): 16662 * runtime/JSNumberCell.h: 16663 (JSC::JSNumberCell::createStructure): 16664 * runtime/JSONObject.h: 16665 (JSC::JSONObject::createStructure): 16666 * runtime/JSObject.h: 16667 (JSC::JSObject::createStructure): 16668 (JSC::JSCell::fastGetOwnPropertySlot): 16669 * runtime/JSStaticScopeObject.h: 16670 (JSC::JSStaticScopeObject::createStructure): 16671 * runtime/JSString.h: 16672 (JSC::JSString::createStructure): 16673 * runtime/JSTypeInfo.h: 16674 (JSC::TypeInfo::overridesGetOwnPropertySlot): 16675 * runtime/JSVariableObject.h: 16676 (JSC::JSVariableObject::createStructure): 16677 * runtime/JSWrapperObject.h: 16678 (JSC::JSWrapperObject::createStructure): 16679 * runtime/MathObject.h: 16680 (JSC::MathObject::createStructure): 16681 * runtime/NumberConstructor.h: 16682 (JSC::NumberConstructor::createStructure): 16683 * runtime/NumberObject.h: 16684 (JSC::NumberObject::createStructure): 16685 * runtime/RegExpConstructor.h: 16686 (JSC::RegExpConstructor::createStructure): 16687 * runtime/RegExpObject.h: 16688 (JSC::RegExpObject::createStructure): 16689 * runtime/StringObject.h: 16690 (JSC::StringObject::createStructure): 16691 * runtime/StringObjectThatMasqueradesAsUndefined.h: 16692 (JSC::StringObjectThatMasqueradesAsUndefined::createStructure): 16693 166942009-10-14 Kevin Ollivier <kevino@theolliviers.com> 166952009-10-14 Darin Adler <darin@apple.com> 16696 16697 Additions so fix for https://bugs.webkit.org/show_bug.cgi?id=18994 16698 can build on Windows. 16699 16700 * wtf/MathExtras.h: Added llround and llroundf for Windows. 16701 167022009-10-14 Kevin Ollivier <kevino@theolliviers.com> 16703 16704 wx build fix. Set ENABLE_PLUGIN_PACKAGE_SIMPLE_HASH for plugins while we're still building stubs. 16705 16706 * wtf/Platform.h: 16707 167082009-10-13 Laszlo Gombos <laszlo.1.gombos@nokia.com> 16709 16710 Reviewed by Simon Hausmann. 16711 16712 Refactor ENABLE_PLUGIN_PACKAGE_SIMPLE_HASH 16713 https://bugs.webkit.org/show_bug.cgi?id=30278 16714 16715 Move the definition of ENABLE_PLUGIN_PACKAGE_SIMPLE_HASH 16716 from the make system into common code. 16717 16718 * wtf/Platform.h: 16719 167202009-10-13 Laszlo Gombos <laszlo.1.gombos@nokia.com> 16721 16722 Reviewed by Darin Adler. 16723 16724 ARM compiler does not understand reinterpret_cast<void*> 16725 https://bugs.webkit.org/show_bug.cgi?id=29034 16726 16727 Change reinterpret_cast<void*> to regular C style (void*) cast 16728 for the ARM RVCT compiler. 16729 16730 * assembler/MacroAssemblerCodeRef.h: 16731 (JSC::FunctionPtr::FunctionPtr): 16732 * jit/JITOpcodes.cpp: Cast to FunctionPtr first 16733 instead of directly casting to reinterpret_cast 16734 * jit/JITStubCall.h: Ditto + change the type of m_stub 16735 from void* to FunctionPtr. 16736 (JSC::JITStubCall::JITStubCall): 16737 (JSC::JITStubCall::call): 16738 * jit/JITStubs.cpp: Ditto. 16739 (JSC::DEFINE_STUB_FUNCTION(EncodedJSValue, op_throw)): 16740 167412009-10-11 Oliver Hunt <oliver@apple.com> 16742 16743 Re-enable the JIT. 16744 16745 * wtf/Platform.h: 16746 167472009-10-10 Oliver Hunt <oliver@apple.com> 16748 16749 Reviewed by Maciej Stachowiak. 16750 16751 Support for String.trim(), String.trimLeft() and String.trimRight() methods 16752 https://bugs.webkit.org/show_bug.cgi?id=26590 16753 16754 Implement trim, trimLeft, and trimRight 16755 16756 * runtime/StringPrototype.cpp: 16757 (JSC::isTrimWhitespace): 16758 Our normal string whitespace function does not include U+200B which 16759 is needed for compatibility with mozilla's implementation of trim. 16760 U+200B does not appear to be expected according to spec, however I am 16761 choosing to be lax, and match mozilla behavior so have added this 16762 exception. 16763 (JSC::trimString): 16764 167652009-10-09 Geoffrey Garen <ggaren@apple.com> 16766 16767 Reviewed by Oliver Hunt. 16768 16769 Eliminated some legacy bytecode weirdness. 16770 16771 Use vPC[x] subscripting instead of ++vPC to access instruction operands. 16772 This is simpler, and often more efficient. 16773 16774 To support this, and to remove use of hard-coded offsets in bytecode and 16775 JIT code generation and dumping, calculate jump offsets from the beginning 16776 of an instruction, rather than the middle or end. 16777 16778 Also, use OPCODE_LENGTH instead of hard-coded constants for the sizes of 16779 opcodes. 16780 16781 SunSpider reports no change in JIT mode, and a 1.01x speedup in Interpreter 16782 mode. 16783 16784 * bytecode/CodeBlock.cpp: 16785 (JSC::printConditionalJump): 16786 (JSC::CodeBlock::dump): 16787 * bytecompiler/BytecodeGenerator.cpp: 16788 (JSC::BytecodeGenerator::emitJump): 16789 (JSC::BytecodeGenerator::emitJumpIfTrue): 16790 (JSC::BytecodeGenerator::emitJumpIfFalse): 16791 (JSC::BytecodeGenerator::emitJumpIfNotFunctionCall): 16792 (JSC::BytecodeGenerator::emitJumpIfNotFunctionApply): 16793 (JSC::BytecodeGenerator::emitComplexJumpScopes): 16794 (JSC::BytecodeGenerator::emitJumpScopes): 16795 (JSC::BytecodeGenerator::emitNextPropertyName): 16796 (JSC::BytecodeGenerator::emitCatch): 16797 (JSC::BytecodeGenerator::emitJumpSubroutine): 16798 (JSC::prepareJumpTableForImmediateSwitch): 16799 (JSC::prepareJumpTableForCharacterSwitch): 16800 (JSC::prepareJumpTableForStringSwitch): 16801 (JSC::BytecodeGenerator::endSwitch): 16802 * bytecompiler/Label.h: 16803 (JSC::Label::setLocation): 16804 (JSC::Label::bind): 16805 * interpreter/Interpreter.cpp: 16806 (JSC::Interpreter::resolve): 16807 (JSC::Interpreter::resolveSkip): 16808 (JSC::Interpreter::resolveGlobal): 16809 (JSC::Interpreter::resolveBase): 16810 (JSC::Interpreter::resolveBaseAndProperty): 16811 (JSC::Interpreter::createExceptionScope): 16812 (JSC::Interpreter::privateExecute): 16813 * interpreter/Interpreter.h: 16814 * jit/JIT.cpp: 16815 (JSC::JIT::privateCompile): 16816 * jit/JITArithmetic.cpp: 16817 (JSC::JIT::emit_op_jnless): 16818 (JSC::JIT::emitSlow_op_jnless): 16819 (JSC::JIT::emit_op_jnlesseq): 16820 (JSC::JIT::emitSlow_op_jnlesseq): 16821 (JSC::JIT::emitBinaryDoubleOp): 16822 * jit/JITOpcodes.cpp: 16823 (JSC::JIT::emit_op_jmp): 16824 (JSC::JIT::emit_op_loop): 16825 (JSC::JIT::emit_op_loop_if_less): 16826 (JSC::JIT::emitSlow_op_loop_if_less): 16827 (JSC::JIT::emit_op_loop_if_lesseq): 16828 (JSC::JIT::emitSlow_op_loop_if_lesseq): 16829 (JSC::JIT::emit_op_loop_if_true): 16830 (JSC::JIT::emitSlow_op_loop_if_true): 16831 (JSC::JIT::emit_op_jfalse): 16832 (JSC::JIT::emitSlow_op_jfalse): 16833 (JSC::JIT::emit_op_jtrue): 16834 (JSC::JIT::emitSlow_op_jtrue): 16835 (JSC::JIT::emit_op_jeq_null): 16836 (JSC::JIT::emit_op_jneq_null): 16837 (JSC::JIT::emit_op_jneq_ptr): 16838 (JSC::JIT::emit_op_jsr): 16839 (JSC::JIT::emit_op_next_pname): 16840 (JSC::JIT::emit_op_jmp_scopes): 16841 168422009-10-09 Geoffrey Garen <ggaren@apple.com> 16843 16844 Reviewed by Sam Weinig. 16845 16846 Migrated some code that didn't belong out of Structure. 16847 16848 SunSpider says maybe 1.03x faster. 16849 16850 * runtime/JSCell.h: Nixed Structure::markAggregate, and made marking of 16851 a Structure's prototype the direct responsility of the object using it. 16852 (Giving Structure a mark function was misleading because it implied that 16853 all live structures get marked during GC, when they don't.) 16854 16855 * runtime/JSGlobalObject.cpp: 16856 (JSC::markIfNeeded): 16857 (JSC::JSGlobalObject::markChildren): Added code to mark prototypes stored 16858 on the global object. Maybe this wasn't necessary, but now we don't have 16859 to wonder. 16860 16861 * runtime/JSObject.cpp: 16862 (JSC::JSObject::getPropertyNames): 16863 (JSC::JSObject::getOwnPropertyNames): 16864 (JSC::JSObject::getEnumerableNamesFromClassInfoTable): 16865 * runtime/JSObject.h: 16866 (JSC::JSObject::markChildrenDirect): 16867 * runtime/PropertyNameArray.h: 16868 * runtime/Structure.cpp: 16869 * runtime/Structure.h: 16870 (JSC::Structure::setEnumerationCache): 16871 (JSC::Structure::enumerationCache): Moved property name gathering code 16872 from Structure to JSObject because having a Structure iterate its JSObject 16873 was a layering violation. A JSObject is implemented using a Structure; not 16874 the other way around. 16875 168762009-10-09 Mark Rowe <mrowe@apple.com> 16877 16878 Attempt to fix the GTK release build. 16879 16880 * GNUmakefile.am: Include Grammar.cpp in release builds now that 16881 AllInOneFile.cpp is gone. 16882 168832009-10-09 Gabor Loki <loki@inf.u-szeged.hu> 16884 16885 Rubber-stamped by Eric Seidel. 16886 16887 Add ARM JIT support for Gtk port (disabled by default) 16888 https://bugs.webkit.org/show_bug.cgi?id=30228 16889 16890 * GNUmakefile.am: 16891 168922009-10-08 Geoffrey Garen <ggaren@apple.com> 16893 16894 Tiger build fix: added a few more variable initializations. 16895 16896 * runtime/StringPrototype.cpp: 16897 (JSC::stringProtoFuncReplace): 16898 (JSC::stringProtoFuncSearch): 16899 169002009-10-08 Geoffrey Garen <ggaren@apple.com> 16901 16902 Qt build fix: added missing #include. 16903 16904 * jsc.cpp: 16905 169062009-10-08 Geoffrey Garen <ggaren@apple.com> 16907 16908 Tiger build fix: initialize variable whose initialization the compiler 16909 can't otherwise figure out. 16910 16911 * runtime/RegExpObject.cpp: 16912 (JSC::RegExpObject::match): 16913 169142009-10-08 Geoffrey Garen <ggaren@apple.com> 16915 16916 Windows build fix: updated exports. 16917 16918 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 16919 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 16920 169212009-10-08 Geoffrey Garen <ggaren@apple.com> 16922 16923 Tiger build fix: fixed file name case. 16924 16925 * JavaScriptCore.xcodeproj/project.pbxproj: 16926 169272009-10-08 Geoffrey Garen <ggaren@apple.com> 16928 16929 Reviewed by Maciej Stachowiak. 16930 16931 At long last, I pronounce the death of AllInOneFile.cpp. 16932 16933 SunSpider reports a 1.01x speedup. 16934 16935 * AllInOneFile.cpp: Removed. 16936 * GNUmakefile.am: 16937 * JavaScriptCore.exp: 16938 * JavaScriptCore.gypi: 16939 * JavaScriptCore.xcodeproj/project.pbxproj: Added missing project files 16940 to compilation stages. 16941 16942 * parser/Grammar.y: 16943 * parser/Lexer.cpp: 16944 * parser/Lexer.h: 16945 (JSC::jscyylex): 16946 * runtime/ArrayConstructor.cpp: 16947 (JSC::constructArrayWithSizeQuirk): 16948 * runtime/Collector.h: 16949 * runtime/JSCell.cpp: 16950 (JSC::JSCell::operator new): 16951 * runtime/JSCell.h: 16952 (JSC::JSCell::operator new): 16953 * runtime/JSGlobalObject.cpp: 16954 (JSC::JSGlobalObject::operator new): 16955 * runtime/JSNumberCell.h: 16956 (JSC::JSNumberCell::operator new): 16957 * runtime/JSString.cpp: 16958 * runtime/JSString.h: 16959 (JSC::jsString): 16960 (JSC::jsSubstring): 16961 (JSC::jsOwnedString): 16962 * runtime/RegExpConstructor.cpp: 16963 * runtime/RegExpConstructor.h: 16964 (JSC::RegExpConstructorPrivate::RegExpConstructorPrivate): 16965 (JSC::RegExpConstructorPrivate::lastOvector): 16966 (JSC::RegExpConstructorPrivate::tempOvector): 16967 (JSC::RegExpConstructorPrivate::changeLastOvector): 16968 (JSC::RegExpConstructor::performMatch): 16969 * runtime/StringPrototype.cpp: 16970 (JSC::stringProtoFuncMatch): 16971 * yarr/RegexJIT.cpp: 16972 * yarr/RegexJIT.h: 16973 (JSC::Yarr::executeRegex): Inlined a few things that Shark said 16974 were hot, on the presumption that AllInOneFile.cpp used to inline them 16975 automatically. 16976 169772009-10-08 Zoltan Herczeg <zherczeg@inf.u-szeged.hu> 16978 16979 Reviewed by Gavin Barraclough. 16980 16981 Fix for JIT'ed op_call instructions (evals, constructs, etc.) 16982 when !ENABLE(JIT_OPTIMIZE_CALL) && USE(JSVALUE32_64) 16983 16984 https://bugs.webkit.org/show_bug.cgi?id=30201 16985 16986 * jit/JITCall.cpp: 16987 (JSC::JIT::compileOpCall): 16988 169892009-10-07 Geoffrey Garen <ggaren@apple.com> 16990 16991 Windows build fix: removed no longer exported symbol. 16992 16993 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 16994 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 16995 169962009-10-07 Geoffrey Garen <ggaren@apple.com> 16997 16998 Reviewed by Oliver Hunt. 16999 17000 Fixed <rdar://problem/5751979> Database code takes JSLock on secondary 17001 thread, permanently slowing down JavaScript 17002 17003 Removed the optional lock from Heap::protect, Heap::unprotect, and friends, 17004 since WebCore no longer uses it. 17005 17006 * JavaScriptCore.exp: 17007 * runtime/Collector.cpp: 17008 (JSC::Heap::protect): 17009 (JSC::Heap::unprotect): 17010 (JSC::Heap::markProtectedObjects): 17011 (JSC::Heap::protectedGlobalObjectCount): 17012 (JSC::Heap::protectedObjectCount): 17013 (JSC::Heap::protectedObjectTypeCounts): 17014 * runtime/Collector.h: 17015 170162009-10-07 Zoltan Horvath <zoltan@webkit.org> 17017 17018 Reviewed by Darin Adler. 17019 17020 Allow custom memory allocation control for JavaScriptCore's IdentifierArena 17021 https://bugs.webkit.org/show_bug.cgi?id=30158 17022 17023 Inherits IdentifierArena class from FastAllocBase because it has been 17024 instantiated by 'new' in JavaScriptCore/parser/ParserArena.cpp:36. 17025 17026 * parser/ParserArena.h: 17027 170282009-10-07 Adam Roben <aroben@apple.com> 17029 17030 Export DateInstance::info in a way that works on Windows 17031 17032 Fixes <http://webkit.org/b/30171> 17033 fast/dom/Window/window-postmessage-clone.html fails on Windows 17034 17035 Reviewed by Anders Carlsson. 17036 17037 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 17038 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 17039 Removed the export of DateInstance::info from here. 17040 17041 * runtime/DateInstance.h: Use JS_EXPORTDATA to export 17042 DateInstance::info, which is the required way of exporting data on 17043 Windows. 17044 170452009-10-07 Jørgen Lind <jorgen.lind@nokia.com> 17046 17047 Reviewed by Simon Hausmann. 17048 17049 When enabling or disabling the JIT through .qmake.cache, make sure 17050 to also toggle ENABLE_YARR_JIT. 17051 17052 * JavaScriptCore.pri: 17053 170542009-10-06 Priit Laes <plaes@plaes.org> 17055 17056 Reviewed by Gavin Barraclough. 17057 17058 Linking fails with "relocation R_X86_64_PC32 against symbol 17059 `cti_vm_throw'" 17060 https://bugs.webkit.org/show_bug.cgi?id=28422 17061 17062 * jit/JITStubs.cpp: 17063 Mark cti_vm_throw symbol as PLT-indirect symbol, so it doesn't end up 17064 in text segment causing relocation errors on amd64 architecture. 17065 Introduced new define SYMBOL_STRING_RELOCATION for such symbols. 17066 170672009-10-06 Oliver Hunt <oliver@apple.com> 17068 17069 Windows linking fix 17070 17071 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 17072 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 17073 170742009-10-06 Oliver Hunt <oliver@apple.com> 17075 17076 Reviewed by NOBODY (build fix). 17077 17078 Windows build fix. 17079 17080 * runtime/DateInstance.cpp: 17081 170822009-10-05 Oliver Hunt <oliver@apple.com> 17083 17084 Reviewed by Gavin Barraclough. 17085 17086 It should be possible to post (clone) built-in JS objects to Workers 17087 https://bugs.webkit.org/show_bug.cgi?id=22878 17088 17089 Expose helpers to throw correct exceptions during object graph walk 17090 used for cloning and add a helper function to create Date instances 17091 without going through the JS Date constructor function. 17092 17093 * JavaScriptCore.exp: 17094 * JavaScriptCore.xcodeproj/project.pbxproj: 17095 * runtime/DateInstance.cpp: 17096 (JSC::DateInstance::DateInstance): 17097 * runtime/DateInstance.h: 17098 * runtime/ExceptionHelpers.cpp: 17099 (JSC::createTypeError): 17100 * runtime/ExceptionHelpers.h: 17101 171022009-10-06 David Levin <levin@chromium.org> 17103 17104 Reviewed by Oliver Hunt. 17105 17106 StringImpl needs a method to get an instance for another thread which doesn't copy the underlying buffer. 17107 https://bugs.webkit.org/show_bug.cgi?id=30095 17108 17109 * wtf/CrossThreadRefCounted.h: 17110 Removed an unused function and assert improvement. 17111 (WTF::CrossThreadRefCounted::isOwnedByCurrentThread): Moved out common code from asserts. 17112 (WTF::CrossThreadRefCounted::ref): Changed assert to use the common method. 17113 (WTF::CrossThreadRefCounted::deref): Changed assert to use the common method. 17114 (WTF::CrossThreadRefCounted::crossThreadCopy): Since this includes a potentially 17115 non-threadsafe operation, add an assert that the class is owned by the current thread. 17116 171172009-10-05 Kevin Ollivier <kevino@theolliviers.com> 17118 17119 wx build fix. Add Symbian files to the list of excludes. 17120 17121 * wscript: 17122 171232009-10-05 Jocelyn Turcotte <jocelyn.turcotte@nokia.com> 17124 17125 Reviewed by Simon Hausmann. 17126 17127 [Qt] Remove precompiled header from JavaScriptCore compilation to 17128 prevent qmake warning during autonomous compilation. 17129 https://bugs.webkit.org/show_bug.cgi?id=30069 17130 17131 * JavaScriptCore.pro: 17132 171332009-10-02 Geoffrey Garen <ggaren@apple.com> 17134 17135 Reviewed by Sam Weinig. 17136 17137 Removed the concept of a "fast access cutoff" in arrays, because it 17138 punished some patterns of array access too much, and made things too 17139 complex for inlining in some cases. 17140 17141 1.3% speedup on SunSpider. 17142 17143 * jit/JITOpcodes.cpp: 17144 (JSC::JIT::emitSlow_op_get_by_val): 17145 (JSC::JIT::emitSlow_op_put_by_val): 17146 * jit/JITPropertyAccess.cpp: 17147 (JSC::JIT::emit_op_get_by_val): 17148 (JSC::JIT::emitSlow_op_get_by_val): 17149 (JSC::JIT::emit_op_put_by_val): 17150 (JSC::JIT::emitSlow_op_put_by_val): 17151 * jit/JITStubs.cpp: 17152 * jit/JITStubs.h: 17153 (JSC::): Check m_vectorLength instead of m_fastAccessCutoff when 17154 getting / putting from / to an array. Inline putting past the end of 17155 the array. 17156 17157 * runtime/JSArray.cpp: 17158 (JSC::JSArray::JSArray): 17159 (JSC::JSArray::getOwnPropertySlot): 17160 (JSC::JSArray::getOwnPropertyDescriptor): 17161 (JSC::JSArray::put): 17162 (JSC::JSArray::putSlowCase): 17163 (JSC::JSArray::deleteProperty): 17164 (JSC::JSArray::getOwnPropertyNames): 17165 (JSC::JSArray::increaseVectorLength): 17166 (JSC::JSArray::setLength): 17167 (JSC::JSArray::pop): 17168 (JSC::JSArray::push): 17169 (JSC::JSArray::sort): 17170 (JSC::JSArray::fillArgList): 17171 (JSC::JSArray::copyToRegisters): 17172 (JSC::JSArray::compactForSorting): 17173 (JSC::JSArray::checkConsistency): 17174 * runtime/JSArray.h: 17175 (JSC::JSArray::canGetIndex): 17176 (JSC::JSArray::canSetIndex): 17177 (JSC::JSArray::setIndex): 17178 (JSC::JSArray::markChildrenDirect): Removed m_fastAccessCutoff, and 17179 replaced with checks for JSValue() to detect reads and writes from / to 17180 uninitialized parts of the array. 17181 171822009-10-02 Jonni Rainisto <jonni.rainisto@nokia.com> 17183 17184 Reviewed by Darin Adler. 17185 17186 Math.random() gives too low values on Win32 when _CRT_RAND_S is not defined 17187 https://bugs.webkit.org/show_bug.cgi?id=29956 17188 17189 * wtf/RandomNumber.cpp: 17190 (WTF::randomNumber): Added PLATFORM(WIN_OS) to handle 15bit rand() 17191 171922009-10-02 Geoffrey Garen <ggaren@apple.com> 17193 17194 Reviewed by Sam Weinig. 17195 17196 Take one branch instead of two to test for JSValue(). 17197 17198 1.1% SunSpider speedup. 17199 17200 * jit/JITCall.cpp: 17201 (JSC::JIT::compileOpCall): 17202 * jit/JITOpcodes.cpp: 17203 (JSC::JIT::emit_op_to_jsnumber): 17204 (JSC::JIT::emit_op_create_arguments): 17205 * jit/JITPropertyAccess.cpp: 17206 (JSC::JIT::emitSlow_op_get_by_val): 17207 (JSC::JIT::emit_op_put_by_val): Test for the empty value tag, instead 17208 of testing for the cell tag with a 0 payload. 17209 17210 * runtime/JSValue.cpp: 17211 (JSC::JSValue::description): Added support for dumping the new empty value, 17212 and deleted values, in debug builds. 17213 17214 * runtime/JSValue.h: 17215 (JSC::JSValue::JSValue()): Construct JSValue() with the empty value tag. 17216 17217 (JSC::JSValue::JSValue(JSCell*)): Convert null pointer to the empty value 17218 tag, to avoid having two different c++ versions of null / empty. 17219 17220 (JSC::JSValue::operator bool): Test for the empty value tag, instead 17221 of testing for the cell tag with a 0 payload. 17222 172232009-10-02 Steve Falkenburg <sfalken@apple.com> 17224 17225 Reviewed by Mark Rowe. 17226 17227 <https://bugs.webkit.org/show_bug.cgi?id=29989> 17228 Safari version number shouldn't be exposed in WebKit code 17229 17230 For a WebKit version of 532.3.4: 17231 Product version is: 5.32.3.4 (was 4.0.3.0) 17232 File version is: 5.32.3.4 (was 4.532.3.4) 17233 17234 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.rc: 17235 172362009-10-02 Tor Arne Vestbø <tor.arne.vestbo@nokia.com> 17237 17238 Rubber-stamped by Simon Hausmann. 17239 17240 Fix the Qt on Mac OS X build. 17241 17242 * wtf/FastMalloc.cpp: 17243 172442009-10-02 Jørgen Lind <jorgen.lind@nokia.com> 17245 17246 Reviewed by Simon Hausmann. 17247 17248 Allow enabling and disabling of the JIT through a qmake variable. 17249 17250 Qt's configure may set this variable through .qmake.cache if a 17251 commandline option is given and/or the compile test for hwcap.h 17252 failed/succeeded. 17253 17254 * JavaScriptCore.pri: 17255 172562009-10-01 Mark Rowe <mrowe@apple.com> 17257 17258 Fix the Tiger build. Don't unconditionally enable 3D canvas as it is not supported on Tiger. 17259 17260 * Configurations/FeatureDefines.xcconfig: 17261 172622009-10-01 Yongjun Zhang <yongjun.zhang@nokia.com> 17263 17264 Reviewed by Darin Adler. 17265 17266 https://bugs.webkit.org/show_bug.cgi?id=29187 17267 17268 Don't inline ~ListRefPtr() to work around winscw compiler forward declaration 17269 bug regarding templated classes. 17270 17271 The compiler bug is reported at: 17272 https://xdabug001.ext.nokia.com/bugzilla/show_bug.cgi?id=9812 17273 17274 The change will be reverted when the above bug is fixed in winscw compiler. 17275 17276 * wtf/ListRefPtr.h: 17277 (WTF::::~ListRefPtr): 17278 172792009-10-01 Zoltan Horvath <zoltan@webkit.org> 17280 17281 Reviewed by Simon Hausmann. 17282 17283 [Qt] Allow custom memory allocation control for the whole JavaScriptCore 17284 https://bugs.webkit.org/show_bug.cgi?id=27029 17285 17286 Since in JavaScriptCore almost every class which has been instantiated by operator new is 17287 inherited from FastAllocBase (bug #20422), we disable customizing global operator new for the Qt-port 17288 when USE_SYSTEM_MALLOC=0. 17289 17290 Add #include <unistd.h> to FastMalloc.cpp because it's used by TCMalloc_PageHeap::scavengerThread(). 17291 (It's needed for the functionality of TCmalloc.) 17292 17293 Add TCSystemAlloc.cpp to JavaScriptCore.pri if USE_SYSTEM_MALLOC is disabled. 17294 17295 * JavaScriptCore.pri: 17296 * wtf/FastMalloc.cpp: 17297 (WTF::sleep): 17298 * wtf/FastMalloc.h: 17299 173002009-09-30 Gabor Loki <loki@inf.u-szeged.hu> 17301 17302 Reviewed by George Staikos. 17303 17304 Defines two pseudo-platforms for ARM and Thumb-2 instruction set. 17305 https://bugs.webkit.org/show_bug.cgi?id=29122 17306 17307 Introduces WTF_PLATFORM_ARM_TRADITIONAL and WTF_PLATFORM_ARM_THUMB2 17308 macros on ARM platforms. The PLATFORM(ARM_THUMB2) should be used 17309 when Thumb-2 instruction set is the required target. The 17310 PLATFORM(ARM_TRADITIONAL) is for generic ARM instruction set. In 17311 case where the code is common the PLATFORM(ARM) have to be used. 17312 17313 Modified by George Wright <gwright@rim.com> to correctly work 17314 with the RVCT-defined __TARGET_ARCH_ARM and __TARGET_ARCH_THUMB 17315 compiler macros, as well as adding readability changes. 17316 17317 * wtf/Platform.h: 17318 173192009-09-30 Oliver Hunt <oliver@apple.com> 17320 17321 Reviewed by Geoff Garen. 17322 17323 Devirtualise array toString conversion 17324 17325 Tweak the implementation of Array.prototype.toString to have a fast path 17326 when acting on a true JSArray. 17327 17328 * runtime/ArrayPrototype.cpp: 17329 (JSC::arrayProtoFuncToString): 17330 173312009-09-30 Csaba Osztrogonac <oszi@inf.u-szeged.hu> 17332 17333 Reviewed by Geoffrey Garen. 17334 17335 Buildfix for platforms using JSVALUE32. 17336 https://bugs.webkit.org/show_bug.cgi?id=29915 17337 17338 After http://trac.webkit.org/changeset/48905 the build broke in JSVALUE32 case. 17339 Also removed unreachable code. 17340 17341 * jit/JITArithmetic.cpp: 17342 (JSC::JIT::emit_op_add): 17343 - Declaration of "OperandTypes types" moved before first use. 17344 - Typos fixed: dst modified to result, regT2 added. 17345 - Unreachable code removed. 17346 (JSC::JIT::emitSlow_op_add): 17347 - Missing declaration of "OperandTypes types" added. 17348 173492009-09-30 Janne Koskinen <janne.p.koskinen@digia.com> 17350 17351 Reviewed by Simon Hausmann. 17352 17353 Reduce heap size on Symbian from 64MB to 8MB. 17354 17355 This is not a perfect fix, it requires more fine tuning. 17356 But this makes it possible again to debug in the emulator, 17357 which is more important in order to be able to fix other 17358 run-time issues. 17359 17360 * runtime/Collector.h: 17361 173622009-09-30 Janne Koskinen <janne.p.koskinen@digia.com> 17363 17364 Reviewed by Simon Hausmann. 17365 17366 Fix CRASH() macro for Symbian build. 17367 17368 * wtf/Assertions.h: Added missing } 17369 173702009-09-29 Geoffrey Garen <ggaren@apple.com> 17371 17372 Reviewed by Gavin Barraclough. 17373 17374 Inlined a few math operations. 17375 17376 ~1% SunSpider speedup. 17377 17378 * jit/JIT.h: 17379 * jit/JITArithmetic.cpp: 17380 (JSC::JIT::compileBinaryArithOpSlowCase): 17381 (JSC::JIT::emitSlow_op_add): 17382 (JSC::JIT::emitSlow_op_mul): 17383 (JSC::JIT::emit_op_sub): 17384 (JSC::JIT::emitSlow_op_sub): Don't take a stub call when operating on 17385 a constant int and a double. 17386 173872009-09-28 Oliver Hunt <oliver@apple.com> 17388 17389 Reviewed by Gavin Barraclough. 17390 17391 Tidy up codeblock sampler 17392 https://bugs.webkit.org/show_bug.cgi?id=29836 17393 17394 Some rather simple refactoring of codeblock sampler so that 17395 it's easier for us to use it to find problems in non-jsc 17396 environments 17397 17398 * JavaScriptCore.exp: 17399 * bytecode/SamplingTool.h: 17400 * debugger/Debugger.cpp: 17401 (JSC::evaluateInGlobalCallFrame): 17402 * debugger/DebuggerCallFrame.cpp: 17403 (JSC::DebuggerCallFrame::evaluate): 17404 * interpreter/Interpreter.cpp: 17405 (JSC::Interpreter::Interpreter): 17406 (JSC::Interpreter::execute): 17407 (JSC::Interpreter::privateExecute): 17408 (JSC::Interpreter::enableSampler): 17409 (JSC::Interpreter::dumpSampleData): 17410 (JSC::Interpreter::startSampling): 17411 (JSC::Interpreter::stopSampling): 17412 * interpreter/Interpreter.h: 17413 (JSC::Interpreter::sampler): 17414 * jit/JIT.h: 17415 * jsc.cpp: 17416 (runWithScripts): 17417 * runtime/Completion.cpp: 17418 (JSC::checkSyntax): 17419 (JSC::evaluate): 17420 * runtime/Executable.h: 17421 (JSC::EvalExecutable::EvalExecutable): 17422 (JSC::ProgramExecutable::create): 17423 (JSC::ProgramExecutable::ProgramExecutable): 17424 * runtime/JSGlobalData.cpp: 17425 (JSC::JSGlobalData::startSampling): 17426 (JSC::JSGlobalData::stopSampling): 17427 (JSC::JSGlobalData::dumpSampleData): 17428 * runtime/JSGlobalData.h: 17429 * runtime/JSGlobalObjectFunctions.cpp: 17430 (JSC::globalFuncEval): 17431 174322009-09-29 Jeremy Orlow <jorlow@chromium.org> 17433 17434 Reviewed by Dimitri Glazkov. 17435 17436 Add GYP generated files to svn:ignore 17437 https://bugs.webkit.org/show_bug.cgi?id=29895 17438 17439 The following files are generated by JavaScriptCore's GYP file and should be ignored: 17440 17441 pcre.mk 17442 wtf.scons 17443 wtf.mk 17444 SConstruct 17445 wtf_config.scons 17446 wtf_config.mk 17447 pcre.scons 17448 17449 * JavaScriptCore.gyp: Changed property svn:ignore. 17450 174512009-09-29 Geoffrey Garen <ggaren@apple.com> 17452 17453 Reviewed by Sam Weinig. 17454 17455 Standardized an optimization for adding non-numbers. 17456 17457 SunSpider says maybe a tiny speedup. 17458 17459 * jit/JITArithmetic.cpp: 17460 (JSC::JIT::emit_op_add): 17461 (JSC::JIT::emitSlow_op_add): 17462 174632009-09-29 Geoffrey Garen <ggaren@apple.com> 17464 17465 Windows build fix: export a new symbol. 17466 17467 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 17468 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 17469 174702009-09-28 Geoffrey Garen <ggaren@apple.com> 17471 17472 Reviewed by Sam Weinig. 17473 17474 Removed virtual destructor from JSGlobalObjectData to eliminate pointer 17475 fix-ups when accessing JSGlobalObject::d. 17476 17477 Replaced with an explicit destructor function pointer. 17478 17479 6% speedup on bench-alloc-nonretained.js. 17480 17481 * JavaScriptCore.exp: 17482 * runtime/JSGlobalObject.cpp: 17483 (JSC::JSGlobalObject::~JSGlobalObject): 17484 (JSC::JSGlobalObject::destroyJSGlobalObjectData): 17485 * runtime/JSGlobalObject.h: 17486 (JSC::JSGlobalObject::JSGlobalObjectData::JSGlobalObjectData): 17487 (JSC::JSGlobalObject::JSGlobalObject): 17488 174892009-09-29 Janne Koskinen <janne.p.koskinen@digia.com> 17490 17491 Reviewed by David Kilzer. 17492 17493 [Qt] Assert messages prints visible in Symbian 17494 https://bugs.webkit.org/show_bug.cgi?id=29808 17495 17496 Asserts use vprintf to print the messages to stderr. 17497 In Symbian Open C it is not possible to see stderr so 17498 I routed the messages to stdout instead. 17499 17500 * wtf/Assertions.cpp: 17501 175022009-09-29 Janne Koskinen <janne.p.koskinen@digia.com> 17503 17504 Reviewed by Darin Adler. 17505 17506 [Qt] Symbian CRASH macro implementation 17507 17508 Added Symbian specific crash macro that 17509 stops to crash line if JIT debugging is used. 17510 Additional differentiation of access violation 17511 (KERN-EXEC 3) and CRASH panic. 17512 17513 * wtf/Assertions.h: 17514 175152009-09-28 Mark Rowe <mrowe@apple.com> 17516 17517 Fix the PowerPC build. 17518 17519 * JavaScriptCore.exp: 17520 175212009-09-28 Mark Rowe <mrowe@apple.com> 17522 17523 Reviewed by Gavin Barraclough. 17524 17525 <rdar://problem/7195704> JavaScriptCore fails to mark registers when built for x86_64 using LLVM GCC. 17526 17527 * runtime/Collector.cpp: 17528 (JSC::Heap::markCurrentThreadConservatively): Force jmp_buf to use the appropriate alignment for a pointer 17529 to ensure that we correctly interpret the contents of registers during marking. 17530 175312009-09-28 Geoffrey Garen <ggaren@apple.com> 17532 17533 Windows build fix: added new exports. 17534 17535 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 17536 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 17537 175382009-09-28 Geoffrey Garen <ggaren@apple.com> 17539 17540 Windows build fix: removed exports that no longer exist. 17541 17542 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 17543 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 17544 175452009-09-28 Geoffrey Garen <ggaren@apple.com> 17546 17547 Reviewed by Darin Adler. 17548 17549 NotNullPassRefPtr: smart pointer optimized for passing references that are not null 17550 https://bugs.webkit.org/show_bug.cgi?id=29822 17551 17552 Added NotNullPassRefPtr, and deployed it in all places that initialize 17553 JavaScript objects. 17554 17555 2.2% speedup on bench-allocate-nonretained.js. 17556 17557 * API/JSCallbackConstructor.cpp: 17558 (JSC::JSCallbackConstructor::JSCallbackConstructor): 17559 * API/JSCallbackConstructor.h: 17560 * API/JSCallbackObject.h: 17561 * API/JSCallbackObjectFunctions.h: 17562 (JSC::JSCallbackObject::JSCallbackObject): 17563 * JavaScriptCore.exp: 17564 * bytecode/CodeBlock.h: 17565 (JSC::CodeBlock::addFunctionDecl): 17566 (JSC::CodeBlock::addFunctionExpr): 17567 * runtime/ArrayConstructor.cpp: 17568 (JSC::ArrayConstructor::ArrayConstructor): 17569 * runtime/ArrayConstructor.h: 17570 * runtime/ArrayPrototype.cpp: 17571 (JSC::ArrayPrototype::ArrayPrototype): 17572 * runtime/ArrayPrototype.h: 17573 * runtime/BooleanConstructor.cpp: 17574 (JSC::BooleanConstructor::BooleanConstructor): 17575 * runtime/BooleanConstructor.h: 17576 * runtime/BooleanObject.cpp: 17577 (JSC::BooleanObject::BooleanObject): 17578 * runtime/BooleanObject.h: 17579 * runtime/BooleanPrototype.cpp: 17580 (JSC::BooleanPrototype::BooleanPrototype): 17581 * runtime/BooleanPrototype.h: 17582 * runtime/DateConstructor.cpp: 17583 (JSC::DateConstructor::DateConstructor): 17584 * runtime/DateConstructor.h: 17585 * runtime/DateInstance.cpp: 17586 (JSC::DateInstance::DateInstance): 17587 * runtime/DateInstance.h: 17588 * runtime/DatePrototype.cpp: 17589 (JSC::DatePrototype::DatePrototype): 17590 * runtime/DatePrototype.h: 17591 * runtime/ErrorConstructor.cpp: 17592 (JSC::ErrorConstructor::ErrorConstructor): 17593 * runtime/ErrorConstructor.h: 17594 * runtime/ErrorInstance.cpp: 17595 (JSC::ErrorInstance::ErrorInstance): 17596 * runtime/ErrorInstance.h: 17597 * runtime/ErrorPrototype.cpp: 17598 (JSC::ErrorPrototype::ErrorPrototype): 17599 * runtime/ErrorPrototype.h: 17600 * runtime/FunctionConstructor.cpp: 17601 (JSC::FunctionConstructor::FunctionConstructor): 17602 * runtime/FunctionConstructor.h: 17603 * runtime/FunctionPrototype.cpp: 17604 (JSC::FunctionPrototype::FunctionPrototype): 17605 * runtime/FunctionPrototype.h: 17606 * runtime/GlobalEvalFunction.cpp: 17607 (JSC::GlobalEvalFunction::GlobalEvalFunction): 17608 * runtime/GlobalEvalFunction.h: 17609 * runtime/InternalFunction.cpp: 17610 (JSC::InternalFunction::InternalFunction): 17611 * runtime/InternalFunction.h: 17612 (JSC::InternalFunction::InternalFunction): 17613 * runtime/JSActivation.cpp: 17614 (JSC::JSActivation::JSActivation): 17615 * runtime/JSActivation.h: 17616 (JSC::JSActivation::JSActivationData::JSActivationData): 17617 * runtime/JSArray.cpp: 17618 (JSC::JSArray::JSArray): 17619 * runtime/JSArray.h: 17620 * runtime/JSByteArray.cpp: 17621 (JSC::JSByteArray::JSByteArray): 17622 * runtime/JSByteArray.h: 17623 * runtime/JSFunction.cpp: 17624 (JSC::JSFunction::JSFunction): 17625 * runtime/JSFunction.h: 17626 * runtime/JSGlobalObject.h: 17627 (JSC::JSGlobalObject::JSGlobalObject): 17628 * runtime/JSONObject.h: 17629 (JSC::JSONObject::JSONObject): 17630 * runtime/JSObject.h: 17631 (JSC::JSObject::JSObject): 17632 (JSC::JSObject::setStructure): 17633 * runtime/JSVariableObject.h: 17634 (JSC::JSVariableObject::JSVariableObject): 17635 * runtime/JSWrapperObject.h: 17636 (JSC::JSWrapperObject::JSWrapperObject): 17637 * runtime/MathObject.cpp: 17638 (JSC::MathObject::MathObject): 17639 * runtime/MathObject.h: 17640 * runtime/NativeErrorConstructor.cpp: 17641 (JSC::NativeErrorConstructor::NativeErrorConstructor): 17642 * runtime/NativeErrorConstructor.h: 17643 * runtime/NativeErrorPrototype.cpp: 17644 (JSC::NativeErrorPrototype::NativeErrorPrototype): 17645 * runtime/NativeErrorPrototype.h: 17646 * runtime/NumberConstructor.cpp: 17647 (JSC::NumberConstructor::NumberConstructor): 17648 * runtime/NumberConstructor.h: 17649 * runtime/NumberObject.cpp: 17650 (JSC::NumberObject::NumberObject): 17651 * runtime/NumberObject.h: 17652 * runtime/NumberPrototype.cpp: 17653 (JSC::NumberPrototype::NumberPrototype): 17654 * runtime/NumberPrototype.h: 17655 * runtime/ObjectConstructor.cpp: 17656 (JSC::ObjectConstructor::ObjectConstructor): 17657 * runtime/ObjectConstructor.h: 17658 * runtime/ObjectPrototype.cpp: 17659 (JSC::ObjectPrototype::ObjectPrototype): 17660 * runtime/ObjectPrototype.h: 17661 * runtime/PropertyNameArray.h: 17662 (JSC::PropertyNameArrayData::setCachedPrototypeChain): 17663 * runtime/PrototypeFunction.cpp: 17664 (JSC::PrototypeFunction::PrototypeFunction): 17665 * runtime/PrototypeFunction.h: 17666 * runtime/RegExpConstructor.cpp: 17667 (JSC::RegExpConstructor::RegExpConstructor): 17668 * runtime/RegExpConstructor.h: 17669 * runtime/RegExpObject.cpp: 17670 (JSC::RegExpObject::RegExpObject): 17671 * runtime/RegExpObject.h: 17672 (JSC::RegExpObject::RegExpObjectData::RegExpObjectData): 17673 * runtime/RegExpPrototype.cpp: 17674 (JSC::RegExpPrototype::RegExpPrototype): 17675 * runtime/RegExpPrototype.h: 17676 * runtime/StringConstructor.cpp: 17677 (JSC::StringConstructor::StringConstructor): 17678 * runtime/StringConstructor.h: 17679 * runtime/StringObject.cpp: 17680 (JSC::StringObject::StringObject): 17681 * runtime/StringObject.h: 17682 * runtime/StringObjectThatMasqueradesAsUndefined.h: 17683 (JSC::StringObjectThatMasqueradesAsUndefined::StringObjectThatMasqueradesAsUndefined): 17684 * runtime/StringPrototype.cpp: 17685 (JSC::StringPrototype::StringPrototype): 17686 * runtime/StringPrototype.h: 17687 * wtf/PassRefPtr.h: 17688 (WTF::NotNullPassRefPtr::NotNullPassRefPtr): 17689 (WTF::NotNullPassRefPtr::~NotNullPassRefPtr): 17690 (WTF::NotNullPassRefPtr::get): 17691 (WTF::NotNullPassRefPtr::clear): 17692 (WTF::NotNullPassRefPtr::releaseRef): 17693 (WTF::NotNullPassRefPtr::operator*): 17694 (WTF::NotNullPassRefPtr::operator->): 17695 (WTF::NotNullPassRefPtr::operator!): 17696 (WTF::NotNullPassRefPtr::operator UnspecifiedBoolType): 17697 * wtf/RefPtr.h: 17698 (WTF::RefPtr::RefPtr): 17699 (WTF::operator==): 17700 177012009-09-28 Oliver Hunt <oliver@apple.com> 17702 17703 Reviewed by Geoff Garen. 17704 17705 Hard dependency on SSE2 instruction set with JIT 17706 https://bugs.webkit.org/show_bug.cgi?id=29779 17707 17708 Add floating point support checks to op_jfalse and op_jtrue, and 17709 fix the logic for the slow case of op_add 17710 17711 * jit/JITArithmetic.cpp: 17712 (JSC::JIT::emitSlow_op_add): 17713 * jit/JITOpcodes.cpp: 17714 (JSC::JIT::emit_op_jfalse): 17715 (JSC::JIT::emit_op_jtrue): 17716 177172009-09-28 Yaar Schnitman <yaar@chromium.org> 17718 17719 Reviewed by Dimitri Glazkov. 17720 17721 Chromium port - recognize we are being built independently 17722 of chromium and look for dependencies under webkit/chromium rather 17723 than chromium/src. 17724 17725 https://bugs.webkit.org/show_bug.cgi?id=29722 17726 17727 * JavaScriptCore.gyp/JavaScriptCore.gyp: 17728 177292009-09-28 Jakub Wieczorek <faw217@gmail.com> 17730 17731 Reviewed by Simon Hausmann. 17732 17733 [Qt] Implement XSLT support with QtXmlPatterns. 17734 https://bugs.webkit.org/show_bug.cgi?id=28303 17735 17736 * wtf/Platform.h: Add a WTF_USE_QXMLQUERY #define. 17737 177382009-09-28 Gabor Loki <loki@inf.u-szeged.hu> 17739 17740 Reviewed by Simon Hausmann. 17741 17742 Remove __clear_cache which is an internal function of GCC 17743 https://bugs.webkit.org/show_bug.cgi?id=28886 17744 17745 Although __clear_cache is exported from GCC, this is an internal 17746 function. GCC makes no promises about it. 17747 17748 * jit/ExecutableAllocator.h: 17749 (JSC::ExecutableAllocator::cacheFlush): 17750 177512009-09-28 Sam Weinig <sam@webkit.org> 17752 17753 Reviewed by Oliver Hunt. 17754 17755 Fix an absolute path to somewhere in Oliver's machine to a relative path 17756 for derived JSONObject.lut.h. 17757 17758 * JavaScriptCore.xcodeproj/project.pbxproj: 17759 177602009-09-28 Joerg Bornemann <joerg.bornemann@nokia.com> 17761 17762 Reviewed by Simon Hausmann. 17763 17764 Add ARM version detection for Windows CE. 17765 17766 * wtf/Platform.h: 17767 177682009-09-26 Yongjun Zhang <yongjun.zhang@nokia.com> 17769 17770 Reviewed by Simon Hausmann. 17771 17772 Add MarkStackSymbian.cpp to build JavascriptCore for Symbian. 17773 17774 Re-use Windows shrinkAllocation implementation because Symbian doesn't 17775 support releasing part of memory region. 17776 17777 Use fastMalloc and fastFree to implement allocateStack and releaseStack 17778 for Symbian port. 17779 17780 * JavaScriptCore.pri: 17781 * runtime/MarkStack.h: 17782 (JSC::MarkStack::MarkStackArray::shrinkAllocation): 17783 * runtime/MarkStackSymbian.cpp: Added. 17784 (JSC::MarkStack::initializePagesize): 17785 (JSC::MarkStack::allocateStack): 17786 (JSC::MarkStack::releaseStack): 17787 177882009-09-25 Gabor Loki <loki@inf.u-szeged.hu> 17789 17790 Reviewed by Gavin Barraclough. 17791 17792 Fix unaligned data access in YARR_JIT on ARMv5 and below. 17793 https://bugs.webkit.org/show_bug.cgi?id=29695 17794 17795 On ARMv5 and below all data access should be naturally aligned. 17796 In the YARR_JIT there is a case when character pairs are 17797 loaded from the input string, but this data access is not 17798 naturally aligned. This fix introduces load32WithUnalignedHalfWords 17799 and branch32WithUnalignedHalfWords functions which contain 17800 naturally aligned memory loads - half word loads - on ARMv5 and below. 17801 17802 * assembler/MacroAssemblerARM.cpp: 17803 (JSC::MacroAssemblerARM::load32WithUnalignedHalfWords): 17804 * assembler/MacroAssemblerARM.h: 17805 (JSC::MacroAssemblerARM::load32WithUnalignedHalfWords): 17806 (JSC::MacroAssemblerARM::branch32WithUnalignedHalfWords): 17807 * assembler/MacroAssemblerARMv7.h: 17808 (JSC::MacroAssemblerARMv7::load32WithUnalignedHalfWords): 17809 (JSC::MacroAssemblerARMv7::branch32): 17810 (JSC::MacroAssemblerARMv7::branch32WithUnalignedHalfWords): 17811 * assembler/MacroAssemblerX86Common.h: 17812 (JSC::MacroAssemblerX86Common::load32WithUnalignedHalfWords): 17813 (JSC::MacroAssemblerX86Common::branch32WithUnalignedHalfWords): 17814 * wtf/Platform.h: 17815 * yarr/RegexJIT.cpp: 17816 (JSC::Yarr::RegexGenerator::generatePatternCharacterPair): 17817 178182009-09-25 Jeremy Orlow <jorlow@chromium.org> 17819 17820 This is breaking Chromium try bots, so I'm counting this as a build fix. 17821 17822 Add more svn:ignore exceptions. On different platforms, these files are 17823 generated with different case for JavaScriptCore. Also there are some 17824 wtf project files that get built apparently. 17825 17826 * JavaScriptCore.gyp: Changed property svn:ignore. 17827 178282009-09-25 Ada Chan <adachan@apple.com> 17829 17830 Build fix. 17831 17832 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 17833 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 17834 178352009-09-25 Geoffrey Garen <ggaren@apple.com> 17836 17837 Reviewed by Darin Adler. 17838 17839 Inlined some object creation code, including lexicalGlobalObject access 17840 https://bugs.webkit.org/show_bug.cgi?id=29750 17841 17842 SunSpider says 0.5% faster. 17843 17844 0.8% speedup on bench-alloc-nonretained.js. 17845 2.5% speedup on v8-splay.js. 17846 17847 * interpreter/CachedCall.h: 17848 (JSC::CachedCall::CachedCall): 17849 * interpreter/CallFrame.h: 17850 (JSC::ExecState::lexicalGlobalObject): 17851 (JSC::ExecState::globalThisValue): 17852 * interpreter/Interpreter.cpp: 17853 (JSC::Interpreter::dumpRegisters): 17854 (JSC::Interpreter::execute): 17855 (JSC::Interpreter::privateExecute): 17856 * jit/JITStubs.cpp: 17857 (JSC::DEFINE_STUB_FUNCTION): 17858 * runtime/FunctionConstructor.cpp: 17859 (JSC::constructFunction): 17860 * runtime/ScopeChain.cpp: 17861 (JSC::ScopeChainNode::print): 17862 * runtime/ScopeChain.h: 17863 (JSC::ScopeChainNode::ScopeChainNode): 17864 (JSC::ScopeChainNode::~ScopeChainNode): 17865 (JSC::ScopeChainNode::push): 17866 (JSC::ScopeChain::ScopeChain): 17867 (JSC::ScopeChain::globalObject): Added a globalObject data member to ScopeChainNode. 17868 Replaced accessor function for globalObject() with data member. Replaced 17869 globalThisObject() accessor with direct access to globalThis, to match. 17870 17871 * runtime/JSGlobalObject.cpp: 17872 (JSC::JSGlobalObject::init): 17873 * runtime/JSGlobalObject.h: Inlined array and object construction. 17874 178752009-09-25 Laszlo Gombos <laszlo.1.gombos@nokia.com> 17876 17877 Reviewed by Gavin Barraclough. 17878 17879 Add ARM version detection rules for Symbian 17880 https://bugs.webkit.org/show_bug.cgi?id=29715 17881 17882 * wtf/Platform.h: 17883 178842009-09-24 Xan Lopez <xlopez@igalia.com> 17885 17886 Reviewed by Mark "Do It!" Rowe. 17887 17888 Some GCC versions don't like C++-style comments in preprocessor 17889 directives, change to C-style to shut them up. 17890 17891 * wtf/Platform.h: 17892 178932009-09-24 Oliver Hunt <oliver@apple.com> 17894 17895 Reviewed by Gavin Barraclough. 17896 17897 Division is needlessly slow in 64-bit 17898 https://bugs.webkit.org/show_bug.cgi?id=29723 17899 17900 Add codegen for op_div on x86-64 17901 17902 * jit/JIT.cpp: 17903 (JSC::JIT::privateCompileMainPass): 17904 (JSC::JIT::privateCompileSlowCases): 17905 * jit/JIT.h: 17906 * jit/JITArithmetic.cpp: 17907 (JSC::JIT::compileBinaryArithOpSlowCase): 17908 (JSC::JIT::emit_op_div): 17909 (JSC::JIT::emitSlow_op_div): 17910 * jit/JITInlineMethods.h: 17911 (JSC::JIT::isOperandConstantImmediateDouble): 17912 (JSC::JIT::addressFor): 17913 (JSC::JIT::emitLoadDouble): 17914 (JSC::JIT::emitLoadInt32ToDouble): 17915 (JSC::JIT::emitJumpSlowCaseIfNotImmediateNumber): 17916 179172009-09-24 Jeremy Orlow <jorlow@chromium.org> 17918 17919 Reviewed by Dimitri Glazkov. 17920 17921 Add GYP generated files to svn:ignore 17922 https://bugs.webkit.org/show_bug.cgi?id=29724 17923 17924 Adding the following files to the svn:ignore list (all in the 17925 JavaScriptCore/JavaScriptCore.gyp directory) 17926 17927 JavaScriptCore.xcodeproj 17928 JavaScriptCore.sln 17929 JavaScriptCore.vcproj 17930 JavaScriptCore_Debug.rules 17931 JavaScriptCore_Release.rules 17932 JavaScriptCore_Release - no tcmalloc.rules 17933 JavaScriptCore_Purify.rules 17934 JavaScriptCore.mk 17935 JavaScriptCore_Debug_rules.mk 17936 JavaScriptCore_Release_rules.mk 17937 JavaScriptCore_Release - no tcmalloc_rules.mk 17938 JavaScriptCore_Purify_rules.mk 17939 JavaScriptCore.scons 17940 JavaScriptCore_main.scons 17941 17942 * JavaScriptCore.gyp: Changed property svn:ignore. 17943 179442009-09-24 Yong Li <yong.li@torchmobile.com> 17945 17946 Reviewed by Adam Barth. 17947 17948 Replace platform-dependent code with WTF::currentTime() 17949 https://bugs.webkit.org/show_bug.cgi?id=29148 17950 17951 * jsc.cpp: 17952 (StopWatch::start): 17953 (StopWatch::stop): 17954 (StopWatch::getElapsedMS): 17955 * runtime/TimeoutChecker.cpp: 17956 (JSC::getCPUTime): 17957 179582009-09-24 Mark Rowe <mrowe@apple.com> 17959 17960 Reviewed by Sam Weinig. 17961 17962 <rdar://problem/7215058> FastMalloc scavenging thread should be named 17963 17964 * wtf/FastMalloc.cpp: 17965 (WTF::TCMalloc_PageHeap::scavengerThread): Set the thread name. 17966 * wtf/Platform.h: Move the knowledge of whether pthread_setname_np exists to here as HAVE(PTHREAD_SETNAME_NP). 17967 * wtf/ThreadingPthreads.cpp: 17968 (WTF::setThreadNameInternal): Use HAVE(PTHREAD_SETNAME_NP). 17969 179702009-09-24 Geoffrey Garen <ggaren@apple.com> 17971 17972 Reviewed by Sam Weinig. 17973 17974 Renamed clear to removeAll, as suggested by Darin Adler. 17975 17976 * wtf/HashCountedSet.h: 17977 (WTF::::removeAll): 17978 179792009-09-24 Mark Rowe <mrowe@apple.com> 17980 17981 Reviewed by Gavin Barraclough. 17982 17983 Fix FastMalloc to build with assertions enabled. 17984 17985 * wtf/FastMalloc.cpp: 17986 (WTF::TCMalloc_Central_FreeList::ReleaseToSpans): 17987 * wtf/TCSpinLock.h: 17988 (TCMalloc_SpinLock::IsHeld): 17989 179902009-09-24 Geoffrey Garen <ggaren@apple.com> 17991 17992 Suggested by Darin Adler. 17993 17994 Removed some unnecessary parameter names. 17995 17996 * wtf/HashCountedSet.h: 17997 179982009-09-24 Janne Koskinen <janne.p.koskinen@digia.com> 17999 18000 Reviewed by Simon Hausmann. 18001 18002 On Windows JSChar is typedef'ed to wchar_t. 18003 18004 When building with WINSCW for Symbian we need to do the 18005 same typedef. 18006 18007 * API/JSStringRef.h: 18008 180092009-09-23 Geoffrey Garen <ggaren@apple.com> 18010 18011 A piece of my last patch that I forgot. 18012 18013 * wtf/HashCountedSet.h: 18014 (WTF::::clear): Added HashCountedSet::clear. 18015 180162009-09-24 Gabor Loki <loki@inf.u-szeged.hu> 18017 18018 Reviewed by Gavin Barraclough. 18019 18020 Avoid __clear_cache built-in function if DISABLE_BUILTIN_CLEAR_CACHE define is set 18021 https://bugs.webkit.org/show_bug.cgi?id=28886 18022 18023 There are some GCC packages (for example GCC-2006q3 from CodeSourcery) 18024 which contain __clear_cache built-in function only for C while the C++ 18025 version of __clear_cache is missing on ARM architectures. 18026 18027 Fixed a small bug in the inline assembly of cacheFlush function on 18028 ARM_TRADITIONAL. 18029 18030 * jit/ExecutableAllocator.h: 18031 (JSC::ExecutableAllocator::cacheFlush): 18032 180332009-09-23 Geoffrey Garen <ggaren@apple.com> 18034 18035 Reviewed by Sam Weinig. 18036 18037 Added the ability to swap vectors with inline capacities, so you can 18038 store a vector with inline capacity in a hash table. 18039 18040 * wtf/Vector.h: 18041 (WTF::swap): 18042 (WTF::VectorBuffer::swap): 18043 180442009-09-23 David Kilzer <ddkilzer@apple.com> 18045 18046 Move definition of USE(PLUGIN_HOST_PROCESS) from WebKitPrefix.h to Platform.h 18047 18048 Reviewed by Mark Rowe. 18049 18050 * wtf/Platform.h: Define WTF_USE_PLUGIN_HOST_PROCESS to 1 when 18051 building on 64-bit SnowLeopard. Define to 0 elsewhere. 18052 180532009-09-22 Oliver Hunt <oliver@apple.com> 18054 18055 Reviewed by Geoff Garen. 18056 18057 Code sampling builds are broken. 18058 https://bugs.webkit.org/show_bug.cgi?id=29662 18059 18060 Fix build. 18061 18062 * bytecode/EvalCodeCache.h: 18063 (JSC::EvalCodeCache::get): 18064 * bytecode/SamplingTool.cpp: 18065 (JSC::ScriptSampleRecord::sample): 18066 (JSC::SamplingTool::doRun): 18067 (JSC::SamplingTool::notifyOfScope): 18068 (JSC::compareScriptSampleRecords): 18069 (JSC::SamplingTool::dump): 18070 * bytecode/SamplingTool.h: 18071 (JSC::ScriptSampleRecord::ScriptSampleRecord): 18072 (JSC::ScriptSampleRecord::~ScriptSampleRecord): 18073 (JSC::SamplingTool::SamplingTool): 18074 * bytecompiler/BytecodeGenerator.cpp: 18075 (JSC::BytecodeGenerator::BytecodeGenerator): 18076 (JSC::BytecodeGenerator::emitNewFunction): 18077 (JSC::BytecodeGenerator::emitNewFunctionExpression): 18078 * bytecompiler/BytecodeGenerator.h: 18079 (JSC::BytecodeGenerator::makeFunction): 18080 * debugger/Debugger.cpp: 18081 (JSC::evaluateInGlobalCallFrame): 18082 * debugger/DebuggerCallFrame.cpp: 18083 (JSC::DebuggerCallFrame::evaluate): 18084 * parser/Nodes.cpp: 18085 (JSC::ScopeNode::ScopeNode): 18086 * runtime/Completion.cpp: 18087 (JSC::checkSyntax): 18088 (JSC::evaluate): 18089 * runtime/Executable.cpp: 18090 (JSC::FunctionExecutable::fromGlobalCode): 18091 * runtime/Executable.h: 18092 (JSC::ScriptExecutable::ScriptExecutable): 18093 (JSC::EvalExecutable::EvalExecutable): 18094 (JSC::EvalExecutable::create): 18095 (JSC::ProgramExecutable::ProgramExecutable): 18096 (JSC::FunctionExecutable::create): 18097 (JSC::FunctionExecutable::FunctionExecutable): 18098 * runtime/JSGlobalObjectFunctions.cpp: 18099 (JSC::globalFuncEval): 18100 181012009-09-22 Darin Adler <darin@apple.com> 18102 18103 Reviewed by Sam Weinig. 18104 18105 * wtf/Forward.h: Added PassOwnPtr. 18106 181072009-09-22 Yaar Schnitman <yaar@chromium.org> 18108 18109 Reviewed by David Levin. 18110 18111 Ported chromium.org's javascriptcore.gyp for the webkit chromium port. 18112 18113 https://bugs.webkit.org/show_bug.cgi?id=29617 18114 18115 * JavaScriptCore.gyp/JavaScriptCore.gyp: Added. 18116 181172009-09-22 Thiago Macieira <thiago.macieira@nokia.com> 18118 18119 Reviewed by Simon Hausmann. 18120 18121 Fix compilation with WINSCW: no varargs macros 18122 18123 Disable variadic arguments for WINSCW just like we do 18124 for MSVC7. 18125 18126 * wtf/Assertions.h: 18127 181282009-09-22 Kent Hansen <khansen@trolltech.com> 18129 18130 Reviewed by Simon Hausmann. 18131 18132 Disable variadic macros on MSVC7. 18133 18134 This was originally added in r26589 but not extended 18135 when LOG_DISABLED/ASSERT_DISABLED was introduced. 18136 18137 * wtf/Assertions.h: 18138 181392009-09-22 Simon Hausmann <simon.hausmann@nokia.com> 18140 18141 Unreviewed build fix for Windows CE < 5 18142 18143 Define WINCEBASIC to disable the IsDebuggerPresent() code in 18144 wtf/Assertions.cpp. 18145 18146 * JavaScriptCore.pri: 18147 181482009-09-22 Joerg Bornemann <joerg.bornemann@nokia.com> 18149 18150 Reviewed by Simon Hausmann. 18151 18152 Fix major memory leak in JavaScriptCore RegisterFile on Windows CE 18153 18154 https://bugs.webkit.org/show_bug.cgi?id=29367 18155 18156 On Widows CE we must decommit all committed pages before we release 18157 them. See VirtualFree documentation. 18158 Desktop Windows behaves much smoother in this situation. 18159 18160 * interpreter/RegisterFile.cpp: 18161 (JSC::RegisterFile::~RegisterFile): 18162 181632009-09-21 Greg Bolsinga <bolsinga@apple.com> 18164 18165 Reviewed by Simon Fraser & Sam Weinig. 18166 18167 Add ENABLE(ORIENTATION_EVENTS) 18168 https://bugs.webkit.org/show_bug.cgi?id=29508 18169 18170 * wtf/Platform.h: Also sort PLATFORM(IPHONE) #defines. 18171 181722009-09-21 Jedrzej Nowacki <jedrzej.nowacki@nokia.com> 18173 18174 Reviewed by Eric Seidel. 18175 18176 [Fix] SourceCode's uninitialized member 18177 18178 Potential source of crashes and bugs was fixed. Default constructor 18179 didn't initialized m_provider member. 18180 18181 https://bugs.webkit.org/show_bug.cgi?id=29364 18182 18183 * parser/SourceCode.h: 18184 (JSC::SourceCode::SourceCode): 18185 181862009-09-21 Oliver Hunt <oliver@apple.com> 18187 18188 Reviewed by Geoff Garen. 18189 18190 REGRESSION (r48582): Crash in StructureStubInfo::initPutByIdTransition when reloading trac.webkit.org 18191 https://bugs.webkit.org/show_bug.cgi?id=29599 18192 18193 It is unsafe to attempt to cache new property transitions on 18194 dictionaries of any type. 18195 18196 * interpreter/Interpreter.cpp: 18197 (JSC::Interpreter::tryCachePutByID): 18198 * jit/JITStubs.cpp: 18199 (JSC::JITThunks::tryCachePutByID): 18200 182012009-09-21 Oliver Hunt <oliver@apple.com> 18202 18203 RS=Maciej Stachowiak. 18204 18205 Re-land SNES fix with corrected assertion. 18206 18207 * interpreter/Interpreter.cpp: 18208 (JSC::Interpreter::resolveGlobal): 18209 (JSC::Interpreter::tryCachePutByID): 18210 (JSC::Interpreter::tryCacheGetByID): 18211 * jit/JITStubs.cpp: 18212 (JSC::JITThunks::tryCachePutByID): 18213 (JSC::JITThunks::tryCacheGetByID): 18214 (JSC::DEFINE_STUB_FUNCTION): 18215 * runtime/BatchedTransitionOptimizer.h: 18216 (JSC::BatchedTransitionOptimizer::BatchedTransitionOptimizer): 18217 * runtime/JSObject.cpp: 18218 (JSC::JSObject::removeDirect): 18219 * runtime/Structure.cpp: 18220 (JSC::Structure::Structure): 18221 (JSC::Structure::getEnumerablePropertyNames): 18222 (JSC::Structure::despecifyDictionaryFunction): 18223 (JSC::Structure::addPropertyTransitionToExistingStructure): 18224 (JSC::Structure::addPropertyTransition): 18225 (JSC::Structure::removePropertyTransition): 18226 (JSC::Structure::toDictionaryTransition): 18227 (JSC::Structure::toCacheableDictionaryTransition): 18228 (JSC::Structure::toUncacheableDictionaryTransition): 18229 (JSC::Structure::fromDictionaryTransition): 18230 (JSC::Structure::removePropertyWithoutTransition): 18231 * runtime/Structure.h: 18232 (JSC::Structure::isDictionary): 18233 (JSC::Structure::isUncacheableDictionary): 18234 (JSC::Structure::): 18235 * runtime/StructureChain.cpp: 18236 (JSC::StructureChain::isCacheable): 18237 182382009-09-21 Adam Roben <aroben@apple.com> 18239 18240 Revert r48573, as it caused many assertion failures 18241 18242 * interpreter/Interpreter.cpp: 18243 * jit/JITStubs.cpp: 18244 * runtime/BatchedTransitionOptimizer.h: 18245 * runtime/JSObject.cpp: 18246 * runtime/Structure.cpp: 18247 * runtime/Structure.h: 18248 * runtime/StructureChain.cpp: 18249 182502009-09-21 Gustavo Noronha Silva <gustavo.noronha@collabora.co.uk> 18251 18252 Unreviewed make dist build fix. Missing files. 18253 18254 * GNUmakefile.am: 18255 182562009-09-19 Gavin Barraclough <barraclough@apple.com> 18257 18258 Reviewed by Sam 'Cabin Boy' Weinig. 18259 18260 Fix stack alignment with ARM THUMB2 JIT. 18261 https://bugs.webkit.org/show_bug.cgi?id=29526 18262 18263 Stack is currently being decremented by 0x3c, bump this to 0x40 to make this a 18264 multiple of 16 bytes. 18265 18266 * jit/JITStubs.cpp: 18267 (JSC::JITThunks::JITThunks): 18268 * jit/JITStubs.h: 18269 182702009-09-20 Oliver Hunt <oliver@apple.com> 18271 18272 Reviewed by Maciej Stachowiak. 18273 18274 SNES is too slow 18275 https://bugs.webkit.org/show_bug.cgi?id=29534 18276 18277 The problem was that the emulator used multiple classes with 18278 more properties than our dictionary cutoff allowed, this resulted 18279 in more or less all critical logic inside the emulator requiring 18280 uncached property access. 18281 18282 Rather than simply bumping the dictionary cutoff, this patch 18283 recognises that there are two ways to create a "dictionary" 18284 structure. Either by adding a large number of properties, or 18285 by removing a property. In the case of adding properties we 18286 know all the existing properties will maintain their existing 18287 offsets, so we could cache access to those properties, if we 18288 know they won't be removed. 18289 18290 To make this possible, this patch adds the logic required to 18291 distinguish a dictionary created by addition from one created 18292 by removal. With this logic in place we can now cache access 18293 to objects with large numbers of properties. 18294 18295 SNES performance improved by more than 6x. 18296 18297 * interpreter/Interpreter.cpp: 18298 (JSC::Interpreter::resolveGlobal): 18299 (JSC::Interpreter::tryCachePutByID): 18300 (JSC::Interpreter::tryCacheGetByID): 18301 * jit/JITStubs.cpp: 18302 (JSC::JITThunks::tryCachePutByID): 18303 (JSC::JITThunks::tryCacheGetByID): 18304 (JSC::DEFINE_STUB_FUNCTION): 18305 * runtime/BatchedTransitionOptimizer.h: 18306 (JSC::BatchedTransitionOptimizer::BatchedTransitionOptimizer): 18307 * runtime/JSObject.cpp: 18308 (JSC::JSObject::removeDirect): 18309 * runtime/Structure.cpp: 18310 (JSC::Structure::Structure): 18311 (JSC::Structure::getEnumerablePropertyNames): 18312 (JSC::Structure::despecifyDictionaryFunction): 18313 (JSC::Structure::addPropertyTransitionToExistingStructure): 18314 (JSC::Structure::addPropertyTransition): 18315 (JSC::Structure::removePropertyTransition): 18316 (JSC::Structure::toDictionaryTransition): 18317 (JSC::Structure::toCacheableDictionaryTransition): 18318 (JSC::Structure::toUncacheableDictionaryTransition): 18319 (JSC::Structure::fromDictionaryTransition): 18320 (JSC::Structure::removePropertyWithoutTransition): 18321 * runtime/Structure.h: 18322 (JSC::Structure::isDictionary): 18323 (JSC::Structure::isUncacheableDictionary): 18324 (JSC::Structure::): 18325 * runtime/StructureChain.cpp: 18326 (JSC::StructureChain::isCacheable): 18327 183282009-09-19 Oliver Hunt <oliver@apple.com> 18329 18330 Reviewed by Maciej Stachowiak. 18331 18332 Implement ES5 Object.create function 18333 https://bugs.webkit.org/show_bug.cgi?id=29524 18334 18335 Implement Object.create. Very simple patch, effectively Object.defineProperties 18336 only creating the target object itself. 18337 18338 * runtime/CommonIdentifiers.h: 18339 * runtime/ObjectConstructor.cpp: 18340 (JSC::ObjectConstructor::ObjectConstructor): 18341 (JSC::objectConstructorCreate): 18342 183432009-09-19 Dan Bernstein <mitz@apple.com> 18344 18345 Fix clean debug builds. 18346 18347 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 18348 183492009-09-19 Joerg Bornemann <joerg.bornemann@nokia.com> 18350 18351 Reviewed by George Staikos. 18352 18353 QtWebKit Windows CE compile fix 18354 18355 https://bugs.webkit.org/show_bug.cgi?id=29379 18356 18357 There is no _aligned_alloc or _aligned_free on Windows CE. 18358 We just use the Windows code that was there before and use VirtualAlloc. 18359 But that also means that the BLOCK_SIZE must be 64K as this function 18360 allocates on 64K boundaries. 18361 18362 * runtime/Collector.cpp: 18363 (JSC::Heap::allocateBlock): 18364 (JSC::Heap::freeBlock): 18365 * runtime/Collector.h: 18366 183672009-09-19 Oliver Hunt <oliver@apple.com> 18368 18369 Reviewed by Sam Weinig. 18370 18371 Implement ES5 Object.defineProperties function 18372 https://bugs.webkit.org/show_bug.cgi?id=29522 18373 18374 Implement Object.defineProperties. Fairly simple patch, simply makes use of 18375 existing functionality used for defineProperty. 18376 18377 * runtime/CommonIdentifiers.h: 18378 * runtime/ObjectConstructor.cpp: 18379 (JSC::ObjectConstructor::ObjectConstructor): 18380 (JSC::defineProperties): 18381 (JSC::objectConstructorDefineProperties): 18382 183832009-09-19 Oliver Hunt <oliver@apple.com> 18384 18385 Reviewed by NOBODY (Build fix). 18386 18387 Windows build fix part2 18388 18389 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 18390 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 18391 183922009-09-19 Oliver Hunt <oliver@apple.com> 18393 18394 Reviewed by NOBODY (Buildfix). 18395 18396 Windows build fix part 1. 18397 18398 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 18399 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 18400 184012009-09-18 Oliver Hunt <oliver@apple.com> 18402 18403 Reviewed by Geoff Garen. 18404 18405 Implement ES5 Object.defineProperty function 18406 https://bugs.webkit.org/show_bug.cgi?id=29503 18407 18408 Implement Object.defineProperty. This requires adding the API to 18409 ObjectConstructor, along with a helper function that implements the 18410 ES5 internal [[ToPropertyDescriptor]] function. It then adds 18411 JSObject::defineOwnProperty that implements the appropriate ES5 semantics. 18412 Currently defineOwnProperty uses a delete followed by a put to redefine 18413 attributes of a property, clearly this is less efficient than it could be 18414 but we can improve this if it needs to be possible in future. 18415 18416 * JavaScriptCore.exp: 18417 * debugger/DebuggerActivation.cpp: 18418 (JSC::DebuggerActivation::defineGetter): 18419 (JSC::DebuggerActivation::defineSetter): 18420 * debugger/DebuggerActivation.h: 18421 * interpreter/Interpreter.cpp: 18422 (JSC::Interpreter::privateExecute): 18423 * jit/JITStubs.cpp: 18424 Update defineGetter/Setter calls 18425 * runtime/CommonIdentifiers.h: 18426 * runtime/JSArray.cpp: 18427 (JSC::JSArray::getOwnPropertySlot): 18428 * runtime/JSGlobalObject.cpp: 18429 (JSC::JSGlobalObject::defineGetter): 18430 (JSC::JSGlobalObject::defineSetter): 18431 * runtime/JSGlobalObject.h: 18432 * runtime/JSObject.cpp: 18433 (JSC::JSObject::defineGetter): 18434 (JSC::JSObject::defineSetter): 18435 (JSC::putDescriptor): 18436 (JSC::JSObject::defineOwnProperty): 18437 * runtime/JSObject.h: 18438 * runtime/ObjectConstructor.cpp: 18439 (JSC::ObjectConstructor::ObjectConstructor): 18440 (JSC::objectConstructorGetOwnPropertyDescriptor): 18441 (JSC::toPropertyDescriptor): 18442 (JSC::objectConstructorDefineProperty): 18443 * runtime/ObjectPrototype.cpp: 18444 (JSC::objectProtoFuncDefineGetter): 18445 (JSC::objectProtoFuncDefineSetter): 18446 * runtime/PropertyDescriptor.cpp: 18447 (JSC::PropertyDescriptor::writable): 18448 (JSC::PropertyDescriptor::enumerable): 18449 (JSC::PropertyDescriptor::configurable): 18450 (JSC::PropertyDescriptor::isDataDescriptor): 18451 (JSC::PropertyDescriptor::isGenericDescriptor): 18452 (JSC::PropertyDescriptor::isAccessorDescriptor): 18453 (JSC::PropertyDescriptor::getter): 18454 (JSC::PropertyDescriptor::setter): 18455 (JSC::PropertyDescriptor::setDescriptor): 18456 (JSC::PropertyDescriptor::setAccessorDescriptor): 18457 (JSC::PropertyDescriptor::setWritable): 18458 (JSC::PropertyDescriptor::setEnumerable): 18459 (JSC::PropertyDescriptor::setConfigurable): 18460 (JSC::PropertyDescriptor::setSetter): 18461 (JSC::PropertyDescriptor::setGetter): 18462 (JSC::PropertyDescriptor::equalTo): 18463 (JSC::PropertyDescriptor::attributesEqual): 18464 (JSC::PropertyDescriptor::attributesWithOverride): 18465 * runtime/PropertyDescriptor.h: 18466 (JSC::PropertyDescriptor::PropertyDescriptor): 18467 (JSC::PropertyDescriptor::value): 18468 (JSC::PropertyDescriptor::setValue): 18469 (JSC::PropertyDescriptor::isEmpty): 18470 (JSC::PropertyDescriptor::writablePresent): 18471 (JSC::PropertyDescriptor::enumerablePresent): 18472 (JSC::PropertyDescriptor::configurablePresent): 18473 (JSC::PropertyDescriptor::setterPresent): 18474 (JSC::PropertyDescriptor::getterPresent): 18475 (JSC::PropertyDescriptor::operator==): 18476 (JSC::PropertyDescriptor::): 18477 184782009-09-18 Gabor Loki <loki@inf.u-szeged.hu> 18479 18480 Reviewed by Gavin Barraclough. 18481 18482 Build fix to enable ARM_THUMB2 on Linux 18483 https://bugs.webkit.org/show_bug.cgi?id= 18484 18485 * jit/ExecutableAllocator.h: 18486 (JSC::ExecutableAllocator::cacheFlush): 18487 * jit/JITStubs.cpp: 18488 * wtf/Platform.h: 18489 184902009-09-18 Gabor Loki <loki@inf.u-szeged.hu> 18491 18492 Reviewed by Gavin Barraclough. 18493 18494 Defines two pseudo-platforms for ARM and Thumb-2 instruction set. 18495 https://bugs.webkit.org/show_bug.cgi?id=29122 18496 18497 Introduces WTF_PLATFORM_ARM_TRADITIONAL and WTF_PLATFORM_ARM_THUMB2 18498 macros on ARM platforms. The PLATFORM(ARM_THUMB2) should be used 18499 when Thumb-2 instruction set is the required target. The 18500 PLATFORM(ARM_TRADITIONAL) is for generic ARM instruction set. In 18501 case where the code is common the PLATFORM(ARM) have to be used. 18502 18503 * assembler/ARMAssembler.cpp: 18504 * assembler/ARMAssembler.h: 18505 * assembler/ARMv7Assembler.h: 18506 * assembler/MacroAssembler.h: 18507 * assembler/MacroAssemblerARM.cpp: 18508 * assembler/MacroAssemblerARM.h: 18509 * assembler/MacroAssemblerCodeRef.h: 18510 (JSC::MacroAssemblerCodePtr::MacroAssemblerCodePtr): 18511 * jit/ExecutableAllocator.h: 18512 * jit/JIT.h: 18513 * jit/JITInlineMethods.h: 18514 (JSC::JIT::beginUninterruptedSequence): 18515 (JSC::JIT::preserveReturnAddressAfterCall): 18516 (JSC::JIT::restoreReturnAddressBeforeReturn): 18517 (JSC::JIT::restoreArgumentReference): 18518 (JSC::JIT::restoreArgumentReferenceForTrampoline): 18519 * jit/JITOpcodes.cpp: 18520 * jit/JITStubs.cpp: 18521 (JSC::JITThunks::JITThunks): 18522 * jit/JITStubs.h: 18523 * wtf/Platform.h: 18524 * yarr/RegexJIT.cpp: 18525 (JSC::Yarr::RegexGenerator::generateEnter): 18526 185272009-09-18 Joerg Bornemann <joerg.bornemann@nokia.com> 18528 18529 Reviewed by Simon Hausmann. 18530 18531 Fix the Qt/Windows CE build. 18532 18533 * JavaScriptCore.pri: Build the ce_time.cpp functions from 18534 within Qt externally. 18535 * wtf/DateMath.cpp: Removed unnecessary Qt #ifdef, for the 18536 Qt build these functions are no external, too. 18537 185382009-09-17 Janne Koskinen <janne.p.koskinen@digia.com> 18539 18540 Reviewed by Simon Hausmann. 18541 18542 Symbian/WINSCW build fox. 18543 18544 Repeat Q_OS_WIN wchar_t hack for WINSCW, similar to 18545 revision 24774. 18546 18547 WINSCW defines wchar_t, thus UChar has to be wchar_t 18548 18549 * wtf/unicode/qt4/UnicodeQt4.h: 18550 185512009-09-17 Janne Koskinen <janne.p.koskinen@digia.com> 18552 18553 Reviewed by Simon Hausmann. 18554 18555 Symbian/WINSCW build fix. 18556 18557 https://bugs.webkit.org/show_bug.cgi?id=29186 18558 18559 WINSCW Template specialisation name in declaration must the be the same as in implementation. 18560 18561 * runtime/LiteralParser.h: 18562 185632009-09-15 Norbert Leser <norbert.leser@nokia.com> 18564 18565 Reviewed by Darin Adler. 18566 18567 https://bugs.webkit.org/show_bug.cgi?id=27060 18568 18569 Symbian compiler for emulator target (WINSCW) fails with 18570 "illegal operand" for m_attributesInPrevious in structure.ccp 18571 (when calling make_pair functions). 18572 This error is apparently due to the compiler not properly 18573 resolving the unsigned type of the declared bitfield. 18574 18575 Initial patch explicitly casted m_attributesInPrevious 18576 to unsigned, but since bitfield optimization is not critical for 18577 the emulator target, this conditional change in header file 18578 appears to be least intrusive. 18579 18580 * runtime/Structure.h: 18581 185822009-09-16 Gabor Loki <loki@inf.u-szeged.hu> 18583 18584 Reviewed by Darin Adler. 18585 18586 Fix GCC warnings on ARM_THUMB2 platform 18587 18588 * assembler/ARMv7Assembler.h: 18589 (JSC::ARMThumbImmediate::countLeadingZerosPartial): 18590 * assembler/MacroAssemblerARMv7.h: 18591 (JSC::MacroAssemblerARMv7::branchTruncateDoubleToInt32): 18592 (JSC::MacroAssemblerARMv7::moveFixedWidthEncoding): 18593 185942009-09-16 Greg Bolsinga <bolsinga@apple.com> 18595 18596 Add ENABLE(INSPECTOR) 18597 https://bugs.webkit.org/show_bug.cgi?id=29260 18598 18599 Reviewed by David Kilzer. 18600 18601 * wtf/Platform.h: 18602 186032009-09-16 Greg Bolsinga <bolsinga@apple.com> 18604 18605 Add ENABLE(CONTEXT_MENUS) 18606 https://bugs.webkit.org/show_bug.cgi?id=29225 18607 18608 Reviewed by David Kilzer. 18609 18610 * wtf/Platform.h: 18611 186122009-09-16 Benjamin C Meyer <benjamin.meyer@torchmobile.com> 18613 18614 Reviewed by Eric Seidel. 18615 18616 The webkit stdint and stdbool headers exists because 18617 the compiler MSVC doesn't include them. The check 18618 should not check for PLATFORM(WIN_OS) but for MSVC. 18619 18620 * os-win32/stdbool.h: 18621 * os-win32/stdint.h: 18622 186232009-09-16 Greg Bolsinga <bolsinga@apple.com> 18624 18625 Add ENABLE(DRAG_SUPPORT) 18626 https://bugs.webkit.org/show_bug.cgi?id=29233 18627 18628 Reviewed by David Kilzer. 18629 18630 * wtf/Platform.h: 18631 186322009-09-16 Kevin Ollivier <kevino@theolliviers.com> 18633 18634 waf build fix after flag was moved to correct place. 18635 18636 * wscript: 18637 186382009-09-16 Tor Arne Vestbø <tor.arne.vestbo@nokia.com> 18639 18640 Reviewed by Simon Hausmann. 18641 18642 [Qt] Build fix for 64-bit Qt on Mac OS X 18643 18644 * wtf/Platform.h: Use JSVALUE64 on DARWIN, not only on MAC 18645 186462009-09-16 Zoltan Herczeg <zherczeg@inf.u-szeged.hu> 18647 18648 Reviewed by Simon Hausmann. 18649 18650 [Qt] Fix wtf/ThreadSpecific.h under Qt to free thread local objects. 18651 https://bugs.webkit.org/show_bug.cgi?id=29295 18652 18653 This is an important fix when JavaScript workers are in use, since 18654 unfreed ThreadGlobalDatas leak a big amount of memory (50-100k each). 18655 QThreadStorage calls the destructor of a given object, which is the 18656 ThreadSpecific::Data. Unlike pthread, Qt is object oriented, and does 18657 not support the calling of a static utility function when the thread 18658 is about to close. In this patch we call the ThreadSpecific::destroy() 18659 utility function from the destructor of ThreadSpecific::Data. Moreover, 18660 since Qt resets all thread local values to 0 before the calling of the 18661 appropriate destructors, we set back the pointer to its original value. 18662 This is necessary because the get() method of the ThreadSpecific 18663 object may be called during the exuction of the destructor. 18664 18665 * wtf/ThreadSpecific.h: 18666 (WTF::ThreadSpecific::Data::~Data): 18667 (WTF::::~ThreadSpecific): 18668 (WTF::::set): 18669 (WTF::::destroy): 18670 186712009-09-10 Oliver Hunt <oliver@apple.com> 18672 18673 Reviewed by Geoff Garen. 18674 18675 Allow anonymous storage inside JSObject 18676 https://bugs.webkit.org/show_bug.cgi?id=29168 18677 18678 Add the concept of anonymous slots to Structures so that it is 18679 possible to store references to values that need marking in the 18680 standard JSObject storage buffer. This allows us to reduce the 18681 malloc overhead of some objects (by allowing them to store JS 18682 values in the inline storage of the object) and reduce the 18683 dependence of custom mark functions (if all an objects children 18684 are in the standard object property storage there's no need to 18685 mark them manually). 18686 18687 * JavaScriptCore.exp: 18688 * runtime/JSObject.h: 18689 (JSC::JSObject::putAnonymousValue): 18690 (JSC::JSObject::getAnonymousValue): 18691 (JSC::JSObject::addAnonymousSlots): 18692 * runtime/JSWrapperObject.h: 18693 (JSC::JSWrapperObject::createStructure): 18694 (JSC::JSWrapperObject::JSWrapperObject): 18695 (JSC::JSWrapperObject::setInternalValue): 18696 * runtime/PropertyMapHashTable.h: 18697 * runtime/Structure.cpp: 18698 (JSC::Structure::~Structure): 18699 (JSC::Structure::materializePropertyMap): 18700 (JSC::Structure::addAnonymousSlotsTransition): 18701 (JSC::Structure::copyPropertyTable): 18702 (JSC::Structure::put): 18703 (JSC::Structure::rehashPropertyMapHashTable): 18704 * runtime/Structure.h: 18705 (JSC::Structure::propertyStorageSize): 18706 (JSC::StructureTransitionTable::reifySingleTransition): 18707 * runtime/StructureTransitionTable.h: 18708 (JSC::StructureTransitionTable::TransitionTable::addSlotTransition): 18709 (JSC::StructureTransitionTable::TransitionTable::removeSlotTransition): 18710 (JSC::StructureTransitionTable::TransitionTable::getSlotTransition): 18711 (JSC::StructureTransitionTable::getAnonymousSlotTransition): 18712 (JSC::StructureTransitionTable::addAnonymousSlotTransition): 18713 (JSC::StructureTransitionTable::removeAnonymousSlotTransition): 18714 187152009-09-15 Alex Milowski <alex@milowski.com> 18716 18717 Reviewed by Tor Arne Vestbø. 18718 18719 Added the ENABLE_MATHML define to the features 18720 18721 * Configurations/FeatureDefines.xcconfig: 18722 187232009-09-15 Csaba Osztrogonac <oszi@inf.u-szeged.hu> 18724 18725 Reviewed by Tor Arne Vestbø. 18726 18727 [Qt] Build fix for windows. 18728 18729 After http://trac.webkit.org/changeset/47795 the MinGW build broke, 18730 because MinGW has __mingw_aligned_malloc instead of _aligned_malloc. 18731 18732 * runtime/Collector.cpp: 18733 (JSC::Heap::allocateBlock): MinGW case added. 18734 (JSC::Heap::freeBlock): MinGW case added. 18735 187362009-09-15 Csaba Osztrogonac <oszi@inf.u-szeged.hu> 18737 18738 Reviewed by Tor Arne Vestbø. 18739 18740 [Qt] Build fix for Windows/MinGW 18741 18742 https://bugs.webkit.org/show_bug.cgi?id=29268 18743 18744 * wtf/Platform.h: JSVALUE32_64 temporarily disabled on PLATFORM(WIN_OS) with COMPILER(MINGW) 18745 187462009-09-14 Gabor Loki <loki@inf.u-szeged.hu> 18747 18748 Reviewed by Gavin Barraclough. 18749 18750 Detect VFP at runtime in generic ARM port on Linux platform. 18751 https://bugs.webkit.org/show_bug.cgi?id=29076 18752 18753 * JavaScriptCore.pri: 18754 * assembler/MacroAssemblerARM.cpp: Added. 18755 (JSC::isVFPPresent): 18756 * assembler/MacroAssemblerARM.h: 18757 (JSC::MacroAssemblerARM::supportsFloatingPoint): 18758 187592009-09-14 Csaba Osztrogonac <oszi@inf.u-szeged.hu> 18760 18761 Reviewed by Tor Arne Vestbø. 18762 18763 [Qt] Build fix for windows build. 18764 18765 * JavaScriptCore.pri: Correct a logic error. 18766 * pcre/dftables: Add missing paranthesis for tmpdir function. 18767 187682009-09-12 Oliver Hunt <oliver@apple.com> 18769 18770 Reviewed by NOBODY (Build fix). 18771 18772 Build fix for windows exports (again). 18773 18774 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 18775 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 18776 187772009-09-12 Oliver Hunt <oliver@apple.com> 18778 18779 Reviewed by NOBODY (Build fix). 18780 18781 Build fix for windows exports. 18782 18783 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 18784 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 18785 187862009-09-12 Oliver Hunt <oliver@apple.com> 18787 18788 Reviewed by NOBODY (Build fix). 18789 18790 Correct fix for non-allinonefile builds 18791 18792 * runtime/ObjectConstructor.cpp: 18793 187942009-09-12 Oliver Hunt <oliver@apple.com> 18795 18796 Reviewed by NOBODY (Build fix). 18797 18798 Fix non-allinonefile builds 18799 18800 * runtime/ObjectConstructor.cpp: 18801 188022009-09-12 Oliver Hunt <oliver@apple.com> 18803 18804 Reviewed by Maciej Stachowiak. 18805 18806 [ES5] Implement Object.keys 18807 https://bugs.webkit.org/show_bug.cgi?id=29170 18808 18809 This patch basically requires two separate steps, the first is to split getPropertyNames 18810 into two functions -- getOwnPropertyNames and getPropertyNames, basically making them behave 18811 in the same way as getOwnPropertySlot and getPropertySlot. In essence getOwnPropertyNames 18812 produces the list of properties on an object excluding its prototype chain and getPropertyNames 18813 just iterates the the object and its prototype chain calling getOwnPropertyNames at each level. 18814 18815 * API/JSCallbackObject.h: 18816 * API/JSCallbackObjectFunctions.h: 18817 (JSC::::getOwnPropertyNames): 18818 * JavaScriptCore.exp: 18819 * debugger/DebuggerActivation.cpp: 18820 (JSC::DebuggerActivation::getOwnPropertyNames): 18821 * debugger/DebuggerActivation.h: 18822 * runtime/CommonIdentifiers.h: 18823 * runtime/JSArray.cpp: 18824 (JSC::JSArray::getOwnPropertyNames): 18825 * runtime/JSArray.h: 18826 * runtime/JSByteArray.cpp: 18827 (JSC::JSByteArray::getOwnPropertyNames): 18828 * runtime/JSByteArray.h: 18829 * runtime/JSNotAnObject.cpp: 18830 (JSC::JSNotAnObject::getOwnPropertyNames): 18831 * runtime/JSNotAnObject.h: 18832 * runtime/JSObject.cpp: 18833 (JSC::JSObject::getOwnPropertyNames): 18834 * runtime/JSObject.h: 18835 * runtime/JSVariableObject.cpp: 18836 (JSC::JSVariableObject::getOwnPropertyNames): 18837 * runtime/JSVariableObject.h: 18838 * runtime/ObjectConstructor.cpp: 18839 (JSC::ObjectConstructor::ObjectConstructor): 18840 (JSC::objectConstructorKeys): 18841 * runtime/RegExpMatchesArray.h: 18842 (JSC::RegExpMatchesArray::getOwnPropertyNames): 18843 * runtime/StringObject.cpp: 18844 (JSC::StringObject::getOwnPropertyNames): 18845 * runtime/StringObject.h: 18846 * runtime/Structure.cpp: 18847 (JSC::Structure::getOwnEnumerablePropertyNames): 18848 (JSC::Structure::getEnumerablePropertyNames): 18849 * runtime/Structure.h: 18850 188512009-09-11 Oliver Hunt <oliver@apple.com> 18852 18853 Reviewed by Sam Weinig. 18854 18855 getPropertyNames caching is invalid when the prototype chain contains objects with custom getPropertyNames 18856 https://bugs.webkit.org/show_bug.cgi?id=29214 18857 18858 Add a flag to TypeInfo to indicate whether a type overrides getPropertyNames. 18859 This flag is used to make sure that caching of the property name data is safe. 18860 18861 * API/JSCallbackConstructor.h: 18862 (JSC::JSCallbackConstructor::createStructure): 18863 * debugger/DebuggerActivation.h: 18864 (JSC::DebuggerActivation::createStructure): 18865 * runtime/BooleanObject.h: 18866 (JSC::BooleanObject::createStructure): 18867 * runtime/DatePrototype.h: 18868 (JSC::DatePrototype::createStructure): 18869 * runtime/FunctionPrototype.h: 18870 (JSC::FunctionPrototype::createStructure): 18871 * runtime/JSONObject.h: 18872 (JSC::JSONObject::createStructure): 18873 * runtime/JSObject.h: 18874 (JSC::JSObject::createStructure): 18875 * runtime/JSTypeInfo.h: 18876 (JSC::TypeInfo::hasDefaultGetPropertyNames): 18877 * runtime/JSVariableObject.h: 18878 (JSC::JSVariableObject::createStructure): 18879 * runtime/JSWrapperObject.h: 18880 (JSC::JSWrapperObject::createStructure): 18881 * runtime/MathObject.h: 18882 (JSC::MathObject::createStructure): 18883 * runtime/NumberConstructor.h: 18884 (JSC::NumberConstructor::createStructure): 18885 * runtime/NumberObject.h: 18886 (JSC::NumberObject::createStructure): 18887 * runtime/RegExpConstructor.h: 18888 (JSC::RegExpConstructor::createStructure): 18889 * runtime/RegExpObject.h: 18890 (JSC::RegExpObject::createStructure): 18891 * runtime/StructureChain.cpp: 18892 (JSC::StructureChain::isCacheable): 18893 188942009-09-11 Alexey Proskuryakov <ap@webkit.org> 18895 18896 Reviewed by Geoff Garen. 18897 18898 https://bugs.webkit.org/show_bug.cgi?id=29207 18899 Add checks for using WebCore JS context on secondary threads 18900 18901 * runtime/JSGlobalData.cpp: (JSC::JSGlobalData::JSGlobalData): 18902 * runtime/JSGlobalData.h: 18903 Added a new mainThreadOnly flag that WebCore would set. 18904 18905 * runtime/Collector.cpp: (JSC::Heap::registerThread): JSC API methods always call this, 18906 so this is a good place to check that the API isn't used form a wrong thread. 18907 189082009-09-11 Jocelyn Turcotte <jocelyn.turcotte@nokia.com> 18909 18910 Reviewed by Simon Hausmann. 18911 18912 Compiling JavaScriptCore on sparc 64 with gcc fails. 18913 18914 ThreadSafeShared uses the atomic __gnu_cxx::__exchange_and_add with an int, 18915 however on sparc 64 the _Atomic_word argument is typedefed to long (8 bytes). 18916 18917 The patch disables WTF_USE_LOCKFREE_THREADSAFESHARED in ThreadSafeShared to use 18918 a mutex instead when compiling for sparc 64 with gcc. 18919 18920 https://bugs.webkit.org/show_bug.cgi?id=29175 18921 18922 * wtf/Platform.h: 18923 __sparc64__ is not defined on all OS. 18924 Uses instead: __sparc__ && __arch64__ || __sparcv9 18925 * wtf/Threading.h: 18926 189272009-09-11 Prasanth Ullattil <prasanth.ullattil@nokia.com> 18928 18929 Reviewed by Simon Hausmann. 18930 18931 Fix compile error on Windows7(64Bit) with latest SDK. 18932 18933 Added the missing include file. 18934 18935 * runtime/UString.cpp: 18936 189372009-09-11 Joerg Bornemann <joerg.bornemann@trolltech.com> 18938 18939 Reviewed by Simon Hausmann. 18940 18941 Qt/Windows CE compile fix, include the executable allocator and 18942 markstack implementation in the windows build. 18943 18944 * JavaScriptCore.pri: 18945 189462009-09-08 John Abd-El-Malek <jam@chromium.org> 18947 18948 Reviewed by Dimitri Glazkov. 18949 18950 Remove unneeded define for ActiveX. 18951 https://bugs.webkit.org/show_bug.cgi?id=29054 18952 18953 * wtf/Platform.h: 18954 189552009-09-10 Mark Rowe <mrowe@apple.com> 18956 18957 Rubber-stamped by Sam Weinig. 18958 18959 Update JavaScriptCore and WebKit's FeatureDefines.xcconfig so that they are in sync with WebCore as they need to be. 18960 18961 * Configurations/FeatureDefines.xcconfig: 18962 189632009-09-10 Fumitoshi Ukai <ukai@chromium.org> 18964 18965 Reviewed by Alexey Proskuryakov. 18966 18967 Export WTF::tryFastMalloc used in WebSocketChannel. 18968 https://bugs.webkit.org/show_bug.cgi?id=28038 18969 18970 * JavaScriptCore.exp: 18971 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 18972 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 18973 189742009-09-10 Oliver Hunt <oliver@apple.com> 18975 18976 Reviewed by NOBODY (Build fix). 18977 18978 Make StructureTransitionTable use an enum for the PtrAndFlags member 18979 used for the single transition slot optimisation. 18980 18981 * runtime/StructureTransitionTable.h: 18982 (JSC::StructureTransitionTable::StructureTransitionTable): 18983 (JSC::StructureTransitionTable::usingSingleTransitionSlot): 18984 (JSC::StructureTransitionTable::): 18985 189862009-09-10 Oliver Hunt <oliver@apple.com> 18987 18988 Reviewed by Geoff Garen. 18989 18990 Refactor StructureTransitionTable and Structure to unify handling of the single slot optimization 18991 https://bugs.webkit.org/show_bug.cgi?id=29141 18992 18993 Make StructureTransitionTable encapsulate the single transition slot optimization. 18994 18995 * runtime/Structure.cpp: 18996 (JSC::Structure::Structure): 18997 (JSC::Structure::~Structure): 18998 (JSC::Structure::addPropertyTransitionToExistingStructure): 18999 (JSC::Structure::addPropertyTransition): 19000 (JSC::Structure::addPropertyWithoutTransition): 19001 (JSC::Structure::removePropertyWithoutTransition): 19002 (JSC::Structure::hasTransition): 19003 * runtime/Structure.h: 19004 (JSC::StructureTransitionTable::contains): 19005 (JSC::StructureTransitionTable::get): 19006 (JSC::StructureTransitionTable::hasTransition): 19007 (JSC::StructureTransitionTable::reifySingleTransition): 19008 * runtime/StructureTransitionTable.h: 19009 (JSC::StructureTransitionTable::StructureTransitionTable): 19010 (JSC::StructureTransitionTable::~StructureTransitionTable): 19011 (JSC::StructureTransitionTable::remove): 19012 (JSC::StructureTransitionTable::add): 19013 (JSC::StructureTransitionTable::table): 19014 (JSC::StructureTransitionTable::singleTransition): 19015 (JSC::StructureTransitionTable::usingSingleTransitionSlot): 19016 (JSC::StructureTransitionTable::setSingleTransition): 19017 (JSC::StructureTransitionTable::setTransitionTable): 19018 (JSC::StructureTransitionTable::): 19019 * wtf/PtrAndFlags.h: 19020 (WTF::PtrAndFlags::PtrAndFlags): 19021 190222009-09-10 Zoltan Horvath <zoltan@webkit.org> 19023 19024 Reviewed by Darin Adler. 19025 19026 Implement fastDeleteSkippingDestructor for FastAllocBase and fastDeleteAllValues for HashSet 19027 https://bugs.webkit.org/show_bug.cgi?id=25930 19028 19029 FastAllocBase has been extended with fastDeleteSkippingDestructor function which 19030 releases memory without destructor call. fastDeleteAllValues has been implemented 19031 similar as deleteAllValues but it uses fastDelete function to release memory. 19032 19033 * wtf/FastAllocBase.h: 19034 (WTF::fastDeleteSkippingDestructor): 19035 * wtf/HashSet.h: 19036 (WTF::fastDeleteAllValues): 19037 190382009-09-10 Laszlo Gombos <laszlo.1.gombos@nokia.com> 19039 19040 Reviewed by Darin Adler. 19041 19042 ARM compiler does not understand GCC visibility attribute 19043 https://bugs.webkit.org/show_bug.cgi?id=29079 19044 19045 * API/JSBase.h: Make the test more specific to hit only 19046 the GCC compiler 19047 190482009-09-10 Adam Barth <abarth@webkit.org> 19049 19050 Unreviewed revert of the previous change. It broke the tests. 19051 19052 * wtf/dtoa.cpp: 19053 (WTF::dtoa): 19054 190552009-09-10 Ben Laurie <benl@google.com> 19056 19057 Reviewed by Adam Barth. 19058 19059 <https://bugs.webkit.org/show_bug.cgi?id=26836> 19060 19061 If dtoa was given a small buffer and the number was either infinite or 19062 NaN, then the buffer would be overflowed. 19063 19064 * wtf/dtoa.cpp: 19065 190662009-09-09 Laszlo Gombos <laszlo.1.gombos@nokia.com> 19067 19068 Reviewed by Darin Adler. 19069 19070 Change reinterpret_cast to static_cast in r48212. 19071 19072 * jit/ExecutableAllocator.h: 19073 (JSC::ExecutableAllocator::cacheFlush): 19074 190752009-09-09 Laszlo Gombos <laszlo.1.gombos@nokia.com> 19076 19077 Reviewed by Darin Adler. 19078 19079 Remove WTF_PLATFORM_FORCE_PACK as it is no longer used 19080 https://bugs.webkit.org/show_bug.cgi?id=29066 19081 19082 * wtf/Platform.h: 19083 190842009-09-09 Laszlo Gombos <laszlo.1.gombos@nokia.com> 19085 19086 Reviewed by Ariya Hidayat. 19087 19088 Implement flushing the instruction cache for Symbian 19089 https://bugs.webkit.org/show_bug.cgi?id=29075 19090 19091 * jit/ExecutableAllocator.h: 19092 (JSC::ExecutableAllocator::cacheFlush): Call IMB_Range to flush 19093 the instruction cache on Symbian 19094 190952009-09-09 Kent Hansen <khansen@trolltech.com> 19096 19097 Reviewed by Darin Adler. 19098 19099 https://bugs.webkit.org/show_bug.cgi?id=29024 19100 Make JavaScriptCore compile on platforms with case-insensitive file systems and typeinfo.h in STL 19101 19102 These platforms include Microsoft Visual Studio 2003, and Symbian with Metrowerks compiler. 19103 19104 * JavaScriptCore.gypi: 19105 * JavaScriptCore.xcodeproj/project.pbxproj: 19106 * runtime/JSTypeInfo.h: Copied from JavaScriptCore/runtime/TypeInfo.h. 19107 * runtime/Structure.h: 19108 * runtime/TypeInfo.h: Removed. 19109 191102009-09-08 Oliver Hunt <oliver@apple.com> 19111 19112 Reviewed by Maciej Stachowiak. 19113 19114 JSON.stringify(Date) loses the milliseconds information 19115 https://bugs.webkit.org/show_bug.cgi?id=29063 19116 19117 Make sure we include milliseconds in the output of toISOString. 19118 19119 * runtime/DatePrototype.cpp: 19120 (JSC::dateProtoFuncToISOString): 19121 191222009-09-08 Kevin Ollivier <kevino@theolliviers.com> 19123 19124 wx build fix, generate derived sources earlier in order to make sure 19125 they're found by the build system when generating the list of sources to build. 19126 19127 * wscript: 19128 191292009-09-08 Laszlo Gombos <laszlo.1.gombos@nokia.com> 19130 19131 Reviewed by Simon Hausmann. 19132 19133 Build fix when USE(LOCKFREE_THREADSAFESHARED) is not defined 19134 https://bugs.webkit.org/show_bug.cgi?id=29011 19135 19136 * wtf/Threading.h: Use LOCKFREE_THREADSAFESHARED guard for 19137 atomicIncrement and atomicDecrement 19138 191392009-09-07 Zoltan Horvath <zoltan@webkit.org> 19140 19141 Reviewed by Darin Adler. 19142 19143 Allow custom memory allocation control in Yarr's RegexInterpreter 19144 https://bugs.webkit.org/show_bug.cgi?id=29025 19145 19146 Inherits RegexInterpreter classes from FastAllocBase (bug #20422), which has 19147 been instantiated by 'new': 19148 19149 class ByteDisjunction 19150 -> instantiated in JavaScriptCore/yarr/RegexInterpreter.cpp:1462 19151 19152 struct BytecodePattern 19153 -> instantiated in JavaScriptCore/yarr/RegexInterpreter.cpp:1279 19154 19155 * yarr/RegexInterpreter.h: 19156 191572009-09-07 Drew Wilson <atwilson@google.com> 19158 19159 Reverting r48121 to fix Windows build errors. 19160 19161 * JavaScriptCore.exp: 19162 191632009-09-07 Drew Wilson <atwilson@google.com> 19164 19165 Reviewed by David Levin. 19166 19167 Enable SHARED_WORKERS by default 19168 https://bugs.webkit.org/show_bug.cgi?id=28959 19169 19170 * Configurations/FeatureDefines.xcconfig: 19171 191722009-09-07 Fumitoshi Ukai <ukai@chromium.org> 19173 19174 Reviewed by Alexey Proskuryakov. 19175 19176 Export WTF::tryFastMalloc used in WebSocketChannel. 19177 https://bugs.webkit.org/show_bug.cgi?id=28038 19178 19179 * JavaScriptCore.exp: 19180 191812009-09-04 Oliver Hunt <oliver@apple.com> 19182 19183 Reviewed by NOBODY (Build fix). 19184 19185 Fix windows export files 19186 19187 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 19188 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 19189 191902009-09-04 Oliver Hunt <oliver@apple.com> 19191 19192 Reviewed by Gavin Barraclough. 19193 19194 [[ToString]] conversion should use the actual toString function for String objects. 19195 19196 Remove incorrect specialisations of toString conversions on StringObject. 19197 19198 * JavaScriptCore.exp: 19199 * runtime/StringObject.cpp: 19200 * runtime/StringObject.h: 19201 192022009-09-04 Steve Falkenburg <sfalken@apple.com> 19203 19204 Windows build fix. 19205 19206 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: Add new export. 19207 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: Add new export. 19208 192092009-09-04 Steve Falkenburg <sfalken@apple.com> 19210 19211 Windows build fix. 19212 19213 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: Remove unneeded export. 19214 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: Remove unneeded export. 19215 192162009-09-04 Darin Adler <darin@apple.com> 19217 19218 Reviewed by Geoff Garen. 19219 19220 DateInstance object collected on ARM JIT (JSValue: WTF_USE_JSVALUE32) 19221 https://bugs.webkit.org/show_bug.cgi?id=28909 19222 19223 Part two. 19224 19225 Make some improvements to garbage collection code: 19226 19227 1) Create a runtime assertion that catches any classes that 19228 override markChildren but have the HasDefaultMark bit set. 19229 2) Remove checks of the mark bit outside the MarkStack::append 19230 function; they are redundant. 19231 3) Improve the efficiency of the asObject and asArray functions 19232 when called on JSCell* to avoid a round trip to JSValue. 19233 4) Make more callers use the checked asCell and asObject 19234 casting functions rather than unchecked casts. 19235 5) Removed the JSCell::marked function and other GC-related 19236 functions because these operations are no longer things that 19237 code other than the core GC code needs to do directly. Fixed 19238 callers that were calling them. 19239 19240 * runtime/Collector.cpp: 19241 (JSC::Heap::markConservatively): Removed unneeded call to MarkStack::drain. 19242 (JSC::Heap::markProtectedObjects): Removed unneeded check of the mark 19243 bit and call to MarkStack::drain. 19244 (JSC::Heap::collect): Removed unneeded checks of the mark bit and also 19245 changed call to SmallStrings::mark to call markChildren instead to match 19246 the rest of the objects. 19247 (JSC::typeName): Removed unneeded cast to JSObject*. 19248 19249 * runtime/JSArray.h: 19250 (JSC::asArray): Added an overload for JSCell* and changed the JSValue 19251 version to call it. Removed some unneeded casts. 19252 (JSC::JSArray::markChildrenDirect): Marked this function inline. It's in 19253 a header, and if not marked inline this could lead to linking problems. 19254 (JSC::MarkStack::markChildren): Added. This helper function is used by 19255 the drain function to avoid repating code. Also added the code here to 19256 check fro default mark violations in debug code. If a markChildren 19257 function adds something to the mark stack, but the type info claimed 19258 hasDefaultMark was true, then we will get an assertion now. Also fixed 19259 the assertion about the mark bit to use the Heap function directly 19260 because we don't have a JSCell::marked function any more. 19261 (JSC::MarkStack::drain): Changed a local variable from "v" to "value", 19262 and from "currentCell" to "cell". Changed to call markChildren in two 19263 places instead of repeating a chain of if statements twice. Changed 19264 code that reads and writes the mark bit to use Heap::isCellMarked and 19265 Heap::markCell so we can eliminate the JSCell::marked and 19266 JSCell::markCellDirect functions. 19267 19268 * runtime/JSCell.h: Removed JSCell's markCellDirect and marked member 19269 functions. Added a comment explaining that asCell should be deprecated 19270 in favor of the JSValue asCell member function. 19271 (JSC::MarkStack::append): Added the assertion that catches callers 19272 that have set the HasDefaultMark bit incorrectly. Changed 19273 code that reads and writes the mark bit to use Heap::isCellMarked and 19274 Heap::markCell so we can eliminate the JSCell::marked and 19275 JSCell::markCellDirect functions. Moved the overload of 19276 MarkStack::append for JSValue here so it can call through to the cell 19277 version. The old version had a copy of all the code instead, but that 19278 repeated the conversion from JSValue to JSCell* and the check for 19279 whether a value is a cell multiple times. 19280 (JSC::Structure::markAggregate): Moved this function here to avoid 19281 dependencies for Structure.h, since this calls MarkStack::append. 19282 19283 * runtime/JSObject.cpp: 19284 (JSC::JSObject::markChildren): Added code to clear 19285 m_isCheckingForDefaultMarkViolation so the marking done by JSObject 19286 doesn't trigger the assertion. 19287 19288 * runtime/JSValue.h: Moved some stray includes that were outside the 19289 header guard inside it. Not sure how that happened! Removed the 19290 GC-related member functions markChildren, hasChildren, marked, and 19291 markDirect. 19292 19293 * runtime/JSWrapperObject.h: Made markChildren private. 19294 (JSC::JSWrapperObject::createStructure): Added. Fixes a bug where the 19295 HasDefaultMark bit was set. 19296 19297 * runtime/MarkStack.h: Added m_isCheckingForDefaultMarkViolation and 19298 initialized it to false. Moved the append function body from here to 19299 JSCell.h. Added a declaration of a private markChildren function used 19300 inside the drain function. 19301 19302 * runtime/SmallStrings.cpp: 19303 (JSC::SmallStrings::markChildren): Changed the name and style of this 19304 function to match other functions. This allows us to share the normal 19305 mark stack code path. 19306 19307 * runtime/SmallStrings.h: Changed the name and interface of mark to 19308 the more-normal markChildren style. 19309 19310 * runtime/Structure.h: Moved the body of markAggregate into the 19311 JSCell.h to avoid a circular dependency with JSCell.h. 19312 193132009-09-04 Darin Adler <darin@apple.com> 19314 19315 Reviewed by Geoff Garen. 19316 19317 DateInstance object collected on ARM JIT (JSValue: WTF_USE_JSVALUE32) 19318 https://bugs.webkit.org/show_bug.cgi?id=28909 19319 19320 Part one. 19321 19322 Make some improvements to garbage collection code: 19323 19324 1) Fix the two classes that had the default mark bit set but 19325 should not. 19326 2) Remove checks of the mark bit outside the MarkStack::append 19327 function; they are redundant. 19328 3) Make more callers use the checked asCell and asObject 19329 casting functions rather than unchecked casts. 19330 4) Removed some GC-related functions because these operations are 19331 no longer things that code other than the core GC code needs 19332 to do directly. Fixed callers that were calling them. 19333 19334 * bytecode/CodeBlock.cpp: 19335 (JSC::CodeBlock::markAggregate): Removed unneeded check of the mark 19336 bit before calling MarkStack::append. 19337 19338 * interpreter/Register.h: Removed unneeded marked and markChildren 19339 functions. 19340 19341 * jit/JITStubs.cpp: 19342 (op_eq): Removed unneeded assertions, instead using checked casting 19343 functions such as asObject. 19344 19345 * runtime/ArgList.h: Added now-needed forward declaration of MarkStack. 19346 19347 * runtime/GetterSetter.cpp: 19348 (JSC::GetterSetter::markChildren): Remmoved unneeded check of the mark bit. 19349 19350 * runtime/GlobalEvalFunction.h: 19351 (JSC::GlobalEvalFunction::createStructure): Added. Fixes a bug where the 19352 HasDefaultMark bit was set. 19353 19354 * runtime/JSCell.cpp: 19355 (JSC::JSCell::getObject): Use asObject to avoid a direct static_cast. 19356 19357 * runtime/JSObject.h: 19358 (JSC::asObject): Added an overload for JSCell* and changed the JSValue 19359 version to call it. 19360 (JSC::JSValue::get): Use asObject to avoid a direct static_cast. 19361 19362 * runtime/JSWrapperObject.h: Made markChildren private. 19363 (JSC::JSWrapperObject::createStructure): Added. Fixes a bug where the 19364 HasDefaultMark bit was set. Later we may want to optimize this for 19365 wrapper types that never have cells in their internal values, but there 19366 is no measured performance regression in SunSpider or V8 doing this 19367 all the time. 19368 19369 * runtime/MarkStack.cpp: Tweaked formatting. 19370 193712009-09-04 Kevin Ollivier <kevino@theolliviers.com> 19372 19373 wx build fix. Switch USE_ defines over to the compiler so that they can be 19374 checked by files not including config.h (like WebCorePrefix.h). 19375 19376 * wtf/Platform.h: 19377 193782009-09-03 Yong Li <yong.li@torchmobile.com> 19379 19380 Reviewed by David Levin. 19381 19382 Remove unnecessary dependency on unistd.h 19383 https://bugs.webkit.org/show_bug.cgi?id=28962 19384 19385 * runtime/Completion.cpp: 19386 193872009-09-03 Fumitoshi Ukai <ukai@chromium.org> 19388 19389 Reviewed by Eric Seidel. 19390 19391 Add strnstr for Linux and Windows in StringExtras.h 19392 https://bugs.webkit.org/show_bug.cgi?id=28901 19393 19394 * wtf/StringExtras.h: 19395 (strnstr): 19396 193972009-09-03 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 19398 19399 Reviewed by Darin Adler. 19400 19401 Allow custom memory allocation control for JavaScriptCore's HashEntry class 19402 https://bugs.webkit.org/show_bug.cgi?id=27830 19403 19404 Inherits HashEntry class from FastAllocBase because it has been 19405 instantiated by 'new' JavaScriptCore/runtime/Lookup.cpp:32. 19406 19407 * runtime/Lookup.h: 19408 194092009-09-02 Gavin Barraclough <barraclough@apple.com> 19410 19411 Should crash if JIT code buffer allocation fails. 19412 19413 https://bugs.webkit.org/show_bug.cgi?id=28926 19414 <rdar://problem/7031922> 19415 19416 * jit/ExecutableAllocatorPosix.cpp: 19417 (JSC::ExecutablePool::systemAlloc): 19418 * jit/ExecutableAllocatorWin.cpp: 19419 (JSC::ExecutablePool::systemAlloc): 19420 194212009-09-02 Kevin Ollivier <kevino@theolliviers.com> 19422 19423 waf build fixes for Windows/MSVC. 19424 19425 * wscript: 19426 194272009-09-02 Kevin Ollivier <kevino@theolliviers.com> 19428 19429 Build fix for building on Windows. 19430 19431 * wtf/ThreadingPthreads.cpp: 19432 194332009-09-02 Norbert Leser <norbert.leser@nokia.com> 19434 19435 Reviewed by Eric Seidel. 19436 19437 Use fastMalloc when neither MMAP nor VIRTUALALLOC are enabled 19438 19439 RegisterFile constructor currently throws #error when both 19440 MMAP and VIRTUALALLOC conditions fail. 19441 On any platform that does not provide these features 19442 (for instance, Symbian), 19443 the fallback should be regular malloc (or fastMalloc). 19444 It is functionally equivalent in this case, even though it may 19445 have certain drawbacks such as lack of dynamic pre-allocation. 19446 19447 * interpreter/RegisterFile.cpp: 19448 (JSC::RegisterFile::~RegisterFile): 19449 * interpreter/RegisterFile.h: 19450 (JSC::RegisterFile::RegisterFile): 19451 194522009-08-31 Robert Agoston <Agoston.Robert@stud.u-szeged.hu> 19453 19454 Reviewed by Gavin Barraclough. 19455 19456 Fixed typo. 19457 https://bugs.webkit.org/show_bug.cgi?id=28691 19458 19459 * parser/Parser.h: 19460 (JSC::Parser::parse): 19461 194622009-08-27 Oliver Hunt <oliver@apple.com> 19463 19464 Reviewed by Maciej Stachowiak. 19465 19466 JSON Stringifier does not follow ES5 spec for handling of Number, String and Boolean objects 19467 https://bugs.webkit.org/show_bug.cgi?id=28797 19468 19469 Fixed unwrapBoxedPrimitive to do the right thing, which necessitated a couple of new exception 19470 checks, and corrected the logic in gap to correctly convert Number and String objects. 19471 19472 * runtime/JSONObject.cpp: 19473 (JSC::unwrapBoxedPrimitive): 19474 (JSC::gap): 19475 (JSC::Stringifier::Stringifier): 19476 (JSC::Stringifier::appendStringifiedValue): 19477 194782009-08-27 Oliver Hunt <oliver@apple.com> 19479 19480 Reviewed by Adam Roben. 19481 19482 JSON.stringify replacer array does not accept values that are not string primitives. 19483 https://bugs.webkit.org/show_bug.cgi?id=28788 19484 19485 Update the JSON stringifier to initialise its replacer array according to the most 19486 recent version of the spec. 19487 19488 * runtime/Identifier.h: 19489 (JSC::Identifier::from): 19490 * runtime/JSONObject.cpp: 19491 (JSC::Stringifier::Stringifier): 19492 194932009-08-27 Alexey Proskuryakov <ap@apple.com> 19494 19495 Reviewed by Oliver Hunt. 19496 19497 https://bugs.webkit.org/show_bug.cgi?id=28753 19498 <rdar://problem/7173448> Excessive number of threads (and a crash) 19499 19500 * wtf/Threading.h: (WTF::atomicIncrement): Changed atomicIncrement to match decrement 19501 and return the new value. Also added using directives for these functions, to match 19502 te rest of WTF. 19503 195042009-08-27 Brent Fulgham <bfulgham@webkit.org> 19505 19506 Reviewed by Adam Roben. 19507 19508 Link the testapi against CFLite when building the WinCairo port. 19509 19510 * JavaScriptCore.vcproj/testapi/testapi.vcproj: Add new Release_CFLite 19511 target. Update all targets to inherit from either the 19512 JavaScriptCF.vsprops (Apple target) or the JavaScriptCFLite.vsprops 19513 file (WinCairo target). 19514 * JavaScriptCore.vcproj/testapi/testapiCommon.vsprops: Remove 19515 input file CoreFoundation.lib. This is provided by either the 19516 JavaScriptCF.vsprops or JavaScriptCFLite.vsprops file. 19517 195182009-08-27 Steve Falkenburg <sfalken@apple.com> 19519 19520 Reviewed by Geoff Garen. 19521 19522 Fix Windows-specific crash due to missing memory clearing call. 19523 19524 * runtime/Collector.cpp: 19525 (JSC::Heap::allocateBlock): 19526 195272009-08-27 Brent Fulgham <bfulgham@webkit.org> 19528 19529 Build fix: JavaScriptCore_debug.def missing some exports. Apple 19530 Windows build does not use this file, so it was not noticed previously. 19531 19532 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 19533 195342009-08-27 Gavin Barraclough <barraclough@apple.com> 19535 19536 Reviewed by Oliver Hunt. 19537 19538 x86-64 GTK broken due to code offsets changing, pointers sometimes packed into immediates. 19539 https://bugs.webkit.org/show_bug.cgi?id=28317 19540 19541 Missed one, fix part II. 19542 19543 * assembler/MacroAssemblerX86Common.h: 19544 (JSC::MacroAssemblerX86Common::move): 19545 * assembler/X86Assembler.h: 19546 (JSC::CAN_SIGN_EXTEND_8_32): 19547 195482009-08-27 Oliver Hunt <oliver@apple.com> 19549 19550 Reviewed by Adam Roben. 19551 19552 JSON.stringify replacer array does not accept values that are not string primitives. 19553 https://bugs.webkit.org/show_bug.cgi?id=28788 19554 19555 Update the JSON stringifier to initialise its replacer array according to the most 19556 recent version of the spec. 19557 19558 * runtime/Identifier.h: 19559 (JSC::Identifier::from): 19560 * runtime/JSONObject.cpp: 19561 (JSC::Stringifier::Stringifier): 19562 195632009-08-27 Oliver Hunt <oliver@apple.com> 19564 19565 Reviewed by Alexey Proskuryakov. 19566 19567 JSON parser accepts trailing comma in array literals 19568 https://bugs.webkit.org/show_bug.cgi?id=28779 19569 19570 Update parser to correctly fail if there's a trailing comma. 19571 19572 * runtime/LiteralParser.cpp: 19573 (JSC::LiteralParser::parse): 19574 195752009-08-26 Oliver Hunt <oliver@apple.com> 19576 19577 Reviewed by Gavin Barraclough. 19578 19579 'this' in JSON.parse reviver is the global object 19580 https://bugs.webkit.org/show_bug.cgi?id=28752 19581 19582 This is a technically simple change, we merely update the code for calling 19583 the reviver function to pass the correct this object. Doing so however 19584 exposes the holder to arbitrary mutation by the reviver function so it is 19585 necessary for us to now guard all property accesses against the possibility 19586 of failure. 19587 19588 * runtime/JSArray.h: 19589 JSON needs to delete a property from the array, so we friend its 19590 Walker class so that we can make a non-virtual call to the arrays 19591 delete and getOwnPropertySlot methods. 19592 * runtime/JSONObject.cpp: 19593 (JSC::Walker::callReviver): 19594 We need to pass the correct this object 19595 (JSC::Walker::walk): 19596 Update calls to callReviver, and update property logic logic 19597 to correctly handle the holder being mutated by the reviver 19598 function. 19599 196002009-08-26 Alice Liu <alice.liu@apple.com> 19601 19602 Windows build fix: added some exported symbols 19603 19604 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 19605 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 19606 196072009-08-26 Geoffrey Garen <ggaren@apple.com> 19608 19609 Windows build fix: Removed some exported symbols that no longer exist. 19610 19611 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 19612 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 19613 196142009-08-26 Gavin Barraclough <barraclough@apple.com> 19615 19616 Reviewed by Olliejver Hunt. 19617 19618 x86-64 GTK broken due to code offsets changing, pointers sometimes packed into immediates. 19619 https://bugs.webkit.org/show_bug.cgi?id=28317 19620 19621 We rely on a slightly OS X specific behaviour, that x86-64 applications have a 4Gb zero page, 19622 so pointers are never representable as a 32-bit integer, and always have to be represented by 19623 a separate immediate load instruction, rather than within the immediate field of an arithmetic 19624 or memory operation. 19625 19626 We explicitly check for a couple of cases where a value might be representable in 32-bit, but 19627 these probably never kick in on Mac OS, and only kick in to hose GTK. Deleting these does not 19628 show a performance degradation on SunSpider. Remove. 19629 19630 * assembler/MacroAssemblerX86_64.h: 19631 (JSC::MacroAssemblerX86_64::storePtr): 19632 (JSC::MacroAssemblerX86_64::branchPtr): 19633 196342009-08-26 Geoffrey Garen <ggaren@apple.com> 19635 19636 Reviewed by Oliver Hunt. 19637 19638 A bit of Collector refatoring. 19639 19640 SunSpider says no change. v8 says 1.003x faster (1.02x faster on splay). 19641 19642 * JavaScriptCore.exp: 19643 19644 * runtime/JSCell.cpp: 19645 (JSC::JSCell::toPrimitive): 19646 (JSC::JSCell::getPrimitiveNumber): 19647 (JSC::JSCell::toBoolean): 19648 (JSC::JSCell::toNumber): 19649 (JSC::JSCell::toString): 19650 (JSC::JSCell::toObject): Removed pure virtual functions from 19651 JSCell, so the collector can construct one. This allowed 19652 me to remove a bunch of ASSERT_NOT_REACHED throughout the 19653 code, too. 19654 19655 * runtime/JSCell.h: 19656 (JSC::JSCell::JSCell): ditto 19657 (JSC::Heap::heap): Inlined this function because it's trivial. 19658 19659 * JavaScriptCore.exp: 19660 19661 * runtime/Collector.cpp: 19662 (JSC::Heap::destroy): 19663 (JSC::Heap::allocateBlock): 19664 (JSC::Heap::freeBlock): 19665 (JSC::Heap::freeBlocks): Renamed freeHeap to freeBlocks, since 19666 it doesn't actually free the Heap object. 19667 (JSC::Heap::heapAllocate): 19668 (JSC::Heap::sweep): 19669 * runtime/Collector.h: Refactored block allocation and destruction 19670 into helper functions. 19671 19672 * runtime/GetterSetter.cpp: 19673 * runtime/JSAPIValueWrapper.cpp: 19674 * runtime/JSPropertyNameIterator.cpp: Removed dummy implementations 19675 of pure virtual functions. (See above.) 19676 19677=== End re-roll-in of r47738:47740 with Windows crash fixed === 19678 196792009-08-26 Geoffrey Garen <ggaren@apple.com> 19680 19681 Build fix: start out with a 32-bit value to avoid a shortening warning. 19682 19683 * runtime/Collector.cpp: 19684 (JSC::Heap::sweep): 19685 196862009-08-24 Geoffrey Garen <ggaren@apple.com> 19687 19688 Reviewed by Oliver Hunt. 19689 19690 Substantially reduced VM thrash in the GC heap. 19691 19692 1.08x faster on v8 (1.60x faster on v8-splay). 19693 19694 1.40x faster on bench-alloc-nonretained. 19695 19696 1.90x faster on bench-alloc-retained. 19697 19698 SunSpider says no change. 19699 19700 * runtime/Collector.cpp: 19701 (JSC::Heap::heapAllocate): Fixed a long-standing bug: update a few local 19702 variables unconditionally after calling collect(), since they may be used 19703 even if we don't "goto scan". (In the bug I saw, usedBlocks got out of 19704 sync with heap.usedBlocks). 19705 (JSC::Heap::sweep): Keep enough free heap space to accomodate 19706 the number of objects we'll allocate before the next GC, plus 25%, for 19707 good measure. 19708 * runtime/Collector.h: Bumped the block size to 256k. This seems to give 19709 the best cache performance, and it prevents us from initiating lots of 19710 VM traffic to recover very small chunks of memory. 19711 19712=== Begin re-roll-in of r47738:47740 with Windows crash fixed === 19713 197142009-08-25 Drew Wilson <atwilson@google.com> 19715 19716 Reviewed by David Levin. 19717 19718 postMessage() spec now supports sending arrays of ports 19719 https://bugs.webkit.org/show_bug.cgi?id=26902 19720 19721 Added OwnPtr to VectorTraits so we can store OwnPtrs in Vectors. 19722 19723 * wtf/VectorTraits.h: 19724 197252009-08-26 Xan Lopez <xlopez@igalia.com> 19726 19727 Rubber-stamped by Gustavo Noronha. 19728 19729 Remove duplicated files from file list. 19730 19731 * GNUmakefile.am: 19732 197332009-08-26 Oliver Hunt <oliver@apple.com> 19734 19735 Reviewed by NOBODY (Build fix). 19736 19737 More export fixes. 19738 19739 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 19740 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 19741 197422009-08-26 Oliver Hunt <oliver@apple.com> 19743 19744 Reviewed by NOBODY (Build fix). 19745 19746 Hopefully fix all the exports from JSC on windows 19747 19748 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 19749 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 19750 197512009-08-26 Oliver Hunt <oliver@apple.com> 19752 19753 Reviewed by NOBODY (Build fixes). 19754 19755 Forgot I added files to JavaScriptCore. 19756 19757 * GNUmakefile.am: 19758 * JavaScriptCore.gypi: 19759 * JavaScriptCore.pri: 19760 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: 19761 * JavaScriptCoreSources.bkl: 19762 197632009-08-25 Oliver Hunt <oliver@apple.com> 19764 19765 Reviewed by Gavin Barraclough. 19766 19767 [ES5] Implement getOwnPropertyDescriptor 19768 https://bugs.webkit.org/show_bug.cgi?id=28724 19769 19770 Implement the core runtime support for getOwnPropertyDescriptor. 19771 This adds a virtual getOwnPropertyDescriptor method to every class 19772 that implements getOwnPropertySlot that shadows the behaviour of 19773 getOwnPropertySlot. The alternative would be to make getOwnPropertySlot 19774 (or PropertySlots in general) provide property attribute information, 19775 but quick testing showed this to be a regression. 19776 19777 * JavaScriptCore.exp: 19778 * JavaScriptCore.xcodeproj/project.pbxproj: 19779 * runtime/Arguments.cpp: 19780 (JSC::Arguments::getOwnPropertyDescriptor): 19781 * runtime/Arguments.h: 19782 * runtime/ArrayPrototype.cpp: 19783 (JSC::ArrayPrototype::getOwnPropertyDescriptor): 19784 * runtime/ArrayPrototype.h: 19785 * runtime/CommonIdentifiers.h: 19786 * runtime/DatePrototype.cpp: 19787 (JSC::DatePrototype::getOwnPropertyDescriptor): 19788 * runtime/DatePrototype.h: 19789 * runtime/JSArray.cpp: 19790 (JSC::JSArray::getOwnPropertyDescriptor): 19791 * runtime/JSArray.h: 19792 * runtime/JSByteArray.cpp: 19793 (JSC::JSByteArray::getOwnPropertyDescriptor): 19794 * runtime/JSByteArray.h: 19795 * runtime/JSFunction.cpp: 19796 (JSC::JSFunction::getOwnPropertyDescriptor): 19797 * runtime/JSFunction.h: 19798 * runtime/JSGlobalObject.h: 19799 (JSC::JSGlobalObject::getOwnPropertyDescriptor): 19800 * runtime/JSNotAnObject.cpp: 19801 (JSC::JSNotAnObject::getOwnPropertyDescriptor): 19802 * runtime/JSNotAnObject.h: 19803 * runtime/JSONObject.cpp: 19804 (JSC::JSONObject::getOwnPropertySlot): 19805 (JSC::JSONObject::getOwnPropertyDescriptor): 19806 * runtime/JSONObject.h: 19807 * runtime/JSObject.cpp: 19808 (JSC::JSObject::getOwnPropertyDescriptor): 19809 (JSC::JSObject::getPropertyDescriptor): 19810 * runtime/JSObject.h: 19811 * runtime/JSString.cpp: 19812 (JSC::JSString::getStringPropertyDescriptor): 19813 (JSC::JSString::getOwnPropertyDescriptor): 19814 * runtime/JSString.h: 19815 * runtime/JSVariableObject.cpp: 19816 (JSC::JSVariableObject::symbolTableGet): 19817 * runtime/JSVariableObject.h: 19818 * runtime/Lookup.h: 19819 (JSC::getStaticPropertyDescriptor): 19820 (JSC::getStaticFunctionDescriptor): 19821 (JSC::getStaticValueDescriptor): 19822 Add property descriptor equivalents of the lookup 19823 table access functions 19824 19825 * runtime/MathObject.cpp: 19826 (JSC::MathObject::getOwnPropertySlot): 19827 (JSC::MathObject::getOwnPropertyDescriptor): 19828 * runtime/MathObject.h: 19829 * runtime/NumberConstructor.cpp: 19830 (JSC::NumberConstructor::getOwnPropertyDescriptor): 19831 * runtime/NumberConstructor.h: 19832 * runtime/ObjectConstructor.cpp: 19833 (JSC::ObjectConstructor::ObjectConstructor): 19834 (JSC::objectConstructorGetOwnPropertyDescriptor): 19835 * runtime/PropertyDescriptor.cpp: Added. 19836 (JSC::PropertyDescriptor::writable): 19837 (JSC::PropertyDescriptor::enumerable): 19838 (JSC::PropertyDescriptor::configurable): 19839 (JSC::PropertyDescriptor::hasAccessors): 19840 (JSC::PropertyDescriptor::setUndefined): 19841 (JSC::PropertyDescriptor::getter): 19842 (JSC::PropertyDescriptor::setter): 19843 (JSC::PropertyDescriptor::setDescriptor): 19844 (JSC::PropertyDescriptor::setAccessorDescriptor): 19845 * runtime/PropertyDescriptor.h: Added. 19846 (JSC::PropertyDescriptor::PropertyDescriptor): 19847 (JSC::PropertyDescriptor::attributes): 19848 (JSC::PropertyDescriptor::isValid): 19849 (JSC::PropertyDescriptor::value): 19850 * runtime/RegExpConstructor.cpp: 19851 (JSC::RegExpConstructor::getOwnPropertyDescriptor): 19852 * runtime/RegExpConstructor.h: 19853 * runtime/RegExpMatchesArray.h: 19854 (JSC::RegExpMatchesArray::getOwnPropertyDescriptor): 19855 * runtime/RegExpObject.cpp: 19856 (JSC::RegExpObject::getOwnPropertyDescriptor): 19857 * runtime/RegExpObject.h: 19858 * runtime/StringObject.cpp: 19859 (JSC::StringObject::getOwnPropertyDescriptor): 19860 * runtime/StringObject.h: 19861 * runtime/StringPrototype.cpp: 19862 (JSC::StringPrototype::getOwnPropertyDescriptor): 19863 * runtime/StringPrototype.h: 19864 198652009-08-24 Gavin Barraclough <barraclough@apple.com> 19866 19867 Reviewed by Darin Adler. 19868 19869 How many copies of the parameters do you need? 19870 https://bugs.webkit.org/show_bug.cgi?id=28701 19871 19872 The function parameters in JSC get copied a lot - and unnecessarily so. 19873 19874 Originally this happened due to duplicating FunctionBodyNodes on recompilation, 19875 though the problem has been exacerbated by copying the parameters from the 19876 original function body onto the executable, then back onto the real body that 19877 will be generated (this happens on every function). And this is all made worse 19878 since the data structures in question are a little ugly - C style arrays of C++ 19879 objects containing ref counts, so they need a full copy-construct (rather than 19880 a simple memcpy). 19881 19882 This can all be greatly simplified by just punting the parameters off into 19883 their own ref-counted object, and forgoing all the copying. 19884 19885 ~no performance change, possible slight progression. 19886 19887 * bytecompiler/BytecodeGenerator.cpp: 19888 (JSC::BytecodeGenerator::BytecodeGenerator): 19889 * bytecompiler/BytecodeGenerator.h: 19890 (JSC::BytecodeGenerator::makeFunction): 19891 * parser/Nodes.cpp: 19892 (JSC::FunctionParameters::FunctionParameters): 19893 (JSC::FunctionBodyNode::FunctionBodyNode): 19894 (JSC::FunctionBodyNode::finishParsing): 19895 * parser/Nodes.h: 19896 (JSC::FunctionBodyNode::parameters): 19897 (JSC::FunctionBodyNode::parameterCount): 19898 * runtime/Executable.cpp: 19899 (JSC::FunctionExecutable::~FunctionExecutable): 19900 (JSC::FunctionExecutable::compile): 19901 (JSC::FunctionExecutable::reparseExceptionInfo): 19902 (JSC::FunctionExecutable::fromGlobalCode): 19903 (JSC::FunctionExecutable::paramString): 19904 * runtime/Executable.h: 19905 (JSC::FunctionExecutable::FunctionExecutable): 19906 (JSC::FunctionExecutable::parameterCount): 19907 199082009-08-25 Brent Fulgham <bfulgham@webkit.org> 19909 19910 Reviewed by NOBODY (Buildfix). 19911 19912 * JavaScriptCore.vcproj/jsc/jsc.vcproj: Add Debug_CFLite target 19913 that inherits from the debug_wincairo property sheet and therefore 19914 links to the proper debug library. 19915 * JavaScriptCore.vcproj/testapi/testapi.vcproj: Add Debug_CFLite target 19916 that inherits from the debug_wincairo property sheet and therefore 19917 links to the proper debug library. 19918 199192009-08-25 Chris Marrin <cmarrin@apple.com> 19920 19921 Reviewed by Simon Fraser. 19922 19923 Export tryFastMalloc for Canvas3D work 19924 https://bugs.webkit.org/show_bug.cgi?id=28018 19925 19926 * JavaScriptCore.exp: 19927 199282009-08-25 David Levin <levin@chromium.org> 19929 19930 Reviewed by Adam Roben. 19931 19932 PLATFORM(CFNETWORK) should be USE(CFNETWORK). 19933 https://bugs.webkit.org/show_bug.cgi?id=28713 19934 19935 * wtf/Platform.h: Added a #define to catch this issue in the 19936 future. The define would generate an error on gcc without the 19937 space in the expansion, but Visual C++ needs the space to cause an error. 19938 199392009-08-24 Brent Fulgham <bfulgham@webkit.org> 19940 19941 Reviewed by Steve Falkenburg. 19942 19943 Revise CFLite Debug build to emit DLL's with _debug label. 19944 https://bugs.webkit.org/show_bug.cgi?id=28695. 19945 19946 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: Modify 19947 Cairo debug build to inherit from new debug_cairo property sheet. 19948 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCoreCFLite.vsprops: 19949 Modify to look for debug CFLite when in debug build. 19950 199512009-08-24 Gavin Barraclough <barraclough@apple.com> 19952 19953 Reviewed by Oliver Adler & Darin Hunt. 19954 19955 https://bugs.webkit.org/show_bug.cgi?id=28691 19956 Do not retain ScopeNodes outside of parsing 19957 19958 There is now no need for these to exist outside of parsing - their use in the runtime is replaced by Executable types. 19959 19960 * bytecode/EvalCodeCache.h: 19961 (JSC::EvalCodeCache::get): 19962 * bytecompiler/BytecodeGenerator.cpp: 19963 (JSC::BytecodeGenerator::BytecodeGenerator): 19964 (JSC::BytecodeGenerator::emitNewFunction): 19965 (JSC::BytecodeGenerator::emitNewFunctionExpression): 19966 * bytecompiler/BytecodeGenerator.h: 19967 (JSC::BytecodeGenerator::makeFunction): 19968 * debugger/Debugger.cpp: 19969 (JSC::Debugger::recompileAllJSFunctions): 19970 (JSC::evaluateInGlobalCallFrame): 19971 * debugger/DebuggerCallFrame.cpp: 19972 (JSC::DebuggerCallFrame::evaluate): 19973 * interpreter/Interpreter.cpp: 19974 (JSC::Interpreter::execute): 19975 (JSC::Interpreter::prepareForRepeatCall): 19976 (JSC::Interpreter::privateExecute): 19977 * jit/JITStubs.cpp: 19978 (JSC::DEFINE_STUB_FUNCTION): 19979 * parser/Nodes.cpp: 19980 (JSC::ScopeNodeData::ScopeNodeData): 19981 (JSC::ProgramNode::create): 19982 (JSC::EvalNode::create): 19983 (JSC::FunctionBodyNode::create): 19984 * parser/Nodes.h: 19985 (JSC::ScopeNode::adoptData): 19986 (JSC::FunctionBodyNode::parameterCount): 19987 * parser/Parser.cpp: 19988 * parser/Parser.h: 19989 (JSC::Parser::arena): 19990 (JSC::Parser::Parser): 19991 (JSC::Parser::parse): 19992 * runtime/ArrayPrototype.cpp: 19993 (JSC::isNumericCompareFunction): 19994 (JSC::arrayProtoFuncSort): 19995 * runtime/Completion.cpp: 19996 (JSC::checkSyntax): 19997 (JSC::evaluate): 19998 * runtime/Executable.cpp: 19999 (JSC::FunctionExecutable::~FunctionExecutable): 20000 (JSC::EvalExecutable::compile): 20001 (JSC::ProgramExecutable::checkSyntax): 20002 (JSC::ProgramExecutable::compile): 20003 (JSC::FunctionExecutable::compile): 20004 (JSC::EvalExecutable::generateJITCode): 20005 (JSC::ProgramExecutable::generateJITCode): 20006 (JSC::FunctionExecutable::generateJITCode): 20007 (JSC::FunctionExecutable::reparseExceptionInfo): 20008 (JSC::EvalExecutable::reparseExceptionInfo): 20009 (JSC::FunctionExecutable::recompile): 20010 (JSC::FunctionExecutable::fromGlobalCode): 20011 (JSC::FunctionExecutable::copyParameters): 20012 (JSC::FunctionExecutable::paramString): 20013 * runtime/Executable.h: 20014 (JSC::ScriptExecutable::ScriptExecutable): 20015 (JSC::ScriptExecutable::sourceID): 20016 (JSC::ScriptExecutable::sourceURL): 20017 (JSC::ScriptExecutable::lineNo): 20018 (JSC::ScriptExecutable::lastLine): 20019 (JSC::ScriptExecutable::usesEval): 20020 (JSC::ScriptExecutable::usesArguments): 20021 (JSC::ScriptExecutable::needsActivation): 20022 (JSC::ScriptExecutable::recordParse): 20023 (JSC::EvalExecutable::bytecode): 20024 (JSC::EvalExecutable::jitCode): 20025 (JSC::ProgramExecutable::bytecode): 20026 (JSC::ProgramExecutable::reparseExceptionInfo): 20027 (JSC::ProgramExecutable::jitCode): 20028 (JSC::FunctionExecutable::FunctionExecutable): 20029 (JSC::FunctionExecutable::make): 20030 (JSC::FunctionExecutable::bytecode): 20031 (JSC::FunctionExecutable::isGenerated): 20032 (JSC::FunctionExecutable::name): 20033 (JSC::FunctionExecutable::parameterCount): 20034 (JSC::FunctionExecutable::jitCode): 20035 * runtime/FunctionConstructor.cpp: 20036 (JSC::constructFunction): 20037 * runtime/JSGlobalData.cpp: 20038 (JSC::JSGlobalData::numericCompareFunction): 20039 * runtime/JSGlobalObjectFunctions.cpp: 20040 (JSC::globalFuncEval): 20041 200422009-08-24 Darin Adler <darin@apple.com> 20043 20044 * runtime/ObjectPrototype.cpp: 20045 (JSC::ObjectPrototype::put): Landed revised version I had tested but forgot 20046 to land. Leave out the branch, since we don't need one. 20047 200482009-08-24 Darin Adler <darin@apple.com> 20049 20050 Reviewed by Geoff Garen. 20051 20052 Array index miss case creates a string every time 20053 https://bugs.webkit.org/show_bug.cgi?id=28664 20054 20055 SunSpider test results I saw: 20056 20057 0.5% faster overall 20058 1% faster on crypto-aes 20059 20% faster on crypto-md5 20060 13% faster on crypto-sha1 20061 20062 * runtime/ObjectPrototype.cpp: 20063 (JSC::ObjectPrototype::ObjectPrototype): Initialize m_hasNoPropertiesWithUInt32Names 20064 to true. 20065 (JSC::ObjectPrototype::put): Clearly m_hasNoPropertiesWithUInt32Names if the new 20066 property has a name that is the string form of a UInt32. 20067 (JSC::ObjectPrototype::getOwnPropertySlot): Don't call JSObject::getOwnPropertySlot 20068 if m_hasNoPropertiesWithUInt32Names is true, and it is highly likely to be true. 20069 20070 * runtime/ObjectPrototype.h: Added declarations for the above. 20071 200722009-08-24 Gustavo Noronha Silva <gustavo.noronha@collabora.co.uk> 20073 20074 Unreviewed. Fix a typo in my distcheck build fix. 20075 20076 * GNUmakefile.am: 20077 200782009-08-23 Gustavo Noronha Silva <gns@gnome.org> 20079 20080 Unreviewed build fix for make distcheck. 20081 20082 * GNUmakefile.am: Added files required for the build. 20083 200842009-08-22 Maciej Stachowiak <mjs@apple.com> 20085 20086 Reviewed by Mark Rowe. 20087 20088 REGRESSION(r47639-r47660): Webkit crashes on launch on PowerPC 20089 https://bugs.webkit.org/show_bug.cgi?id=28655 20090 20091 * runtime/JSFunction.cpp: 20092 (JSC::JSFunction::JSFunction): Initialize properly with a VPtrHackExecutable. 20093 * wtf/Platform.h: 20094 200952009-08-22 Darin Adler <darin@apple.com> 20096 20097 Fix storage leak from syntax tree arena allocation patch. 20098 20099 * parser/Nodes.h: CommaNode needs to inherit from ParserArenaDeletable 20100 because it has a vector. 20101 201022009-08-21 Darin Adler <darin@apple.com> 20103 20104 Fix Qt build. 20105 20106 * parser/Nodes.cpp: 20107 (JSC::ScopeNodeData::ScopeNodeData): Made non-inline again. 20108 This is used outside Nodes.cpp so can't be inline unless 20109 it is in the header. 20110 201112009-08-21 Darin Adler <darin@apple.com> 20112 20113 Two loose ends from the last commit. 20114 20115 * JavaScriptCore.xcodeproj/project.pbxproj: Made ParserArena.h 20116 and create_hash_table project-internal instead of "private". 20117 * runtime/Executable.h: Removed accidentally-added constructor. 20118 201192009-08-21 Darin Adler <darin@apple.com> 20120 20121 Reviewed by Gavin Barraclough. 20122 20123 Syntax tree nodes should use arena allocation 20124 https://bugs.webkit.org/show_bug.cgi?id=25674 20125 20126 Use an actual arena now. 0.6% speedup on SunSpider. 20127 20128 New and improved with 100% less leaking of the universe. 20129 20130 * JavaScriptCore.exp: 20131 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 20132 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 20133 Removed all exports involving the class FunctionBodyNode, which no 20134 longer needs to be used outside JavaScriptCore. 20135 20136 * JavaScriptCore.xcodeproj/project.pbxproj: Made Nodes.h and 20137 Executable.h project-internal instead of "private". 20138 20139 * bytecompiler/BytecodeGenerator.cpp: 20140 (JSC::BytecodeGenerator::BytecodeGenerator): Updated since VarStack 20141 contains const Identifier* now. 20142 20143 * parser/Grammar.y: Made identifiers from the lexer be const 20144 Identifier* and updated since VarStack contains const Identifier* now. 20145 20146 * parser/Lexer.cpp: 20147 (JSC::Lexer::setCode): Pass in ParserArena, used for identifiers. 20148 (JSC::Lexer::makeIdentifier): Changed return type to const Identifier* 20149 and changed to call ParserArena. 20150 (JSC::Lexer::clear): Removed the code to manage m_identifiers and 20151 added code to set m_arena to 0. 20152 * parser/Lexer.h: Updated for changes above. 20153 20154 * parser/NodeConstructors.h: 20155 (JSC::ParserArenaFreeable::operator new): Added. Calls allocateFreeable 20156 on the arena. 20157 (JSC::ParserArenaDeletable::operator new): Changed to call the 20158 allocateDeletable function on the arena instead of deleteWithArena. 20159 (JSC::PropertyNode::PropertyNode): Added new constructor that makes 20160 numeric identifiers. Some day we might want to optimize this for 20161 integers so it doesn't create a string for each one. 20162 (JSC::ContinueNode::ContinueNode): Initialize m_ident to nullIdentifier 20163 since it's now a const Identifier& so it can't be left uninitialized. 20164 (JSC::BreakNode::BreakNode): Ditto. 20165 (JSC::CaseClauseNode::CaseClauseNode): Updated to use SourceElements* 20166 to keep track of the statements rather than a separate statement vector. 20167 (JSC::BlockNode::BlockNode): Ditto. 20168 (JSC::ForInNode::ForInNode): Initialize m_ident to nullIdentifier. 20169 20170 * parser/Nodes.cpp: Moved the comment explaining emitBytecode in here. 20171 It seemed strangely out of place in the header. 20172 (JSC::ThrowableExpressionData::emitThrowError): Added an overload for 20173 UString as well as Identifier. 20174 (JSC::SourceElements::singleStatement): Added. 20175 (JSC::SourceElements::lastStatement): Added. 20176 (JSC::RegExpNode::emitBytecode): Changed the throwError code to use 20177 the substitution mechanism instead of doing a string append. 20178 (JSC::SourceElements::emitBytecode): Added. Replaces the old 20179 statementListEmitCode function, since we now keep the SourceElements 20180 objects around. 20181 (JSC::BlockNode::lastStatement): Added. 20182 (JSC::BlockNode::emitBytecode): Changed to use emitBytecode instead of 20183 statementListEmitCode. 20184 (JSC::CaseClauseNode::emitBytecode): Added. 20185 (JSC::CaseBlockNode::emitBytecodeForBlock): Changed to use emitBytecode 20186 instead of statementListEmitCode. 20187 (JSC::ScopeNodeData::ScopeNodeData): Changed to store the 20188 SourceElements* instead of using releaseContentsIntoVector. 20189 (JSC::ScopeNode::emitStatementsBytecode): Added. 20190 (JSC::ScopeNode::singleStatement): Added. 20191 (JSC::ProgramNode::emitBytecode): Call emitStatementsBytecode instead 20192 of statementListEmitCode. 20193 (JSC::EvalNode::emitBytecode): Ditto. 20194 (JSC::FunctionBodyNode::emitBytecode): Call emitStatementsBytecode 20195 insetad of statementListEmitCode and check for the return node using 20196 the new functions. 20197 20198 * parser/Nodes.h: Changed VarStack to store const Identifier* instead 20199 of Identifier and rely on the arena to control lifetime. Added a new 20200 ParserArenaFreeable class. Made ParserArenaDeletable inherit from 20201 FastAllocBase instead of having its own operator new. Base the Node 20202 class on ParserArenaFreeable. Changed the various Node classes 20203 to use const Identifier& instead of Identifier to avoid the need to 20204 call their destructors and allow them to function as "freeable" in the 20205 arena. Removed extraneous JSC_FAST_CALL on definitions of inline functions. 20206 Changed ElementNode, PropertyNode, ArgumentsNode, ParameterNode, 20207 CaseClauseNode, ClauseListNode, and CaseBlockNode to use ParserArenaFreeable 20208 as a base class since they do not descend from Node. Eliminated the 20209 StatementVector type and instead have various classes use SourceElements* 20210 instead of StatementVector. This prevents those classes from having to 20211 use ParserArenaDeletable to make sure the vector destructor is called. 20212 20213 * parser/Parser.cpp: 20214 (JSC::Parser::parse): Pass the arena to the lexer. 20215 20216 * parser/Parser.h: Added an include of ParserArena.h, which is no longer 20217 included by Nodes.h. 20218 (JSC::Parser::parseFunctionFromGlobalCode): Changed to use the 20219 singleStatement function, since there is no longer any children function. 20220 Removed some unneeded use of RefPtr. 20221 20222 * parser/ParserArena.cpp: 20223 (JSC::ParserArena::ParserArena): Added. Initializes the new members, 20224 m_freeableMemory, m_freeablePoolEnd, and m_identifiers. 20225 (JSC::ParserArena::freeablePool): Added. Computes the pool pointer, 20226 since we store only the current pointer and the end of pool pointer. 20227 (JSC::ParserArena::deallocateObjects): Added. Contains the common 20228 memory-deallocation logic used by both the destructor and the 20229 reset function. 20230 (JSC::ParserArena::~ParserArena): Changed to call deallocateObjects. 20231 (JSC::ParserArena::reset): Ditto. Also added code to zero out the 20232 new structures, and switched to use clear() instead of shrink(0) since 20233 we don't really reuse arenas. 20234 (JSC::ParserArena::makeNumericIdentifier): Added. 20235 (JSC::ParserArena::allocateFreeablePool): Added. Used when the pool 20236 is empty. 20237 (JSC::ParserArena::isEmpty): Added. No longer inline, which is fine 20238 since this is used only for assertions at the moment. 20239 (JSC::ParserArena::derefWithArena): Make non-inline. 20240 20241 * parser/ParserArena.h: Added an actual arena of "freeable" objects, 20242 ones that don't need destructors to be called. Also added a separate 20243 IdentifierArena object, a segmented vector of identifiers that used 20244 to be in the Lexer. 20245 20246 * runtime/Executable.h: Moved the definition of the 20247 FunctionExecutable::make function here. It can't go in JSFunction.h 20248 since that header has to be used outside JavaScriptCore and so can't 20249 include this, which includes Nodes.h. The function could be moved 20250 elswhere if we don't want to include JSFunction.h in this header, but 20251 for now this seems to be the best place. 20252 20253 * runtime/JSFunction.h: Removed the include of Executable.h and 20254 definition of the FunctionExecutable::make function. 20255 20256 * wtf/FastMalloc.cpp: Fixed an incorrect comment. 20257 202582009-08-21 Mark Rowe <mrowe@apple.com> 20259 20260 Fix the non-JIT build. 20261 20262 * runtime/Executable.cpp: 20263 * runtime/Executable.h: 20264 202652009-08-21 Gavin Barraclough <barraclough@apple.com> 20266 20267 Speculative QuickTime build fix. 20268 20269 * runtime/JSArray.cpp: 20270 202712009-08-21 Gavin Barraclough <barraclough@apple.com> 20272 20273 Speculative QT build fix. 20274 20275 * runtime/StringPrototype.cpp: 20276 202772009-08-21 Gavin Barraclough <barraclough@apple.com> 20278 20279 Reviewed by Oliver Hunt. 20280 20281 Restructure Executable types so that host functions do not hold a FunctionExecutable. 20282 https://bugs.webkit.org/show_bug.cgi?id=28621 20283 20284 All JSFunction objects have a pointer to an Executable*. This is currently always a 20285 FunctionExecutable, however this has a couple of drawbacks. Host functions do not 20286 store a range of information that the FunctionExecutable provides (source, name, 20287 CodeBlock & information presently held on the FunctionBodyNode). 20288 20289 [ * nearly all... see below! ] 20290 20291 Instead, make JSFunctions hold a pointer to an ExecutableBase, move fields specific 20292 to JS sourced executable types (source, node) into a new subclass (ScriptExecutable), 20293 and create a new NativeExecutable type. We now provide a new method in JSFunction 20294 to access & downcast to FunctionExecutable, but in doing so we can make an early 20295 check (with an ASSERT) to ensure that the Executable read from a function will only 20296 be treated as a FunctionExecutable (and thus the JS sepcific fields will only be 20297 accessed) if the JSFunction is not a host function. 20298 20299 There is one JSFunction that currently does not have an Executable, which is the 20300 object created to allow us to read out the vtable pointer. By making this change 20301 we can also add a new Executable type fror this object (VPtrHackExecutable). 20302 Since this means that really all JSFunctions have an Executable we no longer have 20303 to null-check m_executable before us it - particularly in isHostFunction(). 20304 20305 This patch removes CacheableEvalExecutable, since all subclasses of ExecutableBase 20306 can now be ref-counted - since both JSFunction holds (and ref-counts) an ExecutableBase 20307 that might be a FunctionExecutable or a NativeExecutable. This does now mean that all 20308 ProgramExecutables and EvalExecutables (unnecessarily) provide an interface to be 20309 ref-counted, however this seems less-bad than host functions unnecessarily providing 20310 interface to access non-host specific information. 20311 20312 The class hierarcy has changed from this: 20313 20314 - ExecutableBase 20315 - ProgramExecutable 20316 - EvalExecutable 20317 - CacheableEvalExecutable (also RefCounted by multiple-inheritance) 20318 - FunctionExecutable (also RefCounted by multiple-inheritance, 'special' FunctionExecutable also used for host functions) 20319 20320 To this: 20321 20322 - RefCounted 20323 - ExecutableBase 20324 - NativeExecutable 20325 - VPtrHackExecutable 20326 - ScriptExecutable 20327 - ProgramExecutable 20328 - EvalExecutable 20329 - FunctionExecutable 20330 20331 This patch speeds up sunspidey by a couple of ms (presumably due to the changes to isHostFunction()). 20332 20333 * bytecode/CodeBlock.cpp: 20334 (JSC::CodeBlock::CodeBlock): 20335 * bytecode/CodeBlock.h: 20336 (JSC::CodeBlock::ownerExecutable): 20337 (JSC::GlobalCodeBlock::GlobalCodeBlock): 20338 * bytecode/EvalCodeCache.h: 20339 (JSC::EvalCodeCache::get): 20340 * debugger/Debugger.cpp: 20341 (JSC::Debugger::recompileAllJSFunctions): 20342 * interpreter/CachedCall.h: 20343 (JSC::CachedCall::CachedCall): 20344 * interpreter/Interpreter.cpp: 20345 (JSC::Interpreter::callEval): 20346 (JSC::Interpreter::privateExecute): 20347 * jit/JITStubs.cpp: 20348 (JSC::DEFINE_STUB_FUNCTION): 20349 * profiler/Profiler.cpp: 20350 (JSC::createCallIdentifierFromFunctionImp): 20351 * runtime/Arguments.h: 20352 (JSC::Arguments::getArgumentsData): 20353 (JSC::Arguments::Arguments): 20354 * runtime/Executable.cpp: 20355 (JSC::NativeExecutable::~NativeExecutable): 20356 (JSC::VPtrHackExecutable::~VPtrHackExecutable): 20357 * runtime/Executable.h: 20358 (JSC::ExecutableBase::ExecutableBase): 20359 (JSC::ExecutableBase::~ExecutableBase): 20360 (JSC::ExecutableBase::isHostFunction): 20361 (JSC::NativeExecutable::NativeExecutable): 20362 (JSC::VPtrHackExecutable::VPtrHackExecutable): 20363 (JSC::ScriptExecutable::ScriptExecutable): 20364 (JSC::ScriptExecutable::source): 20365 (JSC::ScriptExecutable::sourceID): 20366 (JSC::ScriptExecutable::sourceURL): 20367 (JSC::ScriptExecutable::lineNo): 20368 (JSC::ScriptExecutable::lastLine): 20369 (JSC::ScriptExecutable::usesEval): 20370 (JSC::ScriptExecutable::usesArguments): 20371 (JSC::ScriptExecutable::needsActivation): 20372 (JSC::EvalExecutable::EvalExecutable): 20373 (JSC::EvalExecutable::create): 20374 (JSC::ProgramExecutable::ProgramExecutable): 20375 (JSC::FunctionExecutable::FunctionExecutable): 20376 * runtime/FunctionPrototype.cpp: 20377 (JSC::functionProtoFuncToString): 20378 * runtime/JSFunction.cpp: 20379 (JSC::JSFunction::JSFunction): 20380 (JSC::JSFunction::~JSFunction): 20381 (JSC::JSFunction::markChildren): 20382 (JSC::JSFunction::getCallData): 20383 (JSC::JSFunction::call): 20384 (JSC::JSFunction::lengthGetter): 20385 (JSC::JSFunction::getConstructData): 20386 (JSC::JSFunction::construct): 20387 * runtime/JSFunction.h: 20388 (JSC::JSFunction::executable): 20389 (JSC::JSFunction::jsExecutable): 20390 (JSC::JSFunction::isHostFunction): 20391 203922009-08-20 Oliver Hunt <oliver@apple.com> 20393 20394 Reviewed by Maciej Stachowiak. 20395 20396 Browser hangs on opening Web Inspector. 20397 https://bugs.webkit.org/show_bug.cgi?id=28438 20398 20399 Code generation needs to be able to walk the entire scopechain in some 20400 cases, however the symbol table used by activations was a member of the 20401 codeblock. Following recompilation this may no longer exist, leading 20402 to a crash or hang on lookup. 20403 20404 We fix this by introducing a refcounted SymbolTable subclass, SharedSymbolTable, 20405 for the CodeBlocks used by function code. This allows activations to 20406 maintain ownership of a copy of the symbol table even after recompilation so 20407 they can continue to work. 20408 20409 * bytecode/CodeBlock.cpp: 20410 (JSC::CodeBlock::CodeBlock): 20411 * bytecode/CodeBlock.h: 20412 (JSC::CodeBlock::symbolTable): 20413 (JSC::CodeBlock::sharedSymbolTable): 20414 (JSC::GlobalCodeBlock::GlobalCodeBlock): 20415 (JSC::FunctionCodeBlock::FunctionCodeBlock): 20416 (JSC::FunctionCodeBlock::~FunctionCodeBlock): 20417 * interpreter/Interpreter.cpp: 20418 (JSC::Interpreter::retrieveArguments): 20419 * runtime/Executable.cpp: 20420 (JSC::EvalExecutable::generateBytecode): 20421 (JSC::FunctionExecutable::generateBytecode): 20422 (JSC::FunctionExecutable::reparseExceptionInfo): 20423 (JSC::EvalExecutable::reparseExceptionInfo): 20424 * runtime/JSActivation.h: 20425 (JSC::JSActivation::JSActivationData::JSActivationData): 20426 (JSC::JSActivation::JSActivationData::~JSActivationData): 20427 * runtime/SymbolTable.h: 20428 204292009-08-20 Xan Lopez <xlopez@igalia.com> 20430 20431 Add new file to GTK+ build. 20432 20433 * GNUmakefile.am: 20434 204352009-08-20 Geoffrey Garen <ggaren@apple.com> 20436 20437 Reviewed by Maciej Stachowiak. 20438 20439 Added a number => string cache. 20440 20441 1.07x faster on v8 (1.7x faster on v8-splay). 20442 1.004x faster on SunSpider. 20443 20444 * runtime/JSCell.h: Moved JSValue::toString to JSString.h. 20445 * runtime/JSGlobalData.h: Holds the cache. 20446 * runtime/JSNumberCell.cpp: 20447 (JSC::JSNumberCell::toString): 20448 (JSC::JSNumberCell::toThisString): Removed -0 special case. 20449 UString handles this now, since too many clients were 20450 special-casing it. 20451 20452 * runtime/JSString.h: 20453 (JSC::JSValue::toString): Use the cache when converting 20454 an int or double to string. 20455 20456 * runtime/Operations.h: 20457 (JSC::concatenateStrings): Call toString to take advantage 20458 of the cache. 20459 20460 * runtime/SmallStrings.h: 20461 (JSC::NumericStrings::add): 20462 (JSC::NumericStrings::lookup): The cache. 20463 20464 * runtime/UString.cpp: 20465 (JSC::UString::from): Added -0 special case mentioned above. 20466 Removed appendNumeric because it's mutually exclusive with the 20467 cache. 20468 204692009-08-20 Oliver Hunt <oliver@apple.com> 20470 20471 Reviewed by Gavin Barraclough. 20472 20473 REGRESSION: fast/profiler/call.html is crashing occasionally 20474 https://bugs.webkit.org/show_bug.cgi?id=28476 20475 20476 Using the codeblock for information about how many parameters and 20477 locals a function has is unsafe in certain circumstances. The 20478 basic scenario is all function code being cleared in response to 20479 the debugger or profiler being enabled, and then an activation is 20480 marked before its associated function is re-executed. 20481 20482 To deal with this scenario we store the variable count of a function 20483 directly in the FunctionExecutable, and then use that information. 20484 20485 * runtime/Arguments.h: 20486 (JSC::Arguments::getArgumentsData): 20487 * runtime/Executable.cpp: 20488 (JSC::FunctionExecutable::generateBytecode): 20489 * runtime/Executable.h: 20490 (JSC::FunctionExecutable::FunctionExecutable): 20491 (JSC::FunctionExecutable::variableCount): 20492 * runtime/JSActivation.cpp: 20493 (JSC::JSActivation::markChildren): 20494 204952009-08-20 Gavin Barraclough <barraclough@apple.com> 20496 20497 Reviewed by Oliver Hunt. 20498 20499 Numbering of arguments to emitGetJITStubArg/emitPutJITStubArg incorrect 20500 <bug lost in the great bug disasteroony of 08/20/09!> 20501 20502 The argumentNumber argument to emitGetJITStubArg/emitPutJITStubArg should match 20503 the argument number used within the stub functions in JITStubs.cpp, but it doesn't. 20504 20505 Firstly, all the numbers changed when we added a void* 'reserved' as the first slot 20506 (rather than leaving argument 0 unused), and secondly in 32_64 builds the index to 20507 peek/poke needs to be multiplies by 2 (since the argument to peek/poke is a number 20508 of machine words, and on 32_64 build the argument slots to stub functions are two 20509 words wide). 20510 20511 * jit/JIT.h: 20512 * jit/JITCall.cpp: 20513 (JSC::JIT::compileOpCallSetupArgs): 20514 (JSC::JIT::compileOpConstructSetupArgs): 20515 (JSC::JIT::compileOpCallVarargsSetupArgs): 20516 (JSC::JIT::compileOpCall): 20517 * jit/JITInlineMethods.h: 20518 (JSC::JIT::emitPutJITStubArg): 20519 (JSC::JIT::emitPutJITStubArgConstant): 20520 (JSC::JIT::emitGetJITStubArg): 20521 (JSC::JIT::emitPutJITStubArgFromVirtualRegister): 20522 * jit/JITOpcodes.cpp: 20523 (JSC::JIT::privateCompileCTIMachineTrampolines): 20524 * jit/JITPropertyAccess.cpp: 20525 (JSC::JIT::privateCompilePutByIdTransition): 20526 205272009-08-20 Oliver Hunt <oliver@apple.com> 20528 20529 Reviewed by Geoff Garen. 20530 20531 REGRESSION: significant slowdown on Celtic Kane "AJAX declaration" subtest 20532 https://bugs.webkit.org/show_bug.cgi?id=28332 20533 20534 Follow up style fixes that were missed in review. 20535 20536 * runtime/Structure.cpp: 20537 (JSC::Structure::hasTransition): 20538 * runtime/Structure.h: 20539 (JSC::Structure::get): 20540 (JSC::StructureTransitionTable::contains): 20541 * runtime/StructureTransitionTable.h: 20542 (JSC::StructureTransitionTable::add): 20543 205442009-08-20 Oliver Hunt <oliver@apple.com> 20545 20546 Add new exports to windows jsc build 20547 20548 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 20549 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 20550 205512009-08-20 Oliver Hunt <oliver@apple.com> 20552 20553 Reviewed by Gavin Barraclough. 20554 20555 REGRESSION: significant slowdown on Celtic Kane "AJAX declaration" subtest 20556 https://bugs.webkit.org/show_bug.cgi?id=28332 20557 20558 The method check optimisation made transitions aware of the value being 20559 assigned when a transition was assigning a function. This had the side 20560 effect of making every assignment of a function expression result in a 20561 new transition, and thus a new Structure. The net result of this is that 20562 the common JS idiom of 20563 20564 function MyObject() { 20565 this.myFunction = function(...){...}; 20566 } 20567 new MyObject(); 20568 20569 Will produce a unique structure on every iteration, meaning that all 20570 caching is defeated and there is a significant amount of structure churn. 20571 20572 The fix is to return the transition to its original form where it is 20573 keyed off a property name + attributes tuple, but have each transition 20574 support an optional transition on a specific value. 20575 20576 * JavaScriptCore.exp: 20577 * runtime/JSObject.h: 20578 (JSC::JSObject::putDirectInternal): 20579 * runtime/Structure.cpp: 20580 (JSC::Structure::~Structure): 20581 (JSC::Structure::addPropertyTransitionToExistingStructure): 20582 (JSC::Structure::addPropertyTransition): 20583 (JSC::Structure::hasTransition): 20584 * runtime/Structure.h: 20585 (JSC::Structure::transitionedFor): 20586 (JSC::Structure::hasTransition): 20587 (JSC::Structure::): 20588 (JSC::StructureTransitionTable::contains): 20589 (JSC::StructureTransitionTable::get): 20590 * runtime/StructureTransitionTable.h: 20591 (JSC::StructureTransitionTableHashTraits::emptyValue): 20592 (JSC::StructureTransitionTable::hasTransition): 20593 (JSC::StructureTransitionTable::remove): 20594 (JSC::StructureTransitionTable::add): 20595 205962009-08-20 Gavin Barraclough <barraclough@apple.com> 20597 20598 Reviewed by Oliver Hunt. 20599 20600 Remove FunctionCodeBlock. 20601 https://bugs.webkit.org/show_bug.cgi?id=28502 20602 20603 These only exist to allow JIT code to dereference properties off the 20604 CodeBlock for any callee, regardless of whether it is a host function. 20605 20606 Instead just use the FunctionExecutable. Copy the m_parameters field 20607 from the CodeBlock into the Executable, and use this to distinguish 20608 between host functions, functions that have been bytecompiled, and 20609 functions that have not. 20610 20611 m_parameters is moved to ExecutableBase rather than FunctionExecutable 20612 so that (as a separate change) we can move make a separate class of 20613 executable for host code, which is not devived from FunctionExecutable 20614 (host code does not feature any of the properties that normal executable 20615 do and will provide, such as source, attributes, and a parsed name). 20616 20617 1% win on v8 tests, 0.5% on sunspider. 20618 20619 * bytecode/CodeBlock.cpp: 20620 (JSC::CodeBlock::derefStructures): 20621 (JSC::CodeBlock::refStructures): 20622 (JSC::CodeBlock::reparseForExceptionInfoIfNecessary): 20623 (JSC::CodeBlock::handlerForBytecodeOffset): 20624 (JSC::CodeBlock::lineNumberForBytecodeOffset): 20625 (JSC::CodeBlock::expressionRangeForBytecodeOffset): 20626 (JSC::CodeBlock::getByIdExceptionInfoForBytecodeOffset): 20627 (JSC::CodeBlock::functionRegisterForBytecodeOffset): 20628 (JSC::CodeBlock::hasGlobalResolveInstructionAtBytecodeOffset): 20629 (JSC::CodeBlock::hasGlobalResolveInfoAtBytecodeOffset): 20630 * bytecode/CodeBlock.h: 20631 (JSC::): 20632 (JSC::CodeBlock::source): 20633 (JSC::CodeBlock::sourceOffset): 20634 (JSC::CodeBlock::evalCodeCache): 20635 (JSC::CodeBlock::createRareDataIfNecessary): 20636 20637 remove NativeCodeBlocks and the NativeCode code type. 20638 20639 * jit/JIT.cpp: 20640 (JSC::JIT::linkCall): 20641 20642 Revert to previous behaviour (as currently still commented!) that Hhost functions have a null codeblock. 20643 20644 * jit/JITCall.cpp: 20645 (JSC::JIT::compileOpCallInitializeCallFrame): 20646 (JSC::JIT::compileOpCallSetupArgs): 20647 (JSC::JIT::compileOpCallVarargsSetupArgs): 20648 (JSC::JIT::compileOpConstructSetupArgs): 20649 (JSC::JIT::compileOpCallVarargs): 20650 (JSC::JIT::compileOpCall): 20651 (JSC::JIT::compileOpCallSlowCase): 20652 20653 Bring the 32_64 & non-32_64 JITs into line with each other, callee in regT0. 20654 20655 * jit/JITOpcodes.cpp: 20656 (JSC::JIT::privateCompileCTIMachineTrampolines): 20657 20658 Rewrite call trampolines to not use the CodeBlock. 20659 20660 * jit/JITStubs.cpp: 20661 (JSC::DEFINE_STUB_FUNCTION): 20662 20663 Make call_JSFunction & call_arityCheck return the callee, don't expect to be passed the CodeBlock. 20664 20665 * runtime/Executable.cpp: 20666 (JSC::FunctionExecutable::generateBytecode): 20667 (JSC::FunctionExecutable::recompile): 20668 (JSC::FunctionExecutable::FunctionExecutable): 20669 * runtime/Executable.h: 20670 (JSC::ExecutableBase::): 20671 (JSC::ExecutableBase::ExecutableBase): 20672 (JSC::FunctionExecutable::isHostFunction): 20673 20674 Add m_numParameters. 20675 20676 * runtime/JSFunction.cpp: 20677 (JSC::JSFunction::~JSFunction): 20678 20679 Only call generatedBytecode() on JSFunctions non-host FunctionExecutables. 20680 206812009-08-20 Yongjun Zhang <yongjun.zhang@nokia.com> 20682 20683 Reviewed by Eric Seidel. 20684 20685 https://bugs.webkit.org/show_bug.cgi?id=28054 20686 20687 Use a helper function to work around winscw compiler forward declaration bug 20688 regarding templated classes. 20689 20690 Add parenthesis around (PassRefPtr::*UnspecifiedBoolType) to make winscw compiler 20691 work with the default UnSpecifiedBoolType() operator, which removes the winscw 20692 specific bool cast hack. 20693 20694 * wtf/PassRefPtr.h: 20695 (WTF::derefIfNotNull): 20696 (WTF::PassRefPtr::~PassRefPtr): 20697 206982009-08-19 Yong Li <yong.li@torchmobile.com> 20699 20700 Reviewed by Gavin Barraclough. 20701 20702 Change namespace ARM to ARMRegisters 20703 X86 to X86Registers to avoid conflict with macros 20704 https://bugs.webkit.org/show_bug.cgi?id=28428 20705 20706 * assembler/ARMAssembler.cpp: 20707 * assembler/ARMAssembler.h: 20708 * assembler/ARMv7Assembler.h: 20709 * assembler/MacroAssemblerARM.h: 20710 * assembler/MacroAssemblerARMv7.h: 20711 * assembler/MacroAssemblerX86Common.h: 20712 * assembler/MacroAssemblerX86_64.h: 20713 * assembler/X86Assembler.h: 20714 * jit/JIT.h: 20715 * jit/JITArithmetic.cpp: 20716 * jit/JITInlineMethods.h: 20717 * jit/JITOpcodes.cpp: 20718 * wrec/WRECGenerator.cpp: 20719 * wrec/WRECGenerator.h: 20720 * yarr/RegexJIT.cpp: 20721 207222009-08-19 Oliver Hunt <oliver@apple.com> 20723 20724 Reviewed by Gavin Barraclough. 20725 20726 Devirtualise marking 20727 https://bugs.webkit.org/show_bug.cgi?id=28294 20728 20729 We actually need to mark the value in a number object if we're using the 20730 32bit number representation. 20731 20732 * runtime/NumberObject.h: 20733 (JSC::NumberObject::createStructure): 20734 207352009-08-19 Gavin Barraclough <barraclough@apple.com> 20736 20737 Reviewed by Darin Adler. 20738 20739 We probably shouldn't be keeping the AST for eval nodes around forevar. 20740 https://bugs.webkit.org/show_bug.cgi?id=28469 20741 20742 EvalNodes don't destroyData() (delete their parser data) since they need to hold onto 20743 their varStack. Copy a list of variable onto EvalCodeBlock, and this can go away. 20744 20745 * bytecode/CodeBlock.h: 20746 (JSC::EvalCodeBlock::variable): 20747 (JSC::EvalCodeBlock::numVariables): 20748 (JSC::EvalCodeBlock::adoptVariables): 20749 * bytecompiler/BytecodeGenerator.cpp: 20750 (JSC::BytecodeGenerator::BytecodeGenerator): 20751 * interpreter/Interpreter.cpp: 20752 (JSC::Interpreter::execute): 20753 * parser/Nodes.h: 20754 * runtime/Executable.cpp: 20755 (JSC::EvalExecutable::generateBytecode): 20756 * runtime/Executable.h: 20757 207582009-08-19 Jungshik Shin <jshin@chromium.org> 20759 20760 Reviewed by Darin Adler. 20761 20762 http://bugs.webkit.org/show_bug.cgi?id=28441 20763 20764 Fix a build issue with ICU 4.2 or later on Windows with Visual C++. 20765 Instead of defining all isXXX and toupper/tolower as 20766 WTF_Please_use_ASCIICType_instead_of_ctype_see_comment_in_ASCIICType_h, 20767 #define them to be different by prepending 'WTF_...ASCIIType_h' with 20768 the originial names like 'toupper_WTF_...ASCIIType_h'. 20769 20770 * wtf/DisallowCType.h: 20771 207722009-08-18 Oliver Hunt <oliver@apple.com> 20773 20774 Reviewed by Gavin Barraclough. 20775 20776 Assigning a function to an object should always use the existing transition, even if the transition is not specialized 20777 https://bugs.webkit.org/show_bug.cgi?id=28442 20778 20779 Check for an unspecialized transition as an alternative to always failing if specialisation does not match. 20780 20781 * runtime/Structure.cpp: 20782 (JSC::Structure::addPropertyTransitionToExistingStructure): 20783 207842009-08-18 Dirk Schulze <krit@webkit.org> 20785 20786 Reviewed by Oliver Hunt. 20787 20788 Added additional getter to ByteArray with an unsigned char as return. 20789 ByteArray can take unsigned char directly now. 20790 20791 * wtf/ByteArray.h: 20792 (WTF::ByteArray::set): 20793 (WTF::ByteArray::get): 20794 207952009-08-18 Peter Kasting <pkasting@google.com> 20796 20797 Reviewed by Eric Seidel. 20798 20799 https://bugs.webkit.org/show_bug.cgi?id=28415 20800 Set svn:eol-style CRLF on all .sln and .vcproj files that don't already 20801 have it. 20802 20803 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCoreGenerated.vcproj: 20804 * JavaScriptCore.vcproj/testapi/testapi.vcproj: 20805 208062009-08-18 Xan Lopez <xlopez@igalia.com> 20807 20808 Try to fix the GTK+ build. 20809 20810 * GNUmakefile.am: 20811 208122009-08-17 Gavin Barraclough <barraclough@apple.com> 20813 20814 Reviewed by Sam Weinig. 20815 20816 No, silly runtime, AST nodes are not for you. 20817 20818 We still use AST nodes (ScopeNodes, particularly FunctionBodyNodes) within 20819 the runtime, which means that these nodes must be persisted outside of the 20820 arena, contain both parser & runtime data, etc. This is all a bit of a mess. 20821 20822 Move functionality into a new FunctionExecutable class. 20823 20824 * API/JSCallbackFunction.cpp: 20825 * API/JSObjectRef.cpp: 20826 * JavaScriptCore.exp: 20827 * JavaScriptCore.xcodeproj/project.pbxproj: 20828 * bytecode/CodeBlock.cpp: 20829 (JSC::CodeBlock::CodeBlock): 20830 (JSC::CodeBlock::markAggregate): 20831 (JSC::CodeBlock::reparseForExceptionInfoIfNecessary): 20832 (JSC::CodeBlock::lineNumberForBytecodeOffset): 20833 (JSC::CodeBlock::shrinkToFit): 20834 * bytecode/CodeBlock.h: 20835 (JSC::CodeBlock::getBytecodeIndex): 20836 (JSC::CodeBlock::discardBytecode): 20837 (JSC::CodeBlock::instructionCount): 20838 (JSC::CodeBlock::getJITCode): 20839 (JSC::CodeBlock::executablePool): 20840 (JSC::CodeBlock::ownerExecutable): 20841 (JSC::CodeBlock::extractExceptionInfo): 20842 (JSC::CodeBlock::addFunctionDecl): 20843 (JSC::CodeBlock::functionDecl): 20844 (JSC::CodeBlock::numberOfFunctionDecls): 20845 (JSC::CodeBlock::addFunctionExpr): 20846 (JSC::CodeBlock::functionExpr): 20847 (JSC::GlobalCodeBlock::GlobalCodeBlock): 20848 (JSC::ProgramCodeBlock::ProgramCodeBlock): 20849 (JSC::EvalCodeBlock::EvalCodeBlock): 20850 (JSC::FunctionCodeBlock::FunctionCodeBlock): 20851 (JSC::NativeCodeBlock::NativeCodeBlock): 20852 * bytecode/EvalCodeCache.h: 20853 * bytecode/SamplingTool.cpp: 20854 (JSC::SamplingTool::doRun): 20855 * bytecompiler/BytecodeGenerator.cpp: 20856 (JSC::BytecodeGenerator::BytecodeGenerator): 20857 (JSC::BytecodeGenerator::emitNewFunction): 20858 (JSC::BytecodeGenerator::emitNewFunctionExpression): 20859 * bytecompiler/BytecodeGenerator.h: 20860 * debugger/Debugger.cpp: 20861 (JSC::Debugger::recompileAllJSFunctions): 20862 * interpreter/CachedCall.h: 20863 (JSC::CachedCall::CachedCall): 20864 * interpreter/CallFrameClosure.h: 20865 * interpreter/Interpreter.cpp: 20866 (JSC::Interpreter::unwindCallFrame): 20867 (JSC::Interpreter::throwException): 20868 (JSC::Interpreter::execute): 20869 (JSC::Interpreter::prepareForRepeatCall): 20870 (JSC::Interpreter::debug): 20871 (JSC::Interpreter::privateExecute): 20872 (JSC::Interpreter::retrieveLastCaller): 20873 * interpreter/Interpreter.h: 20874 * jit/JIT.cpp: 20875 (JSC::JIT::privateCompile): 20876 * jit/JIT.h: 20877 (JSC::JIT::compile): 20878 * jit/JITOpcodes.cpp: 20879 (JSC::JIT::privateCompileCTIMachineTrampolines): 20880 (JSC::JIT::emit_op_new_func): 20881 (JSC::JIT::emit_op_new_func_exp): 20882 * jit/JITStubs.cpp: 20883 (JSC::DEFINE_STUB_FUNCTION): 20884 * jit/JITStubs.h: 20885 (JSC::): 20886 * parser/Nodes.cpp: 20887 (JSC::FunctionBodyNode::reparseDataIfNecessary): 20888 * parser/Nodes.h: 20889 (JSC::EvalNode::partialDestroyData): 20890 * parser/Parser.h: 20891 * profiler/ProfileGenerator.cpp: 20892 * profiler/Profiler.cpp: 20893 (JSC::Profiler::createCallIdentifier): 20894 (JSC::createCallIdentifierFromFunctionImp): 20895 * runtime/Arguments.h: 20896 (JSC::Arguments::getArgumentsData): 20897 (JSC::Arguments::Arguments): 20898 (JSC::JSActivation::copyRegisters): 20899 * runtime/ArrayPrototype.cpp: 20900 (JSC::isNumericCompareFunction): 20901 * runtime/CallData.h: 20902 (JSC::): 20903 * runtime/Collector.cpp: 20904 (JSC::Heap::collect): 20905 * runtime/ConstructData.h: 20906 (JSC::): 20907 * runtime/ExceptionHelpers.cpp: 20908 (JSC::createUndefinedVariableError): 20909 (JSC::createInvalidParamError): 20910 (JSC::createNotAConstructorError): 20911 (JSC::createNotAFunctionError): 20912 (JSC::createNotAnObjectError): 20913 * runtime/Executable.cpp: Added. 20914 (JSC::EvalExecutable::generateBytecode): 20915 (JSC::ProgramExecutable::generateBytecode): 20916 (JSC::FunctionExecutable::generateBytecode): 20917 (JSC::EvalExecutable::generateJITCode): 20918 (JSC::ProgramExecutable::generateJITCode): 20919 (JSC::FunctionExecutable::generateJITCode): 20920 (JSC::FunctionExecutable::isHostFunction): 20921 (JSC::FunctionExecutable::markAggregate): 20922 (JSC::FunctionExecutable::reparseExceptionInfo): 20923 (JSC::EvalExecutable::reparseExceptionInfo): 20924 (JSC::FunctionExecutable::recompile): 20925 (JSC::FunctionExecutable::FunctionExecutable): 20926 * runtime/Executable.h: 20927 (JSC::ExecutableBase::~ExecutableBase): 20928 (JSC::ExecutableBase::ExecutableBase): 20929 (JSC::ExecutableBase::source): 20930 (JSC::ExecutableBase::sourceID): 20931 (JSC::ExecutableBase::lastLine): 20932 (JSC::ExecutableBase::usesEval): 20933 (JSC::ExecutableBase::usesArguments): 20934 (JSC::ExecutableBase::needsActivation): 20935 (JSC::ExecutableBase::astNode): 20936 (JSC::ExecutableBase::generatedJITCode): 20937 (JSC::ExecutableBase::getExecutablePool): 20938 (JSC::EvalExecutable::EvalExecutable): 20939 (JSC::EvalExecutable::bytecode): 20940 (JSC::EvalExecutable::varStack): 20941 (JSC::EvalExecutable::evalNode): 20942 (JSC::EvalExecutable::jitCode): 20943 (JSC::ProgramExecutable::ProgramExecutable): 20944 (JSC::ProgramExecutable::reparseExceptionInfo): 20945 (JSC::ProgramExecutable::bytecode): 20946 (JSC::ProgramExecutable::programNode): 20947 (JSC::ProgramExecutable::jitCode): 20948 (JSC::FunctionExecutable::FunctionExecutable): 20949 (JSC::FunctionExecutable::name): 20950 (JSC::FunctionExecutable::bytecode): 20951 (JSC::FunctionExecutable::generatedBytecode): 20952 (JSC::FunctionExecutable::usesEval): 20953 (JSC::FunctionExecutable::usesArguments): 20954 (JSC::FunctionExecutable::parameterCount): 20955 (JSC::FunctionExecutable::paramString): 20956 (JSC::FunctionExecutable::isGenerated): 20957 (JSC::FunctionExecutable::body): 20958 (JSC::FunctionExecutable::jitCode): 20959 (JSC::FunctionExecutable::createNativeThunk): 20960 * runtime/FunctionConstructor.cpp: 20961 (JSC::constructFunction): 20962 * runtime/FunctionPrototype.cpp: 20963 (JSC::functionProtoFuncToString): 20964 * runtime/JSActivation.cpp: 20965 (JSC::JSActivation::JSActivation): 20966 (JSC::JSActivation::markChildren): 20967 (JSC::JSActivation::isDynamicScope): 20968 (JSC::JSActivation::argumentsGetter): 20969 * runtime/JSActivation.h: 20970 (JSC::JSActivation::JSActivationData::JSActivationData): 20971 * runtime/JSFunction.cpp: 20972 (JSC::JSFunction::isHostFunction): 20973 (JSC::JSFunction::JSFunction): 20974 (JSC::JSFunction::~JSFunction): 20975 (JSC::JSFunction::markChildren): 20976 (JSC::JSFunction::getCallData): 20977 (JSC::JSFunction::call): 20978 (JSC::JSFunction::lengthGetter): 20979 (JSC::JSFunction::getConstructData): 20980 (JSC::JSFunction::construct): 20981 * runtime/JSFunction.h: 20982 (JSC::JSFunction::executable): 20983 (JSC::FunctionExecutable::make): 20984 * runtime/JSGlobalData.cpp: 20985 (JSC::JSGlobalData::JSGlobalData): 20986 (JSC::JSGlobalData::numericCompareFunction): 20987 * runtime/JSGlobalData.h: 20988 209892009-08-17 Mark Rowe <mrowe@apple.com> 20990 20991 Reviewed by Darin Adler. 20992 20993 Fix 300,000+ leaks seen during the regression tests. 20994 20995 EvalCodeCache::get was heap-allocating an EvalExecutable instance without adopting the initial reference. 20996 While fixing this we noticed that EvalExecutable was a RefCounted type that was sometimes stack allocated. 20997 To make this cleaner and to prevent clients from attempting to ref a stack-allocated instance, we move the 20998 refcounting down to a new CacheableEvalExecutable class that derives from EvalExecutable. EvalCodeCache::get 20999 now uses CacheableEvalExecutable::create and avoids the leak. 21000 21001 * bytecode/EvalCodeCache.h: 21002 (JSC::EvalCodeCache::get): 21003 * interpreter/Interpreter.cpp: 21004 (JSC::Interpreter::callEval): 21005 * runtime/Executable.h: 21006 (JSC::CacheableEvalExecutable::create): 21007 (JSC::CacheableEvalExecutable::CacheableEvalExecutable): 21008 210092009-08-17 Oliver Hunt <oliver@apple.com> 21010 21011 RS=Mark Rowe. 21012 21013 REGRESSION (r47292): Prototype.js is broken by ES5 Arguments changes 21014 https://bugs.webkit.org/show_bug.cgi?id=28341 21015 <rdar://problem/7145615> 21016 21017 Reverting r47292. Alas Prototype.js breaks with Arguments inheriting 21018 from Array as ES5 attempted. Prototype.js defines $A in terms of a 21019 function it places on (among other global objects) the Array prototype, 21020 thus breaking $A for arrays. 21021 21022 * runtime/Arguments.h: 21023 (JSC::Arguments::Arguments): 21024 * runtime/JSGlobalObject.cpp: 21025 (JSC::JSGlobalObject::reset): 21026 (JSC::JSGlobalObject::markChildren): 21027 * runtime/JSGlobalObject.h: 21028 (JSC::JSGlobalObject::JSGlobalObjectData::JSGlobalObjectData): 21029 * runtime/ObjectPrototype.cpp: 21030 (JSC::ObjectPrototype::ObjectPrototype): 21031 * runtime/ObjectPrototype.h: 21032 * tests/mozilla/ecma_3/Function/arguments-001.js: 21033 210342009-08-17 Peter Kasting <pkasting@google.com> 21035 21036 Reviewed by Steve Falkenburg. 21037 21038 https://bugs.webkit.org/show_bug.cgi?id=27323 21039 Only add Cygwin to the path when it isn't already there. This avoids 21040 causing problems for people who purposefully have non-Cygwin versions of 21041 executables like svn in front of the Cygwin ones in their paths. 21042 21043 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCoreCommon.vsprops: 21044 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCoreGenerated.vcproj: 21045 * JavaScriptCore.vcproj/WTF/WTFCommon.vsprops: 21046 * JavaScriptCore.vcproj/jsc/jscCommon.vsprops: 21047 * JavaScriptCore.vcproj/testapi/testapiCommon.vsprops: 21048 210492009-08-17 Xan Lopez <xlopez@igalia.com> 21050 21051 Reviewed by Mark Rowe. 21052 21053 Fix build with FAST_MALLOC_MATCH_VALIDATION enabled. 21054 21055 * wtf/FastMalloc.cpp: 21056 (WTF::fastMalloc): 21057 (WTF::fastCalloc): 21058 (WTF::fastRealloc): 21059 210602009-08-16 Holger Hans Peter Freyther <zecke@selfish.org> 21061 21062 Reviewed by Mark Rowe. 21063 21064 Fix crash on ./ecma_2/RegExp/exec-002.js. 21065 https://bugs.webkit.org/show_bug.cgi?id=28353 21066 21067 Change the order of freeParenthesesDisjunctionContext and 21068 popParenthesesDisjunctionContext on all call sites as the pop 21069 method is accessing backTrack->lastContext which is the context 21070 that is about to be freed. 21071 21072 * yarr/RegexInterpreter.cpp: 21073 (JSC::Yarr::Interpreter::parenthesesDoBacktrack): 21074 (JSC::Yarr::Interpreter::backtrackParentheses): 21075 210762009-08-16 Holger Hans Peter Freyther <zecke@selfish.org> 21077 21078 Reviewed by Mark Rowe. 21079 21080 https://bugs.webkit.org/show_bug.cgi?id=28352 21081 21082 Fix coding style violations. Use m_ for C++ class members. Remove 21083 trailing whitespace on empty lines. 21084 21085 * yarr/RegexInterpreter.cpp: 21086 (JSC::Yarr::Interpreter::ParenthesesDisjunctionContext::ParenthesesDisjunctionContext): 21087 (JSC::Yarr::Interpreter::tryConsumeCharacter): 21088 (JSC::Yarr::Interpreter::tryConsumeBackReference): 21089 (JSC::Yarr::Interpreter::parenthesesDoBacktrack): 21090 (JSC::Yarr::Interpreter::backtrackParentheses): 21091 (JSC::Yarr::ByteCompiler::ByteCompiler): 21092 (JSC::Yarr::ByteCompiler::compile): 21093 (JSC::Yarr::ByteCompiler::checkInput): 21094 (JSC::Yarr::ByteCompiler::assertionBOL): 21095 (JSC::Yarr::ByteCompiler::assertionEOL): 21096 (JSC::Yarr::ByteCompiler::assertionWordBoundary): 21097 (JSC::Yarr::ByteCompiler::atomPatternCharacter): 21098 (JSC::Yarr::ByteCompiler::atomCharacterClass): 21099 (JSC::Yarr::ByteCompiler::atomBackReference): 21100 (JSC::Yarr::ByteCompiler::atomParenthesesSubpatternBegin): 21101 (JSC::Yarr::ByteCompiler::atomParentheticalAssertionBegin): 21102 (JSC::Yarr::ByteCompiler::popParenthesesStack): 21103 (JSC::Yarr::ByteCompiler::closeAlternative): 21104 (JSC::Yarr::ByteCompiler::closeBodyAlternative): 21105 (JSC::Yarr::ByteCompiler::atomParenthesesEnd): 21106 (JSC::Yarr::ByteCompiler::regexBegin): 21107 (JSC::Yarr::ByteCompiler::alterantiveBodyDisjunction): 21108 (JSC::Yarr::ByteCompiler::alterantiveDisjunction): 21109 (JSC::Yarr::ByteCompiler::emitDisjunction): 21110 211112009-08-15 Mark Rowe <mrowe@apple.com> 21112 21113 Fix the build with JIT disabled. 21114 21115 * runtime/Arguments.h: Only compile the jitCode method when the JIT is enabled. 21116 * runtime/Executable.h: Include PrototypeFunction.h so the compiler knows what 21117 NativeFunctionWrapper is when the JIT is disabled. 21118 211192009-08-15 Adam Bergkvist <adam.bergkvist@ericsson.com> 21120 21121 Reviewed by Sam Weinig. 21122 21123 Added ENABLE_EVENTSOURCE flag. 21124 https://bugs.webkit.org/show_bug.cgi?id=14997 21125 21126 * Configurations/FeatureDefines.xcconfig: 21127 211282009-08-14 Gavin Barraclough <barraclough@apple.com> 21129 21130 * parser/Parser.h: 21131 (JSC::EvalExecutable::parse): 21132 (JSC::ProgramExecutable::parse): 21133 * runtime/Executable.h: 21134 211352009-08-14 Gavin Barraclough <barraclough@apple.com> 21136 21137 Reviewed by Oliver Hunt. 21138 21139 Remove AST nodes from use within the Runtime (outside of parsing), stage 1 21140 https://bugs.webkit.org/show_bug.cgi?id=28330 21141 21142 Remove the EvalNode and ProgramNode from use in the runtime. They still exist 21143 after this patch, but are hidden behind EvalExecutable and FunctionExecutable, 21144 and are also still reachable behind CodeBlock::m_ownerNode. 21145 21146 The next step will be to beat back FunctionBodyNode in the same fashion. 21147 Then remove the usage via CodeBlock, then only construct these nodes only on 21148 demand during bytecode generation. 21149 21150 * JavaScriptCore.xcodeproj/project.pbxproj: 21151 * bytecode/CodeBlock.h: 21152 (JSC::GlobalCodeBlock::GlobalCodeBlock): 21153 (JSC::GlobalCodeBlock::~GlobalCodeBlock): 21154 (JSC::ProgramCodeBlock::ProgramCodeBlock): 21155 (JSC::EvalCodeBlock::EvalCodeBlock): 21156 (JSC::FunctionCodeBlock::FunctionCodeBlock): 21157 (JSC::NativeCodeBlock::NativeCodeBlock): 21158 * bytecode/EvalCodeCache.h: 21159 (JSC::EvalCodeCache::get): 21160 * debugger/Debugger.cpp: 21161 (JSC::evaluateInGlobalCallFrame): 21162 * debugger/DebuggerCallFrame.cpp: 21163 (JSC::DebuggerCallFrame::evaluate): 21164 * interpreter/Interpreter.cpp: 21165 (JSC::Interpreter::callEval): 21166 (JSC::Interpreter::execute): 21167 * interpreter/Interpreter.h: 21168 * parser/Nodes.cpp: 21169 (JSC::FunctionBodyNode::createNativeThunk): 21170 (JSC::FunctionBodyNode::generateBytecode): 21171 (JSC::FunctionBodyNode::bytecodeForExceptionInfoReparse): 21172 * parser/Parser.h: 21173 (JSC::Parser::parse): 21174 (JSC::Parser::reparse): 21175 (JSC::Parser::parseFunctionFromGlobalCode): 21176 (JSC::::parse): 21177 * runtime/Completion.cpp: 21178 (JSC::checkSyntax): 21179 (JSC::evaluate): 21180 * runtime/Error.cpp: 21181 (JSC::throwError): 21182 * runtime/Error.h: 21183 * runtime/Executable.h: Added. 21184 (JSC::TemplateExecutable::TemplateExecutable): 21185 (JSC::TemplateExecutable::markAggregate): 21186 (JSC::TemplateExecutable::sourceURL): 21187 (JSC::TemplateExecutable::lineNo): 21188 (JSC::TemplateExecutable::bytecode): 21189 (JSC::TemplateExecutable::jitCode): 21190 (JSC::EvalExecutable::EvalExecutable): 21191 (JSC::ProgramExecutable::ProgramExecutable): 21192 * runtime/FunctionConstructor.cpp: 21193 (JSC::constructFunction): 21194 * runtime/FunctionConstructor.h: 21195 * runtime/JSGlobalData.cpp: 21196 (JSC::JSGlobalData::numericCompareFunction): 21197 * runtime/JSGlobalObject.cpp: 21198 (JSC::JSGlobalObject::~JSGlobalObject): 21199 (JSC::JSGlobalObject::markChildren): 21200 * runtime/JSGlobalObject.h: 21201 (JSC::JSGlobalObject::codeBlocks): 21202 * runtime/JSGlobalObjectFunctions.cpp: 21203 (JSC::globalFuncEval): 21204 212052009-08-14 Darin Adler <darin@apple.com> 21206 21207 Reviewed by Sam Weinig. 21208 21209 Rename the confusing isObject(<class>) to inherits(<class>). 21210 It still works on non-objects, returning false. 21211 21212 * runtime/ArrayConstructor.cpp: 21213 (JSC::arrayConstructorIsArray): Removed unneeded isObject call 21214 and updated remaining isObject call to new name, inherits. 21215 21216 * runtime/JSCell.h: Renamed isObject(<class>) to inherits(<class>) 21217 but more importantly, made it non-virtual (it was already inline) 21218 so it is now as fast as JSObject::inherits was. 21219 21220 * runtime/JSObject.h: Removed inherits function since the one 21221 in the base class is fine as-is. Also made various JSCell functions 21222 that should not be called on JSObject uncallable by making them 21223 both private and not implemented. 21224 (JSC::JSCell::inherits): Updated name. 21225 (JSC::JSValue::inherits): Ditto. 21226 21227 * debugger/Debugger.cpp: 21228 (JSC::Debugger::recompileAllJSFunctions): 21229 * interpreter/Interpreter.cpp: 21230 (JSC::Interpreter::unwindCallFrame): 21231 * runtime/ArrayPrototype.cpp: 21232 (JSC::arrayProtoFuncToString): 21233 (JSC::arrayProtoFuncToLocaleString): 21234 (JSC::arrayProtoFuncConcat): 21235 * runtime/BooleanPrototype.cpp: 21236 (JSC::booleanProtoFuncToString): 21237 (JSC::booleanProtoFuncValueOf): 21238 * runtime/DateConstructor.cpp: 21239 (JSC::constructDate): 21240 * runtime/DatePrototype.cpp: 21241 (JSC::dateProtoFuncToString): 21242 (JSC::dateProtoFuncToUTCString): 21243 (JSC::dateProtoFuncToISOString): 21244 (JSC::dateProtoFuncToDateString): 21245 (JSC::dateProtoFuncToTimeString): 21246 (JSC::dateProtoFuncToLocaleString): 21247 (JSC::dateProtoFuncToLocaleDateString): 21248 (JSC::dateProtoFuncToLocaleTimeString): 21249 (JSC::dateProtoFuncGetTime): 21250 (JSC::dateProtoFuncGetFullYear): 21251 (JSC::dateProtoFuncGetUTCFullYear): 21252 (JSC::dateProtoFuncToGMTString): 21253 (JSC::dateProtoFuncGetMonth): 21254 (JSC::dateProtoFuncGetUTCMonth): 21255 (JSC::dateProtoFuncGetDate): 21256 (JSC::dateProtoFuncGetUTCDate): 21257 (JSC::dateProtoFuncGetDay): 21258 (JSC::dateProtoFuncGetUTCDay): 21259 (JSC::dateProtoFuncGetHours): 21260 (JSC::dateProtoFuncGetUTCHours): 21261 (JSC::dateProtoFuncGetMinutes): 21262 (JSC::dateProtoFuncGetUTCMinutes): 21263 (JSC::dateProtoFuncGetSeconds): 21264 (JSC::dateProtoFuncGetUTCSeconds): 21265 (JSC::dateProtoFuncGetMilliSeconds): 21266 (JSC::dateProtoFuncGetUTCMilliseconds): 21267 (JSC::dateProtoFuncGetTimezoneOffset): 21268 (JSC::dateProtoFuncSetTime): 21269 (JSC::setNewValueFromTimeArgs): 21270 (JSC::setNewValueFromDateArgs): 21271 (JSC::dateProtoFuncSetYear): 21272 (JSC::dateProtoFuncGetYear): 21273 * runtime/FunctionPrototype.cpp: 21274 (JSC::functionProtoFuncToString): 21275 * runtime/JSActivation.cpp: 21276 (JSC::JSActivation::argumentsGetter): 21277 * runtime/JSValue.h: 21278 * runtime/RegExpConstructor.cpp: 21279 (JSC::constructRegExp): 21280 * runtime/RegExpPrototype.cpp: 21281 (JSC::regExpProtoFuncTest): 21282 (JSC::regExpProtoFuncExec): 21283 (JSC::regExpProtoFuncCompile): 21284 (JSC::regExpProtoFuncToString): 21285 * runtime/ScopeChain.cpp: 21286 (JSC::ScopeChain::localDepth): 21287 * runtime/StringPrototype.cpp: 21288 (JSC::stringProtoFuncReplace): 21289 (JSC::stringProtoFuncToString): 21290 (JSC::stringProtoFuncMatch): 21291 (JSC::stringProtoFuncSearch): 21292 (JSC::stringProtoFuncSplit): 21293 Updated to new name, inherits, from old name, isObject. 21294 212952009-07-31 Harald Fernengel <harald.fernengel@nokia.com> 21296 21297 Reviewed by Simon Hausmann. 21298 21299 Adding QNX as a platform. Currently only tested with Qt. 21300 21301 https://bugs.webkit.org/show_bug.cgi?id=27885 21302 21303 * JavaScriptCore/runtime/Collector.cpp: Added retrieving of stack base 21304 since QNX doesn't have the pthread _nt functions 21305 * JavaScriptCore/wtf/Platform.h: Added WTF_PLATFORM_QNX and corresponding 21306 defines 21307 * WebCore/bridge/npapi.h: Build fix for missing typedefs on QNX 21308 213092009-08-14 Gabor Loki <loki@inf.u-szeged.hu> 21310 21311 Reviewed by Simon Hausmann. 21312 21313 Currently generic ARM and ARMv7 platforms work only with JSVALUE32 21314 https://bugs.webkit.org/show_bug.cgi?id=28300 21315 21316 * wtf/Platform.h: 21317 213182009-08-14 Gabor Loki <loki@inf.u-szeged.hu> 21319 21320 Reviewed by Simon Hausmann. 21321 21322 Enable JIT on ARM for QT by default 21323 https://bugs.webkit.org/show_bug.cgi?id=28259 21324 21325 * wtf/Platform.h: 21326 213272009-08-14 Gabor Loki <loki@inf.u-szeged.hu> 21328 21329 Reviewed by Simon Hausmann. 21330 21331 Enable YARR_JIT on ARM for QT by default 21332 https://bugs.webkit.org/show_bug.cgi?id=28259 21333 21334 * wtf/Platform.h: 21335 213362009-08-14 Oliver Hunt <oliver@apple.com> 21337 21338 Reviewed by Gavin Barraclough. 21339 21340 [ES5] Arguments object should inherit from Array 21341 https://bugs.webkit.org/show_bug.cgi?id=28298 21342 21343 Make the Arguments object conform to the behaviour specified in ES5. 21344 The simple portion of this is to make Arguments use Array.prototype 21345 as its prototype rather than Object.prototype. 21346 21347 The spec then requires us to set instance.constructor to the pristine 21348 Object constructor, and instance.toString and instance.toLocaleString 21349 to the pristine versions from Object.prototype. To do this we now 21350 make the ObjectPrototype constructor return its toString and 21351 toLocaleString functions (similar to the call and apply functions 21352 from FunctionPrototype). 21353 21354 Oddly enough this reports itself as a slight win, but given the code 21355 isn't hit in the tests that claim to have improved I put this down to 21356 code motion. 21357 21358 * runtime/Arguments.h: 21359 (JSC::Arguments::Arguments): 21360 (JSC::Arguments::initializeStandardProperties): 21361 * runtime/JSGlobalObject.cpp: 21362 (JSC::JSGlobalObject::reset): 21363 (JSC::JSGlobalObject::markChildren): 21364 * runtime/JSGlobalObject.h: 21365 (JSC::JSGlobalObject::JSGlobalObjectData::JSGlobalObjectData): 21366 (JSC::JSGlobalObject::objectConstructor): 21367 (JSC::JSGlobalObject::objectToStringFunction): 21368 (JSC::JSGlobalObject::objectToLocaleStringFunction): 21369 * runtime/ObjectPrototype.cpp: 21370 (JSC::ObjectPrototype::ObjectPrototype): 21371 * runtime/ObjectPrototype.h: 21372 * tests/mozilla/ecma_3/Function/arguments-001.js: 21373 Update test to new es5 behaviour 21374 213752009-08-14 Oliver Hunt <oliver@apple.com> 21376 21377 Remove MarkStack::drain from the JSC exports file 21378 21379 MarkStack::drain is now marked inline, the including it in the exports file 21380 produces an ld warning 21381 21382 * JavaScriptCore.exp: 21383 213842009-08-13 Sam Weinig <sam@webkit.org> 21385 21386 Reviewed by Oliver Hunt. 21387 21388 Remove accidentally left in debugging statement. 21389 21390 * runtime/JSArray.h: 21391 (JSC::MarkStack::drain): 21392 213932009-08-13 Oliver Hunt <oliver@apple.com> 21394 21395 Reviewed by Maciej Stachowiak. 21396 21397 [ES5] Implement Array.isArray 21398 https://bugs.webkit.org/show_bug.cgi?id=28296 21399 21400 Add support for Array.isArray to the Array constructor 21401 21402 * runtime/ArrayConstructor.cpp: 21403 (JSC::ArrayConstructor::ArrayConstructor): 21404 (JSC::arrayConstructorIsArray): 21405 * runtime/ArrayConstructor.h: 21406 * runtime/CommonIdentifiers.h: 21407 * runtime/JSArray.h: 21408 (JSC::MarkStack::drain): 21409 * runtime/JSGlobalObject.cpp: 21410 (JSC::JSGlobalObject::reset): 21411 214122009-08-13 Oliver Hunt <oliver@apple.com> 21413 21414 Reviewed by NOBODY (Buildfix). 21415 21416 Attempt to fix windows build 21417 21418 * runtime/Collector.cpp: 21419 214202009-08-13 Oliver Hunt <oliver@apple.com> 21421 21422 Reviewed by Maciej Stachowiak. 21423 21424 Devirtualise marking 21425 https://bugs.webkit.org/show_bug.cgi?id=28294 21426 21427 Add a bit to TypeInfo to indicate that an object uses the standard 21428 JSObject::markChildren method. This allows us to devirtualise marking 21429 of most objects (though a branch is still needed). We also add a branch 21430 to identify arrays thus devirtualising marking in that case as well. 21431 21432 In order to make the best use of this devirtualisation I've also reworked 21433 the MarkStack::drain() logic to make the iteration more efficient. 21434 21435 * API/JSCallbackConstructor.h: 21436 (JSC::JSCallbackConstructor::createStructure): 21437 * API/JSCallbackFunction.h: 21438 (JSC::JSCallbackFunction::createStructure): 21439 * JavaScriptCore.exp: 21440 * runtime/BooleanObject.h: 21441 (JSC::BooleanObject::createStructure): 21442 * runtime/FunctionPrototype.h: 21443 (JSC::FunctionPrototype::createStructure): 21444 * runtime/InternalFunction.h: 21445 (JSC::InternalFunction::createStructure): 21446 * runtime/JSAPIValueWrapper.h: 21447 (JSC::JSAPIValueWrapper::JSAPIValueWrapper): 21448 * runtime/JSArray.cpp: 21449 (JSC::JSArray::markChildren): 21450 * runtime/JSArray.h: 21451 (JSC::JSArray::markChildrenDirect): 21452 (JSC::MarkStack::drain): 21453 * runtime/JSByteArray.cpp: 21454 (JSC::JSByteArray::createStructure): 21455 * runtime/JSCell.h: 21456 (JSC::MarkStack::append): 21457 * runtime/JSGlobalData.cpp: 21458 (JSC::JSGlobalData::JSGlobalData): 21459 * runtime/JSNumberCell.h: 21460 (JSC::JSNumberCell::createStructure): 21461 * runtime/JSONObject.h: 21462 (JSC::JSONObject::createStructure): 21463 * runtime/JSObject.cpp: 21464 (JSC::JSObject::markChildren): 21465 * runtime/JSObject.h: 21466 (JSC::JSObject::markChildrenDirect): 21467 (JSC::JSObject::createStructure): 21468 * runtime/JSString.h: 21469 (JSC::JSString::createStructure): 21470 * runtime/JSType.h: 21471 (JSC::): 21472 * runtime/MarkStack.h: 21473 (JSC::MarkStack::MarkStack): 21474 (JSC::MarkStack::MarkSet::MarkSet): 21475 (JSC::MarkStack::MarkStackArray::last): 21476 * runtime/MathObject.h: 21477 (JSC::MathObject::createStructure): 21478 * runtime/NumberConstructor.h: 21479 (JSC::NumberConstructor::createStructure): 21480 * runtime/NumberObject.h: 21481 (JSC::NumberObject::createStructure): 21482 * runtime/RegExpConstructor.h: 21483 (JSC::RegExpConstructor::createStructure): 21484 * runtime/RegExpObject.h: 21485 (JSC::RegExpObject::createStructure): 21486 * runtime/StringObjectThatMasqueradesAsUndefined.h: 21487 (JSC::StringObjectThatMasqueradesAsUndefined::createStructure): 21488 * runtime/TypeInfo.h: 21489 (JSC::TypeInfo::hasDefaultMark): 21490 214912009-08-13 Darin Adler <darin@apple.com> 21492 21493 Reviewed by Mark Rowe. 21494 21495 Some small bits of housekeeping. 21496 21497 * JavaScriptCore.xcodeproj/project.pbxproj: Make Parser.h 21498 project instead of private. Remove JSONObject.lut.h. 21499 21500 * assembler/ARMAssembler.h: Remove unneeded WTF prefix. 21501 * assembler/AssemblerBufferWithConstantPool.h: Ditto. 21502 * bytecompiler/BytecodeGenerator.h: Ditto. 21503 21504 * wtf/SegmentedVector.h: Add a "using" statement as we do 21505 with the other WTF headers. 21506 215072009-08-13 Darin Adler <darin@apple.com> 21508 21509 Fix Tiger build. 21510 21511 * parser/Grammar.y: Use a template function so we can compile 21512 setStatementLocation even if it comes before YYLTYPE is defined. 21513 215142009-08-13 Darin Adler <darin@apple.com> 21515 21516 Reviewed by George Staikos. 21517 21518 Too much use of void* in Grammar.y 21519 https://bugs.webkit.org/show_bug.cgi?id=28287 21520 21521 * parser/Grammar.y: Changed all the helper functions to 21522 take a JSGlobalData* instead of a void*. A couple formatting 21523 tweaks that I missed when breaking this into pieces. 21524 215252009-08-13 Darin Adler <darin@apple.com> 21526 21527 Reviewed by George Staikos. 21528 21529 Another part of https://bugs.webkit.org/show_bug.cgi?id=28287 21530 21531 * parser/Grammar.y: Reduced and sorted includes. Tweaked comment 21532 format. Marked a few more functions inline. 21533 215342009-08-13 Darin Adler <darin@apple.com> 21535 21536 Reviewed by George Staikos. 21537 21538 Another part of https://bugs.webkit.org/show_bug.cgi?id=28287 21539 21540 * parser/Grammar.y: Pass the number to the PropertyNode instead of 21541 first turning it into an Identifier. 21542 21543 * parser/NodeConstructors.h: 21544 (JSC::PropertyNode::PropertyNode): Add an overload that takes a double 21545 so the code to convert to a string can be here instead of Grammar.y. 21546 * parser/Nodes.h: Ditto. 21547 215482009-08-13 Darin Adler <darin@apple.com> 21549 21550 Reviewed by George Staikos. 21551 21552 Another part of https://bugs.webkit.org/show_bug.cgi?id=28287 21553 21554 * parser/Grammar.y: Eliminate the DBG macro. 21555 215562009-08-13 Darin Adler <darin@apple.com> 21557 21558 Reviewed by George Staikos. 21559 21560 Another part of https://bugs.webkit.org/show_bug.cgi?id=28287 21561 21562 * parser/Grammar.y: Eliminate the SET_EXCEPTION_LOCATION macro. 21563 215642009-08-13 Darin Adler <darin@apple.com> 21565 21566 Reviewed by George Staikos. 21567 21568 George asked me to break the patch from 21569 https://bugs.webkit.org/show_bug.cgi?id=28287 21570 into smaller pieces and land it in stages. 21571 21572 * parser/Grammar.y: Eliminate the LEXER macro. 21573 215742009-08-13 Mark Rowe <mrowe@apple.com> 21575 21576 Try some more to fix the Windows build. 21577 21578 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: Export a new symbol. 21579 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: Ditto. 21580 215812009-08-13 Mark Rowe <mrowe@apple.com> 21582 21583 Try and fix the Windows build. 21584 21585 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: Export a new symbol. 21586 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: Ditto. 21587 215882009-08-13 Darin Adler <darin@apple.com> 21589 21590 Reviewed by David Levin. 21591 21592 JavaScriptCore tweaks to get ready for the parser arena 21593 https://bugs.webkit.org/show_bug.cgi?id=28243 21594 21595 Eliminate dependencies on Nodes.h outside JavaScriptCore, 21596 and cut down on them inside JavaScriptCore. 21597 21598 Change regular expression parsing to use identifiers as 21599 with other strings we parse. 21600 21601 Fix a couple things that are needed to use const Identifier 21602 more, which will be part of the parser arena work. 21603 21604 * JavaScriptCore.exp: Resorted and updated. 21605 21606 * JavaScriptCore.xcodeproj/project.pbxproj: Changed 21607 CollectorHeapIterator.h to be project-internal. 21608 21609 * bytecompiler/BytecodeGenerator.cpp: 21610 (JSC::BytecodeGenerator::emitPushNewScope): Added const. 21611 * bytecompiler/BytecodeGenerator.h: Ditto. 21612 21613 * debugger/Debugger.cpp: 21614 (JSC::Debugger::recompileAllJSFunctions): Moved this function 21615 here from WebCore. Here is better since it uses so many internals. 21616 Removed unimportant optimization for the no listener case. 21617 * debugger/Debugger.h: Ditto. Also removed unneeded include 21618 and tweaked formatting and comments. 21619 21620 * debugger/DebuggerCallFrame.cpp: 21621 (JSC::DebuggerCallFrame::functionName): Call asFunction instead 21622 of doing the unchecked static_cast. 21623 (JSC::DebuggerCallFrame::calculatedFunctionName): Ditto. 21624 21625 * jit/JITStubs.cpp: 21626 (JSC::op_call_JSFunction): Call isHostFunction on the body rather 21627 than on the JSFunction. 21628 (JSC::vm_lazyLinkCall): Ditto. 21629 (JSC::op_construct_JSConstruct): Ditto. 21630 21631 * parser/Grammar.y: Changed callers to use new scanRegExp with 21632 out arguments instead of relying on state in the Lexer. And 21633 callers that just want to skip a regular expression to use 21634 skipRegExp. 21635 21636 * parser/Lexer.cpp: 21637 (JSC::Lexer::scanRegExp): Changed to use out arguments, and to 21638 add a prefix argument so we can add in the "=" character as needed. 21639 Also rewrote to streamline the logic a bit inspired by suggestions 21640 by David Levin. 21641 (JSC::Lexer::skipRegExp): Added. Version of the function above that 21642 does not actually put the regular expression into a string. 21643 (JSC::Lexer::clear): Removed code to clear m_pattern and m_flags. 21644 * parser/Lexer.h: Changed scanRegExp to have out arguments. Added 21645 skipRegExp. Eliminated pattern, flags, m_pattern, and m_flags. 21646 21647 * parser/NodeConstructors.h: 21648 (JSC::RegExpNode::RegExpNode): Changed to take const Identifier&. 21649 * parser/Nodes.cpp: 21650 (JSC::RegExpNode::emitBytecode): Changed since m_pattern and 21651 m_flags are now Identifier instead of UString. 21652 (JSC::FunctionBodyNode::make): Moved this function here instead 21653 of putting it in the JSFunction.h header. 21654 * parser/Nodes.h: Changed RegExpNode to use Identifier. 21655 21656 * profiler/Profiler.cpp: 21657 (JSC::Profiler::createCallIdentifier): Changed to use isHostFunction 21658 on the body instead of on the JSFunction object. 21659 * runtime/FunctionPrototype.cpp: 21660 (JSC::functionProtoFuncToString): Ditto. 21661 21662 * runtime/JSFunction.cpp: 21663 (JSC::JSFunction::isHostFunction): Moved here from header. 21664 (JSC::JSFunction::isHostFunctionNonInline): Added. 21665 (JSC::JSFunction::JSFunction): Removed unneeded initialization of 21666 m_body to 0. 21667 (JSC::JSFunction::setBody): Moved here from header. 21668 21669 * runtime/JSFunction.h: Removed unneeded includes. Moved private 21670 constructor down to the private section. Made virtual functions 21671 private. Removed unneeded overload of setBody and moved the body 21672 of the function into the .cpp file. Changed assertions to use 21673 the non-inline version of isHostFunction. 21674 21675 * runtime/PropertySlot.cpp: 21676 (JSC::PropertySlot::functionGetter): Use asFunction instead 21677 of doing the unchecked static_cast. 21678 21679 * wtf/SegmentedVector.h: 21680 (WTF::SegmentedVector::isEmpty): Added. 21681 216822009-08-13 Mark Rowe <mrowe@apple.com> 21683 21684 Rubber-stamped by Darin Adler. 21685 21686 Use the version of operator new that takes a JSGlobalData when allocating FuncDeclNode and FuncExprNode 21687 from within the grammar to prevent these nodes from being leaked. 21688 21689 * parser/Grammar.y: 21690 216912009-08-13 Simon Hausmann <simon.hausmann@nokia.com> 21692 21693 Reviewed by Ariya Hidayat. 21694 21695 Remove the special-case for Qt wrt JSVALUE_32 introduced in 21696 r46709. It must've been a dependency issue on the bot, as 21697 after a manual build all the tests pass on amd64 and ia32. 21698 21699 * wtf/Platform.h: 21700 217012009-08-12 Gabor Loki <loki@inf.u-szeged.hu> 21702 21703 Reviewed by Gavin Barraclough. 21704 21705 Add optimize call and property access support for ARM JIT. 21706 https://bugs.webkit.org/show_bug.cgi?id=24986 21707 21708 For tightly coupled sequences the BEGIN_UNINTERRUPTED_SEQUENCE and 21709 END_UNINTERRUPTED_SEQUENCE macros have been introduced which ensure 21710 space for instructions and constants of the named sequence. This 21711 method is vital for those architecture which are using constant pool. 21712 21713 The 'latePatch' method - which was linked to JmpSrc - is replaced with 21714 a port specific solution (each calls are marked to place their address 21715 on the constant pool). 21716 21717 * assembler/ARMAssembler.cpp: 21718 (JSC::ARMAssembler::linkBranch): 21719 (JSC::ARMAssembler::executableCopy): Add extra align for constant pool. 21720 * assembler/ARMAssembler.h: 21721 (JSC::ARMAssembler::JmpSrc::JmpSrc): 21722 (JSC::ARMAssembler::sizeOfConstantPool): 21723 (JSC::ARMAssembler::jmp): 21724 (JSC::ARMAssembler::linkCall): 21725 * assembler/ARMv7Assembler.h: 21726 * assembler/AbstractMacroAssembler.h: 21727 * assembler/AssemblerBufferWithConstantPool.h: 21728 (JSC::AssemblerBufferWithConstantPool::flushIfNoSpaceFor): Fix the 21729 computation of the remaining space. 21730 * assembler/MacroAssemblerARM.h: 21731 (JSC::MacroAssemblerARM::branch32): 21732 (JSC::MacroAssemblerARM::nearCall): 21733 (JSC::MacroAssemblerARM::call): 21734 (JSC::MacroAssemblerARM::branchPtrWithPatch): 21735 (JSC::MacroAssemblerARM::ensureSpace): 21736 (JSC::MacroAssemblerARM::sizeOfConstantPool): 21737 (JSC::MacroAssemblerARM::prepareCall): 21738 * assembler/X86Assembler.h: 21739 * jit/JIT.h: 21740 * jit/JITCall.cpp: 21741 (JSC::JIT::compileOpCall): 21742 * jit/JITInlineMethods.h: 21743 (JSC::JIT::beginUninterruptedSequence): 21744 (JSC::JIT::endUninterruptedSequence): 21745 * jit/JITPropertyAccess.cpp: 21746 (JSC::JIT::emit_op_method_check): 21747 (JSC::JIT::compileGetByIdHotPath): 21748 (JSC::JIT::compileGetByIdSlowCase): 21749 (JSC::JIT::emit_op_put_by_id): 21750 217512009-08-12 Gavin Barraclough <barraclough@apple.com> 21752 21753 Rubber Stamped by Dave Kilzer. 21754 21755 Disable WTF_USE_JSVALUE32_64 on iPhone for now (support not yet added for ARMv7). 21756 21757 * wtf/Platform.h: 21758 217592009-08-12 Gavin Barraclough <barraclough@apple.com> 21760 21761 Reviewed by Maciej Stachoviak. 21762 21763 Ooops - moved code that had been accidentally added to op_new_func instead of 21764 op_new_func_exp, to where it shoulds be. 21765 21766 * interpreter/Interpreter.cpp: 21767 (JSC::Interpreter::privateExecute): 21768 * wtf/Platform.h: 21769 217702009-08-12 Ada Chan <adachan@apple.com> 21771 21772 Added workaround for the limitation that VirtualFree with MEM_RELEASE 21773 can only accept the base address returned by VirtualAlloc when the region 21774 was reserved and it can only free the entire region, and not a part of it. 21775 21776 Reviewed by Oliver Hunt. 21777 21778 * runtime/MarkStack.h: 21779 (JSC::MarkStack::MarkStackArray::shrinkAllocation): 21780 * runtime/MarkStackWin.cpp: 21781 (JSC::MarkStack::releaseStack): 21782 217832009-08-12 Balazs Kelemen <kelemen.balazs.3@stud.u-szeged.hu> 21784 21785 Reviewed by Ariya Hidayat. 21786 21787 Build fix: use std::numeric_limits<long long>::min() instead of LLONG_MIN 21788 since LLONG_MIN is not defined in standard c++. 21789 21790 * runtime/UString.cpp: 21791 (JSC::UString::from): 21792 217932009-08-12 Benjamin Otte <otte@gnome.org> 21794 21795 Reviewed by Jan Alonzo. 21796 21797 Buildfix for Gtk platforms debug builds. 21798 21799 * GNUmakefile.am: Choose MarkStackPosix.cpp or MarkStackWin.cpp 21800 depending on platform. 21801 218022009-08-12 Simon Hausmann <simon.hausmann@nokia.com> 21803 21804 Prospective build fix for Mac and 32-bit Windows. 21805 21806 * runtime/UString.cpp: Include wtf/StringExtras.h for snprintf. 21807 (JSC::UString::from): Use %lld instead of %I64d for snprintf 21808 on non-windows platforms. 21809 218102009-08-12 Prasanth Ullattil <prasanth.ullattil@nokia.com> 21811 21812 Reviewed by Simon Hausmann. 21813 21814 Fix compile error on 64Bit Windows, when UString::from 21815 is called with an intptr_t. 21816 21817 Added new UString::From overload with long long parameter. 21818 21819 Thanks to Holger for the long long idea. 21820 21821 * runtime/UString.cpp: 21822 (JSC::UString::from): 21823 * runtime/UString.h: 21824 218252009-08-11 Oliver Hunt <oliver@apple.com> 21826 21827 Reviewed by Mark Rowe. 21828 21829 Minor style fixes. 21830 21831 * runtime/UString.h: 21832 (JSC::UString::Rep::createEmptyBuffer): 21833 * wtf/FastMalloc.h: 21834 (WTF::TryMallocReturnValue::getValue): 21835 218362009-08-11 Oliver Hunt <oliver@apple.com> 21837 21838 Reviewed by Gavin Barraclough. 21839 21840 Make it harder to misuse try* allocation routines 21841 https://bugs.webkit.org/show_bug.cgi?id=27469 21842 21843 Jump through a few hoops to make it much harder to accidentally 21844 miss null-checking of values returned by the try-* allocation 21845 routines. 21846 21847 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 21848 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 21849 * JavaScriptCore.xcodeproj/project.pbxproj: 21850 * runtime/JSArray.cpp: 21851 (JSC::JSArray::putSlowCase): 21852 (JSC::JSArray::increaseVectorLength): 21853 * runtime/StringPrototype.cpp: 21854 (JSC::stringProtoFuncFontsize): 21855 (JSC::stringProtoFuncLink): 21856 * runtime/UString.cpp: 21857 (JSC::allocChars): 21858 (JSC::reallocChars): 21859 (JSC::expandCapacity): 21860 (JSC::UString::Rep::reserveCapacity): 21861 (JSC::UString::expandPreCapacity): 21862 (JSC::createRep): 21863 (JSC::concatenate): 21864 (JSC::UString::spliceSubstringsWithSeparators): 21865 (JSC::UString::replaceRange): 21866 (JSC::UString::append): 21867 (JSC::UString::operator=): 21868 * runtime/UString.h: 21869 (JSC::UString::Rep::createEmptyBuffer): 21870 * wtf/FastMalloc.cpp: 21871 (WTF::tryFastZeroedMalloc): 21872 (WTF::tryFastMalloc): 21873 (WTF::tryFastCalloc): 21874 (WTF::tryFastRealloc): 21875 (WTF::TCMallocStats::tryFastMalloc): 21876 (WTF::TCMallocStats::tryFastCalloc): 21877 (WTF::TCMallocStats::tryFastRealloc): 21878 * wtf/FastMalloc.h: 21879 (WTF::TryMallocReturnValue::TryMallocReturnValue): 21880 (WTF::TryMallocReturnValue::~TryMallocReturnValue): 21881 (WTF::TryMallocReturnValue::operator PossiblyNull<T>): 21882 (WTF::TryMallocReturnValue::getValue): 21883 * wtf/Platform.h: 21884 * wtf/PossiblyNull.h: Added. 21885 (WTF::PossiblyNull::PossiblyNull): 21886 (WTF::PossiblyNull::~PossiblyNull): 21887 (WTF::::getValue): 21888 218892009-08-11 Gavin Barraclough <barraclough@apple.com> 21890 21891 Reviewed by NOBODY (build fix part deux). 21892 21893 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 21894 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 21895 218962009-08-11 Gavin Barraclough <barraclough@apple.com> 21897 21898 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 21899 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 21900 219012009-08-11 Gavin Barraclough <barraclough@apple.com> 21902 21903 Reviewed by Oliver Hunt. 21904 21905 Restrict use of FuncDeclNode & FuncExprNode to the parser. 21906 https://bugs.webkit.org/show_bug.cgi?id=28209 21907 21908 These objects were also being referenced from the CodeBlock. By changing this 21909 to just retain pointers to FunctionBodyNodes these classes can be restricted to 21910 use during parsing. 21911 21912 No performance impact (or sub-percent progression). 21913 21914 * JavaScriptCore.exp: 21915 Update symbols. 21916 21917 * bytecode/CodeBlock.cpp: 21918 (JSC::CodeBlock::mark): 21919 (JSC::CodeBlock::reparseForExceptionInfoIfNecessary): 21920 (JSC::CodeBlock::shrinkToFit): 21921 * bytecode/CodeBlock.h: 21922 (JSC::CodeBlock::addFunction): 21923 (JSC::CodeBlock::function): 21924 Unify m_functions & m_functionExpressions into a single Vector<RefPtr<FuncExprNode> >. 21925 21926 * bytecompiler/BytecodeGenerator.cpp: 21927 (JSC::BytecodeGenerator::BytecodeGenerator): 21928 (JSC::BytecodeGenerator::addConstant): 21929 (JSC::BytecodeGenerator::emitNewFunction): 21930 (JSC::BytecodeGenerator::emitNewFunctionExpression): 21931 * bytecompiler/BytecodeGenerator.h: 21932 FunctionStacks now contain FunctionBodyNodes not FuncDeclNodes. 21933 21934 * interpreter/Interpreter.cpp: 21935 (JSC::Interpreter::execute): 21936 (JSC::Interpreter::privateExecute): 21937 Update to reflect chnages in CodeBlock. 21938 21939 * jit/JITOpcodes.cpp: 21940 (JSC::JIT::emit_op_new_func_exp): 21941 * jit/JITStubs.cpp: 21942 (JSC::DEFINE_STUB_FUNCTION): 21943 * jit/JITStubs.h: 21944 (JSC::): 21945 Update to reflect chnages in CodeBlock. 21946 21947 * parser/Grammar.y: 21948 FunctionStacks now contain FunctionBodyNodes not FuncDeclNodes. 21949 21950 * parser/NodeConstructors.h: 21951 (JSC::FuncExprNode::FuncExprNode): 21952 (JSC::FuncDeclNode::FuncDeclNode): 21953 * parser/Nodes.cpp: 21954 (JSC::ScopeNodeData::mark): 21955 (JSC::FunctionBodyNode::finishParsing): 21956 * parser/Nodes.h: 21957 (JSC::FunctionBodyNode::ident): 21958 Move m_ident & make methods from FuncDeclNode & FuncExprNode to FunctionBodyNode. 21959 21960 * runtime/JSFunction.h: 21961 (JSC::FunctionBodyNode::make): 21962 Make this method inline (was FuncDeclNode::makeFunction). 21963 219642009-08-11 Oliver Hunt <oliver@apple.com> 21965 21966 Reviewed by Gavin Barraclough. 21967 21968 Native JSON.stringify does not omit functions 21969 https://bugs.webkit.org/show_bug.cgi?id=28117 21970 21971 Objects that are callable should be treated as undefined when 21972 serialising to JSON. 21973 21974 * runtime/JSONObject.cpp: 21975 (JSC::Stringifier::appendStringifiedValue): 21976 219772009-08-11 Oliver Hunt <oliver@apple.com> 21978 21979 Reviewed by Geoff Garen. 21980 21981 REGRESSION: Hang/crash in BytecodeGenerator::constRegisterFor loading simple page 21982 https://bugs.webkit.org/show_bug.cgi?id=28169 21983 21984 Handle the case where someone has attempted to shadow a property 21985 on the global object with a constant. 21986 21987 * bytecompiler/BytecodeGenerator.cpp: 21988 (JSC::BytecodeGenerator::constRegisterFor): 21989 * parser/Nodes.cpp: 21990 (JSC::ConstDeclNode::emitCodeSingle): 21991 219922009-08-11 John Gregg <johnnyg@google.com> 21993 21994 Reviewed by Maciej Stachowiak. 21995 21996 Desktop Notifications API 21997 https://bugs.webkit.org/show_bug.cgi?id=25463 21998 21999 Adds ENABLE_NOTIFICATION flag. 22000 22001 * Configurations/FeatureDefines.xcconfig: 22002 * wtf/Platform.h: 22003 220042009-08-11 Maxime Simon <simon.maxime@gmail.com> 22005 22006 Reviewed by Eric Seidel. 22007 22008 Modifications on JavaScriptCore to allow Haiku port. 22009 https://bugs.webkit.org/show_bug.cgi?id=28121 22010 22011 * runtime/Collector.cpp: Haiku doesn't have sys/mman.h, using OS.h instead. 22012 (JSC::currentThreadStackBase): Haiku uses its own threading system. 22013 * wtf/Platform.h: Defining all Haiku platform values. 22014 * wtf/haiku/MainThreadHaiku.cpp: Adding a missing header (NotImplemented.h). 22015 220162009-08-11 Jessie Berlin <jberlin@apple.com> 22017 22018 Reviewed by Adam Roben. 22019 22020 Fix windows build. 22021 22022 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 22023 220242009-08-11 Csaba Osztrogonac <oszi@inf.u-szeged.hu> 22025 22026 Reviewed by Tor Arne Vestbø. 22027 22028 Buildfix for Qt-win platforms. 22029 22030 * JavaScriptCore.pri: Choose MarkStackPosix.cpp or MarkStackWin.cpp depend on platform. 22031 220322009-08-10 Oliver Hunt <oliver@apple.com> 22033 22034 Reviewed by NOBODY (And another build fix). 22035 22036 Add new exports for MSVC 22037 22038 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 22039 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 22040 * JavaScriptCore.xcodeproj/project.pbxproj: 22041 220422009-08-10 Oliver Hunt <oliver@apple.com> 22043 22044 Reviewed by NOBODY (yet another build fix). 22045 22046 Remove obsolete entries from MSVC exports file 22047 22048 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 22049 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 22050 220512009-08-10 Oliver Hunt <oliver@apple.com> 22052 22053 Add includes needed for non-allinonefile builds 22054 22055 * runtime/GetterSetter.h: 22056 * runtime/ScopeChain.h: 22057 220582009-08-10 Oliver Hunt <oliver@apple.com> 22059 22060 Fix export file for last build fix 22061 22062 * JavaScriptCore.exp: 22063 220642009-08-10 Oliver Hunt <oliver@apple.com> 22065 22066 Hoist page size initialization into platform specific code. 22067 22068 * jit/ExecutableAllocatorPosix.cpp: 22069 * jit/ExecutableAllocatorWin.cpp: 22070 * runtime/MarkStack.h: 22071 (JSC::MarkStack::pageSize): 22072 * runtime/MarkStackPosix.cpp: 22073 (JSC::MarkStack::initializePagesize): 22074 * runtime/MarkStackWin.cpp: 22075 (JSC::MarkStack::initializePagesize): 22076 220772009-08-07 Oliver Hunt <oliver@apple.com> 22078 22079 Reviewed by Gavin Barraclough. 22080 22081 Stack overflow crash in JavaScript garbage collector mark pass 22082 https://bugs.webkit.org/show_bug.cgi?id=12216 22083 22084 Make the GC mark phase iterative by using an explicit mark stack. 22085 To do this marking any single object is performed in multiple stages 22086 * The object is appended to the MarkStack, this sets the marked 22087 bit for the object using the new markDirect() function, and then 22088 returns 22089 * When the MarkStack is drain()ed the object is popped off the stack 22090 and markChildren(MarkStack&) is called on the object to collect 22091 all of its children. drain() then repeats until the stack is empty. 22092 22093 Additionally I renamed a number of methods from 'mark' to 'markAggregate' 22094 in order to make it more clear that marking of those object was not 22095 going to result in an actual recursive mark. 22096 22097 * GNUmakefile.am 22098 * JavaScriptCore.exp: 22099 * JavaScriptCore.gypi: 22100 * JavaScriptCore.pri: 22101 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: 22102 * JavaScriptCore.xcodeproj/project.pbxproj: 22103 * bytecode/CodeBlock.cpp: 22104 (JSC::CodeBlock::markAggregate): 22105 * bytecode/CodeBlock.h: 22106 * bytecode/EvalCodeCache.h: 22107 (JSC::EvalCodeCache::markAggregate): 22108 * debugger/DebuggerActivation.cpp: 22109 (JSC::DebuggerActivation::markChildren): 22110 * debugger/DebuggerActivation.h: 22111 * interpreter/Register.h: 22112 * interpreter/RegisterFile.h: 22113 (JSC::RegisterFile::markGlobals): 22114 (JSC::RegisterFile::markCallFrames): 22115 * parser/Nodes.cpp: 22116 (JSC::ScopeNodeData::markAggregate): 22117 (JSC::EvalNode::markAggregate): 22118 (JSC::FunctionBodyNode::markAggregate): 22119 * parser/Nodes.h: 22120 (JSC::ScopeNode::markAggregate): 22121 * runtime/ArgList.cpp: 22122 (JSC::MarkedArgumentBuffer::markLists): 22123 * runtime/ArgList.h: 22124 * runtime/Arguments.cpp: 22125 (JSC::Arguments::markChildren): 22126 * runtime/Arguments.h: 22127 * runtime/Collector.cpp: 22128 (JSC::Heap::markConservatively): 22129 (JSC::Heap::markCurrentThreadConservativelyInternal): 22130 (JSC::Heap::markCurrentThreadConservatively): 22131 (JSC::Heap::markOtherThreadConservatively): 22132 (JSC::Heap::markStackObjectsConservatively): 22133 (JSC::Heap::markProtectedObjects): 22134 (JSC::Heap::collect): 22135 * runtime/Collector.h: 22136 * runtime/GetterSetter.cpp: 22137 (JSC::GetterSetter::markChildren): 22138 * runtime/GetterSetter.h: 22139 (JSC::GetterSetter::GetterSetter): 22140 (JSC::GetterSetter::createStructure): 22141 * runtime/GlobalEvalFunction.cpp: 22142 (JSC::GlobalEvalFunction::markChildren): 22143 * runtime/GlobalEvalFunction.h: 22144 * runtime/JSActivation.cpp: 22145 (JSC::JSActivation::markChildren): 22146 * runtime/JSActivation.h: 22147 * runtime/JSArray.cpp: 22148 (JSC::JSArray::markChildren): 22149 * runtime/JSArray.h: 22150 * runtime/JSCell.h: 22151 (JSC::JSCell::markCellDirect): 22152 (JSC::JSCell::markChildren): 22153 (JSC::JSValue::markDirect): 22154 (JSC::JSValue::markChildren): 22155 (JSC::JSValue::hasChildren): 22156 (JSC::MarkStack::append): 22157 (JSC::MarkStack::drain): 22158 * runtime/JSFunction.cpp: 22159 (JSC::JSFunction::markChildren): 22160 * runtime/JSFunction.h: 22161 * runtime/JSGlobalData.cpp: 22162 (JSC::JSGlobalData::JSGlobalData): 22163 * runtime/JSGlobalData.h: 22164 * runtime/JSGlobalObject.cpp: 22165 (JSC::markIfNeeded): 22166 (JSC::JSGlobalObject::markChildren): 22167 * runtime/JSGlobalObject.h: 22168 * runtime/JSNotAnObject.cpp: 22169 (JSC::JSNotAnObject::markChildren): 22170 * runtime/JSNotAnObject.h: 22171 * runtime/JSONObject.cpp: 22172 (JSC::Stringifier::markAggregate): 22173 (JSC::JSONObject::markStringifiers): 22174 * runtime/JSONObject.h: 22175 * runtime/JSObject.cpp: 22176 (JSC::JSObject::markChildren): 22177 (JSC::JSObject::defineGetter): 22178 (JSC::JSObject::defineSetter): 22179 * runtime/JSObject.h: 22180 * runtime/JSPropertyNameIterator.cpp: 22181 (JSC::JSPropertyNameIterator::markChildren): 22182 * runtime/JSPropertyNameIterator.h: 22183 (JSC::JSPropertyNameIterator::createStructure): 22184 (JSC::JSPropertyNameIterator::JSPropertyNameIterator): 22185 (JSC::JSPropertyNameIterator::create): 22186 * runtime/JSStaticScopeObject.cpp: 22187 (JSC::JSStaticScopeObject::markChildren): 22188 * runtime/JSStaticScopeObject.h: 22189 * runtime/JSType.h: 22190 (JSC::): 22191 * runtime/JSValue.h: 22192 * runtime/JSWrapperObject.cpp: 22193 (JSC::JSWrapperObject::markChildren): 22194 * runtime/JSWrapperObject.h: 22195 * runtime/MarkStack.cpp: Added. 22196 (JSC::MarkStack::compact): 22197 * runtime/MarkStack.h: Added. 22198 (JSC::): 22199 (JSC::MarkStack::MarkStack): 22200 (JSC::MarkStack::append): 22201 (JSC::MarkStack::appendValues): 22202 (JSC::MarkStack::~MarkStack): 22203 (JSC::MarkStack::MarkSet::MarkSet): 22204 (JSC::MarkStack::pageSize): 22205 22206 MarkStackArray is a non-shrinking, mmap-based vector type 22207 used for storing objects to be marked. 22208 (JSC::MarkStack::MarkStackArray::MarkStackArray): 22209 (JSC::MarkStack::MarkStackArray::~MarkStackArray): 22210 (JSC::MarkStack::MarkStackArray::expand): 22211 (JSC::MarkStack::MarkStackArray::append): 22212 (JSC::MarkStack::MarkStackArray::removeLast): 22213 (JSC::MarkStack::MarkStackArray::isEmpty): 22214 (JSC::MarkStack::MarkStackArray::size): 22215 (JSC::MarkStack::MarkStackArray::shrinkAllocation): 22216 * runtime/MarkStackPosix.cpp: Added. 22217 (JSC::MarkStack::allocateStack): 22218 (JSC::MarkStack::releaseStack): 22219 * runtime/MarkStackWin.cpp: Added. 22220 (JSC::MarkStack::allocateStack): 22221 (JSC::MarkStack::releaseStack): 22222 22223 * runtime/ScopeChain.h: 22224 * runtime/ScopeChainMark.h: 22225 (JSC::ScopeChain::markAggregate): 22226 * runtime/SmallStrings.cpp: 22227 (JSC::SmallStrings::mark): 22228 * runtime/Structure.h: 22229 (JSC::Structure::markAggregate): 22230 222312009-08-10 Mark Rowe <mrowe@apple.com> 22232 22233 Reviewed by Darin Adler. 22234 22235 Fix hundreds of "pointer being freed was not allocated" errors seen on the build bot. 22236 22237 * wtf/FastMalloc.h: Implement nothrow variants of the delete and delete[] operators since 22238 we implement the nothrow variants of new and new[]. The nothrow variant of delete is called 22239 explicitly in the implementation of std::sort which was resulting in FastMalloc-allocated 22240 memory being passed to the system allocator to free. 22241 222422009-08-10 Jan Michael Alonzo <jmalonzo@webkit.org> 22243 22244 [Gtk] Unreviewed build fix. Move JSAPIValueWrapper.cpp/.h in the debug 22245 section. This file is already part of AllInOneFile in Release builds. 22246 22247 * GNUmakefile.am: 22248 222492009-08-10 Darin Adler <darin@apple.com> 22250 22251 * wtf/FastMalloc.h: Fix build. 22252 222532009-08-10 Darin Adler <darin@apple.com> 22254 22255 Reviewed by Mark Rowe. 22256 22257 FastMalloc.h has cross-platform code but marked as WinCE-only 22258 https://bugs.webkit.org/show_bug.cgi?id=28160 22259 22260 1) The support for nothrow was inside #if PLATFORM(WINCE) even though it is 22261 not platform-specific. 22262 2) The code tried to override operator delete nothrow, which does not exist. 22263 3) The code in the header checks the value of USE_SYSTEM_MALLOC, but the code 22264 in FastMalloc.cpp checks only if the macro is defined. 22265 22266 * wtf/FastMalloc.h: See above. 22267 * wtf/FastMalloc.cpp: Ditto. 22268 222692009-08-10 Sam Weinig <sam@webkit.org> 22270 22271 Reviewed by Anders Carlsson. 22272 22273 Fix an annoying indentation issue. 22274 22275 * runtime/DateConstructor.cpp: 22276 (JSC::constructDate): 22277 222782009-08-10 Xan Lopez <xlopez@igalia.com> 22279 22280 Unreviewed build fix. 22281 22282 Add new files to makefile. 22283 22284 * GNUmakefile.am: 22285 222862009-08-10 Simon Hausmann <simon.hausmann@nokia.com> 22287 22288 Fix compilation with the interpreter instead of the JIT by including 22289 PrototypeFunction.h as forward-declared through NativeFunctionWrapper.h. 22290 22291 * runtime/ObjectConstructor.cpp: 22292 222932009-08-09 Oliver Hunt <oliver@apple.com> 22294 22295 Reviewed by George Staikos. 22296 22297 JSON.stringify replacer returning undefined does not omit object properties 22298 https://bugs.webkit.org/show_bug.cgi?id=28118 22299 22300 Correct behaviour of stringify when using a replacer function that returns 22301 undefined. This is a simple change to move the undefined value check to 22302 after the replacer function is called. This means that the replacer function 22303 is now called for properties with the value undefined, however i've confirmed 22304 that this behaviour is correct. 22305 22306 In addition I've made the cyclic object exception have a more useful error 22307 message. 22308 22309 * runtime/JSONObject.cpp: 22310 (JSC::Stringifier::appendStringifiedValue): 22311 223122009-08-08 Oliver Hunt <oliver@apple.com> 22313 22314 Reviewed by Eric Seidel and Sam Weinig. 22315 22316 [ES5] Implement Object.getPrototypeOf 22317 https://bugs.webkit.org/show_bug.cgi?id=28114 22318 22319 Implement getPrototypeOf 22320 22321 * runtime/CommonIdentifiers.h: 22322 * runtime/JSGlobalObject.cpp: 22323 (JSC::JSGlobalObject::reset): 22324 * runtime/ObjectConstructor.cpp: 22325 (JSC::ObjectConstructor::ObjectConstructor): 22326 (JSC::objectConsGetPrototypeOf): 22327 * runtime/ObjectConstructor.h: 22328 223292009-08-07 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 22330 22331 Reviewed by Eric Seidel. 22332 22333 Allow custom memory allocation control for Noncopyable class 22334 https://bugs.webkit.org/show_bug.cgi?id=27879 22335 22336 Several classes which are inherited from Noncopyable are instantiated by 22337 operator new, so Noncopyable class has been inherited from FastAllocBase. 22338 22339 * wtf/Noncopyable.h: 22340 223412009-08-07 George Staikos <george.staikos@torchmobile.com> 22342 22343 Reviewed by Eric Seidel. 22344 22345 https://bugs.webkit.org/show_bug.cgi?id=27305 22346 Implement WinCE-specific unicode layer. 22347 Written by George Staikos <george.staikos@torchmobile.com> 22348 with bug fixes by Yong Li <yong.li@torchmobile.com> 22349 refactored by Joe Mason <joe.mason@torchmobile.com> 22350 22351 * wtf/Platform.h: 22352 * wtf/unicode/Unicode.h: 22353 * wtf/unicode/wince/UnicodeWince.cpp: Added. 22354 (WTF::Unicode::toLower): 22355 (WTF::Unicode::toUpper): 22356 (WTF::Unicode::foldCase): 22357 (WTF::Unicode::isPrintableChar): 22358 (WTF::Unicode::isSpace): 22359 (WTF::Unicode::isLetter): 22360 (WTF::Unicode::isUpper): 22361 (WTF::Unicode::isLower): 22362 (WTF::Unicode::isDigit): 22363 (WTF::Unicode::isPunct): 22364 (WTF::Unicode::toTitleCase): 22365 (WTF::Unicode::direction): 22366 (WTF::Unicode::category): 22367 (WTF::Unicode::decompositionType): 22368 (WTF::Unicode::combiningClass): 22369 (WTF::Unicode::mirroredChar): 22370 (WTF::Unicode::digitValue): 22371 * wtf/unicode/wince/UnicodeWince.h: Added. 22372 (WTF::Unicode::): 22373 (WTF::Unicode::isSeparatorSpace): 22374 (WTF::Unicode::isHighSurrogate): 22375 (WTF::Unicode::isLowSurrogate): 22376 (WTF::Unicode::isArabicChar): 22377 (WTF::Unicode::hasLineBreakingPropertyComplexContext): 22378 (WTF::Unicode::umemcasecmp): 22379 (WTF::Unicode::surrogateToUcs4): 22380 223812009-08-07 Yongjun Zhang <yongjun.zhang@nokia.com> 22382 22383 Reviewed by Eric Seidel. 22384 22385 https://bugs.webkit.org/show_bug.cgi?id=28069 22386 22387 Add inline to help winscw compiler resolve specialized argument in 22388 templated functions. 22389 22390 * runtime/LiteralParser.cpp: 22391 (JSC::LiteralParser::Lexer::lexString): 22392 223932009-08-07 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 22394 22395 Reviewed by Eric Seidel. 22396 22397 Allow custom memory allocation control for RegExpObjectData struct 22398 http://bugs.webkit.org/show_bug.cgi?id=26750 22399 22400 Inherits RegExpObjectData struct from FastAllocBase because 22401 it has been instantiated by 'new' in JavaScriptCore/runtime/RegExpObject.cpp:62 22402 22403 * runtime/RegExpObject.h: 22404 224052009-08-06 Norbert Leser <norbert.leser@nokia.com> 22406 22407 Reviewed by Darin Adler. 22408 22409 Updated patch for bug #27059: 22410 Symbian platform always uses little endian encoding, 22411 regardless of compiler. 22412 We need to make sure that we correctly detect EABI architecture 22413 for armv5 targets on Symbian, 22414 where __EABI__ is set but not __ARM_EABI__ 22415 22416 * wtf/Platform.h: 22417 224182009-08-06 Adam Barth <abarth@webkit.org> 22419 22420 Unreviewed revert. 22421 22422 http://bugs.webkit.org/show_bug.cgi?id=27879 22423 22424 Revert 46877 because it broke GTK. 22425 22426 * wtf/Noncopyable.h: 22427 224282009-08-06 Gavin Barraclough <barraclough@apple.com> 22429 22430 Reviewed by Oliver Hunt. 22431 22432 Make get_by_id/put_by_id/method_check/call defer optimization using a data flag rather than a code modification. 22433 ( https://bugs.webkit.org/show_bug.cgi?id=27635 ) 22434 22435 This improves performance of ENABLE(ASSEMBLER_WX_EXCLUSIVE) builds by 2-2.5%, reducing the overhead to about 2.5%. 22436 (No performance impact with ASSEMBLER_WX_EXCLUSIVE disabled). 22437 22438 * bytecode/CodeBlock.cpp: 22439 (JSC::printStructureStubInfo): 22440 - Make StructureStubInfo store the type as an integer, rather than an OpcodeID. 22441 22442 * bytecode/CodeBlock.h: 22443 (JSC::): 22444 (JSC::CallLinkInfo::seenOnce): 22445 (JSC::CallLinkInfo::setSeen): 22446 (JSC::MethodCallLinkInfo::seenOnce): 22447 (JSC::MethodCallLinkInfo::setSeen): 22448 - Change a pointer in CallLinkInfo/MethodCallLinkInfo to use a PtrAndFlags, use a flag to track when an op has been executed once. 22449 22450 * bytecode/StructureStubInfo.cpp: 22451 (JSC::StructureStubInfo::deref): 22452 - Make StructureStubInfo store the type as an integer, rather than an OpcodeID. 22453 22454 * bytecode/StructureStubInfo.h: 22455 (JSC::StructureStubInfo::StructureStubInfo): 22456 (JSC::StructureStubInfo::initGetByIdSelf): 22457 (JSC::StructureStubInfo::initGetByIdProto): 22458 (JSC::StructureStubInfo::initGetByIdChain): 22459 (JSC::StructureStubInfo::initGetByIdSelfList): 22460 (JSC::StructureStubInfo::initGetByIdProtoList): 22461 (JSC::StructureStubInfo::initPutByIdTransition): 22462 (JSC::StructureStubInfo::initPutByIdReplace): 22463 (JSC::StructureStubInfo::seenOnce): 22464 (JSC::StructureStubInfo::setSeen): 22465 - Make StructureStubInfo store the type as an integer, rather than an OpcodeID, add a flag to track when an op has been executed once. 22466 22467 * bytecompiler/BytecodeGenerator.cpp: 22468 (JSC::BytecodeGenerator::emitGetById): 22469 (JSC::BytecodeGenerator::emitPutById): 22470 - Make StructureStubInfo store the type as an integer, rather than an OpcodeID. 22471 22472 * jit/JIT.cpp: 22473 (JSC::JIT::privateCompileCTIMachineTrampolines): 22474 (JSC::JIT::unlinkCall): 22475 - Remove the "don't lazy link" stage of calls. 22476 22477 * jit/JIT.h: 22478 (JSC::JIT::compileCTIMachineTrampolines): 22479 - Remove the "don't lazy link" stage of calls. 22480 22481 * jit/JITCall.cpp: 22482 (JSC::JIT::compileOpCallSlowCase): 22483 - Remove the "don't lazy link" stage of calls. 22484 22485 * jit/JITStubs.cpp: 22486 (JSC::JITThunks::JITThunks): 22487 (JSC::JITThunks::tryCachePutByID): 22488 (JSC::JITThunks::tryCacheGetByID): 22489 (JSC::JITStubs::DEFINE_STUB_FUNCTION): 22490 (JSC::JITStubs::getPolymorphicAccessStructureListSlot): 22491 - Remove the "don't lazy link" stage of calls, and the "_second" stage of get_by_id/put_by_id/method_check. 22492 22493 * jit/JITStubs.h: 22494 (JSC::JITThunks::ctiStringLengthTrampoline): 22495 (JSC::JITStubs::): 22496 - Remove the "don't lazy link" stage of calls, and the "_second" stage of get_by_id/put_by_id/method_check. 22497 22498 * wtf/PtrAndFlags.h: 22499 (WTF::PtrAndFlags::PtrAndFlags): 22500 (WTF::PtrAndFlags::operator!): 22501 (WTF::PtrAndFlags::operator->): 22502 - Add ! and -> operators, add constuctor with pointer argument. 22503 225042009-08-06 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 22505 22506 Reviewed by Adam Barth. 22507 22508 Allow custom memory allocation control for Noncopyable class 22509 https://bugs.webkit.org/show_bug.cgi?id=27879 22510 22511 Several classes which inherited from Noncopyable are instantiated by 22512 operator new, so Noncopyable class has been inherited from FastAllocBase. 22513 22514 * wtf/Noncopyable.h: 22515 225162009-08-06 Mark Rowe <mrowe@apple.com> 22517 22518 Rubber-stamped by Sam Weinig. 22519 22520 Add explicit dependencies for our build verification scripts to ensure that they always run after linking has completed. 22521 22522 * JavaScriptCore.xcodeproj/project.pbxproj: 22523 225242009-08-06 Mark Rowe <mrowe@apple.com> 22525 22526 Bring a little order to our otherwise out of control lives. 22527 22528 * JavaScriptCore.xcodeproj/project.pbxproj: 22529 225302009-08-06 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 22531 22532 Reviewed by Darin Adler. 22533 22534 Allow custom memory allocation control for JavaScriptCore's PolymorphicAccessStructureList struct 22535 https://bugs.webkit.org/show_bug.cgi?id=27877 22536 22537 Inherits PolymorphicAccessStructureList struct from FastAllocBase because it has been instantiated by 22538 'new' in JavaScriptCore/jit/JITStubs.cpp:1229. 22539 22540 * bytecode/Instruction.h: 22541 225422009-08-05 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 22543 22544 Reviewed by Darin Adler. 22545 22546 Allow custom memory allocation control for JavaScriptCore's ScopeNodeData struct 22547 https://bugs.webkit.org/show_bug.cgi?id=27875 22548 22549 Inherits ScopeNodeData struct from FastAllocBase because it has been instantiated by 22550 'new' in JavaScriptCore/parser/Nodes.cpp:1848. 22551 22552 * parser/Nodes.h: 22553 225542009-08-05 Zoltan Herczeg <zherczeg@inf.u-szeged.hu> 22555 22556 Reviewed by Gavin Barraclough. 22557 22558 Add floating point support for generic ARM port. 22559 https://bugs.webkit.org/show_bug.cgi?id=24986 22560 22561 * assembler/ARMAssembler.cpp: 22562 (JSC::ARMAssembler::doubleTransfer): 22563 * assembler/ARMAssembler.h: 22564 (JSC::ARM::): 22565 (JSC::ARMAssembler::): 22566 (JSC::ARMAssembler::faddd_r): 22567 (JSC::ARMAssembler::fsubd_r): 22568 (JSC::ARMAssembler::fmuld_r): 22569 (JSC::ARMAssembler::fcmpd_r): 22570 (JSC::ARMAssembler::fdtr_u): 22571 (JSC::ARMAssembler::fdtr_d): 22572 (JSC::ARMAssembler::fmsr_r): 22573 (JSC::ARMAssembler::fsitod_r): 22574 (JSC::ARMAssembler::fmstat): 22575 * assembler/MacroAssemblerARM.h: 22576 (JSC::MacroAssemblerARM::): 22577 (JSC::MacroAssemblerARM::supportsFloatingPoint): 22578 (JSC::MacroAssemblerARM::loadDouble): 22579 (JSC::MacroAssemblerARM::storeDouble): 22580 (JSC::MacroAssemblerARM::addDouble): 22581 (JSC::MacroAssemblerARM::subDouble): 22582 (JSC::MacroAssemblerARM::mulDouble): 22583 (JSC::MacroAssemblerARM::convertInt32ToDouble): 22584 (JSC::MacroAssemblerARM::branchDouble): 22585 * jit/JIT.h: 22586 225872009-08-05 Zoltan Herczeg <zherczeg@inf.u-szeged.hu> 22588 22589 Reviewed by Gavin Barraclough. 22590 22591 Add JIT support for generic ARM port without optimizations. 22592 https://bugs.webkit.org/show_bug.cgi?id=24986 22593 22594 All JIT optimizations are disabled. 22595 22596 Signed off by Zoltan Herczeg <zherczeg@inf.u-szeged.hu> 22597 Signed off by Gabor Loki <loki@inf.u-szeged.hu> 22598 22599 * assembler/ARMAssembler.cpp: 22600 (JSC::ARMAssembler::baseIndexTransfer32): 22601 * assembler/AbstractMacroAssembler.h: 22602 (JSC::AbstractMacroAssembler::Imm32::Imm32): 22603 * assembler/MacroAssemblerARM.h: 22604 (JSC::MacroAssemblerARM::store32): 22605 (JSC::MacroAssemblerARM::move): 22606 (JSC::MacroAssemblerARM::branch32): 22607 (JSC::MacroAssemblerARM::add32): 22608 (JSC::MacroAssemblerARM::sub32): 22609 (JSC::MacroAssemblerARM::load32): 22610 * bytecode/CodeBlock.h: 22611 (JSC::CodeBlock::getBytecodeIndex): 22612 * jit/JIT.h: 22613 * jit/JITInlineMethods.h: 22614 (JSC::JIT::restoreArgumentReference): 22615 * jit/JITOpcodes.cpp: 22616 * jit/JITStubs.cpp: 22617 * jit/JITStubs.h: 22618 (JSC::JITStackFrame::returnAddressSlot): 22619 * wtf/Platform.h: 22620 226212009-08-04 Gavin Barraclough <barraclough@apple.com> 22622 22623 Rubber Stamped by Oiver Hunt. 22624 22625 Revert r46643 since this breaks the Yarr::Interpreter running the v8 tests. 22626 https://bugs.webkit.org/show_bug.cgi?id=27874 22627 22628 * yarr/RegexInterpreter.cpp: 22629 (JSC::Yarr::Interpreter::allocDisjunctionContext): 22630 (JSC::Yarr::Interpreter::freeDisjunctionContext): 22631 (JSC::Yarr::Interpreter::allocParenthesesDisjunctionContext): 22632 (JSC::Yarr::Interpreter::freeParenthesesDisjunctionContext): 22633 226342009-08-04 Oliver Hunt <oliver@apple.com> 22635 22636 PPC64 Build fix 22637 22638 * wtf/Platform.h: 22639 226402009-08-04 Benjamin C Meyer <benjamin.meyer@torchmobile.com> 22641 22642 Reviewed by Adam Treat 22643 22644 Explicitly include limits.h header when using INT_MAX and INT_MIN 22645 22646 * interpreter/Interpreter.cpp 22647 226482009-08-03 Harald Fernengel <harald.fernengel@nokia.com> 22649 22650 Reviewed by Darin Adler. 22651 22652 Fix compile error for ambigous call to abs() 22653 https://bugs.webkit.org/show_bug.cgi?id=27873 22654 22655 Fix ambiguity in abs(long int) call by calling labs() instead 22656 22657 * wtf/DateMath.cpp: replace call to abs() with labs() 22658 226592009-08-03 Laszlo Gombos <laszlo.1.gombos@nokia.com> 22660 22661 Reviewed by Eric Seidel. 22662 22663 [Qt] Consolidate common gcc flags to WebKit.pri 22664 https://bugs.webkit.org/show_bug.cgi?id=27934 22665 22666 * JavaScriptCore.pro: 22667 226682009-08-03 Ada Chan <adachan@apple.com> 22669 22670 Fixed the Tiger build. 22671 22672 * wtf/FastMalloc.cpp: 22673 226742009-08-03 Ada Chan <adachan@apple.com> 22675 22676 Reviewed by Darin Adler. 22677 22678 Don't use background thread to scavenge memory on Tiger until we figure out why it causes a crash. 22679 https://bugs.webkit.org/show_bug.cgi?id=27900 22680 22681 * wtf/FastMalloc.cpp: 22682 226832009-08-03 Fumitoshi Ukai <ukai@chromium.org> 22684 22685 Reviewed by Jan Alonzo. 22686 22687 Fix build break on Gtk/x86_64. 22688 https://bugs.webkit.org/show_bug.cgi?id=27936 22689 22690 Use JSVALUE64 for X86_64 LINUX, except Qt. 22691 22692 * wtf/Platform.h: 22693 226942009-08-02 Xan Lopez <xlopez@igalia.com> 22695 22696 Fix the GTK+ build. 22697 22698 * wtf/Platform.h: 22699 227002009-08-02 Geoffrey Garen <ggaren@apple.com> 22701 22702 Reviewed by Sam Weinig. 22703 22704 Disabled JSVALUE32_64 on Qt builds, since all layout tests mysteriously 22705 crash with it enabled. 22706 22707 * wtf/Platform.h: 22708 227092009-08-02 Geoffrey Garen <ggaren@apple.com> 22710 22711 Qt build fix. 22712 22713 Added JSAPIValueWrapper.cpp to the build. 22714 22715 * JavaScriptCore.pri: 22716 227172009-08-02 Geoffrey Garen <ggaren@apple.com> 22718 22719 Windows build fix. 22720 22721 Exported symbols for JSAPIValueWrapper. 22722 22723 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 22724 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 22725 227262009-08-02 Geoffrey Garen <ggaren@apple.com> 22727 22728 GTK build fix. 22729 22730 * jit/JITStubs.cpp: #include <stdarg.h>, for a definition of va_start. 22731 227322009-08-02 Geoffrey Garen <ggaren@apple.com> 22733 22734 Qt build fix. 22735 22736 * runtime/Collector.cpp: #include <limits.h>, for a definition of ULONG_MAX. 22737 227382009-08-02 Geoffrey Garen <ggaren@apple.com> 22739 22740 Windows build fix: Nixed JSImmediate::prototype, JSImmediate::toObject, 22741 and JSImmediate::toThisObject, and removed their exported symbols. 22742 22743 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 22744 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 22745 * runtime/JSImmediate.cpp: 22746 * runtime/JSImmediate.h: 22747 227482009-08-02 Geoffrey Garen <ggaren@apple.com> 22749 22750 Reviewed by Mark Rowe. 22751 22752 Enabled JSVALUE32_64 by default on all platforms other than x86_64 (which uses JSVALUE64). 22753 22754 * wtf/Platform.h: 22755 227562009-08-02 Kevin Ollivier <kevino@theolliviers.com> 22757 22758 Reviewed by Jan Alonzo. 22759 22760 Script for building the JavaScriptCore library for wx. 22761 https://bugs.webkit.org/show_bug.cgi?id=27619 22762 22763 * wscript: Added. 22764 227652009-08-02 Yong Li <yong.li@torchmobile.com> 22766 22767 Reviewed by George Staikos. 22768 22769 DateMath depends on strftime and localtime, which need to be imported manually on WinCE 22770 https://bugs.webkit.org/show_bug.cgi?id=26558 22771 22772 * wtf/DateMath.cpp: 22773 227742009-08-01 David Kilzer <ddkilzer@apple.com> 22775 22776 wtf/Threading.h: added include of Platform.h 22777 22778 Reviewed by Mark Rowe. 22779 22780 * wtf/Threading.h: Added #include "Platform.h" since this header 22781 uses PLATFORM() and other macros. 22782 227832009-08-01 Mark Rowe <mrowe@apple.com> 22784 22785 Rubber-stamped by Oliver Hunt. 22786 22787 Roll out r46668 as it was misinformed. ScopeChain is only used with placement new. 22788 22789 * runtime/ScopeChain.h: 22790 227912009-08-01 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 22792 22793 Allow custom memory allocation control for JavaScriptCore's HashMap class 22794 http://bugs.webkit.org/show_bug.cgi?id=27871 22795 22796 Inherits HashMap class from FastAllocBase because it has been 22797 instantiated by 'new' in JavaScriptCore/API/JSClassRef.cpp:148. 22798 22799 * wtf/RefPtrHashMap.h: 22800 (WTF::): 22801 228022009-08-01 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 22803 22804 Allow custom memory allocation control for JavaScriptCore's ScopeChain class 22805 https://bugs.webkit.org/show_bug.cgi?id=27834 22806 22807 Inherits ScopeChain class from FastAllocBase because it has been 22808 instantiated by 'new' in JavaScriptCore/runtime/JSFunction.h:109. 22809 22810 * runtime/ScopeChain.h: 22811 228122009-08-01 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 22813 22814 Reviewed by Darin Adler. 22815 22816 Allow custom memory allocation control for JavaScriptCore's RegExpConstructorPrivate struct 22817 https://bugs.webkit.org/show_bug.cgi?id=27833 22818 22819 Inherits RegExpConstructorPrivate class from FastAllocBase because it has been 22820 instantiated by 'new' in JavaScriptCore/runtime/RegExpConstructor.cpp:152. 22821 22822 * runtime/RegExpConstructor.cpp: 22823 228242009-07-31 Yong Li <yong.li@torchmobile.com> 22825 22826 Reviewed by George Staikos. 22827 22828 Resurrect the old GetTickCount implementation of currentTime, controlled by WTF_USE_QUERY_PERFORMANCE_COUNTER 22829 currentSystemTime taken from older WebKit; currentTime written by Yong Li <yong.li@torchmobile.com>; cleanup by Joe Mason <joe.mason@torchmobile.com> 22830 https://bugs.webkit.org/show_bug.cgi?id=27848 22831 22832 * wtf/CurrentTime.cpp: 22833 (WTF::currentSystemTime): get current time with GetCurrentFT 22834 (WTF::currentTime): track msec elapsed since first currentSystemTime call using GetTickCount 22835 * wtf/Platform.h: 22836 228372009-07-31 Ada Chan <adachan@apple.com> 22838 22839 Fixes the Windows release-PGO build. 22840 22841 Reviewed by Jon Honeycutt. 22842 22843 * JavaScriptCore.vcproj/WTF/WTF.vcproj: Suppresses the warning about unreachable code that we get by adding "return 0" to WTF::TCMalloc_PageHeap::runScavengerThread(). 22844 * wtf/FastMalloc.cpp: 22845 (WTF::TCMalloc_PageHeap::runScavengerThread): Fixes the error about the method not returning a value in the release-PGO build. 22846 228472009-07-31 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 22848 22849 Change malloc to fastMalloc and free to fastFree in Yarr's RegexInterpreter.cpp 22850 https://bugs.webkit.org/show_bug.cgi?id=27874 22851 22852 Use fastMalloc and fastFree instead of malloc and free in RegexInterpreter.cpp's methods. 22853 22854 * yarr/RegexInterpreter.cpp: 22855 (JSC::Yarr::Interpreter::allocDisjunctionContext): 22856 (JSC::Yarr::Interpreter::freeDisjunctionContext): 22857 (JSC::Yarr::Interpreter::allocParenthesesDisjunctionContext): 22858 (JSC::Yarr::Interpreter::freeParenthesesDisjunctionContext): 22859 228602009-07-30 Xan Lopez <xlopez@igalia.com> 22861 22862 Reviewed by Jan Alonzo. 22863 22864 Fix compiler warning. 22865 22866 GCC does not like C++-style comments in preprocessor directives. 22867 22868 * wtf/Platform.h: 22869 228702009-07-30 John McCall <rjmccall@apple.com> 22871 22872 Reviewed by Gavin Barraclough. 22873 22874 Optimize the X86_64 trampolines: avoid the need for filler arguments 22875 and move the stub-args area closer to the stack pointer. 22876 22877 * jit/JIT.h: adjust patch offsets because of slight code-size change 22878 * jit/JITCode.h: 22879 (JSC::JITCode::execute): don't pass filler args 22880 * jit/JITStubs.cpp: 22881 (ctiTrampoline): (X86_64): push args onto stack, use args directly 22882 (ctiVMThrowTrampoline): (X86_64): adjust %rsp by correct displacement 22883 (ctiOpThrowNotCaught): (X86_64): adjust %rsp by correct displacement 22884 * jit/JITStubs.h: 22885 (JITStackFrame): (X86_64): move args area earlier 22886 (ctiTrampoline): remove filler args from prototype 22887 228882009-07-30 Gavin Barraclough <barraclough@apple.com> 22889 22890 Temporarily revert r46618 since this is b0rking on Linux. 22891 228922009-07-23 Gavin Barraclough <barraclough@apple.com> 22893 22894 Reviewed by Oliver Hunt. 22895 22896 Make get_by_id/put_by_id/method_check/call defer optimization using a data flag rather than a code modification. 22897 ( https://bugs.webkit.org/show_bug.cgi?id=27635 ) 22898 22899 This improves performance of ENABLE(ASSEMBLER_WX_EXCLUSIVE) builds by 2-2.5%, reducing the overhead to about 2.5%. 22900 (No performance impact with ASSEMBLER_WX_EXCLUSIVE disabled). 22901 22902 * bytecode/CodeBlock.cpp: 22903 (JSC::printStructureStubInfo): 22904 - Make StructureStubInfo store the type as an integer, rather than an OpcodeID. 22905 22906 * bytecode/CodeBlock.h: 22907 (JSC::): 22908 (JSC::CallLinkInfo::seenOnce): 22909 (JSC::CallLinkInfo::setSeen): 22910 (JSC::MethodCallLinkInfo::seenOnce): 22911 (JSC::MethodCallLinkInfo::setSeen): 22912 - Change a pointer in CallLinkInfo/MethodCallLinkInfo to use a PtrAndFlags, use a flag to track when an op has been executed once. 22913 22914 * bytecode/StructureStubInfo.cpp: 22915 (JSC::StructureStubInfo::deref): 22916 - Make StructureStubInfo store the type as an integer, rather than an OpcodeID. 22917 22918 * bytecode/StructureStubInfo.h: 22919 (JSC::StructureStubInfo::StructureStubInfo): 22920 (JSC::StructureStubInfo::initGetByIdSelf): 22921 (JSC::StructureStubInfo::initGetByIdProto): 22922 (JSC::StructureStubInfo::initGetByIdChain): 22923 (JSC::StructureStubInfo::initGetByIdSelfList): 22924 (JSC::StructureStubInfo::initGetByIdProtoList): 22925 (JSC::StructureStubInfo::initPutByIdTransition): 22926 (JSC::StructureStubInfo::initPutByIdReplace): 22927 (JSC::StructureStubInfo::seenOnce): 22928 (JSC::StructureStubInfo::setSeen): 22929 - Make StructureStubInfo store the type as an integer, rather than an OpcodeID, add a flag to track when an op has been executed once. 22930 22931 * bytecompiler/BytecodeGenerator.cpp: 22932 (JSC::BytecodeGenerator::emitGetById): 22933 (JSC::BytecodeGenerator::emitPutById): 22934 - Make StructureStubInfo store the type as an integer, rather than an OpcodeID. 22935 22936 * jit/JIT.cpp: 22937 (JSC::JIT::privateCompileCTIMachineTrampolines): 22938 (JSC::JIT::unlinkCall): 22939 - Remove the "don't lazy link" stage of calls. 22940 22941 * jit/JIT.h: 22942 (JSC::JIT::compileCTIMachineTrampolines): 22943 - Remove the "don't lazy link" stage of calls. 22944 22945 * jit/JITCall.cpp: 22946 (JSC::JIT::compileOpCallSlowCase): 22947 - Remove the "don't lazy link" stage of calls. 22948 22949 * jit/JITStubs.cpp: 22950 (JSC::JITThunks::JITThunks): 22951 (JSC::JITThunks::tryCachePutByID): 22952 (JSC::JITThunks::tryCacheGetByID): 22953 (JSC::JITStubs::DEFINE_STUB_FUNCTION): 22954 (JSC::JITStubs::getPolymorphicAccessStructureListSlot): 22955 - Remove the "don't lazy link" stage of calls, and the "_second" stage of get_by_id/put_by_id/method_check. 22956 22957 * jit/JITStubs.h: 22958 (JSC::JITThunks::ctiStringLengthTrampoline): 22959 (JSC::JITStubs::): 22960 - Remove the "don't lazy link" stage of calls, and the "_second" stage of get_by_id/put_by_id/method_check. 22961 22962 * wtf/PtrAndFlags.h: 22963 (WTF::PtrAndFlags::PtrAndFlags): 22964 (WTF::PtrAndFlags::operator!): 22965 (WTF::PtrAndFlags::operator->): 22966 - Add ! and -> operators, add constuctor with pointer argument. 22967 229682009-07-30 Geoffrey Garen <ggaren@apple.com> 22969 22970 Reviewed by Gavin Barraclough. 22971 22972 Fixed failing tests seen on Windows buildbot. 22973 22974 * jit/JITStubs.cpp: 22975 (JSC::DEFINE_STUB_FUNCTION): 22976 * jit/JITStubs.h: 22977 (JSC::): Use "int" instead of "bool" to guarantee a 32-bit result, 22978 regardless of compiler. gcc on mac uses 32-bit values for bool, 22979 but gcc on linux and MSVC on Windows use 8-bit values. 22980 229812009-07-30 Geoffrey Garen <ggaren@apple.com> 22982 22983 Windows build fix: added missing symbols on Windows. 22984 22985 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 22986 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 22987 229882009-07-30 Geoffrey Garen <ggaren@apple.com> 22989 22990 Windows build fix: removed stale symbols on Windows. 22991 22992 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 22993 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 22994 22995=== End merge of nitro-extreme branch 2009-07-30 === 22996 229972009-07-20 Geoffrey Garen <ggaren@apple.com> 22998 22999 Fixed a post-review typo in r46066 that caused tons of test failures. 23000 23001 SunSpider reports no change. 23002 23003 * runtime/JSArray.cpp: 23004 (JSC::JSArray::JSArray): Initialize the full vector capacity, to avoid 23005 uninitialized members at the end. 23006 230072009-07-20 Geoffrey Garen <ggaren@apple.com> 23008 23009 Windows WebKit build fix: Added some missing exports. 23010 23011 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 23012 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 23013 230142009-07-17 Geoffrey Garen <ggaren@apple.com> 23015 23016 Reviewed by Sam Weinig. 23017 23018 Get the branch working on windows. 23019 https://bugs.webkit.org/show_bug.cgi?id=27391 23020 23021 SunSpider says 0.3% faster. 23022 23023 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 23024 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: Updated 23025 MSVC export lists to fix linker errors. 23026 23027 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: Added / removed 23028 new / old project files. 23029 23030 * jit/JIT.cpp: 23031 (JSC::JIT::privateCompileCTIMachineTrampolines): Used #pragma pack to tell 23032 MSVC that these structures represent actual memory layout, and should not be 23033 automatically aligned. Changed the return value load to load a 64bit quantity 23034 into the canonical registers. 23035 23036 * jit/JIT.h: Moved OBJECT_OFFSETOF definition to StdLibExtras.h because 23037 it's needed by more than just the JIT, and it supplements a standard library 23038 macro (offsetof). 23039 23040 * jit/JITCall.cpp: 23041 (JSC::JIT::compileOpCallInitializeCallFrame): Fixed an incorrectly signed 23042 cast to resolve an MSVC warning. 23043 23044 * jit/JITStubs.h: Used #pragma pack to tell MSVC that these structures 23045 represent actual memory layout, and should not be automatically aligned. 23046 23047 * runtime/JSArray.cpp: 23048 (JSC::JSArray::JSArray): Replaced memset_pattern8 with a for loop, since 23049 memset_pattern8 is not portable. (I verified that this version of the loop 23050 gives the best performance / generated code in GCC.) 23051 23052 * runtime/JSObject.h: 23053 (JSC::JSObject::JSObject): Removed accidental usage of FIELD_OFFSET -- 23054 OBJECT_OFFSETOF is our new macro name. (FIELD_OFFSET conflicts with a 23055 definition in winnt.h.) 23056 23057 * runtime/JSValue.cpp: Added some headers needed by non-all-in-one builds. 23058 23059 * runtime/JSValue.h: 23060 (JSC::JSValue::): Made the tag signed, to match MSVC's signed enum values. 23061 (GCC doesn't seem to care one way or the other.) 23062 23063 * wtf/MainThread.cpp: Moved the StdLibExtras.h #include -- I did this a 23064 while ago to resolve a conflict with winnt.h. I can't remember if it's truly 23065 still needed, but what the heck. 23066 23067 * wtf/StdLibExtras.h: Moved OBJECT_OFFSETOF definition here. 23068 230692009-07-06 Geoffrey Garen <ggaren@apple.com> 23070 23071 Reviewed by Sam Weinig (?). 23072 23073 Fixed an assertion seen during the stress test. 23074 23075 Don't assume that, if op1 is constant, op2 is not, and vice versa. Sadly, 23076 not all constants get folded. 23077 23078 * jit/JITArithmetic.cpp: 23079 (JSC::JIT::emit_op_jnless): 23080 (JSC::JIT::emitSlow_op_jnless): 23081 (JSC::JIT::emit_op_jnlesseq): 23082 (JSC::JIT::emitSlow_op_jnlesseq): 23083 230842009-07-06 Geoffrey Garen <ggaren@apple.com> 23085 23086 Reviewed by Sam Weinig. 23087 23088 Include op_convert_this in result caching. 23089 23090 No change on SunSpider or v8. 23091 23092 * jit/JITOpcodes.cpp: 23093 (JSC::JIT::emit_op_convert_this): 23094 23095 * jit/JITStubs.cpp: 23096 (JSC::DEFINE_STUB_FUNCTION): 23097 * jit/JITStubs.h: 23098 (JSC::): Made the op_convert_this JIT stub return an EncodedJSValue, so 23099 to maintain the result caching contract that { tag, payload } can be 23100 found in { regT1, regT0 }. 23101 231022009-07-06 Geoffrey Garen <ggaren@apple.com> 23103 23104 Reviewed by Sam Weinig. 23105 23106 Implemented result chaining. 23107 23108 1% faster on SunSpider. 4%-5% faster on v8. 23109 23110 * assembler/MacroAssemblerX86Common.h: 23111 (JSC::MacroAssemblerX86Common::move): 23112 * assembler/X86Assembler.h: 23113 (JSC::X86Assembler::movl_rr): Added an optimization to eliminate 23114 no-op mov instructions, to simplify chaining. 23115 23116 * jit/JIT.cpp: 23117 (JSC::JIT::JIT): 23118 * jit/JIT.h: Added data members and helper functions for recording 23119 chained results. We record both a mapping from virtual to machine register 23120 and the opcode for which the mapping is valid, to help ensure that the 23121 mapping isn't used after the mapped register has been stomped by other 23122 instructions. 23123 23124 * jit/JITCall.cpp: 23125 (JSC::JIT::compileOpCallVarargs): 23126 (JSC::JIT::compileOpCallVarargsSlowCase): 23127 (JSC::JIT::emit_op_ret): 23128 (JSC::JIT::emit_op_construct_verify): 23129 (JSC::JIT::compileOpCall): 23130 (JSC::JIT::compileOpCallSlowCase): Chain function call results. 23131 23132 * jit/JITInlineMethods.h: 23133 (JSC::JIT::emitLoadTag): 23134 (JSC::JIT::emitLoadPayload): 23135 (JSC::JIT::emitLoad): 23136 (JSC::JIT::emitLoad2): 23137 (JSC::JIT::isLabeled): 23138 (JSC::JIT::map): 23139 (JSC::JIT::unmap): 23140 (JSC::JIT::isMapped): 23141 (JSC::JIT::getMappedPayload): 23142 (JSC::JIT::getMappedTag): Use helper functions when loading virtual 23143 registers into machine registers, in case the loads can be eliminated 23144 by chaining. 23145 23146 * jit/JITOpcodes.cpp: 23147 (JSC::JIT::emit_op_mov): 23148 (JSC::JIT::emit_op_end): 23149 (JSC::JIT::emit_op_instanceof): 23150 (JSC::JIT::emit_op_get_global_var): 23151 (JSC::JIT::emit_op_put_global_var): 23152 (JSC::JIT::emit_op_get_scoped_var): 23153 (JSC::JIT::emit_op_put_scoped_var): 23154 (JSC::JIT::emit_op_to_primitive): 23155 (JSC::JIT::emit_op_resolve_global): 23156 (JSC::JIT::emit_op_jneq_ptr): 23157 (JSC::JIT::emit_op_next_pname): 23158 (JSC::JIT::emit_op_to_jsnumber): 23159 (JSC::JIT::emit_op_catch): Chain results from these opcodes. 23160 23161 (JSC::JIT::emit_op_profile_will_call): 23162 (JSC::JIT::emit_op_profile_did_call): Load the profiler into regT2 to 23163 avoid stomping a chained result. 23164 23165 * jit/JITPropertyAccess.cpp: 23166 (JSC::JIT::emit_op_method_check): 23167 (JSC::JIT::emit_op_get_by_val): 23168 (JSC::JIT::emit_op_get_by_id): Chain results from these opcodes. 23169 23170 * jit/JITStubCall.h: 23171 (JSC::JITStubCall::addArgument): Always use { regT1, regT0 }, to facilitate 23172 chaining. 23173 23174 (JSC::JITStubCall::call): Unmap all mapped registers, since our callee 23175 stub might stomp them. 23176 231772009-07-01 Sam Weinig <sam@webkit.org> 23178 23179 Reviewed by Gavin Barraclough. 23180 23181 Don't reload values in emitBinaryDoubleOp. 23182 23183 SunSpider reports a 0.6% progression. 23184 23185 * jit/JIT.h: 23186 * jit/JITArithmetic.cpp: 23187 (JSC::JIT::emit_op_jnless): 23188 (JSC::JIT::emit_op_jnlesseq): 23189 (JSC::JIT::emitBinaryDoubleOp): 23190 231912009-07-01 Sam Weinig <sam@webkit.org> 23192 23193 Reviewed by Geoffrey Garen. 23194 23195 Convert op_div to load op1 and op2 up front. 23196 23197 * jit/JITArithmetic.cpp: 23198 (JSC::JIT::emit_op_div): 23199 232002009-07-01 Sam Weinig <sam@webkit.org> 23201 23202 Reviewed by Geoffrey Garen. 23203 23204 Don't emit code in emitBinaryDoubleOp if code is unreachable, observable 23205 via an empty (unlinked) jumplist passed in. This only effects op_jnless 23206 and op_jnlesseq at present. 23207 23208 * jit/JITArithmetic.cpp: 23209 (JSC::JIT::emitSlow_op_jnless): 23210 (JSC::JIT::emitSlow_op_jnlesseq): 23211 (JSC::JIT::emitBinaryDoubleOp): 23212 232132009-07-01 Geoffrey Garen <ggaren@apple.com> 23214 23215 Reviewed by Sam Weinig. 23216 23217 Converted op_mod to put { tag, payload } in { regT1, regT0 }, and 23218 tidied up its constant case. 23219 23220 SunSpider reports a 0.2% regression, but a micro-benchmark of op_mod 23221 shows a 12% speedup, and the SunSpider test that uses op_mod most should 23222 benefit a lot from result caching in the end, since it almost always 23223 performs (expression) % constant. 23224 23225 * jit/JITArithmetic.cpp: 23226 (JSC::JIT::emit_op_mod): 23227 (JSC::JIT::emitSlow_op_mod): 23228 232292009-06-30 Sam Weinig <sam@webkit.org> 23230 23231 Reviewed by Geoffrey Garen. 23232 23233 Converted some more arithmetic ops to put { tag, payload } in 23234 { regT1, regT0 }. 23235 23236 * jit/JITArithmetic.cpp: 23237 (JSC::JIT::emit_op_mul): 23238 (JSC::JIT::emitSlow_op_mul): 23239 232402009-06-30 Geoffrey Garen <ggaren@apple.com> 23241 23242 Reviewed by Sam Weinig. 23243 23244 Converted some more arithmetic ops to put { tag, payload } in 23245 { regT1, regT0 }, and added a case for subtract constant. 23246 23247 SunSpider says no change. v8 says 0.3% slower. 23248 23249 * jit/JIT.h: 23250 * jit/JITArithmetic.cpp: 23251 (JSC::JIT::emit_op_add): 23252 (JSC::JIT::emitAdd32Constant): 23253 (JSC::JIT::emitSlow_op_add): 23254 (JSC::JIT::emit_op_sub): 23255 (JSC::JIT::emitSub32Constant): 23256 (JSC::JIT::emitSlow_op_sub): 23257 232582009-06-30 Gavin Barraclough <barraclough@apple.com> 23259 23260 Reviewed by Sam Weinig. 23261 23262 Remove more uses of addressFor(), load double constants directly from 23263 the constantpool in the CodeBlock, rather than from the register file. 23264 23265 * jit/JITArithmetic.cpp: 23266 (JSC::JIT::emitAdd32Constant): 23267 (JSC::JIT::emitBinaryDoubleOp): 23268 232692009-06-30 Geoffrey Garen <ggaren@apple.com> 23270 23271 Reviewed by Sam Weinig. 23272 23273 Fixed a bug in postfix ops, where we would treat x = x++ and x = x-- 23274 as a no-op, even if x were not an int, and the ++/-- could have side-effects. 23275 23276 * jit/JITArithmetic.cpp: 23277 (JSC::JIT::emit_op_post_inc): 23278 (JSC::JIT::emitSlow_op_post_inc): 23279 (JSC::JIT::emit_op_post_dec): 23280 (JSC::JIT::emitSlow_op_post_dec): 23281 232822009-06-30 Geoffrey Garen <ggaren@apple.com> 23283 23284 Reviewed by Sam Weinig. 23285 23286 Converted some arithmetic ops to put { tag, payload } in 23287 { regT1, regT0 }. 23288 23289 SunSpider says 0.7% faster. v8 says no change. 23290 23291 * jit/JIT.h: 23292 * jit/JITArithmetic.cpp: 23293 (JSC::JIT::emit_op_jnless): 23294 (JSC::JIT::emit_op_jnlesseq): 23295 (JSC::JIT::emit_op_lshift): 23296 (JSC::JIT::emit_op_rshift): 23297 (JSC::JIT::emit_op_bitand): 23298 (JSC::JIT::emit_op_bitor): 23299 (JSC::JIT::emit_op_bitxor): 23300 * jit/JITInlineMethods.h: 23301 (JSC::JIT::isOperandConstantImmediateInt): 23302 (JSC::JIT::getOperandConstantImmediateInt): 23303 233042009-06-30 Gavin Barraclough <barraclough@apple.com> 23305 23306 Reviewed by Sam Weinig. 23307 23308 Start removing cases of addressFor(). 23309 23310 * jit/JIT.h: 23311 * jit/JITArithmetic.cpp: 23312 (JSC::JIT::emitAdd32Constant): 23313 (JSC::JIT::emitBinaryDoubleOp): 23314 (JSC::JIT::emit_op_div): 23315 * jit/JITInlineMethods.h: 23316 (JSC::JIT::emitLoadDouble): 23317 (JSC::JIT::emitLoadInt32ToDouble): 23318 (JSC::JIT::emitStoreDouble): 23319 * jit/JITOpcodes.cpp: 23320 (JSC::JIT::emit_op_jfalse): 23321 (JSC::JIT::emit_op_jtrue): 23322 233232009-06-30 Geoffrey Garen <ggaren@apple.com> 23324 23325 Rolled back in my last patch with regression fixed. 23326 23327 * jit/JIT.cpp: 23328 (JSC::JIT::privateCompileSlowCases): 23329 * jit/JIT.h: 23330 * jit/JITOpcodes.cpp: 23331 (JSC::JIT::emit_op_loop_if_less): 23332 (JSC::JIT::emit_op_loop_if_lesseq): 23333 (JSC::JIT::emit_op_resolve_global): 23334 (JSC::JIT::emitSlow_op_resolve_global): 23335 (JSC::JIT::emit_op_eq): 23336 (JSC::JIT::emitSlow_op_eq): 23337 (JSC::JIT::emit_op_neq): 23338 (JSC::JIT::emitSlow_op_neq): 23339 233402009-06-30 Geoffrey Garen <ggaren@apple.com> 23341 23342 Rolled out my last patch because it was a 2% SunSpider regression. 23343 23344 * jit/JIT.cpp: 23345 (JSC::JIT::privateCompileSlowCases): 23346 * jit/JIT.h: 23347 * jit/JITOpcodes.cpp: 23348 (JSC::JIT::emit_op_loop_if_less): 23349 (JSC::JIT::emit_op_loop_if_lesseq): 23350 (JSC::JIT::emit_op_resolve_global): 23351 (JSC::JIT::emit_op_eq): 23352 (JSC::JIT::emitSlow_op_eq): 23353 (JSC::JIT::emit_op_neq): 23354 (JSC::JIT::emitSlow_op_neq): 23355 233562009-06-30 Geoffrey Garen <ggaren@apple.com> 23357 23358 Reviewed by Gavin "Sam Weinig" Barraclough. 23359 23360 Standardized the rest of our opcodes to put { tag, payload } in 23361 { regT1, regT0 } where possible. 23362 23363 * jit/JIT.cpp: 23364 (JSC::JIT::privateCompileSlowCases): 23365 * jit/JIT.h: 23366 * jit/JITOpcodes.cpp: 23367 (JSC::JIT::emit_op_loop_if_less): 23368 (JSC::JIT::emit_op_loop_if_lesseq): 23369 (JSC::JIT::emit_op_resolve_global): 23370 (JSC::JIT::emitSlow_op_resolve_global): 23371 (JSC::JIT::emit_op_eq): 23372 (JSC::JIT::emitSlow_op_eq): 23373 (JSC::JIT::emit_op_neq): 23374 (JSC::JIT::emitSlow_op_neq): 23375 233762009-06-30 Gavin Barraclough <barraclough@apple.com> 23377 23378 Reviewed by Geoffrey Garen. 23379 23380 Replace calls to store32(tagFor()) and store32(payloadFor()) 23381 with emitStoreInt32(), emitStoreBool(), and emitStoreCell(). 23382 23383 * jit/JIT.h: 23384 * jit/JITArithmetic.cpp: 23385 (JSC::JIT::emit_op_negate): 23386 (JSC::JIT::emit_op_lshift): 23387 (JSC::JIT::emit_op_rshift): 23388 (JSC::JIT::emit_op_bitand): 23389 (JSC::JIT::emitBitAnd32Constant): 23390 (JSC::JIT::emit_op_bitor): 23391 (JSC::JIT::emitBitOr32Constant): 23392 (JSC::JIT::emit_op_bitxor): 23393 (JSC::JIT::emitBitXor32Constant): 23394 (JSC::JIT::emit_op_bitnot): 23395 (JSC::JIT::emit_op_post_inc): 23396 (JSC::JIT::emit_op_post_dec): 23397 (JSC::JIT::emit_op_pre_inc): 23398 (JSC::JIT::emit_op_pre_dec): 23399 (JSC::JIT::emit_op_add): 23400 (JSC::JIT::emitAdd32Constant): 23401 (JSC::JIT::emit_op_sub): 23402 (JSC::JIT::emitSub32ConstantLeft): 23403 (JSC::JIT::emitSub32ConstantRight): 23404 (JSC::JIT::emit_op_mul): 23405 (JSC::JIT::emitSlow_op_mul): 23406 (JSC::JIT::emit_op_div): 23407 (JSC::JIT::emit_op_mod): 23408 * jit/JITCall.cpp: 23409 (JSC::JIT::emit_op_load_varargs): 23410 * jit/JITInlineMethods.h: 23411 (JSC::JIT::emitStoreInt32): 23412 (JSC::JIT::emitStoreCell): 23413 (JSC::JIT::emitStoreBool): 23414 (JSC::JIT::emitStore): 23415 * jit/JITOpcodes.cpp: 23416 (JSC::JIT::emit_op_instanceof): 23417 (JSC::JIT::emit_op_not): 23418 (JSC::JIT::emit_op_eq): 23419 (JSC::JIT::emitSlow_op_eq): 23420 (JSC::JIT::emit_op_neq): 23421 (JSC::JIT::emitSlow_op_neq): 23422 (JSC::JIT::compileOpStrictEq): 23423 (JSC::JIT::emit_op_eq_null): 23424 (JSC::JIT::emit_op_neq_null): 23425 * jit/JITStubCall.h: 23426 (JSC::JITStubCall::call): 23427 234282009-06-30 Geoffrey Garen <ggaren@apple.com> 23429 23430 Reviewed by Sam Weinig. 23431 23432 Standardized the rest of the property access instructions to put { tag, 23433 payload } in { regT1, regT0 }. 23434 23435 Small v8 speedup, 0.2% SunSpider slowdown. 23436 23437 * jit/JIT.h: 23438 * jit/JITInlineMethods.h: 23439 (JSC::JIT::emitLoad): 23440 (JSC::JIT::emitLoad2): 23441 * jit/JITPropertyAccess.cpp: 23442 (JSC::JIT::emit_op_get_by_val): 23443 (JSC::JIT::emitSlow_op_get_by_val): 23444 (JSC::JIT::emit_op_put_by_val): 23445 (JSC::JIT::emitSlow_op_put_by_val): 23446 (JSC::JIT::emit_op_put_by_id): 23447 (JSC::JIT::emitSlow_op_put_by_id): 23448 (JSC::JIT::patchPutByIdReplace): 23449 234502009-06-29 Sam Weinig <sam@webkit.org> 23451 23452 Reviewed by Gavin Barraclough. 23453 23454 Various cleanups. 23455 - Use fpRegT* instead of X86::xmm*. 23456 - Use a switch statement in emitBinaryDoubleOp instead of a bunch of 23457 if/elses. 23458 23459 * jit/JITArithmetic.cpp: 23460 (JSC::JIT::emitAdd32Constant): 23461 (JSC::JIT::emitBinaryDoubleOp): 23462 (JSC::JIT::emit_op_div): 23463 234642009-06-29 Sam Weinig <sam@webkit.org> 23465 23466 Reviewed by Geoffrey Garen. 23467 23468 Add inline code dealing with doubles for op_jfalse and op_jtrue. 23469 23470 * assembler/MacroAssemblerX86Common.h: 23471 (JSC::MacroAssemblerX86Common::): 23472 (JSC::MacroAssemblerX86Common::zeroDouble): 23473 * jit/JITOpcodes.cpp: 23474 (JSC::JIT::emit_op_jfalse): 23475 (JSC::JIT::emit_op_jtrue): 23476 234772009-06-28 Geoffrey Garen <ggaren@apple.com> 23478 23479 Reviewed by Sam Weinig. 23480 23481 Standardized op_get_by_id to put { tag, payload } in { regT1, regT0 }. 23482 23483 SunSpider and v8 report maybe 0.2%-0.4% regressions, but the optimization 23484 this enables will win much more than that back. 23485 23486 * jit/JIT.cpp: 23487 (JSC::JIT::privateCompileCTIMachineTrampolines): 23488 * jit/JIT.h: 23489 * jit/JITPropertyAccess.cpp: 23490 (JSC::JIT::emit_op_method_check): 23491 (JSC::JIT::emit_op_get_by_id): 23492 (JSC::JIT::compileGetByIdHotPath): 23493 (JSC::JIT::compileGetByIdSlowCase): 23494 (JSC::JIT::patchGetByIdSelf): 23495 (JSC::JIT::privateCompilePatchGetArrayLength): 23496 (JSC::JIT::privateCompileGetByIdProto): 23497 (JSC::JIT::privateCompileGetByIdSelfList): 23498 (JSC::JIT::privateCompileGetByIdProtoList): 23499 (JSC::JIT::privateCompileGetByIdChainList): 23500 (JSC::JIT::privateCompileGetByIdChain): 23501 235022009-06-26 Geoffrey Garen <ggaren@apple.com> 23503 23504 Reviewed by Maciej Stachowiak. 23505 23506 Standardized op_call to put { tag, payload } in { regT1, regT0 }. 23507 23508 SunSpider and v8 report no change. 23509 23510 * jit/JIT.cpp: 23511 (JSC::JIT::privateCompileCTIMachineTrampolines): 23512 * jit/JITCall.cpp: 23513 (JSC::JIT::compileOpCallInitializeCallFrame): 23514 (JSC::JIT::compileOpCallSetupArgs): 23515 (JSC::JIT::compileOpConstructSetupArgs): 23516 (JSC::JIT::compileOpCallVarargsSetupArgs): 23517 (JSC::JIT::compileOpCallVarargs): 23518 (JSC::JIT::compileOpCall): 23519 (JSC::JIT::compileOpCallSlowCase): 23520 235212009-06-26 Sam Weinig <sam@webkit.org> 23522 23523 Reviewed by Geoffrey Garen. 23524 23525 Handle multiplying by zero a little better by 23526 inlining the case that both operands are non-negative 23527 into the slowpath. 23528 23529 * assembler/MacroAssemblerX86Common.h: 23530 (JSC::MacroAssemblerX86Common::branchOr32): 23531 * jit/JITArithmetic.cpp: 23532 (JSC::JIT::emit_op_mul): 23533 (JSC::JIT::emitSlow_op_mul): 23534 235352009-06-25 Geoffrey Garen <ggaren@apple.com> 23536 23537 Reviewed by Sam Weinig. 23538 23539 Optimize x++ to ++x inside for loops. 23540 23541 Sadly, no measurable speedup, but this should help with result chaining. 23542 23543 * parser/Nodes.cpp: 23544 (JSC::ForNode::emitBytecode): 23545 235462009-06-25 Geoffrey Garen <ggaren@apple.com> 23547 23548 Reviewed by Sam Weinig. 23549 23550 Standardized some more opcodes to put { tag, payload } in { regT1, regT0 }. 23551 23552 * jit/JITArithmetic.cpp: 23553 (JSC::JIT::emitSlow_op_bitnot): 23554 (JSC::JIT::emit_op_post_inc): 23555 235562009-06-25 Geoffrey Garen <ggaren@apple.com> 23557 23558 Reviewed by Sam Weinig. 23559 23560 Standardized some more opcodes to put { tag, payload } in { regT1, regT0 }. 23561 23562 * jit/JITArithmetic.cpp: 23563 (JSC::JIT::emit_op_bitnot): 23564 (JSC::JIT::emit_op_post_dec): 23565 (JSC::JIT::emit_op_pre_inc): 23566 (JSC::JIT::emitSlow_op_pre_inc): 23567 (JSC::JIT::emit_op_pre_dec): 23568 (JSC::JIT::emitSlow_op_pre_dec): 23569 235702009-06-25 Geoffrey Garen <ggaren@apple.com> 23571 23572 Reviewed by Sam Weinig. 23573 23574 Standardized some more opcodes to put { tag, payload } in { regT1, regT0 }. 23575 23576 * jit/JITArithmetic.cpp: 23577 (JSC::JIT::emit_op_negate): 23578 (JSC::JIT::emitSlow_op_negate): 23579 * jit/JITCall.cpp: 23580 (JSC::JIT::emit_op_construct_verify): 23581 (JSC::JIT::emitSlow_op_construct_verify): 23582 235832009-06-25 Geoffrey Garen <ggaren@apple.com> 23584 23585 Reviewed by Sam Weinig. 23586 23587 Standardized some more opcodes to put { tag, payload } in { regT1, regT0 }. 23588 23589 * jit/JITOpcodes.cpp: 23590 (JSC::JIT::emit_op_loop_if_true): 23591 (JSC::JIT::emit_op_jfalse): 23592 (JSC::JIT::emit_op_jtrue): 23593 (JSC::JIT::emit_op_jeq_null): 23594 (JSC::JIT::emit_op_jneq_null): 23595 (JSC::JIT::emit_op_eq_null): 23596 (JSC::JIT::emit_op_neq_null): 23597 235982009-06-25 Geoffrey Garen <ggaren@apple.com> 23599 23600 Reviewed by Sam Weinig (sort of, maybe). 23601 23602 Fixed some ASSERTs in http/tests/security. 23603 23604 These ASSERTs were introduced by http://trac.webkit.org/changeset/45057, 23605 but the underlying problem was actually older. http://trac.webkit.org/changeset/45057 23606 just exposed the problem by enabling optimization in more cases. 23607 23608 The ASSERTs fired because we tested PropertySlot::slotBase() for validity, 23609 but slotBase() ASSERTs if it's invalid, so we would ASSERT before 23610 the test could happen. Solution: Remove the ASSERT. Maybe it was valid 23611 once, but it clearly goes against a pattern we've deployed of late. 23612 23613 The underlying problem was that WebCore would re-use a PropertySlot in 23614 the case of a forwarding access, and the second use would not completely 23615 overwrite the first use. Solution: Make sure to overwrite m_offset when 23616 setting a value on a PropertySlot. (Other values already get implicitly 23617 overwritten during reuse.) 23618 23619 * runtime/PropertySlot.h: 23620 (JSC::PropertySlot::PropertySlot): 23621 (JSC::PropertySlot::setValueSlot): 23622 (JSC::PropertySlot::setValue): 23623 (JSC::PropertySlot::setRegisterSlot): 23624 (JSC::PropertySlot::setUndefined): 23625 (JSC::PropertySlot::slotBase): 23626 (JSC::PropertySlot::clearOffset): 23627 236282009-06-24 Gavin Barraclough <barraclough@apple.com> 23629 23630 Reviewed by Geoff Garen. 23631 23632 Enable JIT_OPTIMIZE_METHOD_CALLS on the branch, implementation matches current implemenatation in ToT. 23633 23634 * jit/JIT.h: 23635 * jit/JITPropertyAccess.cpp: 23636 (JSC::JIT::emit_op_method_check): 23637 (JSC::JIT::emitSlow_op_method_check): 23638 (JSC::JIT::emit_op_get_by_id): 23639 (JSC::JIT::compileGetByIdHotPath): 23640 (JSC::JIT::emitSlow_op_get_by_id): 23641 (JSC::JIT::compileGetByIdSlowCase): 23642 236432009-06-23 Geoffrey Garen <ggaren@apple.com> 23644 23645 Reviewed by Sam Weinig. 23646 23647 Bit off a tiny bit more of standardizing opcode behavior to help with result 23648 caching. 23649 23650 SunSpider reports no change, v8 maybe a tiny speedup. 23651 23652 * jit/JITOpcodes.cpp: 23653 (JSC::JIT::emit_op_to_jsnumber): 23654 (JSC::JIT::emitSlow_op_to_jsnumber): 23655 (JSC::JIT::emit_op_convert_this): 23656 (JSC::JIT::emitSlow_op_convert_this): 23657 236582009-06-23 Geoffrey Garen <ggaren@apple.com> 23659 23660 Reviewed by Sam Weinig. 23661 23662 Bit off a tiny bit more of standardizing opcode behavior to help with result 23663 caching -- including removing my old enemy, op_resolve_function, because 23664 it was non-standard, and removing it felt better than helping it limp along. 23665 23666 SunSpider reports no change, v8 maybe a tiny speedup. 23667 23668 * bytecode/CodeBlock.cpp: 23669 (JSC::CodeBlock::dump): 23670 * bytecode/Opcode.h: 23671 * bytecompiler/BytecodeGenerator.cpp: 23672 * bytecompiler/BytecodeGenerator.h: 23673 * interpreter/Interpreter.cpp: 23674 (JSC::Interpreter::privateExecute): 23675 * jit/JIT.cpp: 23676 (JSC::JIT::privateCompileMainPass): 23677 * jit/JIT.h: 23678 * jit/JITOpcodes.cpp: 23679 (JSC::JIT::emit_op_get_scoped_var): 23680 (JSC::JIT::emit_op_put_scoped_var): 23681 (JSC::JIT::emit_op_to_primitive): 23682 (JSC::JIT::emitSlow_op_to_primitive): 23683 * jit/JITStubs.cpp: 23684 * jit/JITStubs.h: 23685 * parser/Nodes.cpp: 23686 (JSC::FunctionCallResolveNode::emitBytecode): 23687 236882009-06-23 Geoffrey Garen <ggaren@apple.com> 23689 23690 Reviewed by Sam Weinig. 23691 23692 Bit off a tiny bit of standardizing opcode behavior to help with result 23693 caching. 23694 23695 0.6% SunSpider speedup. 0.3% v8 speedup. 23696 23697 * jit/JITInlineMethods.h: 23698 (JSC::JIT::emitLoad): Accomodate a base register that overlaps with payload 23699 by loading tag before payload, to avoid stomping base/payload. 23700 23701 * jit/JITOpcodes.cpp: 23702 (JSC::JIT::emit_op_mov): Abide by the standard "tag in regT1, payload in 23703 regT0" semantics. 23704 23705 (JSC::JIT::emit_op_get_global_var): 23706 (JSC::JIT::emit_op_put_global_var): Ditto. Also, removed some irrelevent 23707 loads while I was at it. The global object's "d" pointer never changes 23708 after construction. 23709 237102009-06-23 Gavin Barraclough <barraclough@apple.com> 23711 23712 Reviewed by Sam Weinig. 23713 23714 Remove 'arguments' field from Register union (again). 23715 This time do so without breaking tests (radical, I know). 23716 23717 * interpreter/CallFrame.h: 23718 (JSC::ExecState::optionalCalleeArguments): 23719 (JSC::ExecState::setArgumentCount): 23720 (JSC::ExecState::init): 23721 * interpreter/Interpreter.cpp: 23722 (JSC::Interpreter::dumpRegisters): 23723 (JSC::Interpreter::unwindCallFrame): 23724 (JSC::Interpreter::privateExecute): 23725 (JSC::Interpreter::retrieveArguments): 23726 * interpreter/Register.h: 23727 (JSC::Register::withInt): 23728 (JSC::Register::): 23729 (JSC::Register::Register): 23730 (JSC::Register::i): 23731 * jit/JITStubs.cpp: 23732 (JSC::JITStubs::cti_op_tear_off_arguments): 23733 * runtime/Arguments.h: 23734 (JSC::JSActivation::copyRegisters): 23735 (JSC::Register::arguments): 23736 * runtime/JSActivation.cpp: 23737 (JSC::JSActivation::argumentsGetter): 23738 * runtime/JSActivation.h: 23739 237402009-06-23 Geoffrey Garen <ggaren@apple.com> 23741 23742 Reviewed by Sam Weinig. 23743 23744 Removed some result register tracking cruft in preparation for a new 23745 result tracking mechanism. 23746 23747 SunSpider reports no change. 23748 23749 * assembler/AbstractMacroAssembler.h: 23750 * assembler/X86Assembler.h: 23751 (JSC::X86Assembler::JmpDst::JmpDst): No need to track jump targets in 23752 machine code; we already do this in bytecode. 23753 23754 * jit/JIT.cpp: 23755 (JSC::JIT::JIT): 23756 (JSC::JIT::emitTimeoutCheck): Make sure to save and restore the result 23757 registers, so an opcode with a timeout check can still benefit from result 23758 register caching. 23759 23760 (JSC::JIT::privateCompileMainPass): 23761 (JSC::JIT::privateCompileSlowCases): Removed calls to killLastResultRegister() 23762 in preparation for something new. 23763 23764 * jit/JIT.h: 23765 * jit/JITArithmetic.cpp: 23766 (JSC::JIT::emit_op_jnless): 23767 (JSC::JIT::emit_op_jnlesseq): 23768 * jit/JITInlineMethods.h: 23769 (JSC::JIT::emitGetFromCallFrameHeaderPtr): 23770 (JSC::JIT::emitGetFromCallFrameHeader32): 23771 * jit/JITOpcodes.cpp: 23772 (JSC::JIT::emit_op_jmp): 23773 (JSC::JIT::emit_op_jfalse): 23774 (JSC::JIT::emit_op_jtrue): 23775 (JSC::JIT::emit_op_jeq_null): 23776 (JSC::JIT::emit_op_jneq_null): 23777 (JSC::JIT::emit_op_jneq_ptr): 23778 (JSC::JIT::emit_op_jsr): 23779 (JSC::JIT::emit_op_sret): 23780 (JSC::JIT::emit_op_jmp_scopes): ditto 23781 23782 * jit/JITStubCall.h: 23783 (JSC::JITStubCall::JITStubCall): 23784 (JSC::JITStubCall::getArgument): added a mechanism for reloading an argument 23785 you passed to a JIT stub, for use in emitTimeoutCheck. 23786 237872009-06-23 Sam Weinig <sam@webkit.org> 23788 23789 Reviewed by Geoffrey Garen. 23790 23791 Remove now-useless inplace variants of binary ops. 23792 23793 * jit/JIT.h: 23794 * jit/JITArithmetic.cpp: 23795 (JSC::JIT::emit_op_bitand): 23796 (JSC::JIT::emit_op_bitor): 23797 (JSC::JIT::emit_op_bitxor): 23798 (JSC::JIT::emit_op_add): 23799 (JSC::JIT::emit_op_sub): 23800 (JSC::JIT::emit_op_mul): 23801 238022009-06-23 Sam Weinig <sam@webkit.org> 23803 23804 Reviewed by Geoffrey Garen. 23805 23806 Move off memory operands to aid in re-enabling result caching. 23807 23808 - No regression measured. 23809 23810 * jit/JIT.h: 23811 * jit/JITArithmetic.cpp: 23812 (JSC::JIT::emit_op_negate): 23813 (JSC::JIT::emit_op_jnless): 23814 (JSC::JIT::emit_op_jnlesseq): 23815 (JSC::JIT::emit_op_lshift): 23816 (JSC::JIT::emit_op_rshift): 23817 (JSC::JIT::emit_op_bitand): 23818 (JSC::JIT::emitBitAnd32Constant): 23819 (JSC::JIT::emitBitAnd32InPlace): 23820 (JSC::JIT::emit_op_bitor): 23821 (JSC::JIT::emitBitOr32Constant): 23822 (JSC::JIT::emitBitOr32InPlace): 23823 (JSC::JIT::emit_op_bitxor): 23824 (JSC::JIT::emitBitXor32Constant): 23825 (JSC::JIT::emitBitXor32InPlace): 23826 (JSC::JIT::emit_op_bitnot): 23827 (JSC::JIT::emit_op_post_inc): 23828 (JSC::JIT::emit_op_post_dec): 23829 (JSC::JIT::emit_op_pre_inc): 23830 (JSC::JIT::emitSlow_op_pre_inc): 23831 (JSC::JIT::emit_op_pre_dec): 23832 (JSC::JIT::emitSlow_op_pre_dec): 23833 (JSC::JIT::emit_op_add): 23834 (JSC::JIT::emitAdd32Constant): 23835 (JSC::JIT::emitAdd32InPlace): 23836 (JSC::JIT::emitSlow_op_add): 23837 (JSC::JIT::emitSlowAdd32Constant): 23838 (JSC::JIT::emit_op_sub): 23839 (JSC::JIT::emitSlow_op_sub): 23840 (JSC::JIT::emitSub32ConstantLeft): 23841 (JSC::JIT::emitSub32ConstantRight): 23842 (JSC::JIT::emitSub32InPlaceLeft): 23843 (JSC::JIT::emitSub32InPlaceRight): 23844 (JSC::JIT::emitBinaryDoubleOp): 23845 (JSC::JIT::emit_op_mul): 23846 (JSC::JIT::emitMul32InPlace): 23847 (JSC::JIT::emit_op_div): 23848 (JSC::JIT::emit_op_mod): 23849 * jit/JITCall.cpp: 23850 (JSC::JIT::compileOpCallVarargs): 23851 * jit/JITOpcodes.cpp: 23852 (JSC::JIT::emit_op_loop_if_less): 23853 (JSC::JIT::emit_op_loop_if_lesseq): 23854 (JSC::JIT::emit_op_instanceof): 23855 (JSC::JIT::emit_op_to_primitive): 23856 (JSC::JIT::emit_op_not): 23857 (JSC::JIT::emit_op_jneq_ptr): 23858 (JSC::JIT::emit_op_eq): 23859 (JSC::JIT::emit_op_neq): 23860 (JSC::JIT::emit_op_to_jsnumber): 23861 * jit/JITPropertyAccess.cpp: 23862 (JSC::JIT::emit_op_get_by_val): 23863 (JSC::JIT::emit_op_put_by_val): 23864 238652009-06-23 Geoffrey Garen <ggaren@apple.com> 23866 23867 Reviewed by Sam Weinig. 23868 23869 Fixed some missing and/or misplaced labels in bytecode generation, so 23870 we don't have to work around them in JIT code generation. 23871 23872 * bytecompiler/BytecodeGenerator.cpp: 23873 (JSC::BytecodeGenerator::emitJumpSubroutine): 23874 * parser/Nodes.cpp: 23875 (JSC::TryNode::emitBytecode): 23876 238772009-06-22 Geoffrey Garen <ggaren@apple.com> 23878 23879 Reviewed by Sam Weinig. 23880 23881 For member function calls, emit "this" directly into the "this" slot 23882 for the function call, instead of moving it there later. This reduces 23883 time spent in op_mov during certain calls, like "a.b.c()". 23884 23885 1%-2% speedup on v8, mostly richards and delta-blue. 23886 23887 * parser/Nodes.cpp: 23888 (JSC::FunctionCallDotNode::emitBytecode): 23889 238902009-06-22 Gavin Barraclough <barraclough@apple.com> 23891 23892 Reviewed by Sam Weinig. 23893 23894 Remove 'arguments' field from Register union. Having JSCell derived types in the union is 23895 dangerous since it opens the possibility for the field to be written as a raw pointer but 23896 then read as a JSValue. This will lead to statle data being read for the tag, which may 23897 be dangerous. Having removed Arguments* types form Register, all arguments objects must 23898 always explicitly be stored in the register file as JSValues. 23899 23900 * interpreter/CallFrame.h: 23901 (JSC::ExecState::optionalCalleeArguments): 23902 * interpreter/Interpreter.cpp: 23903 (JSC::Interpreter::unwindCallFrame): 23904 (JSC::Interpreter::privateExecute): 23905 (JSC::Interpreter::retrieveArguments): 23906 * interpreter/Register.h: 23907 (JSC::Register::): 23908 * jit/JITStubs.cpp: 23909 (JSC::JITStubs::cti_op_tear_off_arguments): 23910 * runtime/Arguments.h: 23911 (JSC::JSActivation::copyRegisters): 23912 * runtime/JSActivation.cpp: 23913 (JSC::JSActivation::argumentsGetter): 23914 * runtime/JSActivation.h: 23915 239162009-06-03 Sam Weinig <sam@webkit.org> 23917 23918 Reviewed by Geoffrey Garen. 23919 23920 Add back known this value optimization by abstracting 23921 slow case if not JSCell jumps. 23922 23923 * jit/JIT.h: 23924 * jit/JITCall.cpp: 23925 (JSC::JIT::compileOpCallVarargs): 23926 (JSC::JIT::compileOpCallVarargsSlowCase): 23927 (JSC::JIT::compileOpCall): 23928 (JSC::JIT::compileOpCallSlowCase): 23929 * jit/JITInlineMethods.h: 23930 (JSC::JIT::emitJumpSlowCaseIfNotJSCell): 23931 (JSC::JIT::linkSlowCaseIfNotJSCell): 23932 * jit/JITOpcodes.cpp: 23933 (JSC::JIT::emit_op_instanceof): 23934 (JSC::JIT::emitSlow_op_instanceof): 23935 * jit/JITPropertyAccess.cpp: 23936 (JSC::JIT::emit_op_get_by_val): 23937 (JSC::JIT::emitSlow_op_get_by_val): 23938 (JSC::JIT::emit_op_put_by_val): 23939 (JSC::JIT::emitSlow_op_put_by_val): 23940 (JSC::JIT::emit_op_get_by_id): 23941 (JSC::JIT::emitSlow_op_get_by_id): 23942 (JSC::JIT::emit_op_put_by_id): 23943 (JSC::JIT::emitSlow_op_put_by_id): 23944 239452009-06-01 Geoffrey Garen <ggaren@apple.com> 23946 23947 Reviewed by Sam Weinig. 23948 23949 Fixed some of the regression in crypto-aes.js. (8.5% speedup in 23950 crypto-aes.js.) 23951 23952 SunSpider reports no change overall. 23953 23954 Division was producing double results, which took the slow path through 23955 array access code. 23956 23957 Strangely, all my attempts at versions of this patch that modified array 23958 access code to accept ints encoded as doubles along the fast or slow paths 23959 were regressions. So I did this instead. 23960 23961 * jit/JITArithmetic.cpp: 23962 (JSC::JIT::emit_op_div): When dividing an int by an int, go ahead and try 23963 to turn the result into an int. Don't just do int division, though, because 23964 testing shows it to be slower than SSE double division, and the corner 23965 cases are pretty complicated / lengthy on top of that. Also, don't try 23966 to canonicalize division of known tiny numerators into ints, since that's a 23967 waste of time. 23968 239692009-05-26 Geoffrey Garen <ggaren@apple.com> 23970 23971 Reviewed by Oliver Hunt. 23972 23973 Fixed a regression caused by my recent fix for NaN. 23974 23975 * jit/JITArithmetic.cpp: 23976 (JSC::JIT::emitBinaryDoubleOp): Actually do the comparison in reverse 23977 order, like the ChangeLog said we would, bokay? 23978 239792009-05-26 Geoffrey Garen <ggaren@apple.com> 23980 23981 Reviewed by Sam Weinig and Oliver Hunt. 23982 23983 Fixed two edge cases in %: 23984 23985 - Don't do -2147483648 % x as a fast case, since you might do -2147483648 % -1, 23986 which will signal a hardware exception due to overflow. 23987 23988 - In the case of a zero remainder, be sure to store negative zero if the 23989 dividend was zero. 23990 23991 SunSpider reports no change. 23992 23993 * jit/JITArithmetic.cpp: 23994 (JSC::JIT::emit_op_mod): 23995 (JSC::JIT::emitSlow_op_mod): 23996 239972009-05-25 Geoffrey Garen <ggaren@apple.com> 23998 23999 Reviewed by Maciej Stachowiak. 24000 24001 Fixed a regression when comparing to NaN. 24002 24003 * jit/JITArithmetic.cpp: 24004 (JSC::JIT::emitBinaryDoubleOp): For op_jnless and op_jnless_eq, do the 24005 comparison in reverse order, and jump if the result is below or 24006 below-or-equal. This ensures that we do jump in the case of NaN. 24007 240082009-05-25 Geoffrey Garen <ggaren@apple.com> 24009 24010 Reviewed by Oliver Hunt. 24011 24012 SunSpider says no change. 24013 24014 Fixed regressions in fast/js/var-declarations-shadowing.html and 24015 fast/js/equality.html, caused by recent == and != optimizations. 24016 24017 * jit/JITStubs.cpp: 24018 (JSC::JITStubs::cti_op_eq): Don't treat "compare to string" as always 24019 numeric or string comparison. If the second operand is an object, you 24020 need to ToPrimitive it, and start all over again. Also, I wrote out each 24021 of the possible cases explicitly, to cut down on redundant branching. 24022 240232009-05-25 Sam Weinig <sam@webkit.org> 24024 24025 Reviewed by Mark Rowe. 24026 24027 Fix bug in fast/js/constant-folding.html where we were not negating 24028 -0 properly. 24029 24030 * jit/JITArithmetic.cpp: 24031 (JSC::JIT::emit_op_negate): 24032 240332009-05-23 Geoffrey Garen <ggaren@apple.com> 24034 24035 Reviewed by Oliver Hunt. 24036 24037 Refactored new slow case codegen for == and !=. 24038 24039 SunSpider reports no change, maybe a tiny speedup. 24040 24041 * jit/JITOpcodes.cpp: 24042 (JSC::JIT::emitSlow_op_eq): 24043 (JSC::JIT::emitSlow_op_neq): Made a vptr comparison a *Ptr operation, 24044 instead of *32, to make it portable to 64bit. Reorganized the string 24045 and generic cases to make their control flow a little clearer. 24046 240472009-05-23 Geoffrey Garen <ggaren@apple.com> 24048 24049 Reviewed by Maciej Stachowiak. 24050 24051 Optimized == and != for our new value representation -- especially for strings. 24052 24053 14% speedup on date-format-tofte. 24054 24055 * jit/JITOpcodes.cpp: 24056 (JSC::JIT::emit_op_eq): 24057 (JSC::JIT::emitSlow_op_eq): 24058 (JSC::JIT::emit_op_neq): 24059 (JSC::JIT::emitSlow_op_neq): 24060 * jit/JITStubCall.h: 24061 (JSC::JITStubCall::JITStubCall): 24062 * jit/JITStubs.cpp: 24063 (JSC::JITStubs::cti_op_eq): 24064 (JSC::JITStubs::cti_op_eq_strings): 24065 (JSC::JITStubs::cti_op_call_eval): 24066 * jit/JITStubs.h: 24067 (JSC::): 24068 * runtime/JSValue.h: 24069 240702009-05-22 Sam Weinig <sam@webkit.org> 24071 24072 Reviewed by Gavin Barraclough. 24073 24074 Fix non-SSE enabled builds. 24075 24076 * jit/JITArithmetic.cpp: 24077 (JSC::JIT::emitSlow_op_add): Don't early return here, we still need to call the JIT stub. 24078 (JSC::JIT::emitSlow_op_sub): Ditto. 24079 240802009-05-22 Geoffrey Garen <ggaren@apple.com> 24081 24082 Reviewed by Sam Weinig. 24083 24084 Here's a thought: let's not take a jit stub call just to multiply by 1, 24085 bokay? 24086 24087 imul doesn't set the zero flag, so to test for a zero result, we need 24088 an explicit instruction. (Luckily, it does set the overflow flag, so 24089 we can still use that.) 24090 24091 * jit/JIT.h: 24092 * jit/JITArithmetic.cpp: 24093 (JSC::JIT::emit_op_mul): 24094 (JSC::JIT::emitSlow_op_mul): 24095 (JSC::JIT::emitMul32InPlace): 24096 240972009-05-22 Sam Weinig <sam@webkit.org> 24098 24099 Reviewed by Geoffrey "Premature Commit" Garen. 24100 24101 Add back constant integer cases for op_add. 24102 24103 * jit/JIT.h: 24104 * jit/JITArithmetic.cpp: 24105 (JSC::JIT::emit_op_add): 24106 (JSC::JIT::emitAdd32Constant): 24107 (JSC::JIT::emitSlow_op_add): 24108 (JSC::JIT::emitSlowAdd32Constant): 24109 * jit/JITInlineMethods.h: 24110 (JSC::JIT::getConstantOperandImmediateDouble): 24111 (JSC::JIT::isOperandConstantImmediateDouble): 24112 241132009-05-22 Geoffrey Garen <ggaren@apple.com> 24114 24115 Reviewed by Sam Weinig. 24116 24117 Added fast double cases for op_jnless and op_jnlesseq. 24118 24119 * assembler/AbstractMacroAssembler.h: 24120 (JSC::AbstractMacroAssembler::JumpList::jumps): New accesor, used by 24121 addSlowCase. 24122 24123 * assembler/X86Assembler.h: 24124 (JSC::X86Assembler::ucomisd_rm): New method for comparing register to 24125 memory. 24126 24127 * jit/JIT.h: 24128 * jit/JITArithmetic.cpp: 24129 (JSC::JIT::emit_op_jnless): 24130 (JSC::JIT::emitSlow_op_jnless): 24131 (JSC::JIT::emit_op_jnlesseq): 24132 (JSC::JIT::emitSlow_op_jnlesseq): 24133 (JSC::JIT::emit_op_add): 24134 (JSC::JIT::emit_op_sub): 24135 (JSC::JIT::emitBinaryDoubleOp): 24136 (JSC::JIT::emit_op_mul): 24137 (JSC::JIT::emit_op_div): Modified emitBinaryDoubleOp to accept comparison/jump 24138 operations in addition to operations with explicit result registers. 24139 24140 * jit/JITInlineMethods.h: 24141 (JSC::JIT::addSlowCase): Added an "addSlowCase" for JumpLists, so clients 24142 can track multiple jumps to the same slow case condition together. 24143 241442009-05-21 Sam Weinig <sam@webkit.org> 24145 24146 Reviewed by Gavin Barraclough. 24147 24148 Implement op_negate inline fast cases. 24149 24150 * assembler/MacroAssemblerX86Common.h: 24151 (JSC::MacroAssemblerX86Common::neg32): 24152 * assembler/X86Assembler.h: 24153 (JSC::X86Assembler::): 24154 (JSC::X86Assembler::negl_m): 24155 (JSC::X86Assembler::xorpd_rr): 24156 * jit/JIT.cpp: 24157 (JSC::JIT::privateCompileMainPass): 24158 (JSC::JIT::privateCompileSlowCases): 24159 * jit/JIT.h: 24160 * jit/JITArithmetic.cpp: 24161 (JSC::JIT::emit_op_negate): 24162 (JSC::JIT::emitSlow_op_negate): 24163 241642009-05-20 Sam Weinig <sam@webkit.org> 24165 24166 Reviewed by Gavin Barraclough. 24167 24168 Update the patchOffsetGetByIdSlowCaseCall constant for the 24169 case that OPCODE_SAMPLING is enabled. 24170 24171 * jit/JIT.h: 24172 241732009-05-20 Geoffrey Garen <ggaren@apple.com> 24174 24175 Reviewed by Sam Weinig. 24176 24177 Added support for inline subtraction of doubles. 24178 24179 * jit/JITArithmetic.cpp: 24180 (JSC::JIT::emit_op_sub): 24181 (JSC::JIT::emitSlow_op_sub): 24182 (JSC::JIT::emitSlowSub32InPlaceLeft): 24183 (JSC::JIT::emitBinaryDoubleOp): 24184 241852009-05-20 Sam Weinig <sam@webkit.org> 24186 24187 Reviewed by Geoffrey Garen. 24188 24189 Added support for inline division. 24190 24191 * assembler/X86Assembler.h: 24192 (JSC::X86Assembler::): 24193 (JSC::X86Assembler::divsd_rr): 24194 (JSC::X86Assembler::divsd_mr): 24195 * bytecode/CodeBlock.cpp: 24196 (JSC::CodeBlock::dump): 24197 * bytecode/Opcode.h: 24198 * bytecompiler/BytecodeGenerator.cpp: 24199 (JSC::BytecodeGenerator::emitBinaryOp): 24200 * interpreter/Interpreter.cpp: 24201 (JSC::Interpreter::privateExecute): 24202 * jit/JIT.cpp: 24203 (JSC::JIT::privateCompileMainPass): 24204 (JSC::JIT::privateCompileSlowCases): 24205 * jit/JIT.h: 24206 * jit/JITArithmetic.cpp: 24207 (JSC::JIT::emitBinaryDoubleOp): 24208 (JSC::JIT::emit_op_div): 24209 (JSC::JIT::emitSlow_op_div): 24210 242112009-05-20 Geoffrey Garen <ggaren@apple.com> 24212 24213 Reviewed by Sam Weinig. 24214 24215 Added support for inline addition of doubles. 24216 24217 * jit/JITArithmetic.cpp: 24218 (JSC::JIT::emit_op_add): 24219 (JSC::JIT::emitSlow_op_add): 24220 (JSC::JIT::emitSlowAdd32InPlace): 24221 (JSC::JIT::emitBinaryDoubleOp): 24222 (JSC::JIT::emit_op_mul): 24223 (JSC::JIT::emitSlow_op_mul): 24224 242252009-05-20 Geoffrey Garen <ggaren@apple.com> 24226 24227 Reviewed by Sam Weinig. 24228 24229 Factored inline double operations into a helper function, so that we 24230 can reuse this code for other math operations. 24231 24232 * jit/JIT.h: 24233 * jit/JITArithmetic.cpp: 24234 (JSC::JIT::emitBinaryDoubleOp): 24235 (JSC::JIT::emit_op_mul): 24236 * jit/JITCall.cpp: 24237 (JSC::JIT::compileOpCallInitializeCallFrame): 24238 242392009-05-20 Geoffrey Garen <ggaren@apple.com> 24240 24241 Reviewed by Sam Weinig. 24242 24243 Added support for inline multiplication of doubles. 24244 24245 * assembler/X86Assembler.h: 24246 (JSC::X86Assembler::cvtsi2sd_mr): New function, useful for loading an 24247 int32 into a double register. 24248 24249 * jit/JITArithmetic.cpp: 24250 (JSC::JIT::emit_op_mul): 24251 (JSC::JIT::emitSlow_op_mul): Filled out these cases for double arithmetic. 24252 24253 * jit/JIT.h: 24254 * jit/JITInlineMethods.h: 24255 (JSC::JIT::addressFor): New function, useful for addressing a JSValue's 24256 full 64bits as a double. 24257 242582009-05-19 Sam Weinig <sam@webkit.org> 24259 24260 Reviewed by Geoffrey Garen. 24261 24262 Implement and enable optimized calls. 24263 24264 * jit/JIT.cpp: 24265 (JSC::JIT::privateCompileCTIMachineTrampolines): Add ENABLE(JIT_OPTIMIZE_CALL) guards 24266 around the the optimize call only trampolines (virtualCallPreLink and virtualCallLink). 24267 Update the trampolines to account for the new JSValue representation. 24268 (JSC::JIT::unlinkCall): Use NULL instead of JSValue noValue. 24269 24270 * jit/JITCall.cpp: 24271 (JSC::JIT::compileOpCall): Update to account for the new JSValue representation 24272 (JSC::JIT::compileOpCallSlowCase): Ditto. 24273 24274 * jit/JITStubs.h: Remove incorrect !ENABLE(JIT_OPTIMIZE_CALL) guard. 24275 24276 * wtf/Platform.h: Enable ENABLE_JIT_OPTIMIZE_CALL. 24277 242782009-05-19 Sam Weinig <sam@webkit.org> 24279 24280 Reviewed by Geoffrey Garen. 24281 24282 Implement and enable optimized property access. 24283 24284 * assembler/AbstractMacroAssembler.h: Fix comment. 24285 * jit/JIT.cpp: 24286 (JSC::JIT::privateCompileCTIMachineTrampolines): Remove array length trampoline 24287 and implement the string length trampoline. 24288 * jit/JIT.h: Add new constants for patch offsets. 24289 * jit/JITInlineMethods.h: Remove FIELD_OFFSET which is now in StdLibExtras.h. 24290 * jit/JITPropertyAccess.cpp: 24291 (JSC::JIT::emit_op_get_by_id): 24292 (JSC::JIT::emitSlow_op_get_by_id): 24293 (JSC::JIT::emit_op_put_by_id): 24294 (JSC::JIT::emitSlow_op_put_by_id): 24295 (JSC::JIT::compilePutDirectOffset): 24296 (JSC::JIT::compileGetDirectOffset): 24297 (JSC::JIT::privateCompilePutByIdTransition): 24298 (JSC::JIT::patchGetByIdSelf): 24299 (JSC::JIT::patchPutByIdReplace): 24300 (JSC::JIT::privateCompilePatchGetArrayLength): 24301 (JSC::JIT::privateCompileGetByIdProto): 24302 (JSC::JIT::privateCompileGetByIdSelfList): 24303 (JSC::JIT::privateCompileGetByIdProtoList): 24304 (JSC::JIT::privateCompileGetByIdChainList): 24305 (JSC::JIT::privateCompileGetByIdChain): 24306 * jit/JITStubCall.h: 24307 (JSC::JITStubCall::addArgument): Add version of addArgument that takes 24308 two registers for the tag and payload. 24309 * jit/JITStubs.cpp: 24310 (JSC::JITStubs::JITStubs): Remove array length trampoline pointer. 24311 (JSC::JITStubs::cti_op_get_by_id_self_fail): 24312 * jit/JITStubs.h: 24313 * runtime/JSObject.h: 24314 (JSC::JSObject::JSObject): Move m_inheritorID below the property storage 24315 to align it to a 16 byte boundary. 24316 * wtf/Platform.h: Enable ENABLE_JIT_OPTIMIZE_PROPERTY_ACCESS 24317 * wtf/StdLibExtras.h: Move FIELD_OFFSET here. 24318 243192009-05-17 Sam Weinig <sam@webkit.org> 24320 24321 Reviewed by Geoffrey Garen. 24322 24323 Remove unneeded ExecState parameter from the number JSValue constructors. 24324 24325 * runtime/JSValue.h: 24326 (JSC::jsNumber): 24327 (JSC::jsNaN): 24328 (JSC::JSValue::JSValue): 24329 243302009-05-15 Sam Weinig <sam@webkit.org> 24331 24332 Reviewed by Geoffrey Garen. 24333 24334 Implemented fast path for op_put_by_val when putting to arrays. 24335 24336 * jit/JITPropertyAccess.cpp: 24337 (JSC::JIT::emit_op_put_by_val): 24338 (JSC::JIT::emitSlow_op_put_by_val): 24339 243402009-05-15 Geoffrey Garen <ggaren@apple.com> (Mostly by Sam) 24341 24342 Reviewed by Sam Weinig. 24343 24344 Implemented fast path for op_get_by_val when accessing array. 24345 24346 * jit/JIT.cpp: 24347 * jit/JITPropertyAccess.cpp: 24348 (JSC::JIT::emit_op_get_by_val): 24349 (JSC::JIT::emitSlow_op_get_by_val): 24350 243512009-05-14 Geoffrey Garen <ggaren@apple.com> 24352 24353 Reviewed by Sam Weinig. 24354 24355 Fixed a failure in fast/js/math-transforms.html caused by failing to 24356 preserve -0 in multiplication. 24357 24358 * assembler/X86Assembler.h: 24359 (JSC::X86Assembler::jz): 24360 * jit/JITArithmetic.cpp: 24361 (JSC::JIT::emit_op_mul): 24362 (JSC::JIT::emitSlow_op_mul): 24363 (JSC::JIT::emitMul32Constant): 24364 (JSC::JIT::emitMul32InPlace): Check both for overflow and for zero when 24365 doing multiplication. Use a slow case to get these right. 24366 243672009-05-14 Geoffrey Garen <ggaren@apple.com> 24368 24369 Reviewed by Sam Weinig. 24370 24371 Fixed a bug in the varargs calling convention. 24372 24373 * jit/JITCall.cpp: 24374 (JSC::JIT::compileOpCallVarargs): Move the argument count into regT1, 24375 since that's where ctiVirtualCall expects it to be. 24376 243772009-05-14 Geoffrey Garen <ggaren@apple.com> 24378 24379 Reviewed by Sam Weinig. 24380 24381 Fixed a small bug in instanceof's looping code. 24382 24383 * jit/JITOpcodes.cpp: 24384 (JSC::JIT::emit_op_instanceof): NULL means the object has no prototype, 24385 so only loop when *not* equal to NULL. 24386 243872009-05-14 Geoffrey Garen <ggaren@apple.com> 24388 24389 Reviewed by Sam Weinig. 24390 24391 Fixed a small bug in instanceof's result writing code. 24392 24393 * jit/JITOpcodes.cpp: 24394 (JSC::JIT::emit_op_instanceof): Make sure to fill out the payload bits 24395 in all cases. 24396 243972009-05-14 Sam Weinig <sam@webkit.org> 24398 24399 Reviewed by Geoffrey Garen. 24400 24401 Removed an invalid assertion in cti_op_urshift which 24402 depended on a fast path for op_urshift which has 24403 never existed. 24404 24405 * jit/JITStubs.cpp: 24406 (JSC::JITStubs::cti_op_urshift): 24407 244082009-05-14 Geoffrey Garen <ggaren@apple.com> 24409 24410 Reviewed by Sam Weinig. 24411 24412 Fixed loop_if_true, which had the same reversed test that jtrue had. 24413 24414 * jit/JITOpcodes.cpp: 24415 (JSC::JIT::emit_op_loop_if_true): 24416 244172009-05-14 Sam Weinig <sam@webkit.org> 24418 24419 Reviewed by Geoffrey Garen. 24420 24421 In op_neq, we apparently want to check that one value 24422 does *not* equal another. Go figure. 24423 24424 * jit/JITOpcodes.cpp: 24425 (JSC::JIT::emit_op_neq): 24426 244272009-05-14 Sam Weinig <sam@webkit.org> 24428 24429 Reviewed by Geoffrey Garen. 24430 24431 The slow case of op_mod should call op_mod's jit stub, 24432 not op_mul. That would be dumb. 24433 24434 * jit/JITArithmetic.cpp: 24435 (JSC::JIT::emitSlow_op_mod): 24436 244372009-05-14 Geoffrey Garen <ggaren@apple.com> 24438 24439 Reviewed by Sam Weinig. 24440 24441 Fixed problems when using 'arguments' due to a half-initialized register. 24442 24443 * interpreter/CallFrame.h: 24444 (JSC::ExecState::setCalleeArguments): 24445 (JSC::ExecState::init): Require a full JSValue when setting up the 24446 'arguments' virtual register, since this register is accessible from JIT 24447 code and bytecode, and needs to be a true JSValue. 24448 24449 * interpreter/CallFrameClosure.h: 24450 (JSC::CallFrameClosure::resetCallFrame): ditto 24451 24452 * interpreter/Interpreter.cpp: 24453 (JSC::Interpreter::privateExecute): ditto 24454 24455 * interpreter/Register.h: Removed the constructor that allowed assignment 24456 of a JSArguments* to a register. That is not safe. See above. 24457 24458 * jit/JITStubs.cpp: 24459 (JSC::JITStubs::cti_op_create_arguments): 24460 (JSC::JITStubs::cti_op_create_arguments_no_params): ditto 24461 244622009-05-14 Sam Weinig <sam@webkit.org> 24463 24464 Reviewed by Geoffrey Garen. 24465 24466 We really want to go to the slow case in op_jfalse and 24467 op_jtrue if the value is *not* boolean. 24468 24469 * jit/JITOpcodes.cpp: 24470 (JSC::JIT::emit_op_jfalse): 24471 (JSC::JIT::emit_op_jtrue): 24472 244732009-05-14 Sam Weinig <sam@webkit.org> 24474 24475 Reviewed by Geoffrey Garen. 24476 24477 Flipped the condition when emitting a an op_loop_if_less or op_loop_if_lesseq 24478 if the first operand is a constant. 24479 24480 * jit/JITOpcodes.cpp: 24481 (JSC::JIT::emit_op_loop_if_less): 24482 (JSC::JIT::emit_op_loop_if_lesseq): 24483 244842009-05-14 Sam Weinig <sam@webkit.org> 24485 24486 Reviewed by Geoffrey Garen. 24487 24488 Added missing return in op_jnless and op_jnlesseq. 24489 24490 * jit/JITArithmetic.cpp: 24491 (JSC::JIT::emit_op_jnless): 24492 (JSC::JIT::emit_op_jnlesseq): 24493 244942009-05-14 Sam Weinig <sam@webkit.org> 24495 24496 Reviewed by Geoffrey Garen. 24497 24498 Load constants into the the register file as a temporary measure to 24499 aid bring up. This allows us to use to treat constants like any 24500 other virtual register. 24501 24502 * jit/JITOpcodes.cpp: 24503 (JSC::JIT::emit_op_enter): 24504 (JSC::JIT::emit_op_enter_with_activation): 24505 245062009-05-14 Geoffrey Garen <ggaren@apple.com> 24507 24508 Reviewed by Sam Weinig. 24509 24510 Implemented op_strict_eq. Original patch by Snowy, by way of Sam and Gavin. 24511 24512 * assembler/MacroAssemblerX86Common.h: 24513 (JSC::MacroAssemblerX86Common::set8): Added set8, since it's slightly 24514 faster than set32, and the new value representation usually doesn't 24515 need set32. 24516 24517 * jit/JIT.cpp: 24518 * jit/JIT.h: 24519 * jit/JITInlineMethods.h: 24520 (JSC::JIT::emitLoadTag): 24521 (JSC::JIT::emitLoadPayload): Added helper functions for dealing with 24522 constants. Eventually, we should write special cases for all constants, 24523 but these are helpful in the short term. 24524 24525 * jit/JITOpcodes.cpp: 24526 (JSC::JIT::compileOpStrictEq): 24527 (JSC::JIT::emitSlow_op_stricteq): 24528 (JSC::JIT::emitSlow_op_nstricteq): teh opcodez. 24529 24530 * runtime/JSValue.h: 24531 (JSC::JSValue::): 24532 (JSC::JSValue::isDouble): Added a LowestTag for clarity. 24533 245342009-05-13 Geoffrey Garen <ggaren@apple.com> 24535 24536 Reviewed by Sam Weinig. 24537 24538 Fixed some bugs in host function calls. 24539 24540 testapi now passes! 24541 24542 * jit/JIT.cpp: Changed some registers around to avoid overwriting edx:eax, 24543 which is how JSValues are now returned. Also changed the code that 24544 passes thisValue to pass the full 64bits of the value. Also added 24545 an #error compiler directive to other platform builds, since the JSValue 24546 return signature probably won't return in edx:eax on those platforms, 24547 and we'll have to investigate a solution. 24548 245492009-05-13 Geoffrey Garen <ggaren@apple.com> 24550 24551 Reviewed by Sam Weinig. 24552 24553 Removed parameters from functions that are intended never to use their 24554 parameters. 24555 24556 * jit/JITPropertyAccess.cpp: 24557 (JSC::JIT::emitSlow_op_get_by_val): 24558 (JSC::JIT::emitSlow_op_put_by_val): 24559 245602009-05-13 Geoffrey Garen <ggaren@apple.com> 24561 24562 Reviewed by Sam Weinig. 24563 24564 Ported op_instance_of from TOT. It's basically the same, but some register 24565 stuff changed to memory stuff. 24566 24567 * jit/JITInlineMethods.h: 24568 (JSC::JIT::emitPutJITStubArgFromVirtualRegister): 24569 (JSC::JIT::emitStore): Changed to use helper functions. 24570 24571 * jit/JITOpcodes.cpp: 24572 (JSC::JIT::emit_op_instanceof): 24573 (JSC::JIT::emitSlow_op_instanceof): Ported from TOT. 24574 245752009-05-13 Geoffrey Garen <ggaren@apple.com> 24576 24577 Reviewed by Gavin Barraclough. 24578 24579 Added a comment to explain an exception-handling subtelty that we found 24580 hard to remember when reviewing my last patch. 24581 24582 * jit/JITOpcodes.cpp: 24583 (JSC::JIT::emit_op_catch): 24584 245852009-05-13 Geoffrey Garen <ggaren@apple.com> 24586 24587 Reviewed by Sam Weinig. 24588 24589 Implemented try/catch. 24590 24591 * jit/JITOpcodes.cpp: 24592 (JSC::JIT::emit_op_throw): Updated to use JITStackFrame abstraction. 24593 (JSC::JIT::emit_op_catch): Filled out. 24594 245952009-05-13 Sam Weinig <sam@webkit.org> 24596 24597 Reviewed by Geoffrey Garen. 24598 24599 Implemented op_loop_if_true, op_jfalse, op_jtrue, op_jeq_null and op_jneq_null 24600 24601 * jit/JITOpcodes.cpp: 24602 (JSC::JIT::emitSlow_op_instanceof): Moved from below to be next to its 24603 fast brother. 24604 24605 (JSC::JIT::emit_op_loop_if_true): Similar to the old version 24606 in that it tries to do the integer case first and reduce the 24607 number of jumps you might need to take. 24608 (JSC::JIT::emitSlow_op_loop_if_true): 24609 24610 (JSC::JIT::emit_op_jfalse): Very similar to op_loop_if_true, only 24611 the inverse and without a timeout check. 24612 (JSC::JIT::emitSlow_op_jfalse): 24613 24614 (JSC::JIT::emit_op_jtrue): Very similar to op_loop_if_true except 24615 without the timeout check. 24616 (JSC::JIT::emitSlow_op_jtrue): 24617 24618 (JSC::JIT::emit_op_jeq_null): Very similar to the implementation 24619 of op_eq, except it takes jumps instead of copying the condition 24620 to a dst. 24621 (JSC::JIT::emit_op_jneq_null): Ditto but for op_neq. 24622 246232009-05-13 Geoffrey Garen <ggaren@apple.com> 24624 24625 Reviewed by Sam Weinig. 24626 24627 Implemented op_call_varargs. 24628 24629 * jit/JITCall.cpp: 24630 (JSC::JIT::compileOpCallVarargsSetupArgs): 24631 (JSC::JIT::compileOpCallVarargs): 24632 (JSC::JIT::emit_op_call): 24633 (JSC::JIT::emit_op_call_eval): 24634 (JSC::JIT::emit_op_load_varargs): 24635 (JSC::JIT::emit_op_call_varargs): 24636 (JSC::JIT::emit_op_construct): 24637 * jit/JITOpcodes.cpp: 24638 (JSC::JIT::emit_op_jneq_ptr): 24639 246402009-05-13 Geoffrey Garen <ggaren@apple.com> 24641 24642 Reviewed by Sam Weinig. 24643 24644 Implemented op_call_eval. 24645 24646 * jit/JITCall.cpp: 24647 (JSC::JIT::compileOpCallVarargsSetupArgs): 24648 (JSC::JIT::compileOpCall): 24649 * jit/JITStubCall.h: 24650 (JSC::CallEvalJITStub::CallEvalJITStub): 24651 246522009-05-13 Sam Weinig <sam@webkit.org> 24653 24654 Reviewed by Gavin Barraclough. 24655 24656 Implemented op_not. (Gavin did most of the work!) 24657 24658 * jit/JITOpcodes.cpp: 24659 (JSC::JIT::emit_op_not): 24660 (JSC::JIT::emitSlow_op_not): 24661 246622009-05-13 Geoffrey Garen <ggaren@apple.com> 24663 24664 Reviewed by Sam Weinig. 24665 24666 Implemented op_global_resolve. 24667 24668 * jit/JITOpcodes.cpp: 24669 (JSC::JIT::emit_op_loop_if_less): 24670 (JSC::JIT::emit_op_loop_if_lesseq): Added back accidentally removed 24671 early returns. 24672 24673 (JSC::JIT::emit_op_resolve_global): 24674 * jit/JITStubs.cpp: 24675 (JSC::JITStubs::cti_op_resolve_global): Pretty similar to the old code, 24676 but we need two reads and a TimesEight step in order to account for the 24677 64bit value size. 24678 24679 * jit/JITStubs.h: 24680 (JSC::): Slightly tweaked this code to specialize for a JSGlobalObject*, 24681 to avoid having to pass an irrelevant tag pointer to the stub. 24682 246832009-05-13 Sam Weinig <sam@webkit.org> 24684 24685 Reviewed by Geoffrey Garen. 24686 24687 Implemented op_to_jsnumber. 24688 24689 * jit/JITOpcodes.cpp: 24690 (JSC::JIT::emit_op_to_jsnumber): 24691 (JSC::JIT::emitSlow_op_to_jsnumber): 24692 246932009-05-13 Sam Weinig <sam@webkit.org> 24694 24695 Reviewed by Geoffrey Garen. 24696 24697 Implemented op_convert_this. 24698 24699 * jit/JITOpcodes.cpp: 24700 (JSC::JIT::emit_op_convert_this): 24701 (JSC::JIT::emitSlow_op_convert_this): 24702 247032009-05-13 Geoffrey Garen <ggaren@apple.com> 24704 24705 Reviewed by Sam Weinig. 24706 24707 Got basic JS function and constructor calls working. 24708 24709 * jit/JIT.cpp: 24710 (JSC::JIT::privateCompileCTIMachineTrampolines): 24711 * jit/JIT.h: 24712 * jit/JITCall.cpp: 24713 (JSC::JIT::compileOpCallSetupArgs): 24714 (JSC::JIT::compileOpCallVarargsSetupArgs): 24715 (JSC::JIT::compileOpConstructSetupArgs): 24716 (JSC::JIT::emit_op_ret): 24717 (JSC::JIT::emit_op_construct_verify): 24718 (JSC::JIT::emitSlow_op_construct_verify): 24719 (JSC::JIT::emitSlow_op_call): 24720 (JSC::JIT::emitSlow_op_call_eval): 24721 (JSC::JIT::emitSlow_op_call_varargs): 24722 (JSC::JIT::emitSlow_op_construct): 24723 (JSC::JIT::compileOpCall): Filled out these cases, with call_eval #if'd out. 24724 24725 * jit/JITInlineMethods.h: 24726 (JSC::JIT::emitPutJITStubArgFromVirtualRegister): 24727 (JSC::JIT::emitLoad): Restored some legacy "*CTIArg*" functions, 24728 since I wanted to avoid the complexity of revamping the API here while 24729 trying to bring it up. Eventually, we should re-remove all of these functions. 24730 24731 (JSC::JIT::recordJumpTarget): Removed unnecessary macro cruft. You will 24732 not silence me, Sam Weinig! The world will know that you are a crufty, 24733 crufty, crufty programmer!!! 24734 24735 * jit/JITOpcodes.cpp: 24736 * jit/JITStubs.cpp: 24737 (JSC::): 24738 * jit/JITStubs.h: Changed up some offsets in the JITStackFrame class, since 24739 and off-by-one error was causing stack misalignment. 24740 247412009-05-13 Sam Weinig <sam@webkit.org> 24742 24743 Reviewed by Geoffrey Garen. 24744 24745 Implement op_eq_null and op_neq_null. 24746 24747 * assembler/MacroAssemblerX86Common.h: 24748 (JSC::MacroAssemblerX86Common::set8): 24749 (JSC::MacroAssemblerX86Common::setTest8): 24750 * jit/JITOpcodes.cpp: 24751 (JSC::JIT::emit_op_stricteq): 24752 (JSC::JIT::emitSlow_op_stricteq): 24753 (JSC::JIT::emit_op_nstricteq): 24754 (JSC::JIT::emitSlow_op_nstricteq): 24755 (JSC::JIT::emit_op_eq_null): 24756 (JSC::JIT::emit_op_neq_null): 24757 * jsc.cpp: 24758 247592009-05-12 Sam Weinig <sam@webkit.org> 24760 24761 Reviewed by Geoffrey Garen. 24762 24763 Implement op_new_error. 24764 24765 * jit/JITOpcodes.cpp: 24766 (JSC::JIT::emit_op_new_error): 24767 * jit/JITStubCall.h: 24768 (JSC::JITStubCall::addArgument): Add a version of addArgument 24769 that takes a constant JSValue. 24770 247712009-05-12 Sam Weinig <sam@webkit.org> 24772 24773 Reviewed by Geoffrey Garen. 24774 24775 Remove now unused emitGetVariableObjectRegister and emitPutVariableObjectRegister. 24776 24777 * jit/JIT.cpp: 24778 * jit/JIT.h: 24779 247802009-05-12 Sam Weinig <sam@webkit.org> 24781 24782 Reviewed by Geoffrey Garen. 24783 24784 Implement op_to_primitive and op_next_pname. 24785 24786 * jit/JITOpcodes.cpp: 24787 (JSC::JIT::emitSlow_op_construct_verify): 24788 (JSC::JIT::emit_op_to_primitive): 24789 (JSC::JIT::emitSlow_op_to_primitive): 24790 (JSC::JIT::emitSlow_op_loop_if_true): 24791 (JSC::JIT::emit_op_jtrue): 24792 (JSC::JIT::emit_op_next_pname): 24793 247942009-05-12 Sam Weinig <sam@webkit.org> 24795 24796 Reviewed by Geoffrey Garen. 24797 24798 Add op_get_global_var, op_put_global_var, emit_op_get_scoped_var, emit_op_put_scoped_var and 24799 op_unexpected_load. 24800 24801 * jit/JIT.h: 24802 * jit/JITInlineMethods.h: 24803 (JSC::JIT::tagFor): 24804 (JSC::JIT::payloadFor): 24805 (JSC::JIT::emitLoad): 24806 (JSC::JIT::emitStore): 24807 (JSC::JIT::emitLoadReturnValue): 24808 * jit/JITOpcodes.cpp: 24809 (JSC::JIT::emit_op_get_global_var): 24810 (JSC::JIT::emit_op_put_global_var): 24811 (JSC::JIT::emit_op_get_scoped_var): 24812 (JSC::JIT::emit_op_put_scoped_var): 24813 (JSC::JIT::emit_op_unexpected_load): 24814 248152009-05-12 Geoffrey Garen <ggaren@apple.com> 24816 24817 Reviewed by Sam Weinig. 24818 24819 Added overflow handling to op_sub. 24820 24821 * jit/JIT.h: 24822 * jit/JITArithmetic.cpp: 24823 (JSC::JIT::emitSlow_op_sub): 24824 (JSC::JIT::emitSlowSub32InPlaceLeft): 24825 248262009-05-12 Sam Weinig <sam@webkit.org> 24827 24828 Reviewed by Geoffrey Garen. 24829 24830 Remove a function call by folding op_get_by_id and op_put_by_id into 24831 their respective compile functions. 24832 24833 * jit/JIT.h: 24834 * jit/JITPropertyAccess.cpp: 24835 (JSC::JIT::emit_op_get_by_id): 24836 (JSC::JIT::emitSlow_op_get_by_id): 24837 (JSC::JIT::emit_op_put_by_id): 24838 (JSC::JIT::emitSlow_op_put_by_id): 24839 248402009-05-12 Sam Weinig <sam@webkit.org> 24841 24842 Reviewed by Geoffrey Garen. 24843 24844 Make JITStubCall work in 64bit by making the stack index 24845 step dependent on the size of void*. 24846 24847 * jit/JITStubCall.h: 24848 (JSC::JITStubCall::JITStubCall): 24849 (JSC::JITStubCall::addArgument): 24850 248512009-05-12 Sam Weinig <sam@webkit.org> 24852 24853 Reviewed by Geoffrey Garen. 24854 24855 Implement simple version of property access opcodes 24856 which just call a stub functions. 24857 24858 * jit/JITOpcodes.cpp: 24859 * jit/JITPropertyAccess.cpp: 24860 (JSC::JIT::emitSlow_op_put_by_id): 24861 (JSC::JIT::emitSlow_op_get_by_id): 24862 (JSC::JIT::emit_op_get_by_val): 24863 (JSC::JIT::emitSlow_op_get_by_val): 24864 (JSC::JIT::emit_op_put_by_val): 24865 (JSC::JIT::emitSlow_op_put_by_val): 24866 (JSC::JIT::emit_op_put_by_index): 24867 (JSC::JIT::emit_op_put_getter): 24868 (JSC::JIT::emit_op_put_setter): 24869 (JSC::JIT::emit_op_del_by_id): 24870 (JSC::JIT::compileGetByIdHotPath): 24871 (JSC::JIT::compilePutByIdHotPath): 24872 * jit/JITStubCall.h: 24873 (JSC::JITStubCall::addArgument): 24874 * jsc.cpp: 24875 248762009-05-12 Geoffrey Garen <ggaren@apple.com> 24877 24878 Reviewed by Sam Weinig. 24879 24880 Added work-around for XCode debugging echo problem. 24881 24882 * jsc.cpp: 24883 (runInteractive): 24884 248852009-05-12 Geoffrey Garen <ggaren@apple.com> 24886 24887 Reviewed by Sam Weinig. 24888 24889 Added overflow handling to op_add. 24890 24891 * jit/JIT.h: 24892 * jit/JITArithmetic.cpp: 24893 (JSC::JIT::emitSlow_op_add): 24894 (JSC::JIT::emitSlowAdd32InPlace): 24895 248962009-05-12 Sam Weinig <sam@webkit.org> 24897 24898 Reviewed by Geoffrey Garen. 24899 24900 Add slow cases for op_jnless or emit_op_jnlesseq. 24901 24902 * jit/JITArithmetic.cpp: 24903 (JSC::JIT::emitSlow_op_jnless): 24904 (JSC::JIT::emitSlow_op_jnlesseq): 24905 249062009-05-12 Sam Weinig <sam@webkit.org> 24907 24908 Reviewed by Geoffrey Garen. 24909 24910 Add implementations for op_jnless, emit_op_jnlesseq, op_loop_if_less and op_loop_if_lesseq. 24911 No slow cases for op_jnless or emit_op_jnlesseq yet. 24912 24913 * jit/JITArithmetic.cpp: 24914 (JSC::JIT::emit_op_jnless): 24915 (JSC::JIT::emitSlow_op_jnless): 24916 (JSC::JIT::emit_op_jnlesseq): 24917 (JSC::JIT::emitSlow_op_jnlesseq): 24918 * jit/JITOpcodes.cpp: 24919 (JSC::JIT::emit_op_loop_if_less): 24920 (JSC::JIT::emitSlow_op_loop_if_less): 24921 (JSC::JIT::emit_op_loop_if_lesseq): 24922 (JSC::JIT::emitSlow_op_loop_if_lesseq): 24923 249242009-05-12 Sam Weinig <sam@webkit.org> 24925 24926 Reviewed by Geoffrey Garen. 24927 24928 Turn the RECORD_JUMP_TARGET macro into an inline function. 24929 24930 * jit/JIT.h: 24931 * jit/JITInlineMethods.h: 24932 (JSC::JIT::recordJumpTarget): 24933 * jit/JITOpcodes.cpp: 24934 (JSC::JIT::emit_op_jmp): 24935 (JSC::JIT::emit_op_jsr): 24936 (JSC::JIT::emit_op_jmp_scopes): 24937 249382009-05-12 Sam Weinig <sam@webkit.org> 24939 24940 Add MacroAssemblerX86Common::set8 to fix the build. 24941 24942 * assembler/MacroAssemblerX86Common.h: 24943 (JSC::MacroAssemblerX86Common::set8): 24944 249452009-05-12 Geoffrey Garen <ggaren@apple.com> 24946 24947 Reviewed by Sam Weinig. 24948 24949 Added overflow recovery for pre_inc and pre_dec. 24950 24951 Turned some short-circuit code into early returns, as is the WebKit style. 24952 24953 * jit/JITArithmetic.cpp: 24954 (JSC::JIT::emit_op_post_inc): 24955 (JSC::JIT::emitSlow_op_post_inc): 24956 (JSC::JIT::emit_op_post_dec): 24957 (JSC::JIT::emitSlow_op_post_dec): 24958 (JSC::JIT::emitSlow_op_pre_inc): 24959 (JSC::JIT::emitSlow_op_pre_dec): 24960 249612009-05-12 Sam Weinig <sam@webkit.org> 24962 24963 Reviewed by Geoffrey Garen. 24964 24965 Implement op_jmp, op_loop, op_eq and op_neq. 24966 24967 * jit/JITOpcodes.cpp: 24968 (JSC::JIT::emit_op_jmp): 24969 (JSC::JIT::emit_op_loop): 24970 (JSC::JIT::emit_op_eq): 24971 (JSC::JIT::emitSlow_op_eq): 24972 (JSC::JIT::emit_op_neq): 24973 (JSC::JIT::emitSlow_op_neq): 24974 (JSC::JIT::emit_op_enter): 24975 (JSC::JIT::emit_op_enter_with_activation): 24976 249772009-05-12 Sam Weinig <sam@webkit.org> 24978 24979 Reviewed by Geoffrey Garen. 24980 24981 Implement the slow cases for arithmetic opcodes. 24982 24983 * jit/JITArithmetic.cpp: 24984 (JSC::JIT::emitSlow_op_lshift): 24985 (JSC::JIT::emitSlow_op_rshift): 24986 (JSC::JIT::emitSlow_op_bitand): 24987 (JSC::JIT::emitSlow_op_bitor): 24988 (JSC::JIT::emitSlow_op_bitxor): 24989 (JSC::JIT::emitSlow_op_bitnot): 24990 (JSC::JIT::emitSlow_op_sub): 24991 (JSC::JIT::emitSlow_op_mul): 24992 (JSC::JIT::emitSlow_op_mod): 24993 (JSC::JIT::emit_op_mod): 24994 249952009-05-12 Sam Weinig <sam@webkit.org> 24996 24997 Reviewed by Geoffrey Garen. 24998 24999 Implement op_bitnot. 25000 25001 * assembler/MacroAssemblerX86Common.h: 25002 (JSC::MacroAssemblerX86Common::not32): 25003 * assembler/X86Assembler.h: 25004 (JSC::X86Assembler::notl_m): 25005 * jit/JITArithmetic.cpp: 25006 (JSC::JIT::emit_op_bitnot): 25007 250082009-05-12 Sam Weinig <sam@webkit.org> 25009 25010 Reviewed by Geoffrey Garen. 25011 25012 Add arithmetic opcode implementations from the old nitro-extreme branch. 25013 25014 * jit/JIT.h: 25015 * jit/JITArithmetic.cpp: 25016 (JSC::JIT::emit_op_jnless): 25017 (JSC::JIT::emitSlow_op_jnless): 25018 (JSC::JIT::emit_op_jnlesseq): 25019 (JSC::JIT::emitSlow_op_jnlesseq): 25020 (JSC::JIT::emit_op_lshift): 25021 (JSC::JIT::emitSlow_op_lshift): 25022 (JSC::JIT::emit_op_rshift): 25023 (JSC::JIT::emitSlow_op_rshift): 25024 (JSC::JIT::emit_op_bitand): 25025 (JSC::JIT::emitBitAnd32Constant): 25026 (JSC::JIT::emitBitAnd32InPlace): 25027 (JSC::JIT::emit_op_bitor): 25028 (JSC::JIT::emitSlow_op_bitor): 25029 (JSC::JIT::emitBitOr32Constant): 25030 (JSC::JIT::emitBitOr32InPlace): 25031 (JSC::JIT::emit_op_bitxor): 25032 (JSC::JIT::emitSlow_op_bitxor): 25033 (JSC::JIT::emitBitXor32Constant): 25034 (JSC::JIT::emitBitXor32InPlace): 25035 (JSC::JIT::emit_op_bitnot): 25036 (JSC::JIT::emitSlow_op_bitnot): 25037 (JSC::JIT::emit_op_post_inc): 25038 (JSC::JIT::emitSlow_op_post_inc): 25039 (JSC::JIT::emit_op_post_dec): 25040 (JSC::JIT::emitSlow_op_post_dec): 25041 (JSC::JIT::emit_op_pre_inc): 25042 (JSC::JIT::emitSlow_op_pre_inc): 25043 (JSC::JIT::emit_op_pre_dec): 25044 (JSC::JIT::emitSlow_op_pre_dec): 25045 (JSC::JIT::emit_op_add): 25046 (JSC::JIT::emitAdd32Constant): 25047 (JSC::JIT::emitAdd32InPlace): 25048 (JSC::JIT::emitSlow_op_add): 25049 (JSC::JIT::emit_op_sub): 25050 (JSC::JIT::emitSlow_op_sub): 25051 (JSC::JIT::emitSub32ConstantLeft): 25052 (JSC::JIT::emitSub32ConstantRight): 25053 (JSC::JIT::emitSub32InPlaceLeft): 25054 (JSC::JIT::emitSub32InPlaceRight): 25055 (JSC::JIT::emit_op_mul): 25056 (JSC::JIT::emitSlow_op_mul): 25057 (JSC::JIT::emitMul32Constant): 25058 (JSC::JIT::emitMul32InPlace): 25059 (JSC::JIT::emit_op_mod): 25060 (JSC::JIT::emitSlow_op_mod): 25061 * jit/JITOpcodes.cpp: 25062 250632009-05-12 Geoffrey Garen <ggaren@apple.com> 25064 25065 Removed JIT_OPTIMIZE_ARITHMETIC setting, since it was all about 32bit 25066 value representations. 25067 25068 Added JSAPIValueWrapper to the repository. 25069 25070 * jit/JIT.h: 25071 * jit/JITArithmetic.cpp: 25072 * runtime/JSAPIValueWrapper.cpp: Added. 25073 (JSC::JSAPIValueWrapper::toPrimitive): 25074 (JSC::JSAPIValueWrapper::getPrimitiveNumber): 25075 (JSC::JSAPIValueWrapper::toBoolean): 25076 (JSC::JSAPIValueWrapper::toNumber): 25077 (JSC::JSAPIValueWrapper::toString): 25078 (JSC::JSAPIValueWrapper::toObject): 25079 * runtime/JSAPIValueWrapper.h: Added. 25080 (JSC::JSAPIValueWrapper::value): 25081 (JSC::JSAPIValueWrapper::isAPIValueWrapper): 25082 (JSC::JSAPIValueWrapper::JSAPIValueWrapper): 25083 (JSC::jsAPIValueWrapper): 25084 * wtf/Platform.h: 25085 250862009-05-12 Geoffrey Garen <ggaren@apple.com> 25087 25088 Turned on the JIT and got it building and running the most trivial of 25089 programs. 25090 25091 All configurable optimizations are turned off, and a few opcodes are ad 25092 hoc #if'd out. 25093 25094 So far, I've only merged op_mov and op_end, but some stub-reliant 25095 opcodes work as-is from TOT. 25096 25097 * bytecode/CodeBlock.cpp: 25098 (JSC::CodeBlock::~CodeBlock): 25099 * bytecode/CodeBlock.h: 25100 * jit/JIT.cpp: 25101 (JSC::JIT::compileOpStrictEq): 25102 * jit/JIT.h: 25103 * jit/JITArithmetic.cpp: 25104 (JSC::JIT::emit_op_lshift): 25105 (JSC::JIT::emitSlow_op_lshift): 25106 (JSC::JIT::emit_op_rshift): 25107 (JSC::JIT::emitSlow_op_rshift): 25108 (JSC::JIT::emit_op_jnless): 25109 (JSC::JIT::emitSlow_op_jnless): 25110 (JSC::JIT::emit_op_jnlesseq): 25111 (JSC::JIT::emitSlow_op_jnlesseq): 25112 (JSC::JIT::emit_op_bitand): 25113 (JSC::JIT::emitSlow_op_bitand): 25114 (JSC::JIT::emit_op_post_inc): 25115 (JSC::JIT::emitSlow_op_post_inc): 25116 (JSC::JIT::emit_op_post_dec): 25117 (JSC::JIT::emitSlow_op_post_dec): 25118 (JSC::JIT::emit_op_pre_inc): 25119 (JSC::JIT::emitSlow_op_pre_inc): 25120 (JSC::JIT::emit_op_pre_dec): 25121 (JSC::JIT::emitSlow_op_pre_dec): 25122 (JSC::JIT::emit_op_mod): 25123 (JSC::JIT::emitSlow_op_mod): 25124 (JSC::JIT::emit_op_add): 25125 (JSC::JIT::emit_op_mul): 25126 (JSC::JIT::emit_op_sub): 25127 (JSC::JIT::compileBinaryArithOpSlowCase): 25128 (JSC::JIT::emitSlow_op_add): 25129 (JSC::JIT::emitSlow_op_mul): 25130 * jit/JITCall.cpp: 25131 (JSC::JIT::compileOpCallInitializeCallFrame): 25132 (JSC::JIT::compileOpConstructSetupArgs): 25133 (JSC::JIT::compileOpCallVarargs): 25134 (JSC::JIT::compileOpCall): 25135 (JSC::JIT::compileOpCallSlowCase): 25136 * jit/JITInlineMethods.h: 25137 (JSC::JIT::getConstantOperandImmediateInt): 25138 (JSC::JIT::isOperandConstantImmediateInt): 25139 (JSC::JIT::emitInitRegister): 25140 (JSC::JIT::addSlowCase): 25141 (JSC::JIT::addJump): 25142 (JSC::JIT::emitJumpSlowToHot): 25143 (JSC::JIT::tagFor): 25144 (JSC::JIT::payloadFor): 25145 (JSC::JIT::emitLoad): 25146 (JSC::JIT::emitLoadReturnValue): 25147 (JSC::JIT::emitStore): 25148 (JSC::JIT::emitStoreReturnValue): 25149 * jit/JITOpcodes.cpp: 25150 (JSC::JIT::emit_op_mov): 25151 (JSC::JIT::emit_op_end): 25152 (JSC::JIT::emit_op_jmp): 25153 (JSC::JIT::emit_op_loop): 25154 (JSC::JIT::emit_op_loop_if_less): 25155 (JSC::JIT::emit_op_loop_if_lesseq): 25156 (JSC::JIT::emit_op_instanceof): 25157 (JSC::JIT::emit_op_get_global_var): 25158 (JSC::JIT::emit_op_put_global_var): 25159 (JSC::JIT::emit_op_get_scoped_var): 25160 (JSC::JIT::emit_op_put_scoped_var): 25161 (JSC::JIT::emit_op_tear_off_activation): 25162 (JSC::JIT::emit_op_ret): 25163 (JSC::JIT::emit_op_construct_verify): 25164 (JSC::JIT::emit_op_to_primitive): 25165 (JSC::JIT::emit_op_loop_if_true): 25166 (JSC::JIT::emit_op_resolve_global): 25167 (JSC::JIT::emit_op_not): 25168 (JSC::JIT::emit_op_jfalse): 25169 (JSC::JIT::emit_op_jeq_null): 25170 (JSC::JIT::emit_op_jneq_null): 25171 (JSC::JIT::emit_op_jneq_ptr): 25172 (JSC::JIT::emit_op_unexpected_load): 25173 (JSC::JIT::emit_op_eq): 25174 (JSC::JIT::emit_op_bitnot): 25175 (JSC::JIT::emit_op_jtrue): 25176 (JSC::JIT::emit_op_neq): 25177 (JSC::JIT::emit_op_bitxor): 25178 (JSC::JIT::emit_op_bitor): 25179 (JSC::JIT::emit_op_throw): 25180 (JSC::JIT::emit_op_next_pname): 25181 (JSC::JIT::emit_op_push_scope): 25182 (JSC::JIT::emit_op_to_jsnumber): 25183 (JSC::JIT::emit_op_push_new_scope): 25184 (JSC::JIT::emit_op_catch): 25185 (JSC::JIT::emit_op_switch_imm): 25186 (JSC::JIT::emit_op_switch_char): 25187 (JSC::JIT::emit_op_switch_string): 25188 (JSC::JIT::emit_op_new_error): 25189 (JSC::JIT::emit_op_eq_null): 25190 (JSC::JIT::emit_op_neq_null): 25191 (JSC::JIT::emit_op_convert_this): 25192 (JSC::JIT::emit_op_profile_will_call): 25193 (JSC::JIT::emit_op_profile_did_call): 25194 (JSC::JIT::emitSlow_op_construct_verify): 25195 (JSC::JIT::emitSlow_op_get_by_val): 25196 (JSC::JIT::emitSlow_op_loop_if_less): 25197 (JSC::JIT::emitSlow_op_loop_if_lesseq): 25198 (JSC::JIT::emitSlow_op_put_by_val): 25199 (JSC::JIT::emitSlow_op_not): 25200 (JSC::JIT::emitSlow_op_instanceof): 25201 * jit/JITPropertyAccess.cpp: 25202 (JSC::JIT::emit_op_get_by_val): 25203 (JSC::JIT::emit_op_put_by_val): 25204 (JSC::JIT::emit_op_put_by_index): 25205 (JSC::JIT::emit_op_put_getter): 25206 (JSC::JIT::emit_op_put_setter): 25207 (JSC::JIT::emit_op_del_by_id): 25208 (JSC::JIT::compileGetByIdHotPath): 25209 (JSC::JIT::compilePutByIdHotPath): 25210 * jit/JITStubCall.h: 25211 (JSC::JITStubCall::JITStubCall): 25212 (JSC::JITStubCall::addArgument): 25213 (JSC::JITStubCall::call): 25214 (JSC::JITStubCall::): 25215 (JSC::CallEvalJITStub::CallEvalJITStub): 25216 * jit/JITStubs.cpp: 25217 (JSC::): 25218 (JSC::JITStubs::cti_op_add): 25219 (JSC::JITStubs::cti_op_pre_inc): 25220 (JSC::JITStubs::cti_op_mul): 25221 (JSC::JITStubs::cti_op_get_by_val): 25222 (JSC::JITStubs::cti_op_get_by_val_string): 25223 (JSC::JITStubs::cti_op_get_by_val_byte_array): 25224 (JSC::JITStubs::cti_op_sub): 25225 (JSC::JITStubs::cti_op_put_by_val): 25226 (JSC::JITStubs::cti_op_put_by_val_array): 25227 (JSC::JITStubs::cti_op_put_by_val_byte_array): 25228 (JSC::JITStubs::cti_op_negate): 25229 (JSC::JITStubs::cti_op_div): 25230 (JSC::JITStubs::cti_op_pre_dec): 25231 (JSC::JITStubs::cti_op_post_inc): 25232 (JSC::JITStubs::cti_op_eq): 25233 (JSC::JITStubs::cti_op_lshift): 25234 (JSC::JITStubs::cti_op_bitand): 25235 (JSC::JITStubs::cti_op_rshift): 25236 (JSC::JITStubs::cti_op_bitnot): 25237 (JSC::JITStubs::cti_op_mod): 25238 (JSC::JITStubs::cti_op_neq): 25239 (JSC::JITStubs::cti_op_post_dec): 25240 (JSC::JITStubs::cti_op_urshift): 25241 (JSC::JITStubs::cti_op_bitxor): 25242 (JSC::JITStubs::cti_op_bitor): 25243 (JSC::JITStubs::cti_op_switch_imm): 25244 * jit/JITStubs.h: 25245 * runtime/JSArray.cpp: 25246 (JSC::JSArray::JSArray): 25247 * runtime/JSFunction.cpp: 25248 (JSC::JSFunction::~JSFunction): 25249 * runtime/JSValue.h: 25250 (JSC::JSValue::payload): 25251 * wtf/Platform.h: 25252 252532009-05-07 Sam Weinig <sam@webkit.org> 25254 25255 Reviewed by Geoffrey Garen. 25256 25257 Add some new MacroAssembler and assembler functions that will be needed shortly. 25258 25259 * assembler/MacroAssemblerX86Common.h: 25260 (JSC::MacroAssemblerX86Common::add32): 25261 (JSC::MacroAssemblerX86Common::and32): 25262 (JSC::MacroAssemblerX86Common::mul32): 25263 (JSC::MacroAssemblerX86Common::neg32): 25264 (JSC::MacroAssemblerX86Common::or32): 25265 (JSC::MacroAssemblerX86Common::sub32): 25266 (JSC::MacroAssemblerX86Common::xor32): 25267 (JSC::MacroAssemblerX86Common::branchAdd32): 25268 (JSC::MacroAssemblerX86Common::branchMul32): 25269 (JSC::MacroAssemblerX86Common::branchSub32): 25270 * assembler/X86Assembler.h: 25271 (JSC::X86Assembler::): 25272 (JSC::X86Assembler::addl_rm): 25273 (JSC::X86Assembler::andl_mr): 25274 (JSC::X86Assembler::andl_rm): 25275 (JSC::X86Assembler::andl_im): 25276 (JSC::X86Assembler::negl_r): 25277 (JSC::X86Assembler::notl_r): 25278 (JSC::X86Assembler::orl_rm): 25279 (JSC::X86Assembler::orl_im): 25280 (JSC::X86Assembler::subl_rm): 25281 (JSC::X86Assembler::xorl_mr): 25282 (JSC::X86Assembler::xorl_rm): 25283 (JSC::X86Assembler::xorl_im): 25284 (JSC::X86Assembler::imull_mr): 25285 252862009-05-11 Sam Weinig <sam@webkit.org> 25287 25288 Reviewed by Cameron Zwarich. 25289 25290 Remove the NumberHeap. 25291 25292 * JavaScriptCore.exp: 25293 * runtime/Collector.cpp: 25294 (JSC::Heap::Heap): 25295 (JSC::Heap::destroy): 25296 (JSC::Heap::recordExtraCost): 25297 (JSC::Heap::heapAllocate): 25298 (JSC::Heap::markConservatively): 25299 (JSC::Heap::sweep): 25300 (JSC::Heap::collect): 25301 (JSC::Heap::objectCount): 25302 (JSC::Heap::statistics): 25303 (JSC::typeName): 25304 (JSC::Heap::isBusy): 25305 * runtime/Collector.h: 25306 (JSC::Heap::globalData): 25307 * runtime/JSCell.h: 25308 253092009-05-11 Geoffrey Garen <ggaren@apple.com> 25310 25311 Reviewed by Sam Weinig. 25312 25313 Land initial commit of new number representation for 32 bit platforms, 25314 with JIT disabled. 25315 25316 * API/APICast.h: 25317 (toJS): 25318 (toRef): 25319 * API/JSCallbackObjectFunctions.h: 25320 (JSC::::hasInstance): 25321 (JSC::::toNumber): 25322 (JSC::::toString): 25323 * API/tests/testapi.c: 25324 (EvilExceptionObject_convertToType): 25325 * AllInOneFile.cpp: 25326 * JavaScriptCore.exp: 25327 * JavaScriptCore.xcodeproj/project.pbxproj: 25328 * bytecode/CodeBlock.cpp: 25329 (JSC::valueToSourceString): 25330 * bytecompiler/BytecodeGenerator.cpp: 25331 (JSC::BytecodeGenerator::emitLoad): 25332 (JSC::BytecodeGenerator::emitUnexpectedLoad): 25333 (JSC::keyForImmediateSwitch): 25334 * bytecompiler/BytecodeGenerator.h: 25335 * interpreter/Interpreter.cpp: 25336 (JSC::Interpreter::dumpRegisters): 25337 (JSC::Interpreter::privateExecute): 25338 * parser/Nodes.cpp: 25339 (JSC::ArrayNode::emitBytecode): 25340 (JSC::processClauseList): 25341 * runtime/ArgList.h: 25342 * runtime/Collector.h: 25343 (JSC::sizeof): 25344 * runtime/DateMath.cpp: 25345 * runtime/ExceptionHelpers.h: 25346 * runtime/InitializeThreading.cpp: 25347 * runtime/JSArray.cpp: 25348 (JSC::JSArray::JSArray): 25349 * runtime/JSCell.cpp: 25350 * runtime/JSCell.h: 25351 (JSC::JSCell::isAPIValueWrapper): 25352 (JSC::JSValue::isString): 25353 (JSC::JSValue::isGetterSetter): 25354 (JSC::JSValue::isObject): 25355 (JSC::JSValue::getString): 25356 (JSC::JSValue::getObject): 25357 (JSC::JSValue::getCallData): 25358 (JSC::JSValue::getConstructData): 25359 (JSC::JSValue::getUInt32): 25360 (JSC::JSValue::marked): 25361 (JSC::JSValue::toPrimitive): 25362 (JSC::JSValue::getPrimitiveNumber): 25363 (JSC::JSValue::toBoolean): 25364 (JSC::JSValue::toNumber): 25365 (JSC::JSValue::toString): 25366 (JSC::JSValue::needsThisConversion): 25367 (JSC::JSValue::toThisString): 25368 (JSC::JSValue::getJSNumber): 25369 (JSC::JSValue::toObject): 25370 (JSC::JSValue::toThisObject): 25371 * runtime/JSGlobalData.cpp: 25372 (JSC::JSGlobalData::JSGlobalData): 25373 * runtime/JSGlobalData.h: 25374 * runtime/JSGlobalObject.h: 25375 (JSC::Structure::prototypeForLookup): 25376 * runtime/JSGlobalObjectFunctions.cpp: 25377 (JSC::globalFuncParseInt): 25378 * runtime/JSImmediate.h: 25379 * runtime/JSNumberCell.cpp: Removed. 25380 * runtime/JSNumberCell.h: Removed. 25381 * runtime/JSObject.h: 25382 (JSC::JSValue::get): 25383 (JSC::JSValue::put): 25384 * runtime/JSString.h: 25385 (JSC::JSValue::toThisJSString): 25386 * runtime/JSValue.cpp: 25387 (JSC::JSValue::toInteger): 25388 (JSC::JSValue::toIntegerPreserveNaN): 25389 (JSC::JSValue::toObjectSlowCase): 25390 (JSC::JSValue::toThisObjectSlowCase): 25391 (JSC::JSValue::synthesizeObject): 25392 (JSC::JSValue::synthesizePrototype): 25393 (JSC::JSValue::description): 25394 (JSC::nonInlineNaN): 25395 * runtime/JSValue.h: 25396 (JSC::JSValue::): 25397 (JSC::EncodedJSValueHashTraits::emptyValue): 25398 (JSC::jsNaN): 25399 (JSC::operator==): 25400 (JSC::operator!=): 25401 (JSC::toInt32): 25402 (JSC::toUInt32): 25403 (JSC::JSValue::encode): 25404 (JSC::JSValue::decode): 25405 (JSC::JSValue::JSValue): 25406 (JSC::JSValue::operator bool): 25407 (JSC::JSValue::operator==): 25408 (JSC::JSValue::operator!=): 25409 (JSC::JSValue::isUndefined): 25410 (JSC::JSValue::isNull): 25411 (JSC::JSValue::isUndefinedOrNull): 25412 (JSC::JSValue::isCell): 25413 (JSC::JSValue::isInt32): 25414 (JSC::JSValue::isUInt32): 25415 (JSC::JSValue::isDouble): 25416 (JSC::JSValue::isTrue): 25417 (JSC::JSValue::isFalse): 25418 (JSC::JSValue::tag): 25419 (JSC::JSValue::asInt32): 25420 (JSC::JSValue::asUInt32): 25421 (JSC::JSValue::asDouble): 25422 (JSC::JSValue::asCell): 25423 (JSC::JSValue::isNumber): 25424 (JSC::JSValue::isBoolean): 25425 (JSC::JSValue::getBoolean): 25426 (JSC::JSValue::uncheckedGetNumber): 25427 (JSC::JSValue::toJSNumber): 25428 (JSC::JSValue::getNumber): 25429 (JSC::JSValue::toInt32): 25430 (JSC::JSValue::toUInt32): 25431 * runtime/Operations.h: 25432 (JSC::JSValue::equal): 25433 (JSC::JSValue::equalSlowCaseInline): 25434 (JSC::JSValue::strictEqual): 25435 (JSC::JSValue::strictEqualSlowCaseInline): 25436 (JSC::jsLess): 25437 (JSC::jsLessEq): 25438 (JSC::jsAdd): 25439 * runtime/PropertySlot.h: 25440 * runtime/StringPrototype.cpp: 25441 (JSC::stringProtoFuncCharAt): 25442 (JSC::stringProtoFuncCharCodeAt): 25443 (JSC::stringProtoFuncIndexOf): 25444 * wtf/Platform.h: 25445 25446=== Start merge of nitro-extreme branch 2009-07-30 === 25447 254482009-07-29 Laszlo Gombos <laszlo.1.gombos@nokia.com> 25449 25450 Reviewed by George Staikos. 25451 25452 Resolve class/struct mixup in forward declarations 25453 https://bugs.webkit.org/show_bug.cgi?id=27708 25454 25455 * API/JSClassRef.h: 25456 * bytecode/SamplingTool.h: 25457 * interpreter/Interpreter.h: 25458 * jit/JIT.h: 25459 * profiler/ProfileGenerator.h: 25460 * profiler/Profiler.h: 25461 * runtime/ClassInfo.h: 25462 * runtime/ExceptionHelpers.h: 25463 * runtime/JSByteArray.h: 25464 * runtime/JSCell.h: 25465 * runtime/JSFunction.h: 25466 * runtime/JSGlobalData.h: 25467 * runtime/JSObject.h: 25468 * runtime/JSString.h: 25469 254702009-07-28 Ada Chan <adachan@apple.com> 25471 25472 Reviewed by Darin Adler. 25473 25474 https://bugs.webkit.org/show_bug.cgi?id=27236 25475 - Implement TCMalloc_SystemRelease and TCMalloc_SystemCommit for Windows. 25476 - Use a background thread to periodically scavenge memory to release back to the system. 25477 25478 * wtf/FastMalloc.cpp: 25479 (WTF::TCMalloc_PageHeap::init): 25480 (WTF::TCMalloc_PageHeap::runScavengerThread): 25481 (WTF::TCMalloc_PageHeap::scavenge): 25482 (WTF::TCMalloc_PageHeap::shouldContinueScavenging): 25483 (WTF::TCMalloc_PageHeap::New): 25484 (WTF::TCMalloc_PageHeap::AllocLarge): 25485 (WTF::TCMalloc_PageHeap::Delete): 25486 (WTF::TCMalloc_PageHeap::GrowHeap): 25487 (WTF::sleep): 25488 (WTF::TCMalloc_PageHeap::scavengerThread): 25489 * wtf/TCSystemAlloc.cpp: 25490 (TCMalloc_SystemRelease): 25491 (TCMalloc_SystemCommit): 25492 * wtf/TCSystemAlloc.h: 25493 254942009-07-28 Xan Lopez <xlopez@igalia.com> 25495 25496 Add new files, fixes distcheck. 25497 25498 * GNUmakefile.am: 25499 255002009-07-28 Csaba Osztrogonac <oszi@inf.u-szeged.hu> 25501 25502 Reviewed by Simon Hausmann. 25503 25504 [Qt] Determining whether to use JIT or interpreter 25505 moved from JavaScriptCore.pri to Platform.h 25506 25507 * JavaScriptCore.pri: 25508 * wtf/Platform.h: 25509 255102009-07-27 Brian Weinstein <bweinstein@apple.com> 25511 25512 Fix of misuse of sort command. 25513 25514 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 25515 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 25516 255172009-07-27 Brian Weinstein <bweinstein@apple.com> 25518 25519 Build fix for Windows. 25520 25521 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 25522 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 25523 255242009-07-27 Gavin Barraclough <barraclough@apple.com> 25525 25526 Rubber stamped by Oliver Hunt. 25527 25528 Fix tyop in JIT, renamed preverveReturnAddressAfterCall -> preserveReturnAddressAfterCall. 25529 25530 * jit/JIT.cpp: 25531 (JSC::JIT::privateCompile): 25532 (JSC::JIT::privateCompileCTIMachineTrampolines): 25533 * jit/JIT.h: 25534 * jit/JITInlineMethods.h: 25535 (JSC::JIT::preserveReturnAddressAfterCall): 25536 * jit/JITPropertyAccess.cpp: 25537 (JSC::JIT::privateCompilePutByIdTransition): 25538 255392009-07-27 Alexey Proskuryakov <ap@webkit.org> 25540 25541 Gtk build fix. 25542 25543 * runtime/JSLock.cpp: (JSC::JSLock::JSLock): Fix "no threading" case. 25544 255452009-07-27 Alexey Proskuryakov <ap@webkit.org> 25546 25547 Release build fix. 25548 25549 * runtime/JSLock.h: (JSC::JSLock::~JSLock): 25550 255512009-07-27 Alexey Proskuryakov <ap@webkit.org> 25552 25553 Reviewed by Darin Adler. 25554 25555 https://bugs.webkit.org/show_bug.cgi?id=27735 25556 Give a helpful name to JSLock constructor argument 25557 25558 * API/JSBase.cpp: 25559 (JSGarbageCollect): 25560 * API/JSContextRef.cpp: 25561 * API/JSObjectRef.cpp: 25562 (JSPropertyNameArrayRelease): 25563 (JSPropertyNameAccumulatorAddName): 25564 * JavaScriptCore.exp: 25565 * jsc.cpp: 25566 (functionGC): 25567 (cleanupGlobalData): 25568 (jscmain): 25569 * runtime/Collector.cpp: 25570 (JSC::Heap::destroy): 25571 * runtime/JSLock.cpp: 25572 (JSC::JSLock::JSLock): 25573 (JSC::JSLock::lock): 25574 (JSC::JSLock::unlock): 25575 (JSC::JSLock::DropAllLocks::DropAllLocks): 25576 (JSC::JSLock::DropAllLocks::~DropAllLocks): 25577 * runtime/JSLock.h: 25578 (JSC::): 25579 (JSC::JSLock::JSLock): 25580 (JSC::JSLock::~JSLock): 25581 255822009-07-25 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 25583 25584 Reviewed by Eric Seidel. 25585 25586 Allow custom memory allocation control for OpaqueJSPropertyNameArray struct 25587 https://bugs.webkit.org/show_bug.cgi?id=27342 25588 25589 Inherits OpaqueJSPropertyNameArray struct from FastAllocBase because it has been 25590 instantiated by 'new' JavaScriptCore/API/JSObjectRef.cpp:473. 25591 25592 * API/JSObjectRef.cpp: 25593 255942009-07-24 Ada Chan <adachan@apple.com> 25595 25596 In preparation for https://bugs.webkit.org/show_bug.cgi?id=27236: 25597 Remove TCMALLOC_TRACK_DECOMMITED_SPANS. We'll always track decommitted spans. 25598 We have tested this and show it has little impact on performance. 25599 25600 Reviewed by Mark Rowe. 25601 25602 * wtf/FastMalloc.cpp: 25603 (WTF::TCMalloc_PageHeap::New): 25604 (WTF::TCMalloc_PageHeap::AllocLarge): 25605 (WTF::propagateDecommittedState): 25606 (WTF::mergeDecommittedStates): 25607 (WTF::TCMalloc_PageHeap::Delete): 25608 (WTF::TCMalloc_PageHeap::IncrementalScavenge): 25609 256102009-07-24 Csaba Osztrogonac <oszi@inf.u-szeged.hu> 25611 25612 Reviewed by Darin Adler and Adam Barth. 25613 25614 Build fix for x86 platforms. 25615 https://bugs.webkit.org/show_bug.cgi?id=27602 25616 25617 * jit/JIT.cpp: 25618 256192009-07-23 Kevin Ollivier <kevino@theolliviers.com> 25620 25621 wx build fix, adding missing header. 25622 25623 * jit/JIT.cpp: 25624 256252009-07-22 Yong Li <yong.li@torchmobile.com> 25626 25627 Reviewed by George Staikos. 25628 25629 Add wince specific memory files into wtf/wince 25630 https://bugs.webkit.org/show_bug.cgi?id=27550 25631 25632 * wtf/wince/FastMallocWince.h: Added. 25633 * wtf/wince/MemoryManager.cpp: Added. 25634 * wtf/wince/MemoryManager.h: Added. 25635 256362009-07-23 Norbert Leser <norbert.leser@nokia.com> 25637 25638 Reviewed by Simon Hausmann. 25639 25640 Fix for missing mmap features in Symbian 25641 https://bugs.webkit.org/show_bug.cgi?id=24540 25642 25643 Fix, conditionally for PLATFORM(SYMBIAN), as an alternative 25644 to missing support for the MAP_ANON property flag in mmap. 25645 It utilizes Symbian specific memory allocation features. 25646 25647 * runtime/Collector.cpp 25648 256492009-07-22 Gavin Barraclough <barraclough@apple.com> 25650 25651 Reviewed by Sam Weinig. 25652 25653 With ENABLE(ASSEMBLER_WX_EXCLUSIVE), only change permissions once per repatch event. 25654 ( https://bugs.webkit.org/show_bug.cgi?id=27564 ) 25655 25656 Currently we change permissions forwards and backwards for each instruction modified, 25657 instead we should only change permissions once per complete repatching event. 25658 25659 2.5% progression running with ENABLE(ASSEMBLER_WX_EXCLUSIVE) enabled, 25660 which recoups 1/3 of the penalty of running with this mode enabled. 25661 25662 * assembler/ARMAssembler.cpp: 25663 (JSC::ARMAssembler::linkBranch): 25664 - Replace usage of MakeWritable with cacheFlush. 25665 25666 * assembler/ARMAssembler.h: 25667 (JSC::ARMAssembler::patchPointerInternal): 25668 (JSC::ARMAssembler::repatchLoadPtrToLEA): 25669 - Replace usage of MakeWritable with cacheFlush. 25670 25671 * assembler/ARMv7Assembler.h: 25672 (JSC::ARMv7Assembler::relinkJump): 25673 (JSC::ARMv7Assembler::relinkCall): 25674 (JSC::ARMv7Assembler::repatchInt32): 25675 (JSC::ARMv7Assembler::repatchPointer): 25676 (JSC::ARMv7Assembler::repatchLoadPtrToLEA): 25677 (JSC::ARMv7Assembler::setInt32): 25678 - Replace usage of MakeWritable with cacheFlush. 25679 25680 * assembler/LinkBuffer.h: 25681 (JSC::LinkBuffer::performFinalization): 25682 - Make explicit call to cacheFlush. 25683 25684 * assembler/MacroAssemblerCodeRef.h: 25685 (JSC::MacroAssemblerCodeRef::MacroAssemblerCodeRef): 25686 - Make size always available. 25687 25688 * assembler/RepatchBuffer.h: 25689 (JSC::RepatchBuffer::RepatchBuffer): 25690 (JSC::RepatchBuffer::~RepatchBuffer): 25691 - Add calls to MakeWritable & makeExecutable. 25692 25693 * assembler/X86Assembler.h: 25694 (JSC::X86Assembler::relinkJump): 25695 (JSC::X86Assembler::relinkCall): 25696 (JSC::X86Assembler::repatchInt32): 25697 (JSC::X86Assembler::repatchPointer): 25698 (JSC::X86Assembler::repatchLoadPtrToLEA): 25699 - Remove usage of MakeWritable. 25700 25701 * bytecode/CodeBlock.h: 25702 (JSC::CodeBlock::getJITCode): 25703 - Provide access to CodeBlock's JITCode. 25704 25705 * jit/ExecutableAllocator.h: 25706 (JSC::ExecutableAllocator::makeExecutable): 25707 (JSC::ExecutableAllocator::cacheFlush): 25708 - Remove MakeWritable, make cacheFlush public. 25709 25710 * jit/JIT.cpp: 25711 (JSC::ctiPatchNearCallByReturnAddress): 25712 (JSC::ctiPatchCallByReturnAddress): 25713 (JSC::JIT::privateCompile): 25714 (JSC::JIT::unlinkCall): 25715 (JSC::JIT::linkCall): 25716 - Add CodeBlock argument to RepatchBuffer. 25717 25718 * jit/JIT.h: 25719 - Pass CodeBlock argument for use by RepatchBuffer. 25720 25721 * jit/JITCode.h: 25722 (JSC::JITCode::start): 25723 (JSC::JITCode::size): 25724 - Provide access to code start & size. 25725 25726 * jit/JITPropertyAccess.cpp: 25727 (JSC::JIT::privateCompilePutByIdTransition): 25728 (JSC::JIT::patchGetByIdSelf): 25729 (JSC::JIT::patchMethodCallProto): 25730 (JSC::JIT::patchPutByIdReplace): 25731 (JSC::JIT::privateCompilePatchGetArrayLength): 25732 (JSC::JIT::privateCompileGetByIdProto): 25733 (JSC::JIT::privateCompileGetByIdSelfList): 25734 (JSC::JIT::privateCompileGetByIdProtoList): 25735 (JSC::JIT::privateCompileGetByIdChainList): 25736 (JSC::JIT::privateCompileGetByIdChain): 25737 - Add CodeBlock argument to RepatchBuffer. 25738 25739 * jit/JITStubs.cpp: 25740 (JSC::JITThunks::tryCachePutByID): 25741 (JSC::JITThunks::tryCacheGetByID): 25742 (JSC::JITStubs::DEFINE_STUB_FUNCTION): 25743 - Pass CodeBlock argument for use by RepatchBuffer. 25744 257452009-07-21 Zoltan Herczeg <zherczeg@inf.u-szeged.hu> 25746 25747 Reviewed by Gavin Barraclough. 25748 25749 Cache not only the structure of the method, but the 25750 structure of its prototype as well. 25751 https://bugs.webkit.org/show_bug.cgi?id=27077 25752 25753 * bytecode/CodeBlock.cpp: 25754 (JSC::CodeBlock::~CodeBlock): 25755 * bytecode/CodeBlock.h: 25756 (JSC::MethodCallLinkInfo::MethodCallLinkInfo): 25757 * jit/JITPropertyAccess.cpp: 25758 (JSC::JIT::patchMethodCallProto): 25759 257602009-07-21 Gavin Barraclough <barraclough@apple.com> 25761 25762 Reviewed by Sam Weinig. 25763 25764 Move call linking / repatching down from AbstractMacroAssembler into MacroAssemblerARCH classes. 25765 ( https://bugs.webkit.org/show_bug.cgi?id=27527 ) 25766 25767 This allows the implementation to be defined per architecture. Specifically this addresses the 25768 fact that x86-64 MacroAssembler implements far calls as a load to register, followed by a call 25769 to register. Patching the call actually requires the pointer load to be patched, rather than 25770 the call to be patched. This is implementation detail specific to MacroAssemblerX86_64, and as 25771 such is best handled there. 25772 25773 * assembler/AbstractMacroAssembler.h: 25774 * assembler/MacroAssemblerARM.h: 25775 (JSC::MacroAssemblerARM::linkCall): 25776 (JSC::MacroAssemblerARM::repatchCall): 25777 * assembler/MacroAssemblerARMv7.h: 25778 (JSC::MacroAssemblerARMv7::linkCall): 25779 (JSC::MacroAssemblerARMv7::repatchCall): 25780 * assembler/MacroAssemblerX86.h: 25781 (JSC::MacroAssemblerX86::linkCall): 25782 (JSC::MacroAssemblerX86::repatchCall): 25783 * assembler/MacroAssemblerX86_64.h: 25784 (JSC::MacroAssemblerX86_64::linkCall): 25785 (JSC::MacroAssemblerX86_64::repatchCall): 25786 257872009-07-21 Adam Treat <adam.treat@torchmobile.com> 25788 25789 Reviewed by George Staikos. 25790 25791 Every wtf file includes other wtf files with <> style includes 25792 except this one. Fix the exception. 25793 25794 * wtf/ByteArray.h: 25795 257962009-07-21 Gavin Barraclough <barraclough@apple.com> 25797 25798 Reviewed by Oliver Hunt. 25799 25800 Move LinkBuffer/RepatchBuffer out of AbstractMacroAssembler. 25801 ( https://bugs.webkit.org/show_bug.cgi?id=27485 ) 25802 25803 This change is the first step in a process to move code that should be in 25804 the architecture-specific MacroAssembler classes up out of Assmbler and 25805 AbstractMacroAssembler. 25806 25807 * JavaScriptCore.xcodeproj/project.pbxproj: 25808 - added new files 25809 25810 * assembler/ARMAssembler.h: 25811 (JSC::ARMAssembler::linkPointer): 25812 - rename patchPointer to bring it in line with the current link/repatch naming scheme 25813 25814 * assembler/ARMv7Assembler.h: 25815 (JSC::ARMv7Assembler::linkCall): 25816 (JSC::ARMv7Assembler::linkPointer): 25817 (JSC::ARMv7Assembler::relinkCall): 25818 (JSC::ARMv7Assembler::repatchInt32): 25819 (JSC::ARMv7Assembler::repatchPointer): 25820 (JSC::ARMv7Assembler::setInt32): 25821 (JSC::ARMv7Assembler::setPointer): 25822 - rename patchPointer to bring it in line with the current link/repatch naming scheme 25823 25824 * assembler/AbstractMacroAssembler.h: 25825 (JSC::AbstractMacroAssembler::linkJump): 25826 (JSC::AbstractMacroAssembler::linkCall): 25827 (JSC::AbstractMacroAssembler::linkPointer): 25828 (JSC::AbstractMacroAssembler::getLinkerAddress): 25829 (JSC::AbstractMacroAssembler::getLinkerCallReturnOffset): 25830 (JSC::AbstractMacroAssembler::repatchJump): 25831 (JSC::AbstractMacroAssembler::repatchCall): 25832 (JSC::AbstractMacroAssembler::repatchNearCall): 25833 (JSC::AbstractMacroAssembler::repatchInt32): 25834 (JSC::AbstractMacroAssembler::repatchPointer): 25835 (JSC::AbstractMacroAssembler::repatchLoadPtrToLEA): 25836 - remove the LinkBuffer/RepatchBuffer classes, but leave a set of (private, friended) methods to interface to the Assembler 25837 25838 * assembler/LinkBuffer.h: Added. 25839 (JSC::LinkBuffer::LinkBuffer): 25840 (JSC::LinkBuffer::~LinkBuffer): 25841 (JSC::LinkBuffer::link): 25842 (JSC::LinkBuffer::patch): 25843 (JSC::LinkBuffer::locationOf): 25844 (JSC::LinkBuffer::locationOfNearCall): 25845 (JSC::LinkBuffer::returnAddressOffset): 25846 (JSC::LinkBuffer::finalizeCode): 25847 (JSC::LinkBuffer::finalizeCodeAddendum): 25848 (JSC::LinkBuffer::code): 25849 (JSC::LinkBuffer::performFinalization): 25850 - new file containing the LinkBuffer class, previously a member of AbstractMacroAssembler 25851 25852 * assembler/RepatchBuffer.h: Added. 25853 (JSC::RepatchBuffer::RepatchBuffer): 25854 (JSC::RepatchBuffer::relink): 25855 (JSC::RepatchBuffer::repatch): 25856 (JSC::RepatchBuffer::repatchLoadPtrToLEA): 25857 (JSC::RepatchBuffer::relinkCallerToTrampoline): 25858 (JSC::RepatchBuffer::relinkCallerToFunction): 25859 (JSC::RepatchBuffer::relinkNearCallerToTrampoline): 25860 - new file containing the RepatchBuffer class, previously a member of AbstractMacroAssembler 25861 25862 * assembler/X86Assembler.h: 25863 (JSC::X86Assembler::linkJump): 25864 (JSC::X86Assembler::linkCall): 25865 (JSC::X86Assembler::linkPointerForCall): 25866 (JSC::X86Assembler::linkPointer): 25867 (JSC::X86Assembler::relinkJump): 25868 (JSC::X86Assembler::relinkCall): 25869 (JSC::X86Assembler::repatchInt32): 25870 (JSC::X86Assembler::repatchPointer): 25871 (JSC::X86Assembler::setPointer): 25872 (JSC::X86Assembler::setInt32): 25873 (JSC::X86Assembler::setRel32): 25874 - rename patchPointer to bring it in line with the current link/repatch naming scheme 25875 25876 * jit/JIT.cpp: 25877 (JSC::ctiPatchNearCallByReturnAddress): 25878 (JSC::ctiPatchCallByReturnAddress): 25879 - include new headers 25880 - remove MacroAssembler:: specification from RepatchBuffer usage 25881 25882 * jit/JITPropertyAccess.cpp: 25883 * yarr/RegexJIT.cpp: 25884 - include new headers 25885 258862009-07-21 Robert Agoston <Agoston.Robert@stud.u-szeged.hu> 25887 25888 Reviewed by David Levin. 25889 25890 Fixed #undef typo. 25891 https://bugs.webkit.org/show_bug.cgi?id=27506 25892 25893 * bytecode/Opcode.h: 25894 258952009-07-21 Adam Roben <aroben@apple.com> 25896 25897 Roll out r46153, r46154, and r46155 25898 25899 These changes were causing build failures and assertion failures on 25900 Windows. 25901 25902 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 25903 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 25904 * JavaScriptCore.xcodeproj/project.pbxproj: 25905 * runtime/JSArray.cpp: 25906 * runtime/StringPrototype.cpp: 25907 * runtime/UString.cpp: 25908 * runtime/UString.h: 25909 * wtf/FastMalloc.cpp: 25910 * wtf/FastMalloc.h: 25911 * wtf/Platform.h: 25912 * wtf/PossiblyNull.h: Removed. 25913 259142009-07-21 Roland Steiner <rolandsteiner@google.com> 25915 25916 Reviewed by David Levin. 25917 25918 Add ENABLE_RUBY to list of build options 25919 https://bugs.webkit.org/show_bug.cgi?id=27324 25920 25921 * Configurations/FeatureDefines.xcconfig: Added flag ENABLE_RUBY. 25922 259232009-07-20 Oliver Hunt <oliver@apple.com> 25924 25925 Build fix attempt #2 25926 25927 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 25928 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 25929 259302009-07-20 Oliver Hunt <oliver@apple.com> 25931 25932 Build fix attempt #1 25933 25934 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 25935 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 25936 259372009-07-20 Oliver Hunt <oliver@apple.com> 25938 25939 Reviewed by Gavin Barraclough. 25940 25941 Make it harder to misuse try* allocation routines 25942 https://bugs.webkit.org/show_bug.cgi?id=27469 25943 25944 Jump through a few hoops to make it much harder to accidentally 25945 miss null-checking of values returned by the try-* allocation 25946 routines. 25947 25948 * JavaScriptCore.xcodeproj/project.pbxproj: 25949 * runtime/JSArray.cpp: 25950 (JSC::JSArray::putSlowCase): 25951 (JSC::JSArray::increaseVectorLength): 25952 * runtime/StringPrototype.cpp: 25953 (JSC::stringProtoFuncFontsize): 25954 (JSC::stringProtoFuncLink): 25955 * runtime/UString.cpp: 25956 (JSC::allocChars): 25957 (JSC::reallocChars): 25958 (JSC::expandCapacity): 25959 (JSC::UString::Rep::reserveCapacity): 25960 (JSC::UString::expandPreCapacity): 25961 (JSC::createRep): 25962 (JSC::concatenate): 25963 (JSC::UString::spliceSubstringsWithSeparators): 25964 (JSC::UString::replaceRange): 25965 (JSC::UString::append): 25966 (JSC::UString::operator=): 25967 * runtime/UString.h: 25968 (JSC::UString::Rep::createEmptyBuffer): 25969 * wtf/FastMalloc.cpp: 25970 (WTF::tryFastZeroedMalloc): 25971 (WTF::tryFastMalloc): 25972 (WTF::tryFastCalloc): 25973 (WTF::tryFastRealloc): 25974 (WTF::TCMallocStats::tryFastMalloc): 25975 (WTF::TCMallocStats::tryFastCalloc): 25976 (WTF::TCMallocStats::tryFastRealloc): 25977 * wtf/FastMalloc.h: 25978 (WTF::TryMallocReturnValue::TryMallocReturnValue): 25979 (WTF::TryMallocReturnValue::~TryMallocReturnValue): 25980 (WTF::TryMallocReturnValue::operator Maybe<T>): 25981 (WTF::TryMallocReturnValue::getValue): 25982 * wtf/PossiblyNull.h: 25983 (WTF::PossiblyNull::PossiblyNull): 25984 (WTF::PossiblyNull::~PossiblyNull): 25985 (WTF::PossiblyNull::getValue): 25986 * wtf/Platform.h: 25987 259882009-07-20 Gavin Barraclough <barraclough@apple.com> 25989 25990 RS Oliver Hunt. 25991 25992 Add ARM assembler files to xcodeproj, for convenience editing. 25993 25994 * JavaScriptCore.xcodeproj/project.pbxproj: 25995 259962009-07-20 Jessie Berlin <jberlin@apple.com> 25997 25998 Reviewed by David Levin. 25999 26000 Fix an incorrect assertion in Vector::remove. 26001 26002 https://bugs.webkit.org/show_bug.cgi?id=27477 26003 26004 * wtf/Vector.h: 26005 (WTF::::remove): 26006 Assert that the position at which to start removing elements + the 26007 length (the number of elements to remove) is less than or equal to the 26008 size of the entire Vector. 26009 260102009-07-20 Peter Kasting <pkasting@google.com> 26011 26012 Reviewed by Mark Rowe. 26013 26014 https://bugs.webkit.org/show_bug.cgi?id=27468 26015 Back out r46060, which caused problems for some Apple developers. 26016 26017 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCoreCommon.vsprops: 26018 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCoreGenerated.vcproj: 26019 * JavaScriptCore.vcproj/WTF/WTFCommon.vsprops: 26020 * JavaScriptCore.vcproj/jsc/jscCommon.vsprops: 26021 * JavaScriptCore.vcproj/testapi/testapiCommon.vsprops: 26022 260232009-07-20 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 26024 26025 Reviewed by Oliver Hunt. 26026 26027 Allow custom memory allocation control in NewThreadContext 26028 https://bugs.webkit.org/show_bug.cgi?id=27338 26029 26030 Inherits NewThreadContext struct from FastAllocBase because it 26031 has been instantiated by 'new' JavaScriptCore/wtf/Threading.cpp:76. 26032 26033 * wtf/Threading.cpp: 26034 260352009-07-20 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 26036 26037 Reviewed by Oliver Hunt. 26038 26039 Allow custom memory allocation control in JavaScriptCore's JSClassRef.h 26040 https://bugs.webkit.org/show_bug.cgi?id=27340 26041 26042 Inherit StaticValueEntry and StaticFunctionEntry struct from FastAllocBase because these 26043 have been instantiated by 'new' in JavaScriptCore/API/JSClassRef.cpp:153 26044 and in JavaScriptCore/API/JSClassRef.cpp:166. 26045 26046 * API/JSClassRef.h: 26047 260482009-07-20 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 26049 26050 Reviewed by Darin Adler. 26051 26052 Allow custom memory allocation control in JavaScriptCore's RegexPattern.h 26053 https://bugs.webkit.org/show_bug.cgi?id=27343 26054 26055 Inherits RegexPattern.h's structs (which have been instantiated by operator new) from FastAllocBase: 26056 26057 CharacterClass (new call: JavaScriptCore/yarr/RegexCompiler.cpp:144) 26058 PatternAlternative (new call: JavaScriptCore/yarr/RegexPattern.h:221) 26059 PatternDisjunction (new call: JavaScriptCore/yarr/RegexCompiler.cpp:446) 26060 26061 * yarr/RegexPattern.h: 26062 260632009-07-20 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 26064 26065 Reviewed by Darin Adler. 26066 26067 Allow custom memory allocation control for JavaScriptCore's MatchFrame struct 26068 https://bugs.webkit.org/show_bug.cgi?id=27344 26069 26070 Inherits MatchFrame struct from FastAllocBase because it has 26071 been instantiated by 'new' JavaScriptCore/pcre/pcre_exec.cpp:359. 26072 26073 * pcre/pcre_exec.cpp: 26074 260752009-07-20 Laszlo Gombos <laszlo.1.gombos@nokia.com> 26076 26077 Reviewed by Holger Freyther. 26078 26079 Remove some outdated S60 platform specific code 26080 https://bugs.webkit.org/show_bug.cgi?id=27423 26081 26082 * wtf/Platform.h: 26083 260842009-07-20 Csaba Osztrogonac <oszi@inf.u-szeged.hu> 26085 26086 Reviewed by Simon Hausmann. 26087 26088 Qt build fix with MSVC and MinGW. 26089 26090 * jsc.pro: Make sure jsc is a console application, and turn off 26091 exceptions and stl support to fix the build. 26092 260932009-07-20 Xan Lopez <xlopez@igalia.com> 26094 26095 Reviewed by Gustavo Noronha. 26096 26097 Do not use C++-style comments in preprocessor directives. 26098 26099 GCC does not like this in some configurations, using C-style 26100 comments is safer. 26101 26102 * wtf/Platform.h: 26103 261042009-07-17 Peter Kasting <pkasting@google.com> 26105 26106 Reviewed by Steve Falkenburg. 26107 26108 https://bugs.webkit.org/show_bug.cgi?id=27323 26109 Only add Cygwin to the path when it isn't already there. This avoids 26110 causing problems for people who purposefully have non-Cygwin versions of 26111 executables like svn in front of the Cygwin ones in their paths. 26112 26113 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCoreCommon.vsprops: 26114 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCoreGenerated.vcproj: 26115 * JavaScriptCore.vcproj/WTF/WTFCommon.vsprops: 26116 * JavaScriptCore.vcproj/jsc/jscCommon.vsprops: 26117 * JavaScriptCore.vcproj/testapi/testapiCommon.vsprops: 26118 261192009-07-17 Gabor Loki <loki@inf.u-szeged.hu> 26120 26121 Reviewed by Gavin Barraclough. 26122 26123 Add YARR support for generic ARM platforms (disabled by default). 26124 https://bugs.webkit.org/show_bug.cgi?id=24986 26125 26126 Add generic ARM port for MacroAssembler. It supports the whole 26127 MacroAssembler functionality except floating point. 26128 26129 The class JmpSrc is extended with a flag which enables to patch 26130 the jump destination offset during execution. This feature is 26131 required for generic ARM port. 26132 26133 Signed off by Zoltan Herczeg <zherczeg@inf.u-szeged.hu> 26134 Signed off by Gabor Loki <loki@inf.u-szeged.hu> 26135 26136 * JavaScriptCore.pri: 26137 * assembler/ARMAssembler.cpp: Added. 26138 (JSC::ARMAssembler::getLdrImmAddress): 26139 (JSC::ARMAssembler::linkBranch): 26140 (JSC::ARMAssembler::patchConstantPoolLoad): 26141 (JSC::ARMAssembler::getOp2): 26142 (JSC::ARMAssembler::genInt): 26143 (JSC::ARMAssembler::getImm): 26144 (JSC::ARMAssembler::moveImm): 26145 (JSC::ARMAssembler::dataTransfer32): 26146 (JSC::ARMAssembler::baseIndexTransfer32): 26147 (JSC::ARMAssembler::executableCopy): 26148 * assembler/ARMAssembler.h: Added. 26149 (JSC::ARM::): 26150 (JSC::ARMAssembler::ARMAssembler): 26151 (JSC::ARMAssembler::): 26152 (JSC::ARMAssembler::JmpSrc::JmpSrc): 26153 (JSC::ARMAssembler::JmpSrc::enableLatePatch): 26154 (JSC::ARMAssembler::JmpDst::JmpDst): 26155 (JSC::ARMAssembler::JmpDst::isUsed): 26156 (JSC::ARMAssembler::JmpDst::used): 26157 (JSC::ARMAssembler::emitInst): 26158 (JSC::ARMAssembler::and_r): 26159 (JSC::ARMAssembler::ands_r): 26160 (JSC::ARMAssembler::eor_r): 26161 (JSC::ARMAssembler::eors_r): 26162 (JSC::ARMAssembler::sub_r): 26163 (JSC::ARMAssembler::subs_r): 26164 (JSC::ARMAssembler::rsb_r): 26165 (JSC::ARMAssembler::rsbs_r): 26166 (JSC::ARMAssembler::add_r): 26167 (JSC::ARMAssembler::adds_r): 26168 (JSC::ARMAssembler::adc_r): 26169 (JSC::ARMAssembler::adcs_r): 26170 (JSC::ARMAssembler::sbc_r): 26171 (JSC::ARMAssembler::sbcs_r): 26172 (JSC::ARMAssembler::rsc_r): 26173 (JSC::ARMAssembler::rscs_r): 26174 (JSC::ARMAssembler::tst_r): 26175 (JSC::ARMAssembler::teq_r): 26176 (JSC::ARMAssembler::cmp_r): 26177 (JSC::ARMAssembler::orr_r): 26178 (JSC::ARMAssembler::orrs_r): 26179 (JSC::ARMAssembler::mov_r): 26180 (JSC::ARMAssembler::movs_r): 26181 (JSC::ARMAssembler::bic_r): 26182 (JSC::ARMAssembler::bics_r): 26183 (JSC::ARMAssembler::mvn_r): 26184 (JSC::ARMAssembler::mvns_r): 26185 (JSC::ARMAssembler::mul_r): 26186 (JSC::ARMAssembler::muls_r): 26187 (JSC::ARMAssembler::mull_r): 26188 (JSC::ARMAssembler::ldr_imm): 26189 (JSC::ARMAssembler::ldr_un_imm): 26190 (JSC::ARMAssembler::dtr_u): 26191 (JSC::ARMAssembler::dtr_ur): 26192 (JSC::ARMAssembler::dtr_d): 26193 (JSC::ARMAssembler::dtr_dr): 26194 (JSC::ARMAssembler::ldrh_r): 26195 (JSC::ARMAssembler::ldrh_d): 26196 (JSC::ARMAssembler::ldrh_u): 26197 (JSC::ARMAssembler::strh_r): 26198 (JSC::ARMAssembler::push_r): 26199 (JSC::ARMAssembler::pop_r): 26200 (JSC::ARMAssembler::poke_r): 26201 (JSC::ARMAssembler::peek_r): 26202 (JSC::ARMAssembler::clz_r): 26203 (JSC::ARMAssembler::bkpt): 26204 (JSC::ARMAssembler::lsl): 26205 (JSC::ARMAssembler::lsr): 26206 (JSC::ARMAssembler::asr): 26207 (JSC::ARMAssembler::lsl_r): 26208 (JSC::ARMAssembler::lsr_r): 26209 (JSC::ARMAssembler::asr_r): 26210 (JSC::ARMAssembler::size): 26211 (JSC::ARMAssembler::ensureSpace): 26212 (JSC::ARMAssembler::label): 26213 (JSC::ARMAssembler::align): 26214 (JSC::ARMAssembler::jmp): 26215 (JSC::ARMAssembler::patchPointerInternal): 26216 (JSC::ARMAssembler::patchConstantPoolLoad): 26217 (JSC::ARMAssembler::patchPointer): 26218 (JSC::ARMAssembler::repatchInt32): 26219 (JSC::ARMAssembler::repatchPointer): 26220 (JSC::ARMAssembler::repatchLoadPtrToLEA): 26221 (JSC::ARMAssembler::linkJump): 26222 (JSC::ARMAssembler::relinkJump): 26223 (JSC::ARMAssembler::linkCall): 26224 (JSC::ARMAssembler::relinkCall): 26225 (JSC::ARMAssembler::getRelocatedAddress): 26226 (JSC::ARMAssembler::getDifferenceBetweenLabels): 26227 (JSC::ARMAssembler::getCallReturnOffset): 26228 (JSC::ARMAssembler::getOp2Byte): 26229 (JSC::ARMAssembler::placeConstantPoolBarrier): 26230 (JSC::ARMAssembler::RM): 26231 (JSC::ARMAssembler::RS): 26232 (JSC::ARMAssembler::RD): 26233 (JSC::ARMAssembler::RN): 26234 (JSC::ARMAssembler::getConditionalField): 26235 * assembler/ARMv7Assembler.h: 26236 (JSC::ARMv7Assembler::JmpSrc::enableLatePatch): 26237 * assembler/AbstractMacroAssembler.h: 26238 (JSC::AbstractMacroAssembler::Call::enableLatePatch): 26239 (JSC::AbstractMacroAssembler::Jump::enableLatePatch): 26240 * assembler/MacroAssembler.h: 26241 * assembler/MacroAssemblerARM.h: Added. 26242 (JSC::MacroAssemblerARM::): 26243 (JSC::MacroAssemblerARM::add32): 26244 (JSC::MacroAssemblerARM::and32): 26245 (JSC::MacroAssemblerARM::lshift32): 26246 (JSC::MacroAssemblerARM::mul32): 26247 (JSC::MacroAssemblerARM::not32): 26248 (JSC::MacroAssemblerARM::or32): 26249 (JSC::MacroAssemblerARM::rshift32): 26250 (JSC::MacroAssemblerARM::sub32): 26251 (JSC::MacroAssemblerARM::xor32): 26252 (JSC::MacroAssemblerARM::load32): 26253 (JSC::MacroAssemblerARM::load32WithAddressOffsetPatch): 26254 (JSC::MacroAssemblerARM::loadPtrWithPatchToLEA): 26255 (JSC::MacroAssemblerARM::load16): 26256 (JSC::MacroAssemblerARM::store32WithAddressOffsetPatch): 26257 (JSC::MacroAssemblerARM::store32): 26258 (JSC::MacroAssemblerARM::pop): 26259 (JSC::MacroAssemblerARM::push): 26260 (JSC::MacroAssemblerARM::move): 26261 (JSC::MacroAssemblerARM::swap): 26262 (JSC::MacroAssemblerARM::signExtend32ToPtr): 26263 (JSC::MacroAssemblerARM::zeroExtend32ToPtr): 26264 (JSC::MacroAssemblerARM::branch32): 26265 (JSC::MacroAssemblerARM::branch16): 26266 (JSC::MacroAssemblerARM::branchTest32): 26267 (JSC::MacroAssemblerARM::jump): 26268 (JSC::MacroAssemblerARM::branchAdd32): 26269 (JSC::MacroAssemblerARM::mull32): 26270 (JSC::MacroAssemblerARM::branchMul32): 26271 (JSC::MacroAssemblerARM::branchSub32): 26272 (JSC::MacroAssemblerARM::breakpoint): 26273 (JSC::MacroAssemblerARM::nearCall): 26274 (JSC::MacroAssemblerARM::call): 26275 (JSC::MacroAssemblerARM::ret): 26276 (JSC::MacroAssemblerARM::set32): 26277 (JSC::MacroAssemblerARM::setTest32): 26278 (JSC::MacroAssemblerARM::tailRecursiveCall): 26279 (JSC::MacroAssemblerARM::makeTailRecursiveCall): 26280 (JSC::MacroAssemblerARM::moveWithPatch): 26281 (JSC::MacroAssemblerARM::branchPtrWithPatch): 26282 (JSC::MacroAssemblerARM::storePtrWithPatch): 26283 (JSC::MacroAssemblerARM::supportsFloatingPoint): 26284 (JSC::MacroAssemblerARM::supportsFloatingPointTruncate): 26285 (JSC::MacroAssemblerARM::loadDouble): 26286 (JSC::MacroAssemblerARM::storeDouble): 26287 (JSC::MacroAssemblerARM::addDouble): 26288 (JSC::MacroAssemblerARM::subDouble): 26289 (JSC::MacroAssemblerARM::mulDouble): 26290 (JSC::MacroAssemblerARM::convertInt32ToDouble): 26291 (JSC::MacroAssemblerARM::branchDouble): 26292 (JSC::MacroAssemblerARM::branchTruncateDoubleToInt32): 26293 (JSC::MacroAssemblerARM::ARMCondition): 26294 (JSC::MacroAssemblerARM::prepareCall): 26295 (JSC::MacroAssemblerARM::call32): 26296 * assembler/X86Assembler.h: 26297 (JSC::X86Assembler::JmpSrc::enableLatePatch): 26298 * jit/ExecutableAllocator.h: 26299 (JSC::ExecutableAllocator::cacheFlush): 26300 * wtf/Platform.h: 26301 * yarr/RegexJIT.cpp: 26302 (JSC::Yarr::RegexGenerator::generateEnter): 26303 (JSC::Yarr::RegexGenerator::generateReturn): 26304 263052009-07-17 Gabor Loki <loki@inf.u-szeged.hu> 26306 26307 Reviewed by Gavin Barraclough. 26308 26309 Extend AssemblerBuffer with constant pool handling mechanism. 26310 https://bugs.webkit.org/show_bug.cgi?id=24986 26311 26312 Add a platform independed constant pool framework. 26313 This pool can store 32 or 64 bits values which is enough to hold 26314 any integer, pointer or double constant. 26315 26316 * assembler/AssemblerBuffer.h: 26317 (JSC::AssemblerBuffer::putIntUnchecked): 26318 (JSC::AssemblerBuffer::putInt64Unchecked): 26319 (JSC::AssemblerBuffer::append): 26320 (JSC::AssemblerBuffer::grow): 26321 * assembler/AssemblerBufferWithConstantPool.h: Added. 26322 (JSC::): 26323 263242009-07-17 Eric Roman <eroman@chromium.org> 26325 26326 Reviewed by Darin Adler. 26327 26328 Build fix for non-Darwin. 26329 Add a guard for inclusion of RetainPtr.h which includes CoreFoundation.h 26330 26331 https://bugs.webkit.org/show_bug.cgi?id=27382 26332 26333 * wtf/unicode/icu/CollatorICU.cpp: 26334 263352009-07-17 Alexey Proskuryakov <ap@webkit.org> 26336 26337 Reviewed by John Sullivan. 26338 26339 Get user default collation order via a CFLocale API when available. 26340 26341 * wtf/unicode/icu/CollatorICU.cpp: (WTF::Collator::userDefault): 26342 263432009-07-17 Laszlo Gombos <laszlo.1.gombos@nokia.com> 26344 26345 Reviewed by Simon Hausmann. 26346 26347 [Qt] Fix the include path for the Symbian port 26348 https://bugs.webkit.org/show_bug.cgi?id=27358 26349 26350 * JavaScriptCore.pri: 26351 263522009-07-17 Csaba Osztrogonac <oszi@inf.u-szeged.hu> 26353 26354 Reviewed by David Levin. 26355 26356 Build fix on platforms don't have MMAP. 26357 https://bugs.webkit.org/show_bug.cgi?id=27365 26358 26359 * interpreter/RegisterFile.h: Including stdio.h irrespectively of HAVE(MMAP) 26360 263612009-07-16 Fumitoshi Ukai <ukai@chromium.org> 26362 26363 Reviewed by David Levin. 26364 26365 Add --web-sockets flag and ENABLE_WEB_SOCKETS define. 26366 https://bugs.webkit.org/show_bug.cgi?id=27206 26367 26368 Add ENABLE_WEB_SOCKETS 26369 26370 * Configurations/FeatureDefines.xcconfig: add ENABLE_WEB_SOCKETS 26371 263722009-07-16 Maxime Simon <simon.maxime@gmail.com> 26373 26374 Reviewed by Eric Seidel. 26375 26376 Added Haiku-specific files for JavaScriptCore. 26377 https://bugs.webkit.org/show_bug.cgi?id=26620 26378 26379 * wtf/haiku/MainThreadHaiku.cpp: Added. 26380 (WTF::initializeMainThreadPlatform): 26381 (WTF::scheduleDispatchFunctionsOnMainThread): 26382 263832009-07-16 Gavin Barraclough <barraclough@apple.com> 26384 26385 RS by Oliver Hunt. 26386 26387 Revert r45969, this fix does not appear to be valid. 26388 https://bugs.webkit.org/show_bug.cgi?id=27077 26389 26390 * bytecode/CodeBlock.cpp: 26391 (JSC::CodeBlock::~CodeBlock): 26392 (JSC::CodeBlock::unlinkCallers): 26393 * jit/JIT.cpp: 26394 * jit/JIT.h: 26395 263962009-07-16 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 26397 26398 Reviewed by Oliver Hunt. 26399 26400 Allow custom memory allocation control in ExceptionInfo and RareData struct 26401 https://bugs.webkit.org/show_bug.cgi?id=27336 26402 26403 Inherits ExceptionInfo and RareData struct from FastAllocBase because these 26404 have been instantiated by 'new' in JavaScriptCore/bytecode/CodeBlock.cpp:1289 and 26405 in JavaScriptCore/bytecode/CodeBlock.h:453. 26406 26407 Remove unnecessary WTF:: namespace from CodeBlock inheritance. 26408 26409 * bytecode/CodeBlock.h: 26410 264112009-07-16 Mark Rowe <mrowe@apple.com> 26412 26413 Rubber-stamped by Geoff Garen. 26414 26415 Fix FeatureDefines.xcconfig to not be out of sync with the rest of the world. 26416 26417 * Configurations/FeatureDefines.xcconfig: 26418 264192009-07-16 Yong Li <yong.li@torchmobile.com> 26420 26421 Reviewed by George Staikos. 26422 26423 https://bugs.webkit.org/show_bug.cgi?id=27320 26424 _countof is only included in CE6; for CE5 we need to define it ourself 26425 26426 * wtf/Platform.h: 26427 264282009-07-16 Zoltan Herczeg <zherczeg@inf.u-szeged.hu> 26429 26430 Reviewed by Oliver Hunt. 26431 26432 Workers + garbage collector: weird crashes 26433 https://bugs.webkit.org/show_bug.cgi?id=27077 26434 26435 We need to unlink cached method call sites when a function is destroyed. 26436 26437 * JavaScriptCore.xcodeproj/project.pbxproj: 26438 * bytecode/CodeBlock.cpp: 26439 (JSC::CodeBlock::~CodeBlock): 26440 (JSC::CodeBlock::unlinkCallers): 26441 * jit/JIT.cpp: 26442 (JSC::JIT::unlinkMethodCall): 26443 * jit/JIT.h: 26444 264452009-07-15 Steve Falkenburg <sfalken@apple.com> 26446 26447 Windows Build fix. 26448 26449 Visual Studio reset our intermediate directory on us. 26450 This sets it back. 26451 26452 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: 26453 * JavaScriptCore.vcproj/testapi/testapi.vcproj: 26454 264552009-07-15 Kwang Yul Seo <skyul@company100.net> 26456 26457 Reviewed by Eric Seidel. 26458 26459 https://bugs.webkit.org/show_bug.cgi?id=26794 26460 Make Yacc-generated parsers to use fastMalloc/fastFree. 26461 26462 Define YYMALLOC and YYFREE to fastMalloc and fastFree 26463 respectively. 26464 26465 * parser/Grammar.y: 26466 264672009-07-15 Darin Adler <darin@apple.com> 26468 26469 Fix a build for a particular Apple configuration. 26470 26471 * wtf/FastAllocBase.h: Change include to use "" style for 26472 including another wtf header. This is the style we use for 26473 including other public headers in the same directory. 26474 264752009-07-15 George Staikos <george.staikos@torchmobile.com> 26476 26477 Reviewed by Adam Treat. 26478 26479 https://bugs.webkit.org/show_bug.cgi?id=27303 26480 Implement createThreadInternal for WinCE. 26481 Contains changes by George Staikos <george.staikos@torchmobile.com> and Joe Mason <joe.mason@torchmobile.com> 26482 26483 * wtf/ThreadingWin.cpp: 26484 (WTF::createThreadInternal): 26485 264862009-07-15 Joe Mason <joe.mason@torchmobile.com> 26487 26488 Reviewed by George Staikos. 26489 26490 https://bugs.webkit.org/show_bug.cgi?id=27298 26491 Platform defines for WINCE. 26492 Contains changes by Yong Li <yong.li@torchmobile.com>, 26493 George Staikos <george.staikos@torchmobile.com> and Joe Mason <joe.mason@torchmobile.com> 26494 26495 * wtf/Platform.h: 26496 264972009-07-15 Yong Li <yong.li@torchmobile.com> 26498 26499 Reviewed by Adam Treat. 26500 26501 https://bugs.webkit.org/show_bug.cgi?id=27306 26502 Use RegisterClass instead of RegisterClassEx on WinCE. 26503 26504 * wtf/win/MainThreadWin.cpp: 26505 (WTF::initializeMainThreadPlatform): 26506 265072009-07-15 Yong Li <yong.li@torchmobile.com> 26508 26509 Reviewed by George Staikos. 26510 26511 https://bugs.webkit.org/show_bug.cgi?id=27301 26512 Use OutputDebugStringW on WinCE since OutputDebugStringA is not supported 26513 Originally written by Yong Li <yong.li@torchmobile.com> and refactored by 26514 Joe Mason <joe.mason@torchmobile.com> 26515 26516 * wtf/Assertions.cpp: vprintf_stderr_common 26517 265182009-07-15 Yong Li <yong.li@torchmobile.com> 26519 26520 Reviewed by George Staikos. 26521 26522 https://bugs.webkit.org/show_bug.cgi?id=27020 26523 msToGregorianDateTime should set utcOffset to 0 when outputIsUTC is false 26524 26525 * wtf/DateMath.cpp: 26526 (WTF::gregorianDateTimeToMS): 26527 265282009-07-15 Laszlo Gombos <laszlo.1.gombos@nokia.com> 26529 26530 Reviewed by Simon Hausmann. 26531 26532 [Qt] Cleanup - Remove obsolete code from the make system 26533 https://bugs.webkit.org/show_bug.cgi?id=27299 26534 26535 * JavaScriptCore.pro: 26536 * jsc.pro: 26537 265382009-07-07 Norbert Leser <norbert.leser@nokia.com> 26539 26540 Reviewed by Simon Hausmann. 26541 26542 https://bugs.webkit.org/show_bug.cgi?id=27056 26543 26544 Alternate bool operator for codewarrior compiler (WINSCW). 26545 Compiler (latest b482) reports error for UnspecifiedBoolType construct: 26546 "illegal explicit conversion from 'WTF::OwnArrayPtr<JSC::Register>' to 'bool'" 26547 26548 Same fix as in r38391. 26549 26550 * JavaScriptCore/wtf/OwnArrayPtr.h: 26551 265522009-07-15 Norbert Leser <norbert.leser@nokia.com> 26553 26554 Reviewed by Darin Adler. 26555 26556 Qualify include path with wtf to fix compilation 26557 on Symbian. 26558 https://bugs.webkit.org/show_bug.cgi?id=27055 26559 26560 * interpreter/Interpreter.h: 26561 265622009-07-15 Laszlo Gombos <laszlo.1.gombos@nokia.com> 26563 26564 Reviewed by Dave Kilzer. 26565 26566 Turn off non-portable date manipulations for SYMBIAN 26567 https://bugs.webkit.org/show_bug.cgi?id=27064 26568 26569 Introduce HAVE(TM_GMTOFF), HAVE(TM_ZONE) and HAVE(TIMEGM) guards 26570 and place the rules for controlling the guards in Platform.h. 26571 Turn off these newly introduced guards for SYMBIAN. 26572 26573 * wtf/DateMath.cpp: 26574 (WTF::calculateUTCOffset): 26575 * wtf/DateMath.h: 26576 (WTF::GregorianDateTime::GregorianDateTime): 26577 (WTF::GregorianDateTime::operator tm): 26578 * wtf/Platform.h: 26579 265802009-07-15 Norbert Leser <norbert.leser@nokia.com> 26581 26582 Reviewed by Simon Hausmann. 26583 26584 Undef ASSERT on Symbian, to avoid excessive warnings 26585 https://bugs.webkit.org/show_bug.cgi?id=27052 26586 26587 * wtf/Assertions.h: 26588 265892009-07-15 Oliver Hunt <oliver@apple.com> 26590 26591 Reviewed by Simon Hausmann. 26592 26593 REGRESSION: fast/js/postfix-syntax.html fails with interpreter 26594 https://bugs.webkit.org/show_bug.cgi?id=27294 26595 26596 When postfix operators operating on locals assign to the same local 26597 the order of operations has to be to store the incremented value, then 26598 store the unmodified number. Rather than implementing this subtle 26599 semantic in the interpreter I've just made the logic explicit in the 26600 bytecode generator, so x=x++ effectively becomes x=ToNumber(x) (for a 26601 local var x). 26602 26603 * parser/Nodes.cpp: 26604 (JSC::emitPostIncOrDec): 26605 266062009-07-15 Oliver Hunt <oliver@apple.com> 26607 26608 Reviewed by Simon Hausmann. 26609 26610 REGRESSION(43559): fast/js/kde/arguments-scope.html fails with interpreter 26611 https://bugs.webkit.org/show_bug.cgi?id=27259 26612 26613 The interpreter was incorrectly basing its need to create the arguments object 26614 based on the presence of the callframe's argument reference rather than the local 26615 arguments reference. Based on this it then overrode the local variable reference. 26616 26617 * interpreter/Interpreter.cpp: 26618 (JSC::Interpreter::privateExecute): 26619 266202009-07-14 Steve Falkenburg <sfalken@apple.com> 26621 26622 Reorganize JavaScriptCore headers into: 26623 API: include/JavaScriptCore/ 26624 Private: include/private/JavaScriptCore/ 26625 26626 Reviewed by Darin Adler. 26627 26628 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: 26629 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCoreCommon.vsprops: 26630 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCoreGenerated.make: 26631 * JavaScriptCore.vcproj/testapi/testapi.vcproj: 26632 * JavaScriptCore.vcproj/testapi/testapiCommon.vsprops: 26633 266342009-07-14 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 26635 26636 Reviewed by Darin Adler. 26637 26638 Change JSCell's superclass to NoncopyableCustomAllocated 26639 https://bugs.webkit.org/show_bug.cgi?id=27248 26640 26641 JSCell class customizes operator new, since Noncopyable will be 26642 inherited from FastAllocBase, NoncopyableCustomAllocated has 26643 to be used. 26644 26645 * runtime/JSCell.h: 26646 266472009-07-14 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 26648 26649 Reviewed by Darin Adler. 26650 26651 Change all Noncopyable inheriting visibility to public. 26652 https://bugs.webkit.org/show_bug.cgi?id=27225 26653 26654 Change all Noncopyable inheriting visibility to public because 26655 it is needed to the custom allocation framework (bug #20422). 26656 26657 * bytecode/SamplingTool.h: 26658 * bytecompiler/RegisterID.h: 26659 * interpreter/CachedCall.h: 26660 * interpreter/RegisterFile.h: 26661 * parser/Lexer.h: 26662 * parser/Parser.h: 26663 * runtime/ArgList.h: 26664 * runtime/BatchedTransitionOptimizer.h: 26665 * runtime/Collector.h: 26666 * runtime/CommonIdentifiers.h: 26667 * runtime/JSCell.h: 26668 * runtime/JSGlobalObject.h: 26669 * runtime/JSLock.h: 26670 * runtime/JSONObject.cpp: 26671 * runtime/SmallStrings.cpp: 26672 * runtime/SmallStrings.h: 26673 * wtf/CrossThreadRefCounted.h: 26674 * wtf/GOwnPtr.h: 26675 * wtf/Locker.h: 26676 * wtf/MessageQueue.h: 26677 * wtf/OwnArrayPtr.h: 26678 * wtf/OwnFastMallocPtr.h: 26679 * wtf/OwnPtr.h: 26680 * wtf/RefCounted.h: 26681 * wtf/ThreadSpecific.h: 26682 * wtf/Threading.h: 26683 * wtf/Vector.h: 26684 * wtf/unicode/Collator.h: 26685 266862009-07-14 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 26687 26688 Reviewed by Darin Adler. 26689 26690 Change ParserArenaRefCounted's superclass to RefCountedCustomAllocated 26691 https://bugs.webkit.org/show_bug.cgi?id=27249 26692 26693 ParserArenaDeletable customizes operator new, to avoid double inheritance 26694 ParserArenaDeletable's superclass has been changed to RefCountedCustomAllocated. 26695 26696 * parser/Nodes.h: 26697 266982009-07-14 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 26699 26700 Reviewed by Darin Adler. 26701 26702 Add RefCountedCustomAllocated to RefCounted.h 26703 https://bugs.webkit.org/show_bug.cgi?id=27232 26704 26705 Some class which are inherited from RefCounted customize 26706 operator new, but RefCounted is inherited from Noncopyable 26707 which will be inherited from FastAllocBase. To avoid 26708 conflicts Noncopyable inheriting was moved down to RefCounted 26709 and to avoid double inheritance this class has been added. 26710 26711 * wtf/RefCounted.h: 26712 (WTF::RefCountedCustomAllocated::deref): 26713 (WTF::RefCountedCustomAllocated::~RefCountedCustomAllocated): 26714 267152009-07-14 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 26716 26717 Reviewed by Darin Adler. 26718 26719 Add NoncopyableCustomAllocated to Noncopyable.h. 26720 https://bugs.webkit.org/show_bug.cgi?id=27228 26721 26722 Some classes which inherited from Noncopyable overrides operator new 26723 since Noncopyable'll be inherited from FastAllocBase, Noncopyable.h 26724 needs to be extended with this new class to support the overriding. 26725 26726 * wtf/Noncopyable.h: 26727 (WTFNoncopyable::NoncopyableCustomAllocated::NoncopyableCustomAllocated): 26728 (WTFNoncopyable::NoncopyableCustomAllocated::~NoncopyableCustomAllocated): 26729 267302009-07-14 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 26731 26732 Reviewed by Darin Adler. 26733 26734 Allow custom memory allocation control for JavaScriptCore's IdentifierTable class 26735 https://bugs.webkit.org/show_bug.cgi?id=27260 26736 26737 Inherits IdentifierTable class from FastAllocBase because it has been 26738 instantiated by 'new' in JavaScriptCore/runtime/Identifier.cpp:70. 26739 26740 * runtime/Identifier.cpp: 26741 267422009-07-14 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 26743 26744 Reviewed by Darin Adler. 26745 26746 Allow custom memory allocation control for JavaScriptCore's Profiler class 26747 https://bugs.webkit.org/show_bug.cgi?id=27253 26748 26749 Inherits Profiler class from FastAllocBase because it has been instantiated by 26750 'new' in JavaScriptCore/profiler/Profiler.cpp:56. 26751 26752 * profiler/Profiler.h: 26753 267542009-07-06 George Staikos <george.staikos@torchmobile.com> 26755 26756 Reviewed by Adam Treat. 26757 26758 Authors: George Staikos <george.staikos@torchmobile.com>, Joe Mason <joe.mason@torchmobile.com>, Makoto Matsumoto <matumoto@math.keio.ac.jp>, Takuji Nishimura 26759 26760 https://bugs.webkit.org/show_bug.cgi?id=27030 26761 Implement custom RNG for WinCE using Mersenne Twister 26762 26763 * wtf/RandomNumber.cpp: 26764 (WTF::randomNumber): 26765 * wtf/RandomNumberSeed.h: 26766 (WTF::initializeRandomNumberGenerator): 26767 * wtf/wince/mt19937ar.c: Added. 26768 (init_genrand): 26769 (init_by_array): 26770 (genrand_int32): 26771 (genrand_int31): 26772 (genrand_real1): 26773 (genrand_real2): 26774 (genrand_real3): 26775 (genrand_res53): 26776 267772009-07-13 Gustavo Noronha Silva <gustavo.noronha@collabora.co.uk> 26778 26779 Unreviewed make dist build fix. 26780 26781 * GNUmakefile.am: 26782 267832009-07-13 Drew Wilson <atwilson@google.com> 26784 26785 Reviewed by David Levin. 26786 26787 Add ENABLE(SHARED_WORKERS) flag and define SharedWorker APIs 26788 https://bugs.webkit.org/show_bug.cgi?id=26932 26789 26790 Added ENABLE(SHARED_WORKERS) flag (off by default). 26791 26792 * Configurations/FeatureDefines.xcconfig: 26793 267942009-07-07 Norbert Leser <norbert.leser@nokia.com> 26795 26796 Reviewed by Maciej Stachoviak. 26797 26798 https://bugs.webkit.org/show_bug.cgi?id=27058 26799 26800 Removed superfluous parenthesis around single expression. 26801 Compilers on Symbian platform fail to properly parse and compile. 26802 26803 * JavaScriptCore/wtf/Platform.h: 26804 268052009-07-13 Norbert Leser <norbert.leser@nokia.com> 26806 26807 Reviewed by Maciej Stachoviak. 26808 26809 https://bugs.webkit.org/show_bug.cgi?id=27054 26810 26811 Renamed Translator to HashTranslator 26812 26813 Codewarrior compiler (WINSCW) latest b482 cannot resolve typename 26814 mismatch between template declaration and definition 26815 (HashTranslator / Translator) 26816 26817 * wtf/HashSet.h: 26818 268192009-07-13 Norbert Leser <norbert.leser@nokia.com> 26820 26821 Reviewed by Eric Seidel. 26822 26823 https://bugs.webkit.org/show_bug.cgi?id=27053 26824 26825 Ambiguity in LabelScope initialization 26826 26827 Codewarrior compiler (WINSCW) latest b482 on Symbian cannot resolve 26828 type of "0" unambiguously. Set expression explicitly to 26829 PassRefPtr<Label>::PassRefPtr() 26830 26831 * bytecompiler/BytecodeGenerator.cpp 26832 268332009-07-11 Simon Fraser <simon.fraser@apple.com> 26834 26835 Enable support for accelerated compositing and 3d transforms on Leopard. 26836 <https://bugs.webkit.org/show_bug.cgi?id=20166> 26837 <rdar://problem/6120614> 26838 26839 Reviewed by Oliver Hunt. 26840 26841 * Configurations/FeatureDefines.xcconfig: 26842 * wtf/Platform.h: 26843 268442009-07-10 Mark Rowe <mrowe@apple.com> 26845 26846 Second part of the "make Windows happier" dance. 26847 26848 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 26849 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 26850 268512009-07-10 Mark Rowe <mrowe@apple.com> 26852 26853 Try and make the Windows build happy. 26854 26855 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 26856 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 26857 268582009-07-10 Kevin McCullough <kmccullough@apple.com> 26859 26860 Reviewed by Geoffrey Garen. 26861 26862 * debugger/Debugger.h: Made this function virtual for use in WebCore's 26863 WebInspector. 26864 268652009-07-10 Kwang Yul Seo <skyul@company100.net> 26866 26867 Reviewed by Darin Adler. 26868 26869 ParserArenaDeletable should override delete 26870 https://bugs.webkit.org/show_bug.cgi?id=26790 26871 26872 ParserArenaDeletable overrides new, but it does not override delete. 26873 ParserArenaDeletable must be freed by fastFree 26874 because it is allocated by fastMalloc. 26875 26876 * parser/NodeConstructors.h: 26877 (JSC::ParserArenaDeletable::operator delete): 26878 * parser/Nodes.h: 26879 268802009-07-10 Adam Roben <aroben@apple.com> 26881 26882 Sort all our Xcode projects 26883 26884 Accomplished using sort-Xcode-project-file. 26885 26886 Requested by Dave Kilzer. 26887 26888 * JavaScriptCore.xcodeproj/project.pbxproj: 26889 268902009-07-09 Maciej Stachowiak <mjs@apple.com> 26891 26892 Not reviewed, build fix. 26893 26894 Windows build fix for the last change. 26895 26896 * wtf/dtoa.cpp: Forgot to include Vector.h 26897 268982009-07-09 Maciej Stachowiak <mjs@apple.com> 26899 26900 Reviewed by Darin Adler. 26901 26902 REGRESSION: crash in edge cases of floating point parsing. 26903 https://bugs.webkit.org/show_bug.cgi?id=27110 26904 <rdar://problem/7044458> 26905 26906 Tests: fast/css/number-parsing-crash.html 26907 fast/css/number-parsing-crash.html 26908 fast/js/number-parsing-crash.html 26909 26910 * wtf/dtoa.cpp: 26911 (WTF::BigInt::BigInt): Converted this to more a proper class, using a Vector 26912 with inline capacity 26913 26914 (WTF::lshift): Rearranged logic somewhat nontrivially to deal with the new way of sizing BigInts. 26915 Added an assertion to verify that invariants are maintained. 26916 26917 All other functions are adapted fairly mechanically to the above changes. 26918 (WTF::BigInt::clear): 26919 (WTF::BigInt::size): 26920 (WTF::BigInt::resize): 26921 (WTF::BigInt::words): 26922 (WTF::BigInt::append): 26923 (WTF::multadd): 26924 (WTF::s2b): 26925 (WTF::i2b): 26926 (WTF::mult): 26927 (WTF::cmp): 26928 (WTF::diff): 26929 (WTF::b2d): 26930 (WTF::d2b): 26931 (WTF::ratio): 26932 (WTF::strtod): 26933 (WTF::quorem): 26934 (WTF::dtoa): 26935 269362009-07-09 Drew Wilson <atwilson@google.com> 26937 26938 Reviewed by Alexey Proskuryakov. 26939 26940 Turned on CHANNEL_MESSAGING by default because the MessageChannel API 26941 can now be implemented for Web Workers and is reasonably stable. 26942 26943 * Configurations/FeatureDefines.xcconfig: 26944 269452009-07-09 Oliver Hunt <oliver@apple.com> 26946 26947 * interpreter/Interpreter.cpp: 26948 (JSC::Interpreter::privateExecute): 26949 269502009-07-09 Oliver Hunt <oliver@apple.com> 26951 26952 Reviewed by Darin Adler. 26953 26954 Bug 27016 - Interpreter crashes due to invalid array indexes 26955 <https://bugs.webkit.org/show_bug.cgi?id=27016> 26956 26957 Unsigned vs signed conversions results in incorrect behaviour in 26958 64bit interpreter builds. 26959 26960 * interpreter/Interpreter.cpp: 26961 (JSC::Interpreter::privateExecute): 26962 269632009-07-09 Dimitri Glazkov <dglazkov@chromium.org> 26964 26965 Reviewed by Darin Fisher. 26966 26967 [Chromium] Upstream JavaScriptCore.gypi, the project file for Chromium build. 26968 https://bugs.webkit.org/show_bug.cgi?id=27135 26969 26970 * JavaScriptCore.gypi: Added. 26971 269722009-07-09 Joe Mason <joe.mason@torchmobile.com> 26973 26974 Reviewed by George Staikos. 26975 26976 Authors: Yong Li <yong.li@torchmobile.com>, Joe Mason <joe.mason@torchmobile.com> 26977 26978 https://bugs.webkit.org/show_bug.cgi?id=27031 26979 Add an override for deleteOwnedPtr(HDC) on Windows 26980 26981 * wtf/OwnPtrCommon.h: 26982 * wtf/OwnPtrWin.cpp: 26983 (WTF::deleteOwnedPtr): 26984 269852009-07-09 Laszlo Gombos <laszlo.1.gombos@nokia.com> 26986 26987 Reviewed by Darin Adler. 26988 26989 Guard singal.h dependency with HAVE(SIGNAL_H) to enable building jsc 26990 on SYMBIAN. 26991 26992 https://bugs.webkit.org/show_bug.cgi?id=27026 26993 26994 Based on Norbert Leser's work. 26995 26996 * jsc.cpp: 26997 (printUsageStatement): 26998 (parseArguments): 26999 * wtf/Platform.h: 27000 270012009-07-07 Gavin Barraclough <barraclough@apple.com> 27002 27003 Reviewed by Sam Weinig. 27004 27005 Stop loading constants into the register file. 27006 27007 Instead, use high register values (highest bit bar the sign bit set) to indicate 27008 constants in the instruction stream, and when we encounter such a value load it 27009 directly from the CodeBlock. 27010 27011 Since constants are no longer copied into the register file, this patch renders 27012 the 'unexpected constant' mechanism redundant, and removes it. 27013 27014 2% improvement, thanks to Sam Weinig. 27015 27016 * bytecode/CodeBlock.cpp: 27017 (JSC::CodeBlock::dump): 27018 (JSC::CodeBlock::CodeBlock): 27019 (JSC::CodeBlock::mark): 27020 (JSC::CodeBlock::shrinkToFit): 27021 * bytecode/CodeBlock.h: 27022 (JSC::CodeBlock::isTemporaryRegisterIndex): 27023 (JSC::CodeBlock::constantRegister): 27024 (JSC::CodeBlock::isConstantRegisterIndex): 27025 (JSC::CodeBlock::getConstant): 27026 (JSC::ExecState::r): 27027 * bytecode/Opcode.h: 27028 * bytecompiler/BytecodeGenerator.cpp: 27029 (JSC::BytecodeGenerator::preserveLastVar): 27030 (JSC::BytecodeGenerator::BytecodeGenerator): 27031 (JSC::BytecodeGenerator::addConstantValue): 27032 (JSC::BytecodeGenerator::emitEqualityOp): 27033 (JSC::BytecodeGenerator::emitLoad): 27034 (JSC::BytecodeGenerator::emitResolveBase): 27035 (JSC::BytecodeGenerator::emitResolveWithBase): 27036 (JSC::BytecodeGenerator::emitNewError): 27037 * bytecompiler/BytecodeGenerator.h: 27038 (JSC::BytecodeGenerator::emitNode): 27039 * interpreter/CallFrame.h: 27040 (JSC::ExecState::noCaller): 27041 (JSC::ExecState::hasHostCallFrameFlag): 27042 (JSC::ExecState::addHostCallFrameFlag): 27043 (JSC::ExecState::removeHostCallFrameFlag): 27044 * interpreter/Interpreter.cpp: 27045 (JSC::Interpreter::resolve): 27046 (JSC::Interpreter::resolveSkip): 27047 (JSC::Interpreter::resolveGlobal): 27048 (JSC::Interpreter::resolveBase): 27049 (JSC::Interpreter::resolveBaseAndProperty): 27050 (JSC::Interpreter::resolveBaseAndFunc): 27051 (JSC::Interpreter::dumpRegisters): 27052 (JSC::Interpreter::throwException): 27053 (JSC::Interpreter::createExceptionScope): 27054 (JSC::Interpreter::privateExecute): 27055 (JSC::Interpreter::retrieveArguments): 27056 * jit/JIT.cpp: 27057 (JSC::JIT::privateCompileMainPass): 27058 * jit/JITInlineMethods.h: 27059 (JSC::JIT::emitLoadDouble): 27060 (JSC::JIT::emitLoadInt32ToDouble): 27061 * jit/JITOpcodes.cpp: 27062 (JSC::JIT::emit_op_new_error): 27063 (JSC::JIT::emit_op_enter): 27064 (JSC::JIT::emit_op_enter_with_activation): 27065 * parser/Nodes.cpp: 27066 (JSC::DeleteResolveNode::emitBytecode): 27067 (JSC::DeleteValueNode::emitBytecode): 27068 (JSC::PrefixResolveNode::emitBytecode): 27069 * runtime/JSActivation.cpp: 27070 (JSC::JSActivation::JSActivation): 27071 * wtf/Platform.h: 27072 270732009-07-07 Mark Rowe <mrowe@apple.com> 27074 27075 Reviewed by Darin Adler. 27076 27077 Fix <https://bugs.webkit.org/show_bug.cgi?id=27025> / <rdar://problem/7033448>. 27078 Bug 27025: Crashes and regression test failures related to regexps in 64-bit 27079 27080 For x86_64 RegexGenerator uses rbx, a callee-save register, as a scratch register but 27081 neglects to save and restore it. The change in handling of the output vector in r45545 27082 altered code generation so that the RegExp::match was now storing important data in rbx, 27083 which caused crashes and bogus results when it was clobbered. 27084 27085 * yarr/RegexJIT.cpp: 27086 (JSC::Yarr::RegexGenerator::generateEnter): Save rbx. 27087 (JSC::Yarr::RegexGenerator::generateReturn): Restore rbx. 27088 270892009-07-06 Ada Chan <adachan@apple.com> 27090 27091 Reviewed by Darin Adler and Mark Rowe. 27092 27093 Decommitted spans are added to the list of normal spans rather than 27094 the returned spans in TCMalloc_PageHeap::Delete(). 27095 https://bugs.webkit.org/show_bug.cgi?id=26998 27096 27097 In TCMalloc_PageHeap::Delete(), the deleted span can be decommitted in 27098 the process of merging with neighboring spans that are also decommitted. 27099 The merged span needs to be placed in the list of returned spans (spans 27100 whose memory has been returned to the system). Right now it's always added 27101 to the list of the normal spans which can theoretically cause thrashing. 27102 27103 * wtf/FastMalloc.cpp: 27104 (WTF::TCMalloc_PageHeap::Delete): 27105 271062009-07-05 Lars Knoll <lars.knoll@nokia.com> 27107 27108 Reviewed by Maciej Stachowiak. 27109 27110 https://bugs.webkit.org/show_bug.cgi?id=26843 27111 27112 Fix run-time crashes in JavaScriptCore with the Metrowerks compiler on Symbian. 27113 27114 The Metrowerks compiler on the Symbian platform moves the globally 27115 defined Hashtables into read-only memory, despite one of the members 27116 being mutable. This causes crashes at run-time due to write access to 27117 read-only memory. 27118 27119 Avoid the use of const with this compiler by introducing the 27120 JSC_CONST_HASHTABLE macro. 27121 27122 Based on idea by Norbert Leser. 27123 27124 * runtime/Lookup.h: Define JSC_CONST_HASHTABLE as const for !WINSCW. 27125 * create_hash_table: Use JSC_CONST_HASHTABLE for hashtables. 27126 * runtime/JSGlobalData.cpp: Import various global hashtables via the macro. 27127 271282009-07-04 Dan Bernstein <mitz@apple.com> 27129 27130 - debug build fix 27131 27132 * runtime/RegExpConstructor.cpp: 27133 (JSC::RegExpConstructor::getLastParen): 27134 271352009-07-03 Yong Li <yong.li@torchmobile.com> 27136 27137 Reviewed by Maciej Stachowiak (and revised slightly) 27138 27139 RegExp::match to be optimized 27140 https://bugs.webkit.org/show_bug.cgi?id=26957 27141 27142 Allow regexp matching to use Vectors with inline capacity instead of 27143 allocating a new ovector buffer every time. 27144 27145 ~5% speedup on SunSpider string-unpack-code test, 0.3% on SunSpider overall. 27146 27147 * runtime/RegExp.cpp: 27148 (JSC::RegExp::match): 27149 * runtime/RegExp.h: 27150 * runtime/RegExpConstructor.cpp: 27151 (JSC::RegExpConstructorPrivate::RegExpConstructorPrivate): 27152 (JSC::RegExpConstructorPrivate::lastOvector): 27153 (JSC::RegExpConstructorPrivate::tempOvector): 27154 (JSC::RegExpConstructorPrivate::changeLastOvector): 27155 (JSC::RegExpConstructor::performMatch): 27156 (JSC::RegExpMatchesArray::RegExpMatchesArray): 27157 (JSC::RegExpMatchesArray::fillArrayInstance): 27158 (JSC::RegExpConstructor::getBackref): 27159 (JSC::RegExpConstructor::getLastParen): 27160 (JSC::RegExpConstructor::getLeftContext): 27161 (JSC::RegExpConstructor::getRightContext): 27162 * runtime/StringPrototype.cpp: 27163 (JSC::stringProtoFuncSplit): 27164 271652009-06-30 Kwang Yul Seo <skyul@company100.net> 27166 27167 Reviewed by Eric Seidel. 27168 27169 Override operator new/delete with const std::nothrow_t& as the second 27170 argument. 27171 https://bugs.webkit.org/show_bug.cgi?id=26792 27172 27173 On Windows CE, operator new/delete, new[]/delete[] with const 27174 std::nothrow_t& must be overrided because some standard template 27175 libraries use these operators. 27176 27177 The problem occurs when memory allocated by new(size_t s, const 27178 std::nothrow_t&) is freed by delete(void* p). This causes the umatched 27179 malloc/free problem. 27180 27181 The patch overrides all new, delete, new[] and delete[] to use 27182 fastMaloc and fastFree consistently. 27183 27184 * wtf/FastMalloc.h: 27185 (throw): 27186 271872009-06-30 Gabor Loki <loki@inf.u-szeged.hu> 27188 27189 Reviewed by Sam Weinig. 27190 27191 <https://bugs.webkit.org/show_bug.cgi?id=24986> 27192 27193 Remove unnecessary references to AssemblerBuffer. 27194 27195 * interpreter/Interpreter.cpp: 27196 * interpreter/Interpreter.h: 27197 271982009-06-29 David Levin <levin@chromium.org> 27199 27200 Reviewed by Oliver Hunt. 27201 27202 Still seeing occasional leaks from UString::sharedBuffer code 27203 https://bugs.webkit.org/show_bug.cgi?id=26420 27204 27205 The problem is that the pointer to the memory allocation isn't visible 27206 by "leaks" due to the lower bits being used as flags. The fix is to 27207 make the pointer visible in memory (in debug only). The downside of 27208 this fix that the memory allocated by sharedBuffer will still look like 27209 a leak in non-debug builds when any flags are set. 27210 27211 * wtf/PtrAndFlags.h: 27212 (WTF::PtrAndFlags::set): 27213 272142009-06-29 Sam Weinig <sam@webkit.org> 27215 27216 Reviewed by Mark Rowe. 27217 27218 Remove more unused scons support. 27219 27220 * SConstruct: Removed. 27221 272222009-06-29 Oliver Hunt <oliver@apple.com> 27223 27224 Reviewed by Gavin Barraclough. 27225 27226 <rdar://problem/7016214> JSON.parse fails to parse valid JSON with most Unicode characters 27227 <https://bugs.webkit.org/show_bug.cgi?id=26802> 27228 27229 In the original JSON.parse patch unicode was handled correctly, however in some last 27230 minute "clean up" I oversimplified isSafeStringCharacter. This patch corrects this bug. 27231 27232 * runtime/LiteralParser.cpp: 27233 (JSC::isSafeStringCharacter): 27234 (JSC::LiteralParser::Lexer::lexString): 27235 272362009-06-26 Oliver Hunt <oliver@apple.com> 27237 27238 Reviewed by Dan Bernstein. 27239 27240 <rdar://problem/7009684> REGRESSION(r45039): Crashes inside JSEvent::put on PowerPC (26746) 27241 <https://bugs.webkit.org/show_bug.cgi?id=26746> 27242 27243 Fix for r45039 incorrectly uncached a get_by_id by converting it to put_by_id. Clearly this 27244 is less than correct. This patch corrects that error. 27245 27246 * interpreter/Interpreter.cpp: 27247 (JSC::Interpreter::tryCacheGetByID): 27248 272492009-06-26 Eric Seidel <eric@webkit.org> 27250 27251 No review, only rolling out r45259. 27252 27253 Roll out r45259 after crash appeared on the bots: 27254 plugins/undefined-property-crash.html 27255 ASSERTION FAILED: s <= HeapConstants<heapType>::cellSize 27256 (leopard-intel-debug-tests/build/JavaScriptCore/runtime/Collector.cpp:278 27257 void* JSC::Heap::heapAllocate(size_t) [with JSC::HeapType heapType = PrimaryHeap]) 27258 27259 * runtime/DateInstance.cpp: 27260 * runtime/Identifier.cpp: 27261 * runtime/Lookup.h: 27262 * runtime/RegExpConstructor.cpp: 27263 * runtime/RegExpObject.h: 27264 * runtime/ScopeChain.h: 27265 * runtime/UString.h: 27266 272672009-06-26 Jedrzej Nowacki <jedrzej.nowacki@nokia.com> 27268 27269 Reviewed by Simon Hausmann. 27270 27271 Add support for QDataStream operators to Vector. 27272 27273 * wtf/Vector.h: 27274 (WTF::operator<<): 27275 (WTF::operator>>): 27276 272772009-06-24 Sam Weinig <sam@webkit.org> 27278 27279 Reviewed by Gavin Barraclough. 27280 27281 Make the opcode sampler work once again. 27282 27283 * jit/JIT.h: 27284 (JSC::JIT::compileGetByIdProto): 27285 (JSC::JIT::compileGetByIdSelfList): 27286 (JSC::JIT::compileGetByIdProtoList): 27287 (JSC::JIT::compileGetByIdChainList): 27288 (JSC::JIT::compileGetByIdChain): 27289 (JSC::JIT::compilePutByIdTransition): 27290 (JSC::JIT::compileCTIMachineTrampolines): 27291 (JSC::JIT::compilePatchGetArrayLength): 27292 * jit/JITStubCall.h: 27293 (JSC::JITStubCall::call): 27294 272952009-06-24 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 27296 27297 Reviewed by Maciej Stachowiak. 27298 27299 Extend FastAllocBase.h with 'using WTF::FastAllocBase' to avoid 27300 unnecessary WTF:: usings. 27301 Remove existing unnecessary WTF:: usings. 27302 27303 * interpreter/Interpreter.h: 27304 * profiler/CallIdentifier.h: 27305 * runtime/ScopeChain.h: 27306 * wtf/FastAllocBase.h: 27307 273082009-06-24 David Levin <levin@chromium.org> 27309 27310 Fix all builds. 27311 27312 * bytecode/CodeBlock.h: 27313 * bytecompiler/BytecodeGenerator.h: 27314 * interpreter/Register.h: 27315 273162009-06-24 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 27317 27318 Reviewed by Maciej Stachowiak. 27319 27320 https://bugs.webkit.org/show_bug.cgi?id=26677 27321 27322 Inherits CodeBlock class from FastAllocBase because it 27323 has been instantiated by 'new' in JavaScriptCore/bytecode/CodeBlock.h:217. 27324 27325 * bytecode/CodeBlock.h: 27326 273272009-06-24 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 27328 27329 Reviewed by Maciej Stachowiak. 27330 27331 https://bugs.webkit.org/show_bug.cgi?id=26676 27332 27333 Inherits BytecodeGenerator class from FastAllocBase because it has been 27334 instantiated by 'new' in JavaScriptCore/parser/Nodes.cpp:1892. 27335 27336 * bytecompiler/BytecodeGenerator.h: 27337 273382009-06-24 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 27339 27340 Reviewed by Maciej Stachowiak. 27341 27342 https://bugs.webkit.org/show_bug.cgi?id=26675 27343 27344 Inherits Register class from FastAllocBase because it has been 27345 instantiated by 'new' in JavaScriptCore/runtime/JSVariableObject.h:149. 27346 27347 * interpreter/Register.h: 27348 273492009-06-24 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 27350 27351 Reviewed by Darin Adler. 27352 27353 https://bugs.webkit.org/show_bug.cgi?id=26674 27354 27355 Inherits HashMap class from FastAllocBase because it has been 27356 instantiated by 'new' in JavaScriptCore/runtime/Structure.cpp:458. 27357 27358 * wtf/HashMap.h: 27359 273602009-06-24 Oliver Hunt <oliver@apple.com> 27361 27362 Reviewed by Darin Adler. 27363 27364 <rdar://problem/6940519> REGRESSION (Safari 4 Public Beta - TOT): google.com/adplanner shows blank page instead of site details in "basic research' 27365 27366 The problem was caused by the page returned with a function using a 27367 var declaration list containing around ~3000 variables. The solution 27368 to this is to flatten the comma expression representation and make 27369 codegen comma expressions and initializer lists iterative rather than 27370 recursive. 27371 27372 * parser/Grammar.y: 27373 * parser/NodeConstructors.h: 27374 (JSC::CommaNode::CommaNode): 27375 * parser/Nodes.cpp: 27376 (JSC::CommaNode::emitBytecode): 27377 * parser/Nodes.h: 27378 (JSC::ExpressionNode::isCommaNode): 27379 (JSC::CommaNode::isCommaNode): 27380 (JSC::CommaNode::append): 27381 273822009-06-24 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 27383 27384 Reviewed by Maciej Stachowiak. 27385 27386 https://bugs.webkit.org/show_bug.cgi?id=26645 27387 27388 Inherits ScopeChainNode class from FastAllocBase because it has been 27389 instantiated by 'new' in JavaScriptCore/runtime/ScopeChain.h:95. 27390 27391 * wtf/RefPtr.h: 27392 273932009-06-24 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 27394 27395 Reviewed by Darin Adler. 27396 27397 https://bugs.webkit.org/show_bug.cgi?id=26648 27398 27399 Inherits Deque class from FastAllocBase because it has been 27400 instantiated by 'new' with DEFINE_STATIC_LOCAL macro in 27401 JavaScriptCore/wtf/MainThread.cpp:62. 27402 27403 * wtf/Deque.h: 27404 274052009-06-24 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 27406 27407 Reviewed by Darin Adler. 27408 27409 https://bugs.webkit.org/show_bug.cgi?id=26644 27410 27411 Inherits RefPtr class from FastAllocBase because it has been 27412 instantiated by 'new' in JavaScriptCore/runtime/StructureChain.cpp:41. 27413 27414 * wtf/RefPtr.h: 27415 274162009-06-24 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 27417 27418 Reviewed by Darin Adler. 27419 27420 Inherits HashSet class from FastAllocBase, because it has been 27421 instantiated by 'new' in JavaScriptCore/runtime/Collector.h:116. 27422 27423 * wtf/HashSet.h: 27424 274252009-06-24 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 27426 27427 Reviewed by Darin Adler. 27428 27429 Inherits Vector class from FastAllocBase because it has been 27430 instantiated by 'new' in JavaScriptCore/runtime/Structure.cpp:633. 27431 27432 * wtf/Vector.h: 27433 274342009-06-24 Norbert Leser <norbert.leser@nokia.com> 27435 27436 Reviewed by Maciej Stachoviak. 27437 27438 The BytecodeGenerator objects were instantiated on stack, which takes up ~38kB per instance 27439 (each instance includes copy of JSC::CodeBlock with large SymbolTable, etc.). 27440 Specifically, since there is nested invocation (e.g., GlobalCode --> FunctionCode), 27441 the stack overflows immediately on Symbian hardware (max. 80 kB). 27442 Proposed change allocates generator objects on heap. 27443 Performance impact (if any) should be negligible and change is proposed as general fix, 27444 rather than ifdef'd for SYMBIAN. 27445 27446 * parser/Nodes.cpp: 27447 (JSC::ProgramNode::generateBytecode): 27448 (JSC::EvalNode::generateBytecode): 27449 (JSC::EvalNode::bytecodeForExceptionInfoReparse): 27450 (JSC::FunctionBodyNode::generateBytecode): 27451 (JSC::FunctionBodyNode::bytecodeForExceptionInfoReparse): 27452 274532009-06-23 Oliver Hunt <oliver@apple.com> 27454 27455 Reviewed by Gavin Barraclough. 27456 27457 <rdar://problem/6992806> REGRESSION: Enumeration can skip new properties in cases of prototypes that have more than 64 (26593) 27458 <https://bugs.webkit.org/show_bug.cgi?id=26593> 27459 27460 Do not attempt to cache structure chains if they contain a dictionary at any level. 27461 27462 * interpreter/Interpreter.cpp: 27463 (JSC::Interpreter::tryCachePutByID): 27464 (JSC::Interpreter::tryCacheGetByID): 27465 * jit/JITStubs.cpp: 27466 (JSC::JITThunks::tryCachePutByID): 27467 * runtime/Structure.cpp: 27468 (JSC::Structure::getEnumerablePropertyNames): 27469 (JSC::Structure::addPropertyTransition): 27470 * runtime/StructureChain.cpp: 27471 (JSC::StructureChain::isCacheable): 27472 * runtime/StructureChain.h: 27473 274742009-06-23 Yong Li <yong.li@torchmobile.com> 27475 27476 Reviewed by George Staikos. 27477 27478 https://bugs.webkit.org/show_bug.cgi?id=26654 27479 Add the proper export define for the JavaScriptCore API when building for WINCE. 27480 27481 * API/JSBase.h: 27482 274832009-06-23 Joe Mason <joe.mason@torchmobile.com> 27484 27485 Reviewed by Adam Treat. 27486 27487 Authors: Yong Li <yong.li@torchmobile.com>, Joe Mason <joe.mason@torchmobile.com> 27488 27489 https://bugs.webkit.org/show_bug.cgi?id=26611 27490 Implement currentThreadStackBase on WINCE by adding a global, 27491 g_stackBase, which must be set to the address of a local variable 27492 by the caller before calling any WebKit function that invokes JSC. 27493 27494 * runtime/Collector.cpp: 27495 (JSC::isPageWritable): 27496 (JSC::getStackBase): 27497 Starts at the top of the stack and returns the entire range of 27498 consecutive writable pages as an estimate of the actual stack. 27499 This will be much bigger than the actual stack range, so some 27500 dead objects can't be collected, but it guarantees live objects 27501 aren't collected prematurely. 27502 27503 (JSC::currentThreadStackBase): 27504 On WinCE, returns g_stackBase if set or call getStackBase as a 27505 fallback if not. 27506 275072009-06-23 Oliver Hunt <oliver@apple.com> 27508 27509 Reviewed by Alexey Proskuryakov. 27510 27511 Fix stupid performance problem in the LiteralParser 27512 27513 The LiteralParser was making a new UString in order to use 27514 toDouble, however UString's toDouble allows a much wider range 27515 of numberic strings than the LiteralParser accepts, and requires 27516 an additional heap allocation or two for the construciton of the 27517 UString. To rectify this we just call WTF::dtoa directly using 27518 a stack allocated buffer to hold the validated numeric literal. 27519 27520 * runtime/LiteralParser.cpp: 27521 (JSC::LiteralParser::Lexer::lexNumber): 27522 (JSC::LiteralParser::parse): 27523 * runtime/LiteralParser.h: 27524 275252009-06-22 Oliver Hunt <oliver@apple.com> 27526 27527 Reviewed by Alexey Proskuryakov. 27528 27529 Bug 26640: JSON.stringify needs to special case Boolean objects 27530 <https://bugs.webkit.org/show_bug.cgi?id=26640> 27531 27532 Add special case handling of the Boolean object so we match current 27533 ES5 errata. 27534 27535 * runtime/JSONObject.cpp: 27536 (JSC::unwrapBoxedPrimitive): renamed from unwrapNumberOrString 27537 (JSC::gap): 27538 (JSC::Stringifier::appendStringifiedValue): 27539 275402009-06-22 Oliver Hunt <oliver@apple.com> 27541 27542 Reviewed by Darin Adler. 27543 27544 Bug 26591: Support revivers in JSON.parse 27545 <https://bugs.webkit.org/show_bug.cgi?id=26591> 27546 27547 Add reviver support to JSON.parse. This completes the JSON object. 27548 27549 * runtime/JSONObject.cpp: 27550 (JSC::Walker::Walker): 27551 (JSC::Walker::callReviver): 27552 (JSC::Walker::walk): 27553 (JSC::JSONProtoFuncParse): 27554 275552009-06-21 Oliver Hunt <oliver@apple.com> 27556 27557 Reviewed by Darin Adler. 27558 27559 Bug 26592: Support standard toJSON functions 27560 <https://bugs.webkit.org/show_bug.cgi?id=26592> 27561 27562 Add support for the standard Date.toJSON function. 27563 27564 * runtime/DatePrototype.cpp: 27565 (JSC::dateProtoFuncToJSON): 27566 275672009-06-21 Oliver Hunt <oliver@apple.com> 27568 27569 Reviewed by Sam Weinig. 27570 27571 Bug 26594: JSC needs to support Date.toISOString 27572 <https://bugs.webkit.org/show_bug.cgi?id=26594> 27573 27574 Add support for Date.toISOString. 27575 27576 * runtime/DatePrototype.cpp: 27577 (JSC::dateProtoFuncToISOString): 27578 275792009-06-21 Oliver Hunt <oliver@apple.com> 27580 27581 Remove dead code. 27582 27583 * runtime/LiteralParser.cpp: 27584 (JSC::LiteralParser::parse): 27585 275862009-06-21 Oliver Hunt <oliver@apple.com> 27587 27588 Reviewed by Darin Adler and Cameron Zwarich. 27589 27590 Bug 26587: Support JSON.parse 27591 <https://bugs.webkit.org/show_bug.cgi?id=26587> 27592 27593 Extend the LiteralParser to support the full strict JSON 27594 grammar, fix a few places where the grammar was incorrectly 27595 lenient. Doesn't yet support the JSON.parse reviver function 27596 but that does not block the JSON.parse functionality itself. 27597 27598 * interpreter/Interpreter.cpp: 27599 (JSC::Interpreter::callEval): 27600 * runtime/JSGlobalObjectFunctions.cpp: 27601 (JSC::globalFuncEval): 27602 * runtime/JSONObject.cpp: 27603 (JSC::JSONProtoFuncParse): 27604 * runtime/LiteralParser.cpp: 27605 (JSC::LiteralParser::Lexer::lex): 27606 (JSC::isSafeStringCharacter): 27607 (JSC::LiteralParser::Lexer::lexString): 27608 (JSC::LiteralParser::parse): 27609 * runtime/LiteralParser.h: 27610 (JSC::LiteralParser::LiteralParser): 27611 (JSC::LiteralParser::tryJSONParse): 27612 (JSC::LiteralParser::): 27613 (JSC::LiteralParser::Lexer::Lexer): 27614 276152009-06-21 David Levin <levin@chromium.org> 27616 27617 Reviewed by NOBODY (speculative build fix for windows). 27618 27619 Simply removed some whitespace form this file to make windows build wtf and 27620 hopefully copy the new MessageQueque.h so that WebCore picks it up. 27621 27622 * wtf/Assertions.cpp: 27623 276242009-06-21 Drew Wilson <atwilson@google.com> 27625 27626 Reviewed by David Levin. 27627 27628 <https://bugs.webkit.org/show_bug.cgi?id=25043> 27629 Added support for multi-threaded MessagePorts. 27630 27631 * wtf/MessageQueue.h: 27632 (WTF::::appendAndCheckEmpty): 27633 Added API to test whether the queue was empty before adding an element. 27634 276352009-06-20 David D. Kilzer <ddkilzer@webkit.org> 27636 27637 Fix namespace comment in SegmentedVector.h 27638 27639 * wtf/SegmentedVector.h: Updated namespace comment to reflect 27640 new namespace after r44897. 27641 276422009-06-20 Zoltan Herczeg <zherczeg@inf.u-szeged.hu> 27643 27644 Bug 24986: ARM JIT port 27645 <https://bugs.webkit.org/show_bug.cgi?id=24986> 27646 27647 Reviewed by Oliver Hunt. 27648 27649 An Iterator added for SegmentedVector. Currently 27650 only the pre ++ operator is supported. 27651 27652 * wtf/SegmentedVector.h: 27653 (WTF::SegmentedVectorIterator::~SegmentedVectorIterator): 27654 (WTF::SegmentedVectorIterator::operator*): 27655 (WTF::SegmentedVectorIterator::operator->): 27656 (WTF::SegmentedVectorIterator::operator++): 27657 (WTF::SegmentedVectorIterator::operator==): 27658 (WTF::SegmentedVectorIterator::operator!=): 27659 (WTF::SegmentedVectorIterator::operator=): 27660 (WTF::SegmentedVectorIterator::SegmentedVectorIterator): 27661 (WTF::SegmentedVector::alloc): 27662 (WTF::SegmentedVector::begin): 27663 (WTF::SegmentedVector::end): 27664 276652009-06-20 Zoltan Herczeg <zherczeg@inf.u-szeged.hu> 27666 27667 Bug 24986: ARM JIT port 27668 <https://bugs.webkit.org/show_bug.cgi?id=24986> 27669 27670 Reviewed by Oliver Hunt. 27671 27672 Move SegmentedVector to /wtf subdirectory 27673 and change "namespace JSC" to "namespace WTF" 27674 27675 Additional build file updates by David Kilzer. 27676 27677 * GNUmakefile.am: Updated path to SegmentedVector.h. 27678 * JavaScriptCore.order: Updated SegmentedVector namespace from 27679 JSC to WTF in mangled C++ method name. 27680 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: 27681 Removed reference to bytecompiler\SegmentedVector.h. 27682 * JavaScriptCore.vcproj/WTF/WTF.vcproj: Added reference to 27683 wtf\SegmentedVector.h. 27684 * JavaScriptCore.xcodeproj/project.pbxproj: Moved 27685 SegmentedVector.h definition from bytecompiler subdirectory to 27686 wtf subdirectory. 27687 * bytecompiler/BytecodeGenerator.h: Updated #include path to 27688 SegmentedVector.h and prepended WTF:: namespace to its use. 27689 * parser/Lexer.h: Ditto. 27690 * wtf/SegmentedVector.h: Renamed from JavaScriptCore/bytecompiler/SegmentedVector.h. 27691 (WTF::SegmentedVector::SegmentedVector): 27692 (WTF::SegmentedVector::~SegmentedVector): 27693 (WTF::SegmentedVector::size): 27694 (WTF::SegmentedVector::at): 27695 (WTF::SegmentedVector::operator[]): 27696 (WTF::SegmentedVector::last): 27697 (WTF::SegmentedVector::append): 27698 (WTF::SegmentedVector::removeLast): 27699 (WTF::SegmentedVector::grow): 27700 (WTF::SegmentedVector::clear): 27701 (WTF::SegmentedVector::deleteAllSegments): 27702 (WTF::SegmentedVector::segmentExistsFor): 27703 (WTF::SegmentedVector::segmentFor): 27704 (WTF::SegmentedVector::subscriptFor): 27705 (WTF::SegmentedVector::ensureSegmentsFor): 27706 (WTF::SegmentedVector::ensureSegment): 27707 277082009-06-19 Gavin Barraclough <barraclough@apple.com> 27709 27710 Reviewed by NOBODY (build fix take 2 - rename FIELD_OFFSET to something that doesn't conflict with winnt.h). 27711 27712 * jit/JIT.cpp: 27713 (JSC::JIT::privateCompile): 27714 (JSC::JIT::privateCompileCTIMachineTrampolines): 27715 (JSC::JIT::emitGetVariableObjectRegister): 27716 (JSC::JIT::emitPutVariableObjectRegister): 27717 * jit/JIT.h: 27718 * jit/JITArithmetic.cpp: 27719 (JSC::JIT::emit_op_rshift): 27720 (JSC::JIT::emitSlow_op_jnless): 27721 (JSC::JIT::emitSlow_op_jnlesseq): 27722 (JSC::JIT::compileBinaryArithOp): 27723 * jit/JITCall.cpp: 27724 (JSC::JIT::compileOpCallInitializeCallFrame): 27725 (JSC::JIT::compileOpCall): 27726 * jit/JITInlineMethods.h: 27727 (JSC::JIT::restoreArgumentReference): 27728 (JSC::JIT::checkStructure): 27729 * jit/JITOpcodes.cpp: 27730 (JSC::JIT::emit_op_instanceof): 27731 (JSC::JIT::emit_op_get_scoped_var): 27732 (JSC::JIT::emit_op_put_scoped_var): 27733 (JSC::JIT::emit_op_construct_verify): 27734 (JSC::JIT::emit_op_resolve_global): 27735 (JSC::JIT::emit_op_jeq_null): 27736 (JSC::JIT::emit_op_jneq_null): 27737 (JSC::JIT::emit_op_to_jsnumber): 27738 (JSC::JIT::emit_op_catch): 27739 (JSC::JIT::emit_op_eq_null): 27740 (JSC::JIT::emit_op_neq_null): 27741 (JSC::JIT::emit_op_convert_this): 27742 (JSC::JIT::emit_op_profile_will_call): 27743 (JSC::JIT::emit_op_profile_did_call): 27744 (JSC::JIT::emitSlow_op_get_by_val): 27745 * jit/JITPropertyAccess.cpp: 27746 (JSC::JIT::emit_op_get_by_val): 27747 (JSC::JIT::emit_op_put_by_val): 27748 (JSC::JIT::emit_op_method_check): 27749 (JSC::JIT::compileGetByIdHotPath): 27750 (JSC::JIT::emit_op_put_by_id): 27751 (JSC::JIT::compilePutDirectOffset): 27752 (JSC::JIT::compileGetDirectOffset): 27753 (JSC::JIT::privateCompilePutByIdTransition): 27754 (JSC::JIT::privateCompilePatchGetArrayLength): 27755 * jit/JITStubs.cpp: 27756 (JSC::JITThunks::JITThunks): 27757 277582009-06-19 Gavin Barraclough <barraclough@apple.com> 27759 27760 Reviewed by NOBODY (Windows build fix). 27761 27762 * jit/JIT.h: 27763 * jit/JITInlineMethods.h: 27764 277652009-06-19 Gabor Loki <loki@inf.u-szeged.hu> 27766 27767 Reviewed by Gavin Barraclough. 27768 27769 Reorganize ARM architecture specific macros. 27770 Use PLATFORM_ARM_ARCH(7) instead of PLATFORM(ARM_V7). 27771 27772 Bug 24986: ARM JIT port 27773 <https://bugs.webkit.org/show_bug.cgi?id=24986> 27774 27775 * assembler/ARMv7Assembler.h: 27776 * assembler/AbstractMacroAssembler.h: 27777 (JSC::AbstractMacroAssembler::Imm32::Imm32): 27778 * assembler/MacroAssembler.h: 27779 * assembler/MacroAssemblerCodeRef.h: 27780 (JSC::MacroAssemblerCodePtr::MacroAssemblerCodePtr): 27781 * jit/ExecutableAllocator.h: 27782 (JSC::ExecutableAllocator::cacheFlush): 27783 * jit/JIT.h: 27784 * jit/JITInlineMethods.h: 27785 (JSC::JIT::restoreArgumentReferenceForTrampoline): 27786 * jit/JITStubs.cpp: 27787 * jit/JITStubs.h: 27788 * wtf/Platform.h: 27789 * yarr/RegexJIT.cpp: 27790 (JSC::Yarr::RegexGenerator::generateEnter): 27791 (JSC::Yarr::RegexGenerator::generateReturn): 27792 277932009-06-19 Gavin Barraclough <barraclough@apple.com> 27794 27795 Reviewed by Oliver Hunt. 27796 27797 Fix armv7 JIT build issues. 27798 27799 Unfortunate the arm compiler does not like the use of offsetof on JITStackFrame (since it now contains non POD types), 27800 and the FIELD_OFFSET macro does not appear constantish enough for it to be happy with its use in COMPILE_ASSERT macros. 27801 27802 * Replace offsetofs with FIELD_OFFSETs (safe on C++ objects). 27803 * Move COMPILE_ASSERTs defending layout of JITStackFrame structure on armv7 into JITThunks constructor. 27804 27805 * jit/JIT.cpp: 27806 * jit/JIT.h: 27807 * jit/JITInlineMethods.h: 27808 (JSC::JIT::restoreArgumentReference): 27809 * jit/JITOpcodes.cpp: 27810 (JSC::JIT::emit_op_catch): 27811 * jit/JITStubs.cpp: 27812 (JSC::JITThunks::JITThunks): 27813 278142009-06-19 Adam Treat <adam.treat@torchmobile.com> 27815 27816 Blind attempt at build fix. 27817 27818 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 27819 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 27820 278212009-06-19 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 27822 27823 Reviewed by Oliver Hunt. 27824 27825 Inherits CallIdentifier struct from FastAllocBase because it has been 27826 instantiated by 'new' in JavaScriptCore/profiler/CallIdentifier.h:86. 27827 27828 * wtf/HashCountedSet.h: 27829 278302009-06-19 Adam Treat <adam.treat@torchmobile.com> 27831 27832 Reviewed by Oliver Hunt. 27833 27834 https://bugs.webkit.org/show_bug.cgi?id=26540 27835 Modify the test shell to add a new function 'checkSyntax' that will 27836 only parse the source instead of executing it. In this way we can test 27837 pure parsing performance against some of the larger scripts in the wild. 27838 27839 * jsc.cpp: 27840 (GlobalObject::GlobalObject): 27841 (functionCheckSyntax): 27842 278432009-06-19 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 27844 27845 Reviewed by Darin Adler. 27846 27847 Inherits HashCountedSet class from FastAllocBase because it has been 27848 instantiated by 'new' in JavaScriptCore/runtime/Collector.cpp:1095. 27849 27850 * wtf/HashCountedSet.h: 27851 278522009-06-19 Yong Li <yong.li@torchmobile.com> 27853 27854 Reviewed by George Staikos. 27855 27856 https://bugs.webkit.org/show_bug.cgi?id=26558 27857 Declare these symbols extern for WINCE as they are provided by libce. 27858 27859 * runtime/DateConstructor.cpp: 27860 * runtime/DatePrototype.cpp: 27861 (JSC::formatLocaleDate): 27862 278632009-06-19 Oliver Hunt <oliver@apple.com> 27864 27865 Reviewed by Maciej Stachowiak. 27866 27867 <rdar://problem/6988973> ScopeChain leak in interpreter builds 27868 27869 Move the Scopechain destruction code in JSFunction outside of the ENABLE(JIT) 27870 path. 27871 27872 * runtime/JSFunction.cpp: 27873 (JSC::JSFunction::~JSFunction): 27874 * wtf/Platform.h: 27875 278762009-06-19 Yong Li <yong.li@torchmobile.com> 27877 27878 Reviewed by George Staikos. 27879 27880 https://bugs.webkit.org/show_bug.cgi?id=26543 27881 Windows CE uses 'GetLastError' instead of 'errno.' 27882 27883 * interpreter/RegisterFile.h: 27884 (JSC::RegisterFile::RegisterFile): 27885 (JSC::RegisterFile::grow): 27886 278872009-06-19 David Levin <levin@chromium.org> 27888 27889 Reviewed by NOBODY (Windows build fix). 27890 27891 Add export for Windows corresponding to OSX export done in r44844. 27892 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 27893 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 27894 278952009-06-18 Oliver Hunt <oliver@apple.com> 27896 27897 Reviewed by Gavin "Viceroy of Venezuela" Barraclough. 27898 27899 Bug 26532: Native functions do not correctly unlink from optimised callsites when they're collected 27900 <https://bugs.webkit.org/show_bug.cgi?id=26532> <rdar://problem/6625385> 27901 27902 We need to make sure that each native function instance correctly unlinks any references to it 27903 when it is collected. Allowing this to happen required a few changes: 27904 * Every native function needs a codeblock to track the link information 27905 * To have this codeblock, every function now also needs its own functionbodynode 27906 so we no longer get to have a single shared instance. 27907 * Identifying a host function is now done by looking for CodeBlock::codeType() == NativeCode 27908 27909 * JavaScriptCore.exp: 27910 * bytecode/CodeBlock.cpp: 27911 (JSC::CodeBlock::CodeBlock): 27912 Constructor for NativeCode CodeBlock 27913 (JSC::CodeBlock::derefStructures): 27914 (JSC::CodeBlock::refStructures): 27915 (JSC::CodeBlock::reparseForExceptionInfoIfNecessary): 27916 (JSC::CodeBlock::handlerForBytecodeOffset): 27917 (JSC::CodeBlock::lineNumberForBytecodeOffset): 27918 (JSC::CodeBlock::expressionRangeForBytecodeOffset): 27919 (JSC::CodeBlock::getByIdExceptionInfoForBytecodeOffset): 27920 (JSC::CodeBlock::functionRegisterForBytecodeOffset): 27921 (JSC::CodeBlock::hasGlobalResolveInstructionAtBytecodeOffset): 27922 (JSC::CodeBlock::hasGlobalResolveInfoAtBytecodeOffset): 27923 (JSC::CodeBlock::setJITCode): 27924 Add assertions to ensure we don't try and use NativeCode CodeBlocks as 27925 a normal codeblock. 27926 27927 * bytecode/CodeBlock.h: 27928 (JSC::): 27929 (JSC::CodeBlock::source): 27930 (JSC::CodeBlock::sourceOffset): 27931 (JSC::CodeBlock::evalCodeCache): 27932 (JSC::CodeBlock::createRareDataIfNecessary): 27933 More assertions. 27934 27935 * jit/JIT.cpp: 27936 (JSC::JIT::privateCompileCTIMachineTrampolines): 27937 (JSC::JIT::linkCall): 27938 Update logic to allow native function caching 27939 27940 * jit/JITStubs.cpp: 27941 * parser/Nodes.cpp: 27942 (JSC::FunctionBodyNode::createNativeThunk): 27943 (JSC::FunctionBodyNode::isHostFunction): 27944 * parser/Nodes.h: 27945 * runtime/JSFunction.cpp: 27946 (JSC::JSFunction::JSFunction): 27947 (JSC::JSFunction::~JSFunction): 27948 (JSC::JSFunction::mark): 27949 * runtime/JSGlobalData.cpp: 27950 (JSC::JSGlobalData::~JSGlobalData): 27951 * runtime/JSGlobalData.h: 27952 279532009-06-18 Gavin Barraclough <barraclough@apple.com> 27954 27955 Reviewed by NOBODY (Windows build fix). 27956 27957 * wtf/DateMath.cpp: 27958 (WTF::calculateUTCOffset): 27959 279602009-06-18 Gavin Barraclough <barraclough@apple.com> 27961 27962 Reviewed by Geoff Garen. 27963 27964 Timezone calculation incorrect in Venezuela. 27965 27966 https://bugs.webkit.org/show_bug.cgi?id=26531 27967 <rdar://problem/6646169> Time is incorrectly reported to JavaScript in both Safari 3 and Firefox 3 27968 27969 The problem is that we're calculating the timezone relative to 01/01/2000, 27970 but the VET timezone changed from -4 hours to -4:30 hours on 12/09/2007. 27971 According to the spec, section 15.9.1.9 states "the time since the beginning 27972 of the year", presumably meaning the *current* year. Change the calculation 27973 to be based on whatever the current year is, rather than a canned date. 27974 27975 No performance impact. 27976 27977 * wtf/DateMath.cpp: 27978 (WTF::calculateUTCOffset): 27979 279802009-06-18 Gavin Barraclough <barraclough@apple.com> 27981 27982 Rubber Stamped by Mark Rowe (originally reviewed by Sam Weinig). 27983 27984 (Reintroducing patch added in r44492, and reverted in r44796.) 27985 27986 Change the implementation of op_throw so the stub function always modifies its 27987 return address - if it doesn't find a 'catch' it will switch to a trampoline 27988 to force a return from JIT execution. This saves memory, by avoiding the need 27989 for a unique return for every op_throw. 27990 27991 * jit/JITOpcodes.cpp: 27992 (JSC::JIT::emit_op_throw): 27993 JITStubs::cti_op_throw now always changes its return address, 27994 remove return code generated after the stub call (this is now 27995 handled by ctiOpThrowNotCaught). 27996 * jit/JITStubs.cpp: 27997 (JSC::): 27998 Add ctiOpThrowNotCaught definitions. 27999 (JSC::JITStubs::DEFINE_STUB_FUNCTION): 28000 Change cti_op_throw to always change its return address. 28001 * jit/JITStubs.h: 28002 Add ctiOpThrowNotCaught declaration. 28003 280042009-06-18 Kevin McCullough <kmccullough@apple.com> 28005 28006 Reviewed by Oliver Hunt. 28007 28008 <rdar://problem/6940880> REGRESSION: Breakpoints don't break in 64-bit 28009 28010 - Exposed functions now needed by WebCore. 28011 28012 * JavaScriptCore.exp: 28013 280142009-06-17 Darin Adler <darin@apple.com> 28015 28016 Reviewed by Oliver Hunt. 28017 28018 Bug 26429: Make JSON.stringify non-recursive so it can handle objects 28019 of arbitrary complexity 28020 https://bugs.webkit.org/show_bug.cgi?id=26429 28021 28022 For marking I decided not to use gcProtect, because this is inside the engine 28023 so it's easy enough to just do marking. And that darned gcProtect does locking! 28024 Oliver tried to convince me to used MarkedArgumentBuffer, but the constructor 28025 for that class says "FIXME: Remove all clients of this API, then remove this API." 28026 28027 * runtime/Collector.cpp: 28028 (JSC::Heap::collect): Add a call to JSONObject::markStringifiers. 28029 28030 * runtime/CommonIdentifiers.cpp: 28031 (JSC::CommonIdentifiers::CommonIdentifiers): Added emptyIdentifier. 28032 * runtime/CommonIdentifiers.h: Ditto. 28033 28034 * runtime/JSGlobalData.cpp: 28035 (JSC::JSGlobalData::JSGlobalData): Initialize firstStringifierToMark to 0. 28036 * runtime/JSGlobalData.h: Added firstStringifierToMark. 28037 28038 * runtime/JSONObject.cpp: Cut down the includes to the needed ones only. 28039 (JSC::unwrapNumberOrString): Added. Helper for unwrapping number and string 28040 objects to get their number and string values. 28041 (JSC::ReplacerPropertyName::ReplacerPropertyName): Added. The class is used 28042 to wrap an identifier or integer so we don't have to do any work unless we 28043 actually call a replacer. 28044 (JSC::ReplacerPropertyName::value): Added. 28045 (JSC::gap): Added. Helper function for the Stringifier constructor. 28046 (JSC::PropertyNameForFunctionCall::PropertyNameForFunctionCall): Added. 28047 The class is used to wrap an identifier or integer so we don't have to 28048 allocate a number or string until we actually call toJSON or a replacer. 28049 (JSC::PropertyNameForFunctionCall::asJSValue): Added. 28050 (JSC::Stringifier::Stringifier): Updated and moved out of the class 28051 definition. Added code to hook this into a singly linked list for marking. 28052 (JSC::Stringifier::~Stringifier): Remove from the singly linked list. 28053 (JSC::Stringifier::mark): Mark all the objects in the holder stacks. 28054 (JSC::Stringifier::stringify): Updated. 28055 (JSC::Stringifier::appendQuotedString): Tweaked and streamlined a bit. 28056 (JSC::Stringifier::toJSON): Renamed from toJSONValue. 28057 (JSC::Stringifier::appendStringifiedValue): Renamed from stringify. 28058 Added code to use the m_holderStack to do non-recursive stringify of 28059 objects and arrays. This code also uses the timeout checker since in 28060 pathological cases it could be slow even without calling into the 28061 JavaScript virtual machine. 28062 (JSC::Stringifier::willIndent): Added. 28063 (JSC::Stringifier::indent): Added. 28064 (JSC::Stringifier::unindent): Added. 28065 (JSC::Stringifier::startNewLine): Added. 28066 (JSC::Stringifier::Holder::Holder): Added. 28067 (JSC::Stringifier::Holder::appendNextProperty): Added. This is the 28068 function that handles the format of arrays and objects. 28069 (JSC::JSONObject::getOwnPropertySlot): Moved this down to the bottom 28070 of the file so the JSONObject class is not interleaved with the 28071 Stringifier class. 28072 (JSC::JSONObject::markStringifiers): Added. Calls mark. 28073 (JSC::JSONProtoFuncStringify): Streamlined the code here. The code 28074 to compute the gap string is now a separate function. 28075 28076 * runtime/JSONObject.h: Made everything private. Added markStringifiers. 28077 280782009-06-17 Oliver Hunt <oliver@apple.com> 28079 28080 Reviewed by Gavin Barraclough. 28081 28082 <rdar://problem/6974140> REGRESSION(r43849): Crash in cti_op_call_NotJSFunction when getting directions on maps.google.com 28083 28084 Roll out r43849 as it appears that we cannot rely on the address of 28085 an objects property storage being constant even if the structure is 28086 unchanged. 28087 28088 * jit/JIT.h: 28089 * jit/JITPropertyAccess.cpp: 28090 (JSC::JIT::compileGetDirectOffset): 28091 (JSC::JIT::privateCompileGetByIdProto): 28092 (JSC::JIT::privateCompileGetByIdProtoList): 28093 (JSC::JIT::privateCompileGetByIdChainList): 28094 (JSC::JIT::privateCompileGetByIdChain): 28095 280962009-06-17 Gavin Barraclough <barraclough@apple.com> 28097 28098 Rubber Stamped by Mark Rowe. 28099 28100 Fully revert r44492 & r44748 while we fix a bug they cause on internal builds <rdar://problem/6955963>. 28101 28102 * jit/JITOpcodes.cpp: 28103 (JSC::JIT::emit_op_throw): 28104 * jit/JITStubs.cpp: 28105 (JSC::): 28106 (JSC::JITStubs::DEFINE_STUB_FUNCTION): 28107 * jit/JITStubs.h: 28108 281092009-06-17 Gavin Barraclough <barraclough@apple.com> 28110 28111 Reviewed by Mark Rowe. 28112 28113 <rdar://problem/6947426> sunspider math-cordic.js exhibits different intermediate results running 32-bit vs. 64-bit 28114 28115 On 64-bit, NaN-encoded values must be detagged before they can be used in rshift. 28116 28117 No performance impact. 28118 28119 * jit/JITArithmetic.cpp: 28120 (JSC::JIT::emit_op_rshift): 28121 281222009-06-17 Adam Treat <adam.treat@torchmobile.com> 28123 28124 Reviewed by George Staikos. 28125 28126 https://bugs.webkit.org/show_bug.cgi?id=23155 28127 Move WIN_CE -> WINCE as previously discussed with Qt WINCE folks. 28128 28129 * jsc.cpp: 28130 (main): 28131 281322009-06-17 George Staikos <george.staikos@torchmobile.com> 28133 28134 Reviewed by Adam Treat. 28135 28136 https://bugs.webkit.org/show_bug.cgi?id=23155 28137 Move WIN_CE -> WINCE as previously discussed with Qt WINCE folks. 28138 28139 * config.h: 28140 * jsc.cpp: 28141 * wtf/Assertions.cpp: 28142 * wtf/Assertions.h: 28143 * wtf/CurrentTime.cpp: 28144 (WTF::lowResUTCTime): 28145 * wtf/DateMath.cpp: 28146 (WTF::getLocalTime): 28147 * wtf/MathExtras.h: 28148 * wtf/Platform.h: 28149 * wtf/StringExtras.h: 28150 * wtf/Threading.h: 28151 * wtf/win/MainThreadWin.cpp: 28152 281532009-06-17 Gavin Barraclough <barraclough@apple.com> 28154 28155 Reviewed by Oliver Hunt. 28156 28157 <rdar://problem/6974175> ASSERT in JITStubs.cpp at appsaccess.apple.com 28158 28159 Remove PropertySlot::putValue - PropertySlots should only be used for getting, 28160 not putting. Rename JSGlobalObject::getOwnPropertySlot to hasOwnPropertyForWrite, 28161 which is what it really was being used to ask, and remove some other getOwnPropertySlot 28162 & getOwnPropertySlotForWrite methods, which were unused and likely to lead to confusion. 28163 28164 * runtime/JSGlobalObject.h: 28165 (JSC::JSGlobalObject::hasOwnPropertyForWrite): 28166 * runtime/JSObject.h: 28167 * runtime/JSStaticScopeObject.cpp: 28168 * runtime/JSStaticScopeObject.h: 28169 * runtime/PropertySlot.h: 28170 281712009-06-16 Gavin Barraclough <barraclough@apple.com> 28172 28173 Reviewed by Oliver hunt. 28174 28175 Temporarily partially disable r44492, since this is causing some problems on internal builds. 28176 28177 * jit/JITOpcodes.cpp: 28178 (JSC::JIT::emit_op_throw): 28179 * jit/JITStubs.cpp: 28180 (JSC::JITStubs::DEFINE_STUB_FUNCTION): 28181 281822009-06-16 Sam Weinig <sam@webkit.org> 28183 28184 Fix windows build. 28185 28186 * jit/JIT.cpp: 28187 (JSC::JIT::JIT): 28188 281892009-06-16 Sam Weinig <sam@webkit.org> 28190 28191 Reviewed by Oliver Hunt. 28192 28193 Initialize m_bytecodeIndex to -1 in JIT, and correctly initialize 28194 it for each type of stub using the return address to find the correct 28195 offset. 28196 28197 * jit/JIT.cpp: 28198 (JSC::JIT::JIT): 28199 * jit/JIT.h: 28200 (JSC::JIT::compileGetByIdProto): 28201 (JSC::JIT::compileGetByIdSelfList): 28202 (JSC::JIT::compileGetByIdProtoList): 28203 (JSC::JIT::compileGetByIdChainList): 28204 (JSC::JIT::compileGetByIdChain): 28205 (JSC::JIT::compilePutByIdTransition): 28206 (JSC::JIT::compileCTIMachineTrampolines): 28207 (JSC::JIT::compilePatchGetArrayLength): 28208 * jit/JITStubCall.h: 28209 (JSC::JITStubCall::call): 28210 28211== Rolled over to ChangeLog-2009-06-16 == 28212