12009-08-10 Oliver Hunt <oliver@apple.com> 2 3 Reviewed by NOBODY (And another build fix). 4 5 Add new exports for MSVC 6 7 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 8 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 9 * JavaScriptCore.xcodeproj/project.pbxproj: 10 112009-08-10 Oliver Hunt <oliver@apple.com> 12 13 Reviewed by NOBODY (yet another build fix). 14 15 Remove obsolete entries from MSVC exports file 16 17 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 18 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 19 202009-08-10 Oliver Hunt <oliver@apple.com> 21 22 Reviewed by NOBODY (Build fix). 23 24 Add includes needed for non-allinonefile builds 25 26 * runtime/GetterSetter.h: 27 * runtime/ScopeChain.h: 28 292009-08-10 Oliver Hunt <oliver@apple.com> 30 31 Reviewed by NOBODY (Build fix). 32 33 Fix export file for last build fix 34 35 * JavaScriptCore.exp: 36 372009-08-10 Oliver Hunt <oliver@apple.com> 38 39 Reviewed by NOBODY (Build fix). 40 41 Hoist page size initialization into platform specific code. 42 43 * jit/ExecutableAllocatorPosix.cpp: 44 * jit/ExecutableAllocatorWin.cpp: 45 * runtime/MarkStack.h: 46 (JSC::MarkStack::pageSize): 47 * runtime/MarkStackPosix.cpp: 48 (JSC::MarkStack::initializePagesize): 49 * runtime/MarkStackWin.cpp: 50 (JSC::MarkStack::initializePagesize): 51 522009-08-07 Oliver Hunt <oliver@apple.com> 53 54 Reviewed by Gavin Barraclough. 55 56 Stack overflow crash in JavaScript garbage collector mark pass 57 https://bugs.webkit.org/show_bug.cgi?id=12216 58 59 Make the GC mark phase iterative by using an explicit mark stack. 60 To do this marking any single object is performed in multiple stages 61 * The object is appended to the MarkStack, this sets the marked 62 bit for the object using the new markDirect() function, and then 63 returns 64 * When the MarkStack is drain()ed the object is popped off the stack 65 and markChildren(MarkStack&) is called on the object to collect 66 all of its children. drain() then repeats until the stack is empty. 67 68 Additionally I renamed a number of methods from 'mark' to 'markAggregate' 69 in order to make it more clear that marking of those object was not 70 going to result in an actual recursive mark. 71 72 * GNUmakefile.am 73 * JavaScriptCore.exp: 74 * JavaScriptCore.gypi: 75 * JavaScriptCore.pri: 76 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: 77 * JavaScriptCore.xcodeproj/project.pbxproj: 78 * bytecode/CodeBlock.cpp: 79 (JSC::CodeBlock::markAggregate): 80 * bytecode/CodeBlock.h: 81 * bytecode/EvalCodeCache.h: 82 (JSC::EvalCodeCache::markAggregate): 83 * debugger/DebuggerActivation.cpp: 84 (JSC::DebuggerActivation::markChildren): 85 * debugger/DebuggerActivation.h: 86 * interpreter/Register.h: 87 * interpreter/RegisterFile.h: 88 (JSC::RegisterFile::markGlobals): 89 (JSC::RegisterFile::markCallFrames): 90 * parser/Nodes.cpp: 91 (JSC::ScopeNodeData::markAggregate): 92 (JSC::EvalNode::markAggregate): 93 (JSC::FunctionBodyNode::markAggregate): 94 * parser/Nodes.h: 95 (JSC::ScopeNode::markAggregate): 96 * runtime/ArgList.cpp: 97 (JSC::MarkedArgumentBuffer::markLists): 98 * runtime/ArgList.h: 99 * runtime/Arguments.cpp: 100 (JSC::Arguments::markChildren): 101 * runtime/Arguments.h: 102 * runtime/Collector.cpp: 103 (JSC::Heap::markConservatively): 104 (JSC::Heap::markCurrentThreadConservativelyInternal): 105 (JSC::Heap::markCurrentThreadConservatively): 106 (JSC::Heap::markOtherThreadConservatively): 107 (JSC::Heap::markStackObjectsConservatively): 108 (JSC::Heap::markProtectedObjects): 109 (JSC::Heap::collect): 110 * runtime/Collector.h: 111 * runtime/GetterSetter.cpp: 112 (JSC::GetterSetter::markChildren): 113 * runtime/GetterSetter.h: 114 (JSC::GetterSetter::GetterSetter): 115 (JSC::GetterSetter::createStructure): 116 * runtime/GlobalEvalFunction.cpp: 117 (JSC::GlobalEvalFunction::markChildren): 118 * runtime/GlobalEvalFunction.h: 119 * runtime/JSActivation.cpp: 120 (JSC::JSActivation::markChildren): 121 * runtime/JSActivation.h: 122 * runtime/JSArray.cpp: 123 (JSC::JSArray::markChildren): 124 * runtime/JSArray.h: 125 * runtime/JSCell.h: 126 (JSC::JSCell::markCellDirect): 127 (JSC::JSCell::markChildren): 128 (JSC::JSValue::markDirect): 129 (JSC::JSValue::markChildren): 130 (JSC::JSValue::hasChildren): 131 (JSC::MarkStack::append): 132 (JSC::MarkStack::drain): 133 * runtime/JSFunction.cpp: 134 (JSC::JSFunction::markChildren): 135 * runtime/JSFunction.h: 136 * runtime/JSGlobalData.cpp: 137 (JSC::JSGlobalData::JSGlobalData): 138 * runtime/JSGlobalData.h: 139 * runtime/JSGlobalObject.cpp: 140 (JSC::markIfNeeded): 141 (JSC::JSGlobalObject::markChildren): 142 * runtime/JSGlobalObject.h: 143 * runtime/JSNotAnObject.cpp: 144 (JSC::JSNotAnObject::markChildren): 145 * runtime/JSNotAnObject.h: 146 * runtime/JSONObject.cpp: 147 (JSC::Stringifier::markAggregate): 148 (JSC::JSONObject::markStringifiers): 149 * runtime/JSONObject.h: 150 * runtime/JSObject.cpp: 151 (JSC::JSObject::markChildren): 152 (JSC::JSObject::defineGetter): 153 (JSC::JSObject::defineSetter): 154 * runtime/JSObject.h: 155 * runtime/JSPropertyNameIterator.cpp: 156 (JSC::JSPropertyNameIterator::markChildren): 157 * runtime/JSPropertyNameIterator.h: 158 (JSC::JSPropertyNameIterator::createStructure): 159 (JSC::JSPropertyNameIterator::JSPropertyNameIterator): 160 (JSC::JSPropertyNameIterator::create): 161 * runtime/JSStaticScopeObject.cpp: 162 (JSC::JSStaticScopeObject::markChildren): 163 * runtime/JSStaticScopeObject.h: 164 * runtime/JSType.h: 165 (JSC::): 166 * runtime/JSValue.h: 167 * runtime/JSWrapperObject.cpp: 168 (JSC::JSWrapperObject::markChildren): 169 * runtime/JSWrapperObject.h: 170 * runtime/MarkStack.cpp: Added. 171 (JSC::MarkStack::compact): 172 * runtime/MarkStack.h: Added. 173 (JSC::): 174 (JSC::MarkStack::MarkStack): 175 (JSC::MarkStack::append): 176 (JSC::MarkStack::appendValues): 177 (JSC::MarkStack::~MarkStack): 178 (JSC::MarkStack::MarkSet::MarkSet): 179 (JSC::MarkStack::pageSize): 180 181 MarkStackArray is a non-shrinking, mmap-based vector type 182 used for storing objects to be marked. 183 (JSC::MarkStack::MarkStackArray::MarkStackArray): 184 (JSC::MarkStack::MarkStackArray::~MarkStackArray): 185 (JSC::MarkStack::MarkStackArray::expand): 186 (JSC::MarkStack::MarkStackArray::append): 187 (JSC::MarkStack::MarkStackArray::removeLast): 188 (JSC::MarkStack::MarkStackArray::isEmpty): 189 (JSC::MarkStack::MarkStackArray::size): 190 (JSC::MarkStack::MarkStackArray::shrinkAllocation): 191 * runtime/MarkStackPosix.cpp: Added. 192 (JSC::MarkStack::allocateStack): 193 (JSC::MarkStack::releaseStack): 194 * runtime/MarkStackWin.cpp: Added. 195 (JSC::MarkStack::allocateStack): 196 (JSC::MarkStack::releaseStack): 197 198 * runtime/ScopeChain.h: 199 * runtime/ScopeChainMark.h: 200 (JSC::ScopeChain::markAggregate): 201 * runtime/SmallStrings.cpp: 202 (JSC::SmallStrings::mark): 203 * runtime/Structure.h: 204 (JSC::Structure::markAggregate): 205 2062009-08-10 Mark Rowe <mrowe@apple.com> 207 208 Reviewed by Darin Adler. 209 210 Fix hundreds of "pointer being freed was not allocated" errors seen on the build bot. 211 212 * wtf/FastMalloc.h: Implement nothrow variants of the delete and delete[] operators since 213 we implement the nothrow variants of new and new[]. The nothrow variant of delete is called 214 explicitly in the implementation of std::sort which was resulting in FastMalloc-allocated 215 memory being passed to the system allocator to free. 216 2172009-08-10 Jan Michael Alonzo <jmalonzo@webkit.org> 218 219 [Gtk] Unreviewed build fix. Move JSAPIValueWrapper.cpp/.h in the debug 220 section. This file is already part of AllInOneFile in Release builds. 221 222 * GNUmakefile.am: 223 2242009-08-10 Darin Adler <darin@apple.com> 225 226 * wtf/FastMalloc.h: Fix build. 227 2282009-08-10 Darin Adler <darin@apple.com> 229 230 Reviewed by Mark Rowe. 231 232 FastMalloc.h has cross-platform code but marked as WinCE-only 233 https://bugs.webkit.org/show_bug.cgi?id=28160 234 235 1) The support for nothrow was inside #if PLATFORM(WINCE) even though it is 236 not platform-specific. 237 2) The code tried to override operator delete nothrow, which does not exist. 238 3) The code in the header checks the value of USE_SYSTEM_MALLOC, but the code 239 in FastMalloc.cpp checks only if the macro is defined. 240 241 * wtf/FastMalloc.h: See above. 242 * wtf/FastMalloc.cpp: Ditto. 243 2442009-08-10 Sam Weinig <sam@webkit.org> 245 246 Reviewed by Anders Carlsson. 247 248 Fix an annoying indentation issue. 249 250 * runtime/DateConstructor.cpp: 251 (JSC::constructDate): 252 2532009-08-10 Xan Lopez <xlopez@igalia.com> 254 255 Unreviewed build fix. 256 257 Add new files to makefile. 258 259 * GNUmakefile.am: 260 2612009-08-10 Simon Hausmann <simon.hausmann@nokia.com> 262 263 Fix compilation with the interpreter instead of the JIT by including 264 PrototypeFunction.h as forward-declared through NativeFunctionWrapper.h. 265 266 * runtime/ObjectConstructor.cpp: 267 2682009-08-09 Oliver Hunt <oliver@apple.com> 269 270 Reviewed by George Staikos. 271 272 JSON.stringify replacer returning undefined does not omit object properties 273 https://bugs.webkit.org/show_bug.cgi?id=28118 274 275 Correct behaviour of stringify when using a replacer function that returns 276 undefined. This is a simple change to move the undefined value check to 277 after the replacer function is called. This means that the replacer function 278 is now called for properties with the value undefined, however i've confirmed 279 that this behaviour is correct. 280 281 In addition I've made the cyclic object exception have a more useful error 282 message. 283 284 * runtime/JSONObject.cpp: 285 (JSC::Stringifier::appendStringifiedValue): 286 2872009-08-08 Oliver Hunt <oliver@apple.com> 288 289 Reviewed by Eric Seidel and Sam Weinig. 290 291 [ES5] Implement Object.getPrototypeOf 292 https://bugs.webkit.org/show_bug.cgi?id=28114 293 294 Implement getPrototypeOf 295 296 * runtime/CommonIdentifiers.h: 297 * runtime/JSGlobalObject.cpp: 298 (JSC::JSGlobalObject::reset): 299 * runtime/ObjectConstructor.cpp: 300 (JSC::ObjectConstructor::ObjectConstructor): 301 (JSC::objectConsGetPrototypeOf): 302 * runtime/ObjectConstructor.h: 303 3042009-08-07 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 305 306 Reviewed by Eric Seidel. 307 308 Allow custom memory allocation control for Noncopyable class 309 https://bugs.webkit.org/show_bug.cgi?id=27879 310 311 Several classes which are inherited from Noncopyable are instantiated by 312 operator new, so Noncopyable class has been inherited from FastAllocBase. 313 314 * wtf/Noncopyable.h: 315 3162009-08-07 George Staikos <george.staikos@torchmobile.com> 317 318 Reviewed by Eric Seidel. 319 320 https://bugs.webkit.org/show_bug.cgi?id=27305 321 Implement WinCE-specific unicode layer. 322 Written by George Staikos <george.staikos@torchmobile.com> 323 with bug fixes by Yong Li <yong.li@torchmobile.com> 324 refactored by Joe Mason <joe.mason@torchmobile.com> 325 326 * wtf/Platform.h: 327 * wtf/unicode/Unicode.h: 328 * wtf/unicode/wince/UnicodeWince.cpp: Added. 329 (WTF::Unicode::toLower): 330 (WTF::Unicode::toUpper): 331 (WTF::Unicode::foldCase): 332 (WTF::Unicode::isPrintableChar): 333 (WTF::Unicode::isSpace): 334 (WTF::Unicode::isLetter): 335 (WTF::Unicode::isUpper): 336 (WTF::Unicode::isLower): 337 (WTF::Unicode::isDigit): 338 (WTF::Unicode::isPunct): 339 (WTF::Unicode::toTitleCase): 340 (WTF::Unicode::direction): 341 (WTF::Unicode::category): 342 (WTF::Unicode::decompositionType): 343 (WTF::Unicode::combiningClass): 344 (WTF::Unicode::mirroredChar): 345 (WTF::Unicode::digitValue): 346 * wtf/unicode/wince/UnicodeWince.h: Added. 347 (WTF::Unicode::): 348 (WTF::Unicode::isSeparatorSpace): 349 (WTF::Unicode::isHighSurrogate): 350 (WTF::Unicode::isLowSurrogate): 351 (WTF::Unicode::isArabicChar): 352 (WTF::Unicode::hasLineBreakingPropertyComplexContext): 353 (WTF::Unicode::umemcasecmp): 354 (WTF::Unicode::surrogateToUcs4): 355 3562009-08-07 Yongjun Zhang <yongjun.zhang@nokia.com> 357 358 Reviewed by Eric Seidel. 359 360 https://bugs.webkit.org/show_bug.cgi?id=28069 361 362 Add inline to help winscw compiler resolve specialized argument in 363 templated functions. 364 365 * runtime/LiteralParser.cpp: 366 (JSC::LiteralParser::Lexer::lexString): 367 3682009-08-07 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 369 370 Reviewed by Eric Seidel. 371 372 Allow custom memory allocation control for RegExpObjectData struct 373 http://bugs.webkit.org/show_bug.cgi?id=26750 374 375 Inherits RegExpObjectData struct from FastAllocBase because 376 it has been instantiated by 'new' in JavaScriptCore/runtime/RegExpObject.cpp:62 377 378 * runtime/RegExpObject.h: 379 3802009-08-06 Norbert Leser <norbert.leser@nokia.com> 381 382 Reviewed by Darin Adler. 383 384 Updated patch for bug #27059: 385 Symbian platform always uses little endian encoding, 386 regardless of compiler. 387 We need to make sure that we correctly detect EABI architecture 388 for armv5 targets on Symbian, 389 where __EABI__ is set but not __ARM_EABI__ 390 391 * wtf/Platform.h: 392 3932009-08-06 Adam Barth <abarth@webkit.org> 394 395 Unreviewed revert. 396 397 http://bugs.webkit.org/show_bug.cgi?id=27879 398 399 Revert 46877 because it broke GTK. 400 401 * wtf/Noncopyable.h: 402 4032009-08-06 Gavin Barraclough <barraclough@apple.com> 404 405 Reviewed by Oliver Hunt. 406 407 Make get_by_id/put_by_id/method_check/call defer optimization using a data flag rather than a code modification. 408 ( https://bugs.webkit.org/show_bug.cgi?id=27635 ) 409 410 This improves performance of ENABLE(ASSEMBLER_WX_EXCLUSIVE) builds by 2-2.5%, reducing the overhead to about 2.5%. 411 (No performance impact with ASSEMBLER_WX_EXCLUSIVE disabled). 412 413 * bytecode/CodeBlock.cpp: 414 (JSC::printStructureStubInfo): 415 - Make StructureStubInfo store the type as an integer, rather than an OpcodeID. 416 417 * bytecode/CodeBlock.h: 418 (JSC::): 419 (JSC::CallLinkInfo::seenOnce): 420 (JSC::CallLinkInfo::setSeen): 421 (JSC::MethodCallLinkInfo::seenOnce): 422 (JSC::MethodCallLinkInfo::setSeen): 423 - Change a pointer in CallLinkInfo/MethodCallLinkInfo to use a PtrAndFlags, use a flag to track when an op has been executed once. 424 425 * bytecode/StructureStubInfo.cpp: 426 (JSC::StructureStubInfo::deref): 427 - Make StructureStubInfo store the type as an integer, rather than an OpcodeID. 428 429 * bytecode/StructureStubInfo.h: 430 (JSC::StructureStubInfo::StructureStubInfo): 431 (JSC::StructureStubInfo::initGetByIdSelf): 432 (JSC::StructureStubInfo::initGetByIdProto): 433 (JSC::StructureStubInfo::initGetByIdChain): 434 (JSC::StructureStubInfo::initGetByIdSelfList): 435 (JSC::StructureStubInfo::initGetByIdProtoList): 436 (JSC::StructureStubInfo::initPutByIdTransition): 437 (JSC::StructureStubInfo::initPutByIdReplace): 438 (JSC::StructureStubInfo::seenOnce): 439 (JSC::StructureStubInfo::setSeen): 440 - Make StructureStubInfo store the type as an integer, rather than an OpcodeID, add a flag to track when an op has been executed once. 441 442 * bytecompiler/BytecodeGenerator.cpp: 443 (JSC::BytecodeGenerator::emitGetById): 444 (JSC::BytecodeGenerator::emitPutById): 445 - Make StructureStubInfo store the type as an integer, rather than an OpcodeID. 446 447 * jit/JIT.cpp: 448 (JSC::JIT::privateCompileCTIMachineTrampolines): 449 (JSC::JIT::unlinkCall): 450 - Remove the "don't lazy link" stage of calls. 451 452 * jit/JIT.h: 453 (JSC::JIT::compileCTIMachineTrampolines): 454 - Remove the "don't lazy link" stage of calls. 455 456 * jit/JITCall.cpp: 457 (JSC::JIT::compileOpCallSlowCase): 458 - Remove the "don't lazy link" stage of calls. 459 460 * jit/JITStubs.cpp: 461 (JSC::JITThunks::JITThunks): 462 (JSC::JITThunks::tryCachePutByID): 463 (JSC::JITThunks::tryCacheGetByID): 464 (JSC::JITStubs::DEFINE_STUB_FUNCTION): 465 (JSC::JITStubs::getPolymorphicAccessStructureListSlot): 466 - Remove the "don't lazy link" stage of calls, and the "_second" stage of get_by_id/put_by_id/method_check. 467 468 * jit/JITStubs.h: 469 (JSC::JITThunks::ctiStringLengthTrampoline): 470 (JSC::JITStubs::): 471 - Remove the "don't lazy link" stage of calls, and the "_second" stage of get_by_id/put_by_id/method_check. 472 473 * wtf/PtrAndFlags.h: 474 (WTF::PtrAndFlags::PtrAndFlags): 475 (WTF::PtrAndFlags::operator!): 476 (WTF::PtrAndFlags::operator->): 477 - Add ! and -> operators, add constuctor with pointer argument. 478 4792009-08-06 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 480 481 Reviewed by Adam Barth. 482 483 Allow custom memory allocation control for Noncopyable class 484 https://bugs.webkit.org/show_bug.cgi?id=27879 485 486 Several classes which inherited from Noncopyable are instantiated by 487 operator new, so Noncopyable class has been inherited from FastAllocBase. 488 489 * wtf/Noncopyable.h: 490 4912009-08-06 Mark Rowe <mrowe@apple.com> 492 493 Rubber-stamped by Sam Weinig. 494 495 Add explicit dependencies for our build verification scripts to ensure that they always run after linking has completed. 496 497 * JavaScriptCore.xcodeproj/project.pbxproj: 498 4992009-08-06 Mark Rowe <mrowe@apple.com> 500 501 Bring a little order to our otherwise out of control lives. 502 503 * JavaScriptCore.xcodeproj/project.pbxproj: 504 5052009-08-06 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 506 507 Reviewed by Darin Adler. 508 509 Allow custom memory allocation control for JavaScriptCore's PolymorphicAccessStructureList struct 510 https://bugs.webkit.org/show_bug.cgi?id=27877 511 512 Inherits PolymorphicAccessStructureList struct from FastAllocBase because it has been instantiated by 513 'new' in JavaScriptCore/jit/JITStubs.cpp:1229. 514 515 * bytecode/Instruction.h: 516 5172009-08-05 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 518 519 Reviewed by Darin Adler. 520 521 Allow custom memory allocation control for JavaScriptCore's ScopeNodeData struct 522 https://bugs.webkit.org/show_bug.cgi?id=27875 523 524 Inherits ScopeNodeData struct from FastAllocBase because it has been instantiated by 525 'new' in JavaScriptCore/parser/Nodes.cpp:1848. 526 527 * parser/Nodes.h: 528 5292009-08-05 Zoltan Herczeg <zherczeg@inf.u-szeged.hu> 530 531 Reviewed by Gavin Barraclough. 532 533 Add floating point support for generic ARM port. 534 https://bugs.webkit.org/show_bug.cgi?id=24986 535 536 * assembler/ARMAssembler.cpp: 537 (JSC::ARMAssembler::doubleTransfer): 538 * assembler/ARMAssembler.h: 539 (JSC::ARM::): 540 (JSC::ARMAssembler::): 541 (JSC::ARMAssembler::faddd_r): 542 (JSC::ARMAssembler::fsubd_r): 543 (JSC::ARMAssembler::fmuld_r): 544 (JSC::ARMAssembler::fcmpd_r): 545 (JSC::ARMAssembler::fdtr_u): 546 (JSC::ARMAssembler::fdtr_d): 547 (JSC::ARMAssembler::fmsr_r): 548 (JSC::ARMAssembler::fsitod_r): 549 (JSC::ARMAssembler::fmstat): 550 * assembler/MacroAssemblerARM.h: 551 (JSC::MacroAssemblerARM::): 552 (JSC::MacroAssemblerARM::supportsFloatingPoint): 553 (JSC::MacroAssemblerARM::loadDouble): 554 (JSC::MacroAssemblerARM::storeDouble): 555 (JSC::MacroAssemblerARM::addDouble): 556 (JSC::MacroAssemblerARM::subDouble): 557 (JSC::MacroAssemblerARM::mulDouble): 558 (JSC::MacroAssemblerARM::convertInt32ToDouble): 559 (JSC::MacroAssemblerARM::branchDouble): 560 * jit/JIT.h: 561 5622009-08-05 Zoltan Herczeg <zherczeg@inf.u-szeged.hu> 563 564 Reviewed by Gavin Barraclough. 565 566 Add JIT support for generic ARM port without optimizations. 567 https://bugs.webkit.org/show_bug.cgi?id=24986 568 569 All JIT optimizations are disabled. 570 571 Signed off by Zoltan Herczeg <zherczeg@inf.u-szeged.hu> 572 Signed off by Gabor Loki <loki@inf.u-szeged.hu> 573 574 * assembler/ARMAssembler.cpp: 575 (JSC::ARMAssembler::baseIndexTransfer32): 576 * assembler/AbstractMacroAssembler.h: 577 (JSC::AbstractMacroAssembler::Imm32::Imm32): 578 * assembler/MacroAssemblerARM.h: 579 (JSC::MacroAssemblerARM::store32): 580 (JSC::MacroAssemblerARM::move): 581 (JSC::MacroAssemblerARM::branch32): 582 (JSC::MacroAssemblerARM::add32): 583 (JSC::MacroAssemblerARM::sub32): 584 (JSC::MacroAssemblerARM::load32): 585 * bytecode/CodeBlock.h: 586 (JSC::CodeBlock::getBytecodeIndex): 587 * jit/JIT.h: 588 * jit/JITInlineMethods.h: 589 (JSC::JIT::restoreArgumentReference): 590 * jit/JITOpcodes.cpp: 591 * jit/JITStubs.cpp: 592 * jit/JITStubs.h: 593 (JSC::JITStackFrame::returnAddressSlot): 594 * wtf/Platform.h: 595 5962009-08-04 Gavin Barraclough <barraclough@apple.com> 597 598 Rubber Stamped by Oiver Hunt. 599 600 Revert r46643 since this breaks the Yarr::Interpreter running the v8 tests. 601 https://bugs.webkit.org/show_bug.cgi?id=27874 602 603 * yarr/RegexInterpreter.cpp: 604 (JSC::Yarr::Interpreter::allocDisjunctionContext): 605 (JSC::Yarr::Interpreter::freeDisjunctionContext): 606 (JSC::Yarr::Interpreter::allocParenthesesDisjunctionContext): 607 (JSC::Yarr::Interpreter::freeParenthesesDisjunctionContext): 608 6092009-08-04 Oliver Hunt <oliver@apple.com> 610 611 Reviewed by NOBODY (build fix). 612 613 PPC64 Build fix 614 615 * wtf/Platform.h: 616 6172009-08-04 Benjamin C Meyer <benjamin.meyer@torchmobile.com> 618 619 Reviewed by Adam Treat 620 621 Explicitly include limits.h header when using INT_MAX and INT_MIN 622 623 * interpreter/Interpreter.cpp 624 6252009-08-03 Harald Fernengel <harald.fernengel@nokia.com> 626 627 Reviewed by Darin Adler. 628 629 Fix compile error for ambigous call to abs() 630 https://bugs.webkit.org/show_bug.cgi?id=27873 631 632 Fix ambiguity in abs(long int) call by calling labs() instead 633 634 * wtf/DateMath.cpp: replace call to abs() with labs() 635 6362009-08-03 Laszlo Gombos <laszlo.1.gombos@nokia.com> 637 638 Reviewed by Eric Seidel. 639 640 [Qt] Consolidate common gcc flags to WebKit.pri 641 https://bugs.webkit.org/show_bug.cgi?id=27934 642 643 * JavaScriptCore.pro: 644 6452009-08-03 Ada Chan <adachan@apple.com> 646 647 Fixed the Tiger build. 648 649 * wtf/FastMalloc.cpp: 650 6512009-08-03 Ada Chan <adachan@apple.com> 652 653 Reviewed by Darin Adler. 654 655 Don't use background thread to scavenge memory on Tiger until we figure out why it causes a crash. 656 https://bugs.webkit.org/show_bug.cgi?id=27900 657 658 * wtf/FastMalloc.cpp: 659 6602009-08-03 Fumitoshi Ukai <ukai@chromium.org> 661 662 Reviewed by Jan Alonzo. 663 664 Fix build break on Gtk/x86_64. 665 https://bugs.webkit.org/show_bug.cgi?id=27936 666 667 Use JSVALUE64 for X86_64 LINUX, except Qt. 668 669 * wtf/Platform.h: 670 6712009-08-02 Xan Lopez <xlopez@igalia.com> 672 673 Fix the GTK+ build. 674 675 * wtf/Platform.h: 676 6772009-08-02 Geoffrey Garen <ggaren@apple.com> 678 679 Reviewed by Sam Weinig. 680 681 Disabled JSVALUE32_64 on Qt builds, since all layout tests mysteriously 682 crash with it enabled. 683 684 * wtf/Platform.h: 685 6862009-08-02 Geoffrey Garen <ggaren@apple.com> 687 688 Qt build fix. 689 690 Added JSAPIValueWrapper.cpp to the build. 691 692 * JavaScriptCore.pri: 693 6942009-08-02 Geoffrey Garen <ggaren@apple.com> 695 696 Windows build fix. 697 698 Exported symbols for JSAPIValueWrapper. 699 700 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 701 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 702 7032009-08-02 Geoffrey Garen <ggaren@apple.com> 704 705 GTK build fix. 706 707 * jit/JITStubs.cpp: #include <stdarg.h>, for a definition of va_start. 708 7092009-08-02 Geoffrey Garen <ggaren@apple.com> 710 711 Qt build fix. 712 713 * runtime/Collector.cpp: #include <limits.h>, for a definition of ULONG_MAX. 714 7152009-08-02 Geoffrey Garen <ggaren@apple.com> 716 717 Windows build fix: Nixed JSImmediate::prototype, JSImmediate::toObject, 718 and JSImmediate::toThisObject, and removed their exported symbols. 719 720 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 721 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 722 * runtime/JSImmediate.cpp: 723 * runtime/JSImmediate.h: 724 7252009-08-02 Geoffrey Garen <ggaren@apple.com> 726 727 Reviewed by Mark Rowe. 728 729 Enabled JSVALUE32_64 by default on all platforms other than x86_64 (which uses JSVALUE64). 730 731 * wtf/Platform.h: 732 7332009-08-02 Kevin Ollivier <kevino@theolliviers.com> 734 735 Reviewed by Jan Alonzo. 736 737 Script for building the JavaScriptCore library for wx. 738 https://bugs.webkit.org/show_bug.cgi?id=27619 739 740 * wscript: Added. 741 7422009-08-02 Yong Li <yong.li@torchmobile.com> 743 744 Reviewed by George Staikos. 745 746 DateMath depends on strftime and localtime, which need to be imported manually on WinCE 747 https://bugs.webkit.org/show_bug.cgi?id=26558 748 749 * wtf/DateMath.cpp: 750 7512009-08-01 David Kilzer <ddkilzer@apple.com> 752 753 wtf/Threading.h: added include of Platform.h 754 755 Reviewed by Mark Rowe. 756 757 * wtf/Threading.h: Added #include "Platform.h" since this header 758 uses PLATFORM() and other macros. 759 7602009-08-01 Mark Rowe <mrowe@apple.com> 761 762 Rubber-stamped by Oliver Hunt. 763 764 Roll out r46668 as it was misinformed. ScopeChain is only used with placement new. 765 766 * runtime/ScopeChain.h: 767 7682009-08-01 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 769 770 Allow custom memory allocation control for JavaScriptCore's HashMap class 771 http://bugs.webkit.org/show_bug.cgi?id=27871 772 773 Inherits HashMap class from FastAllocBase because it has been 774 instantiated by 'new' in JavaScriptCore/API/JSClassRef.cpp:148. 775 776 * wtf/RefPtrHashMap.h: 777 (WTF::): 778 7792009-08-01 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 780 781 Allow custom memory allocation control for JavaScriptCore's ScopeChain class 782 https://bugs.webkit.org/show_bug.cgi?id=27834 783 784 Inherits ScopeChain class from FastAllocBase because it has been 785 instantiated by 'new' in JavaScriptCore/runtime/JSFunction.h:109. 786 787 * runtime/ScopeChain.h: 788 7892009-08-01 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 790 791 Reviewed by Darin Adler. 792 793 Allow custom memory allocation control for JavaScriptCore's RegExpConstructorPrivate struct 794 https://bugs.webkit.org/show_bug.cgi?id=27833 795 796 Inherits RegExpConstructorPrivate class from FastAllocBase because it has been 797 instantiated by 'new' in JavaScriptCore/runtime/RegExpConstructor.cpp:152. 798 799 * runtime/RegExpConstructor.cpp: 800 8012009-07-31 Yong Li <yong.li@torchmobile.com> 802 803 Reviewed by George Staikos. 804 805 Resurrect the old GetTickCount implementation of currentTime, controlled by WTF_USE_QUERY_PERFORMANCE_COUNTER 806 currentSystemTime taken from older WebKit; currentTime written by Yong Li <yong.li@torchmobile.com>; cleanup by Joe Mason <joe.mason@torchmobile.com> 807 https://bugs.webkit.org/show_bug.cgi?id=27848 808 809 * wtf/CurrentTime.cpp: 810 (WTF::currentSystemTime): get current time with GetCurrentFT 811 (WTF::currentTime): track msec elapsed since first currentSystemTime call using GetTickCount 812 * wtf/Platform.h: 813 8142009-07-31 Ada Chan <adachan@apple.com> 815 816 Fixes the Windows release-PGO build. 817 818 Reviewed by Jon Honeycutt. 819 820 * JavaScriptCore.vcproj/WTF/WTF.vcproj: Suppresses the warning about unreachable code that we get by adding "return 0" to WTF::TCMalloc_PageHeap::runScavengerThread(). 821 * wtf/FastMalloc.cpp: 822 (WTF::TCMalloc_PageHeap::runScavengerThread): Fixes the error about the method not returning a value in the release-PGO build. 823 8242009-07-31 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 825 826 Change malloc to fastMalloc and free to fastFree in Yarr's RegexInterpreter.cpp 827 https://bugs.webkit.org/show_bug.cgi?id=27874 828 829 Use fastMalloc and fastFree instead of malloc and free in RegexInterpreter.cpp's methods. 830 831 * yarr/RegexInterpreter.cpp: 832 (JSC::Yarr::Interpreter::allocDisjunctionContext): 833 (JSC::Yarr::Interpreter::freeDisjunctionContext): 834 (JSC::Yarr::Interpreter::allocParenthesesDisjunctionContext): 835 (JSC::Yarr::Interpreter::freeParenthesesDisjunctionContext): 836 8372009-07-30 Xan Lopez <xlopez@igalia.com> 838 839 Reviewed by Jan Alonzo. 840 841 Fix compiler warning. 842 843 GCC does not like C++-style comments in preprocessor directives. 844 845 * wtf/Platform.h: 846 8472009-07-30 John McCall <rjmccall@apple.com> 848 849 Reviewed by Gavin Barraclough. 850 851 Optimize the X86_64 trampolines: avoid the need for filler arguments 852 and move the stub-args area closer to the stack pointer. 853 854 * jit/JIT.h: adjust patch offsets because of slight code-size change 855 * jit/JITCode.h: 856 (JSC::JITCode::execute): don't pass filler args 857 * jit/JITStubs.cpp: 858 (ctiTrampoline): (X86_64): push args onto stack, use args directly 859 (ctiVMThrowTrampoline): (X86_64): adjust %rsp by correct displacement 860 (ctiOpThrowNotCaught): (X86_64): adjust %rsp by correct displacement 861 * jit/JITStubs.h: 862 (JITStackFrame): (X86_64): move args area earlier 863 (ctiTrampoline): remove filler args from prototype 864 8652009-07-30 Gavin Barraclough <barraclough@apple.com> 866 867 Reviewed by NOBODY (build fix). 868 869 Temporarily revert r46618 since this is b0rking on Linux. 870 8712009-07-23 Gavin Barraclough <barraclough@apple.com> 872 873 Reviewed by Oliver Hunt. 874 875 Make get_by_id/put_by_id/method_check/call defer optimization using a data flag rather than a code modification. 876 ( https://bugs.webkit.org/show_bug.cgi?id=27635 ) 877 878 This improves performance of ENABLE(ASSEMBLER_WX_EXCLUSIVE) builds by 2-2.5%, reducing the overhead to about 2.5%. 879 (No performance impact with ASSEMBLER_WX_EXCLUSIVE disabled). 880 881 * bytecode/CodeBlock.cpp: 882 (JSC::printStructureStubInfo): 883 - Make StructureStubInfo store the type as an integer, rather than an OpcodeID. 884 885 * bytecode/CodeBlock.h: 886 (JSC::): 887 (JSC::CallLinkInfo::seenOnce): 888 (JSC::CallLinkInfo::setSeen): 889 (JSC::MethodCallLinkInfo::seenOnce): 890 (JSC::MethodCallLinkInfo::setSeen): 891 - Change a pointer in CallLinkInfo/MethodCallLinkInfo to use a PtrAndFlags, use a flag to track when an op has been executed once. 892 893 * bytecode/StructureStubInfo.cpp: 894 (JSC::StructureStubInfo::deref): 895 - Make StructureStubInfo store the type as an integer, rather than an OpcodeID. 896 897 * bytecode/StructureStubInfo.h: 898 (JSC::StructureStubInfo::StructureStubInfo): 899 (JSC::StructureStubInfo::initGetByIdSelf): 900 (JSC::StructureStubInfo::initGetByIdProto): 901 (JSC::StructureStubInfo::initGetByIdChain): 902 (JSC::StructureStubInfo::initGetByIdSelfList): 903 (JSC::StructureStubInfo::initGetByIdProtoList): 904 (JSC::StructureStubInfo::initPutByIdTransition): 905 (JSC::StructureStubInfo::initPutByIdReplace): 906 (JSC::StructureStubInfo::seenOnce): 907 (JSC::StructureStubInfo::setSeen): 908 - Make StructureStubInfo store the type as an integer, rather than an OpcodeID, add a flag to track when an op has been executed once. 909 910 * bytecompiler/BytecodeGenerator.cpp: 911 (JSC::BytecodeGenerator::emitGetById): 912 (JSC::BytecodeGenerator::emitPutById): 913 - Make StructureStubInfo store the type as an integer, rather than an OpcodeID. 914 915 * jit/JIT.cpp: 916 (JSC::JIT::privateCompileCTIMachineTrampolines): 917 (JSC::JIT::unlinkCall): 918 - Remove the "don't lazy link" stage of calls. 919 920 * jit/JIT.h: 921 (JSC::JIT::compileCTIMachineTrampolines): 922 - Remove the "don't lazy link" stage of calls. 923 924 * jit/JITCall.cpp: 925 (JSC::JIT::compileOpCallSlowCase): 926 - Remove the "don't lazy link" stage of calls. 927 928 * jit/JITStubs.cpp: 929 (JSC::JITThunks::JITThunks): 930 (JSC::JITThunks::tryCachePutByID): 931 (JSC::JITThunks::tryCacheGetByID): 932 (JSC::JITStubs::DEFINE_STUB_FUNCTION): 933 (JSC::JITStubs::getPolymorphicAccessStructureListSlot): 934 - Remove the "don't lazy link" stage of calls, and the "_second" stage of get_by_id/put_by_id/method_check. 935 936 * jit/JITStubs.h: 937 (JSC::JITThunks::ctiStringLengthTrampoline): 938 (JSC::JITStubs::): 939 - Remove the "don't lazy link" stage of calls, and the "_second" stage of get_by_id/put_by_id/method_check. 940 941 * wtf/PtrAndFlags.h: 942 (WTF::PtrAndFlags::PtrAndFlags): 943 (WTF::PtrAndFlags::operator!): 944 (WTF::PtrAndFlags::operator->): 945 - Add ! and -> operators, add constuctor with pointer argument. 946 9472009-07-30 Geoffrey Garen <ggaren@apple.com> 948 949 Reviewed by Gavin Barraclough. 950 951 Fixed failing tests seen on Windows buildbot. 952 953 * jit/JITStubs.cpp: 954 (JSC::DEFINE_STUB_FUNCTION): 955 * jit/JITStubs.h: 956 (JSC::): Use "int" instead of "bool" to guarantee a 32-bit result, 957 regardless of compiler. gcc on mac uses 32-bit values for bool, 958 but gcc on linux and MSVC on Windows use 8-bit values. 959 9602009-07-30 Geoffrey Garen <ggaren@apple.com> 961 962 Windows build fix: added missing symbols on Windows. 963 964 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 965 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 966 9672009-07-30 Geoffrey Garen <ggaren@apple.com> 968 969 Windows build fix: removed stale symbols on Windows. 970 971 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 972 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 973 974=== End merge of nitro-extreme branch 2009-07-30 === 975 9762009-07-20 Geoffrey Garen <ggaren@apple.com> 977 978 Fixed a post-review typo in r46066 that caused tons of test failures. 979 980 SunSpider reports no change. 981 982 * runtime/JSArray.cpp: 983 (JSC::JSArray::JSArray): Initialize the full vector capacity, to avoid 984 uninitialized members at the end. 985 9862009-07-20 Geoffrey Garen <ggaren@apple.com> 987 988 Windows WebKit build fix: Added some missing exports. 989 990 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 991 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 992 9932009-07-17 Geoffrey Garen <ggaren@apple.com> 994 995 Reviewed by Sam Weinig. 996 997 Get the branch working on windows. 998 https://bugs.webkit.org/show_bug.cgi?id=27391 999 1000 SunSpider says 0.3% faster. 1001 1002 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 1003 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: Updated 1004 MSVC export lists to fix linker errors. 1005 1006 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: Added / removed 1007 new / old project files. 1008 1009 * jit/JIT.cpp: 1010 (JSC::JIT::privateCompileCTIMachineTrampolines): Used #pragma pack to tell 1011 MSVC that these structures represent actual memory layout, and should not be 1012 automatically aligned. Changed the return value load to load a 64bit quantity 1013 into the canonical registers. 1014 1015 * jit/JIT.h: Moved OBJECT_OFFSETOF definition to StdLibExtras.h because 1016 it's needed by more than just the JIT, and it supplements a standard library 1017 macro (offsetof). 1018 1019 * jit/JITCall.cpp: 1020 (JSC::JIT::compileOpCallInitializeCallFrame): Fixed an incorrectly signed 1021 cast to resolve an MSVC warning. 1022 1023 * jit/JITStubs.h: Used #pragma pack to tell MSVC that these structures 1024 represent actual memory layout, and should not be automatically aligned. 1025 1026 * runtime/JSArray.cpp: 1027 (JSC::JSArray::JSArray): Replaced memset_pattern8 with a for loop, since 1028 memset_pattern8 is not portable. (I verified that this version of the loop 1029 gives the best performance / generated code in GCC.) 1030 1031 * runtime/JSObject.h: 1032 (JSC::JSObject::JSObject): Removed accidental usage of FIELD_OFFSET -- 1033 OBJECT_OFFSETOF is our new macro name. (FIELD_OFFSET conflicts with a 1034 definition in winnt.h.) 1035 1036 * runtime/JSValue.cpp: Added some headers needed by non-all-in-one builds. 1037 1038 * runtime/JSValue.h: 1039 (JSC::JSValue::): Made the tag signed, to match MSVC's signed enum values. 1040 (GCC doesn't seem to care one way or the other.) 1041 1042 * wtf/MainThread.cpp: Moved the StdLibExtras.h #include -- I did this a 1043 while ago to resolve a conflict with winnt.h. I can't remember if it's truly 1044 still needed, but what the heck. 1045 1046 * wtf/StdLibExtras.h: Moved OBJECT_OFFSETOF definition here. 1047 10482009-07-06 Geoffrey Garen <ggaren@apple.com> 1049 1050 Reviewed by Sam Weinig (?). 1051 1052 Fixed an assertion seen during the stress test. 1053 1054 Don't assume that, if op1 is constant, op2 is not, and vice versa. Sadly, 1055 not all constants get folded. 1056 1057 * jit/JITArithmetic.cpp: 1058 (JSC::JIT::emit_op_jnless): 1059 (JSC::JIT::emitSlow_op_jnless): 1060 (JSC::JIT::emit_op_jnlesseq): 1061 (JSC::JIT::emitSlow_op_jnlesseq): 1062 10632009-07-06 Geoffrey Garen <ggaren@apple.com> 1064 1065 Reviewed by Sam Weinig. 1066 1067 Include op_convert_this in result caching. 1068 1069 No change on SunSpider or v8. 1070 1071 * jit/JITOpcodes.cpp: 1072 (JSC::JIT::emit_op_convert_this): 1073 1074 * jit/JITStubs.cpp: 1075 (JSC::DEFINE_STUB_FUNCTION): 1076 * jit/JITStubs.h: 1077 (JSC::): Made the op_convert_this JIT stub return an EncodedJSValue, so 1078 to maintain the result caching contract that { tag, payload } can be 1079 found in { regT1, regT0 }. 1080 10812009-07-06 Geoffrey Garen <ggaren@apple.com> 1082 1083 Reviewed by Sam Weinig. 1084 1085 Implemented result chaining. 1086 1087 1% faster on SunSpider. 4%-5% faster on v8. 1088 1089 * assembler/MacroAssemblerX86Common.h: 1090 (JSC::MacroAssemblerX86Common::move): 1091 * assembler/X86Assembler.h: 1092 (JSC::X86Assembler::movl_rr): Added an optimization to eliminate 1093 no-op mov instructions, to simplify chaining. 1094 1095 * jit/JIT.cpp: 1096 (JSC::JIT::JIT): 1097 * jit/JIT.h: Added data members and helper functions for recording 1098 chained results. We record both a mapping from virtual to machine register 1099 and the opcode for which the mapping is valid, to help ensure that the 1100 mapping isn't used after the mapped register has been stomped by other 1101 instructions. 1102 1103 * jit/JITCall.cpp: 1104 (JSC::JIT::compileOpCallVarargs): 1105 (JSC::JIT::compileOpCallVarargsSlowCase): 1106 (JSC::JIT::emit_op_ret): 1107 (JSC::JIT::emit_op_construct_verify): 1108 (JSC::JIT::compileOpCall): 1109 (JSC::JIT::compileOpCallSlowCase): Chain function call results. 1110 1111 * jit/JITInlineMethods.h: 1112 (JSC::JIT::emitLoadTag): 1113 (JSC::JIT::emitLoadPayload): 1114 (JSC::JIT::emitLoad): 1115 (JSC::JIT::emitLoad2): 1116 (JSC::JIT::isLabeled): 1117 (JSC::JIT::map): 1118 (JSC::JIT::unmap): 1119 (JSC::JIT::isMapped): 1120 (JSC::JIT::getMappedPayload): 1121 (JSC::JIT::getMappedTag): Use helper functions when loading virtual 1122 registers into machine registers, in case the loads can be eliminated 1123 by chaining. 1124 1125 * jit/JITOpcodes.cpp: 1126 (JSC::JIT::emit_op_mov): 1127 (JSC::JIT::emit_op_end): 1128 (JSC::JIT::emit_op_instanceof): 1129 (JSC::JIT::emit_op_get_global_var): 1130 (JSC::JIT::emit_op_put_global_var): 1131 (JSC::JIT::emit_op_get_scoped_var): 1132 (JSC::JIT::emit_op_put_scoped_var): 1133 (JSC::JIT::emit_op_to_primitive): 1134 (JSC::JIT::emit_op_resolve_global): 1135 (JSC::JIT::emit_op_jneq_ptr): 1136 (JSC::JIT::emit_op_next_pname): 1137 (JSC::JIT::emit_op_to_jsnumber): 1138 (JSC::JIT::emit_op_catch): Chain results from these opcodes. 1139 1140 (JSC::JIT::emit_op_profile_will_call): 1141 (JSC::JIT::emit_op_profile_did_call): Load the profiler into regT2 to 1142 avoid stomping a chained result. 1143 1144 * jit/JITPropertyAccess.cpp: 1145 (JSC::JIT::emit_op_method_check): 1146 (JSC::JIT::emit_op_get_by_val): 1147 (JSC::JIT::emit_op_get_by_id): Chain results from these opcodes. 1148 1149 * jit/JITStubCall.h: 1150 (JSC::JITStubCall::addArgument): Always use { regT1, regT0 }, to facilitate 1151 chaining. 1152 1153 (JSC::JITStubCall::call): Unmap all mapped registers, since our callee 1154 stub might stomp them. 1155 11562009-07-01 Sam Weinig <sam@webkit.org> 1157 1158 Reviewed by Gavin Barraclough. 1159 1160 Don't reload values in emitBinaryDoubleOp. 1161 1162 SunSpider reports a 0.6% progression. 1163 1164 * jit/JIT.h: 1165 * jit/JITArithmetic.cpp: 1166 (JSC::JIT::emit_op_jnless): 1167 (JSC::JIT::emit_op_jnlesseq): 1168 (JSC::JIT::emitBinaryDoubleOp): 1169 11702009-07-01 Sam Weinig <sam@webkit.org> 1171 1172 Reviewed by Geoffrey Garen. 1173 1174 Convert op_div to load op1 and op2 up front. 1175 1176 * jit/JITArithmetic.cpp: 1177 (JSC::JIT::emit_op_div): 1178 11792009-07-01 Sam Weinig <sam@webkit.org> 1180 1181 Reviewed by Geoffrey Garen. 1182 1183 Don't emit code in emitBinaryDoubleOp if code is unreachable, observable 1184 via an empty (unlinked) jumplist passed in. This only effects op_jnless 1185 and op_jnlesseq at present. 1186 1187 * jit/JITArithmetic.cpp: 1188 (JSC::JIT::emitSlow_op_jnless): 1189 (JSC::JIT::emitSlow_op_jnlesseq): 1190 (JSC::JIT::emitBinaryDoubleOp): 1191 11922009-07-01 Geoffrey Garen <ggaren@apple.com> 1193 1194 Reviewed by Sam Weinig. 1195 1196 Converted op_mod to put { tag, payload } in { regT1, regT0 }, and 1197 tidied up its constant case. 1198 1199 SunSpider reports a 0.2% regression, but a micro-benchmark of op_mod 1200 shows a 12% speedup, and the SunSpider test that uses op_mod most should 1201 benefit a lot from result caching in the end, since it almost always 1202 performs (expression) % constant. 1203 1204 * jit/JITArithmetic.cpp: 1205 (JSC::JIT::emit_op_mod): 1206 (JSC::JIT::emitSlow_op_mod): 1207 12082009-06-30 Sam Weinig <sam@webkit.org> 1209 1210 Reviewed by Geoffrey Garen. 1211 1212 Converted some more arithmetic ops to put { tag, payload } in 1213 { regT1, regT0 }. 1214 1215 * jit/JITArithmetic.cpp: 1216 (JSC::JIT::emit_op_mul): 1217 (JSC::JIT::emitSlow_op_mul): 1218 12192009-06-30 Geoffrey Garen <ggaren@apple.com> 1220 1221 Reviewed by Sam Weinig. 1222 1223 Converted some more arithmetic ops to put { tag, payload } in 1224 { regT1, regT0 }, and added a case for subtract constant. 1225 1226 SunSpider says no change. v8 says 0.3% slower. 1227 1228 * jit/JIT.h: 1229 * jit/JITArithmetic.cpp: 1230 (JSC::JIT::emit_op_add): 1231 (JSC::JIT::emitAdd32Constant): 1232 (JSC::JIT::emitSlow_op_add): 1233 (JSC::JIT::emit_op_sub): 1234 (JSC::JIT::emitSub32Constant): 1235 (JSC::JIT::emitSlow_op_sub): 1236 12372009-06-30 Gavin Barraclough <barraclough@apple.com> 1238 1239 Reviewed by Sam Weinig. 1240 1241 Remove more uses of addressFor(), load double constants directly from 1242 the constantpool in the CodeBlock, rather than from the register file. 1243 1244 * jit/JITArithmetic.cpp: 1245 (JSC::JIT::emitAdd32Constant): 1246 (JSC::JIT::emitBinaryDoubleOp): 1247 12482009-06-30 Geoffrey Garen <ggaren@apple.com> 1249 1250 Reviewed by Sam Weinig. 1251 1252 Fixed a bug in postfix ops, where we would treat x = x++ and x = x-- 1253 as a no-op, even if x were not an int, and the ++/-- could have side-effects. 1254 1255 * jit/JITArithmetic.cpp: 1256 (JSC::JIT::emit_op_post_inc): 1257 (JSC::JIT::emitSlow_op_post_inc): 1258 (JSC::JIT::emit_op_post_dec): 1259 (JSC::JIT::emitSlow_op_post_dec): 1260 12612009-06-30 Geoffrey Garen <ggaren@apple.com> 1262 1263 Reviewed by Sam Weinig. 1264 1265 Converted some arithmetic ops to put { tag, payload } in 1266 { regT1, regT0 }. 1267 1268 SunSpider says 0.7% faster. v8 says no change. 1269 1270 * jit/JIT.h: 1271 * jit/JITArithmetic.cpp: 1272 (JSC::JIT::emit_op_jnless): 1273 (JSC::JIT::emit_op_jnlesseq): 1274 (JSC::JIT::emit_op_lshift): 1275 (JSC::JIT::emit_op_rshift): 1276 (JSC::JIT::emit_op_bitand): 1277 (JSC::JIT::emit_op_bitor): 1278 (JSC::JIT::emit_op_bitxor): 1279 * jit/JITInlineMethods.h: 1280 (JSC::JIT::isOperandConstantImmediateInt): 1281 (JSC::JIT::getOperandConstantImmediateInt): 1282 12832009-06-30 Gavin Barraclough <barraclough@apple.com> 1284 1285 Reviewed by Sam Weinig. 1286 1287 Start removing cases of addressFor(). 1288 1289 * jit/JIT.h: 1290 * jit/JITArithmetic.cpp: 1291 (JSC::JIT::emitAdd32Constant): 1292 (JSC::JIT::emitBinaryDoubleOp): 1293 (JSC::JIT::emit_op_div): 1294 * jit/JITInlineMethods.h: 1295 (JSC::JIT::emitLoadDouble): 1296 (JSC::JIT::emitLoadInt32ToDouble): 1297 (JSC::JIT::emitStoreDouble): 1298 * jit/JITOpcodes.cpp: 1299 (JSC::JIT::emit_op_jfalse): 1300 (JSC::JIT::emit_op_jtrue): 1301 13022009-06-30 Geoffrey Garen <ggaren@apple.com> 1303 1304 Rolled back in my last patch with regression fixed. 1305 1306 * jit/JIT.cpp: 1307 (JSC::JIT::privateCompileSlowCases): 1308 * jit/JIT.h: 1309 * jit/JITOpcodes.cpp: 1310 (JSC::JIT::emit_op_loop_if_less): 1311 (JSC::JIT::emit_op_loop_if_lesseq): 1312 (JSC::JIT::emit_op_resolve_global): 1313 (JSC::JIT::emitSlow_op_resolve_global): 1314 (JSC::JIT::emit_op_eq): 1315 (JSC::JIT::emitSlow_op_eq): 1316 (JSC::JIT::emit_op_neq): 1317 (JSC::JIT::emitSlow_op_neq): 1318 13192009-06-30 Geoffrey Garen <ggaren@apple.com> 1320 1321 Rolled out my last patch because it was a 2% SunSpider regression. 1322 1323 * jit/JIT.cpp: 1324 (JSC::JIT::privateCompileSlowCases): 1325 * jit/JIT.h: 1326 * jit/JITOpcodes.cpp: 1327 (JSC::JIT::emit_op_loop_if_less): 1328 (JSC::JIT::emit_op_loop_if_lesseq): 1329 (JSC::JIT::emit_op_resolve_global): 1330 (JSC::JIT::emit_op_eq): 1331 (JSC::JIT::emitSlow_op_eq): 1332 (JSC::JIT::emit_op_neq): 1333 (JSC::JIT::emitSlow_op_neq): 1334 13352009-06-30 Geoffrey Garen <ggaren@apple.com> 1336 1337 Reviewed by Gavin "Sam Weinig" Barraclough. 1338 1339 Standardized the rest of our opcodes to put { tag, payload } in 1340 { regT1, regT0 } where possible. 1341 1342 * jit/JIT.cpp: 1343 (JSC::JIT::privateCompileSlowCases): 1344 * jit/JIT.h: 1345 * jit/JITOpcodes.cpp: 1346 (JSC::JIT::emit_op_loop_if_less): 1347 (JSC::JIT::emit_op_loop_if_lesseq): 1348 (JSC::JIT::emit_op_resolve_global): 1349 (JSC::JIT::emitSlow_op_resolve_global): 1350 (JSC::JIT::emit_op_eq): 1351 (JSC::JIT::emitSlow_op_eq): 1352 (JSC::JIT::emit_op_neq): 1353 (JSC::JIT::emitSlow_op_neq): 1354 13552009-06-30 Gavin Barraclough <barraclough@apple.com> 1356 1357 Reviewed by Geoffrey Garen. 1358 1359 Replace calls to store32(tagFor()) and store32(payloadFor()) 1360 with emitStoreInt32(), emitStoreBool(), and emitStoreCell(). 1361 1362 * jit/JIT.h: 1363 * jit/JITArithmetic.cpp: 1364 (JSC::JIT::emit_op_negate): 1365 (JSC::JIT::emit_op_lshift): 1366 (JSC::JIT::emit_op_rshift): 1367 (JSC::JIT::emit_op_bitand): 1368 (JSC::JIT::emitBitAnd32Constant): 1369 (JSC::JIT::emit_op_bitor): 1370 (JSC::JIT::emitBitOr32Constant): 1371 (JSC::JIT::emit_op_bitxor): 1372 (JSC::JIT::emitBitXor32Constant): 1373 (JSC::JIT::emit_op_bitnot): 1374 (JSC::JIT::emit_op_post_inc): 1375 (JSC::JIT::emit_op_post_dec): 1376 (JSC::JIT::emit_op_pre_inc): 1377 (JSC::JIT::emit_op_pre_dec): 1378 (JSC::JIT::emit_op_add): 1379 (JSC::JIT::emitAdd32Constant): 1380 (JSC::JIT::emit_op_sub): 1381 (JSC::JIT::emitSub32ConstantLeft): 1382 (JSC::JIT::emitSub32ConstantRight): 1383 (JSC::JIT::emit_op_mul): 1384 (JSC::JIT::emitSlow_op_mul): 1385 (JSC::JIT::emit_op_div): 1386 (JSC::JIT::emit_op_mod): 1387 * jit/JITCall.cpp: 1388 (JSC::JIT::emit_op_load_varargs): 1389 * jit/JITInlineMethods.h: 1390 (JSC::JIT::emitStoreInt32): 1391 (JSC::JIT::emitStoreCell): 1392 (JSC::JIT::emitStoreBool): 1393 (JSC::JIT::emitStore): 1394 * jit/JITOpcodes.cpp: 1395 (JSC::JIT::emit_op_instanceof): 1396 (JSC::JIT::emit_op_not): 1397 (JSC::JIT::emit_op_eq): 1398 (JSC::JIT::emitSlow_op_eq): 1399 (JSC::JIT::emit_op_neq): 1400 (JSC::JIT::emitSlow_op_neq): 1401 (JSC::JIT::compileOpStrictEq): 1402 (JSC::JIT::emit_op_eq_null): 1403 (JSC::JIT::emit_op_neq_null): 1404 * jit/JITStubCall.h: 1405 (JSC::JITStubCall::call): 1406 14072009-06-30 Geoffrey Garen <ggaren@apple.com> 1408 1409 Reviewed by Sam Weinig. 1410 1411 Standardized the rest of the property access instructions to put { tag, 1412 payload } in { regT1, regT0 }. 1413 1414 Small v8 speedup, 0.2% SunSpider slowdown. 1415 1416 * jit/JIT.h: 1417 * jit/JITInlineMethods.h: 1418 (JSC::JIT::emitLoad): 1419 (JSC::JIT::emitLoad2): 1420 * jit/JITPropertyAccess.cpp: 1421 (JSC::JIT::emit_op_get_by_val): 1422 (JSC::JIT::emitSlow_op_get_by_val): 1423 (JSC::JIT::emit_op_put_by_val): 1424 (JSC::JIT::emitSlow_op_put_by_val): 1425 (JSC::JIT::emit_op_put_by_id): 1426 (JSC::JIT::emitSlow_op_put_by_id): 1427 (JSC::JIT::patchPutByIdReplace): 1428 14292009-06-29 Sam Weinig <sam@webkit.org> 1430 1431 Reviewed by Gavin Barraclough. 1432 1433 Various cleanups. 1434 - Use fpRegT* instead of X86::xmm*. 1435 - Use a switch statement in emitBinaryDoubleOp instead of a bunch of 1436 if/elses. 1437 1438 * jit/JITArithmetic.cpp: 1439 (JSC::JIT::emitAdd32Constant): 1440 (JSC::JIT::emitBinaryDoubleOp): 1441 (JSC::JIT::emit_op_div): 1442 14432009-06-29 Sam Weinig <sam@webkit.org> 1444 1445 Reviewed by Geoffrey Garen. 1446 1447 Add inline code dealing with doubles for op_jfalse and op_jtrue. 1448 1449 * assembler/MacroAssemblerX86Common.h: 1450 (JSC::MacroAssemblerX86Common::): 1451 (JSC::MacroAssemblerX86Common::zeroDouble): 1452 * jit/JITOpcodes.cpp: 1453 (JSC::JIT::emit_op_jfalse): 1454 (JSC::JIT::emit_op_jtrue): 1455 14562009-06-28 Geoffrey Garen <ggaren@apple.com> 1457 1458 Reviewed by Sam Weinig. 1459 1460 Standardized op_get_by_id to put { tag, payload } in { regT1, regT0 }. 1461 1462 SunSpider and v8 report maybe 0.2%-0.4% regressions, but the optimization 1463 this enables will win much more than that back. 1464 1465 * jit/JIT.cpp: 1466 (JSC::JIT::privateCompileCTIMachineTrampolines): 1467 * jit/JIT.h: 1468 * jit/JITPropertyAccess.cpp: 1469 (JSC::JIT::emit_op_method_check): 1470 (JSC::JIT::emit_op_get_by_id): 1471 (JSC::JIT::compileGetByIdHotPath): 1472 (JSC::JIT::compileGetByIdSlowCase): 1473 (JSC::JIT::patchGetByIdSelf): 1474 (JSC::JIT::privateCompilePatchGetArrayLength): 1475 (JSC::JIT::privateCompileGetByIdProto): 1476 (JSC::JIT::privateCompileGetByIdSelfList): 1477 (JSC::JIT::privateCompileGetByIdProtoList): 1478 (JSC::JIT::privateCompileGetByIdChainList): 1479 (JSC::JIT::privateCompileGetByIdChain): 1480 14812009-06-26 Geoffrey Garen <ggaren@apple.com> 1482 1483 Reviewed by Maciej Stachowiak. 1484 1485 Standardized op_call to put { tag, payload } in { regT1, regT0 }. 1486 1487 SunSpider and v8 report no change. 1488 1489 * jit/JIT.cpp: 1490 (JSC::JIT::privateCompileCTIMachineTrampolines): 1491 * jit/JITCall.cpp: 1492 (JSC::JIT::compileOpCallInitializeCallFrame): 1493 (JSC::JIT::compileOpCallSetupArgs): 1494 (JSC::JIT::compileOpConstructSetupArgs): 1495 (JSC::JIT::compileOpCallVarargsSetupArgs): 1496 (JSC::JIT::compileOpCallVarargs): 1497 (JSC::JIT::compileOpCall): 1498 (JSC::JIT::compileOpCallSlowCase): 1499 15002009-06-26 Sam Weinig <sam@webkit.org> 1501 1502 Reviewed by Geoffrey Garen. 1503 1504 Handle multiplying by zero a little better by 1505 inlining the case that both operands are non-negative 1506 into the slowpath. 1507 1508 * assembler/MacroAssemblerX86Common.h: 1509 (JSC::MacroAssemblerX86Common::branchOr32): 1510 * jit/JITArithmetic.cpp: 1511 (JSC::JIT::emit_op_mul): 1512 (JSC::JIT::emitSlow_op_mul): 1513 15142009-06-25 Geoffrey Garen <ggaren@apple.com> 1515 1516 Reviewed by Sam Weinig. 1517 1518 Optimize x++ to ++x inside for loops. 1519 1520 Sadly, no measurable speedup, but this should help with result chaining. 1521 1522 * parser/Nodes.cpp: 1523 (JSC::ForNode::emitBytecode): 1524 15252009-06-25 Geoffrey Garen <ggaren@apple.com> 1526 1527 Reviewed by Sam Weinig. 1528 1529 Standardized some more opcodes to put { tag, payload } in { regT1, regT0 }. 1530 1531 * jit/JITArithmetic.cpp: 1532 (JSC::JIT::emitSlow_op_bitnot): 1533 (JSC::JIT::emit_op_post_inc): 1534 15352009-06-25 Geoffrey Garen <ggaren@apple.com> 1536 1537 Reviewed by Sam Weinig. 1538 1539 Standardized some more opcodes to put { tag, payload } in { regT1, regT0 }. 1540 1541 * jit/JITArithmetic.cpp: 1542 (JSC::JIT::emit_op_bitnot): 1543 (JSC::JIT::emit_op_post_dec): 1544 (JSC::JIT::emit_op_pre_inc): 1545 (JSC::JIT::emitSlow_op_pre_inc): 1546 (JSC::JIT::emit_op_pre_dec): 1547 (JSC::JIT::emitSlow_op_pre_dec): 1548 15492009-06-25 Geoffrey Garen <ggaren@apple.com> 1550 1551 Reviewed by Sam Weinig. 1552 1553 Standardized some more opcodes to put { tag, payload } in { regT1, regT0 }. 1554 1555 * jit/JITArithmetic.cpp: 1556 (JSC::JIT::emit_op_negate): 1557 (JSC::JIT::emitSlow_op_negate): 1558 * jit/JITCall.cpp: 1559 (JSC::JIT::emit_op_construct_verify): 1560 (JSC::JIT::emitSlow_op_construct_verify): 1561 15622009-06-25 Geoffrey Garen <ggaren@apple.com> 1563 1564 Reviewed by Sam Weinig. 1565 1566 Standardized some more opcodes to put { tag, payload } in { regT1, regT0 }. 1567 1568 * jit/JITOpcodes.cpp: 1569 (JSC::JIT::emit_op_loop_if_true): 1570 (JSC::JIT::emit_op_jfalse): 1571 (JSC::JIT::emit_op_jtrue): 1572 (JSC::JIT::emit_op_jeq_null): 1573 (JSC::JIT::emit_op_jneq_null): 1574 (JSC::JIT::emit_op_eq_null): 1575 (JSC::JIT::emit_op_neq_null): 1576 15772009-06-25 Geoffrey Garen <ggaren@apple.com> 1578 1579 Reviewed by Sam Weinig (sort of, maybe). 1580 1581 Fixed some ASSERTs in http/tests/security. 1582 1583 These ASSERTs were introduced by http://trac.webkit.org/changeset/45057, 1584 but the underlying problem was actually older. http://trac.webkit.org/changeset/45057 1585 just exposed the problem by enabling optimization in more cases. 1586 1587 The ASSERTs fired because we tested PropertySlot::slotBase() for validity, 1588 but slotBase() ASSERTs if it's invalid, so we would ASSERT before 1589 the test could happen. Solution: Remove the ASSERT. Maybe it was valid 1590 once, but it clearly goes against a pattern we've deployed of late. 1591 1592 The underlying problem was that WebCore would re-use a PropertySlot in 1593 the case of a forwarding access, and the second use would not completely 1594 overwrite the first use. Solution: Make sure to overwrite m_offset when 1595 setting a value on a PropertySlot. (Other values already get implicitly 1596 overwritten during reuse.) 1597 1598 * runtime/PropertySlot.h: 1599 (JSC::PropertySlot::PropertySlot): 1600 (JSC::PropertySlot::setValueSlot): 1601 (JSC::PropertySlot::setValue): 1602 (JSC::PropertySlot::setRegisterSlot): 1603 (JSC::PropertySlot::setUndefined): 1604 (JSC::PropertySlot::slotBase): 1605 (JSC::PropertySlot::clearOffset): 1606 16072009-06-24 Gavin Barraclough <barraclough@apple.com> 1608 1609 Reviewed by Geoff Garen. 1610 1611 Enable JIT_OPTIMIZE_METHOD_CALLS on the branch, implementation matches current implemenatation in ToT. 1612 1613 * jit/JIT.h: 1614 * jit/JITPropertyAccess.cpp: 1615 (JSC::JIT::emit_op_method_check): 1616 (JSC::JIT::emitSlow_op_method_check): 1617 (JSC::JIT::emit_op_get_by_id): 1618 (JSC::JIT::compileGetByIdHotPath): 1619 (JSC::JIT::emitSlow_op_get_by_id): 1620 (JSC::JIT::compileGetByIdSlowCase): 1621 16222009-06-23 Geoffrey Garen <ggaren@apple.com> 1623 1624 Reviewed by Sam Weinig. 1625 1626 Bit off a tiny bit more of standardizing opcode behavior to help with result 1627 caching. 1628 1629 SunSpider reports no change, v8 maybe a tiny speedup. 1630 1631 * jit/JITOpcodes.cpp: 1632 (JSC::JIT::emit_op_to_jsnumber): 1633 (JSC::JIT::emitSlow_op_to_jsnumber): 1634 (JSC::JIT::emit_op_convert_this): 1635 (JSC::JIT::emitSlow_op_convert_this): 1636 16372009-06-23 Geoffrey Garen <ggaren@apple.com> 1638 1639 Reviewed by Sam Weinig. 1640 1641 Bit off a tiny bit more of standardizing opcode behavior to help with result 1642 caching -- including removing my old enemy, op_resolve_function, because 1643 it was non-standard, and removing it felt better than helping it limp along. 1644 1645 SunSpider reports no change, v8 maybe a tiny speedup. 1646 1647 * bytecode/CodeBlock.cpp: 1648 (JSC::CodeBlock::dump): 1649 * bytecode/Opcode.h: 1650 * bytecompiler/BytecodeGenerator.cpp: 1651 * bytecompiler/BytecodeGenerator.h: 1652 * interpreter/Interpreter.cpp: 1653 (JSC::Interpreter::privateExecute): 1654 * jit/JIT.cpp: 1655 (JSC::JIT::privateCompileMainPass): 1656 * jit/JIT.h: 1657 * jit/JITOpcodes.cpp: 1658 (JSC::JIT::emit_op_get_scoped_var): 1659 (JSC::JIT::emit_op_put_scoped_var): 1660 (JSC::JIT::emit_op_to_primitive): 1661 (JSC::JIT::emitSlow_op_to_primitive): 1662 * jit/JITStubs.cpp: 1663 * jit/JITStubs.h: 1664 * parser/Nodes.cpp: 1665 (JSC::FunctionCallResolveNode::emitBytecode): 1666 16672009-06-23 Geoffrey Garen <ggaren@apple.com> 1668 1669 Reviewed by Sam Weinig. 1670 1671 Bit off a tiny bit of standardizing opcode behavior to help with result 1672 caching. 1673 1674 0.6% SunSpider speedup. 0.3% v8 speedup. 1675 1676 * jit/JITInlineMethods.h: 1677 (JSC::JIT::emitLoad): Accomodate a base register that overlaps with payload 1678 by loading tag before payload, to avoid stomping base/payload. 1679 1680 * jit/JITOpcodes.cpp: 1681 (JSC::JIT::emit_op_mov): Abide by the standard "tag in regT1, payload in 1682 regT0" semantics. 1683 1684 (JSC::JIT::emit_op_get_global_var): 1685 (JSC::JIT::emit_op_put_global_var): Ditto. Also, removed some irrelevent 1686 loads while I was at it. The global object's "d" pointer never changes 1687 after construction. 1688 16892009-06-23 Gavin Barraclough <barraclough@apple.com> 1690 1691 Reviewed by Sam Weinig. 1692 1693 Remove 'arguments' field from Register union (again). 1694 This time do so without breaking tests (radical, I know). 1695 1696 * interpreter/CallFrame.h: 1697 (JSC::ExecState::optionalCalleeArguments): 1698 (JSC::ExecState::setArgumentCount): 1699 (JSC::ExecState::init): 1700 * interpreter/Interpreter.cpp: 1701 (JSC::Interpreter::dumpRegisters): 1702 (JSC::Interpreter::unwindCallFrame): 1703 (JSC::Interpreter::privateExecute): 1704 (JSC::Interpreter::retrieveArguments): 1705 * interpreter/Register.h: 1706 (JSC::Register::withInt): 1707 (JSC::Register::): 1708 (JSC::Register::Register): 1709 (JSC::Register::i): 1710 * jit/JITStubs.cpp: 1711 (JSC::JITStubs::cti_op_tear_off_arguments): 1712 * runtime/Arguments.h: 1713 (JSC::JSActivation::copyRegisters): 1714 (JSC::Register::arguments): 1715 * runtime/JSActivation.cpp: 1716 (JSC::JSActivation::argumentsGetter): 1717 * runtime/JSActivation.h: 1718 17192009-06-23 Geoffrey Garen <ggaren@apple.com> 1720 1721 Reviewed by Sam Weinig. 1722 1723 Removed some result register tracking cruft in preparation for a new 1724 result tracking mechanism. 1725 1726 SunSpider reports no change. 1727 1728 * assembler/AbstractMacroAssembler.h: 1729 * assembler/X86Assembler.h: 1730 (JSC::X86Assembler::JmpDst::JmpDst): No need to track jump targets in 1731 machine code; we already do this in bytecode. 1732 1733 * jit/JIT.cpp: 1734 (JSC::JIT::JIT): 1735 (JSC::JIT::emitTimeoutCheck): Make sure to save and restore the result 1736 registers, so an opcode with a timeout check can still benefit from result 1737 register caching. 1738 1739 (JSC::JIT::privateCompileMainPass): 1740 (JSC::JIT::privateCompileSlowCases): Removed calls to killLastResultRegister() 1741 in preparation for something new. 1742 1743 * jit/JIT.h: 1744 * jit/JITArithmetic.cpp: 1745 (JSC::JIT::emit_op_jnless): 1746 (JSC::JIT::emit_op_jnlesseq): 1747 * jit/JITInlineMethods.h: 1748 (JSC::JIT::emitGetFromCallFrameHeaderPtr): 1749 (JSC::JIT::emitGetFromCallFrameHeader32): 1750 * jit/JITOpcodes.cpp: 1751 (JSC::JIT::emit_op_jmp): 1752 (JSC::JIT::emit_op_jfalse): 1753 (JSC::JIT::emit_op_jtrue): 1754 (JSC::JIT::emit_op_jeq_null): 1755 (JSC::JIT::emit_op_jneq_null): 1756 (JSC::JIT::emit_op_jneq_ptr): 1757 (JSC::JIT::emit_op_jsr): 1758 (JSC::JIT::emit_op_sret): 1759 (JSC::JIT::emit_op_jmp_scopes): ditto 1760 1761 * jit/JITStubCall.h: 1762 (JSC::JITStubCall::JITStubCall): 1763 (JSC::JITStubCall::getArgument): added a mechanism for reloading an argument 1764 you passed to a JIT stub, for use in emitTimeoutCheck. 1765 17662009-06-23 Sam Weinig <sam@webkit.org> 1767 1768 Reviewed by Geoffrey Garen. 1769 1770 Remove now-useless inplace variants of binary ops. 1771 1772 * jit/JIT.h: 1773 * jit/JITArithmetic.cpp: 1774 (JSC::JIT::emit_op_bitand): 1775 (JSC::JIT::emit_op_bitor): 1776 (JSC::JIT::emit_op_bitxor): 1777 (JSC::JIT::emit_op_add): 1778 (JSC::JIT::emit_op_sub): 1779 (JSC::JIT::emit_op_mul): 1780 17812009-06-23 Sam Weinig <sam@webkit.org> 1782 1783 Reviewed by Geoffrey Garen. 1784 1785 Move off memory operands to aid in re-enabling result caching. 1786 1787 - No regression measured. 1788 1789 * jit/JIT.h: 1790 * jit/JITArithmetic.cpp: 1791 (JSC::JIT::emit_op_negate): 1792 (JSC::JIT::emit_op_jnless): 1793 (JSC::JIT::emit_op_jnlesseq): 1794 (JSC::JIT::emit_op_lshift): 1795 (JSC::JIT::emit_op_rshift): 1796 (JSC::JIT::emit_op_bitand): 1797 (JSC::JIT::emitBitAnd32Constant): 1798 (JSC::JIT::emitBitAnd32InPlace): 1799 (JSC::JIT::emit_op_bitor): 1800 (JSC::JIT::emitBitOr32Constant): 1801 (JSC::JIT::emitBitOr32InPlace): 1802 (JSC::JIT::emit_op_bitxor): 1803 (JSC::JIT::emitBitXor32Constant): 1804 (JSC::JIT::emitBitXor32InPlace): 1805 (JSC::JIT::emit_op_bitnot): 1806 (JSC::JIT::emit_op_post_inc): 1807 (JSC::JIT::emit_op_post_dec): 1808 (JSC::JIT::emit_op_pre_inc): 1809 (JSC::JIT::emitSlow_op_pre_inc): 1810 (JSC::JIT::emit_op_pre_dec): 1811 (JSC::JIT::emitSlow_op_pre_dec): 1812 (JSC::JIT::emit_op_add): 1813 (JSC::JIT::emitAdd32Constant): 1814 (JSC::JIT::emitAdd32InPlace): 1815 (JSC::JIT::emitSlow_op_add): 1816 (JSC::JIT::emitSlowAdd32Constant): 1817 (JSC::JIT::emit_op_sub): 1818 (JSC::JIT::emitSlow_op_sub): 1819 (JSC::JIT::emitSub32ConstantLeft): 1820 (JSC::JIT::emitSub32ConstantRight): 1821 (JSC::JIT::emitSub32InPlaceLeft): 1822 (JSC::JIT::emitSub32InPlaceRight): 1823 (JSC::JIT::emitBinaryDoubleOp): 1824 (JSC::JIT::emit_op_mul): 1825 (JSC::JIT::emitMul32InPlace): 1826 (JSC::JIT::emit_op_div): 1827 (JSC::JIT::emit_op_mod): 1828 * jit/JITCall.cpp: 1829 (JSC::JIT::compileOpCallVarargs): 1830 * jit/JITOpcodes.cpp: 1831 (JSC::JIT::emit_op_loop_if_less): 1832 (JSC::JIT::emit_op_loop_if_lesseq): 1833 (JSC::JIT::emit_op_instanceof): 1834 (JSC::JIT::emit_op_to_primitive): 1835 (JSC::JIT::emit_op_not): 1836 (JSC::JIT::emit_op_jneq_ptr): 1837 (JSC::JIT::emit_op_eq): 1838 (JSC::JIT::emit_op_neq): 1839 (JSC::JIT::emit_op_to_jsnumber): 1840 * jit/JITPropertyAccess.cpp: 1841 (JSC::JIT::emit_op_get_by_val): 1842 (JSC::JIT::emit_op_put_by_val): 1843 18442009-06-23 Geoffrey Garen <ggaren@apple.com> 1845 1846 Reviewed by Sam Weinig. 1847 1848 Fixed some missing and/or misplaced labels in bytecode generation, so 1849 we don't have to work around them in JIT code generation. 1850 1851 * bytecompiler/BytecodeGenerator.cpp: 1852 (JSC::BytecodeGenerator::emitJumpSubroutine): 1853 * parser/Nodes.cpp: 1854 (JSC::TryNode::emitBytecode): 1855 18562009-06-22 Geoffrey Garen <ggaren@apple.com> 1857 1858 Reviewed by Sam Weinig. 1859 1860 For member function calls, emit "this" directly into the "this" slot 1861 for the function call, instead of moving it there later. This reduces 1862 time spent in op_mov during certain calls, like "a.b.c()". 1863 1864 1%-2% speedup on v8, mostly richards and delta-blue. 1865 1866 * parser/Nodes.cpp: 1867 (JSC::FunctionCallDotNode::emitBytecode): 1868 18692009-06-22 Gavin Barraclough <barraclough@apple.com> 1870 1871 Reviewed by Sam Weinig. 1872 1873 Remove 'arguments' field from Register union. Having JSCell derived types in the union is 1874 dangerous since it opens the possibility for the field to be written as a raw pointer but 1875 then read as a JSValue. This will lead to statle data being read for the tag, which may 1876 be dangerous. Having removed Arguments* types form Register, all arguments objects must 1877 always explicitly be stored in the register file as JSValues. 1878 1879 * interpreter/CallFrame.h: 1880 (JSC::ExecState::optionalCalleeArguments): 1881 * interpreter/Interpreter.cpp: 1882 (JSC::Interpreter::unwindCallFrame): 1883 (JSC::Interpreter::privateExecute): 1884 (JSC::Interpreter::retrieveArguments): 1885 * interpreter/Register.h: 1886 (JSC::Register::): 1887 * jit/JITStubs.cpp: 1888 (JSC::JITStubs::cti_op_tear_off_arguments): 1889 * runtime/Arguments.h: 1890 (JSC::JSActivation::copyRegisters): 1891 * runtime/JSActivation.cpp: 1892 (JSC::JSActivation::argumentsGetter): 1893 * runtime/JSActivation.h: 1894 18952009-06-03 Sam Weinig <sam@webkit.org> 1896 1897 Reviewed by Geoffrey Garen. 1898 1899 Add back known this value optimization by abstracting 1900 slow case if not JSCell jumps. 1901 1902 * jit/JIT.h: 1903 * jit/JITCall.cpp: 1904 (JSC::JIT::compileOpCallVarargs): 1905 (JSC::JIT::compileOpCallVarargsSlowCase): 1906 (JSC::JIT::compileOpCall): 1907 (JSC::JIT::compileOpCallSlowCase): 1908 * jit/JITInlineMethods.h: 1909 (JSC::JIT::emitJumpSlowCaseIfNotJSCell): 1910 (JSC::JIT::linkSlowCaseIfNotJSCell): 1911 * jit/JITOpcodes.cpp: 1912 (JSC::JIT::emit_op_instanceof): 1913 (JSC::JIT::emitSlow_op_instanceof): 1914 * jit/JITPropertyAccess.cpp: 1915 (JSC::JIT::emit_op_get_by_val): 1916 (JSC::JIT::emitSlow_op_get_by_val): 1917 (JSC::JIT::emit_op_put_by_val): 1918 (JSC::JIT::emitSlow_op_put_by_val): 1919 (JSC::JIT::emit_op_get_by_id): 1920 (JSC::JIT::emitSlow_op_get_by_id): 1921 (JSC::JIT::emit_op_put_by_id): 1922 (JSC::JIT::emitSlow_op_put_by_id): 1923 19242009-06-01 Geoffrey Garen <ggaren@apple.com> 1925 1926 Reviewed by Sam Weinig. 1927 1928 Fixed some of the regression in crypto-aes.js. (8.5% speedup in 1929 crypto-aes.js.) 1930 1931 SunSpider reports no change overall. 1932 1933 Division was producing double results, which took the slow path through 1934 array access code. 1935 1936 Strangely, all my attempts at versions of this patch that modified array 1937 access code to accept ints encoded as doubles along the fast or slow paths 1938 were regressions. So I did this instead. 1939 1940 * jit/JITArithmetic.cpp: 1941 (JSC::JIT::emit_op_div): When dividing an int by an int, go ahead and try 1942 to turn the result into an int. Don't just do int division, though, because 1943 testing shows it to be slower than SSE double division, and the corner 1944 cases are pretty complicated / lengthy on top of that. Also, don't try 1945 to canonicalize division of known tiny numerators into ints, since that's a 1946 waste of time. 1947 19482009-05-26 Geoffrey Garen <ggaren@apple.com> 1949 1950 Reviewed by Oliver Hunt. 1951 1952 Fixed a regression caused by my recent fix for NaN. 1953 1954 * jit/JITArithmetic.cpp: 1955 (JSC::JIT::emitBinaryDoubleOp): Actually do the comparison in reverse 1956 order, like the ChangeLog said we would, bokay? 1957 19582009-05-26 Geoffrey Garen <ggaren@apple.com> 1959 1960 Reviewed by Sam Weinig and Oliver Hunt. 1961 1962 Fixed two edge cases in %: 1963 1964 - Don't do -2147483648 % x as a fast case, since you might do -2147483648 % -1, 1965 which will signal a hardware exception due to overflow. 1966 1967 - In the case of a zero remainder, be sure to store negative zero if the 1968 dividend was zero. 1969 1970 SunSpider reports no change. 1971 1972 * jit/JITArithmetic.cpp: 1973 (JSC::JIT::emit_op_mod): 1974 (JSC::JIT::emitSlow_op_mod): 1975 19762009-05-25 Geoffrey Garen <ggaren@apple.com> 1977 1978 Reviewed by Maciej Stachowiak. 1979 1980 Fixed a regression when comparing to NaN. 1981 1982 * jit/JITArithmetic.cpp: 1983 (JSC::JIT::emitBinaryDoubleOp): For op_jnless and op_jnless_eq, do the 1984 comparison in reverse order, and jump if the result is below or 1985 below-or-equal. This ensures that we do jump in the case of NaN. 1986 19872009-05-25 Geoffrey Garen <ggaren@apple.com> 1988 1989 Reviewed by Oliver Hunt. 1990 1991 SunSpider says no change. 1992 1993 Fixed regressions in fast/js/var-declarations-shadowing.html and 1994 fast/js/equality.html, caused by recent == and != optimizations. 1995 1996 * jit/JITStubs.cpp: 1997 (JSC::JITStubs::cti_op_eq): Don't treat "compare to string" as always 1998 numeric or string comparison. If the second operand is an object, you 1999 need to ToPrimitive it, and start all over again. Also, I wrote out each 2000 of the possible cases explicitly, to cut down on redundant branching. 2001 20022009-05-25 Sam Weinig <sam@webkit.org> 2003 2004 Reviewed by Mark Rowe. 2005 2006 Fix bug in fast/js/constant-folding.html where we were not negating 2007 -0 properly. 2008 2009 * jit/JITArithmetic.cpp: 2010 (JSC::JIT::emit_op_negate): 2011 20122009-05-23 Geoffrey Garen <ggaren@apple.com> 2013 2014 Reviewed by Oliver Hunt. 2015 2016 Refactored new slow case codegen for == and !=. 2017 2018 SunSpider reports no change, maybe a tiny speedup. 2019 2020 * jit/JITOpcodes.cpp: 2021 (JSC::JIT::emitSlow_op_eq): 2022 (JSC::JIT::emitSlow_op_neq): Made a vptr comparison a *Ptr operation, 2023 instead of *32, to make it portable to 64bit. Reorganized the string 2024 and generic cases to make their control flow a little clearer. 2025 20262009-05-23 Geoffrey Garen <ggaren@apple.com> 2027 2028 Reviewed by Maciej Stachowiak. 2029 2030 Optimized == and != for our new value representation -- especially for strings. 2031 2032 14% speedup on date-format-tofte. 2033 2034 * jit/JITOpcodes.cpp: 2035 (JSC::JIT::emit_op_eq): 2036 (JSC::JIT::emitSlow_op_eq): 2037 (JSC::JIT::emit_op_neq): 2038 (JSC::JIT::emitSlow_op_neq): 2039 * jit/JITStubCall.h: 2040 (JSC::JITStubCall::JITStubCall): 2041 * jit/JITStubs.cpp: 2042 (JSC::JITStubs::cti_op_eq): 2043 (JSC::JITStubs::cti_op_eq_strings): 2044 (JSC::JITStubs::cti_op_call_eval): 2045 * jit/JITStubs.h: 2046 (JSC::): 2047 * runtime/JSValue.h: 2048 20492009-05-22 Sam Weinig <sam@webkit.org> 2050 2051 Reviewed by Gavin Barraclough. 2052 2053 Fix non-SSE enabled builds. 2054 2055 * jit/JITArithmetic.cpp: 2056 (JSC::JIT::emitSlow_op_add): Don't early return here, we still need to call the JIT stub. 2057 (JSC::JIT::emitSlow_op_sub): Ditto. 2058 20592009-05-22 Geoffrey Garen <ggaren@apple.com> 2060 2061 Reviewed by Sam Weinig. 2062 2063 Here's a thought: let's not take a jit stub call just to multiply by 1, 2064 bokay? 2065 2066 imul doesn't set the zero flag, so to test for a zero result, we need 2067 an explicit instruction. (Luckily, it does set the overflow flag, so 2068 we can still use that.) 2069 2070 * jit/JIT.h: 2071 * jit/JITArithmetic.cpp: 2072 (JSC::JIT::emit_op_mul): 2073 (JSC::JIT::emitSlow_op_mul): 2074 (JSC::JIT::emitMul32InPlace): 2075 20762009-05-22 Sam Weinig <sam@webkit.org> 2077 2078 Reviewed by Geoffrey "Premature Commit" Garen. 2079 2080 Add back constant integer cases for op_add. 2081 2082 * jit/JIT.h: 2083 * jit/JITArithmetic.cpp: 2084 (JSC::JIT::emit_op_add): 2085 (JSC::JIT::emitAdd32Constant): 2086 (JSC::JIT::emitSlow_op_add): 2087 (JSC::JIT::emitSlowAdd32Constant): 2088 * jit/JITInlineMethods.h: 2089 (JSC::JIT::getConstantOperandImmediateDouble): 2090 (JSC::JIT::isOperandConstantImmediateDouble): 2091 20922009-05-22 Geoffrey Garen <ggaren@apple.com> 2093 2094 Reviewed by Sam Weinig. 2095 2096 Added fast double cases for op_jnless and op_jnlesseq. 2097 2098 * assembler/AbstractMacroAssembler.h: 2099 (JSC::AbstractMacroAssembler::JumpList::jumps): New accesor, used by 2100 addSlowCase. 2101 2102 * assembler/X86Assembler.h: 2103 (JSC::X86Assembler::ucomisd_rm): New method for comparing register to 2104 memory. 2105 2106 * jit/JIT.h: 2107 * jit/JITArithmetic.cpp: 2108 (JSC::JIT::emit_op_jnless): 2109 (JSC::JIT::emitSlow_op_jnless): 2110 (JSC::JIT::emit_op_jnlesseq): 2111 (JSC::JIT::emitSlow_op_jnlesseq): 2112 (JSC::JIT::emit_op_add): 2113 (JSC::JIT::emit_op_sub): 2114 (JSC::JIT::emitBinaryDoubleOp): 2115 (JSC::JIT::emit_op_mul): 2116 (JSC::JIT::emit_op_div): Modified emitBinaryDoubleOp to accept comparison/jump 2117 operations in addition to operations with explicit result registers. 2118 2119 * jit/JITInlineMethods.h: 2120 (JSC::JIT::addSlowCase): Added an "addSlowCase" for JumpLists, so clients 2121 can track multiple jumps to the same slow case condition together. 2122 21232009-05-21 Sam Weinig <sam@webkit.org> 2124 2125 Reviewed by Gavin Barraclough. 2126 2127 Implement op_negate inline fast cases. 2128 2129 * assembler/MacroAssemblerX86Common.h: 2130 (JSC::MacroAssemblerX86Common::neg32): 2131 * assembler/X86Assembler.h: 2132 (JSC::X86Assembler::): 2133 (JSC::X86Assembler::negl_m): 2134 (JSC::X86Assembler::xorpd_rr): 2135 * jit/JIT.cpp: 2136 (JSC::JIT::privateCompileMainPass): 2137 (JSC::JIT::privateCompileSlowCases): 2138 * jit/JIT.h: 2139 * jit/JITArithmetic.cpp: 2140 (JSC::JIT::emit_op_negate): 2141 (JSC::JIT::emitSlow_op_negate): 2142 21432009-05-20 Sam Weinig <sam@webkit.org> 2144 2145 Reviewed by Gavin Barraclough. 2146 2147 Update the patchOffsetGetByIdSlowCaseCall constant for the 2148 case that OPCODE_SAMPLING is enabled. 2149 2150 * jit/JIT.h: 2151 21522009-05-20 Geoffrey Garen <ggaren@apple.com> 2153 2154 Reviewed by Sam Weinig. 2155 2156 Added support for inline subtraction of doubles. 2157 2158 * jit/JITArithmetic.cpp: 2159 (JSC::JIT::emit_op_sub): 2160 (JSC::JIT::emitSlow_op_sub): 2161 (JSC::JIT::emitSlowSub32InPlaceLeft): 2162 (JSC::JIT::emitBinaryDoubleOp): 2163 21642009-05-20 Sam Weinig <sam@webkit.org> 2165 2166 Reviewed by Geoffrey Garen. 2167 2168 Added support for inline division. 2169 2170 * assembler/X86Assembler.h: 2171 (JSC::X86Assembler::): 2172 (JSC::X86Assembler::divsd_rr): 2173 (JSC::X86Assembler::divsd_mr): 2174 * bytecode/CodeBlock.cpp: 2175 (JSC::CodeBlock::dump): 2176 * bytecode/Opcode.h: 2177 * bytecompiler/BytecodeGenerator.cpp: 2178 (JSC::BytecodeGenerator::emitBinaryOp): 2179 * interpreter/Interpreter.cpp: 2180 (JSC::Interpreter::privateExecute): 2181 * jit/JIT.cpp: 2182 (JSC::JIT::privateCompileMainPass): 2183 (JSC::JIT::privateCompileSlowCases): 2184 * jit/JIT.h: 2185 * jit/JITArithmetic.cpp: 2186 (JSC::JIT::emitBinaryDoubleOp): 2187 (JSC::JIT::emit_op_div): 2188 (JSC::JIT::emitSlow_op_div): 2189 21902009-05-20 Geoffrey Garen <ggaren@apple.com> 2191 2192 Reviewed by Sam Weinig. 2193 2194 Added support for inline addition of doubles. 2195 2196 * jit/JITArithmetic.cpp: 2197 (JSC::JIT::emit_op_add): 2198 (JSC::JIT::emitSlow_op_add): 2199 (JSC::JIT::emitSlowAdd32InPlace): 2200 (JSC::JIT::emitBinaryDoubleOp): 2201 (JSC::JIT::emit_op_mul): 2202 (JSC::JIT::emitSlow_op_mul): 2203 22042009-05-20 Geoffrey Garen <ggaren@apple.com> 2205 2206 Reviewed by Sam Weinig. 2207 2208 Factored inline double operations into a helper function, so that we 2209 can reuse this code for other math operations. 2210 2211 * jit/JIT.h: 2212 * jit/JITArithmetic.cpp: 2213 (JSC::JIT::emitBinaryDoubleOp): 2214 (JSC::JIT::emit_op_mul): 2215 * jit/JITCall.cpp: 2216 (JSC::JIT::compileOpCallInitializeCallFrame): 2217 22182009-05-20 Geoffrey Garen <ggaren@apple.com> 2219 2220 Reviewed by Sam Weinig. 2221 2222 Added support for inline multiplication of doubles. 2223 2224 * assembler/X86Assembler.h: 2225 (JSC::X86Assembler::cvtsi2sd_mr): New function, useful for loading an 2226 int32 into a double register. 2227 2228 * jit/JITArithmetic.cpp: 2229 (JSC::JIT::emit_op_mul): 2230 (JSC::JIT::emitSlow_op_mul): Filled out these cases for double arithmetic. 2231 2232 * jit/JIT.h: 2233 * jit/JITInlineMethods.h: 2234 (JSC::JIT::addressFor): New function, useful for addressing a JSValue's 2235 full 64bits as a double. 2236 22372009-05-19 Sam Weinig <sam@webkit.org> 2238 2239 Reviewed by Geoffrey Garen. 2240 2241 Implement and enable optimized calls. 2242 2243 * jit/JIT.cpp: 2244 (JSC::JIT::privateCompileCTIMachineTrampolines): Add ENABLE(JIT_OPTIMIZE_CALL) guards 2245 around the the optimize call only trampolines (virtualCallPreLink and virtualCallLink). 2246 Update the trampolines to account for the new JSValue representation. 2247 (JSC::JIT::unlinkCall): Use NULL instead of JSValue noValue. 2248 2249 * jit/JITCall.cpp: 2250 (JSC::JIT::compileOpCall): Update to account for the new JSValue representation 2251 (JSC::JIT::compileOpCallSlowCase): Ditto. 2252 2253 * jit/JITStubs.h: Remove incorrect !ENABLE(JIT_OPTIMIZE_CALL) guard. 2254 2255 * wtf/Platform.h: Enable ENABLE_JIT_OPTIMIZE_CALL. 2256 22572009-05-19 Sam Weinig <sam@webkit.org> 2258 2259 Reviewed by Geoffrey Garen. 2260 2261 Implement and enable optimized property access. 2262 2263 * assembler/AbstractMacroAssembler.h: Fix comment. 2264 * jit/JIT.cpp: 2265 (JSC::JIT::privateCompileCTIMachineTrampolines): Remove array length trampoline 2266 and implement the string length trampoline. 2267 * jit/JIT.h: Add new constants for patch offsets. 2268 * jit/JITInlineMethods.h: Remove FIELD_OFFSET which is now in StdLibExtras.h. 2269 * jit/JITPropertyAccess.cpp: 2270 (JSC::JIT::emit_op_get_by_id): 2271 (JSC::JIT::emitSlow_op_get_by_id): 2272 (JSC::JIT::emit_op_put_by_id): 2273 (JSC::JIT::emitSlow_op_put_by_id): 2274 (JSC::JIT::compilePutDirectOffset): 2275 (JSC::JIT::compileGetDirectOffset): 2276 (JSC::JIT::privateCompilePutByIdTransition): 2277 (JSC::JIT::patchGetByIdSelf): 2278 (JSC::JIT::patchPutByIdReplace): 2279 (JSC::JIT::privateCompilePatchGetArrayLength): 2280 (JSC::JIT::privateCompileGetByIdProto): 2281 (JSC::JIT::privateCompileGetByIdSelfList): 2282 (JSC::JIT::privateCompileGetByIdProtoList): 2283 (JSC::JIT::privateCompileGetByIdChainList): 2284 (JSC::JIT::privateCompileGetByIdChain): 2285 * jit/JITStubCall.h: 2286 (JSC::JITStubCall::addArgument): Add version of addArgument that takes 2287 two registers for the tag and payload. 2288 * jit/JITStubs.cpp: 2289 (JSC::JITStubs::JITStubs): Remove array length trampoline pointer. 2290 (JSC::JITStubs::cti_op_get_by_id_self_fail): 2291 * jit/JITStubs.h: 2292 * runtime/JSObject.h: 2293 (JSC::JSObject::JSObject): Move m_inheritorID below the property storage 2294 to align it to a 16 byte boundary. 2295 * wtf/Platform.h: Enable ENABLE_JIT_OPTIMIZE_PROPERTY_ACCESS 2296 * wtf/StdLibExtras.h: Move FIELD_OFFSET here. 2297 22982009-05-17 Sam Weinig <sam@webkit.org> 2299 2300 Reviewed by Geoffrey Garen. 2301 2302 Remove unneeded ExecState parameter from the number JSValue constructors. 2303 2304 * runtime/JSValue.h: 2305 (JSC::jsNumber): 2306 (JSC::jsNaN): 2307 (JSC::JSValue::JSValue): 2308 23092009-05-15 Sam Weinig <sam@webkit.org> 2310 2311 Reviewed by Geoffrey Garen. 2312 2313 Implemented fast path for op_put_by_val when putting to arrays. 2314 2315 * jit/JITPropertyAccess.cpp: 2316 (JSC::JIT::emit_op_put_by_val): 2317 (JSC::JIT::emitSlow_op_put_by_val): 2318 23192009-05-15 Geoffrey Garen <ggaren@apple.com> (Mostly by Sam) 2320 2321 Reviewed by Sam Weinig. 2322 2323 Implemented fast path for op_get_by_val when accessing array. 2324 2325 * jit/JIT.cpp: 2326 * jit/JITPropertyAccess.cpp: 2327 (JSC::JIT::emit_op_get_by_val): 2328 (JSC::JIT::emitSlow_op_get_by_val): 2329 23302009-05-14 Geoffrey Garen <ggaren@apple.com> 2331 2332 Reviewed by Sam Weinig. 2333 2334 Fixed a failure in fast/js/math-transforms.html caused by failing to 2335 preserve -0 in multiplication. 2336 2337 * assembler/X86Assembler.h: 2338 (JSC::X86Assembler::jz): 2339 * jit/JITArithmetic.cpp: 2340 (JSC::JIT::emit_op_mul): 2341 (JSC::JIT::emitSlow_op_mul): 2342 (JSC::JIT::emitMul32Constant): 2343 (JSC::JIT::emitMul32InPlace): Check both for overflow and for zero when 2344 doing multiplication. Use a slow case to get these right. 2345 23462009-05-14 Geoffrey Garen <ggaren@apple.com> 2347 2348 Reviewed by Sam Weinig. 2349 2350 Fixed a bug in the varargs calling convention. 2351 2352 * jit/JITCall.cpp: 2353 (JSC::JIT::compileOpCallVarargs): Move the argument count into regT1, 2354 since that's where ctiVirtualCall expects it to be. 2355 23562009-05-14 Geoffrey Garen <ggaren@apple.com> 2357 2358 Reviewed by Sam Weinig. 2359 2360 Fixed a small bug in instanceof's looping code. 2361 2362 * jit/JITOpcodes.cpp: 2363 (JSC::JIT::emit_op_instanceof): NULL means the object has no prototype, 2364 so only loop when *not* equal to NULL. 2365 23662009-05-14 Geoffrey Garen <ggaren@apple.com> 2367 2368 Reviewed by Sam Weinig. 2369 2370 Fixed a small bug in instanceof's result writing code. 2371 2372 * jit/JITOpcodes.cpp: 2373 (JSC::JIT::emit_op_instanceof): Make sure to fill out the payload bits 2374 in all cases. 2375 23762009-05-14 Sam Weinig <sam@webkit.org> 2377 2378 Reviewed by Geoffrey Garen. 2379 2380 Removed an invalid assertion in cti_op_urshift which 2381 depended on a fast path for op_urshift which has 2382 never existed. 2383 2384 * jit/JITStubs.cpp: 2385 (JSC::JITStubs::cti_op_urshift): 2386 23872009-05-14 Geoffrey Garen <ggaren@apple.com> 2388 2389 Reviewed by Sam Weinig. 2390 2391 Fixed loop_if_true, which had the same reversed test that jtrue had. 2392 2393 * jit/JITOpcodes.cpp: 2394 (JSC::JIT::emit_op_loop_if_true): 2395 23962009-05-14 Sam Weinig <sam@webkit.org> 2397 2398 Reviewed by Geoffrey Garen. 2399 2400 In op_neq, we apparently want to check that one value 2401 does *not* equal another. Go figure. 2402 2403 * jit/JITOpcodes.cpp: 2404 (JSC::JIT::emit_op_neq): 2405 24062009-05-14 Sam Weinig <sam@webkit.org> 2407 2408 Reviewed by Geoffrey Garen. 2409 2410 The slow case of op_mod should call op_mod's jit stub, 2411 not op_mul. That would be dumb. 2412 2413 * jit/JITArithmetic.cpp: 2414 (JSC::JIT::emitSlow_op_mod): 2415 24162009-05-14 Geoffrey Garen <ggaren@apple.com> 2417 2418 Reviewed by Sam Weinig. 2419 2420 Fixed problems when using 'arguments' due to a half-initialized register. 2421 2422 * interpreter/CallFrame.h: 2423 (JSC::ExecState::setCalleeArguments): 2424 (JSC::ExecState::init): Require a full JSValue when setting up the 2425 'arguments' virtual register, since this register is accessible from JIT 2426 code and bytecode, and needs to be a true JSValue. 2427 2428 * interpreter/CallFrameClosure.h: 2429 (JSC::CallFrameClosure::resetCallFrame): ditto 2430 2431 * interpreter/Interpreter.cpp: 2432 (JSC::Interpreter::privateExecute): ditto 2433 2434 * interpreter/Register.h: Removed the constructor that allowed assignment 2435 of a JSArguments* to a register. That is not safe. See above. 2436 2437 * jit/JITStubs.cpp: 2438 (JSC::JITStubs::cti_op_create_arguments): 2439 (JSC::JITStubs::cti_op_create_arguments_no_params): ditto 2440 24412009-05-14 Sam Weinig <sam@webkit.org> 2442 2443 Reviewed by Geoffrey Garen. 2444 2445 We really want to go to the slow case in op_jfalse and 2446 op_jtrue if the value is *not* boolean. 2447 2448 * jit/JITOpcodes.cpp: 2449 (JSC::JIT::emit_op_jfalse): 2450 (JSC::JIT::emit_op_jtrue): 2451 24522009-05-14 Sam Weinig <sam@webkit.org> 2453 2454 Reviewed by Geoffrey Garen. 2455 2456 Flipped the condition when emitting a an op_loop_if_less or op_loop_if_lesseq 2457 if the first operand is a constant. 2458 2459 * jit/JITOpcodes.cpp: 2460 (JSC::JIT::emit_op_loop_if_less): 2461 (JSC::JIT::emit_op_loop_if_lesseq): 2462 24632009-05-14 Sam Weinig <sam@webkit.org> 2464 2465 Reviewed by Geoffrey Garen. 2466 2467 Added missing return in op_jnless and op_jnlesseq. 2468 2469 * jit/JITArithmetic.cpp: 2470 (JSC::JIT::emit_op_jnless): 2471 (JSC::JIT::emit_op_jnlesseq): 2472 24732009-05-14 Sam Weinig <sam@webkit.org> 2474 2475 Reviewed by Geoffrey Garen. 2476 2477 Load constants into the the register file as a temporary measure to 2478 aid bring up. This allows us to use to treat constants like any 2479 other virtual register. 2480 2481 * jit/JITOpcodes.cpp: 2482 (JSC::JIT::emit_op_enter): 2483 (JSC::JIT::emit_op_enter_with_activation): 2484 24852009-05-14 Geoffrey Garen <ggaren@apple.com> 2486 2487 Reviewed by Sam Weinig. 2488 2489 Implemented op_strict_eq. Original patch by Snowy, by way of Sam and Gavin. 2490 2491 * assembler/MacroAssemblerX86Common.h: 2492 (JSC::MacroAssemblerX86Common::set8): Added set8, since it's slightly 2493 faster than set32, and the new value representation usually doesn't 2494 need set32. 2495 2496 * jit/JIT.cpp: 2497 * jit/JIT.h: 2498 * jit/JITInlineMethods.h: 2499 (JSC::JIT::emitLoadTag): 2500 (JSC::JIT::emitLoadPayload): Added helper functions for dealing with 2501 constants. Eventually, we should write special cases for all constants, 2502 but these are helpful in the short term. 2503 2504 * jit/JITOpcodes.cpp: 2505 (JSC::JIT::compileOpStrictEq): 2506 (JSC::JIT::emitSlow_op_stricteq): 2507 (JSC::JIT::emitSlow_op_nstricteq): teh opcodez. 2508 2509 * runtime/JSValue.h: 2510 (JSC::JSValue::): 2511 (JSC::JSValue::isDouble): Added a LowestTag for clarity. 2512 25132009-05-13 Geoffrey Garen <ggaren@apple.com> 2514 2515 Reviewed by Sam Weinig. 2516 2517 Fixed some bugs in host function calls. 2518 2519 testapi now passes! 2520 2521 * jit/JIT.cpp: Changed some registers around to avoid overwriting edx:eax, 2522 which is how JSValues are now returned. Also changed the code that 2523 passes thisValue to pass the full 64bits of the value. Also added 2524 an #error compiler directive to other platform builds, since the JSValue 2525 return signature probably won't return in edx:eax on those platforms, 2526 and we'll have to investigate a solution. 2527 25282009-05-13 Geoffrey Garen <ggaren@apple.com> 2529 2530 Reviewed by Sam Weinig. 2531 2532 Removed parameters from functions that are intended never to use their 2533 parameters. 2534 2535 * jit/JITPropertyAccess.cpp: 2536 (JSC::JIT::emitSlow_op_get_by_val): 2537 (JSC::JIT::emitSlow_op_put_by_val): 2538 25392009-05-13 Geoffrey Garen <ggaren@apple.com> 2540 2541 Reviewed by Sam Weinig. 2542 2543 Ported op_instance_of from TOT. It's basically the same, but some register 2544 stuff changed to memory stuff. 2545 2546 * jit/JITInlineMethods.h: 2547 (JSC::JIT::emitPutJITStubArgFromVirtualRegister): 2548 (JSC::JIT::emitStore): Changed to use helper functions. 2549 2550 * jit/JITOpcodes.cpp: 2551 (JSC::JIT::emit_op_instanceof): 2552 (JSC::JIT::emitSlow_op_instanceof): Ported from TOT. 2553 25542009-05-13 Geoffrey Garen <ggaren@apple.com> 2555 2556 Reviewed by Gavin Barraclough. 2557 2558 Added a comment to explain an exception-handling subtelty that we found 2559 hard to remember when reviewing my last patch. 2560 2561 * jit/JITOpcodes.cpp: 2562 (JSC::JIT::emit_op_catch): 2563 25642009-05-13 Geoffrey Garen <ggaren@apple.com> 2565 2566 Reviewed by Sam Weinig. 2567 2568 Implemented try/catch. 2569 2570 * jit/JITOpcodes.cpp: 2571 (JSC::JIT::emit_op_throw): Updated to use JITStackFrame abstraction. 2572 (JSC::JIT::emit_op_catch): Filled out. 2573 25742009-05-13 Sam Weinig <sam@webkit.org> 2575 2576 Reviewed by Geoffrey Garen. 2577 2578 Implemented op_loop_if_true, op_jfalse, op_jtrue, op_jeq_null and op_jneq_null 2579 2580 * jit/JITOpcodes.cpp: 2581 (JSC::JIT::emitSlow_op_instanceof): Moved from below to be next to its 2582 fast brother. 2583 2584 (JSC::JIT::emit_op_loop_if_true): Similar to the old version 2585 in that it tries to do the integer case first and reduce the 2586 number of jumps you might need to take. 2587 (JSC::JIT::emitSlow_op_loop_if_true): 2588 2589 (JSC::JIT::emit_op_jfalse): Very similar to op_loop_if_true, only 2590 the inverse and without a timeout check. 2591 (JSC::JIT::emitSlow_op_jfalse): 2592 2593 (JSC::JIT::emit_op_jtrue): Very similar to op_loop_if_true except 2594 without the timeout check. 2595 (JSC::JIT::emitSlow_op_jtrue): 2596 2597 (JSC::JIT::emit_op_jeq_null): Very similar to the implementation 2598 of op_eq, except it takes jumps instead of copying the condition 2599 to a dst. 2600 (JSC::JIT::emit_op_jneq_null): Ditto but for op_neq. 2601 26022009-05-13 Geoffrey Garen <ggaren@apple.com> 2603 2604 Reviewed by Sam Weinig. 2605 2606 Implemented op_call_varargs. 2607 2608 * jit/JITCall.cpp: 2609 (JSC::JIT::compileOpCallVarargsSetupArgs): 2610 (JSC::JIT::compileOpCallVarargs): 2611 (JSC::JIT::emit_op_call): 2612 (JSC::JIT::emit_op_call_eval): 2613 (JSC::JIT::emit_op_load_varargs): 2614 (JSC::JIT::emit_op_call_varargs): 2615 (JSC::JIT::emit_op_construct): 2616 * jit/JITOpcodes.cpp: 2617 (JSC::JIT::emit_op_jneq_ptr): 2618 26192009-05-13 Geoffrey Garen <ggaren@apple.com> 2620 2621 Reviewed by Sam Weinig. 2622 2623 Implemented op_call_eval. 2624 2625 * jit/JITCall.cpp: 2626 (JSC::JIT::compileOpCallVarargsSetupArgs): 2627 (JSC::JIT::compileOpCall): 2628 * jit/JITStubCall.h: 2629 (JSC::CallEvalJITStub::CallEvalJITStub): 2630 26312009-05-13 Sam Weinig <sam@webkit.org> 2632 2633 Reviewed by Gavin Barraclough. 2634 2635 Implemented op_not. (Gavin did most of the work!) 2636 2637 * jit/JITOpcodes.cpp: 2638 (JSC::JIT::emit_op_not): 2639 (JSC::JIT::emitSlow_op_not): 2640 26412009-05-13 Geoffrey Garen <ggaren@apple.com> 2642 2643 Reviewed by Sam Weinig. 2644 2645 Implemented op_global_resolve. 2646 2647 * jit/JITOpcodes.cpp: 2648 (JSC::JIT::emit_op_loop_if_less): 2649 (JSC::JIT::emit_op_loop_if_lesseq): Added back accidentally removed 2650 early returns. 2651 2652 (JSC::JIT::emit_op_resolve_global): 2653 * jit/JITStubs.cpp: 2654 (JSC::JITStubs::cti_op_resolve_global): Pretty similar to the old code, 2655 but we need two reads and a TimesEight step in order to account for the 2656 64bit value size. 2657 2658 * jit/JITStubs.h: 2659 (JSC::): Slightly tweaked this code to specialize for a JSGlobalObject*, 2660 to avoid having to pass an irrelevant tag pointer to the stub. 2661 26622009-05-13 Sam Weinig <sam@webkit.org> 2663 2664 Reviewed by Geoffrey Garen. 2665 2666 Implemented op_to_jsnumber. 2667 2668 * jit/JITOpcodes.cpp: 2669 (JSC::JIT::emit_op_to_jsnumber): 2670 (JSC::JIT::emitSlow_op_to_jsnumber): 2671 26722009-05-13 Sam Weinig <sam@webkit.org> 2673 2674 Reviewed by Geoffrey Garen. 2675 2676 Implemented op_convert_this. 2677 2678 * jit/JITOpcodes.cpp: 2679 (JSC::JIT::emit_op_convert_this): 2680 (JSC::JIT::emitSlow_op_convert_this): 2681 26822009-05-13 Geoffrey Garen <ggaren@apple.com> 2683 2684 Reviewed by Sam Weinig. 2685 2686 Got basic JS function and constructor calls working. 2687 2688 * jit/JIT.cpp: 2689 (JSC::JIT::privateCompileCTIMachineTrampolines): 2690 * jit/JIT.h: 2691 * jit/JITCall.cpp: 2692 (JSC::JIT::compileOpCallSetupArgs): 2693 (JSC::JIT::compileOpCallVarargsSetupArgs): 2694 (JSC::JIT::compileOpConstructSetupArgs): 2695 (JSC::JIT::emit_op_ret): 2696 (JSC::JIT::emit_op_construct_verify): 2697 (JSC::JIT::emitSlow_op_construct_verify): 2698 (JSC::JIT::emitSlow_op_call): 2699 (JSC::JIT::emitSlow_op_call_eval): 2700 (JSC::JIT::emitSlow_op_call_varargs): 2701 (JSC::JIT::emitSlow_op_construct): 2702 (JSC::JIT::compileOpCall): Filled out these cases, with call_eval #if'd out. 2703 2704 * jit/JITInlineMethods.h: 2705 (JSC::JIT::emitPutJITStubArgFromVirtualRegister): 2706 (JSC::JIT::emitLoad): Restored some legacy "*CTIArg*" functions, 2707 since I wanted to avoid the complexity of revamping the API here while 2708 trying to bring it up. Eventually, we should re-remove all of these functions. 2709 2710 (JSC::JIT::recordJumpTarget): Removed unnecessary macro cruft. You will 2711 not silence me, Sam Weinig! The world will know that you are a crufty, 2712 crufty, crufty programmer!!! 2713 2714 * jit/JITOpcodes.cpp: 2715 * jit/JITStubs.cpp: 2716 (JSC::): 2717 * jit/JITStubs.h: Changed up some offsets in the JITStackFrame class, since 2718 and off-by-one error was causing stack misalignment. 2719 27202009-05-13 Sam Weinig <sam@webkit.org> 2721 2722 Reviewed by Geoffrey Garen. 2723 2724 Implement op_eq_null and op_neq_null. 2725 2726 * assembler/MacroAssemblerX86Common.h: 2727 (JSC::MacroAssemblerX86Common::set8): 2728 (JSC::MacroAssemblerX86Common::setTest8): 2729 * jit/JITOpcodes.cpp: 2730 (JSC::JIT::emit_op_stricteq): 2731 (JSC::JIT::emitSlow_op_stricteq): 2732 (JSC::JIT::emit_op_nstricteq): 2733 (JSC::JIT::emitSlow_op_nstricteq): 2734 (JSC::JIT::emit_op_eq_null): 2735 (JSC::JIT::emit_op_neq_null): 2736 * jsc.cpp: 2737 27382009-05-12 Sam Weinig <sam@webkit.org> 2739 2740 Reviewed by Geoffrey Garen. 2741 2742 Implement op_new_error. 2743 2744 * jit/JITOpcodes.cpp: 2745 (JSC::JIT::emit_op_new_error): 2746 * jit/JITStubCall.h: 2747 (JSC::JITStubCall::addArgument): Add a version of addArgument 2748 that takes a constant JSValue. 2749 27502009-05-12 Sam Weinig <sam@webkit.org> 2751 2752 Reviewed by Geoffrey Garen. 2753 2754 Remove now unused emitGetVariableObjectRegister and emitPutVariableObjectRegister. 2755 2756 * jit/JIT.cpp: 2757 * jit/JIT.h: 2758 27592009-05-12 Sam Weinig <sam@webkit.org> 2760 2761 Reviewed by Geoffrey Garen. 2762 2763 Implement op_to_primitive and op_next_pname. 2764 2765 * jit/JITOpcodes.cpp: 2766 (JSC::JIT::emitSlow_op_construct_verify): 2767 (JSC::JIT::emit_op_to_primitive): 2768 (JSC::JIT::emitSlow_op_to_primitive): 2769 (JSC::JIT::emitSlow_op_loop_if_true): 2770 (JSC::JIT::emit_op_jtrue): 2771 (JSC::JIT::emit_op_next_pname): 2772 27732009-05-12 Sam Weinig <sam@webkit.org> 2774 2775 Reviewed by Geoffrey Garen. 2776 2777 Add op_get_global_var, op_put_global_var, emit_op_get_scoped_var, emit_op_put_scoped_var and 2778 op_unexpected_load. 2779 2780 * jit/JIT.h: 2781 * jit/JITInlineMethods.h: 2782 (JSC::JIT::tagFor): 2783 (JSC::JIT::payloadFor): 2784 (JSC::JIT::emitLoad): 2785 (JSC::JIT::emitStore): 2786 (JSC::JIT::emitLoadReturnValue): 2787 * jit/JITOpcodes.cpp: 2788 (JSC::JIT::emit_op_get_global_var): 2789 (JSC::JIT::emit_op_put_global_var): 2790 (JSC::JIT::emit_op_get_scoped_var): 2791 (JSC::JIT::emit_op_put_scoped_var): 2792 (JSC::JIT::emit_op_unexpected_load): 2793 27942009-05-12 Geoffrey Garen <ggaren@apple.com> 2795 2796 Reviewed by Sam Weinig. 2797 2798 Added overflow handling to op_sub. 2799 2800 * jit/JIT.h: 2801 * jit/JITArithmetic.cpp: 2802 (JSC::JIT::emitSlow_op_sub): 2803 (JSC::JIT::emitSlowSub32InPlaceLeft): 2804 28052009-05-12 Sam Weinig <sam@webkit.org> 2806 2807 Reviewed by Geoffrey Garen. 2808 2809 Remove a function call by folding op_get_by_id and op_put_by_id into 2810 their respective compile functions. 2811 2812 * jit/JIT.h: 2813 * jit/JITPropertyAccess.cpp: 2814 (JSC::JIT::emit_op_get_by_id): 2815 (JSC::JIT::emitSlow_op_get_by_id): 2816 (JSC::JIT::emit_op_put_by_id): 2817 (JSC::JIT::emitSlow_op_put_by_id): 2818 28192009-05-12 Sam Weinig <sam@webkit.org> 2820 2821 Reviewed by Geoffrey Garen. 2822 2823 Make JITStubCall work in 64bit by making the stack index 2824 step dependent on the size of void*. 2825 2826 * jit/JITStubCall.h: 2827 (JSC::JITStubCall::JITStubCall): 2828 (JSC::JITStubCall::addArgument): 2829 28302009-05-12 Sam Weinig <sam@webkit.org> 2831 2832 Reviewed by Geoffrey Garen. 2833 2834 Implement simple version of property access opcodes 2835 which just call a stub functions. 2836 2837 * jit/JITOpcodes.cpp: 2838 * jit/JITPropertyAccess.cpp: 2839 (JSC::JIT::emitSlow_op_put_by_id): 2840 (JSC::JIT::emitSlow_op_get_by_id): 2841 (JSC::JIT::emit_op_get_by_val): 2842 (JSC::JIT::emitSlow_op_get_by_val): 2843 (JSC::JIT::emit_op_put_by_val): 2844 (JSC::JIT::emitSlow_op_put_by_val): 2845 (JSC::JIT::emit_op_put_by_index): 2846 (JSC::JIT::emit_op_put_getter): 2847 (JSC::JIT::emit_op_put_setter): 2848 (JSC::JIT::emit_op_del_by_id): 2849 (JSC::JIT::compileGetByIdHotPath): 2850 (JSC::JIT::compilePutByIdHotPath): 2851 * jit/JITStubCall.h: 2852 (JSC::JITStubCall::addArgument): 2853 * jsc.cpp: 2854 28552009-05-12 Geoffrey Garen <ggaren@apple.com> 2856 2857 Reviewed by Sam Weinig. 2858 2859 Added work-around for XCode debugging echo problem. 2860 2861 * jsc.cpp: 2862 (runInteractive): 2863 28642009-05-12 Geoffrey Garen <ggaren@apple.com> 2865 2866 Reviewed by Sam Weinig. 2867 2868 Added overflow handling to op_add. 2869 2870 * jit/JIT.h: 2871 * jit/JITArithmetic.cpp: 2872 (JSC::JIT::emitSlow_op_add): 2873 (JSC::JIT::emitSlowAdd32InPlace): 2874 28752009-05-12 Sam Weinig <sam@webkit.org> 2876 2877 Reviewed by Geoffrey Garen. 2878 2879 Add slow cases for op_jnless or emit_op_jnlesseq. 2880 2881 * jit/JITArithmetic.cpp: 2882 (JSC::JIT::emitSlow_op_jnless): 2883 (JSC::JIT::emitSlow_op_jnlesseq): 2884 28852009-05-12 Sam Weinig <sam@webkit.org> 2886 2887 Reviewed by Geoffrey Garen. 2888 2889 Add implementations for op_jnless, emit_op_jnlesseq, op_loop_if_less and op_loop_if_lesseq. 2890 No slow cases for op_jnless or emit_op_jnlesseq yet. 2891 2892 * jit/JITArithmetic.cpp: 2893 (JSC::JIT::emit_op_jnless): 2894 (JSC::JIT::emitSlow_op_jnless): 2895 (JSC::JIT::emit_op_jnlesseq): 2896 (JSC::JIT::emitSlow_op_jnlesseq): 2897 * jit/JITOpcodes.cpp: 2898 (JSC::JIT::emit_op_loop_if_less): 2899 (JSC::JIT::emitSlow_op_loop_if_less): 2900 (JSC::JIT::emit_op_loop_if_lesseq): 2901 (JSC::JIT::emitSlow_op_loop_if_lesseq): 2902 29032009-05-12 Sam Weinig <sam@webkit.org> 2904 2905 Reviewed by Geoffrey Garen. 2906 2907 Turn the RECORD_JUMP_TARGET macro into an inline function. 2908 2909 * jit/JIT.h: 2910 * jit/JITInlineMethods.h: 2911 (JSC::JIT::recordJumpTarget): 2912 * jit/JITOpcodes.cpp: 2913 (JSC::JIT::emit_op_jmp): 2914 (JSC::JIT::emit_op_jsr): 2915 (JSC::JIT::emit_op_jmp_scopes): 2916 29172009-05-12 Sam Weinig <sam@webkit.org> 2918 2919 Add MacroAssemblerX86Common::set8 to fix the build. 2920 2921 * assembler/MacroAssemblerX86Common.h: 2922 (JSC::MacroAssemblerX86Common::set8): 2923 29242009-05-12 Geoffrey Garen <ggaren@apple.com> 2925 2926 Reviewed by Sam Weinig. 2927 2928 Added overflow recovery for pre_inc and pre_dec. 2929 2930 Turned some short-circuit code into early returns, as is the WebKit style. 2931 2932 * jit/JITArithmetic.cpp: 2933 (JSC::JIT::emit_op_post_inc): 2934 (JSC::JIT::emitSlow_op_post_inc): 2935 (JSC::JIT::emit_op_post_dec): 2936 (JSC::JIT::emitSlow_op_post_dec): 2937 (JSC::JIT::emitSlow_op_pre_inc): 2938 (JSC::JIT::emitSlow_op_pre_dec): 2939 29402009-05-12 Sam Weinig <sam@webkit.org> 2941 2942 Reviewed by Geoffrey Garen. 2943 2944 Implement op_jmp, op_loop, op_eq and op_neq. 2945 2946 * jit/JITOpcodes.cpp: 2947 (JSC::JIT::emit_op_jmp): 2948 (JSC::JIT::emit_op_loop): 2949 (JSC::JIT::emit_op_eq): 2950 (JSC::JIT::emitSlow_op_eq): 2951 (JSC::JIT::emit_op_neq): 2952 (JSC::JIT::emitSlow_op_neq): 2953 (JSC::JIT::emit_op_enter): 2954 (JSC::JIT::emit_op_enter_with_activation): 2955 29562009-05-12 Sam Weinig <sam@webkit.org> 2957 2958 Reviewed by Geoffrey Garen. 2959 2960 Implement the slow cases for arithmetic opcodes. 2961 2962 * jit/JITArithmetic.cpp: 2963 (JSC::JIT::emitSlow_op_lshift): 2964 (JSC::JIT::emitSlow_op_rshift): 2965 (JSC::JIT::emitSlow_op_bitand): 2966 (JSC::JIT::emitSlow_op_bitor): 2967 (JSC::JIT::emitSlow_op_bitxor): 2968 (JSC::JIT::emitSlow_op_bitnot): 2969 (JSC::JIT::emitSlow_op_sub): 2970 (JSC::JIT::emitSlow_op_mul): 2971 (JSC::JIT::emitSlow_op_mod): 2972 (JSC::JIT::emit_op_mod): 2973 29742009-05-12 Sam Weinig <sam@webkit.org> 2975 2976 Reviewed by Geoffrey Garen. 2977 2978 Implement op_bitnot. 2979 2980 * assembler/MacroAssemblerX86Common.h: 2981 (JSC::MacroAssemblerX86Common::not32): 2982 * assembler/X86Assembler.h: 2983 (JSC::X86Assembler::notl_m): 2984 * jit/JITArithmetic.cpp: 2985 (JSC::JIT::emit_op_bitnot): 2986 29872009-05-12 Sam Weinig <sam@webkit.org> 2988 2989 Reviewed by Geoffrey Garen. 2990 2991 Add arithmetic opcode implementations from the old nitro-extreme branch. 2992 2993 * jit/JIT.h: 2994 * jit/JITArithmetic.cpp: 2995 (JSC::JIT::emit_op_jnless): 2996 (JSC::JIT::emitSlow_op_jnless): 2997 (JSC::JIT::emit_op_jnlesseq): 2998 (JSC::JIT::emitSlow_op_jnlesseq): 2999 (JSC::JIT::emit_op_lshift): 3000 (JSC::JIT::emitSlow_op_lshift): 3001 (JSC::JIT::emit_op_rshift): 3002 (JSC::JIT::emitSlow_op_rshift): 3003 (JSC::JIT::emit_op_bitand): 3004 (JSC::JIT::emitBitAnd32Constant): 3005 (JSC::JIT::emitBitAnd32InPlace): 3006 (JSC::JIT::emit_op_bitor): 3007 (JSC::JIT::emitSlow_op_bitor): 3008 (JSC::JIT::emitBitOr32Constant): 3009 (JSC::JIT::emitBitOr32InPlace): 3010 (JSC::JIT::emit_op_bitxor): 3011 (JSC::JIT::emitSlow_op_bitxor): 3012 (JSC::JIT::emitBitXor32Constant): 3013 (JSC::JIT::emitBitXor32InPlace): 3014 (JSC::JIT::emit_op_bitnot): 3015 (JSC::JIT::emitSlow_op_bitnot): 3016 (JSC::JIT::emit_op_post_inc): 3017 (JSC::JIT::emitSlow_op_post_inc): 3018 (JSC::JIT::emit_op_post_dec): 3019 (JSC::JIT::emitSlow_op_post_dec): 3020 (JSC::JIT::emit_op_pre_inc): 3021 (JSC::JIT::emitSlow_op_pre_inc): 3022 (JSC::JIT::emit_op_pre_dec): 3023 (JSC::JIT::emitSlow_op_pre_dec): 3024 (JSC::JIT::emit_op_add): 3025 (JSC::JIT::emitAdd32Constant): 3026 (JSC::JIT::emitAdd32InPlace): 3027 (JSC::JIT::emitSlow_op_add): 3028 (JSC::JIT::emit_op_sub): 3029 (JSC::JIT::emitSlow_op_sub): 3030 (JSC::JIT::emitSub32ConstantLeft): 3031 (JSC::JIT::emitSub32ConstantRight): 3032 (JSC::JIT::emitSub32InPlaceLeft): 3033 (JSC::JIT::emitSub32InPlaceRight): 3034 (JSC::JIT::emit_op_mul): 3035 (JSC::JIT::emitSlow_op_mul): 3036 (JSC::JIT::emitMul32Constant): 3037 (JSC::JIT::emitMul32InPlace): 3038 (JSC::JIT::emit_op_mod): 3039 (JSC::JIT::emitSlow_op_mod): 3040 * jit/JITOpcodes.cpp: 3041 30422009-05-12 Geoffrey Garen <ggaren@apple.com> 3043 3044 Removed JIT_OPTIMIZE_ARITHMETIC setting, since it was all about 32bit 3045 value representations. 3046 3047 Added JSAPIValueWrapper to the repository. 3048 3049 * jit/JIT.h: 3050 * jit/JITArithmetic.cpp: 3051 * runtime/JSAPIValueWrapper.cpp: Added. 3052 (JSC::JSAPIValueWrapper::toPrimitive): 3053 (JSC::JSAPIValueWrapper::getPrimitiveNumber): 3054 (JSC::JSAPIValueWrapper::toBoolean): 3055 (JSC::JSAPIValueWrapper::toNumber): 3056 (JSC::JSAPIValueWrapper::toString): 3057 (JSC::JSAPIValueWrapper::toObject): 3058 * runtime/JSAPIValueWrapper.h: Added. 3059 (JSC::JSAPIValueWrapper::value): 3060 (JSC::JSAPIValueWrapper::isAPIValueWrapper): 3061 (JSC::JSAPIValueWrapper::JSAPIValueWrapper): 3062 (JSC::jsAPIValueWrapper): 3063 * wtf/Platform.h: 3064 30652009-05-12 Geoffrey Garen <ggaren@apple.com> 3066 3067 Turned on the JIT and got it building and running the most trivial of 3068 programs. 3069 3070 All configurable optimizations are turned off, and a few opcodes are ad 3071 hoc #if'd out. 3072 3073 So far, I've only merged op_mov and op_end, but some stub-reliant 3074 opcodes work as-is from TOT. 3075 3076 * bytecode/CodeBlock.cpp: 3077 (JSC::CodeBlock::~CodeBlock): 3078 * bytecode/CodeBlock.h: 3079 * jit/JIT.cpp: 3080 (JSC::JIT::compileOpStrictEq): 3081 * jit/JIT.h: 3082 * jit/JITArithmetic.cpp: 3083 (JSC::JIT::emit_op_lshift): 3084 (JSC::JIT::emitSlow_op_lshift): 3085 (JSC::JIT::emit_op_rshift): 3086 (JSC::JIT::emitSlow_op_rshift): 3087 (JSC::JIT::emit_op_jnless): 3088 (JSC::JIT::emitSlow_op_jnless): 3089 (JSC::JIT::emit_op_jnlesseq): 3090 (JSC::JIT::emitSlow_op_jnlesseq): 3091 (JSC::JIT::emit_op_bitand): 3092 (JSC::JIT::emitSlow_op_bitand): 3093 (JSC::JIT::emit_op_post_inc): 3094 (JSC::JIT::emitSlow_op_post_inc): 3095 (JSC::JIT::emit_op_post_dec): 3096 (JSC::JIT::emitSlow_op_post_dec): 3097 (JSC::JIT::emit_op_pre_inc): 3098 (JSC::JIT::emitSlow_op_pre_inc): 3099 (JSC::JIT::emit_op_pre_dec): 3100 (JSC::JIT::emitSlow_op_pre_dec): 3101 (JSC::JIT::emit_op_mod): 3102 (JSC::JIT::emitSlow_op_mod): 3103 (JSC::JIT::emit_op_add): 3104 (JSC::JIT::emit_op_mul): 3105 (JSC::JIT::emit_op_sub): 3106 (JSC::JIT::compileBinaryArithOpSlowCase): 3107 (JSC::JIT::emitSlow_op_add): 3108 (JSC::JIT::emitSlow_op_mul): 3109 * jit/JITCall.cpp: 3110 (JSC::JIT::compileOpCallInitializeCallFrame): 3111 (JSC::JIT::compileOpConstructSetupArgs): 3112 (JSC::JIT::compileOpCallVarargs): 3113 (JSC::JIT::compileOpCall): 3114 (JSC::JIT::compileOpCallSlowCase): 3115 * jit/JITInlineMethods.h: 3116 (JSC::JIT::getConstantOperandImmediateInt): 3117 (JSC::JIT::isOperandConstantImmediateInt): 3118 (JSC::JIT::emitInitRegister): 3119 (JSC::JIT::addSlowCase): 3120 (JSC::JIT::addJump): 3121 (JSC::JIT::emitJumpSlowToHot): 3122 (JSC::JIT::tagFor): 3123 (JSC::JIT::payloadFor): 3124 (JSC::JIT::emitLoad): 3125 (JSC::JIT::emitLoadReturnValue): 3126 (JSC::JIT::emitStore): 3127 (JSC::JIT::emitStoreReturnValue): 3128 * jit/JITOpcodes.cpp: 3129 (JSC::JIT::emit_op_mov): 3130 (JSC::JIT::emit_op_end): 3131 (JSC::JIT::emit_op_jmp): 3132 (JSC::JIT::emit_op_loop): 3133 (JSC::JIT::emit_op_loop_if_less): 3134 (JSC::JIT::emit_op_loop_if_lesseq): 3135 (JSC::JIT::emit_op_instanceof): 3136 (JSC::JIT::emit_op_get_global_var): 3137 (JSC::JIT::emit_op_put_global_var): 3138 (JSC::JIT::emit_op_get_scoped_var): 3139 (JSC::JIT::emit_op_put_scoped_var): 3140 (JSC::JIT::emit_op_tear_off_activation): 3141 (JSC::JIT::emit_op_ret): 3142 (JSC::JIT::emit_op_construct_verify): 3143 (JSC::JIT::emit_op_to_primitive): 3144 (JSC::JIT::emit_op_loop_if_true): 3145 (JSC::JIT::emit_op_resolve_global): 3146 (JSC::JIT::emit_op_not): 3147 (JSC::JIT::emit_op_jfalse): 3148 (JSC::JIT::emit_op_jeq_null): 3149 (JSC::JIT::emit_op_jneq_null): 3150 (JSC::JIT::emit_op_jneq_ptr): 3151 (JSC::JIT::emit_op_unexpected_load): 3152 (JSC::JIT::emit_op_eq): 3153 (JSC::JIT::emit_op_bitnot): 3154 (JSC::JIT::emit_op_jtrue): 3155 (JSC::JIT::emit_op_neq): 3156 (JSC::JIT::emit_op_bitxor): 3157 (JSC::JIT::emit_op_bitor): 3158 (JSC::JIT::emit_op_throw): 3159 (JSC::JIT::emit_op_next_pname): 3160 (JSC::JIT::emit_op_push_scope): 3161 (JSC::JIT::emit_op_to_jsnumber): 3162 (JSC::JIT::emit_op_push_new_scope): 3163 (JSC::JIT::emit_op_catch): 3164 (JSC::JIT::emit_op_switch_imm): 3165 (JSC::JIT::emit_op_switch_char): 3166 (JSC::JIT::emit_op_switch_string): 3167 (JSC::JIT::emit_op_new_error): 3168 (JSC::JIT::emit_op_eq_null): 3169 (JSC::JIT::emit_op_neq_null): 3170 (JSC::JIT::emit_op_convert_this): 3171 (JSC::JIT::emit_op_profile_will_call): 3172 (JSC::JIT::emit_op_profile_did_call): 3173 (JSC::JIT::emitSlow_op_construct_verify): 3174 (JSC::JIT::emitSlow_op_get_by_val): 3175 (JSC::JIT::emitSlow_op_loop_if_less): 3176 (JSC::JIT::emitSlow_op_loop_if_lesseq): 3177 (JSC::JIT::emitSlow_op_put_by_val): 3178 (JSC::JIT::emitSlow_op_not): 3179 (JSC::JIT::emitSlow_op_instanceof): 3180 * jit/JITPropertyAccess.cpp: 3181 (JSC::JIT::emit_op_get_by_val): 3182 (JSC::JIT::emit_op_put_by_val): 3183 (JSC::JIT::emit_op_put_by_index): 3184 (JSC::JIT::emit_op_put_getter): 3185 (JSC::JIT::emit_op_put_setter): 3186 (JSC::JIT::emit_op_del_by_id): 3187 (JSC::JIT::compileGetByIdHotPath): 3188 (JSC::JIT::compilePutByIdHotPath): 3189 * jit/JITStubCall.h: 3190 (JSC::JITStubCall::JITStubCall): 3191 (JSC::JITStubCall::addArgument): 3192 (JSC::JITStubCall::call): 3193 (JSC::JITStubCall::): 3194 (JSC::CallEvalJITStub::CallEvalJITStub): 3195 * jit/JITStubs.cpp: 3196 (JSC::): 3197 (JSC::JITStubs::cti_op_add): 3198 (JSC::JITStubs::cti_op_pre_inc): 3199 (JSC::JITStubs::cti_op_mul): 3200 (JSC::JITStubs::cti_op_get_by_val): 3201 (JSC::JITStubs::cti_op_get_by_val_string): 3202 (JSC::JITStubs::cti_op_get_by_val_byte_array): 3203 (JSC::JITStubs::cti_op_sub): 3204 (JSC::JITStubs::cti_op_put_by_val): 3205 (JSC::JITStubs::cti_op_put_by_val_array): 3206 (JSC::JITStubs::cti_op_put_by_val_byte_array): 3207 (JSC::JITStubs::cti_op_negate): 3208 (JSC::JITStubs::cti_op_div): 3209 (JSC::JITStubs::cti_op_pre_dec): 3210 (JSC::JITStubs::cti_op_post_inc): 3211 (JSC::JITStubs::cti_op_eq): 3212 (JSC::JITStubs::cti_op_lshift): 3213 (JSC::JITStubs::cti_op_bitand): 3214 (JSC::JITStubs::cti_op_rshift): 3215 (JSC::JITStubs::cti_op_bitnot): 3216 (JSC::JITStubs::cti_op_mod): 3217 (JSC::JITStubs::cti_op_neq): 3218 (JSC::JITStubs::cti_op_post_dec): 3219 (JSC::JITStubs::cti_op_urshift): 3220 (JSC::JITStubs::cti_op_bitxor): 3221 (JSC::JITStubs::cti_op_bitor): 3222 (JSC::JITStubs::cti_op_switch_imm): 3223 * jit/JITStubs.h: 3224 * runtime/JSArray.cpp: 3225 (JSC::JSArray::JSArray): 3226 * runtime/JSFunction.cpp: 3227 (JSC::JSFunction::~JSFunction): 3228 * runtime/JSValue.h: 3229 (JSC::JSValue::payload): 3230 * wtf/Platform.h: 3231 32322009-05-07 Sam Weinig <sam@webkit.org> 3233 3234 Reviewed by Geoffrey Garen. 3235 3236 Add some new MacroAssembler and assembler functions that will be needed shortly. 3237 3238 * assembler/MacroAssemblerX86Common.h: 3239 (JSC::MacroAssemblerX86Common::add32): 3240 (JSC::MacroAssemblerX86Common::and32): 3241 (JSC::MacroAssemblerX86Common::mul32): 3242 (JSC::MacroAssemblerX86Common::neg32): 3243 (JSC::MacroAssemblerX86Common::or32): 3244 (JSC::MacroAssemblerX86Common::sub32): 3245 (JSC::MacroAssemblerX86Common::xor32): 3246 (JSC::MacroAssemblerX86Common::branchAdd32): 3247 (JSC::MacroAssemblerX86Common::branchMul32): 3248 (JSC::MacroAssemblerX86Common::branchSub32): 3249 * assembler/X86Assembler.h: 3250 (JSC::X86Assembler::): 3251 (JSC::X86Assembler::addl_rm): 3252 (JSC::X86Assembler::andl_mr): 3253 (JSC::X86Assembler::andl_rm): 3254 (JSC::X86Assembler::andl_im): 3255 (JSC::X86Assembler::negl_r): 3256 (JSC::X86Assembler::notl_r): 3257 (JSC::X86Assembler::orl_rm): 3258 (JSC::X86Assembler::orl_im): 3259 (JSC::X86Assembler::subl_rm): 3260 (JSC::X86Assembler::xorl_mr): 3261 (JSC::X86Assembler::xorl_rm): 3262 (JSC::X86Assembler::xorl_im): 3263 (JSC::X86Assembler::imull_mr): 3264 32652009-05-11 Sam Weinig <sam@webkit.org> 3266 3267 Reviewed by Cameron Zwarich. 3268 3269 Remove the NumberHeap. 3270 3271 * JavaScriptCore.exp: 3272 * runtime/Collector.cpp: 3273 (JSC::Heap::Heap): 3274 (JSC::Heap::destroy): 3275 (JSC::Heap::recordExtraCost): 3276 (JSC::Heap::heapAllocate): 3277 (JSC::Heap::markConservatively): 3278 (JSC::Heap::sweep): 3279 (JSC::Heap::collect): 3280 (JSC::Heap::objectCount): 3281 (JSC::Heap::statistics): 3282 (JSC::typeName): 3283 (JSC::Heap::isBusy): 3284 * runtime/Collector.h: 3285 (JSC::Heap::globalData): 3286 * runtime/JSCell.h: 3287 32882009-05-11 Geoffrey Garen <ggaren@apple.com> 3289 3290 Reviewed by Sam Weinig. 3291 3292 Land initial commit of new number representation for 32 bit platforms, 3293 with JIT disabled. 3294 3295 * API/APICast.h: 3296 (toJS): 3297 (toRef): 3298 * API/JSCallbackObjectFunctions.h: 3299 (JSC::::hasInstance): 3300 (JSC::::toNumber): 3301 (JSC::::toString): 3302 * API/tests/testapi.c: 3303 (EvilExceptionObject_convertToType): 3304 * AllInOneFile.cpp: 3305 * JavaScriptCore.exp: 3306 * JavaScriptCore.xcodeproj/project.pbxproj: 3307 * bytecode/CodeBlock.cpp: 3308 (JSC::valueToSourceString): 3309 * bytecompiler/BytecodeGenerator.cpp: 3310 (JSC::BytecodeGenerator::emitLoad): 3311 (JSC::BytecodeGenerator::emitUnexpectedLoad): 3312 (JSC::keyForImmediateSwitch): 3313 * bytecompiler/BytecodeGenerator.h: 3314 * interpreter/Interpreter.cpp: 3315 (JSC::Interpreter::dumpRegisters): 3316 (JSC::Interpreter::privateExecute): 3317 * parser/Nodes.cpp: 3318 (JSC::ArrayNode::emitBytecode): 3319 (JSC::processClauseList): 3320 * runtime/ArgList.h: 3321 * runtime/Collector.h: 3322 (JSC::sizeof): 3323 * runtime/DateMath.cpp: 3324 * runtime/ExceptionHelpers.h: 3325 * runtime/InitializeThreading.cpp: 3326 * runtime/JSArray.cpp: 3327 (JSC::JSArray::JSArray): 3328 * runtime/JSCell.cpp: 3329 * runtime/JSCell.h: 3330 (JSC::JSCell::isAPIValueWrapper): 3331 (JSC::JSValue::isString): 3332 (JSC::JSValue::isGetterSetter): 3333 (JSC::JSValue::isObject): 3334 (JSC::JSValue::getString): 3335 (JSC::JSValue::getObject): 3336 (JSC::JSValue::getCallData): 3337 (JSC::JSValue::getConstructData): 3338 (JSC::JSValue::getUInt32): 3339 (JSC::JSValue::marked): 3340 (JSC::JSValue::toPrimitive): 3341 (JSC::JSValue::getPrimitiveNumber): 3342 (JSC::JSValue::toBoolean): 3343 (JSC::JSValue::toNumber): 3344 (JSC::JSValue::toString): 3345 (JSC::JSValue::needsThisConversion): 3346 (JSC::JSValue::toThisString): 3347 (JSC::JSValue::getJSNumber): 3348 (JSC::JSValue::toObject): 3349 (JSC::JSValue::toThisObject): 3350 * runtime/JSGlobalData.cpp: 3351 (JSC::JSGlobalData::JSGlobalData): 3352 * runtime/JSGlobalData.h: 3353 * runtime/JSGlobalObject.h: 3354 (JSC::Structure::prototypeForLookup): 3355 * runtime/JSGlobalObjectFunctions.cpp: 3356 (JSC::globalFuncParseInt): 3357 * runtime/JSImmediate.h: 3358 * runtime/JSNumberCell.cpp: Removed. 3359 * runtime/JSNumberCell.h: Removed. 3360 * runtime/JSObject.h: 3361 (JSC::JSValue::get): 3362 (JSC::JSValue::put): 3363 * runtime/JSString.h: 3364 (JSC::JSValue::toThisJSString): 3365 * runtime/JSValue.cpp: 3366 (JSC::JSValue::toInteger): 3367 (JSC::JSValue::toIntegerPreserveNaN): 3368 (JSC::JSValue::toObjectSlowCase): 3369 (JSC::JSValue::toThisObjectSlowCase): 3370 (JSC::JSValue::synthesizeObject): 3371 (JSC::JSValue::synthesizePrototype): 3372 (JSC::JSValue::description): 3373 (JSC::nonInlineNaN): 3374 * runtime/JSValue.h: 3375 (JSC::JSValue::): 3376 (JSC::EncodedJSValueHashTraits::emptyValue): 3377 (JSC::jsNaN): 3378 (JSC::operator==): 3379 (JSC::operator!=): 3380 (JSC::toInt32): 3381 (JSC::toUInt32): 3382 (JSC::JSValue::encode): 3383 (JSC::JSValue::decode): 3384 (JSC::JSValue::JSValue): 3385 (JSC::JSValue::operator bool): 3386 (JSC::JSValue::operator==): 3387 (JSC::JSValue::operator!=): 3388 (JSC::JSValue::isUndefined): 3389 (JSC::JSValue::isNull): 3390 (JSC::JSValue::isUndefinedOrNull): 3391 (JSC::JSValue::isCell): 3392 (JSC::JSValue::isInt32): 3393 (JSC::JSValue::isUInt32): 3394 (JSC::JSValue::isDouble): 3395 (JSC::JSValue::isTrue): 3396 (JSC::JSValue::isFalse): 3397 (JSC::JSValue::tag): 3398 (JSC::JSValue::asInt32): 3399 (JSC::JSValue::asUInt32): 3400 (JSC::JSValue::asDouble): 3401 (JSC::JSValue::asCell): 3402 (JSC::JSValue::isNumber): 3403 (JSC::JSValue::isBoolean): 3404 (JSC::JSValue::getBoolean): 3405 (JSC::JSValue::uncheckedGetNumber): 3406 (JSC::JSValue::toJSNumber): 3407 (JSC::JSValue::getNumber): 3408 (JSC::JSValue::toInt32): 3409 (JSC::JSValue::toUInt32): 3410 * runtime/Operations.h: 3411 (JSC::JSValue::equal): 3412 (JSC::JSValue::equalSlowCaseInline): 3413 (JSC::JSValue::strictEqual): 3414 (JSC::JSValue::strictEqualSlowCaseInline): 3415 (JSC::jsLess): 3416 (JSC::jsLessEq): 3417 (JSC::jsAdd): 3418 * runtime/PropertySlot.h: 3419 * runtime/StringPrototype.cpp: 3420 (JSC::stringProtoFuncCharAt): 3421 (JSC::stringProtoFuncCharCodeAt): 3422 (JSC::stringProtoFuncIndexOf): 3423 * wtf/Platform.h: 3424 3425=== Start merge of nitro-extreme branch 2009-07-30 === 3426 34272009-07-29 Laszlo Gombos <laszlo.1.gombos@nokia.com> 3428 3429 Reviewed by George Staikos. 3430 3431 Resolve class/struct mixup in forward declarations 3432 https://bugs.webkit.org/show_bug.cgi?id=27708 3433 3434 * API/JSClassRef.h: 3435 * bytecode/SamplingTool.h: 3436 * interpreter/Interpreter.h: 3437 * jit/JIT.h: 3438 * profiler/ProfileGenerator.h: 3439 * profiler/Profiler.h: 3440 * runtime/ClassInfo.h: 3441 * runtime/ExceptionHelpers.h: 3442 * runtime/JSByteArray.h: 3443 * runtime/JSCell.h: 3444 * runtime/JSFunction.h: 3445 * runtime/JSGlobalData.h: 3446 * runtime/JSObject.h: 3447 * runtime/JSString.h: 3448 34492009-07-28 Ada Chan <adachan@apple.com> 3450 3451 Reviewed by Darin Adler. 3452 3453 https://bugs.webkit.org/show_bug.cgi?id=27236 3454 - Implement TCMalloc_SystemRelease and TCMalloc_SystemCommit for Windows. 3455 - Use a background thread to periodically scavenge memory to release back to the system. 3456 3457 * wtf/FastMalloc.cpp: 3458 (WTF::TCMalloc_PageHeap::init): 3459 (WTF::TCMalloc_PageHeap::runScavengerThread): 3460 (WTF::TCMalloc_PageHeap::scavenge): 3461 (WTF::TCMalloc_PageHeap::shouldContinueScavenging): 3462 (WTF::TCMalloc_PageHeap::New): 3463 (WTF::TCMalloc_PageHeap::AllocLarge): 3464 (WTF::TCMalloc_PageHeap::Delete): 3465 (WTF::TCMalloc_PageHeap::GrowHeap): 3466 (WTF::sleep): 3467 (WTF::TCMalloc_PageHeap::scavengerThread): 3468 * wtf/TCSystemAlloc.cpp: 3469 (TCMalloc_SystemRelease): 3470 (TCMalloc_SystemCommit): 3471 * wtf/TCSystemAlloc.h: 3472 34732009-07-28 Xan Lopez <xlopez@igalia.com> 3474 3475 Add new files, fixes distcheck. 3476 3477 * GNUmakefile.am: 3478 34792009-07-28 Csaba Osztrogonac <oszi@inf.u-szeged.hu> 3480 3481 Reviewed by Simon Hausmann. 3482 3483 [Qt] Determining whether to use JIT or interpreter 3484 moved from JavaScriptCore.pri to Platform.h 3485 3486 * JavaScriptCore.pri: 3487 * wtf/Platform.h: 3488 34892009-07-27 Brian Weinstein <bweinstein@apple.com> 3490 3491 Fix of misuse of sort command. 3492 3493 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 3494 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 3495 34962009-07-27 Brian Weinstein <bweinstein@apple.com> 3497 3498 Build fix for Windows. 3499 3500 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 3501 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 3502 35032009-07-27 Gavin Barraclough <barraclough@apple.com> 3504 3505 Rubber stamped by Oliver Hunt. 3506 3507 Fix tyop in JIT, renamed preverveReturnAddressAfterCall -> preserveReturnAddressAfterCall. 3508 3509 * jit/JIT.cpp: 3510 (JSC::JIT::privateCompile): 3511 (JSC::JIT::privateCompileCTIMachineTrampolines): 3512 * jit/JIT.h: 3513 * jit/JITInlineMethods.h: 3514 (JSC::JIT::preserveReturnAddressAfterCall): 3515 * jit/JITPropertyAccess.cpp: 3516 (JSC::JIT::privateCompilePutByIdTransition): 3517 35182009-07-27 Alexey Proskuryakov <ap@webkit.org> 3519 3520 Gtk build fix. 3521 3522 * runtime/JSLock.cpp: (JSC::JSLock::JSLock): Fix "no threading" case. 3523 35242009-07-27 Alexey Proskuryakov <ap@webkit.org> 3525 3526 Release build fix. 3527 3528 * runtime/JSLock.h: (JSC::JSLock::~JSLock): 3529 35302009-07-27 Alexey Proskuryakov <ap@webkit.org> 3531 3532 Reviewed by Darin Adler. 3533 3534 https://bugs.webkit.org/show_bug.cgi?id=27735 3535 Give a helpful name to JSLock constructor argument 3536 3537 * API/JSBase.cpp: 3538 (JSGarbageCollect): 3539 * API/JSContextRef.cpp: 3540 * API/JSObjectRef.cpp: 3541 (JSPropertyNameArrayRelease): 3542 (JSPropertyNameAccumulatorAddName): 3543 * JavaScriptCore.exp: 3544 * jsc.cpp: 3545 (functionGC): 3546 (cleanupGlobalData): 3547 (jscmain): 3548 * runtime/Collector.cpp: 3549 (JSC::Heap::destroy): 3550 * runtime/JSLock.cpp: 3551 (JSC::JSLock::JSLock): 3552 (JSC::JSLock::lock): 3553 (JSC::JSLock::unlock): 3554 (JSC::JSLock::DropAllLocks::DropAllLocks): 3555 (JSC::JSLock::DropAllLocks::~DropAllLocks): 3556 * runtime/JSLock.h: 3557 (JSC::): 3558 (JSC::JSLock::JSLock): 3559 (JSC::JSLock::~JSLock): 3560 35612009-07-25 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 3562 3563 Reviewed by Eric Seidel. 3564 3565 Allow custom memory allocation control for OpaqueJSPropertyNameArray struct 3566 https://bugs.webkit.org/show_bug.cgi?id=27342 3567 3568 Inherits OpaqueJSPropertyNameArray struct from FastAllocBase because it has been 3569 instantiated by 'new' JavaScriptCore/API/JSObjectRef.cpp:473. 3570 3571 * API/JSObjectRef.cpp: 3572 35732009-07-24 Ada Chan <adachan@apple.com> 3574 3575 In preparation for https://bugs.webkit.org/show_bug.cgi?id=27236: 3576 Remove TCMALLOC_TRACK_DECOMMITED_SPANS. We'll always track decommitted spans. 3577 We have tested this and show it has little impact on performance. 3578 3579 Reviewed by Mark Rowe. 3580 3581 * wtf/FastMalloc.cpp: 3582 (WTF::TCMalloc_PageHeap::New): 3583 (WTF::TCMalloc_PageHeap::AllocLarge): 3584 (WTF::propagateDecommittedState): 3585 (WTF::mergeDecommittedStates): 3586 (WTF::TCMalloc_PageHeap::Delete): 3587 (WTF::TCMalloc_PageHeap::IncrementalScavenge): 3588 35892009-07-24 Csaba Osztrogonac <oszi@inf.u-szeged.hu> 3590 3591 Reviewed by Darin Adler and Adam Barth. 3592 3593 Build fix for x86 platforms. 3594 https://bugs.webkit.org/show_bug.cgi?id=27602 3595 3596 * jit/JIT.cpp: 3597 35982009-07-23 Kevin Ollivier <kevino@theolliviers.com> 3599 3600 wx build fix, adding missing header. 3601 3602 * jit/JIT.cpp: 3603 36042009-07-22 Yong Li <yong.li@torchmobile.com> 3605 3606 Reviewed by George Staikos. 3607 3608 Add wince specific memory files into wtf/wince 3609 https://bugs.webkit.org/show_bug.cgi?id=27550 3610 3611 * wtf/wince/FastMallocWince.h: Added. 3612 * wtf/wince/MemoryManager.cpp: Added. 3613 * wtf/wince/MemoryManager.h: Added. 3614 36152009-07-23 Norbert Leser <norbert.leser@nokia.com> 3616 3617 Reviewed by Simon Hausmann. 3618 3619 Fix for missing mmap features in Symbian 3620 https://bugs.webkit.org/show_bug.cgi?id=24540 3621 3622 Fix, conditionally for PLATFORM(SYMBIAN), as an alternative 3623 to missing support for the MAP_ANON property flag in mmap. 3624 It utilizes Symbian specific memory allocation features. 3625 3626 * runtime/Collector.cpp 3627 36282009-07-22 Gavin Barraclough <barraclough@apple.com> 3629 3630 Reviewed by Sam Weinig. 3631 3632 With ENABLE(ASSEMBLER_WX_EXCLUSIVE), only change permissions once per repatch event. 3633 ( https://bugs.webkit.org/show_bug.cgi?id=27564 ) 3634 3635 Currently we change permissions forwards and backwards for each instruction modified, 3636 instead we should only change permissions once per complete repatching event. 3637 3638 2.5% progression running with ENABLE(ASSEMBLER_WX_EXCLUSIVE) enabled, 3639 which recoups 1/3 of the penalty of running with this mode enabled. 3640 3641 * assembler/ARMAssembler.cpp: 3642 (JSC::ARMAssembler::linkBranch): 3643 - Replace usage of MakeWritable with cacheFlush. 3644 3645 * assembler/ARMAssembler.h: 3646 (JSC::ARMAssembler::patchPointerInternal): 3647 (JSC::ARMAssembler::repatchLoadPtrToLEA): 3648 - Replace usage of MakeWritable with cacheFlush. 3649 3650 * assembler/ARMv7Assembler.h: 3651 (JSC::ARMv7Assembler::relinkJump): 3652 (JSC::ARMv7Assembler::relinkCall): 3653 (JSC::ARMv7Assembler::repatchInt32): 3654 (JSC::ARMv7Assembler::repatchPointer): 3655 (JSC::ARMv7Assembler::repatchLoadPtrToLEA): 3656 (JSC::ARMv7Assembler::setInt32): 3657 - Replace usage of MakeWritable with cacheFlush. 3658 3659 * assembler/LinkBuffer.h: 3660 (JSC::LinkBuffer::performFinalization): 3661 - Make explicit call to cacheFlush. 3662 3663 * assembler/MacroAssemblerCodeRef.h: 3664 (JSC::MacroAssemblerCodeRef::MacroAssemblerCodeRef): 3665 - Make size always available. 3666 3667 * assembler/RepatchBuffer.h: 3668 (JSC::RepatchBuffer::RepatchBuffer): 3669 (JSC::RepatchBuffer::~RepatchBuffer): 3670 - Add calls to MakeWritable & makeExecutable. 3671 3672 * assembler/X86Assembler.h: 3673 (JSC::X86Assembler::relinkJump): 3674 (JSC::X86Assembler::relinkCall): 3675 (JSC::X86Assembler::repatchInt32): 3676 (JSC::X86Assembler::repatchPointer): 3677 (JSC::X86Assembler::repatchLoadPtrToLEA): 3678 - Remove usage of MakeWritable. 3679 3680 * bytecode/CodeBlock.h: 3681 (JSC::CodeBlock::getJITCode): 3682 - Provide access to CodeBlock's JITCode. 3683 3684 * jit/ExecutableAllocator.h: 3685 (JSC::ExecutableAllocator::makeExecutable): 3686 (JSC::ExecutableAllocator::cacheFlush): 3687 - Remove MakeWritable, make cacheFlush public. 3688 3689 * jit/JIT.cpp: 3690 (JSC::ctiPatchNearCallByReturnAddress): 3691 (JSC::ctiPatchCallByReturnAddress): 3692 (JSC::JIT::privateCompile): 3693 (JSC::JIT::unlinkCall): 3694 (JSC::JIT::linkCall): 3695 - Add CodeBlock argument to RepatchBuffer. 3696 3697 * jit/JIT.h: 3698 - Pass CodeBlock argument for use by RepatchBuffer. 3699 3700 * jit/JITCode.h: 3701 (JSC::JITCode::start): 3702 (JSC::JITCode::size): 3703 - Provide access to code start & size. 3704 3705 * jit/JITPropertyAccess.cpp: 3706 (JSC::JIT::privateCompilePutByIdTransition): 3707 (JSC::JIT::patchGetByIdSelf): 3708 (JSC::JIT::patchMethodCallProto): 3709 (JSC::JIT::patchPutByIdReplace): 3710 (JSC::JIT::privateCompilePatchGetArrayLength): 3711 (JSC::JIT::privateCompileGetByIdProto): 3712 (JSC::JIT::privateCompileGetByIdSelfList): 3713 (JSC::JIT::privateCompileGetByIdProtoList): 3714 (JSC::JIT::privateCompileGetByIdChainList): 3715 (JSC::JIT::privateCompileGetByIdChain): 3716 - Add CodeBlock argument to RepatchBuffer. 3717 3718 * jit/JITStubs.cpp: 3719 (JSC::JITThunks::tryCachePutByID): 3720 (JSC::JITThunks::tryCacheGetByID): 3721 (JSC::JITStubs::DEFINE_STUB_FUNCTION): 3722 - Pass CodeBlock argument for use by RepatchBuffer. 3723 37242009-07-21 Zoltan Herczeg <zherczeg@inf.u-szeged.hu> 3725 3726 Reviewed by Gavin Barraclough. 3727 3728 Cache not only the structure of the method, but the 3729 structure of its prototype as well. 3730 https://bugs.webkit.org/show_bug.cgi?id=27077 3731 3732 * bytecode/CodeBlock.cpp: 3733 (JSC::CodeBlock::~CodeBlock): 3734 * bytecode/CodeBlock.h: 3735 (JSC::MethodCallLinkInfo::MethodCallLinkInfo): 3736 * jit/JITPropertyAccess.cpp: 3737 (JSC::JIT::patchMethodCallProto): 3738 37392009-07-21 Gavin Barraclough <barraclough@apple.com> 3740 3741 Reviewed by Sam Weinig. 3742 3743 Move call linking / repatching down from AbstractMacroAssembler into MacroAssemblerARCH classes. 3744 ( https://bugs.webkit.org/show_bug.cgi?id=27527 ) 3745 3746 This allows the implementation to be defined per architecture. Specifically this addresses the 3747 fact that x86-64 MacroAssembler implements far calls as a load to register, followed by a call 3748 to register. Patching the call actually requires the pointer load to be patched, rather than 3749 the call to be patched. This is implementation detail specific to MacroAssemblerX86_64, and as 3750 such is best handled there. 3751 3752 * assembler/AbstractMacroAssembler.h: 3753 * assembler/MacroAssemblerARM.h: 3754 (JSC::MacroAssemblerARM::linkCall): 3755 (JSC::MacroAssemblerARM::repatchCall): 3756 * assembler/MacroAssemblerARMv7.h: 3757 (JSC::MacroAssemblerARMv7::linkCall): 3758 (JSC::MacroAssemblerARMv7::repatchCall): 3759 * assembler/MacroAssemblerX86.h: 3760 (JSC::MacroAssemblerX86::linkCall): 3761 (JSC::MacroAssemblerX86::repatchCall): 3762 * assembler/MacroAssemblerX86_64.h: 3763 (JSC::MacroAssemblerX86_64::linkCall): 3764 (JSC::MacroAssemblerX86_64::repatchCall): 3765 37662009-07-21 Adam Treat <adam.treat@torchmobile.com> 3767 3768 Reviewed by George Staikos. 3769 3770 Every wtf file includes other wtf files with <> style includes 3771 except this one. Fix the exception. 3772 3773 * wtf/ByteArray.h: 3774 37752009-07-21 Gavin Barraclough <barraclough@apple.com> 3776 3777 Reviewed by Oliver Hunt. 3778 3779 Move LinkBuffer/RepatchBuffer out of AbstractMacroAssembler. 3780 ( https://bugs.webkit.org/show_bug.cgi?id=27485 ) 3781 3782 This change is the first step in a process to move code that should be in 3783 the architecture-specific MacroAssembler classes up out of Assmbler and 3784 AbstractMacroAssembler. 3785 3786 * JavaScriptCore.xcodeproj/project.pbxproj: 3787 - added new files 3788 3789 * assembler/ARMAssembler.h: 3790 (JSC::ARMAssembler::linkPointer): 3791 - rename patchPointer to bring it in line with the current link/repatch naming scheme 3792 3793 * assembler/ARMv7Assembler.h: 3794 (JSC::ARMv7Assembler::linkCall): 3795 (JSC::ARMv7Assembler::linkPointer): 3796 (JSC::ARMv7Assembler::relinkCall): 3797 (JSC::ARMv7Assembler::repatchInt32): 3798 (JSC::ARMv7Assembler::repatchPointer): 3799 (JSC::ARMv7Assembler::setInt32): 3800 (JSC::ARMv7Assembler::setPointer): 3801 - rename patchPointer to bring it in line with the current link/repatch naming scheme 3802 3803 * assembler/AbstractMacroAssembler.h: 3804 (JSC::AbstractMacroAssembler::linkJump): 3805 (JSC::AbstractMacroAssembler::linkCall): 3806 (JSC::AbstractMacroAssembler::linkPointer): 3807 (JSC::AbstractMacroAssembler::getLinkerAddress): 3808 (JSC::AbstractMacroAssembler::getLinkerCallReturnOffset): 3809 (JSC::AbstractMacroAssembler::repatchJump): 3810 (JSC::AbstractMacroAssembler::repatchCall): 3811 (JSC::AbstractMacroAssembler::repatchNearCall): 3812 (JSC::AbstractMacroAssembler::repatchInt32): 3813 (JSC::AbstractMacroAssembler::repatchPointer): 3814 (JSC::AbstractMacroAssembler::repatchLoadPtrToLEA): 3815 - remove the LinkBuffer/RepatchBuffer classes, but leave a set of (private, friended) methods to interface to the Assembler 3816 3817 * assembler/LinkBuffer.h: Added. 3818 (JSC::LinkBuffer::LinkBuffer): 3819 (JSC::LinkBuffer::~LinkBuffer): 3820 (JSC::LinkBuffer::link): 3821 (JSC::LinkBuffer::patch): 3822 (JSC::LinkBuffer::locationOf): 3823 (JSC::LinkBuffer::locationOfNearCall): 3824 (JSC::LinkBuffer::returnAddressOffset): 3825 (JSC::LinkBuffer::finalizeCode): 3826 (JSC::LinkBuffer::finalizeCodeAddendum): 3827 (JSC::LinkBuffer::code): 3828 (JSC::LinkBuffer::performFinalization): 3829 - new file containing the LinkBuffer class, previously a member of AbstractMacroAssembler 3830 3831 * assembler/RepatchBuffer.h: Added. 3832 (JSC::RepatchBuffer::RepatchBuffer): 3833 (JSC::RepatchBuffer::relink): 3834 (JSC::RepatchBuffer::repatch): 3835 (JSC::RepatchBuffer::repatchLoadPtrToLEA): 3836 (JSC::RepatchBuffer::relinkCallerToTrampoline): 3837 (JSC::RepatchBuffer::relinkCallerToFunction): 3838 (JSC::RepatchBuffer::relinkNearCallerToTrampoline): 3839 - new file containing the RepatchBuffer class, previously a member of AbstractMacroAssembler 3840 3841 * assembler/X86Assembler.h: 3842 (JSC::X86Assembler::linkJump): 3843 (JSC::X86Assembler::linkCall): 3844 (JSC::X86Assembler::linkPointerForCall): 3845 (JSC::X86Assembler::linkPointer): 3846 (JSC::X86Assembler::relinkJump): 3847 (JSC::X86Assembler::relinkCall): 3848 (JSC::X86Assembler::repatchInt32): 3849 (JSC::X86Assembler::repatchPointer): 3850 (JSC::X86Assembler::setPointer): 3851 (JSC::X86Assembler::setInt32): 3852 (JSC::X86Assembler::setRel32): 3853 - rename patchPointer to bring it in line with the current link/repatch naming scheme 3854 3855 * jit/JIT.cpp: 3856 (JSC::ctiPatchNearCallByReturnAddress): 3857 (JSC::ctiPatchCallByReturnAddress): 3858 - include new headers 3859 - remove MacroAssembler:: specification from RepatchBuffer usage 3860 3861 * jit/JITPropertyAccess.cpp: 3862 * yarr/RegexJIT.cpp: 3863 - include new headers 3864 38652009-07-21 Robert Agoston <Agoston.Robert@stud.u-szeged.hu> 3866 3867 Reviewed by David Levin. 3868 3869 Fixed #undef typo. 3870 https://bugs.webkit.org/show_bug.cgi?id=27506 3871 3872 * bytecode/Opcode.h: 3873 38742009-07-21 Adam Roben <aroben@apple.com> 3875 3876 Roll out r46153, r46154, and r46155 3877 3878 These changes were causing build failures and assertion failures on 3879 Windows. 3880 3881 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 3882 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 3883 * JavaScriptCore.xcodeproj/project.pbxproj: 3884 * runtime/JSArray.cpp: 3885 * runtime/StringPrototype.cpp: 3886 * runtime/UString.cpp: 3887 * runtime/UString.h: 3888 * wtf/FastMalloc.cpp: 3889 * wtf/FastMalloc.h: 3890 * wtf/Platform.h: 3891 * wtf/PossiblyNull.h: Removed. 3892 38932009-07-21 Roland Steiner <rolandsteiner@google.com> 3894 3895 Reviewed by David Levin. 3896 3897 Add ENABLE_RUBY to list of build options 3898 https://bugs.webkit.org/show_bug.cgi?id=27324 3899 3900 * Configurations/FeatureDefines.xcconfig: Added flag ENABLE_RUBY. 3901 39022009-07-20 Oliver Hunt <oliver@apple.com> 3903 3904 Reviewed by NOBODY (Build fix). 3905 3906 Build fix attempt #2 3907 3908 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 3909 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 3910 39112009-07-20 Oliver Hunt <oliver@apple.com> 3912 3913 Reviewed by NOBODY (Build fix). 3914 3915 Build fix attempt #1 3916 3917 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 3918 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 3919 39202009-07-20 Oliver Hunt <oliver@apple.com> 3921 3922 Reviewed by Gavin Barraclough. 3923 3924 Make it harder to misuse try* allocation routines 3925 https://bugs.webkit.org/show_bug.cgi?id=27469 3926 3927 Jump through a few hoops to make it much harder to accidentally 3928 miss null-checking of values returned by the try-* allocation 3929 routines. 3930 3931 * JavaScriptCore.xcodeproj/project.pbxproj: 3932 * runtime/JSArray.cpp: 3933 (JSC::JSArray::putSlowCase): 3934 (JSC::JSArray::increaseVectorLength): 3935 * runtime/StringPrototype.cpp: 3936 (JSC::stringProtoFuncFontsize): 3937 (JSC::stringProtoFuncLink): 3938 * runtime/UString.cpp: 3939 (JSC::allocChars): 3940 (JSC::reallocChars): 3941 (JSC::expandCapacity): 3942 (JSC::UString::Rep::reserveCapacity): 3943 (JSC::UString::expandPreCapacity): 3944 (JSC::createRep): 3945 (JSC::concatenate): 3946 (JSC::UString::spliceSubstringsWithSeparators): 3947 (JSC::UString::replaceRange): 3948 (JSC::UString::append): 3949 (JSC::UString::operator=): 3950 * runtime/UString.h: 3951 (JSC::UString::Rep::createEmptyBuffer): 3952 * wtf/FastMalloc.cpp: 3953 (WTF::tryFastZeroedMalloc): 3954 (WTF::tryFastMalloc): 3955 (WTF::tryFastCalloc): 3956 (WTF::tryFastRealloc): 3957 (WTF::TCMallocStats::tryFastMalloc): 3958 (WTF::TCMallocStats::tryFastCalloc): 3959 (WTF::TCMallocStats::tryFastRealloc): 3960 * wtf/FastMalloc.h: 3961 (WTF::TryMallocReturnValue::TryMallocReturnValue): 3962 (WTF::TryMallocReturnValue::~TryMallocReturnValue): 3963 (WTF::TryMallocReturnValue::operator Maybe<T>): 3964 (WTF::TryMallocReturnValue::getValue): 3965 * wtf/PossiblyNull.h: 3966 (WTF::PossiblyNull::PossiblyNull): 3967 (WTF::PossiblyNull::~PossiblyNull): 3968 (WTF::PossiblyNull::getValue): 3969 * wtf/Platform.h: 3970 39712009-07-20 Gavin Barraclough <barraclough@apple.com> 3972 3973 RS Oliver Hunt. 3974 3975 Add ARM assembler files to xcodeproj, for convenience editing. 3976 3977 * JavaScriptCore.xcodeproj/project.pbxproj: 3978 39792009-07-20 Jessie Berlin <jberlin@apple.com> 3980 3981 Reviewed by David Levin. 3982 3983 Fix an incorrect assertion in Vector::remove. 3984 3985 https://bugs.webkit.org/show_bug.cgi?id=27477 3986 3987 * wtf/Vector.h: 3988 (WTF::::remove): 3989 Assert that the position at which to start removing elements + the 3990 length (the number of elements to remove) is less than or equal to the 3991 size of the entire Vector. 3992 39932009-07-20 Peter Kasting <pkasting@google.com> 3994 3995 Reviewed by Mark Rowe. 3996 3997 https://bugs.webkit.org/show_bug.cgi?id=27468 3998 Back out r46060, which caused problems for some Apple developers. 3999 4000 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCoreCommon.vsprops: 4001 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCoreGenerated.vcproj: 4002 * JavaScriptCore.vcproj/WTF/WTFCommon.vsprops: 4003 * JavaScriptCore.vcproj/jsc/jscCommon.vsprops: 4004 * JavaScriptCore.vcproj/testapi/testapiCommon.vsprops: 4005 40062009-07-20 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 4007 4008 Reviewed by Oliver Hunt. 4009 4010 Allow custom memory allocation control in NewThreadContext 4011 https://bugs.webkit.org/show_bug.cgi?id=27338 4012 4013 Inherits NewThreadContext struct from FastAllocBase because it 4014 has been instantiated by 'new' JavaScriptCore/wtf/Threading.cpp:76. 4015 4016 * wtf/Threading.cpp: 4017 40182009-07-20 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 4019 4020 Reviewed by Oliver Hunt. 4021 4022 Allow custom memory allocation control in JavaScriptCore's JSClassRef.h 4023 https://bugs.webkit.org/show_bug.cgi?id=27340 4024 4025 Inherit StaticValueEntry and StaticFunctionEntry struct from FastAllocBase because these 4026 have been instantiated by 'new' in JavaScriptCore/API/JSClassRef.cpp:153 4027 and in JavaScriptCore/API/JSClassRef.cpp:166. 4028 4029 * API/JSClassRef.h: 4030 40312009-07-20 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 4032 4033 Reviewed by Darin Adler. 4034 4035 Allow custom memory allocation control in JavaScriptCore's RegexPattern.h 4036 https://bugs.webkit.org/show_bug.cgi?id=27343 4037 4038 Inherits RegexPattern.h's structs (which have been instantiated by operator new) from FastAllocBase: 4039 4040 CharacterClass (new call: JavaScriptCore/yarr/RegexCompiler.cpp:144) 4041 PatternAlternative (new call: JavaScriptCore/yarr/RegexPattern.h:221) 4042 PatternDisjunction (new call: JavaScriptCore/yarr/RegexCompiler.cpp:446) 4043 4044 * yarr/RegexPattern.h: 4045 40462009-07-20 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 4047 4048 Reviewed by Darin Adler. 4049 4050 Allow custom memory allocation control for JavaScriptCore's MatchFrame struct 4051 https://bugs.webkit.org/show_bug.cgi?id=27344 4052 4053 Inherits MatchFrame struct from FastAllocBase because it has 4054 been instantiated by 'new' JavaScriptCore/pcre/pcre_exec.cpp:359. 4055 4056 * pcre/pcre_exec.cpp: 4057 40582009-07-20 Laszlo Gombos <laszlo.1.gombos@nokia.com> 4059 4060 Reviewed by Holger Freyther. 4061 4062 Remove some outdated S60 platform specific code 4063 https://bugs.webkit.org/show_bug.cgi?id=27423 4064 4065 * wtf/Platform.h: 4066 40672009-07-20 Csaba Osztrogonac <oszi@inf.u-szeged.hu> 4068 4069 Reviewed by Simon Hausmann. 4070 4071 Qt build fix with MSVC and MinGW. 4072 4073 * jsc.pro: Make sure jsc is a console application, and turn off 4074 exceptions and stl support to fix the build. 4075 40762009-07-20 Xan Lopez <xlopez@igalia.com> 4077 4078 Reviewed by Gustavo Noronha. 4079 4080 Do not use C++-style comments in preprocessor directives. 4081 4082 GCC does not like this in some configurations, using C-style 4083 comments is safer. 4084 4085 * wtf/Platform.h: 4086 40872009-07-17 Peter Kasting <pkasting@google.com> 4088 4089 Reviewed by Steve Falkenburg. 4090 4091 https://bugs.webkit.org/show_bug.cgi?id=27323 4092 Only add Cygwin to the path when it isn't already there. This avoids 4093 causing problems for people who purposefully have non-Cygwin versions of 4094 executables like svn in front of the Cygwin ones in their paths. 4095 4096 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCoreCommon.vsprops: 4097 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCoreGenerated.vcproj: 4098 * JavaScriptCore.vcproj/WTF/WTFCommon.vsprops: 4099 * JavaScriptCore.vcproj/jsc/jscCommon.vsprops: 4100 * JavaScriptCore.vcproj/testapi/testapiCommon.vsprops: 4101 41022009-07-17 Gabor Loki <loki@inf.u-szeged.hu> 4103 4104 Reviewed by Gavin Barraclough. 4105 4106 Add YARR support for generic ARM platforms (disabled by default). 4107 https://bugs.webkit.org/show_bug.cgi?id=24986 4108 4109 Add generic ARM port for MacroAssembler. It supports the whole 4110 MacroAssembler functionality except floating point. 4111 4112 The class JmpSrc is extended with a flag which enables to patch 4113 the jump destination offset during execution. This feature is 4114 required for generic ARM port. 4115 4116 Signed off by Zoltan Herczeg <zherczeg@inf.u-szeged.hu> 4117 Signed off by Gabor Loki <loki@inf.u-szeged.hu> 4118 4119 * JavaScriptCore.pri: 4120 * assembler/ARMAssembler.cpp: Added. 4121 (JSC::ARMAssembler::getLdrImmAddress): 4122 (JSC::ARMAssembler::linkBranch): 4123 (JSC::ARMAssembler::patchConstantPoolLoad): 4124 (JSC::ARMAssembler::getOp2): 4125 (JSC::ARMAssembler::genInt): 4126 (JSC::ARMAssembler::getImm): 4127 (JSC::ARMAssembler::moveImm): 4128 (JSC::ARMAssembler::dataTransfer32): 4129 (JSC::ARMAssembler::baseIndexTransfer32): 4130 (JSC::ARMAssembler::executableCopy): 4131 * assembler/ARMAssembler.h: Added. 4132 (JSC::ARM::): 4133 (JSC::ARMAssembler::ARMAssembler): 4134 (JSC::ARMAssembler::): 4135 (JSC::ARMAssembler::JmpSrc::JmpSrc): 4136 (JSC::ARMAssembler::JmpSrc::enableLatePatch): 4137 (JSC::ARMAssembler::JmpDst::JmpDst): 4138 (JSC::ARMAssembler::JmpDst::isUsed): 4139 (JSC::ARMAssembler::JmpDst::used): 4140 (JSC::ARMAssembler::emitInst): 4141 (JSC::ARMAssembler::and_r): 4142 (JSC::ARMAssembler::ands_r): 4143 (JSC::ARMAssembler::eor_r): 4144 (JSC::ARMAssembler::eors_r): 4145 (JSC::ARMAssembler::sub_r): 4146 (JSC::ARMAssembler::subs_r): 4147 (JSC::ARMAssembler::rsb_r): 4148 (JSC::ARMAssembler::rsbs_r): 4149 (JSC::ARMAssembler::add_r): 4150 (JSC::ARMAssembler::adds_r): 4151 (JSC::ARMAssembler::adc_r): 4152 (JSC::ARMAssembler::adcs_r): 4153 (JSC::ARMAssembler::sbc_r): 4154 (JSC::ARMAssembler::sbcs_r): 4155 (JSC::ARMAssembler::rsc_r): 4156 (JSC::ARMAssembler::rscs_r): 4157 (JSC::ARMAssembler::tst_r): 4158 (JSC::ARMAssembler::teq_r): 4159 (JSC::ARMAssembler::cmp_r): 4160 (JSC::ARMAssembler::orr_r): 4161 (JSC::ARMAssembler::orrs_r): 4162 (JSC::ARMAssembler::mov_r): 4163 (JSC::ARMAssembler::movs_r): 4164 (JSC::ARMAssembler::bic_r): 4165 (JSC::ARMAssembler::bics_r): 4166 (JSC::ARMAssembler::mvn_r): 4167 (JSC::ARMAssembler::mvns_r): 4168 (JSC::ARMAssembler::mul_r): 4169 (JSC::ARMAssembler::muls_r): 4170 (JSC::ARMAssembler::mull_r): 4171 (JSC::ARMAssembler::ldr_imm): 4172 (JSC::ARMAssembler::ldr_un_imm): 4173 (JSC::ARMAssembler::dtr_u): 4174 (JSC::ARMAssembler::dtr_ur): 4175 (JSC::ARMAssembler::dtr_d): 4176 (JSC::ARMAssembler::dtr_dr): 4177 (JSC::ARMAssembler::ldrh_r): 4178 (JSC::ARMAssembler::ldrh_d): 4179 (JSC::ARMAssembler::ldrh_u): 4180 (JSC::ARMAssembler::strh_r): 4181 (JSC::ARMAssembler::push_r): 4182 (JSC::ARMAssembler::pop_r): 4183 (JSC::ARMAssembler::poke_r): 4184 (JSC::ARMAssembler::peek_r): 4185 (JSC::ARMAssembler::clz_r): 4186 (JSC::ARMAssembler::bkpt): 4187 (JSC::ARMAssembler::lsl): 4188 (JSC::ARMAssembler::lsr): 4189 (JSC::ARMAssembler::asr): 4190 (JSC::ARMAssembler::lsl_r): 4191 (JSC::ARMAssembler::lsr_r): 4192 (JSC::ARMAssembler::asr_r): 4193 (JSC::ARMAssembler::size): 4194 (JSC::ARMAssembler::ensureSpace): 4195 (JSC::ARMAssembler::label): 4196 (JSC::ARMAssembler::align): 4197 (JSC::ARMAssembler::jmp): 4198 (JSC::ARMAssembler::patchPointerInternal): 4199 (JSC::ARMAssembler::patchConstantPoolLoad): 4200 (JSC::ARMAssembler::patchPointer): 4201 (JSC::ARMAssembler::repatchInt32): 4202 (JSC::ARMAssembler::repatchPointer): 4203 (JSC::ARMAssembler::repatchLoadPtrToLEA): 4204 (JSC::ARMAssembler::linkJump): 4205 (JSC::ARMAssembler::relinkJump): 4206 (JSC::ARMAssembler::linkCall): 4207 (JSC::ARMAssembler::relinkCall): 4208 (JSC::ARMAssembler::getRelocatedAddress): 4209 (JSC::ARMAssembler::getDifferenceBetweenLabels): 4210 (JSC::ARMAssembler::getCallReturnOffset): 4211 (JSC::ARMAssembler::getOp2Byte): 4212 (JSC::ARMAssembler::placeConstantPoolBarrier): 4213 (JSC::ARMAssembler::RM): 4214 (JSC::ARMAssembler::RS): 4215 (JSC::ARMAssembler::RD): 4216 (JSC::ARMAssembler::RN): 4217 (JSC::ARMAssembler::getConditionalField): 4218 * assembler/ARMv7Assembler.h: 4219 (JSC::ARMv7Assembler::JmpSrc::enableLatePatch): 4220 * assembler/AbstractMacroAssembler.h: 4221 (JSC::AbstractMacroAssembler::Call::enableLatePatch): 4222 (JSC::AbstractMacroAssembler::Jump::enableLatePatch): 4223 * assembler/MacroAssembler.h: 4224 * assembler/MacroAssemblerARM.h: Added. 4225 (JSC::MacroAssemblerARM::): 4226 (JSC::MacroAssemblerARM::add32): 4227 (JSC::MacroAssemblerARM::and32): 4228 (JSC::MacroAssemblerARM::lshift32): 4229 (JSC::MacroAssemblerARM::mul32): 4230 (JSC::MacroAssemblerARM::not32): 4231 (JSC::MacroAssemblerARM::or32): 4232 (JSC::MacroAssemblerARM::rshift32): 4233 (JSC::MacroAssemblerARM::sub32): 4234 (JSC::MacroAssemblerARM::xor32): 4235 (JSC::MacroAssemblerARM::load32): 4236 (JSC::MacroAssemblerARM::load32WithAddressOffsetPatch): 4237 (JSC::MacroAssemblerARM::loadPtrWithPatchToLEA): 4238 (JSC::MacroAssemblerARM::load16): 4239 (JSC::MacroAssemblerARM::store32WithAddressOffsetPatch): 4240 (JSC::MacroAssemblerARM::store32): 4241 (JSC::MacroAssemblerARM::pop): 4242 (JSC::MacroAssemblerARM::push): 4243 (JSC::MacroAssemblerARM::move): 4244 (JSC::MacroAssemblerARM::swap): 4245 (JSC::MacroAssemblerARM::signExtend32ToPtr): 4246 (JSC::MacroAssemblerARM::zeroExtend32ToPtr): 4247 (JSC::MacroAssemblerARM::branch32): 4248 (JSC::MacroAssemblerARM::branch16): 4249 (JSC::MacroAssemblerARM::branchTest32): 4250 (JSC::MacroAssemblerARM::jump): 4251 (JSC::MacroAssemblerARM::branchAdd32): 4252 (JSC::MacroAssemblerARM::mull32): 4253 (JSC::MacroAssemblerARM::branchMul32): 4254 (JSC::MacroAssemblerARM::branchSub32): 4255 (JSC::MacroAssemblerARM::breakpoint): 4256 (JSC::MacroAssemblerARM::nearCall): 4257 (JSC::MacroAssemblerARM::call): 4258 (JSC::MacroAssemblerARM::ret): 4259 (JSC::MacroAssemblerARM::set32): 4260 (JSC::MacroAssemblerARM::setTest32): 4261 (JSC::MacroAssemblerARM::tailRecursiveCall): 4262 (JSC::MacroAssemblerARM::makeTailRecursiveCall): 4263 (JSC::MacroAssemblerARM::moveWithPatch): 4264 (JSC::MacroAssemblerARM::branchPtrWithPatch): 4265 (JSC::MacroAssemblerARM::storePtrWithPatch): 4266 (JSC::MacroAssemblerARM::supportsFloatingPoint): 4267 (JSC::MacroAssemblerARM::supportsFloatingPointTruncate): 4268 (JSC::MacroAssemblerARM::loadDouble): 4269 (JSC::MacroAssemblerARM::storeDouble): 4270 (JSC::MacroAssemblerARM::addDouble): 4271 (JSC::MacroAssemblerARM::subDouble): 4272 (JSC::MacroAssemblerARM::mulDouble): 4273 (JSC::MacroAssemblerARM::convertInt32ToDouble): 4274 (JSC::MacroAssemblerARM::branchDouble): 4275 (JSC::MacroAssemblerARM::branchTruncateDoubleToInt32): 4276 (JSC::MacroAssemblerARM::ARMCondition): 4277 (JSC::MacroAssemblerARM::prepareCall): 4278 (JSC::MacroAssemblerARM::call32): 4279 * assembler/X86Assembler.h: 4280 (JSC::X86Assembler::JmpSrc::enableLatePatch): 4281 * jit/ExecutableAllocator.h: 4282 (JSC::ExecutableAllocator::cacheFlush): 4283 * wtf/Platform.h: 4284 * yarr/RegexJIT.cpp: 4285 (JSC::Yarr::RegexGenerator::generateEnter): 4286 (JSC::Yarr::RegexGenerator::generateReturn): 4287 42882009-07-17 Gabor Loki <loki@inf.u-szeged.hu> 4289 4290 Reviewed by Gavin Barraclough. 4291 4292 Extend AssemblerBuffer with constant pool handling mechanism. 4293 https://bugs.webkit.org/show_bug.cgi?id=24986 4294 4295 Add a platform independed constant pool framework. 4296 This pool can store 32 or 64 bits values which is enough to hold 4297 any integer, pointer or double constant. 4298 4299 * assembler/AssemblerBuffer.h: 4300 (JSC::AssemblerBuffer::putIntUnchecked): 4301 (JSC::AssemblerBuffer::putInt64Unchecked): 4302 (JSC::AssemblerBuffer::append): 4303 (JSC::AssemblerBuffer::grow): 4304 * assembler/AssemblerBufferWithConstantPool.h: Added. 4305 (JSC::): 4306 43072009-07-17 Eric Roman <eroman@chromium.org> 4308 4309 Reviewed by Darin Adler. 4310 4311 Build fix for non-Darwin. 4312 Add a guard for inclusion of RetainPtr.h which includes CoreFoundation.h 4313 4314 https://bugs.webkit.org/show_bug.cgi?id=27382 4315 4316 * wtf/unicode/icu/CollatorICU.cpp: 4317 43182009-07-17 Alexey Proskuryakov <ap@webkit.org> 4319 4320 Reviewed by John Sullivan. 4321 4322 Get user default collation order via a CFLocale API when available. 4323 4324 * wtf/unicode/icu/CollatorICU.cpp: (WTF::Collator::userDefault): 4325 43262009-07-17 Laszlo Gombos <laszlo.1.gombos@nokia.com> 4327 4328 Reviewed by Simon Hausmann. 4329 4330 [Qt] Fix the include path for the Symbian port 4331 https://bugs.webkit.org/show_bug.cgi?id=27358 4332 4333 * JavaScriptCore.pri: 4334 43352009-07-17 Csaba Osztrogonac <oszi@inf.u-szeged.hu> 4336 4337 Reviewed by David Levin. 4338 4339 Build fix on platforms don't have MMAP. 4340 https://bugs.webkit.org/show_bug.cgi?id=27365 4341 4342 * interpreter/RegisterFile.h: Including stdio.h irrespectively of HAVE(MMAP) 4343 43442009-07-16 Fumitoshi Ukai <ukai@chromium.org> 4345 4346 Reviewed by David Levin. 4347 4348 Add --web-sockets flag and ENABLE_WEB_SOCKETS define. 4349 https://bugs.webkit.org/show_bug.cgi?id=27206 4350 4351 Add ENABLE_WEB_SOCKETS 4352 4353 * Configurations/FeatureDefines.xcconfig: add ENABLE_WEB_SOCKETS 4354 43552009-07-16 Maxime Simon <simon.maxime@gmail.com> 4356 4357 Reviewed by Eric Seidel. 4358 4359 Added Haiku-specific files for JavaScriptCore. 4360 https://bugs.webkit.org/show_bug.cgi?id=26620 4361 4362 * wtf/haiku/MainThreadHaiku.cpp: Added. 4363 (WTF::initializeMainThreadPlatform): 4364 (WTF::scheduleDispatchFunctionsOnMainThread): 4365 43662009-07-16 Gavin Barraclough <barraclough@apple.com> 4367 4368 RS by Oliver Hunt. 4369 4370 Revert r45969, this fix does not appear to be valid. 4371 https://bugs.webkit.org/show_bug.cgi?id=27077 4372 4373 * bytecode/CodeBlock.cpp: 4374 (JSC::CodeBlock::~CodeBlock): 4375 (JSC::CodeBlock::unlinkCallers): 4376 * jit/JIT.cpp: 4377 * jit/JIT.h: 4378 43792009-07-16 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 4380 4381 Reviewed by Oliver Hunt. 4382 4383 Allow custom memory allocation control in ExceptionInfo and RareData struct 4384 https://bugs.webkit.org/show_bug.cgi?id=27336 4385 4386 Inherits ExceptionInfo and RareData struct from FastAllocBase because these 4387 have been instantiated by 'new' in JavaScriptCore/bytecode/CodeBlock.cpp:1289 and 4388 in JavaScriptCore/bytecode/CodeBlock.h:453. 4389 4390 Remove unnecessary WTF:: namespace from CodeBlock inheritance. 4391 4392 * bytecode/CodeBlock.h: 4393 43942009-07-16 Mark Rowe <mrowe@apple.com> 4395 4396 Rubber-stamped by Geoff Garen. 4397 4398 Fix FeatureDefines.xcconfig to not be out of sync with the rest of the world. 4399 4400 * Configurations/FeatureDefines.xcconfig: 4401 44022009-07-16 Yong Li <yong.li@torchmobile.com> 4403 4404 Reviewed by George Staikos. 4405 4406 https://bugs.webkit.org/show_bug.cgi?id=27320 4407 _countof is only included in CE6; for CE5 we need to define it ourself 4408 4409 * wtf/Platform.h: 4410 44112009-07-16 Zoltan Herczeg <zherczeg@inf.u-szeged.hu> 4412 4413 Reviewed by Oliver Hunt. 4414 4415 Workers + garbage collector: weird crashes 4416 https://bugs.webkit.org/show_bug.cgi?id=27077 4417 4418 We need to unlink cached method call sites when a function is destroyed. 4419 4420 * JavaScriptCore.xcodeproj/project.pbxproj: 4421 * bytecode/CodeBlock.cpp: 4422 (JSC::CodeBlock::~CodeBlock): 4423 (JSC::CodeBlock::unlinkCallers): 4424 * jit/JIT.cpp: 4425 (JSC::JIT::unlinkMethodCall): 4426 * jit/JIT.h: 4427 44282009-07-15 Steve Falkenburg <sfalken@apple.com> 4429 4430 Windows Build fix. 4431 4432 Visual Studio reset our intermediate directory on us. 4433 This sets it back. 4434 4435 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: 4436 * JavaScriptCore.vcproj/testapi/testapi.vcproj: 4437 44382009-07-15 Kwang Yul Seo <skyul@company100.net> 4439 4440 Reviewed by Eric Seidel. 4441 4442 https://bugs.webkit.org/show_bug.cgi?id=26794 4443 Make Yacc-generated parsers to use fastMalloc/fastFree. 4444 4445 Define YYMALLOC and YYFREE to fastMalloc and fastFree 4446 respectively. 4447 4448 * parser/Grammar.y: 4449 44502009-07-15 Darin Adler <darin@apple.com> 4451 4452 Fix a build for a particular Apple configuration. 4453 4454 * wtf/FastAllocBase.h: Change include to use "" style for 4455 including another wtf header. This is the style we use for 4456 including other public headers in the same directory. 4457 44582009-07-15 George Staikos <george.staikos@torchmobile.com> 4459 4460 Reviewed by Adam Treat. 4461 4462 https://bugs.webkit.org/show_bug.cgi?id=27303 4463 Implement createThreadInternal for WinCE. 4464 Contains changes by George Staikos <george.staikos@torchmobile.com> and Joe Mason <joe.mason@torchmobile.com> 4465 4466 * wtf/ThreadingWin.cpp: 4467 (WTF::createThreadInternal): 4468 44692009-07-15 Joe Mason <joe.mason@torchmobile.com> 4470 4471 Reviewed by George Staikos. 4472 4473 https://bugs.webkit.org/show_bug.cgi?id=27298 4474 Platform defines for WINCE. 4475 Contains changes by Yong Li <yong.li@torchmobile.com>, 4476 George Staikos <george.staikos@torchmobile.com> and Joe Mason <joe.mason@torchmobile.com> 4477 4478 * wtf/Platform.h: 4479 44802009-07-15 Yong Li <yong.li@torchmobile.com> 4481 4482 Reviewed by Adam Treat. 4483 4484 https://bugs.webkit.org/show_bug.cgi?id=27306 4485 Use RegisterClass instead of RegisterClassEx on WinCE. 4486 4487 * wtf/win/MainThreadWin.cpp: 4488 (WTF::initializeMainThreadPlatform): 4489 44902009-07-15 Yong Li <yong.li@torchmobile.com> 4491 4492 Reviewed by George Staikos. 4493 4494 https://bugs.webkit.org/show_bug.cgi?id=27301 4495 Use OutputDebugStringW on WinCE since OutputDebugStringA is not supported 4496 Originally written by Yong Li <yong.li@torchmobile.com> and refactored by 4497 Joe Mason <joe.mason@torchmobile.com> 4498 4499 * wtf/Assertions.cpp: vprintf_stderr_common 4500 45012009-07-15 Yong Li <yong.li@torchmobile.com> 4502 4503 Reviewed by George Staikos. 4504 4505 https://bugs.webkit.org/show_bug.cgi?id=27020 4506 msToGregorianDateTime should set utcOffset to 0 when outputIsUTC is false 4507 4508 * wtf/DateMath.cpp: 4509 (WTF::gregorianDateTimeToMS): 4510 45112009-07-15 Laszlo Gombos <laszlo.1.gombos@nokia.com> 4512 4513 Reviewed by Simon Hausmann. 4514 4515 [Qt] Cleanup - Remove obsolete code from the make system 4516 https://bugs.webkit.org/show_bug.cgi?id=27299 4517 4518 * JavaScriptCore.pro: 4519 * jsc.pro: 4520 45212009-07-07 Norbert Leser <norbert.leser@nokia.com> 4522 4523 Reviewed by Simon Hausmann. 4524 4525 https://bugs.webkit.org/show_bug.cgi?id=27056 4526 4527 Alternate bool operator for codewarrior compiler (WINSCW). 4528 Compiler (latest b482) reports error for UnspecifiedBoolType construct: 4529 "illegal explicit conversion from 'WTF::OwnArrayPtr<JSC::Register>' to 'bool'" 4530 4531 Same fix as in r38391. 4532 4533 * JavaScriptCore/wtf/OwnArrayPtr.h: 4534 45352009-07-15 Norbert Leser <norbert.leser@nokia.com> 4536 4537 Reviewed by Darin Adler. 4538 4539 Qualify include path with wtf to fix compilation 4540 on Symbian. 4541 https://bugs.webkit.org/show_bug.cgi?id=27055 4542 4543 * interpreter/Interpreter.h: 4544 45452009-07-15 Laszlo Gombos <laszlo.1.gombos@nokia.com> 4546 4547 Reviewed by Dave Kilzer. 4548 4549 Turn off non-portable date manipulations for SYMBIAN 4550 https://bugs.webkit.org/show_bug.cgi?id=27064 4551 4552 Introduce HAVE(TM_GMTOFF), HAVE(TM_ZONE) and HAVE(TIMEGM) guards 4553 and place the rules for controlling the guards in Platform.h. 4554 Turn off these newly introduced guards for SYMBIAN. 4555 4556 * wtf/DateMath.cpp: 4557 (WTF::calculateUTCOffset): 4558 * wtf/DateMath.h: 4559 (WTF::GregorianDateTime::GregorianDateTime): 4560 (WTF::GregorianDateTime::operator tm): 4561 * wtf/Platform.h: 4562 45632009-07-15 Norbert Leser <norbert.leser@nokia.com> 4564 4565 Reviewed by Simon Hausmann. 4566 4567 Undef ASSERT on Symbian, to avoid excessive warnings 4568 https://bugs.webkit.org/show_bug.cgi?id=27052 4569 4570 * wtf/Assertions.h: 4571 45722009-07-15 Oliver Hunt <oliver@apple.com> 4573 4574 Reviewed by Simon Hausmann. 4575 4576 REGRESSION: fast/js/postfix-syntax.html fails with interpreter 4577 https://bugs.webkit.org/show_bug.cgi?id=27294 4578 4579 When postfix operators operating on locals assign to the same local 4580 the order of operations has to be to store the incremented value, then 4581 store the unmodified number. Rather than implementing this subtle 4582 semantic in the interpreter I've just made the logic explicit in the 4583 bytecode generator, so x=x++ effectively becomes x=ToNumber(x) (for a 4584 local var x). 4585 4586 * parser/Nodes.cpp: 4587 (JSC::emitPostIncOrDec): 4588 45892009-07-15 Oliver Hunt <oliver@apple.com> 4590 4591 Reviewed by Simon Hausmann. 4592 4593 REGRESSION(43559): fast/js/kde/arguments-scope.html fails with interpreter 4594 https://bugs.webkit.org/show_bug.cgi?id=27259 4595 4596 The interpreter was incorrectly basing its need to create the arguments object 4597 based on the presence of the callframe's argument reference rather than the local 4598 arguments reference. Based on this it then overrode the local variable reference. 4599 4600 * interpreter/Interpreter.cpp: 4601 (JSC::Interpreter::privateExecute): 4602 46032009-07-14 Steve Falkenburg <sfalken@apple.com> 4604 4605 Reorganize JavaScriptCore headers into: 4606 API: include/JavaScriptCore/ 4607 Private: include/private/JavaScriptCore/ 4608 4609 Reviewed by Darin Adler. 4610 4611 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: 4612 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCoreCommon.vsprops: 4613 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCoreGenerated.make: 4614 * JavaScriptCore.vcproj/testapi/testapi.vcproj: 4615 * JavaScriptCore.vcproj/testapi/testapiCommon.vsprops: 4616 46172009-07-14 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 4618 4619 Reviewed by Darin Adler. 4620 4621 Change JSCell's superclass to NoncopyableCustomAllocated 4622 https://bugs.webkit.org/show_bug.cgi?id=27248 4623 4624 JSCell class customizes operator new, since Noncopyable will be 4625 inherited from FastAllocBase, NoncopyableCustomAllocated has 4626 to be used. 4627 4628 * runtime/JSCell.h: 4629 46302009-07-14 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 4631 4632 Reviewed by Darin Adler. 4633 4634 Change all Noncopyable inheriting visibility to public. 4635 https://bugs.webkit.org/show_bug.cgi?id=27225 4636 4637 Change all Noncopyable inheriting visibility to public because 4638 it is needed to the custom allocation framework (bug #20422). 4639 4640 * bytecode/SamplingTool.h: 4641 * bytecompiler/RegisterID.h: 4642 * interpreter/CachedCall.h: 4643 * interpreter/RegisterFile.h: 4644 * parser/Lexer.h: 4645 * parser/Parser.h: 4646 * runtime/ArgList.h: 4647 * runtime/BatchedTransitionOptimizer.h: 4648 * runtime/Collector.h: 4649 * runtime/CommonIdentifiers.h: 4650 * runtime/JSCell.h: 4651 * runtime/JSGlobalObject.h: 4652 * runtime/JSLock.h: 4653 * runtime/JSONObject.cpp: 4654 * runtime/SmallStrings.cpp: 4655 * runtime/SmallStrings.h: 4656 * wtf/CrossThreadRefCounted.h: 4657 * wtf/GOwnPtr.h: 4658 * wtf/Locker.h: 4659 * wtf/MessageQueue.h: 4660 * wtf/OwnArrayPtr.h: 4661 * wtf/OwnFastMallocPtr.h: 4662 * wtf/OwnPtr.h: 4663 * wtf/RefCounted.h: 4664 * wtf/ThreadSpecific.h: 4665 * wtf/Threading.h: 4666 * wtf/Vector.h: 4667 * wtf/unicode/Collator.h: 4668 46692009-07-14 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 4670 4671 Reviewed by Darin Adler. 4672 4673 Change ParserArenaRefCounted's superclass to RefCountedCustomAllocated 4674 https://bugs.webkit.org/show_bug.cgi?id=27249 4675 4676 ParserArenaDeletable customizes operator new, to avoid double inheritance 4677 ParserArenaDeletable's superclass has been changed to RefCountedCustomAllocated. 4678 4679 * parser/Nodes.h: 4680 46812009-07-14 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 4682 4683 Reviewed by Darin Adler. 4684 4685 Add RefCountedCustomAllocated to RefCounted.h 4686 https://bugs.webkit.org/show_bug.cgi?id=27232 4687 4688 Some class which are inherited from RefCounted customize 4689 operator new, but RefCounted is inherited from Noncopyable 4690 which will be inherited from FastAllocBase. To avoid 4691 conflicts Noncopyable inheriting was moved down to RefCounted 4692 and to avoid double inheritance this class has been added. 4693 4694 * wtf/RefCounted.h: 4695 (WTF::RefCountedCustomAllocated::deref): 4696 (WTF::RefCountedCustomAllocated::~RefCountedCustomAllocated): 4697 46982009-07-14 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 4699 4700 Reviewed by Darin Adler. 4701 4702 Add NoncopyableCustomAllocated to Noncopyable.h. 4703 https://bugs.webkit.org/show_bug.cgi?id=27228 4704 4705 Some classes which inherited from Noncopyable overrides operator new 4706 since Noncopyable'll be inherited from FastAllocBase, Noncopyable.h 4707 needs to be extended with this new class to support the overriding. 4708 4709 * wtf/Noncopyable.h: 4710 (WTFNoncopyable::NoncopyableCustomAllocated::NoncopyableCustomAllocated): 4711 (WTFNoncopyable::NoncopyableCustomAllocated::~NoncopyableCustomAllocated): 4712 47132009-07-14 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 4714 4715 Reviewed by Darin Adler. 4716 4717 Allow custom memory allocation control for JavaScriptCore's IdentifierTable class 4718 https://bugs.webkit.org/show_bug.cgi?id=27260 4719 4720 Inherits IdentifierTable class from FastAllocBase because it has been 4721 instantiated by 'new' in JavaScriptCore/runtime/Identifier.cpp:70. 4722 4723 * runtime/Identifier.cpp: 4724 47252009-07-14 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 4726 4727 Reviewed by Darin Adler. 4728 4729 Allow custom memory allocation control for JavaScriptCore's Profiler class 4730 https://bugs.webkit.org/show_bug.cgi?id=27253 4731 4732 Inherits Profiler class from FastAllocBase because it has been instantiated by 4733 'new' in JavaScriptCore/profiler/Profiler.cpp:56. 4734 4735 * profiler/Profiler.h: 4736 47372009-07-06 George Staikos <george.staikos@torchmobile.com> 4738 4739 Reviewed by Adam Treat. 4740 4741 Authors: George Staikos <george.staikos@torchmobile.com>, Joe Mason <joe.mason@torchmobile.com>, Makoto Matsumoto <matumoto@math.keio.ac.jp>, Takuji Nishimura 4742 4743 https://bugs.webkit.org/show_bug.cgi?id=27030 4744 Implement custom RNG for WinCE using Mersenne Twister 4745 4746 * wtf/RandomNumber.cpp: 4747 (WTF::randomNumber): 4748 * wtf/RandomNumberSeed.h: 4749 (WTF::initializeRandomNumberGenerator): 4750 * wtf/wince/mt19937ar.c: Added. 4751 (init_genrand): 4752 (init_by_array): 4753 (genrand_int32): 4754 (genrand_int31): 4755 (genrand_real1): 4756 (genrand_real2): 4757 (genrand_real3): 4758 (genrand_res53): 4759 47602009-07-13 Gustavo Noronha Silva <gustavo.noronha@collabora.co.uk> 4761 4762 Unreviewed make dist build fix. 4763 4764 * GNUmakefile.am: 4765 47662009-07-13 Drew Wilson <atwilson@google.com> 4767 4768 Reviewed by David Levin. 4769 4770 Add ENABLE(SHARED_WORKERS) flag and define SharedWorker APIs 4771 https://bugs.webkit.org/show_bug.cgi?id=26932 4772 4773 Added ENABLE(SHARED_WORKERS) flag (off by default). 4774 4775 * Configurations/FeatureDefines.xcconfig: 4776 47772009-07-07 Norbert Leser <norbert.leser@nokia.com> 4778 4779 Reviewed by Maciej Stachoviak. 4780 4781 https://bugs.webkit.org/show_bug.cgi?id=27058 4782 4783 Removed superfluous parenthesis around single expression. 4784 Compilers on Symbian platform fail to properly parse and compile. 4785 4786 * JavaScriptCore/wtf/Platform.h: 4787 47882009-07-13 Norbert Leser <norbert.leser@nokia.com> 4789 4790 Reviewed by Maciej Stachoviak. 4791 4792 https://bugs.webkit.org/show_bug.cgi?id=27054 4793 4794 Renamed Translator to HashTranslator 4795 4796 Codewarrior compiler (WINSCW) latest b482 cannot resolve typename 4797 mismatch between template declaration and definition 4798 (HashTranslator / Translator) 4799 4800 * wtf/HashSet.h: 4801 48022009-07-13 Norbert Leser <norbert.leser@nokia.com> 4803 4804 Reviewed by Eric Seidel. 4805 4806 https://bugs.webkit.org/show_bug.cgi?id=27053 4807 4808 Ambiguity in LabelScope initialization 4809 4810 Codewarrior compiler (WINSCW) latest b482 on Symbian cannot resolve 4811 type of "0" unambiguously. Set expression explicitly to 4812 PassRefPtr<Label>::PassRefPtr() 4813 4814 * bytecompiler/BytecodeGenerator.cpp 4815 48162009-07-11 Simon Fraser <simon.fraser@apple.com> 4817 4818 Enable support for accelerated compositing and 3d transforms on Leopard. 4819 <https://bugs.webkit.org/show_bug.cgi?id=20166> 4820 <rdar://problem/6120614> 4821 4822 Reviewed by Oliver Hunt. 4823 4824 * Configurations/FeatureDefines.xcconfig: 4825 * wtf/Platform.h: 4826 48272009-07-10 Mark Rowe <mrowe@apple.com> 4828 4829 Second part of the "make Windows happier" dance. 4830 4831 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 4832 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 4833 48342009-07-10 Mark Rowe <mrowe@apple.com> 4835 4836 Try and make the Windows build happy. 4837 4838 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 4839 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 4840 48412009-07-10 Kevin McCullough <kmccullough@apple.com> 4842 4843 Reviewed by Geoffrey Garen. 4844 4845 * debugger/Debugger.h: Made this function virtual for use in WebCore's 4846 WebInspector. 4847 48482009-07-10 Kwang Yul Seo <skyul@company100.net> 4849 4850 Reviewed by Darin Adler. 4851 4852 ParserArenaDeletable should override delete 4853 https://bugs.webkit.org/show_bug.cgi?id=26790 4854 4855 ParserArenaDeletable overrides new, but it does not override delete. 4856 ParserArenaDeletable must be freed by fastFree 4857 because it is allocated by fastMalloc. 4858 4859 * parser/NodeConstructors.h: 4860 (JSC::ParserArenaDeletable::operator delete): 4861 * parser/Nodes.h: 4862 48632009-07-10 Adam Roben <aroben@apple.com> 4864 4865 Sort all our Xcode projects 4866 4867 Accomplished using sort-Xcode-project-file. 4868 4869 Requested by Dave Kilzer. 4870 4871 * JavaScriptCore.xcodeproj/project.pbxproj: 4872 48732009-07-09 Maciej Stachowiak <mjs@apple.com> 4874 4875 Not reviewed, build fix. 4876 4877 Windows build fix for the last change. 4878 4879 * wtf/dtoa.cpp: Forgot to include Vector.h 4880 48812009-07-09 Maciej Stachowiak <mjs@apple.com> 4882 4883 Reviewed by Darin Adler. 4884 4885 REGRESSION: crash in edge cases of floating point parsing. 4886 https://bugs.webkit.org/show_bug.cgi?id=27110 4887 <rdar://problem/7044458> 4888 4889 Tests: fast/css/number-parsing-crash.html 4890 fast/css/number-parsing-crash.html 4891 fast/js/number-parsing-crash.html 4892 4893 * wtf/dtoa.cpp: 4894 (WTF::BigInt::BigInt): Converted this to more a proper class, using a Vector 4895 with inline capacity 4896 4897 (WTF::lshift): Rearranged logic somewhat nontrivially to deal with the new way of sizing BigInts. 4898 Added an assertion to verify that invariants are maintained. 4899 4900 All other functions are adapted fairly mechanically to the above changes. 4901 (WTF::BigInt::clear): 4902 (WTF::BigInt::size): 4903 (WTF::BigInt::resize): 4904 (WTF::BigInt::words): 4905 (WTF::BigInt::append): 4906 (WTF::multadd): 4907 (WTF::s2b): 4908 (WTF::i2b): 4909 (WTF::mult): 4910 (WTF::cmp): 4911 (WTF::diff): 4912 (WTF::b2d): 4913 (WTF::d2b): 4914 (WTF::ratio): 4915 (WTF::strtod): 4916 (WTF::quorem): 4917 (WTF::dtoa): 4918 49192009-07-09 Drew Wilson <atwilson@google.com> 4920 4921 Reviewed by Alexey Proskuryakov. 4922 4923 Turned on CHANNEL_MESSAGING by default because the MessageChannel API 4924 can now be implemented for Web Workers and is reasonably stable. 4925 4926 * Configurations/FeatureDefines.xcconfig: 4927 49282009-07-09 Oliver Hunt <oliver@apple.com> 4929 4930 Reviewed by NOBODY (Build fix). 4931 4932 * interpreter/Interpreter.cpp: 4933 (JSC::Interpreter::privateExecute): 4934 49352009-07-09 Oliver Hunt <oliver@apple.com> 4936 4937 Reviewed by Darin Adler. 4938 4939 Bug 27016 - Interpreter crashes due to invalid array indexes 4940 <https://bugs.webkit.org/show_bug.cgi?id=27016> 4941 4942 Unsigned vs signed conversions results in incorrect behaviour in 4943 64bit interpreter builds. 4944 4945 * interpreter/Interpreter.cpp: 4946 (JSC::Interpreter::privateExecute): 4947 49482009-07-09 Dimitri Glazkov <dglazkov@chromium.org> 4949 4950 Reviewed by Darin Fisher. 4951 4952 [Chromium] Upstream JavaScriptCore.gypi, the project file for Chromium build. 4953 https://bugs.webkit.org/show_bug.cgi?id=27135 4954 4955 * JavaScriptCore.gypi: Added. 4956 49572009-07-09 Joe Mason <joe.mason@torchmobile.com> 4958 4959 Reviewed by George Staikos. 4960 4961 Authors: Yong Li <yong.li@torchmobile.com>, Joe Mason <joe.mason@torchmobile.com> 4962 4963 https://bugs.webkit.org/show_bug.cgi?id=27031 4964 Add an override for deleteOwnedPtr(HDC) on Windows 4965 4966 * wtf/OwnPtrCommon.h: 4967 * wtf/OwnPtrWin.cpp: 4968 (WTF::deleteOwnedPtr): 4969 49702009-07-09 Laszlo Gombos <laszlo.1.gombos@nokia.com> 4971 4972 Reviewed by Darin Adler. 4973 4974 Guard singal.h dependency with HAVE(SIGNAL_H) to enable building jsc 4975 on SYMBIAN. 4976 4977 https://bugs.webkit.org/show_bug.cgi?id=27026 4978 4979 Based on Norbert Leser's work. 4980 4981 * jsc.cpp: 4982 (printUsageStatement): 4983 (parseArguments): 4984 * wtf/Platform.h: 4985 49862009-07-07 Gavin Barraclough <barraclough@apple.com> 4987 4988 Reviewed by Sam Weinig. 4989 4990 Stop loading constants into the register file. 4991 4992 Instead, use high register values (highest bit bar the sign bit set) to indicate 4993 constants in the instruction stream, and when we encounter such a value load it 4994 directly from the CodeBlock. 4995 4996 Since constants are no longer copied into the register file, this patch renders 4997 the 'unexpected constant' mechanism redundant, and removes it. 4998 4999 2% improvement, thanks to Sam Weinig. 5000 5001 * bytecode/CodeBlock.cpp: 5002 (JSC::CodeBlock::dump): 5003 (JSC::CodeBlock::CodeBlock): 5004 (JSC::CodeBlock::mark): 5005 (JSC::CodeBlock::shrinkToFit): 5006 * bytecode/CodeBlock.h: 5007 (JSC::CodeBlock::isTemporaryRegisterIndex): 5008 (JSC::CodeBlock::constantRegister): 5009 (JSC::CodeBlock::isConstantRegisterIndex): 5010 (JSC::CodeBlock::getConstant): 5011 (JSC::ExecState::r): 5012 * bytecode/Opcode.h: 5013 * bytecompiler/BytecodeGenerator.cpp: 5014 (JSC::BytecodeGenerator::preserveLastVar): 5015 (JSC::BytecodeGenerator::BytecodeGenerator): 5016 (JSC::BytecodeGenerator::addConstantValue): 5017 (JSC::BytecodeGenerator::emitEqualityOp): 5018 (JSC::BytecodeGenerator::emitLoad): 5019 (JSC::BytecodeGenerator::emitResolveBase): 5020 (JSC::BytecodeGenerator::emitResolveWithBase): 5021 (JSC::BytecodeGenerator::emitNewError): 5022 * bytecompiler/BytecodeGenerator.h: 5023 (JSC::BytecodeGenerator::emitNode): 5024 * interpreter/CallFrame.h: 5025 (JSC::ExecState::noCaller): 5026 (JSC::ExecState::hasHostCallFrameFlag): 5027 (JSC::ExecState::addHostCallFrameFlag): 5028 (JSC::ExecState::removeHostCallFrameFlag): 5029 * interpreter/Interpreter.cpp: 5030 (JSC::Interpreter::resolve): 5031 (JSC::Interpreter::resolveSkip): 5032 (JSC::Interpreter::resolveGlobal): 5033 (JSC::Interpreter::resolveBase): 5034 (JSC::Interpreter::resolveBaseAndProperty): 5035 (JSC::Interpreter::resolveBaseAndFunc): 5036 (JSC::Interpreter::dumpRegisters): 5037 (JSC::Interpreter::throwException): 5038 (JSC::Interpreter::createExceptionScope): 5039 (JSC::Interpreter::privateExecute): 5040 (JSC::Interpreter::retrieveArguments): 5041 * jit/JIT.cpp: 5042 (JSC::JIT::privateCompileMainPass): 5043 * jit/JITInlineMethods.h: 5044 (JSC::JIT::emitLoadDouble): 5045 (JSC::JIT::emitLoadInt32ToDouble): 5046 * jit/JITOpcodes.cpp: 5047 (JSC::JIT::emit_op_new_error): 5048 (JSC::JIT::emit_op_enter): 5049 (JSC::JIT::emit_op_enter_with_activation): 5050 * parser/Nodes.cpp: 5051 (JSC::DeleteResolveNode::emitBytecode): 5052 (JSC::DeleteValueNode::emitBytecode): 5053 (JSC::PrefixResolveNode::emitBytecode): 5054 * runtime/JSActivation.cpp: 5055 (JSC::JSActivation::JSActivation): 5056 * wtf/Platform.h: 5057 50582009-07-07 Mark Rowe <mrowe@apple.com> 5059 5060 Reviewed by Darin Adler. 5061 5062 Fix <https://bugs.webkit.org/show_bug.cgi?id=27025> / <rdar://problem/7033448>. 5063 Bug 27025: Crashes and regression test failures related to regexps in 64-bit 5064 5065 For x86_64 RegexGenerator uses rbx, a callee-save register, as a scratch register but 5066 neglects to save and restore it. The change in handling of the output vector in r45545 5067 altered code generation so that the RegExp::match was now storing important data in rbx, 5068 which caused crashes and bogus results when it was clobbered. 5069 5070 * yarr/RegexJIT.cpp: 5071 (JSC::Yarr::RegexGenerator::generateEnter): Save rbx. 5072 (JSC::Yarr::RegexGenerator::generateReturn): Restore rbx. 5073 50742009-07-06 Ada Chan <adachan@apple.com> 5075 5076 Reviewed by Darin Adler and Mark Rowe. 5077 5078 Decommitted spans are added to the list of normal spans rather than 5079 the returned spans in TCMalloc_PageHeap::Delete(). 5080 https://bugs.webkit.org/show_bug.cgi?id=26998 5081 5082 In TCMalloc_PageHeap::Delete(), the deleted span can be decommitted in 5083 the process of merging with neighboring spans that are also decommitted. 5084 The merged span needs to be placed in the list of returned spans (spans 5085 whose memory has been returned to the system). Right now it's always added 5086 to the list of the normal spans which can theoretically cause thrashing. 5087 5088 * wtf/FastMalloc.cpp: 5089 (WTF::TCMalloc_PageHeap::Delete): 5090 50912009-07-05 Lars Knoll <lars.knoll@nokia.com> 5092 5093 Reviewed by Maciej Stachowiak. 5094 5095 https://bugs.webkit.org/show_bug.cgi?id=26843 5096 5097 Fix run-time crashes in JavaScriptCore with the Metrowerks compiler on Symbian. 5098 5099 The Metrowerks compiler on the Symbian platform moves the globally 5100 defined Hashtables into read-only memory, despite one of the members 5101 being mutable. This causes crashes at run-time due to write access to 5102 read-only memory. 5103 5104 Avoid the use of const with this compiler by introducing the 5105 JSC_CONST_HASHTABLE macro. 5106 5107 Based on idea by Norbert Leser. 5108 5109 * runtime/Lookup.h: Define JSC_CONST_HASHTABLE as const for !WINSCW. 5110 * create_hash_table: Use JSC_CONST_HASHTABLE for hashtables. 5111 * runtime/JSGlobalData.cpp: Import various global hashtables via the macro. 5112 51132009-07-04 Dan Bernstein <mitz@apple.com> 5114 5115 - debug build fix 5116 5117 * runtime/RegExpConstructor.cpp: 5118 (JSC::RegExpConstructor::getLastParen): 5119 51202009-07-03 Yong Li <yong.li@torchmobile.com> 5121 5122 Reviewed by Maciej Stachowiak (and revised slightly) 5123 5124 RegExp::match to be optimized 5125 https://bugs.webkit.org/show_bug.cgi?id=26957 5126 5127 Allow regexp matching to use Vectors with inline capacity instead of 5128 allocating a new ovector buffer every time. 5129 5130 ~5% speedup on SunSpider string-unpack-code test, 0.3% on SunSpider overall. 5131 5132 * runtime/RegExp.cpp: 5133 (JSC::RegExp::match): 5134 * runtime/RegExp.h: 5135 * runtime/RegExpConstructor.cpp: 5136 (JSC::RegExpConstructorPrivate::RegExpConstructorPrivate): 5137 (JSC::RegExpConstructorPrivate::lastOvector): 5138 (JSC::RegExpConstructorPrivate::tempOvector): 5139 (JSC::RegExpConstructorPrivate::changeLastOvector): 5140 (JSC::RegExpConstructor::performMatch): 5141 (JSC::RegExpMatchesArray::RegExpMatchesArray): 5142 (JSC::RegExpMatchesArray::fillArrayInstance): 5143 (JSC::RegExpConstructor::getBackref): 5144 (JSC::RegExpConstructor::getLastParen): 5145 (JSC::RegExpConstructor::getLeftContext): 5146 (JSC::RegExpConstructor::getRightContext): 5147 * runtime/StringPrototype.cpp: 5148 (JSC::stringProtoFuncSplit): 5149 51502009-06-30 Kwang Yul Seo <skyul@company100.net> 5151 5152 Reviewed by Eric Seidel. 5153 5154 Override operator new/delete with const std::nothrow_t& as the second 5155 argument. 5156 https://bugs.webkit.org/show_bug.cgi?id=26792 5157 5158 On Windows CE, operator new/delete, new[]/delete[] with const 5159 std::nothrow_t& must be overrided because some standard template 5160 libraries use these operators. 5161 5162 The problem occurs when memory allocated by new(size_t s, const 5163 std::nothrow_t&) is freed by delete(void* p). This causes the umatched 5164 malloc/free problem. 5165 5166 The patch overrides all new, delete, new[] and delete[] to use 5167 fastMaloc and fastFree consistently. 5168 5169 * wtf/FastMalloc.h: 5170 (throw): 5171 51722009-06-30 Gabor Loki <loki@inf.u-szeged.hu> 5173 5174 Reviewed by Sam Weinig. 5175 5176 <https://bugs.webkit.org/show_bug.cgi?id=24986> 5177 5178 Remove unnecessary references to AssemblerBuffer. 5179 5180 * interpreter/Interpreter.cpp: 5181 * interpreter/Interpreter.h: 5182 51832009-06-29 David Levin <levin@chromium.org> 5184 5185 Reviewed by Oliver Hunt. 5186 5187 Still seeing occasional leaks from UString::sharedBuffer code 5188 https://bugs.webkit.org/show_bug.cgi?id=26420 5189 5190 The problem is that the pointer to the memory allocation isn't visible 5191 by "leaks" due to the lower bits being used as flags. The fix is to 5192 make the pointer visible in memory (in debug only). The downside of 5193 this fix that the memory allocated by sharedBuffer will still look like 5194 a leak in non-debug builds when any flags are set. 5195 5196 * wtf/PtrAndFlags.h: 5197 (WTF::PtrAndFlags::set): 5198 51992009-06-29 Sam Weinig <sam@webkit.org> 5200 5201 Reviewed by Mark Rowe. 5202 5203 Remove more unused scons support. 5204 5205 * SConstruct: Removed. 5206 52072009-06-29 Oliver Hunt <oliver@apple.com> 5208 5209 Reviewed by Gavin Barraclough. 5210 5211 <rdar://problem/7016214> JSON.parse fails to parse valid JSON with most Unicode characters 5212 <https://bugs.webkit.org/show_bug.cgi?id=26802> 5213 5214 In the original JSON.parse patch unicode was handled correctly, however in some last 5215 minute "clean up" I oversimplified isSafeStringCharacter. This patch corrects this bug. 5216 5217 * runtime/LiteralParser.cpp: 5218 (JSC::isSafeStringCharacter): 5219 (JSC::LiteralParser::Lexer::lexString): 5220 52212009-06-26 Oliver Hunt <oliver@apple.com> 5222 5223 Reviewed by Dan Bernstein. 5224 5225 <rdar://problem/7009684> REGRESSION(r45039): Crashes inside JSEvent::put on PowerPC (26746) 5226 <https://bugs.webkit.org/show_bug.cgi?id=26746> 5227 5228 Fix for r45039 incorrectly uncached a get_by_id by converting it to put_by_id. Clearly this 5229 is less than correct. This patch corrects that error. 5230 5231 * interpreter/Interpreter.cpp: 5232 (JSC::Interpreter::tryCacheGetByID): 5233 52342009-06-26 Eric Seidel <eric@webkit.org> 5235 5236 No review, only rolling out r45259. 5237 5238 Roll out r45259 after crash appeared on the bots: 5239 plugins/undefined-property-crash.html 5240 ASSERTION FAILED: s <= HeapConstants<heapType>::cellSize 5241 (leopard-intel-debug-tests/build/JavaScriptCore/runtime/Collector.cpp:278 5242 void* JSC::Heap::heapAllocate(size_t) [with JSC::HeapType heapType = PrimaryHeap]) 5243 5244 * runtime/DateInstance.cpp: 5245 * runtime/Identifier.cpp: 5246 * runtime/Lookup.h: 5247 * runtime/RegExpConstructor.cpp: 5248 * runtime/RegExpObject.h: 5249 * runtime/ScopeChain.h: 5250 * runtime/UString.h: 5251 52522009-06-26 Jedrzej Nowacki <jedrzej.nowacki@nokia.com> 5253 5254 Reviewed by Simon Hausmann. 5255 5256 Add support for QDataStream operators to Vector. 5257 5258 * wtf/Vector.h: 5259 (WTF::operator<<): 5260 (WTF::operator>>): 5261 52622009-06-24 Sam Weinig <sam@webkit.org> 5263 5264 Reviewed by Gavin Barraclough. 5265 5266 Make the opcode sampler work once again. 5267 5268 * jit/JIT.h: 5269 (JSC::JIT::compileGetByIdProto): 5270 (JSC::JIT::compileGetByIdSelfList): 5271 (JSC::JIT::compileGetByIdProtoList): 5272 (JSC::JIT::compileGetByIdChainList): 5273 (JSC::JIT::compileGetByIdChain): 5274 (JSC::JIT::compilePutByIdTransition): 5275 (JSC::JIT::compileCTIMachineTrampolines): 5276 (JSC::JIT::compilePatchGetArrayLength): 5277 * jit/JITStubCall.h: 5278 (JSC::JITStubCall::call): 5279 52802009-06-24 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 5281 5282 Reviewed by Maciej Stachowiak. 5283 5284 Extend FastAllocBase.h with 'using WTF::FastAllocBase' to avoid 5285 unnecessary WTF:: usings. 5286 Remove existing unnecessary WTF:: usings. 5287 5288 * interpreter/Interpreter.h: 5289 * profiler/CallIdentifier.h: 5290 * runtime/ScopeChain.h: 5291 * wtf/FastAllocBase.h: 5292 52932009-06-24 David Levin <levin@chromium.org> 5294 5295 Fix all builds. 5296 5297 * bytecode/CodeBlock.h: 5298 * bytecompiler/BytecodeGenerator.h: 5299 * interpreter/Register.h: 5300 53012009-06-24 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 5302 5303 Reviewed by Maciej Stachowiak. 5304 5305 https://bugs.webkit.org/show_bug.cgi?id=26677 5306 5307 Inherits CodeBlock class from FastAllocBase because it 5308 has been instantiated by 'new' in JavaScriptCore/bytecode/CodeBlock.h:217. 5309 5310 * bytecode/CodeBlock.h: 5311 53122009-06-24 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 5313 5314 Reviewed by Maciej Stachowiak. 5315 5316 https://bugs.webkit.org/show_bug.cgi?id=26676 5317 5318 Inherits BytecodeGenerator class from FastAllocBase because it has been 5319 instantiated by 'new' in JavaScriptCore/parser/Nodes.cpp:1892. 5320 5321 * bytecompiler/BytecodeGenerator.h: 5322 53232009-06-24 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 5324 5325 Reviewed by Maciej Stachowiak. 5326 5327 https://bugs.webkit.org/show_bug.cgi?id=26675 5328 5329 Inherits Register class from FastAllocBase because it has been 5330 instantiated by 'new' in JavaScriptCore/runtime/JSVariableObject.h:149. 5331 5332 * interpreter/Register.h: 5333 53342009-06-24 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 5335 5336 Reviewed by Darin Adler. 5337 5338 https://bugs.webkit.org/show_bug.cgi?id=26674 5339 5340 Inherits HashMap class from FastAllocBase because it has been 5341 instantiated by 'new' in JavaScriptCore/runtime/Structure.cpp:458. 5342 5343 * wtf/HashMap.h: 5344 53452009-06-24 Oliver Hunt <oliver@apple.com> 5346 5347 Reviewed by Darin Adler. 5348 5349 <rdar://problem/6940519> REGRESSION (Safari 4 Public Beta - TOT): google.com/adplanner shows blank page instead of site details in "basic research' 5350 5351 The problem was caused by the page returned with a function using a 5352 var declaration list containing around ~3000 variables. The solution 5353 to this is to flatten the comma expression representation and make 5354 codegen comma expressions and initializer lists iterative rather than 5355 recursive. 5356 5357 * parser/Grammar.y: 5358 * parser/NodeConstructors.h: 5359 (JSC::CommaNode::CommaNode): 5360 * parser/Nodes.cpp: 5361 (JSC::CommaNode::emitBytecode): 5362 * parser/Nodes.h: 5363 (JSC::ExpressionNode::isCommaNode): 5364 (JSC::CommaNode::isCommaNode): 5365 (JSC::CommaNode::append): 5366 53672009-06-24 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 5368 5369 Reviewed by Maciej Stachowiak. 5370 5371 https://bugs.webkit.org/show_bug.cgi?id=26645 5372 5373 Inherits ScopeChainNode class from FastAllocBase because it has been 5374 instantiated by 'new' in JavaScriptCore/runtime/ScopeChain.h:95. 5375 5376 * wtf/RefPtr.h: 5377 53782009-06-24 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 5379 5380 Reviewed by Darin Adler. 5381 5382 https://bugs.webkit.org/show_bug.cgi?id=26648 5383 5384 Inherits Deque class from FastAllocBase because it has been 5385 instantiated by 'new' with DEFINE_STATIC_LOCAL macro in 5386 JavaScriptCore/wtf/MainThread.cpp:62. 5387 5388 * wtf/Deque.h: 5389 53902009-06-24 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 5391 5392 Reviewed by Darin Adler. 5393 5394 https://bugs.webkit.org/show_bug.cgi?id=26644 5395 5396 Inherits RefPtr class from FastAllocBase because it has been 5397 instantiated by 'new' in JavaScriptCore/runtime/StructureChain.cpp:41. 5398 5399 * wtf/RefPtr.h: 5400 54012009-06-24 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 5402 5403 Reviewed by Darin Adler. 5404 5405 Inherits HashSet class from FastAllocBase, because it has been 5406 instantiated by 'new' in JavaScriptCore/runtime/Collector.h:116. 5407 5408 * wtf/HashSet.h: 5409 54102009-06-24 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 5411 5412 Reviewed by Darin Adler. 5413 5414 Inherits Vector class from FastAllocBase because it has been 5415 instantiated by 'new' in JavaScriptCore/runtime/Structure.cpp:633. 5416 5417 * wtf/Vector.h: 5418 54192009-06-24 Norbert Leser <norbert.leser@nokia.com> 5420 5421 Reviewed by Maciej Stachoviak. 5422 5423 The BytecodeGenerator objects were instantiated on stack, which takes up ~38kB per instance 5424 (each instance includes copy of JSC::CodeBlock with large SymbolTable, etc.). 5425 Specifically, since there is nested invocation (e.g., GlobalCode --> FunctionCode), 5426 the stack overflows immediately on Symbian hardware (max. 80 kB). 5427 Proposed change allocates generator objects on heap. 5428 Performance impact (if any) should be negligible and change is proposed as general fix, 5429 rather than ifdef'd for SYMBIAN. 5430 5431 * parser/Nodes.cpp: 5432 (JSC::ProgramNode::generateBytecode): 5433 (JSC::EvalNode::generateBytecode): 5434 (JSC::EvalNode::bytecodeForExceptionInfoReparse): 5435 (JSC::FunctionBodyNode::generateBytecode): 5436 (JSC::FunctionBodyNode::bytecodeForExceptionInfoReparse): 5437 54382009-06-23 Oliver Hunt <oliver@apple.com> 5439 5440 Reviewed by Gavin Barraclough. 5441 5442 <rdar://problem/6992806> REGRESSION: Enumeration can skip new properties in cases of prototypes that have more than 64 (26593) 5443 <https://bugs.webkit.org/show_bug.cgi?id=26593> 5444 5445 Do not attempt to cache structure chains if they contain a dictionary at any level. 5446 5447 * interpreter/Interpreter.cpp: 5448 (JSC::Interpreter::tryCachePutByID): 5449 (JSC::Interpreter::tryCacheGetByID): 5450 * jit/JITStubs.cpp: 5451 (JSC::JITThunks::tryCachePutByID): 5452 * runtime/Structure.cpp: 5453 (JSC::Structure::getEnumerablePropertyNames): 5454 (JSC::Structure::addPropertyTransition): 5455 * runtime/StructureChain.cpp: 5456 (JSC::StructureChain::isCacheable): 5457 * runtime/StructureChain.h: 5458 54592009-06-23 Yong Li <yong.li@torchmobile.com> 5460 5461 Reviewed by George Staikos. 5462 5463 https://bugs.webkit.org/show_bug.cgi?id=26654 5464 Add the proper export define for the JavaScriptCore API when building for WINCE. 5465 5466 * API/JSBase.h: 5467 54682009-06-23 Joe Mason <joe.mason@torchmobile.com> 5469 5470 Reviewed by Adam Treat. 5471 5472 Authors: Yong Li <yong.li@torchmobile.com>, Joe Mason <joe.mason@torchmobile.com> 5473 5474 https://bugs.webkit.org/show_bug.cgi?id=26611 5475 Implement currentThreadStackBase on WINCE by adding a global, 5476 g_stackBase, which must be set to the address of a local variable 5477 by the caller before calling any WebKit function that invokes JSC. 5478 5479 * runtime/Collector.cpp: 5480 (JSC::isPageWritable): 5481 (JSC::getStackBase): 5482 Starts at the top of the stack and returns the entire range of 5483 consecutive writable pages as an estimate of the actual stack. 5484 This will be much bigger than the actual stack range, so some 5485 dead objects can't be collected, but it guarantees live objects 5486 aren't collected prematurely. 5487 5488 (JSC::currentThreadStackBase): 5489 On WinCE, returns g_stackBase if set or call getStackBase as a 5490 fallback if not. 5491 54922009-06-23 Oliver Hunt <oliver@apple.com> 5493 5494 Reviewed by Alexey Proskuryakov. 5495 5496 Fix stupid performance problem in the LiteralParser 5497 5498 The LiteralParser was making a new UString in order to use 5499 toDouble, however UString's toDouble allows a much wider range 5500 of numberic strings than the LiteralParser accepts, and requires 5501 an additional heap allocation or two for the construciton of the 5502 UString. To rectify this we just call WTF::dtoa directly using 5503 a stack allocated buffer to hold the validated numeric literal. 5504 5505 * runtime/LiteralParser.cpp: 5506 (JSC::LiteralParser::Lexer::lexNumber): 5507 (JSC::LiteralParser::parse): 5508 * runtime/LiteralParser.h: 5509 55102009-06-22 Oliver Hunt <oliver@apple.com> 5511 5512 Reviewed by Alexey Proskuryakov. 5513 5514 Bug 26640: JSON.stringify needs to special case Boolean objects 5515 <https://bugs.webkit.org/show_bug.cgi?id=26640> 5516 5517 Add special case handling of the Boolean object so we match current 5518 ES5 errata. 5519 5520 * runtime/JSONObject.cpp: 5521 (JSC::unwrapBoxedPrimitive): renamed from unwrapNumberOrString 5522 (JSC::gap): 5523 (JSC::Stringifier::appendStringifiedValue): 5524 55252009-06-22 Oliver Hunt <oliver@apple.com> 5526 5527 Reviewed by Darin Adler. 5528 5529 Bug 26591: Support revivers in JSON.parse 5530 <https://bugs.webkit.org/show_bug.cgi?id=26591> 5531 5532 Add reviver support to JSON.parse. This completes the JSON object. 5533 5534 * runtime/JSONObject.cpp: 5535 (JSC::Walker::Walker): 5536 (JSC::Walker::callReviver): 5537 (JSC::Walker::walk): 5538 (JSC::JSONProtoFuncParse): 5539 55402009-06-21 Oliver Hunt <oliver@apple.com> 5541 5542 Reviewed by Darin Adler. 5543 5544 Bug 26592: Support standard toJSON functions 5545 <https://bugs.webkit.org/show_bug.cgi?id=26592> 5546 5547 Add support for the standard Date.toJSON function. 5548 5549 * runtime/DatePrototype.cpp: 5550 (JSC::dateProtoFuncToJSON): 5551 55522009-06-21 Oliver Hunt <oliver@apple.com> 5553 5554 Reviewed by Sam Weinig. 5555 5556 Bug 26594: JSC needs to support Date.toISOString 5557 <https://bugs.webkit.org/show_bug.cgi?id=26594> 5558 5559 Add support for Date.toISOString. 5560 5561 * runtime/DatePrototype.cpp: 5562 (JSC::dateProtoFuncToISOString): 5563 55642009-06-21 Oliver Hunt <oliver@apple.com> 5565 5566 Reviewed by NOBODY (Build fix). 5567 5568 Remove dead code. 5569 5570 * runtime/LiteralParser.cpp: 5571 (JSC::LiteralParser::parse): 5572 55732009-06-21 Oliver Hunt <oliver@apple.com> 5574 5575 Reviewed by Darin Adler and Cameron Zwarich. 5576 5577 Bug 26587: Support JSON.parse 5578 <https://bugs.webkit.org/show_bug.cgi?id=26587> 5579 5580 Extend the LiteralParser to support the full strict JSON 5581 grammar, fix a few places where the grammar was incorrectly 5582 lenient. Doesn't yet support the JSON.parse reviver function 5583 but that does not block the JSON.parse functionality itself. 5584 5585 * interpreter/Interpreter.cpp: 5586 (JSC::Interpreter::callEval): 5587 * runtime/JSGlobalObjectFunctions.cpp: 5588 (JSC::globalFuncEval): 5589 * runtime/JSONObject.cpp: 5590 (JSC::JSONProtoFuncParse): 5591 * runtime/LiteralParser.cpp: 5592 (JSC::LiteralParser::Lexer::lex): 5593 (JSC::isSafeStringCharacter): 5594 (JSC::LiteralParser::Lexer::lexString): 5595 (JSC::LiteralParser::parse): 5596 * runtime/LiteralParser.h: 5597 (JSC::LiteralParser::LiteralParser): 5598 (JSC::LiteralParser::tryJSONParse): 5599 (JSC::LiteralParser::): 5600 (JSC::LiteralParser::Lexer::Lexer): 5601 56022009-06-21 David Levin <levin@chromium.org> 5603 5604 Reviewed by NOBODY (speculative build fix for windows). 5605 5606 Simply removed some whitespace form this file to make windows build wtf and 5607 hopefully copy the new MessageQueque.h so that WebCore picks it up. 5608 5609 * wtf/Assertions.cpp: 5610 56112009-06-21 Drew Wilson <atwilson@google.com> 5612 5613 Reviewed by David Levin. 5614 5615 <https://bugs.webkit.org/show_bug.cgi?id=25043> 5616 Added support for multi-threaded MessagePorts. 5617 5618 * wtf/MessageQueue.h: 5619 (WTF::::appendAndCheckEmpty): 5620 Added API to test whether the queue was empty before adding an element. 5621 56222009-06-20 David D. Kilzer <ddkilzer@webkit.org> 5623 5624 Fix namespace comment in SegmentedVector.h 5625 5626 * wtf/SegmentedVector.h: Updated namespace comment to reflect 5627 new namespace after r44897. 5628 56292009-06-20 Zoltan Herczeg <zherczeg@inf.u-szeged.hu> 5630 5631 Bug 24986: ARM JIT port 5632 <https://bugs.webkit.org/show_bug.cgi?id=24986> 5633 5634 Reviewed by Oliver Hunt. 5635 5636 An Iterator added for SegmentedVector. Currently 5637 only the pre ++ operator is supported. 5638 5639 * wtf/SegmentedVector.h: 5640 (WTF::SegmentedVectorIterator::~SegmentedVectorIterator): 5641 (WTF::SegmentedVectorIterator::operator*): 5642 (WTF::SegmentedVectorIterator::operator->): 5643 (WTF::SegmentedVectorIterator::operator++): 5644 (WTF::SegmentedVectorIterator::operator==): 5645 (WTF::SegmentedVectorIterator::operator!=): 5646 (WTF::SegmentedVectorIterator::operator=): 5647 (WTF::SegmentedVectorIterator::SegmentedVectorIterator): 5648 (WTF::SegmentedVector::alloc): 5649 (WTF::SegmentedVector::begin): 5650 (WTF::SegmentedVector::end): 5651 56522009-06-20 Zoltan Herczeg <zherczeg@inf.u-szeged.hu> 5653 5654 Bug 24986: ARM JIT port 5655 <https://bugs.webkit.org/show_bug.cgi?id=24986> 5656 5657 Reviewed by Oliver Hunt. 5658 5659 Move SegmentedVector to /wtf subdirectory 5660 and change "namespace JSC" to "namespace WTF" 5661 5662 Additional build file updates by David Kilzer. 5663 5664 * GNUmakefile.am: Updated path to SegmentedVector.h. 5665 * JavaScriptCore.order: Updated SegmentedVector namespace from 5666 JSC to WTF in mangled C++ method name. 5667 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: 5668 Removed reference to bytecompiler\SegmentedVector.h. 5669 * JavaScriptCore.vcproj/WTF/WTF.vcproj: Added reference to 5670 wtf\SegmentedVector.h. 5671 * JavaScriptCore.xcodeproj/project.pbxproj: Moved 5672 SegmentedVector.h definition from bytecompiler subdirectory to 5673 wtf subdirectory. 5674 * bytecompiler/BytecodeGenerator.h: Updated #include path to 5675 SegmentedVector.h and prepended WTF:: namespace to its use. 5676 * parser/Lexer.h: Ditto. 5677 * wtf/SegmentedVector.h: Renamed from JavaScriptCore/bytecompiler/SegmentedVector.h. 5678 (WTF::SegmentedVector::SegmentedVector): 5679 (WTF::SegmentedVector::~SegmentedVector): 5680 (WTF::SegmentedVector::size): 5681 (WTF::SegmentedVector::at): 5682 (WTF::SegmentedVector::operator[]): 5683 (WTF::SegmentedVector::last): 5684 (WTF::SegmentedVector::append): 5685 (WTF::SegmentedVector::removeLast): 5686 (WTF::SegmentedVector::grow): 5687 (WTF::SegmentedVector::clear): 5688 (WTF::SegmentedVector::deleteAllSegments): 5689 (WTF::SegmentedVector::segmentExistsFor): 5690 (WTF::SegmentedVector::segmentFor): 5691 (WTF::SegmentedVector::subscriptFor): 5692 (WTF::SegmentedVector::ensureSegmentsFor): 5693 (WTF::SegmentedVector::ensureSegment): 5694 56952009-06-19 Gavin Barraclough <barraclough@apple.com> 5696 5697 Reviewed by NOBODY (build fix take 2 - rename FIELD_OFFSET to something that doesn't conflict with winnt.h). 5698 5699 * jit/JIT.cpp: 5700 (JSC::JIT::privateCompile): 5701 (JSC::JIT::privateCompileCTIMachineTrampolines): 5702 (JSC::JIT::emitGetVariableObjectRegister): 5703 (JSC::JIT::emitPutVariableObjectRegister): 5704 * jit/JIT.h: 5705 * jit/JITArithmetic.cpp: 5706 (JSC::JIT::emit_op_rshift): 5707 (JSC::JIT::emitSlow_op_jnless): 5708 (JSC::JIT::emitSlow_op_jnlesseq): 5709 (JSC::JIT::compileBinaryArithOp): 5710 * jit/JITCall.cpp: 5711 (JSC::JIT::compileOpCallInitializeCallFrame): 5712 (JSC::JIT::compileOpCall): 5713 * jit/JITInlineMethods.h: 5714 (JSC::JIT::restoreArgumentReference): 5715 (JSC::JIT::checkStructure): 5716 * jit/JITOpcodes.cpp: 5717 (JSC::JIT::emit_op_instanceof): 5718 (JSC::JIT::emit_op_get_scoped_var): 5719 (JSC::JIT::emit_op_put_scoped_var): 5720 (JSC::JIT::emit_op_construct_verify): 5721 (JSC::JIT::emit_op_resolve_global): 5722 (JSC::JIT::emit_op_jeq_null): 5723 (JSC::JIT::emit_op_jneq_null): 5724 (JSC::JIT::emit_op_to_jsnumber): 5725 (JSC::JIT::emit_op_catch): 5726 (JSC::JIT::emit_op_eq_null): 5727 (JSC::JIT::emit_op_neq_null): 5728 (JSC::JIT::emit_op_convert_this): 5729 (JSC::JIT::emit_op_profile_will_call): 5730 (JSC::JIT::emit_op_profile_did_call): 5731 (JSC::JIT::emitSlow_op_get_by_val): 5732 * jit/JITPropertyAccess.cpp: 5733 (JSC::JIT::emit_op_get_by_val): 5734 (JSC::JIT::emit_op_put_by_val): 5735 (JSC::JIT::emit_op_method_check): 5736 (JSC::JIT::compileGetByIdHotPath): 5737 (JSC::JIT::emit_op_put_by_id): 5738 (JSC::JIT::compilePutDirectOffset): 5739 (JSC::JIT::compileGetDirectOffset): 5740 (JSC::JIT::privateCompilePutByIdTransition): 5741 (JSC::JIT::privateCompilePatchGetArrayLength): 5742 * jit/JITStubs.cpp: 5743 (JSC::JITThunks::JITThunks): 5744 57452009-06-19 Gavin Barraclough <barraclough@apple.com> 5746 5747 Reviewed by NOBODY (Windows build fix). 5748 5749 * jit/JIT.h: 5750 * jit/JITInlineMethods.h: 5751 57522009-06-19 Gabor Loki <loki@inf.u-szeged.hu> 5753 5754 Reviewed by Gavin Barraclough. 5755 5756 Reorganize ARM architecture specific macros. 5757 Use PLATFORM_ARM_ARCH(7) instead of PLATFORM(ARM_V7). 5758 5759 Bug 24986: ARM JIT port 5760 <https://bugs.webkit.org/show_bug.cgi?id=24986> 5761 5762 * assembler/ARMv7Assembler.h: 5763 * assembler/AbstractMacroAssembler.h: 5764 (JSC::AbstractMacroAssembler::Imm32::Imm32): 5765 * assembler/MacroAssembler.h: 5766 * assembler/MacroAssemblerCodeRef.h: 5767 (JSC::MacroAssemblerCodePtr::MacroAssemblerCodePtr): 5768 * jit/ExecutableAllocator.h: 5769 (JSC::ExecutableAllocator::cacheFlush): 5770 * jit/JIT.h: 5771 * jit/JITInlineMethods.h: 5772 (JSC::JIT::restoreArgumentReferenceForTrampoline): 5773 * jit/JITStubs.cpp: 5774 * jit/JITStubs.h: 5775 * wtf/Platform.h: 5776 * yarr/RegexJIT.cpp: 5777 (JSC::Yarr::RegexGenerator::generateEnter): 5778 (JSC::Yarr::RegexGenerator::generateReturn): 5779 57802009-06-19 Gavin Barraclough <barraclough@apple.com> 5781 5782 Reviewed by Oliver Hunt. 5783 5784 Fix armv7 JIT build issues. 5785 5786 Unfortunate the arm compiler does not like the use of offsetof on JITStackFrame (since it now contains non POD types), 5787 and the FIELD_OFFSET macro does not appear constantish enough for it to be happy with its use in COMPILE_ASSERT macros. 5788 5789 * Replace offsetofs with FIELD_OFFSETs (safe on C++ objects). 5790 * Move COMPILE_ASSERTs defending layout of JITStackFrame structure on armv7 into JITThunks constructor. 5791 5792 * jit/JIT.cpp: 5793 * jit/JIT.h: 5794 * jit/JITInlineMethods.h: 5795 (JSC::JIT::restoreArgumentReference): 5796 * jit/JITOpcodes.cpp: 5797 (JSC::JIT::emit_op_catch): 5798 * jit/JITStubs.cpp: 5799 (JSC::JITThunks::JITThunks): 5800 58012009-06-19 Adam Treat <adam.treat@torchmobile.com> 5802 5803 Blind attempt at build fix. 5804 5805 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 5806 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 5807 58082009-06-19 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 5809 5810 Reviewed by Oliver Hunt. 5811 5812 Inherits CallIdentifier struct from FastAllocBase because it has been 5813 instantiated by 'new' in JavaScriptCore/profiler/CallIdentifier.h:86. 5814 5815 * wtf/HashCountedSet.h: 5816 58172009-06-19 Adam Treat <adam.treat@torchmobile.com> 5818 5819 Reviewed by Oliver Hunt. 5820 5821 https://bugs.webkit.org/show_bug.cgi?id=26540 5822 Modify the test shell to add a new function 'checkSyntax' that will 5823 only parse the source instead of executing it. In this way we can test 5824 pure parsing performance against some of the larger scripts in the wild. 5825 5826 * jsc.cpp: 5827 (GlobalObject::GlobalObject): 5828 (functionCheckSyntax): 5829 58302009-06-19 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 5831 5832 Reviewed by Darin Adler. 5833 5834 Inherits HashCountedSet class from FastAllocBase because it has been 5835 instantiated by 'new' in JavaScriptCore/runtime/Collector.cpp:1095. 5836 5837 * wtf/HashCountedSet.h: 5838 58392009-06-19 Yong Li <yong.li@torchmobile.com> 5840 5841 Reviewed by George Staikos. 5842 5843 https://bugs.webkit.org/show_bug.cgi?id=26558 5844 Declare these symbols extern for WINCE as they are provided by libce. 5845 5846 * runtime/DateConstructor.cpp: 5847 * runtime/DatePrototype.cpp: 5848 (JSC::formatLocaleDate): 5849 58502009-06-19 Oliver Hunt <oliver@apple.com> 5851 5852 Reviewed by Maciej Stachowiak. 5853 5854 <rdar://problem/6988973> ScopeChain leak in interpreter builds 5855 5856 Move the Scopechain destruction code in JSFunction outside of the ENABLE(JIT) 5857 path. 5858 5859 * runtime/JSFunction.cpp: 5860 (JSC::JSFunction::~JSFunction): 5861 * wtf/Platform.h: 5862 58632009-06-19 Yong Li <yong.li@torchmobile.com> 5864 5865 Reviewed by George Staikos. 5866 5867 https://bugs.webkit.org/show_bug.cgi?id=26543 5868 Windows CE uses 'GetLastError' instead of 'errno.' 5869 5870 * interpreter/RegisterFile.h: 5871 (JSC::RegisterFile::RegisterFile): 5872 (JSC::RegisterFile::grow): 5873 58742009-06-19 David Levin <levin@chromium.org> 5875 5876 Reviewed by NOBODY (Windows build fix). 5877 5878 Add export for Windows corresponding to OSX export done in r44844. 5879 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 5880 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 5881 58822009-06-18 Oliver Hunt <oliver@apple.com> 5883 5884 Reviewed by Gavin "Viceroy of Venezuela" Barraclough. 5885 5886 Bug 26532: Native functions do not correctly unlink from optimised callsites when they're collected 5887 <https://bugs.webkit.org/show_bug.cgi?id=26532> <rdar://problem/6625385> 5888 5889 We need to make sure that each native function instance correctly unlinks any references to it 5890 when it is collected. Allowing this to happen required a few changes: 5891 * Every native function needs a codeblock to track the link information 5892 * To have this codeblock, every function now also needs its own functionbodynode 5893 so we no longer get to have a single shared instance. 5894 * Identifying a host function is now done by looking for CodeBlock::codeType() == NativeCode 5895 5896 * JavaScriptCore.exp: 5897 * bytecode/CodeBlock.cpp: 5898 (JSC::CodeBlock::CodeBlock): 5899 Constructor for NativeCode CodeBlock 5900 (JSC::CodeBlock::derefStructures): 5901 (JSC::CodeBlock::refStructures): 5902 (JSC::CodeBlock::reparseForExceptionInfoIfNecessary): 5903 (JSC::CodeBlock::handlerForBytecodeOffset): 5904 (JSC::CodeBlock::lineNumberForBytecodeOffset): 5905 (JSC::CodeBlock::expressionRangeForBytecodeOffset): 5906 (JSC::CodeBlock::getByIdExceptionInfoForBytecodeOffset): 5907 (JSC::CodeBlock::functionRegisterForBytecodeOffset): 5908 (JSC::CodeBlock::hasGlobalResolveInstructionAtBytecodeOffset): 5909 (JSC::CodeBlock::hasGlobalResolveInfoAtBytecodeOffset): 5910 (JSC::CodeBlock::setJITCode): 5911 Add assertions to ensure we don't try and use NativeCode CodeBlocks as 5912 a normal codeblock. 5913 5914 * bytecode/CodeBlock.h: 5915 (JSC::): 5916 (JSC::CodeBlock::source): 5917 (JSC::CodeBlock::sourceOffset): 5918 (JSC::CodeBlock::evalCodeCache): 5919 (JSC::CodeBlock::createRareDataIfNecessary): 5920 More assertions. 5921 5922 * jit/JIT.cpp: 5923 (JSC::JIT::privateCompileCTIMachineTrampolines): 5924 (JSC::JIT::linkCall): 5925 Update logic to allow native function caching 5926 5927 * jit/JITStubs.cpp: 5928 * parser/Nodes.cpp: 5929 (JSC::FunctionBodyNode::createNativeThunk): 5930 (JSC::FunctionBodyNode::isHostFunction): 5931 * parser/Nodes.h: 5932 * runtime/JSFunction.cpp: 5933 (JSC::JSFunction::JSFunction): 5934 (JSC::JSFunction::~JSFunction): 5935 (JSC::JSFunction::mark): 5936 * runtime/JSGlobalData.cpp: 5937 (JSC::JSGlobalData::~JSGlobalData): 5938 * runtime/JSGlobalData.h: 5939 59402009-06-18 Gavin Barraclough <barraclough@apple.com> 5941 5942 Reviewed by NOBODY (Windows build fix). 5943 5944 * wtf/DateMath.cpp: 5945 (WTF::calculateUTCOffset): 5946 59472009-06-18 Gavin Barraclough <barraclough@apple.com> 5948 5949 Reviewed by Geoff Garen. 5950 5951 Timezone calculation incorrect in Venezuela. 5952 5953 https://bugs.webkit.org/show_bug.cgi?id=26531 5954 <rdar://problem/6646169> Time is incorrectly reported to JavaScript in both Safari 3 and Firefox 3 5955 5956 The problem is that we're calculating the timezone relative to 01/01/2000, 5957 but the VET timezone changed from -4 hours to -4:30 hours on 12/09/2007. 5958 According to the spec, section 15.9.1.9 states "the time since the beginning 5959 of the year", presumably meaning the *current* year. Change the calculation 5960 to be based on whatever the current year is, rather than a canned date. 5961 5962 No performance impact. 5963 5964 * wtf/DateMath.cpp: 5965 (WTF::calculateUTCOffset): 5966 59672009-06-18 Gavin Barraclough <barraclough@apple.com> 5968 5969 Rubber Stamped by Mark Rowe (originally reviewed by Sam Weinig). 5970 5971 (Reintroducing patch added in r44492, and reverted in r44796.) 5972 5973 Change the implementation of op_throw so the stub function always modifies its 5974 return address - if it doesn't find a 'catch' it will switch to a trampoline 5975 to force a return from JIT execution. This saves memory, by avoiding the need 5976 for a unique return for every op_throw. 5977 5978 * jit/JITOpcodes.cpp: 5979 (JSC::JIT::emit_op_throw): 5980 JITStubs::cti_op_throw now always changes its return address, 5981 remove return code generated after the stub call (this is now 5982 handled by ctiOpThrowNotCaught). 5983 * jit/JITStubs.cpp: 5984 (JSC::): 5985 Add ctiOpThrowNotCaught definitions. 5986 (JSC::JITStubs::DEFINE_STUB_FUNCTION): 5987 Change cti_op_throw to always change its return address. 5988 * jit/JITStubs.h: 5989 Add ctiOpThrowNotCaught declaration. 5990 59912009-06-18 Kevin McCullough <kmccullough@apple.com> 5992 5993 Reviewed by Oliver Hunt. 5994 5995 <rdar://problem/6940880> REGRESSION: Breakpoints don't break in 64-bit 5996 5997 - Exposed functions now needed by WebCore. 5998 5999 * JavaScriptCore.exp: 6000 60012009-06-17 Darin Adler <darin@apple.com> 6002 6003 Reviewed by Oliver Hunt. 6004 6005 Bug 26429: Make JSON.stringify non-recursive so it can handle objects 6006 of arbitrary complexity 6007 https://bugs.webkit.org/show_bug.cgi?id=26429 6008 6009 For marking I decided not to use gcProtect, because this is inside the engine 6010 so it's easy enough to just do marking. And that darned gcProtect does locking! 6011 Oliver tried to convince me to used MarkedArgumentBuffer, but the constructor 6012 for that class says "FIXME: Remove all clients of this API, then remove this API." 6013 6014 * runtime/Collector.cpp: 6015 (JSC::Heap::collect): Add a call to JSONObject::markStringifiers. 6016 6017 * runtime/CommonIdentifiers.cpp: 6018 (JSC::CommonIdentifiers::CommonIdentifiers): Added emptyIdentifier. 6019 * runtime/CommonIdentifiers.h: Ditto. 6020 6021 * runtime/JSGlobalData.cpp: 6022 (JSC::JSGlobalData::JSGlobalData): Initialize firstStringifierToMark to 0. 6023 * runtime/JSGlobalData.h: Added firstStringifierToMark. 6024 6025 * runtime/JSONObject.cpp: Cut down the includes to the needed ones only. 6026 (JSC::unwrapNumberOrString): Added. Helper for unwrapping number and string 6027 objects to get their number and string values. 6028 (JSC::ReplacerPropertyName::ReplacerPropertyName): Added. The class is used 6029 to wrap an identifier or integer so we don't have to do any work unless we 6030 actually call a replacer. 6031 (JSC::ReplacerPropertyName::value): Added. 6032 (JSC::gap): Added. Helper function for the Stringifier constructor. 6033 (JSC::PropertyNameForFunctionCall::PropertyNameForFunctionCall): Added. 6034 The class is used to wrap an identifier or integer so we don't have to 6035 allocate a number or string until we actually call toJSON or a replacer. 6036 (JSC::PropertyNameForFunctionCall::asJSValue): Added. 6037 (JSC::Stringifier::Stringifier): Updated and moved out of the class 6038 definition. Added code to hook this into a singly linked list for marking. 6039 (JSC::Stringifier::~Stringifier): Remove from the singly linked list. 6040 (JSC::Stringifier::mark): Mark all the objects in the holder stacks. 6041 (JSC::Stringifier::stringify): Updated. 6042 (JSC::Stringifier::appendQuotedString): Tweaked and streamlined a bit. 6043 (JSC::Stringifier::toJSON): Renamed from toJSONValue. 6044 (JSC::Stringifier::appendStringifiedValue): Renamed from stringify. 6045 Added code to use the m_holderStack to do non-recursive stringify of 6046 objects and arrays. This code also uses the timeout checker since in 6047 pathological cases it could be slow even without calling into the 6048 JavaScript virtual machine. 6049 (JSC::Stringifier::willIndent): Added. 6050 (JSC::Stringifier::indent): Added. 6051 (JSC::Stringifier::unindent): Added. 6052 (JSC::Stringifier::startNewLine): Added. 6053 (JSC::Stringifier::Holder::Holder): Added. 6054 (JSC::Stringifier::Holder::appendNextProperty): Added. This is the 6055 function that handles the format of arrays and objects. 6056 (JSC::JSONObject::getOwnPropertySlot): Moved this down to the bottom 6057 of the file so the JSONObject class is not interleaved with the 6058 Stringifier class. 6059 (JSC::JSONObject::markStringifiers): Added. Calls mark. 6060 (JSC::JSONProtoFuncStringify): Streamlined the code here. The code 6061 to compute the gap string is now a separate function. 6062 6063 * runtime/JSONObject.h: Made everything private. Added markStringifiers. 6064 60652009-06-17 Oliver Hunt <oliver@apple.com> 6066 6067 Reviewed by Gavin Barraclough. 6068 6069 <rdar://problem/6974140> REGRESSION(r43849): Crash in cti_op_call_NotJSFunction when getting directions on maps.google.com 6070 6071 Roll out r43849 as it appears that we cannot rely on the address of 6072 an objects property storage being constant even if the structure is 6073 unchanged. 6074 6075 * jit/JIT.h: 6076 * jit/JITPropertyAccess.cpp: 6077 (JSC::JIT::compileGetDirectOffset): 6078 (JSC::JIT::privateCompileGetByIdProto): 6079 (JSC::JIT::privateCompileGetByIdProtoList): 6080 (JSC::JIT::privateCompileGetByIdChainList): 6081 (JSC::JIT::privateCompileGetByIdChain): 6082 60832009-06-17 Gavin Barraclough <barraclough@apple.com> 6084 6085 Rubber Stamped by Mark Rowe. 6086 6087 Fully revert r44492 & r44748 while we fix a bug they cause on internal builds <rdar://problem/6955963>. 6088 6089 * jit/JITOpcodes.cpp: 6090 (JSC::JIT::emit_op_throw): 6091 * jit/JITStubs.cpp: 6092 (JSC::): 6093 (JSC::JITStubs::DEFINE_STUB_FUNCTION): 6094 * jit/JITStubs.h: 6095 60962009-06-17 Gavin Barraclough <barraclough@apple.com> 6097 6098 Reviewed by Mark Rowe. 6099 6100 <rdar://problem/6947426> sunspider math-cordic.js exhibits different intermediate results running 32-bit vs. 64-bit 6101 6102 On 64-bit, NaN-encoded values must be detagged before they can be used in rshift. 6103 6104 No performance impact. 6105 6106 * jit/JITArithmetic.cpp: 6107 (JSC::JIT::emit_op_rshift): 6108 61092009-06-17 Adam Treat <adam.treat@torchmobile.com> 6110 6111 Reviewed by George Staikos. 6112 6113 https://bugs.webkit.org/show_bug.cgi?id=23155 6114 Move WIN_CE -> WINCE as previously discussed with Qt WINCE folks. 6115 6116 * jsc.cpp: 6117 (main): 6118 61192009-06-17 George Staikos <george.staikos@torchmobile.com> 6120 6121 Reviewed by Adam Treat. 6122 6123 https://bugs.webkit.org/show_bug.cgi?id=23155 6124 Move WIN_CE -> WINCE as previously discussed with Qt WINCE folks. 6125 6126 * config.h: 6127 * jsc.cpp: 6128 * wtf/Assertions.cpp: 6129 * wtf/Assertions.h: 6130 * wtf/CurrentTime.cpp: 6131 (WTF::lowResUTCTime): 6132 * wtf/DateMath.cpp: 6133 (WTF::getLocalTime): 6134 * wtf/MathExtras.h: 6135 * wtf/Platform.h: 6136 * wtf/StringExtras.h: 6137 * wtf/Threading.h: 6138 * wtf/win/MainThreadWin.cpp: 6139 61402009-06-17 Gavin Barraclough <barraclough@apple.com> 6141 6142 Reviewed by Oliver Hunt. 6143 6144 <rdar://problem/6974175> ASSERT in JITStubs.cpp at appsaccess.apple.com 6145 6146 Remove PropertySlot::putValue - PropertySlots should only be used for getting, 6147 not putting. Rename JSGlobalObject::getOwnPropertySlot to hasOwnPropertyForWrite, 6148 which is what it really was being used to ask, and remove some other getOwnPropertySlot 6149 & getOwnPropertySlotForWrite methods, which were unused and likely to lead to confusion. 6150 6151 * runtime/JSGlobalObject.h: 6152 (JSC::JSGlobalObject::hasOwnPropertyForWrite): 6153 * runtime/JSObject.h: 6154 * runtime/JSStaticScopeObject.cpp: 6155 * runtime/JSStaticScopeObject.h: 6156 * runtime/PropertySlot.h: 6157 61582009-06-16 Gavin Barraclough <barraclough@apple.com> 6159 6160 Reviewed by Oliver hunt. 6161 6162 Temporarily partially disable r44492, since this is causing some problems on internal builds. 6163 6164 * jit/JITOpcodes.cpp: 6165 (JSC::JIT::emit_op_throw): 6166 * jit/JITStubs.cpp: 6167 (JSC::JITStubs::DEFINE_STUB_FUNCTION): 6168 61692009-06-16 Sam Weinig <sam@webkit.org> 6170 6171 Fix windows build. 6172 6173 * jit/JIT.cpp: 6174 (JSC::JIT::JIT): 6175 61762009-06-16 Sam Weinig <sam@webkit.org> 6177 6178 Reviewed by Oliver Hunt. 6179 6180 Initialize m_bytecodeIndex to -1 in JIT, and correctly initialize 6181 it for each type of stub using the return address to find the correct 6182 offset. 6183 6184 * jit/JIT.cpp: 6185 (JSC::JIT::JIT): 6186 * jit/JIT.h: 6187 (JSC::JIT::compileGetByIdProto): 6188 (JSC::JIT::compileGetByIdSelfList): 6189 (JSC::JIT::compileGetByIdProtoList): 6190 (JSC::JIT::compileGetByIdChainList): 6191 (JSC::JIT::compileGetByIdChain): 6192 (JSC::JIT::compilePutByIdTransition): 6193 (JSC::JIT::compileCTIMachineTrampolines): 6194 (JSC::JIT::compilePatchGetArrayLength): 6195 * jit/JITStubCall.h: 6196 (JSC::JITStubCall::call): 6197 6198== Rolled over to ChangeLog-2009-06-16 == 6199