12010-02-12 Janne Koskinen <janne.p.koskinen@digia.com> 2 3 Reviewed by Tor Arne Vestbø. 4 5 Additional refptr/passrefptr workarounds for WINSCW compiler 6 https://bugs.webkit.org/show_bug.cgi?id=28054 7 8 * wtf/PassRefPtr.h: 9 (WTF::refIfNotNull): 10 (WTF::PassRefPtr::PassRefPtr): 11 (WTF::PassRefPtr::~PassRefPtr): 12 (WTF::PassRefPtr::clear): 13 (WTF::::operator): 14 * wtf/RefPtr.h: 15 (WTF::RefPtr::RefPtr): 16 (WTF::::operator): 17 182010-02-12 Janne Koskinen <janne.p.koskinen@digia.com> 19 20 Reviewed by Simon Hausmann. 21 22 Don't import the cmath functions from std:: for WINSCW. 23 24 * wtf/MathExtras.h: 25 262010-02-12 Kwang Yul Seo <skyul@company100.net> 27 28 Reviewed by Adam Barth. 29 30 Typedef both JSChar and UChar to wchar_t in RVCT. 31 https://bugs.webkit.org/show_bug.cgi?id=34560 32 33 Define both JSChar and UChar to wchar_t as the size 34 of wchar_t is 2 bytes in RVCT. 35 36 * API/JSStringRef.h: 37 * wtf/unicode/qt4/UnicodeQt4.h: 38 392010-02-11 Geoffrey Garen <ggaren@apple.com> 40 41 Reviewed by Oliver Hunt and Darin Adler. 42 43 The rest of the fix for 44 https://bugs.webkit.org/show_bug.cgi?id=34864 | <rdar://problem/7594198> 45 Many objects left uncollected after visiting mail.google.com and closing 46 window 47 48 Don't unconditionally hang onto small strings. Instead, hang onto all 49 small strings as long as any small string is still referenced. 50 51 SunSpider reports no change. 52 53 * runtime/Collector.cpp: 54 (JSC::Heap::markRoots): Mark the small strings cache last, so it can 55 check if anything else has kept any strings alive. 56 57 * runtime/SmallStrings.cpp: 58 (JSC::isMarked): 59 (JSC::SmallStrings::markChildren): Only keep our strings alive if some 60 other reference to at least one of them exists, too. 61 622010-02-11 Geoffrey Garen <ggaren@apple.com> 63 64 Reviewed by Gavin Barraclough. 65 66 Some progress toward fixing 67 https://bugs.webkit.org/show_bug.cgi?id=34864 | <rdar://problem/7594198> 68 Many objects left uncollected after visiting mail.google.com and closing 69 window 70 71 SunSpider reports no change. 72 73 Keep weak references, rather than protected references, to cached for-in 74 property name enumerators. 75 76 One problem with protected references is that a chain like 77 [ gc object 1 ] => [ non-gc object ] => [ gc object 2 ] 78 takes two GC passes to break, since the first pass collects [ gc object 1 ], 79 releasing [ non-gc object ] and unprotecting [ gc object 2 ], and only 80 then can a second pass collect [ gc object 2 ]. 81 82 Another problem with protected references is that they can keep a bunch 83 of strings alive long after they're useful. In SunSpider and a few popular 84 websites, the size-speed tradeoff seems to favor weak references. 85 86 * runtime/JSPropertyNameIterator.cpp: 87 (JSC::JSPropertyNameIterator::JSPropertyNameIterator): Moved this constructor 88 into the .cpp file, since it's not used elsewhere. 89 90 (JSC::JSPropertyNameIterator::~JSPropertyNameIterator): Added a destructor 91 to support our weak reference. 92 93 * runtime/JSPropertyNameIterator.h: 94 (JSC::Structure::setEnumerationCache): 95 (JSC::Structure::clearEnumerationCache): 96 (JSC::Structure::enumerationCache): Added a function for clearing a 97 Structure's enumeration cache, used by our new destructor. Also fixed 98 indentation to match the rest of the file. 99 100 * runtime/Structure.h: Changed from protected pointer to weak pointer. 101 1022010-02-11 Chris Rogers <crogers@google.com> 103 104 Reviewed by David Levin. 105 106 audio engine: add Complex number class 107 https://bugs.webkit.org/show_bug.cgi?id=34538 108 109 * wtf/Complex.h: Added. 110 (WebCore::complexFromMagnitudePhase): 111 1122010-02-10 Geoffrey Garen <ggaren@apple.com> 113 114 Reviewed by Oliver Hunt. 115 116 Added an SPI for asking about all the different live objects on the heap. 117 Useful for memory debugging. 118 119 * JavaScriptCore.exp: Export the new SPI. 120 121 * runtime/Collector.cpp: 122 (JSC::typeName): Use a little capitalization. Don't crash in the case of 123 a non-object cell, since it might just be an uninitialized cell. 124 125 (JSC::Heap::objectTypeCounts): The new SPI. 126 127 * runtime/Collector.h: 128 * runtime/CollectorHeapIterator.h: 129 (JSC::CollectorHeapIterator::advance): 130 (JSC::LiveObjectIterator::operator++): 131 (JSC::DeadObjectIterator::operator++): 132 (JSC::ObjectIterator::operator++): Made 2 tweaks to these iterators: 133 (1) Skip the last cell in the block, since it's a dummy sentinel, and 134 we don't want it to confuse the object count; (2) Fixed a logic error 135 in LiveObjectIterator that could cause it to iterate dead objects if 136 m_block were equal to m_heap.nextBlock and m_cell were less than 137 m_heap.nextCell. No test for this since I can't think of a way that this 138 could make WebKit behave badly. 139 1402010-02-11 Steve Block <steveblock@google.com> 141 142 Reviewed by Darin Adler. 143 144 Guard cmath using declarations in MathExtras.h on Android 145 https://bugs.webkit.org/show_bug.cgi?id=34840 146 147 Android does not provide these functions. 148 149 * wtf/MathExtras.h: 150 1512010-02-08 Maciej Stachowiak <mjs@apple.com> 152 153 Reviewed by Cameron Zwarich. 154 155 Restore ENABLE_RUBY flag so vendors can ship with Ruby disabled if they choose. 156 https://bugs.webkit.org/show_bug.cgi?id=34698 157 158 * Configurations/FeatureDefines.xcconfig: 159 1602010-02-10 Kevin Watters <kevinwatters@gmail.com> 161 162 Reviewed by Kevin Ollivier. 163 164 [wx] Add Windows complex text support and Mac support for containsCharacters. 165 166 https://bugs.webkit.org/show_bug.cgi?id=34759 167 168 * wscript: 169 1702010-02-10 Alexey Proskuryakov <ap@apple.com> 171 172 Addressing issues found by style bot. 173 174 * wtf/ValueCheck.h: Renamed header guard to match final file name. 175 176 * wtf/Vector.h: (WTF::::checkConsistency): Remove braces around a one-line clause. 177 1782010-02-09 Alexey Proskuryakov <ap@apple.com> 179 180 Reviewed by Geoffrey Garen. 181 182 https://bugs.webkit.org/show_bug.cgi?id=34490 183 WebCore::ImageEventSender::dispatchPendingEvents() crashes in certain conditions 184 185 * GNUmakefile.am: 186 * JavaScriptCore.gypi: 187 * JavaScriptCore.vcproj/WTF/WTF.vcproj: 188 * JavaScriptCore.xcodeproj/project.pbxproj: 189 Added ValueCheck.h. 190 191 * wtf/ValueCheck.h: Added. Moved code out of HashTraits, since it would be awkward to 192 include that from Vector.h. 193 (WTF::ValueCheck::checkConsistency): Allow null pointers, those are pretty consistent. 194 195 * wtf/HashTraits.h: Moved value checking code out of here. 196 197 * wtf/HashTable.h: (WTF::::checkTableConsistencyExceptSize): Updated for the above changes. 198 199 * wtf/Vector.h: 200 (WTF::::checkConsistency): Check all vector elements. 201 (WTF::ValueCheck): Support checking a Vector as an element in other containers. Currently 202 unused. 203 2042010-02-10 Jedrzej Nowacki <jedrzej.nowacki@nokia.com> 205 206 Reviewed by Simon Hausmann. 207 208 Fix QScriptValue::toBool. 209 210 Fix ECMA compliance in the QScriptValue for values like 0, NaN and 211 empty strings. 212 213 [Qt] QScriptValue::toBool problem 214 https://bugs.webkit.org/show_bug.cgi?id=34793 215 216 * qt/api/qscriptvalue_p.h: 217 (QScriptValuePrivate::toBool): 218 * qt/tests/qscriptvalue/tst_qscriptvalue.h: 219 * qt/tests/qscriptvalue/tst_qscriptvalue_generated.cpp: 220 (tst_QScriptValue::toBool_initData): 221 (tst_QScriptValue::toBool_makeData): 222 (tst_QScriptValue::toBool_test): 223 (tst_QScriptValue::toBoolean_initData): 224 (tst_QScriptValue::toBoolean_makeData): 225 (tst_QScriptValue::toBoolean_test): 226 2272009-10-06 Yongjun Zhang <yongjun.zhang@nokia.com> 228 229 Reviewed by Simon Hausmann. 230 231 Use derefIfNotNull() to work around WINSCW compiler forward declaration bug 232 233 The compiler bug is reported at 234 https://xdabug001.ext.nokia.com/bugzilla/show_bug.cgi?id=9812. 235 236 The change should be reverted when the above bug is fixed in WINSCW compiler. 237 238 https://bugs.webkit.org/show_bug.cgi?id=28054 239 2402009-10-06 Yongjun Zhang <yongjun.zhang@nokia.com> 241 242 Reviewed by Simon Hausmann. 243 244 Get rid of WINSCW hack for UnSpecifiedBoolType 245 246 Add parenthesis around (RefPtr::*UnspecifiedBoolType) to make the WINSCW 247 compiler work with the default UnSpecifiedBoolType() operator. 248 249 https://bugs.webkit.org/show_bug.cgi?id=28054 250 251 * wtf/RefPtr.h: 252 2532010-02-09 Jedrzej Nowacki <jedrzej.nowacki@nokia.com> 254 255 Reviewed by Simon Hausmann. 256 257 New functions nullValue() and undefinedValue(). 258 259 [Qt] QScriptEngine should contain nullValue and undefinedValue methods 260 https://bugs.webkit.org/show_bug.cgi?id=34749 261 262 * qt/api/qscriptengine.cpp: 263 (QScriptEngine::nullValue): 264 (QScriptEngine::undefinedValue): 265 * qt/api/qscriptengine.h: 266 * qt/tests/qscriptengine/tst_qscriptengine.cpp: 267 (tst_QScriptEngine::nullValue): 268 (tst_QScriptEngine::undefinedValue): 269 2702010-02-09 Jedrzej Nowacki <jedrzej.nowacki@nokia.com> 271 272 Reviewed by Simon Hausmann. 273 274 Fixes for QScriptValue::toNumber(). 275 276 Fix ECMA compliance in QScriptValue for values unbound 277 to a QScriptEngine. 278 279 [Qt] QScriptValue::toNumber() is broken 280 https://bugs.webkit.org/show_bug.cgi?id=34592 281 282 * qt/api/qscriptvalue_p.h: 283 (QScriptValuePrivate::toNumber): 284 * qt/tests/qscriptvalue/tst_qscriptvalue.h: 285 * qt/tests/qscriptvalue/tst_qscriptvalue_generated.cpp: 286 (tst_QScriptValue::toNumber_initData): 287 (tst_QScriptValue::toNumber_makeData): 288 (tst_QScriptValue::toNumber_test): 289 2902010-02-09 Jedrzej Nowacki <jedrzej.nowacki@nokia.com> 291 292 Reviewed by Simon Hausmann. 293 294 Fix QScriptValue::isNumber(). 295 296 The isNumber() should return 'true' if the value is in the CNumber 297 state. 298 299 [Qt] QScriptValue::isNumber() returns an incorrect value 300 https://bugs.webkit.org/show_bug.cgi?id=34575 301 302 * qt/api/qscriptvalue_p.h: 303 (QScriptValuePrivate::isNumber): 304 * qt/tests/qscriptvalue/tst_qscriptvalue.h: 305 * qt/tests/qscriptvalue/tst_qscriptvalue_generated.cpp: 306 (tst_QScriptValue::isNumber_initData): 307 (tst_QScriptValue::isNumber_makeData): 308 (tst_QScriptValue::isNumber_test): 309 3102010-02-09 Geoffrey Garen <ggaren@apple.com> 311 312 Reviewed by Oliver Hunt. 313 314 Small refactoring to the small strings cache to allow it to be cleared 315 dynamically. 316 317 * runtime/SmallStrings.cpp: 318 (JSC::SmallStrings::SmallStrings): 319 (JSC::SmallStrings::clear): 320 * runtime/SmallStrings.h: Moved initialization code into a shared function, 321 and changed the constructor to call it. 322 3232010-02-09 Gavin Barraclough <barraclough@apple.com> 324 325 Rubber Stamped by Geoff Garen. 326 327 Rename StringBuilder::release && JSStringBuilder::releaseJSString 328 to 'build()'. 329 330 * runtime/ArrayPrototype.cpp: 331 (JSC::arrayProtoFuncToLocaleString): 332 (JSC::arrayProtoFuncJoin): 333 * runtime/Executable.cpp: 334 (JSC::FunctionExecutable::paramString): 335 * runtime/FunctionConstructor.cpp: 336 (JSC::constructFunction): 337 * runtime/JSGlobalObjectFunctions.cpp: 338 (JSC::encode): 339 (JSC::decode): 340 (JSC::globalFuncEscape): 341 (JSC::globalFuncUnescape): 342 * runtime/JSONObject.cpp: 343 (JSC::Stringifier::stringify): 344 * runtime/JSStringBuilder.h: 345 (JSC::JSStringBuilder::build): 346 * runtime/LiteralParser.cpp: 347 (JSC::LiteralParser::Lexer::lexString): 348 * runtime/NumberPrototype.cpp: 349 (JSC::integerPartNoExp): 350 (JSC::numberProtoFuncToFixed): 351 * runtime/StringBuilder.h: 352 (JSC::StringBuilder::build): 353 3542010-02-09 John Sullivan <sullivan@apple.com> 355 356 https://bugs.webkit.org/show_bug.cgi?id=34772 357 Overzealous new assertion in URStringImpl::adopt() 358 359 Reviewed by Adam Barth. 360 361 * runtime/UStringImpl.h: 362 (JSC::UStringImpl::adopt): 363 Only assert that vector.data() is non-zero if vector.size() is non-zero. 364 3652010-02-09 Nikolas Zimmermann <nzimmermann@rim.com> 366 367 Not reviewed. Try to fix build problem on SnowLeopard slaves to bring them back. 368 369 * API/JSClassRef.cpp: 370 (tryCreateStringFromUTF8): Mark method as 'static inline' to suppress "warning: no previous prototype for ..." 371 3722010-02-09 Gavin Barraclough <barraclough@apple.com> 373 374 Reviewed by Oliver Hunt. 375 376 Three small string fixes: 377 (1) StringBuilder::release should CRASH if the buffer allocation failed. 378 (2) Remove weird, dead code from JSString::tryGetValue, replace with an ASSERT. 379 (3) Move UString::createFromUTF8 out to the API, as tryCreateStringFromUTF8. 380 This is only used from the API, and (now) unlike other UString::create 381 methods may return UString::null() to indicate failure cases. Better 382 handle these in the API. 383 384 * API/JSClassRef.cpp: 385 (tryCreateStringFromUTF8): 386 (OpaqueJSClass::OpaqueJSClass): 387 (OpaqueJSClassContextData::OpaqueJSClassContextData): 388 * runtime/JSString.h: 389 (JSC::Fiber::tryGetValue): 390 * runtime/StringBuilder.h: 391 (JSC::StringBuilder::release): 392 * runtime/UString.cpp: 393 (JSC::UString::UString): 394 (JSC::UString::from): 395 (JSC::UString::find): 396 * runtime/UString.h: 397 3982010-02-09 Janne Koskinen <janne.p.koskinen@digia.com> 399 400 Reviewed by Laszlo Gombos. 401 402 [Qt] use nanval() for Symbian as nonInlineNaN 403 https://bugs.webkit.org/show_bug.cgi?id=34170 404 405 numeric_limits<double>::quiet_NaN is broken in Symbian 406 causing NaN to be evaluated as a number. 407 408 * runtime/JSValue.cpp: 409 (JSC::nonInlineNaN): 410 4112010-02-09 Tamas Szirbucz <szirbucz@inf.u-szeged.hu> 412 413 Reviewed by Gavin Barraclough. 414 415 Add a soft modulo operation to ARM JIT using a trampoline function. 416 The performance progression is about ~1.8% on ARMv7 417 https://bugs.webkit.org/show_bug.cgi?id=34424 418 419 Developed in cooperation with Gabor Loki. 420 421 * jit/JIT.h: 422 * jit/JITArithmetic.cpp: 423 (JSC::JIT::emit_op_mod): 424 (JSC::JIT::emitSlow_op_mod): 425 * jit/JITOpcodes.cpp: 426 (JSC::JIT::softModulo): 427 * jit/JITStubs.h: 428 (JSC::JITThunks::ctiSoftModulo): 429 * wtf/Platform.h: 430 4312010-02-08 Gavin Barraclough <barraclough@apple.com> 432 433 Reviewed by NOBODY (SL/win build fixes). 434 435 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 436 * runtime/StringPrototype.cpp: 437 4382010-02-08 Gavin Barraclough <barraclough@apple.com> 439 440 Reviewed by Oliver Hunt 441 442 Make String.replace throw an exception on out-of-memory, rather than 443 returning a null (err, empty-ish) string. Move String::replaceRange 444 and String::spliceSubstringsWithSeparators out to StringPrototype - 445 these were fairly specific use anyway, and we can better integrate 446 throwing the JS expcetion this way. 447 448 Also removes redundant assignment operator from UString. 449 450 * JavaScriptCore.exp: 451 * runtime/StringPrototype.cpp: 452 (JSC::StringRange::StringRange): 453 (JSC::jsSpliceSubstringsWithSeparators): 454 (JSC::jsReplaceRange): 455 (JSC::stringProtoFuncReplace): 456 * runtime/UString.cpp: 457 * runtime/UString.h: 458 4592010-02-08 Kwang Yul Seo <skyul@company100.net> 460 461 Reviewed by Eric Seidel. 462 463 [BREWMP] Undefine WTF_OS_WINDOWS and WTF_PLATFORM_WIN 464 https://bugs.webkit.org/show_bug.cgi?id=34561 465 466 As the binary for simulator is built with MSVC 2005, 467 WTF_OS_WINDOWS and WTF_PLATFORM_WIN are defined. 468 Undefine them as we don't target Windows. 469 470 * wtf/Platform.h: 471 4722010-02-08 Chris Rogers <crogers@google.com> 473 474 Reviewed by Darin Adler. 475 476 audio engine: add Vector3 class 477 https://bugs.webkit.org/show_bug.cgi?id=34548 478 479 * wtf/Vector3.h: Added. 480 (WebCore::Vector3::Vector3): 481 (WebCore::Vector3::abs): 482 (WebCore::Vector3::isZero): 483 (WebCore::Vector3::normalize): 484 (WebCore::Vector3::x): 485 (WebCore::Vector3::y): 486 (WebCore::Vector3::z): 487 (WebCore::operator+): 488 (WebCore::operator-): 489 (WebCore::operator*): 490 (WebCore::dot): 491 (WebCore::cross): 492 (WebCore::distance): 493 4942010-02-08 Oliver Hunt <oliver@apple.com> 495 496 Reviewed by Gavin Barraclough. 497 498 Fix warning in clang++ 499 500 * runtime/Structure.h: 501 (JSC::Structure::propertyStorageSize): 502 5032010-02-08 Gavin Barraclough <barraclough@apple.com> 504 505 Reviewed by Geoff Garen. 506 507 Make makeString CRASH if we fail to allocate a string. 508 509 (tryMakeString or jsMakeNontrivialString can be used where we 510 expect allocation may fail and want to handle the error). 511 512 * runtime/JSStringBuilder.h: 513 (JSC::jsMakeNontrivialString): 514 * runtime/UString.h: 515 (JSC::tryMakeString): 516 (JSC::makeString): 517 5182010-02-08 Gavin Barraclough <barraclough@apple.com> 519 520 Rubber Stamped by Oliver Hunt. 521 522 Remove a couple of unnecesary C-style casts spotted by Darin. 523 524 * runtime/JSGlobalObjectFunctions.cpp: 525 (JSC::encode): 526 (JSC::globalFuncEscape): 527 5282010-02-08 Gavin Barraclough <barraclough@apple.com> 529 530 Reviewed by Geoff Garen. 531 532 Switch some more StringBuilder/jsNontrivialString code to use 533 JSStringBuilder/jsMakeNontrivialString - these methods will 534 throw an exception if we hit out-of-memory, rather than just 535 CRASHing. 536 537 * runtime/FunctionPrototype.cpp: 538 (JSC::functionProtoFuncToString): 539 * runtime/JSGlobalObjectFunctions.cpp: 540 (JSC::encode): 541 (JSC::decode): 542 (JSC::globalFuncEscape): 543 5442010-02-08 Gavin Barraclough <barraclough@apple.com> 545 546 Reviewed by Sam Weinig. 547 548 Use an empty identifier instead of a null identifier for parse 549 tokens without an identifier. 550 551 This helps encapsulate the null UStringImpl within UString. 552 553 * parser/Grammar.y: 554 * parser/NodeConstructors.h: 555 (JSC::ContinueNode::ContinueNode): 556 (JSC::BreakNode::BreakNode): 557 (JSC::ForInNode::ForInNode): 558 * runtime/CommonIdentifiers.cpp: 559 (JSC::CommonIdentifiers::CommonIdentifiers): 560 * runtime/CommonIdentifiers.h: 561 * runtime/FunctionPrototype.cpp: 562 (JSC::FunctionPrototype::FunctionPrototype): 563 5642010-02-08 Gustavo Noronha Silva <gustavo.noronha@collabora.co.uk> 565 566 Build fix for make distcheck. 567 568 * GNUmakefile.am: 569 5702010-02-08 Simon Hausmann <simon.hausmann@nokia.com> 571 572 Unreviewed RVCT build fix. 573 574 Similar to r54391, don't import the cmath functions from std:: for RVCT. 575 576 * wtf/MathExtras.h: 577 5782010-02-05 Gavin Barraclough <barraclough@apple.com> 579 580 Reviewed by Geoff Garen. 581 582 Change UStringImpl::create to CRASH if the string cannot be allocated, 583 rather than returning a null string (which will behave like a zero-length 584 string if used). 585 586 Also move createRep function from UString to become new overloaded 587 UStringImpl::create methods. In doing so, bring their behaviour closer to 588 being in line with WebCore::StringImpl, in removing the behaviour that they 589 can be used to produce null UStrings (ASSERT the char* provided is non-null). 590 This behaviour of converting null C-strings to null UStrings is inefficient 591 (cmompared to just using UString::null()), incompatible with WebCore::StringImpl's 592 behaviour, and may generate unexpected behaviour, since in many cases a null 593 UString can be used like an empty string. 594 595 With these changes UStringImpl need not have a concept of null impls, we can 596 start transitioning this to become an implementation detail of UString, that 597 internally it chooses to use a null-object rather than an actually zero impl 598 pointer. 599 600 * JavaScriptCore.exp: 601 * debugger/Debugger.cpp: 602 (JSC::Debugger::recompileAllJSFunctions): 603 * debugger/DebuggerCallFrame.cpp: 604 (JSC::DebuggerCallFrame::calculatedFunctionName): 605 * parser/Parser.cpp: 606 (JSC::Parser::parse): 607 * profiler/Profile.cpp: 608 (JSC::Profile::Profile): 609 * profiler/ProfileGenerator.cpp: 610 (JSC::ProfileGenerator::stopProfiling): 611 * runtime/Error.cpp: 612 (JSC::Error::create): 613 (JSC::throwError): 614 * runtime/ExceptionHelpers.cpp: 615 (JSC::createError): 616 * runtime/Identifier.cpp: 617 (JSC::Identifier::add): 618 * runtime/PropertyNameArray.cpp: 619 (JSC::PropertyNameArray::add): 620 * runtime/UString.cpp: 621 (JSC::initializeUString): 622 (JSC::UString::UString): 623 (JSC::UString::operator=): 624 * runtime/UString.h: 625 (JSC::UString::isNull): 626 (JSC::UString::null): 627 (JSC::UString::rep): 628 (JSC::UString::UString): 629 * runtime/UStringImpl.cpp: 630 (JSC::UStringImpl::create): 631 * runtime/UStringImpl.h: 632 6332010-02-05 Kwang Yul Seo <skyul@company100.net> 634 635 Reviewed by Eric Seidel. 636 637 [BREWMP] Define SYSTEM_MALLOC 1 638 https://bugs.webkit.org/show_bug.cgi?id=34640 639 640 Make BREWMP use system malloc because FastMalloc is not ported. 641 642 * wtf/Platform.h: 643 6442010-02-05 Kwang Yul Seo <skyul@company100.net> 645 646 Reviewed by Alexey Proskuryakov. 647 648 Don't call CRASH() in fastMalloc and fastCalloc when the requested memory size is 0 649 https://bugs.webkit.org/show_bug.cgi?id=34569 650 651 With USE_SYSTEM_MALLOC=1, fastMalloc and fastCalloc call CRASH() 652 if the return value of malloc and calloc is 0. 653 654 However, these functions can return 0 when the request size is 0. 655 Libc manual says, "If size is 0, then malloc() returns either NULL, 656 or a unique pointer value that can later be successfully passed to free()." 657 Though malloc returns a unique pointer in most systems, 658 0 can be returned in some systems. For instance, BREW's MALLOC returns 0 659 when size is 0. 660 661 If malloc or calloc returns 0 due to allocation size, increase the size 662 to 1 and try again. 663 664 * wtf/FastMalloc.cpp: 665 (WTF::fastMalloc): 666 (WTF::fastCalloc): 667 6682010-02-04 Mark Rowe <mrowe@apple.com> 669 670 Reviewed by Timothy Hatcher. 671 672 Build fix. Remove a symbol corresponding to an inline function from the linker export 673 file to prevent a weak external failure. 674 675 * JavaScriptCore.xcodeproj/project.pbxproj: Accommodate rename of script. 676 6772010-02-04 Daniel Bates <dbates@webkit.org> 678 679 [Qt] Unreviewed, build fix for Qt bot. 680 681 * runtime/JSStringBuilder.h: Changed #include <X.h> notation #include "X.h". 682 6832010-02-04 Geoffrey Garen <ggaren@apple.com> 684 685 Reviewed by Oliver Hunt. 686 687 Clearing a WeakGCPtr is weird 688 https://bugs.webkit.org/show_bug.cgi?id=34627 689 690 Added a WeakGCPtr::clear interface. 691 692 As discussed in https://bugs.webkit.org/show_bug.cgi?id=33383, the old 693 interface made it pretty weird for a client to conditionally clear a 694 WeakGCPtr, which is exactly what clients want to do when objects are 695 finalized. 696 697 * API/JSClassRef.cpp: 698 (clearReferenceToPrototype): Use the new WeakGCPtr::clear() interface. 699 700 * runtime/WeakGCPtr.h: 701 (JSC::WeakGCPtr::clear): Added an interface for clearing a WeakGCPtr, 702 iff its current value is the value passed in. It's cumbersome for the 703 client to do this test, since WeakGCPtr sometimes pretends to be null. 704 7052010-02-04 Geoffrey Garen <ggaren@apple.com> 706 707 Build fix: export a header. 708 709 * JavaScriptCore.xcodeproj/project.pbxproj: 710 7112010-02-04 Gavin Barraclough <barraclough@apple.com> 712 713 Reviewed by Oliver Hunt. 714 715 Add a JSStringBuilder class (similar-to, and derived-from StringBuilder) to 716 construct JSStrings, throwing a JS exception should we run out of memory whilst 717 allocating storage for the string. 718 719 Similarly, add jsMakeNontrivialString methods to use in cases where previously 720 we were calling makeString & passing the result to jsNontrivialString. Again, 721 these new methods throw if we hit an out of memory condition. 722 723 Move throwOutOfMemoryError into ExceptionHelpers, to make it more widely available. 724 725 * JavaScriptCore.xcodeproj/project.pbxproj: 726 * runtime/ArrayPrototype.cpp: 727 (JSC::arrayProtoFuncToString): 728 (JSC::arrayProtoFuncToLocaleString): 729 (JSC::arrayProtoFuncJoin): 730 * runtime/DateConstructor.cpp: 731 (JSC::callDate): 732 * runtime/DatePrototype.cpp: 733 (JSC::dateProtoFuncToString): 734 (JSC::dateProtoFuncToUTCString): 735 (JSC::dateProtoFuncToGMTString): 736 * runtime/ErrorPrototype.cpp: 737 (JSC::errorProtoFuncToString): 738 * runtime/ExceptionHelpers.cpp: 739 (JSC::throwOutOfMemoryError): 740 * runtime/ExceptionHelpers.h: 741 * runtime/JSStringBuilder.h: Added. 742 (JSC::JSStringBuilder::releaseJSString): 743 (JSC::jsMakeNontrivialString): 744 * runtime/NumberPrototype.cpp: 745 (JSC::numberProtoFuncToPrecision): 746 * runtime/ObjectPrototype.cpp: 747 (JSC::objectProtoFuncToString): 748 * runtime/Operations.cpp: 749 * runtime/Operations.h: 750 * runtime/RegExpPrototype.cpp: 751 (JSC::regExpProtoFuncToString): 752 * runtime/StringBuilder.h: 753 (JSC::StringBuilder::append): 754 * runtime/StringPrototype.cpp: 755 (JSC::stringProtoFuncBig): 756 (JSC::stringProtoFuncSmall): 757 (JSC::stringProtoFuncBlink): 758 (JSC::stringProtoFuncBold): 759 (JSC::stringProtoFuncFixed): 760 (JSC::stringProtoFuncItalics): 761 (JSC::stringProtoFuncStrike): 762 (JSC::stringProtoFuncSub): 763 (JSC::stringProtoFuncSup): 764 (JSC::stringProtoFuncFontcolor): 765 (JSC::stringProtoFuncFontsize): 766 (JSC::stringProtoFuncAnchor): 767 7682010-02-04 Steve Falkenburg <sfalken@apple.com> 769 770 Windows build fix. 771 772 * wtf/MathExtras.h: 773 7742010-02-04 Darin Adler <darin@apple.com> 775 776 Reviewed by David Levin. 777 778 Make MathExtras.h compatible with <cmath> 779 https://bugs.webkit.org/show_bug.cgi?id=34618 780 781 * wtf/MathExtras.h: Include <cmath> instead of <math.h>. 782 Use "using" as we do elsewhere in WTF for the four functions from <cmath> 783 we want to use without the prefix. Later we could consider making the std 784 explicit at call sites instead. 785 7862010-02-04 Tamas Szirbucz <szirbucz@inf.u-szeged.hu> 787 788 Reviewed by Gavin Barraclough. 789 790 Use an easily appendable structure for trampolines instead of pointer parameters. 791 https://bugs.webkit.org/show_bug.cgi?id=34424 792 793 * assembler/ARMAssembler.cpp: 794 (JSC::ARMAssembler::executableCopy): 795 * jit/JIT.h: 796 (JSC::JIT::compileCTIMachineTrampolines): 797 * jit/JITOpcodes.cpp: 798 (JSC::JIT::privateCompileCTIMachineTrampolines): 799 * jit/JITStubs.cpp: 800 (JSC::JITThunks::JITThunks): 801 * jit/JITStubs.h: 802 (JSC::JITThunks::ctiStringLengthTrampoline): 803 (JSC::JITThunks::ctiVirtualCallLink): 804 (JSC::JITThunks::ctiVirtualCall): 805 (JSC::JITThunks::ctiNativeCallThunk): 806 8072010-02-04 Jedrzej Nowacki <jedrzej.nowacki@nokia.com> 808 809 Reviewed by Simon Hausmann. 810 811 Increase test coverage for the QScriptValue. 812 813 https://bugs.webkit.org/show_bug.cgi?id=34533 814 815 * qt/tests/qscriptvalue/qscriptvalue.pro: 816 * qt/tests/qscriptvalue/tst_qscriptvalue.cpp: 817 (tst_QScriptValue::tst_QScriptValue): 818 (tst_QScriptValue::~tst_QScriptValue): 819 (tst_QScriptValue::dataHelper): 820 (tst_QScriptValue::newRow): 821 (tst_QScriptValue::testHelper): 822 (tst_QScriptValue::ctor): 823 * qt/tests/qscriptvalue/tst_qscriptvalue.h: Added. 824 * qt/tests/qscriptvalue/tst_qscriptvalue_generated.cpp: Added. 825 (tst_QScriptValue::initScriptValues): 826 (tst_QScriptValue::isValid_initData): 827 (tst_QScriptValue::isValid_makeData): 828 (tst_QScriptValue::isValid_test): 829 (tst_QScriptValue::isBool_initData): 830 (tst_QScriptValue::isBool_makeData): 831 (tst_QScriptValue::isBool_test): 832 (tst_QScriptValue::isBoolean_initData): 833 (tst_QScriptValue::isBoolean_makeData): 834 (tst_QScriptValue::isBoolean_test): 835 (tst_QScriptValue::isFunction_initData): 836 (tst_QScriptValue::isFunction_makeData): 837 (tst_QScriptValue::isFunction_test): 838 (tst_QScriptValue::isNull_initData): 839 (tst_QScriptValue::isNull_makeData): 840 (tst_QScriptValue::isNull_test): 841 (tst_QScriptValue::isString_initData): 842 (tst_QScriptValue::isString_makeData): 843 (tst_QScriptValue::isString_test): 844 (tst_QScriptValue::isUndefined_initData): 845 (tst_QScriptValue::isUndefined_makeData): 846 (tst_QScriptValue::isUndefined_test): 847 (tst_QScriptValue::isObject_initData): 848 (tst_QScriptValue::isObject_makeData): 849 (tst_QScriptValue::isObject_test): 850 8512010-02-03 Kwang Yul Seo <skyul@company100.net> 852 853 Reviewed by Eric Seidel. 854 855 [BREWMP] Define WTF_PLATFORM_BREWMP_SIMULATOR when AEE_SIMULATOR is defined 856 https://bugs.webkit.org/show_bug.cgi?id=34514 857 858 PLATFORM(BREWMP_SIMULATOR) guard is needed to make distinction between BREWMP 859 and BREWMP simulator. 860 861 * wtf/Platform.h: 862 8632010-02-03 Kwang Yul Seo <skyul@company100.net> 864 865 Reviewed by Eric Seidel. 866 867 [BREWMP] Remove COMPILE_ASSERT conflict with the underlying PLATFORM 868 https://bugs.webkit.org/show_bug.cgi?id=34190 869 870 COMPILE_ASSERT conflicts with the underlying PLATFORM because it is defined 871 both in WTF's Assertions.h and BREWMP's AEEClassIDs.h. Include AEEClassIDs.h 872 in Assertions.h and undef COMPILE_ASSERT to avoid redefining COMPILE_ASSERT. 873 874 * wtf/Assertions.h: 875 8762010-02-03 Kwang Yul Seo <skyul@company100.net> 877 878 Reviewed by Eric Seidel. 879 880 [BREWMP] Implement OwnPtrBrew to make sure BREW instances are freed. 881 https://bugs.webkit.org/show_bug.cgi?id=34518 882 883 Add OwnPtrBrew to release IFile, IFileMgr and IBitmap instances. 884 885 * wtf/brew/OwnPtrBrew.cpp: Added. 886 (WTF::IFileMgr): 887 (WTF::IFile): 888 (WTF::IBitmap): 889 (WTF::freeOwnedPtrBrew): 890 * wtf/brew/OwnPtrBrew.h: Added. 891 (WTF::OwnPtrBrew::OwnPtrBrew): 892 (WTF::OwnPtrBrew::~OwnPtrBrew): 893 (WTF::OwnPtrBrew::get): 894 (WTF::OwnPtrBrew::release): 895 (WTF::OwnPtrBrew::outPtr): 896 (WTF::OwnPtrBrew::set): 897 (WTF::OwnPtrBrew::clear): 898 (WTF::OwnPtrBrew::operator*): 899 (WTF::OwnPtrBrew::operator->): 900 (WTF::OwnPtrBrew::operator!): 901 (WTF::OwnPtrBrew::operator UnspecifiedBoolType): 902 (WTF::OwnPtrBrew::swap): 903 (WTF::swap): 904 (WTF::operator==): 905 (WTF::operator!=): 906 (WTF::getPtr): 907 9082010-02-03 Kwang Yul Seo <skyul@company100.net> 909 910 Reviewed by Darin Adler. 911 912 Export WTF::fastStrDup symbol 913 https://bugs.webkit.org/show_bug.cgi?id=34526 914 915 * JavaScriptCore.exp: 916 9172010-02-03 Kevin Watters <kevinwatters@gmail.com> 918 919 Reviewed by Kevin Ollivier. 920 921 [wx] Enable JIT compilation for wx. 922 923 https://bugs.webkit.org/show_bug.cgi?id=34536 924 925 * wtf/Platform.h: 926 9272010-02-02 Oliver Hunt <oliver@apple.com> 928 929 Reviewed by Geoffrey Garen. 930 931 Crash in CollectorBitmap::get at nbcolympics.com 932 https://bugs.webkit.org/show_bug.cgi?id=34504 933 934 This was caused by the use of m_offset to determine the offset of 935 a new property into the property storage. This patch corrects 936 the effected cases by incorporating the anonymous slot count. It 937 also removes the duplicate copy of anonymous slot count from the 938 property table as keeping this up to date merely increased the 939 chance of a mismatch. Finally I've added a large number of 940 assertions in an attempt to prevent such a bug from happening 941 again. 942 943 With the new assertions in place the existing anonymous slot tests 944 all fail without the m_offset fixes. 945 946 * runtime/PropertyMapHashTable.h: 947 * runtime/Structure.cpp: 948 (JSC::Structure::materializePropertyMap): 949 (JSC::Structure::addPropertyTransitionToExistingStructure): 950 (JSC::Structure::addPropertyTransition): 951 (JSC::Structure::removePropertyTransition): 952 (JSC::Structure::flattenDictionaryStructure): 953 (JSC::Structure::addPropertyWithoutTransition): 954 (JSC::Structure::removePropertyWithoutTransition): 955 (JSC::Structure::copyPropertyTable): 956 (JSC::Structure::get): 957 (JSC::Structure::put): 958 (JSC::Structure::remove): 959 (JSC::Structure::insertIntoPropertyMapHashTable): 960 (JSC::Structure::createPropertyMapHashTable): 961 (JSC::Structure::rehashPropertyMapHashTable): 962 (JSC::Structure::checkConsistency): 963 9642010-02-02 Steve Falkenburg <sfalken@apple.com> 965 966 Reviewed by Darin Adler. 967 968 Copyright year updating for Windows version resources should be automatic 969 https://bugs.webkit.org/show_bug.cgi?id=34503 970 971 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.rc: 972 9732010-02-02 Kwang Yul Seo <skyul@company100.net> 974 975 Reviewed by Eric Seidel. 976 977 [BREWMP] Add dummy main thread functions 978 https://bugs.webkit.org/show_bug.cgi?id=33569 979 980 Add dummy initializeMainThreadPlatform and 981 scheduleDispatchFunctionsOnMainThread. 982 983 * wtf/brew/MainThreadBrew.cpp: Added. 984 (WTF::initializeMainThreadPlatform): 985 (WTF::scheduleDispatchFunctionsOnMainThread): 986 9872010-02-02 Kwang Yul Seo <skyul@company100.net> 988 989 Reviewed by Darin Adler. 990 991 Add using WTF::getLocalTime to CurrentTime.h 992 https://bugs.webkit.org/show_bug.cgi?id=34493 993 994 * wtf/CurrentTime.h: 995 9962010-02-02 Kwang Yul Seo <skyul@company100.net> 997 998 Reviewed by Eric Seidel. 999 1000 [BREWMP] Add HAVE_XXX definitions 1001 https://bugs.webkit.org/show_bug.cgi?id=34414 1002 1003 Add HAVE_ERRNO_H=1 1004 1005 * wtf/Platform.h: 1006 10072010-02-02 Kwang Yul Seo <skyul@company100.net> 1008 1009 Reviewed by Eric Seidel. 1010 1011 [BREWMP] Don't define HAVE_TM_GMTOFF, HAVE_TM_ZONE and HAVE_TIMEGM 1012 https://bugs.webkit.org/show_bug.cgi?id=34388 1013 1014 BREWMP does not have these features. 1015 1016 * wtf/Platform.h: 1017 10182010-02-02 Kwang Yul Seo <skyul@company100.net> 1019 1020 Reviewed by Eric Seidel. 1021 1022 [BREWMP] Define WTF_PLATFORM_BREWMP=1 when BUILDING_BREWMP is defined 1023 https://bugs.webkit.org/show_bug.cgi?id=34386 1024 1025 Define WTF_PLATFORM_BREWMP=1 so that PLATFORM(BREWMP) guard can be used. 1026 1027 * wtf/Platform.h: 1028 10292010-02-01 Kent Tamura <tkent@chromium.org> 1030 1031 Reviewed by Darin Adler. 1032 1033 Date.UTC() should apply TimeClip operation. 1034 https://bugs.webkit.org/show_bug.cgi?id=34461 1035 1036 ECMAScript 5 15.9.4.3: 1037 > 9 Return TimeClip(MakeDate(MakeDay(yr, m, dt), MakeTime(h, min, s, milli))). 1038 1039 * runtime/DateConstructor.cpp: 1040 (JSC::dateUTC): Calls WTF::timeClip(). 1041 10422010-02-01 Kent Tamura <tkent@chromium.org> 1043 1044 Reviewed by Darin Adler. 1045 1046 Fix a bug that Math.round() retunrs incorrect results for huge integers 1047 https://bugs.webkit.org/show_bug.cgi?id=34462 1048 1049 * runtime/MathObject.cpp: 1050 (JSC::mathProtoFuncRound): Avoid "arg + 0.5". 1051 10522010-02-01 Kwang Yul Seo <skyul@company100.net> 1053 1054 Reviewed by Eric Seidel. 1055 1056 [BREWMP] Port WTF's currentTime 1057 https://bugs.webkit.org/show_bug.cgi?id=33567 1058 1059 Combine GETUTCSECONDS and GETTIMEMS to calculate the number 1060 of milliseconds since 1970/01/01 00:00:00 UTC. 1061 1062 * wtf/CurrentTime.cpp: 1063 (WTF::currentTime): 1064 10652010-02-01 Patrick Gansterer <paroga@paroga.com> 1066 1067 Reviewed by Darin Adler. 1068 1069 [Qt] WinCE buildfix after r52729 and fix for Q_BIG_ENDIAN typo. 1070 https://bugs.webkit.org/show_bug.cgi?id=34378 1071 1072 * wtf/Platform.h: 1073 10742010-02-01 Oliver Hunt <oliver@apple.com> 1075 1076 Reviewed by Gavin Barraclough. 1077 1078 Structure not accounting for anonymous slots when computing property storage size 1079 https://bugs.webkit.org/show_bug.cgi?id=34441 1080 1081 Previously any Structure with anonymous storage would have a property map, so we 1082 were only including anonymous slot size if there was a property map. Given this 1083 is no longer the case we should always include the anonymous slot count in the 1084 property storage size. 1085 1086 * runtime/Structure.h: 1087 (JSC::Structure::propertyStorageSize): 1088 10892010-02-01 Oliver Hunt <oliver@apple.com> 1090 1091 Windows build fix, update exports file (again) 1092 1093 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 1094 10952010-02-01 Oliver Hunt <oliver@apple.com> 1096 1097 Windows build fix, update exports file 1098 1099 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 1100 11012010-01-31 Oliver Hunt <oliver@apple.com> 1102 1103 Reviewed by Maciej Stachowiak. 1104 1105 JSC is failing to propagate anonymous slot count on some transitions 1106 https://bugs.webkit.org/show_bug.cgi?id=34321 1107 1108 Remove secondary Structure constructor, and make Structure store a copy 1109 of the number of anonymous slots directly so saving an immediate allocation 1110 of a property map for all structures with anonymous storage, which also 1111 avoids the leaked property map on new property transition in the original 1112 version of this patch. 1113 1114 We need to propagate the the anonymous slot count otherwise we can end up 1115 with a structure recording incorrect information about the available and 1116 needed space for property storage, or alternatively incorrectly reusing 1117 some slots. 1118 1119 * JavaScriptCore.exp: 1120 * runtime/Structure.cpp: 1121 (JSC::Structure::Structure): 1122 (JSC::Structure::materializePropertyMap): 1123 (JSC::Structure::addPropertyTransition): 1124 (JSC::Structure::changePrototypeTransition): 1125 (JSC::Structure::despecifyFunctionTransition): 1126 (JSC::Structure::getterSetterTransition): 1127 (JSC::Structure::toDictionaryTransition): 1128 (JSC::Structure::flattenDictionaryStructure): 1129 (JSC::Structure::copyPropertyTable): 1130 (JSC::Structure::put): 1131 (JSC::Structure::remove): 1132 (JSC::Structure::insertIntoPropertyMapHashTable): 1133 (JSC::Structure::createPropertyMapHashTable): 1134 * runtime/Structure.h: 1135 (JSC::Structure::create): 1136 (JSC::Structure::hasAnonymousSlots): 1137 (JSC::Structure::anonymousSlotCount): 1138 11392010-01-31 Patrick Gansterer <paroga@paroga.com> 1140 1141 Reviewed by Darin Adler. 1142 1143 Buildfix for WinCE + style fixes (TLS_OUT_OF_INDEXES is not defined). 1144 https://bugs.webkit.org/show_bug.cgi?id=34380 1145 1146 * wtf/ThreadSpecific.h: 1147 11482010-01-31 Kent Tamura <tkent@chromium.org> 1149 1150 Reviewed by Darin Adler. 1151 1152 [Windows] Fix a bug of round() with huge integral numbers 1153 https://bugs.webkit.org/show_bug.cgi?id=34297 1154 1155 Fix a bug that round() for huge integral numbers returns incorrect 1156 results. For example, round(8639999913600001) returns 1157 8639999913600002 without this change though the double type can 1158 represent 8639999913600001 precisely. 1159 1160 Math.round() of JavaScript has a similar problem. But this change 1161 doesn't fix it because Math.round() doesn't use round() of 1162 MathExtra.h. 1163 1164 * wtf/MathExtras.h: 1165 (round): Avoid to do "num + 0.5" or "num - 0.5". 1166 (roundf): Fixed similarly. 1167 (llround): Calls round(). 1168 (llroundf): Calls roundf(). 1169 (lround): Calls round(). 1170 (lroundf): Calls roundf(). 1171 11722010-01-29 Mark Rowe <mrowe@apple.com> 1173 1174 Sort Xcode projects. 1175 1176 * JavaScriptCore.xcodeproj/project.pbxproj: 1177 11782010-01-29 Mark Rowe <mrowe@apple.com> 1179 1180 Fix the Mac build. 1181 1182 Disable ENABLE_INDEXED_DATABASE since it is "completely non-functional". 1183 1184 As the comment in FeatureDefines.xcconfig notes, the list of feature defines 1185 needs to be kept in sync across the various files. The default values also 1186 need to be kept in sync between these files and build-webkit. 1187 1188 * Configurations/FeatureDefines.xcconfig: 1189 11902010-01-29 Simon Hausmann <simon.hausmann@nokia.com> 1191 1192 Rubber-stamped by Maciej Stachowiak. 1193 1194 Fix the ARM build. 1195 1196 * runtime/JSNumberCell.h: 1197 (JSC::JSNumberCell::createStructure): Call the right Structure::create overload. 1198 11992010-01-28 Kevin Ollivier <kevino@theolliviers.com> 1200 1201 [wx] Build fix for MSW, use ThreadingWin.cpp as the Windows pthreads implementation 1202 implements pthread_t in a way that makes it impossible to check its validity, 1203 which is needed by ThreadingPthreads.cpp. 1204 1205 * wscript: 1206 12072010-01-28 Oliver Hunt <oliver@apple.com> 1208 1209 Reviewed by Gavin Barraclough. 1210 1211 DOM Objects shouldn't all require custom mark functions 1212 https://bugs.webkit.org/show_bug.cgi?id=34291 1213 1214 Make getAnonymousValue const-friendly 1215 1216 * runtime/JSObject.h: 1217 (JSC::JSObject::getAnonymousValue): 1218 12192010-01-28 Oliver Hunt <oliver@apple.com> 1220 1221 Reviewed by Gavin Barraclough. 1222 1223 Simplify anonymous slot implementation 1224 https://bugs.webkit.org/show_bug.cgi?id=34282 1225 1226 A class must now specify the number of slots it needs at construction time 1227 rather than later on with a transition. This makes many things simpler, 1228 we no longer need to need an additional transition on object creation to 1229 add the anonymous slots, and we remove the need for a number of transition 1230 type checks. 1231 1232 * API/JSCallbackConstructor.h: 1233 (JSC::JSCallbackConstructor::createStructure): 1234 * API/JSCallbackFunction.h: 1235 (JSC::JSCallbackFunction::createStructure): 1236 * API/JSCallbackObject.h: 1237 (JSC::JSCallbackObject::createStructure): 1238 * JavaScriptCore.exp: 1239 * debugger/DebuggerActivation.h: 1240 (JSC::DebuggerActivation::createStructure): 1241 * runtime/Arguments.h: 1242 (JSC::Arguments::createStructure): 1243 * runtime/BooleanObject.h: 1244 (JSC::BooleanObject::createStructure): 1245 * runtime/DateInstance.h: 1246 (JSC::DateInstance::createStructure): 1247 * runtime/DatePrototype.h: 1248 (JSC::DatePrototype::createStructure): 1249 * runtime/FunctionPrototype.h: 1250 (JSC::FunctionPrototype::createStructure): 1251 * runtime/GetterSetter.h: 1252 (JSC::GetterSetter::createStructure): 1253 * runtime/GlobalEvalFunction.h: 1254 (JSC::GlobalEvalFunction::createStructure): 1255 * runtime/InternalFunction.h: 1256 (JSC::InternalFunction::createStructure): 1257 * runtime/JSAPIValueWrapper.h: 1258 (JSC::JSAPIValueWrapper::createStructure): 1259 * runtime/JSActivation.h: 1260 (JSC::JSActivation::createStructure): 1261 * runtime/JSArray.h: 1262 (JSC::JSArray::createStructure): 1263 * runtime/JSByteArray.cpp: 1264 (JSC::JSByteArray::createStructure): 1265 * runtime/JSCell.h: 1266 (JSC::JSCell::createDummyStructure): 1267 * runtime/JSFunction.h: 1268 (JSC::JSFunction::createStructure): 1269 * runtime/JSGlobalObject.h: 1270 (JSC::JSGlobalObject::createStructure): 1271 * runtime/JSNotAnObject.h: 1272 (JSC::JSNotAnObject::createStructure): 1273 * runtime/JSONObject.h: 1274 (JSC::JSONObject::createStructure): 1275 * runtime/JSObject.h: 1276 (JSC::JSObject::createStructure): 1277 (JSC::JSObject::putAnonymousValue): 1278 (JSC::JSObject::getAnonymousValue): 1279 * runtime/JSPropertyNameIterator.h: 1280 (JSC::JSPropertyNameIterator::createStructure): 1281 * runtime/JSStaticScopeObject.h: 1282 (JSC::JSStaticScopeObject::createStructure): 1283 * runtime/JSString.h: 1284 (JSC::Fiber::createStructure): 1285 * runtime/JSVariableObject.h: 1286 (JSC::JSVariableObject::createStructure): 1287 * runtime/JSWrapperObject.h: 1288 (JSC::JSWrapperObject::createStructure): 1289 (JSC::JSWrapperObject::JSWrapperObject): 1290 * runtime/MathObject.h: 1291 (JSC::MathObject::createStructure): 1292 * runtime/NumberConstructor.h: 1293 (JSC::NumberConstructor::createStructure): 1294 * runtime/NumberObject.h: 1295 (JSC::NumberObject::createStructure): 1296 * runtime/RegExpConstructor.h: 1297 (JSC::RegExpConstructor::createStructure): 1298 * runtime/RegExpObject.h: 1299 (JSC::RegExpObject::createStructure): 1300 * runtime/StringObject.h: 1301 (JSC::StringObject::createStructure): 1302 * runtime/StringObjectThatMasqueradesAsUndefined.h: 1303 (JSC::StringObjectThatMasqueradesAsUndefined::createStructure): 1304 * runtime/Structure.cpp: 1305 (JSC::Structure::~Structure): 1306 (JSC::Structure::materializePropertyMap): 1307 * runtime/Structure.h: 1308 (JSC::Structure::create): 1309 (JSC::Structure::anonymousSlotCount): 1310 * runtime/StructureTransitionTable.h: 1311 13122010-01-27 Oliver Hunt <oliver@apple.com> 1313 1314 Windows build fix. 1315 1316 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 1317 13182010-01-27 Oliver Hunt <oliver@apple.com> 1319 1320 Reviewed by Maciej Stachowiak. 1321 1322 MessageEvent.data should deserialize in the context of the MessageEvent's global object 1323 https://bugs.webkit.org/show_bug.cgi?id=34227 1324 1325 Add logic to allow us to create an Object, Array, or Date instance 1326 so we can create them in the context of a specific global object, 1327 rather than just using the current lexical global object. 1328 1329 * JavaScriptCore.exp: 1330 * runtime/DateInstance.cpp: 1331 (JSC::DateInstance::DateInstance): 1332 * runtime/DateInstance.h: 1333 * runtime/JSGlobalObject.h: 1334 (JSC::constructEmptyObject): 1335 (JSC::constructEmptyArray): 1336 13372010-01-27 Alexey Proskuryakov <ap@apple.com> 1338 1339 Reviewed by Darin Adler. 1340 1341 https://bugs.webkit.org/show_bug.cgi?id=34150 1342 WebKit needs a mechanism to catch stale HashMap entries 1343 1344 It is very difficult to catch stale pointers that are HashMap keys - since a pointer's hash 1345 is just its value, it is very unlikely that any observable problem is reproducible. 1346 1347 This extends hash table consistency checks to check that pointers are referencing allocated 1348 memory blocks, and makes it possible to invoke the checks explicitly (it is not feasible 1349 to enable CHECK_HASHTABLE_CONSISTENCY by default, because that affects performance too much). 1350 1351 * wtf/HashMap.h: (WTF::::checkConsistency): Call through to HashTable implementation. We can 1352 add similar calls to HashSet and HashCountedSet, but I haven't seen hard to debug problems 1353 with those yet. 1354 1355 * wtf/HashSet.h: (WTF::::remove): The version of checkTableConsistency that's guarded by 1356 CHECK_HASHTABLE_CONSISTENCY is now called internalCheckTableConsistency(). 1357 1358 * wtf/HashTable.h: 1359 (WTF::HashTable::internalCheckTableConsistency): 1360 (WTF::HashTable::internalCheckTableConsistencyExceptSize): 1361 (WTF::HashTable::checkTableConsistencyExceptSize): 1362 Expose checkTableConsistency() even if CHECK_HASHTABLE_CONSISTENCY is off. 1363 (WTF::::add): Updated for checkTableConsistency renaming. 1364 (WTF::::addPassingHashCode): Ditto. 1365 (WTF::::removeAndInvalidate): Ditto. 1366 (WTF::::remove): Ditto. 1367 (WTF::::rehash): Ditto. 1368 (WTF::::checkTableConsistency): The assertion for !shouldExpand() was not correct - this 1369 function returns true for tables with m_table == 0. 1370 (WTF::::checkTableConsistencyExceptSize): Call checkValueConsistency for key. Potentially, 1371 we could do the same for values. 1372 1373 * wtf/HashTraits.h: 1374 (WTF::GenericHashTraits::checkValueConsistency): An empty function that can be overridden 1375 to add checks. Currently, the only override is for pointer hashes. 1376 1377 * wtf/RefPtrHashMap.h: (WTF::::remove): Updated for checkTableConsistency renaming. 1378 13792010-01-27 Anton Muhin <antonm@chromium.org> 1380 1381 Reviewed by Darin Adler. 1382 1383 Remove trailing \ from inline function code 1384 https://bugs.webkit.org/show_bug.cgi?id=34223 1385 1386 * assembler/ARMv7Assembler.h: 1387 (JSC::ARMThumbImmediate::countLeadingZerosPartial): 1388 13892010-01-27 Kwang Yul Seo <skyul@company100.net> 1390 1391 Reviewed by Eric Seidel. 1392 1393 [BREWMP] Port WTF's randomNumber 1394 https://bugs.webkit.org/show_bug.cgi?id=33566 1395 1396 Use GETRAND to generate 4 byte random byte sequence to implement 1397 weakRandomNumber. Create a secure random number generator with 1398 AEECLSID_RANDOM to implement randomNumber. 1399 1400 * wtf/RandomNumber.cpp: 1401 (WTF::weakRandomNumber): 1402 (WTF::randomNumber): 1403 14042010-01-27 Kwang Yul Seo <skyul@company100.net> 1405 1406 Reviewed by Eric Seidel. 1407 1408 [BREWMP] Port getCPUTime 1409 https://bugs.webkit.org/show_bug.cgi?id=33572 1410 1411 Use GETUPTIMEMS which returns a continuously and 1412 linearly increasing millisecond timer from the time the device 1413 was powered on. This function is enough to implement getCPUTime. 1414 1415 * runtime/TimeoutChecker.cpp: 1416 (JSC::getCPUTime): 1417 14182010-01-27 Kwang Yul Seo <skyul@company100.net> 1419 1420 Reviewed by Oliver Hunt. 1421 1422 [BREWMP] Add MarkStack fastMalloc implementation for platforms without VirtualAlloc or mmap. 1423 https://bugs.webkit.org/show_bug.cgi?id=33582 1424 1425 Use fastMalloc and fastFree to implement MarkStack::allocateStack and 1426 MarkStack::releaseStack for platforms without page level allocation. 1427 1428 * runtime/MarkStack.h: 1429 (JSC::MarkStack::MarkStackArray::shrinkAllocation): 1430 * runtime/MarkStackNone.cpp: Added. 1431 (JSC::MarkStack::initializePagesize): 1432 (JSC::MarkStack::allocateStack): 1433 (JSC::MarkStack::releaseStack): 1434 14352010-01-27 Kwang Yul Seo <skyul@company100.net> 1436 1437 Reviewed by Eric Seidel. 1438 1439 [BREWMP] Don't use time function 1440 https://bugs.webkit.org/show_bug.cgi?id=33577 1441 1442 Calling time(0) in BREW devices causes a crash because time 1443 is not properly ported in most devices. Cast currentTime() to 1444 time_t to get the same result as time(0). 1445 1446 * wtf/DateMath.cpp: 1447 (WTF::calculateUTCOffset): 1448 14492010-01-27 Alexey Proskuryakov <ap@apple.com> 1450 1451 Revert r53899 (HashMap<AtomicStringImpl*, Value> key checks) and subsequent build fixes, 1452 because they make SVG tests crash in release builds. 1453 1454 * wtf/HashMap.h: 1455 (WTF::::remove): 1456 * wtf/HashSet.h: 1457 (WTF::::remove): 1458 * wtf/HashTable.h: 1459 (WTF::::add): 1460 (WTF::::addPassingHashCode): 1461 (WTF::::removeAndInvalidate): 1462 (WTF::::remove): 1463 (WTF::::rehash): 1464 (WTF::::checkTableConsistency): 1465 (WTF::::checkTableConsistencyExceptSize): 1466 * wtf/HashTraits.h: 1467 (WTF::GenericHashTraits::emptyValue): 1468 (WTF::): 1469 * wtf/RefPtrHashMap.h: 1470 (WTF::::remove): 1471 14722010-01-26 Alexey Proskuryakov <ap@apple.com> 1473 1474 More Windows build fixing. 1475 1476 * wtf/HashTraits.h: _msize takes void*, remove const qualifier from type. 1477 14782010-01-26 Alexey Proskuryakov <ap@apple.com> 1479 1480 Windows build fix. 1481 1482 * wtf/HashTraits.h: Include malloc.h for _msize(). 1483 14842010-01-26 Alexey Proskuryakov <ap@apple.com> 1485 1486 Build fix. 1487 1488 * wtf/HashTable.h: (WTF::HashTable::checkTableConsistencyExceptSize): Remove const from a 1489 static (empty) version of this function. 1490 14912010-01-26 Alexey Proskuryakov <ap@apple.com> 1492 1493 Reviewed by Darin Adler. 1494 1495 https://bugs.webkit.org/show_bug.cgi?id=34150 1496 WebKit needs a mechanism to catch stale HashMap entries 1497 1498 It is very difficult to catch stale pointers that are HashMap keys - since a pointer's hash 1499 is just its value, it is very unlikely that any observable problem is reproducible. 1500 1501 This extends hash table consistency checks to check that pointers are referencing allocated 1502 memory blocks, and makes it possible to invoke the checks explicitly (it is not feasible 1503 to enable CHECK_HASHTABLE_CONSISTENCY by default, because that affects performance too much). 1504 1505 * wtf/HashMap.h: (WTF::::checkConsistency): Call through to HashTable implementation. We can 1506 add similar calls to HashSet and HashCountedSet, but I haven't seen hard to debug problems 1507 with those yet. 1508 1509 * wtf/HashSet.h: (WTF::::remove): The version of checkTableConsistency that's guarded by 1510 CHECK_HASHTABLE_CONSISTENCY is now called internalCheckTableConsistency(). 1511 1512 * wtf/HashTable.h: 1513 (WTF::HashTable::internalCheckTableConsistency): 1514 (WTF::HashTable::internalCheckTableConsistencyExceptSize): 1515 (WTF::HashTable::checkTableConsistencyExceptSize): 1516 Expose checkTableConsistency() even if CHECK_HASHTABLE_CONSISTENCY is off. 1517 (WTF::::add): Updated for checkTableConsistency renaming. 1518 (WTF::::addPassingHashCode): Ditto. 1519 (WTF::::removeAndInvalidate): Ditto. 1520 (WTF::::remove): Ditto. 1521 (WTF::::rehash): Ditto. 1522 (WTF::::checkTableConsistency): The assertion for !shouldExpand() was not correct - this 1523 function returns true for tables with m_table == 0. 1524 (WTF::::checkTableConsistencyExceptSize): Call checkValueConsistency for key. Potentially, 1525 we could do the same for values. 1526 1527 * wtf/HashTraits.h: 1528 (WTF::GenericHashTraits::checkValueConsistency): An empty function that can be overridden 1529 to add checks. Currently, the only override is for pointer hashes. 1530 1531 * wtf/RefPtrHashMap.h: (WTF::::remove): Updated for checkTableConsistency renaming. 1532 15332010-01-26 Lyon Chen <liachen@rim.com> 1534 1535 Reviewed by Maciej Stachowiak. 1536 1537 Opcode.h use const void* for Opcode cause error #1211 for RVCT compiler 1538 https://bugs.webkit.org/show_bug.cgi?id=33902 1539 1540 * bytecode/Opcode.h: 1541 15422010-01-26 Steve Falkenburg <sfalken@apple.com> 1543 1544 Reviewed by Oliver Hunt. 1545 1546 Windows build references non-existent include paths 1547 https://bugs.webkit.org/show_bug.cgi?id=34175 1548 1549 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: 1550 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCoreCommon.vsprops: 1551 * JavaScriptCore.vcproj/WTF/WTFCommon.vsprops: 1552 * JavaScriptCore.vcproj/jsc/jscCommon.vsprops: 1553 * JavaScriptCore.vcproj/testapi/testapi.vcproj: 1554 * JavaScriptCore.vcproj/testapi/testapiCommon.vsprops: 1555 15562010-01-26 Oliver Hunt <oliver@apple.com> 1557 1558 Reviewed by Geoffrey Garen. 1559 1560 Using JavaScriptCore API with a webkit vended context can result in slow script dialog 1561 https://bugs.webkit.org/show_bug.cgi?id=34172 1562 1563 Make the APIShim correctly increment and decrement the timeout 1564 entry counter. 1565 1566 * API/APIShims.h: 1567 (JSC::APIEntryShimWithoutLock::APIEntryShimWithoutLock): 1568 (JSC::APIEntryShimWithoutLock::~APIEntryShimWithoutLock): 1569 (JSC::APICallbackShim::APICallbackShim): 1570 (JSC::APICallbackShim::~APICallbackShim): 1571 15722010-01-26 Simon Hausmann <simon.hausmann@nokia.com> 1573 1574 [Qt] Fix compilation of QtScript with non-gcc compilers 1575 1576 Variable length stack arrays are a gcc extension. Use QVarLengthArray 1577 as a more portable solution that still tries to allocate on the stack 1578 first. 1579 1580 * qt/api/qscriptvalue_p.h: 1581 (QScriptValuePrivate::call): 1582 15832010-01-26 Simon Hausmann <simon.hausmann@nokia.com> 1584 1585 Reviewed by Tor Arne Vestbø. 1586 1587 [Qt] Fix the build on platforms without JIT support. 1588 1589 The JIT support should be determined at compile-time via wtf/Platform.h 1590 1591 * qt/api/QtScript.pro: 1592 15932010-01-26 Jedrzej Nowacki <jedrzej.nowacki@nokia.com> 1594 1595 Reviewed by Simon Hausmann. 1596 1597 First steps of the QtScript API. 1598 1599 Two new classes were created; QScriptEngine and QScriptValue. 1600 The first should encapsulate a javascript context and the second a script 1601 value. 1602 1603 This API is still in development, so it isn't compiled by default. 1604 To trigger compilation, pass --qmakearg="CONFIG+=build-qtscript" to 1605 build-webkit. 1606 1607 https://bugs.webkit.org/show_bug.cgi?id=32565 1608 1609 * qt/api/QtScript.pro: Added. 1610 * qt/api/qscriptconverter_p.h: Added. 1611 (QScriptConverter::toString): 1612 * qt/api/qscriptengine.cpp: Added. 1613 (QScriptEngine::QScriptEngine): 1614 (QScriptEngine::~QScriptEngine): 1615 (QScriptEngine::evaluate): 1616 (QScriptEngine::collectGarbage): 1617 * qt/api/qscriptengine.h: Added. 1618 * qt/api/qscriptengine_p.cpp: Added. 1619 (QScriptEnginePrivate::QScriptEnginePrivate): 1620 (QScriptEnginePrivate::~QScriptEnginePrivate): 1621 (QScriptEnginePrivate::evaluate): 1622 * qt/api/qscriptengine_p.h: Added. 1623 (QScriptEnginePrivate::get): 1624 (QScriptEnginePrivate::collectGarbage): 1625 (QScriptEnginePrivate::makeJSValue): 1626 (QScriptEnginePrivate::context): 1627 * qt/api/qscriptvalue.cpp: Added. 1628 (QScriptValue::QScriptValue): 1629 (QScriptValue::~QScriptValue): 1630 (QScriptValue::isValid): 1631 (QScriptValue::isBool): 1632 (QScriptValue::isBoolean): 1633 (QScriptValue::isNumber): 1634 (QScriptValue::isNull): 1635 (QScriptValue::isString): 1636 (QScriptValue::isUndefined): 1637 (QScriptValue::isError): 1638 (QScriptValue::isObject): 1639 (QScriptValue::isFunction): 1640 (QScriptValue::toString): 1641 (QScriptValue::toNumber): 1642 (QScriptValue::toBool): 1643 (QScriptValue::toBoolean): 1644 (QScriptValue::toInteger): 1645 (QScriptValue::toInt32): 1646 (QScriptValue::toUInt32): 1647 (QScriptValue::toUInt16): 1648 (QScriptValue::call): 1649 (QScriptValue::engine): 1650 (QScriptValue::operator=): 1651 (QScriptValue::equals): 1652 (QScriptValue::strictlyEquals): 1653 * qt/api/qscriptvalue.h: Added. 1654 (QScriptValue::): 1655 * qt/api/qscriptvalue_p.h: Added. 1656 (QScriptValuePrivate::): 1657 (QScriptValuePrivate::get): 1658 (QScriptValuePrivate::QScriptValuePrivate): 1659 (QScriptValuePrivate::isValid): 1660 (QScriptValuePrivate::isBool): 1661 (QScriptValuePrivate::isNumber): 1662 (QScriptValuePrivate::isNull): 1663 (QScriptValuePrivate::isString): 1664 (QScriptValuePrivate::isUndefined): 1665 (QScriptValuePrivate::isError): 1666 (QScriptValuePrivate::isObject): 1667 (QScriptValuePrivate::isFunction): 1668 (QScriptValuePrivate::toString): 1669 (QScriptValuePrivate::toNumber): 1670 (QScriptValuePrivate::toBool): 1671 (QScriptValuePrivate::toInteger): 1672 (QScriptValuePrivate::toInt32): 1673 (QScriptValuePrivate::toUInt32): 1674 (QScriptValuePrivate::toUInt16): 1675 (QScriptValuePrivate::equals): 1676 (QScriptValuePrivate::strictlyEquals): 1677 (QScriptValuePrivate::assignEngine): 1678 (QScriptValuePrivate::call): 1679 (QScriptValuePrivate::engine): 1680 (QScriptValuePrivate::context): 1681 (QScriptValuePrivate::value): 1682 (QScriptValuePrivate::object): 1683 (QScriptValuePrivate::inherits): 1684 (QScriptValuePrivate::isJSBased): 1685 (QScriptValuePrivate::isNumberBased): 1686 (QScriptValuePrivate::isStringBased): 1687 * qt/api/qtscriptglobal.h: Added. 1688 * qt/tests/qscriptengine/qscriptengine.pro: Added. 1689 * qt/tests/qscriptengine/tst_qscriptengine.cpp: Added. 1690 (tst_QScriptEngine::tst_QScriptEngine): 1691 (tst_QScriptEngine::~tst_QScriptEngine): 1692 (tst_QScriptEngine::init): 1693 (tst_QScriptEngine::cleanup): 1694 (tst_QScriptEngine::collectGarbage): 1695 (tst_QScriptEngine::evaluate): 1696 * qt/tests/qscriptvalue/qscriptvalue.pro: Added. 1697 * qt/tests/qscriptvalue/tst_qscriptvalue.cpp: Added. 1698 (tst_QScriptValue::tst_QScriptValue): 1699 (tst_QScriptValue::~tst_QScriptValue): 1700 (tst_QScriptValue::init): 1701 (tst_QScriptValue::cleanup): 1702 (tst_QScriptValue::ctor): 1703 (tst_QScriptValue::toString_data): 1704 (tst_QScriptValue::toString): 1705 (tst_QScriptValue::copyConstructor_data): 1706 (tst_QScriptValue::copyConstructor): 1707 (tst_QScriptValue::assignOperator_data): 1708 (tst_QScriptValue::assignOperator): 1709 (tst_QScriptValue::dataSharing): 1710 (tst_QScriptValue::constructors_data): 1711 (tst_QScriptValue::constructors): 1712 (tst_QScriptValue::call): 1713 * qt/tests/tests.pri: Added. 1714 * qt/tests/tests.pro: Added. 1715 17162010-01-25 Dmitry Titov <dimich@chromium.org> 1717 1718 Reviewed by David Levin. 1719 1720 Fix Chromium Linux tests: the pthread functions on Linux produce segfault if they receive 0 thread handle. 1721 After r53714, we can have 0 thread handles passed to pthread_join and pthread_detach if corresponding threads 1722 were already terminated and their threadMap entries cleared. 1723 Add a 0 check. 1724 1725 * wtf/ThreadingPthreads.cpp: 1726 (WTF::waitForThreadCompletion): 1727 (WTF::detachThread): 1728 17292010-01-24 Laszlo Gombos <laszlo.1.gombos@nokia.com> 1730 1731 Reviewed by Maciej Stachowiak. 1732 1733 Refactor JITStubs.cpp so that DEFINE_STUB_FUNCTION is only used once for each function 1734 https://bugs.webkit.org/show_bug.cgi?id=33866 1735 1736 Place the guard USE(JSVALUE32_64) inside the body of the DEFINE_STUB_FUNCTION 1737 macro for those functions that are always present. 1738 1739 * jit/JITStubs.cpp: 1740 (JSC::DEFINE_STUB_FUNCTION): 1741 17422010-01-22 Kevin Watters <kevinwatters@gmail.com> 1743 1744 Reviewed by Kevin Ollivier. 1745 1746 [wx] Remove the Bakefile build system, which is no longer being used. 1747 1748 https://bugs.webkit.org/show_bug.cgi?id=34022 1749 1750 * JavaScriptCoreSources.bkl: Removed. 1751 * jscore.bkl: Removed. 1752 17532010-01-22 Steve Falkenburg <sfalken@apple.com> 1754 1755 Reviewed by Darin Adler. 1756 1757 https://bugs.webkit.org/show_bug.cgi?id=34025 1758 Enable client-based Geolocation abstraction for Mac, Windows AppleWebKit targets. 1759 1760 * Configurations/FeatureDefines.xcconfig: 1761 17622010-01-22 Dmitry Titov <dimich@chromium.org> 1763 1764 Not reviewed, attempted Snow Leopard build fix. 1765 1766 * wtf/ThreadingPthreads.cpp: Add a forward declaration of a function which is not 'static'. 1767 17682009-01-22 Dmitry Titov <dimich@chromium.org> 1769 1770 Reviewed by Maciej Stachowiak. 1771 1772 Fix the leak of ThreadIdentifiers in threadMap across threads. 1773 https://bugs.webkit.org/show_bug.cgi?id=32689 1774 1775 Test is added to DumpRenderTree.mm. 1776 1777 * Android.mk: Added file ThreadIdentifierDataPthreads.(h|cpp) to build. 1778 * Android.v8.wtf.mk: Ditto. 1779 * GNUmakefile.am: Ditto. 1780 * JavaScriptCore.gyp/JavaScriptCore.gyp: Ditto. 1781 * JavaScriptCore.gypi: Ditto. 1782 * JavaScriptCore.xcodeproj/project.pbxproj: Ditto. 1783 1784 * wtf/ThreadIdentifierDataPthreads.cpp: Added. Contains custom implementation of thread-specific data that uses custom destructor. 1785 (WTF::ThreadIdentifierData::~ThreadIdentifierData): Removes the ThreadIdentifier from the threadMap. 1786 (WTF::ThreadIdentifierData::identifier): 1787 (WTF::ThreadIdentifierData::initialize): 1788 (WTF::ThreadIdentifierData::destruct): Custom thread-specific destructor. Resets the value for the key again to cause second invoke. 1789 (WTF::ThreadIdentifierData::initializeKeyOnceHelper): 1790 (WTF::ThreadIdentifierData::initializeKeyOnce): Need to use pthread_once since initialization may come on any thread(s). 1791 * wtf/ThreadIdentifierDataPthreads.h: Added. 1792 (WTF::ThreadIdentifierData::ThreadIdentifierData): 1793 1794 * wtf/Threading.cpp: 1795 (WTF::threadEntryPoint): Move initializeCurrentThreadInternal to after the lock to make 1796 sure it is invoked when ThreadIdentifier is already established. 1797 1798 * wtf/Threading.h: Rename setThreadNameInternal -> initializeCurrentThreadInternal since it does more then only set the name now. 1799 * wtf/ThreadingNone.cpp: 1800 (WTF::initializeCurrentThreadInternal): Ditto. 1801 * wtf/ThreadingWin.cpp: 1802 (WTF::initializeCurrentThreadInternal): Ditto. 1803 (WTF::initializeThreading): Ditto. 1804 * wtf/gtk/ThreadingGtk.cpp: 1805 (WTF::initializeCurrentThreadInternal): Ditto. 1806 * wtf/qt/ThreadingQt.cpp: 1807 (WTF::initializeCurrentThreadInternal): Ditto. 1808 1809 * wtf/ThreadingPthreads.cpp: 1810 (WTF::establishIdentifierForPthreadHandle): 1811 (WTF::clearPthreadHandleForIdentifier): Make it not 'static' so the ~ThreadIdentifierData() in another file can call it. 1812 (WTF::initializeCurrentThreadInternal): Set the thread-specific data. The ThreadIdentifier is already established by creating thread. 1813 (WTF::waitForThreadCompletion): Remove call to clearPthreadHandleForIdentifier(threadID) since it is now done in ~ThreadIdentifierData(). 1814 (WTF::detachThread): Ditto. 1815 (WTF::currentThread): Use the thread-specific data to get the ThreadIdentifier. It's many times faster then Mutex-protected iteration through the map. 1816 Also, set the thread-specific data if called first time on the thread. 1817 18182010-01-21 Kwang Yul Seo <skyul@company100.net> 1819 1820 Reviewed by Alexey Proskuryakov. 1821 1822 Add ThreadSpecific for ENABLE(SINGLE_THREADED) 1823 https://bugs.webkit.org/show_bug.cgi?id=33878 1824 1825 Implement ThreadSpecific with a simple getter/setter 1826 when ENABLE(SINGLE_THREADED) is true. 1827 1828 Due to the change in https://bugs.webkit.org/show_bug.cgi?id=33236, 1829 an implementation of ThreadSpecific must be available to build WebKit. 1830 This causes a build failure for platforms without a proper 1831 ThreadSpecific implementation. 1832 1833 * wtf/ThreadSpecific.h: 1834 (WTF::::ThreadSpecific): 1835 (WTF::::~ThreadSpecific): 1836 (WTF::::get): 1837 (WTF::::set): 1838 (WTF::::destroy): 1839 18402010-01-21 Kwang Yul Seo <skyul@company100.net> 1841 1842 Reviewed by Maciej Stachowiak. 1843 1844 Add fastStrDup to FastMalloc 1845 https://bugs.webkit.org/show_bug.cgi?id=33937 1846 1847 The new string returned by fastStrDup is obtained with fastMalloc, 1848 and can be freed with fastFree. This makes the memory management 1849 more consistent because we don't need to keep strdup allocated pointers 1850 and free them with free(). Instead we can use fastFree everywhere. 1851 1852 * wtf/FastMalloc.cpp: 1853 (WTF::fastStrDup): 1854 * wtf/FastMalloc.h: 1855 18562010-01-21 Brady Eidson <beidson@apple.com> 1857 1858 Reviewed by Maciej Stachowiak. 1859 1860 history.back() for same-document history traversals isn't synchronous as the specification states. 1861 <rdar://problem/7535011> and https://bugs.webkit.org/show_bug.cgi?id=33538 1862 1863 * wtf/Platform.h: Add a "HISTORY_ALWAYS_ASYNC" enable and turn it on for Chromium. 1864 18652010-01-21 Geoffrey Garen <ggaren@apple.com> 1866 1867 Reviewed by Oliver Hunt. 1868 1869 Always create a prototype for automatically managed classes. 1870 1871 This fixes some errors where prototype chains were not correctly hooked 1872 up, and also ensures that API classes work correctly with features like 1873 instanceof. 1874 1875 * API/JSClassRef.cpp: 1876 (OpaqueJSClass::create): Cleaned up some of this code. Also changed it 1877 to always create a prototype class. 1878 1879 * API/tests/testapi.c: 1880 (Derived2_class): 1881 (main): Fixed a null value crash in the exception checking code. 1882 * API/tests/testapi.js: Added some tests for the case where a prototype 1883 chain would not be hooked up correctly. 1884 18852010-01-21 Oliver Hunt <oliver@apple.com> 1886 1887 Reviewed by Geoff Garen. 1888 1889 Force JSC to create a prototype chain for API classes with a 1890 parent class but no static functions. 1891 1892 * API/JSClassRef.cpp: 1893 (OpaqueJSClass::create): 1894 18952010-01-21 Kent Hansen <kent.hansen@nokia.com> 1896 1897 Reviewed by Geoffrey Garen. 1898 1899 Object.getOwnPropertyDescriptor always returns undefined for JS API objects 1900 https://bugs.webkit.org/show_bug.cgi?id=33946 1901 1902 Ideally the getOwnPropertyDescriptor() reimplementation should return an 1903 access descriptor that wraps the property getter and setter callbacks, but 1904 that approach is much more involved than returning a value descriptor. 1905 Keep it simple for now. 1906 1907 * API/JSCallbackObject.h: 1908 * API/JSCallbackObjectFunctions.h: 1909 (JSC::::getOwnPropertyDescriptor): 1910 * API/tests/testapi.js: 1911 19122010-01-20 Mark Rowe <mrowe@apple.com> 1913 1914 Build fix. 1915 1916 * wtf/FastMalloc.cpp: 1917 (WTF::TCMalloc_PageHeap::initializeScavenger): Remove unnecessary function call. 1918 19192010-01-20 Mark Rowe <mrowe@apple.com> 1920 1921 Reviewed by Oliver Hunt. 1922 1923 Use the inline i386 assembly for x86_64 as well rather than falling back to using pthread mutexes. 1924 1925 * wtf/TCSpinLock.h: 1926 (TCMalloc_SpinLock::Lock): 1927 (TCMalloc_SpinLock::Unlock): 1928 (TCMalloc_SlowLock): 1929 19302010-01-20 Mark Rowe <mrowe@apple.com> 1931 1932 Reviewed by Oliver Hunt. 1933 1934 <rdar://problem/7215063> Use GCD instead of an extra thread for FastMalloc scavenging on platforms where it is supported 1935 1936 Abstract the background scavenging slightly so that an alternate implementation that uses GCD can be used on platforms 1937 where it is supported. 1938 1939 * wtf/FastMalloc.cpp: 1940 (WTF::TCMalloc_PageHeap::init): 1941 (WTF::TCMalloc_PageHeap::initializeScavenger): 1942 (WTF::TCMalloc_PageHeap::signalScavenger): 1943 (WTF::TCMalloc_PageHeap::shouldContinueScavenging): 1944 (WTF::TCMalloc_PageHeap::Delete): 1945 (WTF::TCMalloc_PageHeap::periodicScavenge): 1946 * wtf/Platform.h: 1947 19482010-01-20 Geoffrey Garen <ggaren@apple.com> 1949 1950 Reviewed by Oliver Hunt. 1951 1952 <rdar://problem/7562708> REGRESSION(53460): Heap::destroy may not run 1953 all destructors 1954 1955 * runtime/Collector.cpp: 1956 (JSC::Heap::freeBlocks): Instead of fully marking protected objects, 1957 just set their mark bits. This prevents protected objects from keeping 1958 unprotected objects alive. Destructor order is not guaranteed, so it's 1959 OK to destroy objects pointed to by protected objects before destroying 1960 protected objects. 1961 19622010-01-19 David Levin <levin@chromium.org> 1963 1964 Reviewed by Oliver Hunt. 1965 1966 CrossThreadCopier needs to support ThreadSafeShared better. 1967 https://bugs.webkit.org/show_bug.cgi?id=33698 1968 1969 * wtf/TypeTraits.cpp: Added tests for the new type traits. 1970 * wtf/TypeTraits.h: 1971 (WTF::IsSubclass): Determines if a class is a derived from another class. 1972 (WTF::IsSubclassOfTemplate): Determines if a class is a derived from a 1973 template class (with one parameter that is unknown). 1974 (WTF::RemoveTemplate): Reveals the type for a template parameter. 1975 19762010-01-20 Steve Falkenburg <sfalken@apple.com> 1977 1978 Reviewed by Darin Adler and Adam Roben. 1979 1980 Feature defines are difficult to maintain on Windows builds 1981 https://bugs.webkit.org/show_bug.cgi?id=33883 1982 1983 FeatureDefines.vsprops are now maintained in a way similar to 1984 Configurations/FeatureDefines.xcconfig, with the added advantage 1985 of having a single FeatureDefines file across all projects. 1986 1987 * Configurations/FeatureDefines.xcconfig: Add comments about keeping feature definitions in sync. 1988 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: Add FeatureDefines.vsprops inherited property sheet. 1989 * JavaScriptCore.vcproj/WTF/WTF.vcproj: Add FeatureDefines.vsprops inherited property sheet. 1990 19912010-01-20 Csaba Osztrogonác <ossy@webkit.org> 1992 1993 [Qt] Unreviewed buildfix for r53547. 1994 1995 * DerivedSources.pro: 1996 19972010-01-20 Tor Arne Vestbø <tor.arne.vestbo@nokia.com> 1998 1999 Reviewed by Simon Hausmann. 2000 2001 [Qt] Make extraCompilers for generated sources depend on their scripts 2002 2003 * DerivedSources.pro: 2004 20052010-01-19 Brian Weinstein <bweinstein@apple.com> 2006 2007 Reviewed by Tim Hatcher. 2008 2009 When JavaScriptCore calls Debugger::Exception, have it pass a 2010 hasHandler variable that represents if exception is being handled 2011 in the same function (not in a parent on the call stack). 2012 2013 This just adds a new parameter, no behavior is changed. 2014 2015 * debugger/Debugger.h: 2016 * interpreter/Interpreter.cpp: 2017 (JSC::Interpreter::throwException): 2018 20192010-01-18 Maciej Stachowiak <mjs@apple.com> 2020 2021 Reviewed by Adam Barth. 2022 2023 Inline functions that are hot in DOM manipulation 2024 https://bugs.webkit.org/show_bug.cgi?id=33820 2025 2026 (3% speedup on Dromaeo DOM Core tests) 2027 2028 * runtime/WeakGCMap.h: 2029 (JSC::::get): inline 2030 20312010-01-19 Laszlo Gombos <laszlo.1.gombos@nokia.com> 2032 2033 Unreviewed build fix for JIT with RVCT. 2034 2035 Remove IMPORT statement; cti_vm_throw is already defined in JITStubs.h. 2036 Remove extra ')'. 2037 2038 * jit/JITStubs.cpp: 2039 (JSC::ctiVMThrowTrampoline): 2040 20412010-01-19 Geoffrey Garen <ggaren@apple.com> 2042 2043 Reviewed by Oliver Hunt. 2044 2045 REGRESSION (52082): Crash on worker thread when reloading http://radnan.public.iastate.edu/procedural/ 2046 https://bugs.webkit.org/show_bug.cgi?id=33826 2047 2048 This bug was caused by a GC-protected object being destroyed early by 2049 Heap::destroy. Clients of the GC protect APIs (reasonably) expect pointers 2050 to GC-protected memory to be valid. 2051 2052 The solution is to do two passes of tear-down in Heap::destroy. The first 2053 pass tears down all unprotected objects. The second pass ASSERTs that all 2054 previously protected objects are now unprotected, and then tears down 2055 all perviously protected objects. These two passes simulate the two passes 2056 that would have been required to free a protected object during normal GC. 2057 2058 * API/JSContextRef.cpp: Removed some ASSERTs that have moved into Heap. 2059 2060 * runtime/Collector.cpp: 2061 (JSC::Heap::destroy): Moved ASSERTs to here. 2062 (JSC::Heap::freeBlock): Tidied up the use of didShrink by moving its 2063 setter to the function that does the shrinking. 2064 (JSC::Heap::freeBlocks): Implemented above algorithm. 2065 (JSC::Heap::shrinkBlocks): Tidied up the use of didShrink. 2066 20672010-01-19 Gavin Barraclough <barraclough@apple.com> 2068 2069 Reviewed by NOBODY (build fix). 2070 2071 Reverting r53455, breaks 2 javascriptcore tests. 2072 2073 * API/JSContextRef.cpp: 2074 * runtime/Collector.cpp: 2075 (JSC::Heap::destroy): 2076 (JSC::Heap::freeBlock): 2077 (JSC::Heap::freeBlocks): 2078 (JSC::Heap::shrinkBlocks): 2079 20802010-01-18 Gavin Barraclough <barraclough@apple.com> 2081 2082 Reviewed by NOBODY (build fix). 2083 2084 Revert r53454, since it causes much sadness in this world. 2085 2086 * runtime/UString.cpp: 2087 (JSC::UString::spliceSubstringsWithSeparators): 2088 (JSC::UString::replaceRange): 2089 * runtime/UStringImpl.cpp: 2090 (JSC::UStringImpl::baseSharedBuffer): 2091 (JSC::UStringImpl::sharedBuffer): 2092 (JSC::UStringImpl::~UStringImpl): 2093 * runtime/UStringImpl.h: 2094 (JSC::UntypedPtrAndBitfield::UntypedPtrAndBitfield): 2095 (JSC::UntypedPtrAndBitfield::asPtr): 2096 (JSC::UntypedPtrAndBitfield::operator&=): 2097 (JSC::UntypedPtrAndBitfield::operator|=): 2098 (JSC::UntypedPtrAndBitfield::operator&): 2099 (JSC::UStringImpl::create): 2100 (JSC::UStringImpl::cost): 2101 (JSC::UStringImpl::isIdentifier): 2102 (JSC::UStringImpl::setIsIdentifier): 2103 (JSC::UStringImpl::ref): 2104 (JSC::UStringImpl::deref): 2105 (JSC::UStringImpl::checkConsistency): 2106 (JSC::UStringImpl::UStringImpl): 2107 (JSC::UStringImpl::bufferOwnerString): 2108 (JSC::UStringImpl::bufferOwnership): 2109 (JSC::UStringImpl::isStatic): 2110 * wtf/StringHashFunctions.h: 2111 (WTF::stringHash): 2112 21132010-01-18 Geoffrey Garen <ggaren@apple.com> 2114 2115 Reviewed by Oliver Hunt. 2116 2117 REGRESSION (52082): Crash on worker thread when reloading http://radnan.public.iastate.edu/procedural/ 2118 https://bugs.webkit.org/show_bug.cgi?id=33826 2119 2120 This bug was caused by a GC-protected object being destroyed early by 2121 Heap::destroy. Clients of the GC protect APIs (reasonably) expect pointers 2122 to GC-protected memory to be valid. 2123 2124 The solution is to do two passes of tear-down in Heap::destroy. The first 2125 pass tears down all unprotected objects. The second pass ASSERTs that all 2126 previously protected objects are now unprotected, and then tears down 2127 all perviously protected objects. These two passes simulate the two passes 2128 that would have been required to free a protected object during normal GC. 2129 2130 * API/JSContextRef.cpp: Removed some ASSERTs that have moved into Heap. 2131 2132 * runtime/Collector.cpp: 2133 (JSC::Heap::destroy): Moved ASSERTs to here. 2134 (JSC::Heap::freeBlock): Tidied up the use of didShrink by moving its 2135 setter to the function that does the shrinking. 2136 (JSC::Heap::freeBlocks): Implemented above algorithm. 2137 (JSC::Heap::shrinkBlocks): Tidied up the use of didShrink. 2138 21392010-01-18 Gavin Barraclough <barraclough@apple.com> 2140 2141 Reviewed by Oliver Hunt. 2142 2143 https://bugs.webkit.org/show_bug.cgi?id=33731 2144 Remove UntypedPtrAndBitfield from UStringImpl (akin to PtrAndFlags). 2145 2146 This break the OS X Leaks tool. Instead, free up some more bits from the refCount. 2147 2148 * runtime/UStringImpl.cpp: 2149 (JSC::UStringImpl::sharedBuffer): 2150 (JSC::UStringImpl::~UStringImpl): 2151 * runtime/UStringImpl.h: 2152 (JSC::UStringImpl::cost): 2153 (JSC::UStringImpl::checkConsistency): 2154 (JSC::UStringImpl::UStringImpl): 2155 (JSC::UStringImpl::bufferOwnerString): 2156 (JSC::UStringImpl::): 2157 * wtf/StringHashFunctions.h: 2158 (WTF::stringHash): 2159 21602010-01-18 Kent Tamura <tkent@chromium.org> 2161 2162 Reviewed by Darin Adler. 2163 2164 HTMLInputElement::valueAsDate setter support for type=month. 2165 https://bugs.webkit.org/show_bug.cgi?id=33021 2166 2167 Expose the following functions to be used by WebCore: 2168 - WTF::msToyear() 2169 - WTF::dayInYear() 2170 - WTF::monthFromDayInYear() 2171 - WTF::dayInMonthFromDayInYear() 2172 2173 * JavaScriptCore.exp: 2174 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 2175 * wtf/DateMath.cpp: 2176 (WTF::msToYear): Remove "static inline". 2177 (WTF::dayInYear): Remove "static inline". 2178 (WTF::monthFromDayInYear): Remove "static inline". 2179 (WTF::dayInMonthFromDayInYear): Remove "static inline". 2180 * wtf/DateMath.h: Declare the above functions. 2181 21822010-01-18 Darin Adler <darin@apple.com> 2183 2184 Fix build by reverting the previous change. 2185 2186 * runtime/UString.h: Rolled out the FastAllocBase base class. 2187 It was making UString larger, and therefore JSString larger, 2188 and too big for a garbage collection cell. 2189 2190 This raises the unpleasant possibility that many classes became 2191 larger because we added the FastAllocBase base class. I am 2192 worried about this, and it needs to be investigated. 2193 21942010-01-18 Zoltan Horvath <zoltan@webkit.org> 2195 2196 Reviewed by Darin Adler. 2197 2198 Allow custom memory allocation control for UString class 2199 https://bugs.webkit.org/show_bug.cgi?id=27831 2200 2201 Inherits the following class from FastAllocBase because it is 2202 instantiated by 'new' and no need to be copyable: 2203 2204 class name - instantiated at: 2205 classs UString - JavaScriptCore/runtime/UString.cpp:160 2206 2207 * runtime/UString.h: 2208 22092010-01-18 Evan Cheng <evan.cheng@apple.com> 2210 2211 Reviewed by Darin Adler. 2212 2213 Add some ALWAYS_INLINE for key functions not inlined by some versions of GCC. 2214 rdar://problem/7553780 2215 2216 * runtime/JSObject.h: 2217 (JSC::JSObject::getPropertySlot): ALWAYS_INLINE both overloads. 2218 * runtime/JSString.h: 2219 (JSC::JSString::JSString): ALWAYS_INLINE the version that takes a UString. 2220 * runtime/UString.h: 2221 (JSC::operator==): ALWAYS_INLINE the version that compares two UString objects. 2222 22232010-01-18 Csaba Osztrogonác <ossy@webkit.org> 2224 2225 Reviewed by Darin Adler. 2226 2227 Delete dftables-xxxxxxxx.in files automatically. 2228 https://bugs.webkit.org/show_bug.cgi?id=33796 2229 2230 * pcre/dftables: unlink unnecessary temporary file. 2231 22322010-01-18 Tor Arne Vestbø <tor.arne.vestbo@nokia.com> 2233 2234 Reviewed by Simon Hausmann. 2235 2236 [Qt] Force qmake to generate a single makefile for DerivedSources.pro 2237 2238 * DerivedSources.pro: 2239 22402010-01-18 Csaba Osztrogonác <ossy@webkit.org> 2241 2242 Rubber-stamped by Gustavo Noronha Silva. 2243 2244 Rolling out r53391 and r53392 because of random crashes on buildbots. 2245 https://bugs.webkit.org/show_bug.cgi?id=33731 2246 2247 * bytecode/CodeBlock.h: 2248 (JSC::CallLinkInfo::seenOnce): 2249 (JSC::CallLinkInfo::setSeen): 2250 (JSC::MethodCallLinkInfo::MethodCallLinkInfo): 2251 (JSC::MethodCallLinkInfo::seenOnce): 2252 (JSC::MethodCallLinkInfo::setSeen): 2253 * jit/JIT.cpp: 2254 (JSC::JIT::unlinkCall): 2255 * jit/JITPropertyAccess.cpp: 2256 (JSC::JIT::patchMethodCallProto): 2257 * runtime/UString.cpp: 2258 (JSC::UString::spliceSubstringsWithSeparators): 2259 (JSC::UString::replaceRange): 2260 * runtime/UString.h: 2261 * runtime/UStringImpl.cpp: 2262 (JSC::UStringImpl::baseSharedBuffer): 2263 (JSC::UStringImpl::sharedBuffer): 2264 (JSC::UStringImpl::~UStringImpl): 2265 * runtime/UStringImpl.h: 2266 (JSC::UntypedPtrAndBitfield::UntypedPtrAndBitfield): 2267 (JSC::UntypedPtrAndBitfield::asPtr): 2268 (JSC::UntypedPtrAndBitfield::operator&=): 2269 (JSC::UntypedPtrAndBitfield::operator|=): 2270 (JSC::UntypedPtrAndBitfield::operator&): 2271 (JSC::UStringImpl::create): 2272 (JSC::UStringImpl::cost): 2273 (JSC::UStringImpl::isIdentifier): 2274 (JSC::UStringImpl::setIsIdentifier): 2275 (JSC::UStringImpl::ref): 2276 (JSC::UStringImpl::deref): 2277 (JSC::UStringImpl::checkConsistency): 2278 (JSC::UStringImpl::UStringImpl): 2279 (JSC::UStringImpl::bufferOwnerString): 2280 (JSC::UStringImpl::bufferOwnership): 2281 (JSC::UStringImpl::isStatic): 2282 * wtf/StringHashFunctions.h: 2283 (WTF::stringHash): 2284 22852010-01-18 Simon Hausmann <simon.hausmann@nokia.com> 2286 2287 Reviewed by Kenneth Rohde Christiansen. 2288 2289 Fix the build with strict gcc and RVCT versions: It's not legal to cast a 2290 pointer to a function to a void* without an intermediate cast to a non-pointer 2291 type. A cast to a ptrdiff_t inbetween fixes it. 2292 2293 * runtime/JSString.h: 2294 (JSC::Fiber::JSString): 2295 22962010-01-15 Gavin Barraclough <barraclough@apple.com> 2297 2298 Reviewed by Oliver Hunt. 2299 2300 https://bugs.webkit.org/show_bug.cgi?id=33731 2301 Remove UntypedPtrAndBitfield from UStringImpl (akin to PtrAndFlags). 2302 2303 This break the OS X Leaks tool. Instead, free up some more bits from the refCount. 2304 2305 * runtime/UStringImpl.cpp: 2306 (JSC::UStringImpl::sharedBuffer): 2307 (JSC::UStringImpl::~UStringImpl): 2308 * runtime/UStringImpl.h: 2309 (JSC::UStringImpl::cost): 2310 (JSC::UStringImpl::checkConsistency): 2311 (JSC::UStringImpl::UStringImpl): 2312 (JSC::UStringImpl::bufferOwnerString): 2313 (JSC::UStringImpl::): 2314 * wtf/StringHashFunctions.h: 2315 (WTF::stringHash): 2316 23172010-01-15 Gavin Barraclough <barraclough@apple.com> 2318 2319 Reviewed by Oliver Hunt. 2320 2321 https://bugs.webkit.org/show_bug.cgi?id=33731 2322 Remove uses of PtrAndFlags from JIT data stuctures. 2323 2324 These break the OS X Leaks tool. Free up a bit in CallLinkInfo, and invalid 2325 permutation of pointer states in MethodCallLinkInfo to represent the removed bits. 2326 2327 * bytecode/CodeBlock.h: 2328 (JSC::CallLinkInfo::seenOnce): 2329 (JSC::CallLinkInfo::setSeen): 2330 (JSC::MethodCallLinkInfo::MethodCallLinkInfo): 2331 (JSC::MethodCallLinkInfo::seenOnce): 2332 (JSC::MethodCallLinkInfo::setSeen): 2333 * jit/JIT.cpp: 2334 (JSC::JIT::unlinkCall): 2335 * jit/JITPropertyAccess.cpp: 2336 (JSC::JIT::patchMethodCallProto): 2337 * runtime/UString.h: 2338 23392010-01-16 Maciej Stachowiak <mjs@apple.com> 2340 2341 Reviewed by Oliver Hunt. 2342 2343 Cache JS string values made from DOM strings (Dromaeo speedup) 2344 https://bugs.webkit.org/show_bug.cgi?id=33768 2345 <rdar://problem/7353576> 2346 2347 * runtime/JSString.h: 2348 (JSC::jsStringWithFinalizer): Added new mechanism for a string to have an optional 2349 finalizer callback, for the benefit of weak-referencing caches. 2350 (JSC::): 2351 (JSC::Fiber::JSString): 2352 (JSC::Fiber::~JSString): 2353 * runtime/JSString.cpp: 2354 (JSC::JSString::resolveRope): Clear fibers so this doesn't look like a string with a finalizer. 2355 * runtime/WeakGCMap.h: Include "Collector.h" to make this header includable by itself. 2356 23572010-01-15 Sam Weinig <sam@webkit.org> 2358 2359 Reviewed by Maciej Stachowiak. 2360 2361 Fix for <rdar://problem/7548432> 2362 Add ALWAYS_INLINE to jsLess for a 1% speedup on llvm-gcc. 2363 2364 * runtime/Operations.h: 2365 (JSC::jsLess): 2366 23672010-01-14 Geoffrey Garen <ggaren@apple.com> 2368 2369 Reviewed by Oliver Hunt. 2370 2371 REGRESISON: Google maps buttons not working properly 2372 https://bugs.webkit.org/show_bug.cgi?id=31871 2373 2374 REGRESSION(r52948): JavaScript exceptions thrown on Google Maps when 2375 getting directions for a second time 2376 https://bugs.webkit.org/show_bug.cgi?id=33446 2377 2378 SunSpider and v8 report no change. 2379 2380 * interpreter/Interpreter.cpp: 2381 (JSC::Interpreter::tryCacheGetByID): Update our cached offset in case 2382 flattening the dictionary changed any of its offsets. 2383 2384 * jit/JITStubs.cpp: 2385 (JSC::JITThunks::tryCacheGetByID): 2386 (JSC::DEFINE_STUB_FUNCTION): 2387 * runtime/Operations.h: 2388 (JSC::normalizePrototypeChain): ditto 2389 23902010-01-14 Gavin Barraclough <barraclough@apple.com> 2391 2392 Reviewed by Oliver Hunt. 2393 2394 https://bugs.webkit.org/show_bug.cgi?id=33705 2395 UStringImpl::create() should use internal storage 2396 2397 When creating a UStringImpl copying of a UChar*, we can use an internal buffer, 2398 by calling UStringImpl::tryCreateUninitialized(). 2399 2400 Also, remove duplicate of copyChars from JSString, call UStringImpl's version. 2401 2402 Small (max 0.5%) progression on Sunspidey. 2403 2404 * runtime/JSString.cpp: 2405 (JSC::JSString::resolveRope): 2406 * runtime/UStringImpl.h: 2407 (JSC::UStringImpl::create): 2408 24092010-01-14 Gavin Barraclough <barraclough@apple.com> 2410 2411 Reviewed by Sam Weinig. 2412 2413 Make naming & behaviour of UString[Impl] methods more consistent. 2414 https://bugs.webkit.org/show_bug.cgi?id=33702 2415 2416 UString::create() creates a copy of the UChar* passed, but UStringImpl::create() assumes 2417 that it should assume ownership of the provided buffer (with UString::createNonCopying() 2418 and UStringImpl::createCopying() providing the alternate behaviours). Unify on create() 2419 taking a copy of the provided buffer. For non-copying cases, use the name 'adopt', and 2420 make this method take a Vector<UChar>&. For cases where non-copying construction was being 2421 used, other than from a Vector<UChar>, change the code to allocate the storage along with 2422 the UStringImpl using UStringImpl::createUninitialized(). (The adopt() method also more 2423 closely matches that of WebCore::StringImpl). 2424 2425 Also, UString::createUninitialized() and UStringImpl::createUninitialized() have incompatible 2426 behaviours, in that the UString form sets the provided UChar* to a null or non-null value to 2427 indicate success or failure, but UStringImpl uses the returned PassRefPtr<UStringImpl> to 2428 indicate when allocation has failed (potentially leaving the output Char* uninitialized). 2429 This is also incompatible with WebCore::StringImpl's behaviour, in that 2430 StringImpl::createUninitialized() will CRASH() if unable to allocate. Some uses of 2431 createUninitialized() in JSC are unsafe, since they do not test the result for null. 2432 UStringImpl's indication is preferable, since we may want a successful call to set the result 2433 buffer to 0 (specifically, StringImpl returns 0 for the buffer where createUninitialized() 2434 returns the empty string, which seems reasonable to catch bugs early). UString's method 2435 cannot support UStringImpl's behaviour directly, since it returns an object rather than a 2436 pointer. 2437 - remove UString::createUninitialized(), replace with calls to UStringImpl::createUninitialized() 2438 - create a UStringImpl::tryCreateUninitialized() form UStringImpl::createUninitialized(), 2439 with current behaviour, make createUninitialized() crash on failure to allocate. 2440 - make cases in JSC that do not check the result call createUninitialized(), and cases that do 2441 check call tryCreateUninitialized(). 2442 2443 Rename computedHash() to existingHash(), to bring this in line wih WebCore::StringImpl. 2444 2445 * API/JSClassRef.cpp: 2446 (OpaqueJSClassContextData::OpaqueJSClassContextData): 2447 * JavaScriptCore.exp: 2448 * runtime/ArrayPrototype.cpp: 2449 (JSC::arrayProtoFuncToString): 2450 * runtime/Identifier.cpp: 2451 (JSC::CStringTranslator::translate): 2452 (JSC::UCharBufferTranslator::translate): 2453 * runtime/JSString.cpp: 2454 (JSC::JSString::resolveRope): 2455 * runtime/Lookup.cpp: 2456 (JSC::HashTable::createTable): 2457 * runtime/Lookup.h: 2458 (JSC::HashTable::entry): 2459 * runtime/StringBuilder.h: 2460 (JSC::StringBuilder::release): 2461 * runtime/StringConstructor.cpp: 2462 (JSC::stringFromCharCodeSlowCase): 2463 * runtime/StringPrototype.cpp: 2464 (JSC::substituteBackreferencesSlow): 2465 (JSC::stringProtoFuncToLowerCase): 2466 (JSC::stringProtoFuncToUpperCase): 2467 (JSC::stringProtoFuncFontsize): 2468 (JSC::stringProtoFuncLink): 2469 * runtime/Structure.cpp: 2470 (JSC::Structure::despecifyDictionaryFunction): 2471 (JSC::Structure::get): 2472 (JSC::Structure::despecifyFunction): 2473 (JSC::Structure::put): 2474 (JSC::Structure::remove): 2475 (JSC::Structure::insertIntoPropertyMapHashTable): 2476 (JSC::Structure::checkConsistency): 2477 * runtime/Structure.h: 2478 (JSC::Structure::get): 2479 * runtime/StructureTransitionTable.h: 2480 (JSC::StructureTransitionTableHash::hash): 2481 * runtime/UString.cpp: 2482 (JSC::createRep): 2483 (JSC::UString::UString): 2484 (JSC::UString::spliceSubstringsWithSeparators): 2485 (JSC::UString::replaceRange): 2486 (JSC::UString::operator=): 2487 * runtime/UString.h: 2488 (JSC::UString::adopt): 2489 (JSC::IdentifierRepHash::hash): 2490 (JSC::makeString): 2491 * runtime/UStringImpl.h: 2492 (JSC::UStringImpl::adopt): 2493 (JSC::UStringImpl::create): 2494 (JSC::UStringImpl::createUninitialized): 2495 (JSC::UStringImpl::tryCreateUninitialized): 2496 (JSC::UStringImpl::existingHash): 2497 24982010-01-13 Kent Hansen <kent.hansen@nokia.com> 2499 2500 Reviewed by Oliver Hunt. 2501 2502 JSON.stringify and JSON.parse needlessly process properties in the prototype chain 2503 https://bugs.webkit.org/show_bug.cgi?id=33053 2504 2505 * runtime/JSONObject.cpp: 2506 (JSC::Stringifier::Holder::appendNextProperty): 2507 (JSC::Walker::walk): 2508 25092010-01-13 Gavin Barraclough <barraclough@apple.com> 2510 2511 Reviewed by NOBODY (buildfix). 2512 2513 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 2514 25152010-01-13 Alexey Proskuryakov <ap@apple.com> 2516 2517 Reviewed by Darin Adler. 2518 2519 https://bugs.webkit.org/show_bug.cgi?id=33641 2520 Assertion failure in Lexer.cpp if input stream ends while in string escape 2521 2522 Test: fast/js/end-in-string-escape.html 2523 2524 * parser/Lexer.cpp: (JSC::Lexer::lex): Bail out quickly on end of stream, not giving the 2525 assertion a chance to fire. 2526 25272010-01-13 Gavin Barraclough <barraclough@apple.com> 2528 2529 Reviewed by NOBODY (buildfix). 2530 2531 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 2532 25332010-01-13 Gavin Barraclough <barraclough@apple.com> 2534 2535 Rubber stamped by Sam Weinig & Darin Adler. 2536 2537 Three quick fixes to UStringImpl. 2538 - The destroy() method can be switched back to a normal destructor; since we've switched 2539 the way we protect static strings to be using an odd ref-count the destroy() won't abort. 2540 - The cost() calculation logic was wrong. If you have multiple JSStrings wrapping substrings 2541 of a base string, they would each report the full cost of the base string to the heap. 2542 Instead we should only be reporting once for the base string. 2543 - Remove the overloaded new operator calling fastMalloc, replace this with a 'using' to pick 2544 up the implementation from the parent class. 2545 2546 * JavaScriptCore.exp: 2547 * runtime/UStringImpl.cpp: 2548 (JSC::UStringImpl::~UStringImpl): 2549 * runtime/UStringImpl.h: 2550 (JSC::UStringImpl::cost): 2551 (JSC::UStringImpl::deref): 2552 25532010-01-13 Jocelyn Turcotte <jocelyn.turcotte@nokia.com> 2554 2555 Reviewed by Simon Hausmann. 2556 2557 [Qt] Split the build process in two different .pro files. 2558 This allows qmake to be run once all source files are available. 2559 2560 * DerivedSources.pro: Added. 2561 * JavaScriptCore.pri: Moved source generation to DerivedSources.pro 2562 * pcre/pcre.pri: Moved source generation to DerivedSources.pro 2563 25642010-01-12 Kent Hansen <kent.hansen@nokia.com> 2565 2566 Reviewed by Geoffrey Garen. 2567 2568 [ES5] Implement Object.getOwnPropertyNames 2569 https://bugs.webkit.org/show_bug.cgi?id=32242 2570 2571 Add an extra argument to getPropertyNames() and getOwnPropertyNames() 2572 (and all reimplementations thereof) that indicates whether non-enumerable 2573 properties should be added. 2574 2575 * API/JSCallbackObject.h: 2576 * API/JSCallbackObjectFunctions.h: 2577 (JSC::::getOwnPropertyNames): 2578 * JavaScriptCore.exp: 2579 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 2580 * debugger/DebuggerActivation.cpp: 2581 (JSC::DebuggerActivation::getOwnPropertyNames): 2582 * debugger/DebuggerActivation.h: 2583 * runtime/Arguments.cpp: 2584 (JSC::Arguments::getOwnPropertyNames): 2585 * runtime/Arguments.h: 2586 * runtime/CommonIdentifiers.h: 2587 * runtime/JSArray.cpp: 2588 (JSC::JSArray::getOwnPropertyNames): 2589 * runtime/JSArray.h: 2590 * runtime/JSByteArray.cpp: 2591 (JSC::JSByteArray::getOwnPropertyNames): 2592 * runtime/JSByteArray.h: 2593 * runtime/JSFunction.cpp: 2594 (JSC::JSFunction::getOwnPropertyNames): 2595 * runtime/JSFunction.h: 2596 * runtime/JSNotAnObject.cpp: 2597 (JSC::JSNotAnObject::getOwnPropertyNames): 2598 * runtime/JSNotAnObject.h: 2599 * runtime/JSObject.cpp: 2600 (JSC::getClassPropertyNames): 2601 (JSC::JSObject::getPropertyNames): 2602 (JSC::JSObject::getOwnPropertyNames): 2603 * runtime/JSObject.h: 2604 * runtime/JSVariableObject.cpp: 2605 (JSC::JSVariableObject::getOwnPropertyNames): 2606 * runtime/JSVariableObject.h: 2607 * runtime/ObjectConstructor.cpp: 2608 (JSC::ObjectConstructor::ObjectConstructor): 2609 (JSC::objectConstructorGetOwnPropertyNames): 2610 * runtime/RegExpMatchesArray.h: 2611 (JSC::RegExpMatchesArray::getOwnPropertyNames): 2612 * runtime/StringObject.cpp: 2613 (JSC::StringObject::getOwnPropertyNames): 2614 * runtime/StringObject.h: 2615 * runtime/Structure.cpp: Rename getEnumerablePropertyNames() to getPropertyNames(), which takes an extra argument. 2616 (JSC::Structure::getPropertyNames): 2617 * runtime/Structure.h: 2618 (JSC::): 2619 26202010-01-12 Alexey Proskuryakov <ap@apple.com> 2621 2622 Reviewed by Darin Adler. 2623 2624 https://bugs.webkit.org/show_bug.cgi?id=33540 2625 Make it possible to build in debug mode with assertions disabled 2626 2627 * jit/JITStubs.cpp: (JSC::DEFINE_STUB_FUNCTION): 2628 * runtime/Identifier.cpp: (JSC::Identifier::checkSameIdentifierTable): 2629 * wtf/FastMalloc.cpp: 2630 * wtf/HashTable.h: (WTF::HashTableConstIterator::checkValidity): 2631 * yarr/RegexCompiler.cpp: (JSC::Yarr::compileRegex): 2632 26332009-11-23 Yong Li <yoli@rim.com> 2634 2635 Reviewed by Adam Treat. 2636 2637 Make GIF decoder support down-sampling 2638 https://bugs.webkit.org/show_bug.cgi?id=31806 2639 2640 * platform/image-decoders/ImageDecoder.cpp: 2641 (WebCore::ImageDecoder::upperBoundScaledY): 2642 (WebCore::ImageDecoder::lowerBoundScaledY): 2643 * platform/image-decoders/ImageDecoder.h: 2644 (WebCore::RGBA32Buffer::scaledRect): 2645 (WebCore::RGBA32Buffer::setScaledRect): 2646 (WebCore::ImageDecoder::scaledSize): 2647 * platform/image-decoders/gif/GIFImageDecoder.cpp: 2648 (WebCore::GIFImageDecoder::sizeNowAvailable): 2649 (WebCore::GIFImageDecoder::initFrameBuffer): 2650 (WebCore::copyOnePixel): 2651 (WebCore::GIFImageDecoder::haveDecodedRow): 2652 (WebCore::GIFImageDecoder::frameComplete): 2653 26542010-01-12 Adam Barth <abarth@webkit.org> 2655 2656 Reviewed by Eric Seidel. 2657 2658 ecma/Date/15.9.5.12-1.js fails every night at midnight 2659 https://bugs.webkit.org/show_bug.cgi?id=28041 2660 2661 Change the test to use a concrete time instead of "now". 2662 2663 * tests/mozilla/ecma/Date/15.9.5.10-1.js: 2664 * tests/mozilla/ecma/Date/15.9.5.12-1.js: 2665 26662010-01-11 Csaba Osztrogonác <ossy@webkit.org> 2667 2668 Reviewed by Ariya Hidayat. 2669 2670 [Qt] Enable JIT and YARR_JIT if (CPU(X86_64) && OS(LINUX) && GCC_VERSION >= 40100) 2671 2672 * wtf/Platform.h: 2673 26742010-01-11 Geoffrey Garen <ggaren@apple.com> 2675 2676 Reviewed by Alexey Proskuryakov. 2677 2678 https://bugs.webkit.org/show_bug.cgi?id=33481 2679 Uninitialized data members in ArrayStorage 2680 2681 SunSpider reports no change. 2682 2683 * runtime/JSArray.cpp: 2684 (JSC::JSArray::JSArray): Initialize missing data members in the two cases 2685 where we don't use fastZeroedMalloc, so it doesn't happen automatically. 2686 26872010-01-11 Steve Falkenburg <sfalken@apple.com> 2688 2689 Reviewed by Sam Weinig. 2690 2691 https://bugs.webkit.org/show_bug.cgi?id=33480 2692 2693 Improve debugging reliability for WTF on Windows. 2694 Store WTF static library's PDB file into a better location. 2695 2696 * JavaScriptCore.vcproj/WTF/WTF.vcproj: 2697 26982010-01-11 Steve Falkenburg <sfalken@apple.com> 2699 2700 Windows build fix. 2701 Remove extraneous entries from def file causing build warning. 2702 2703 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 2704 27052010-01-10 Kent Hansen <kent.hansen@nokia.com> 2706 2707 Reviewed by Darin Adler. 2708 2709 RegExp.prototype.toString returns "//" for empty regular expressions 2710 https://bugs.webkit.org/show_bug.cgi?id=33319 2711 2712 "//" starts a single-line comment, hence "/(?:)/" should be used, according to ECMA. 2713 2714 * runtime/RegExpPrototype.cpp: 2715 (JSC::regExpProtoFuncToString): 2716 2717 * tests/mozilla/ecma_2/RegExp/properties-001.js: 2718 (AddRegExpCases): 2719 * tests/mozilla/js1_2/regexp/toString.js: 2720 Update relevant Mozilla tests (Mozilla has had this behavior since November 2003). 2721 27222010-01-10 Darin Adler <darin@apple.com> 2723 2724 * tests/mozilla/ecma/Array/15.4.1.1.js: Added property allow-tabs. 2725 * tests/mozilla/ecma/Array/15.4.1.2.js: Added property allow-tabs. 2726 * tests/mozilla/ecma/Array/15.4.2.1-1.js: Added property allow-tabs. 2727 * tests/mozilla/ecma/Array/15.4.2.2-1.js: Added property allow-tabs. 2728 * tests/mozilla/ecma/Array/15.4.2.2-2.js: Added property allow-tabs. 2729 * tests/mozilla/ecma/Array/15.4.2.3.js: Added property allow-tabs. 2730 * tests/mozilla/ecma/Array/15.4.3.2.js: Added property allow-tabs. 2731 * tests/mozilla/ecma/Array/15.4.3.js: Added property allow-tabs. 2732 * tests/mozilla/ecma/Array/15.4.4.1.js: Added property allow-tabs. 2733 * tests/mozilla/ecma/Array/15.4.4.js: Added property allow-tabs. 2734 * tests/mozilla/ecma/LexicalConventions/7.7.4.js: Added property allow-tabs. 2735 * tests/mozilla/ecma/Math/15.8.2.13.js: Added property allow-tabs. 2736 * tests/mozilla/ecma/Math/15.8.2.16.js: Added property allow-tabs. 2737 * tests/mozilla/ecma/Math/15.8.2.18.js: Added property allow-tabs. 2738 * tests/mozilla/ecma/Math/15.8.2.2.js: Added property allow-tabs. 2739 * tests/mozilla/ecma/Math/15.8.2.4.js: Added property allow-tabs. 2740 * tests/mozilla/ecma/Math/15.8.2.5.js: Added property allow-tabs. 2741 * tests/mozilla/ecma/Math/15.8.2.7.js: Added property allow-tabs. 2742 * tests/mozilla/ecma/String/15.5.1.js: Added property allow-tabs. 2743 * tests/mozilla/ecma/String/15.5.2.js: Added property allow-tabs. 2744 * tests/mozilla/ecma/String/15.5.3.1-3.js: Added property allow-tabs. 2745 * tests/mozilla/ecma/String/15.5.3.1-4.js: Added property allow-tabs. 2746 * tests/mozilla/ecma/String/15.5.3.js: Added property allow-tabs. 2747 * tests/mozilla/ecma/TypeConversion/9.5-2.js: Added property allow-tabs. 2748 * tests/mozilla/ecma/jsref.js: Modified property allow-tabs. 2749 * tests/mozilla/ecma/shell.js: Modified property allow-tabs. 2750 * tests/mozilla/ecma_2/LexicalConventions/keywords-001.js: Added property allow-tabs. 2751 * tests/mozilla/ecma_2/RegExp/exec-001.js: Added property allow-tabs. 2752 * tests/mozilla/ecma_2/String/match-004.js: Added property allow-tabs. 2753 * tests/mozilla/ecma_2/String/replace-001.js: Added property allow-tabs. 2754 * tests/mozilla/ecma_2/String/split-002.js: Added property allow-tabs. 2755 * tests/mozilla/ecma_2/jsref.js: Modified property allow-tabs. 2756 * tests/mozilla/ecma_2/shell.js: Added property allow-tabs. 2757 * tests/mozilla/ecma_3/Date/shell.js: Modified property allow-tabs. 2758 * tests/mozilla/ecma_3/Exceptions/regress-181654.js: Added property allow-tabs. 2759 * tests/mozilla/ecma_3/RegExp/regress-209067.js: Added property allow-tabs. 2760 * tests/mozilla/ecma_3/RegExp/regress-85721.js: Added property allow-tabs. 2761 * tests/mozilla/importList.html: Added property allow-tabs. 2762 * tests/mozilla/js1_1/shell.js: Added property allow-tabs. 2763 * tests/mozilla/js1_2/Array/general1.js: Added property allow-tabs. 2764 * tests/mozilla/js1_2/Array/general2.js: Added property allow-tabs. 2765 * tests/mozilla/js1_2/Array/slice.js: Added property allow-tabs. 2766 * tests/mozilla/js1_2/Array/splice1.js: Added property allow-tabs. 2767 * tests/mozilla/js1_2/Array/splice2.js: Added property allow-tabs. 2768 * tests/mozilla/js1_2/Objects/toString-001.js: Added property allow-tabs. 2769 * tests/mozilla/js1_2/String/charCodeAt.js: Added property allow-tabs. 2770 * tests/mozilla/js1_2/String/concat.js: Modified property allow-tabs. 2771 * tests/mozilla/js1_2/String/match.js: Added property allow-tabs. 2772 * tests/mozilla/js1_2/String/slice.js: Added property allow-tabs. 2773 * tests/mozilla/js1_2/function/Function_object.js: Added property allow-tabs. 2774 * tests/mozilla/js1_2/function/Number.js: Modified property allow-tabs. 2775 * tests/mozilla/js1_2/function/String.js: Modified property allow-tabs. 2776 * tests/mozilla/js1_2/function/nesting.js: Added property allow-tabs. 2777 * tests/mozilla/js1_2/function/regexparg-1.js: Added property allow-tabs. 2778 * tests/mozilla/js1_2/function/regexparg-2-n.js: Added property allow-tabs. 2779 * tests/mozilla/js1_2/jsref.js: Added property allow-tabs. 2780 * tests/mozilla/js1_2/operator/equality.js: Added property allow-tabs. 2781 * tests/mozilla/js1_2/operator/strictEquality.js: Added property allow-tabs. 2782 * tests/mozilla/js1_2/regexp/RegExp_dollar_number.js: Added property allow-tabs. 2783 * tests/mozilla/js1_2/regexp/RegExp_input.js: Added property allow-tabs. 2784 * tests/mozilla/js1_2/regexp/RegExp_input_as_array.js: Added property allow-tabs. 2785 * tests/mozilla/js1_2/regexp/RegExp_lastIndex.js: Added property allow-tabs. 2786 * tests/mozilla/js1_2/regexp/RegExp_lastMatch.js: Added property allow-tabs. 2787 * tests/mozilla/js1_2/regexp/RegExp_lastMatch_as_array.js: Added property allow-tabs. 2788 * tests/mozilla/js1_2/regexp/RegExp_lastParen.js: Added property allow-tabs. 2789 * tests/mozilla/js1_2/regexp/RegExp_lastParen_as_array.js: Added property allow-tabs. 2790 * tests/mozilla/js1_2/regexp/RegExp_leftContext.js: Added property allow-tabs. 2791 * tests/mozilla/js1_2/regexp/RegExp_leftContext_as_array.js: Added property allow-tabs. 2792 * tests/mozilla/js1_2/regexp/RegExp_multiline.js: Added property allow-tabs. 2793 * tests/mozilla/js1_2/regexp/RegExp_multiline_as_array.js: Added property allow-tabs. 2794 * tests/mozilla/js1_2/regexp/RegExp_object.js: Added property allow-tabs. 2795 * tests/mozilla/js1_2/regexp/RegExp_rightContext.js: Added property allow-tabs. 2796 * tests/mozilla/js1_2/regexp/RegExp_rightContext_as_array.js: Added property allow-tabs. 2797 * tests/mozilla/js1_2/regexp/alphanumeric.js: Added property allow-tabs. 2798 * tests/mozilla/js1_2/regexp/asterisk.js: Added property allow-tabs. 2799 * tests/mozilla/js1_2/regexp/backslash.js: Added property allow-tabs. 2800 * tests/mozilla/js1_2/regexp/backspace.js: Added property allow-tabs. 2801 * tests/mozilla/js1_2/regexp/beginLine.js: Added property allow-tabs. 2802 * tests/mozilla/js1_2/regexp/character_class.js: Added property allow-tabs. 2803 * tests/mozilla/js1_2/regexp/compile.js: Added property allow-tabs. 2804 * tests/mozilla/js1_2/regexp/control_characters.js: Added property allow-tabs. 2805 * tests/mozilla/js1_2/regexp/digit.js: Added property allow-tabs. 2806 * tests/mozilla/js1_2/regexp/dot.js: Added property allow-tabs. 2807 * tests/mozilla/js1_2/regexp/endLine.js: Added property allow-tabs. 2808 * tests/mozilla/js1_2/regexp/everything.js: Added property allow-tabs. 2809 * tests/mozilla/js1_2/regexp/exec.js: Added property allow-tabs. 2810 * tests/mozilla/js1_2/regexp/flags.js: Added property allow-tabs. 2811 * tests/mozilla/js1_2/regexp/global.js: Added property allow-tabs. 2812 * tests/mozilla/js1_2/regexp/hexadecimal.js: Added property allow-tabs. 2813 * tests/mozilla/js1_2/regexp/ignoreCase.js: Added property allow-tabs. 2814 * tests/mozilla/js1_2/regexp/interval.js: Added property allow-tabs. 2815 * tests/mozilla/js1_2/regexp/octal.js: Added property allow-tabs. 2816 * tests/mozilla/js1_2/regexp/parentheses.js: Added property allow-tabs. 2817 * tests/mozilla/js1_2/regexp/plus.js: Added property allow-tabs. 2818 * tests/mozilla/js1_2/regexp/question_mark.js: Added property allow-tabs. 2819 * tests/mozilla/js1_2/regexp/simple_form.js: Added property allow-tabs. 2820 * tests/mozilla/js1_2/regexp/source.js: Added property allow-tabs. 2821 * tests/mozilla/js1_2/regexp/special_characters.js: Added property allow-tabs. 2822 * tests/mozilla/js1_2/regexp/string_replace.js: Added property allow-tabs. 2823 * tests/mozilla/js1_2/regexp/string_search.js: Added property allow-tabs. 2824 * tests/mozilla/js1_2/regexp/string_split.js: Added property allow-tabs. 2825 * tests/mozilla/js1_2/regexp/test.js: Added property allow-tabs. 2826 * tests/mozilla/js1_2/regexp/toString.js: Added property allow-tabs. 2827 * tests/mozilla/js1_2/regexp/vertical_bar.js: Added property allow-tabs. 2828 * tests/mozilla/js1_2/regexp/whitespace.js: Added property allow-tabs. 2829 * tests/mozilla/js1_2/regexp/word_boundary.js: Added property allow-tabs. 2830 * tests/mozilla/js1_2/shell.js: Added property allow-tabs. 2831 * tests/mozilla/js1_2/statements/break.js: Added property allow-tabs. 2832 * tests/mozilla/js1_2/statements/continue.js: Added property allow-tabs. 2833 * tests/mozilla/js1_2/statements/do_while.js: Added property allow-tabs. 2834 * tests/mozilla/js1_2/statements/switch.js: Added property allow-tabs. 2835 * tests/mozilla/js1_2/statements/switch2.js: Added property allow-tabs. 2836 * tests/mozilla/js1_3/shell.js: Added property allow-tabs. 2837 * tests/mozilla/js1_4/shell.js: Added property allow-tabs. 2838 * tests/mozilla/js1_5/Regress/regress-111557.js: Added property allow-tabs. 2839 * tests/mozilla/js1_5/Regress/regress-216320.js: Added property allow-tabs. 2840 * tests/mozilla/menuhead.html: Added property allow-tabs. 2841 * tests/mozilla/mklistpage.pl: Added property allow-tabs. 2842 * tests/mozilla/runtests.pl: Added property allow-tabs. 2843 28442010-01-08 Daniel Bates <dbates@webkit.org> 2845 2846 Reviewed by Adam Barth. 2847 2848 https://bugs.webkit.org/show_bug.cgi?id=33417 2849 2850 Cleans up style errors exposed by the patch for bug #33198. 2851 Moreover, fixes all "Weird number of spaces at line-start. Are you using a 4-space indent?" 2852 errors reported by check-webkit-style. 2853 2854 No functionality was changed. So, no new tests. 2855 2856 * wtf/Platform.h: 2857 28582010-01-08 Kent Hansen <kent.hansen@nokia.com> 2859 2860 Reviewed by Eric Seidel. 2861 2862 Don't store RegExp flags string representation 2863 https://bugs.webkit.org/show_bug.cgi?id=33321 2864 2865 It's unused; the string representation is reconstructed from flags. 2866 2867 * runtime/RegExp.cpp: 2868 (JSC::RegExp::RegExp): 2869 * runtime/RegExp.h: 2870 28712010-01-08 Geoffrey Garen <ggaren@apple.com> 2872 2873 Reviewed by Oliver Hunt. 2874 2875 Memory use grows grows possibly unbounded in this JavaScript Array test case 2876 https://bugs.webkit.org/show_bug.cgi?id=31675 2877 2878 This fixes one observed bug in this test case, which is that 2879 arrays don't report extra cost for the sparse value maps. 2880 2881 SunSpider reports a small speedup. 2882 2883 * runtime/JSArray.cpp: 2884 (JSC::JSArray::putSlowCase): Report extra memory cost for 2885 the sparse value map. 2886 * runtime/JSArray.h: 2887 28882010-01-08 Yong Li <yoli@rim.com> 2889 2890 Reviewed by Darin Adler. 2891 2892 Remove unnecessary #include from FastMalloc.cpp 2893 https://bugs.webkit.org/show_bug.cgi?id=33393 2894 2895 * wtf/FastMalloc.cpp: 2896 28972010-01-08 Eric Seidel <eric@webkit.org> 2898 2899 No review, rolling out r52983. 2900 http://trac.webkit.org/changeset/52983 2901 https://bugs.webkit.org/show_bug.cgi?id=33321 2902 2903 Broke 59 JavaScriptCore tests. I don't think Kent knew about 2904 run-javascriptcore-tests. Sadly neither does the commit-bot, 2905 yet. 2906 2907 * runtime/RegExp.cpp: 2908 (JSC::RegExp::RegExp): 2909 * runtime/RegExp.h: 2910 (JSC::RegExp::flags): 2911 29122010-01-08 Eric Seidel <eric@webkit.org> 2913 2914 No review, rolling out r52981. 2915 http://trac.webkit.org/changeset/52981 2916 https://bugs.webkit.org/show_bug.cgi?id=33319 2917 2918 Caused two JS tests to start failing: 2919 ecma_2/RegExp/properties-001.js and js1_2/regexp/toString.js 2920 2921 * runtime/RegExpPrototype.cpp: 2922 (JSC::regExpProtoFuncToString): 2923 29242010-01-08 Kent Hansen <kent.hansen@nokia.com> 2925 2926 Reviewed by Darin Adler. 2927 2928 Don't store RegExp flags string representation 2929 https://bugs.webkit.org/show_bug.cgi?id=33321 2930 2931 It's unused; the string representation is reconstructed from flags. 2932 2933 * runtime/RegExp.cpp: 2934 (JSC::RegExp::RegExp): 2935 * runtime/RegExp.h: 2936 29372010-01-08 Kent Hansen <kent.hansen@nokia.com> 2938 2939 Reviewed by Darin Adler. 2940 2941 RegExp.prototype.toString returns "//" for empty regular expressions 2942 https://bugs.webkit.org/show_bug.cgi?id=33319 2943 2944 "//" starts a single-line comment, hence "/(?:)/" should be used, according to ECMA. 2945 2946 * runtime/RegExpPrototype.cpp: 2947 (JSC::regExpProtoFuncToString): 2948 29492010-01-08 Norbert Leser <norbert.leser@nokia.com> 2950 2951 Reviewed by Darin Adler. 2952 2953 RVCT compiler with "-Otime -O3" optimization tries to optimize out 2954 inline new'ed pointers that are passed as arguments. 2955 Proposed patch assigns new'ed pointer explicitly outside function call. 2956 2957 https://bugs.webkit.org/show_bug.cgi?id=33084 2958 2959 * API/JSClassRef.cpp: 2960 (OpaqueJSClass::OpaqueJSClass): 2961 (OpaqueJSClassContextData::OpaqueJSClassContextData): 2962 29632010-01-08 Gabor Loki <loki@webkit.org> 2964 2965 Reviewed by Gavin Barraclough. 2966 2967 Remove an unnecessary cacheFlush from ARM_TRADITIONAL JIT 2968 https://bugs.webkit.org/show_bug.cgi?id=33203 2969 2970 * assembler/ARMAssembler.cpp: Remove obsolete linkBranch function. 2971 (JSC::ARMAssembler::executableCopy): Inline a clean linkBranch code. 2972 * assembler/ARMAssembler.h: 2973 (JSC::ARMAssembler::getLdrImmAddress): Use inline function. 2974 (JSC::ARMAssembler::getLdrImmAddressOnPool): Ditto. 2975 (JSC::ARMAssembler::patchPointerInternal): Remove an unnecessary cacheFlush. 2976 (JSC::ARMAssembler::linkJump): Use patchPointerInternal instead of linkBranch. 2977 (JSC::ARMAssembler::linkCall): Ditto. 2978 (JSC::ARMAssembler::relinkCall): Ditto. 2979 29802010-01-07 Gabor Loki <loki@webkit.org> 2981 2982 Reviewed by Gavin Barraclough. 2983 2984 Build fix for JSVALUE32 when ENABLE_JIT_OPTIMIZE* are disabled 2985 https://bugs.webkit.org/show_bug.cgi?id=33311 2986 2987 Move compileGetDirectOffset function to common part of JSVALUE32 2988 2989 * jit/JITPropertyAccess.cpp: 2990 (JSC::JIT::compileGetDirectOffset): 2991 29922010-01-07 Laszlo Gombos <laszlo.1.gombos@nokia.com> 2993 2994 Reviewed by Maciej Stachowiak. 2995 2996 Allow call sites to determine if ASSERT_* and LOG_* macros are operational 2997 https://bugs.webkit.org/show_bug.cgi?id=33020 2998 2999 * wtf/Assertions.h: Set ASSERT_MSG_DISABLED, FATAL_DISABLED, 3000 ERROR_DISABLED, LOG_DISABLED to 1 if the compiler does not support 3001 variadic macros. Refactor for better readibility. 3002 30032010-01-07 Daniel Bates <dbates@rim.com> 3004 3005 Reviewed by Eric Seidel. 3006 3007 https://bugs.webkit.org/show_bug.cgi?id=32987 3008 3009 Added ENABLE_XHTMLMP flag. Disabled by default. 3010 3011 * Configurations/FeatureDefines.xcconfig: 3012 30132010-01-07 Laszlo Gombos <laszlo.1.gombos@nokia.com> 3014 3015 Reviewed by Gavin Barraclough. 3016 3017 [Symbian] Port ARM traditional JIT Trampolines to RVCT 3018 https://bugs.webkit.org/show_bug.cgi?id=30552 3019 3020 Take the GCC implementation and mechanically convert 3021 it to RVCT syntax. 3022 3023 Use 'bx rX' instead of 'mov pc, rX' when it is available. 3024 3025 Developed in cooperation with Iain Campbell and Gabor Loki. 3026 3027 * JavaScriptCore.pri: Extra step to generate RVCT stubs. The 3028 script generation intentionally executed all the time not just 3029 for RVCT targets. 3030 3031 * create_rvct_stubs: Added. Perl script to expand precompiler macros 3032 for RVCT assembler - the template is defined in JITStubs.cpp. 3033 3034 * jit/JITStubs.cpp: 3035 (JSC::ctiTrampoline): 3036 (JSC::ctiVMThrowTrampoline): 3037 (JSC::ctiOpThrowNotCaught): 3038 30392010-01-07 Geoffrey Garen <ggaren@apple.com> 3040 3041 Reviewed by Sam Weinig. 3042 3043 Fix a crash seen on the buildbots. 3044 3045 * runtime/JSGlobalObject.cpp: 3046 (JSC::JSGlobalObject::init): Disable specific function tracking here, 3047 instead of in WebCore, to ensure that the disabling happens before a 3048 specific function can be registered. 3049 30502010-01-07 Alexey Proskuryakov <ap@apple.com> 3051 3052 Mac build fix. 3053 3054 * JavaScriptCore.exp: Export new JSGlobalData static data members. 3055 30562010-01-07 Alexey Proskuryakov <ap@apple.com> 3057 3058 Reviewed by Geoffrey Garen. 3059 3060 https://bugs.webkit.org/show_bug.cgi?id=33057 3061 REGRESSION(r49365): typeof(xhr.responseText) != "string" in Windows 3062 3063 <rdar://problem/7296920> REGRESSION: WebKit fails to start PeaceKeeper benchmark 3064 3065 Test: fast/js/webcore-string-comparison.html 3066 3067 In r49365, some code was moved from JSString.cpp to JSString.h, and as a result, WebCore 3068 got a way to directly instantiate JSStrings over DLL borders. Since vftable for JSString was 3069 not exported, objects created from WebCore got a different vptr, and JavaScriptCore 3070 optimizations that relied on vptr of all JSString objects being equal failed. 3071 3072 * config.h: Added a JS_EXPORTCLASS macro for exporting classes. It's currently the same as 3073 JS_EXPORTDATA, but it clearly needed a new name. 3074 3075 * runtime/InitializeThreading.cpp: 3076 (JSC::initializeThreadingOnce): 3077 * runtime/JSGlobalData.cpp: 3078 (JSC::JSGlobalData::storeVPtrs): 3079 (JSC::JSGlobalData::JSGlobalData): 3080 (JSC::JSGlobalData::createNonDefault): 3081 (JSC::JSGlobalData::create): 3082 (JSC::JSGlobalData::sharedInstance): 3083 * runtime/JSGlobalData.h: 3084 Store vptrs just once, no need to repeatedly pick and copy them. This makes it possible to 3085 assert vptr correctness in object destructors (which don't have access to JSGlobalData, 3086 and even Heap::heap(this) will fail for fake objects created from storeVPtrs()). 3087 3088 * runtime/JSArray.cpp: (JSC::JSArray::~JSArray): Assert that vptr is what we expect it to be. 3089 It's important to assert in destructor, because MSVC changes the vptr after constructor 3090 is invoked. 3091 * runtime/JSByteArray.cpp: (JSC::JSByteArray::~JSByteArray): Ditto. 3092 * runtime/JSByteArray.h: Ditto. 3093 * runtime/JSFunction.h: Ditto. 3094 * runtime/JSFunction.cpp: (JSC::JSFunction::~JSFunction): Ditto. 3095 3096 * runtime/JSCell.h: (JSC::JSCell::setVPtr): Added a method to substitute vptr for another 3097 one. 3098 3099 * runtime/JSString.h: Export JSString class together with its vftable, and tell other 3100 libraries tp import it. This is needed on platforms that have a separate JavaScriptCore 3101 dynamic library - and on Mac, we already did the export via JavaScriptCore.exp. 3102 (JSC::JSString::~JSString): Assert tha vptr is what we expect it to be. 3103 (JSC::fixupVPtr): Store a previously saved primary vftable pointer (do nothing if building 3104 JavaScriptCore itself). 3105 (JSC::jsSingleCharacterString): Call fixupVPtr in case this is call across DLL boundary. 3106 (JSC::jsSingleCharacterSubstring): Ditto. 3107 (JSC::jsNontrivialString): Ditto. 3108 (JSC::jsString): Ditto. 3109 (JSC::jsSubstring): Ditto. 3110 (JSC::jsOwnedString): Ditto. 3111 3112 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: Export the new static 3113 JSGlobalData members that are used in WebCore via inline functions. 3114 31152010-01-07 Geoffrey Garen <ggaren@apple.com> 3116 3117 Reviewed by Sam Weinig. 3118 3119 Safari memory usage skyrockets using new Google AdWords interface 3120 https://bugs.webkit.org/show_bug.cgi?id=33343 3121 3122 The memory use was caused by the global object creating too many structures 3123 as it thrashed between different specific functions. 3124 3125 * runtime/Structure.cpp: 3126 (JSC::Structure::Structure): 3127 (JSC::Structure::addPropertyTransition): 3128 (JSC::Structure::changePrototypeTransition): 3129 (JSC::Structure::despecifyFunctionTransition): 3130 (JSC::Structure::addAnonymousSlotsTransition): 3131 (JSC::Structure::getterSetterTransition): 3132 (JSC::Structure::toDictionaryTransition): 3133 (JSC::Structure::addPropertyWithoutTransition): 3134 (JSC::Structure::despecifyAllFunctions): 3135 * runtime/Structure.h: 3136 (JSC::Structure::disableSpecificFunctionTracking): Track a thrash count 3137 for specific functions. Disable specific function tracking once the 3138 thrash count has been hit. 3139 31402010-01-07 Csaba Osztrogonác <ossy@webkit.org> 3141 3142 Reviewed by Simon Hausmann. 3143 3144 [Qt] Enable JIT in debug mode on win32 after r51141 fixed the crashes. 3145 3146 * JavaScriptCore.pri: 3147 31482010-01-07 Zoltan Horvath <zoltan@webkit.org> 3149 3150 Reviewed by Holger Freyther. 3151 3152 [Mac] Build fix when FAST_MALLOC_MATCH_VALIDATION=1 3153 https://bugs.webkit.org/show_bug.cgi?id=33312 3154 3155 Using of operator += cause compile error on Mac, so it is changed to 3156 "= static_cast<AllocAlignmentInteger*>(old_ptr) + 1". 3157 3158 * wtf/FastMalloc.cpp: 3159 (WTF::TCMallocStats::realloc): 3160 31612010-01-07 Zoltan Horvath <zoltan@webkit.org> 3162 3163 Reviewed by Holger Freyther. 3164 3165 [Qt] Build fix when FAST_MALLOC_MATCH_VALIDATION=1 3166 https://bugs.webkit.org/show_bug.cgi?id=33312 3167 3168 Remove pByte (committed in r42344 from #20422), because pByte doesn't 3169 exist and it is unnecessary. 3170 3171 * wtf/FastMalloc.cpp: 3172 (WTF::TCMallocStats::realloc): 3173 31742010-01-06 Gavin Barraclough <barraclough@apple.com> 3175 3176 QT build fix. 3177 3178 * runtime/Identifier.cpp: 3179 (JSC::createIdentifierTableSpecific): 3180 31812010-01-06 Gavin Barraclough <barraclough@apple.com> 3182 3183 Windows build fix part I. 3184 3185 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 3186 31872010-01-06 Dan Bernstein <mitz@apple.com> 3188 3189 Build fix 3190 3191 * runtime/Identifier.cpp: 3192 (JSC::createIdentifierTableSpecificCallback): 3193 31942010-01-05 Gavin Barraclough <barraclough@apple.com> 3195 3196 Reviewed by Sam Weinig. 3197 3198 https://bugs.webkit.org/show_bug.cgi?id=33236 3199 Remove m_identifierTable pointer from UString 3200 3201 Currently every string holds a pointer so that during destruction, 3202 if a string has been used as an identifier, it can remove itself 3203 from the table. By instead accessing the identifierTable via a 3204 thread specific tracking the table associated with the current 3205 globaldata, we can save the memory cost of this pointer. 3206 3207 * API/APIShims.h: 3208 (JSC::APIEntryShimWithoutLock::APIEntryShimWithoutLock): 3209 (JSC::APIEntryShimWithoutLock::~APIEntryShimWithoutLock): 3210 (JSC::APICallbackShim::APICallbackShim): 3211 (JSC::APICallbackShim::~APICallbackShim): 3212 3213 - change the API shims to track the identifierTable of the current JSGlobalData. 3214 3215 * API/JSContextRef.cpp: 3216 (JSContextGroupCreate): 3217 3218 - update creation of JSGlobalData for API usage to use new create method. 3219 - fix shim instanciation bug in JSGlobalContextCreateInGroup. 3220 3221 * JavaScriptCore.exp: 3222 * runtime/Completion.cpp: 3223 (JSC::checkSyntax): 3224 (JSC::evaluate): 3225 3226 - add asserts to check the identifierTable is being tracked correctly. 3227 3228 * runtime/Identifier.cpp: 3229 (JSC::IdentifierTable::~IdentifierTable): 3230 (JSC::IdentifierTable::add): 3231 (JSC::Identifier::remove): 3232 (JSC::Identifier::checkSameIdentifierTable): 3233 (JSC::createIdentifierTableSpecificCallback): 3234 (JSC::createIdentifierTableSpecific): 3235 (JSC::createDefaultDataSpecific): 3236 3237 - Use currentIdentifierTable() instead of UStringImpl::m_identifierTable. 3238 - Define methods to access the thread specific identifier tables. 3239 3240 * runtime/Identifier.h: 3241 (JSC::ThreadIdentifierTableData::ThreadIdentifierTableData): 3242 (JSC::defaultIdentifierTable): 3243 (JSC::setDefaultIdentifierTable): 3244 (JSC::currentIdentifierTable): 3245 (JSC::setCurrentIdentifierTable): 3246 (JSC::resetCurrentIdentifierTable): 3247 3248 - Declare methods to access the thread specific identifier tables. 3249 3250 * runtime/JSGlobalData.cpp: 3251 (JSC::JSGlobalData::createNonDefault): 3252 (JSC::JSGlobalData::create): 3253 (JSC::JSGlobalData::sharedInstance): 3254 3255 - creation of JSGlobalData objects, other than for API usage, associate themselves with the current thread. 3256 3257 * runtime/JSGlobalData.h: 3258 * runtime/UStringImpl.cpp: 3259 (JSC::UStringImpl::destroy): 3260 3261 - destroy() method should be using isIdentifier(). 3262 3263 * runtime/UStringImpl.h: 3264 (JSC::UStringImpl::isIdentifier): 3265 (JSC::UStringImpl::setIsIdentifier): 3266 (JSC::UStringImpl::checkConsistency): 3267 (JSC::UStringImpl::UStringImpl): 3268 3269 - replace m_identifierTable with a single m_isIdentifier bit. 3270 3271 * wtf/StringHashFunctions.h: 3272 (WTF::stringHash): 3273 3274 - change string hash result from 32-bit to 31-bit, to free a bit in UStringImpl for m_isIdentifier. 3275 32762009-12-25 Patrick Gansterer <paroga@paroga.com> 3277 3278 Reviewed by Eric Seidel. 3279 3280 Buildfix for WinCE + style fixes. 3281 https://bugs.webkit.org/show_bug.cgi?id=32939 3282 3283 * jsc.cpp: 3284 (functionPrint): 3285 (functionQuit): 3286 (parseArguments): 3287 (fillBufferWithContentsOfFile): 3288 32892010-01-05 Patrick Gansterer <paroga@paroga.com> 3290 3291 Reviewed by Eric Seidel. 3292 3293 WinCE buildfix after r52791 (renamed PLATFORM(WINCE) to OS(WINCE)). 3294 https://bugs.webkit.org/show_bug.cgi?id=33205 3295 3296 * jit/ExecutableAllocator.h: 3297 32982010-01-05 Patrick Gansterer <paroga@paroga.com> 3299 3300 Reviewed by Darin Adler. 3301 3302 Added compiler error for unsupported platforms. 3303 https://bugs.webkit.org/show_bug.cgi?id=33112 3304 3305 * jit/JITStubs.cpp: 3306 33072010-01-05 Gabor Loki <loki@webkit.org> 3308 3309 Reviewed by Maciej Stachowiak. 3310 3311 Follow r52729 in ARMAssembler. 3312 https://bugs.webkit.org/show_bug.cgi?id=33208 3313 3314 Use WTF_ARM_ARCH_AT_LEAST instead of ARM_ARCH_VERSION 3315 3316 * assembler/ARMAssembler.cpp: 3317 (JSC::ARMAssembler::encodeComplexImm): Move tmp declaration to ARMv7 3318 * assembler/ARMAssembler.h: 3319 (JSC::ARMAssembler::): 3320 (JSC::ARMAssembler::bkpt): 3321 33222010-01-05 Maciej Stachowiak <mjs@apple.com> 3323 3324 Unreviewed build fix for Gtk+ 3325 3326 Don't use // comments in Platform.h, at least some of them seem to make the version of GCC 3327 used on the Gtk buildbot unhappy. 3328 3329 * wtf/Platform.h: 3330 33312010-01-04 Maciej Stachowiak <mjs@apple.com> 3332 3333 Reviewed by Darin Fisher. 3334 3335 Reorganize, document and rename OS() platform macros. 3336 https://bugs.webkit.org/show_bug.cgi?id=33198 3337 3338 * wtf/Platform.h: Rename, reorganize and document OS() macros. 3339 3340 Adapt to name changes. Also fixed a few incorrect OS checks. 3341 3342 * API/JSContextRef.cpp: 3343 * assembler/MacroAssemblerARM.cpp: 3344 (JSC::isVFPPresent): 3345 * assembler/MacroAssemblerX86Common.h: 3346 * bytecode/SamplingTool.cpp: 3347 * config.h: 3348 * interpreter/RegisterFile.cpp: 3349 (JSC::RegisterFile::~RegisterFile): 3350 * interpreter/RegisterFile.h: 3351 (JSC::RegisterFile::RegisterFile): 3352 (JSC::RegisterFile::grow): 3353 * jit/ExecutableAllocator.h: 3354 * jit/ExecutableAllocatorFixedVMPool.cpp: 3355 * jit/ExecutableAllocatorPosix.cpp: 3356 * jit/ExecutableAllocatorSymbian.cpp: 3357 * jit/ExecutableAllocatorWin.cpp: 3358 * jit/JITOpcodes.cpp: 3359 (JSC::JIT::privateCompileCTIMachineTrampolines): 3360 * jit/JITStubs.cpp: 3361 * jsc.cpp: 3362 (main): 3363 * parser/Grammar.y: 3364 * profiler/ProfileNode.cpp: 3365 (JSC::getCount): 3366 * runtime/Collector.cpp: 3367 (JSC::Heap::Heap): 3368 (JSC::Heap::allocateBlock): 3369 (JSC::Heap::freeBlockPtr): 3370 (JSC::currentThreadStackBase): 3371 (JSC::getCurrentPlatformThread): 3372 (JSC::suspendThread): 3373 (JSC::resumeThread): 3374 (JSC::getPlatformThreadRegisters): 3375 (JSC::otherThreadStackPointer): 3376 * runtime/Collector.h: 3377 * runtime/DateConstructor.cpp: 3378 * runtime/DatePrototype.cpp: 3379 (JSC::formatLocaleDate): 3380 * runtime/InitializeThreading.cpp: 3381 (JSC::initializeThreading): 3382 * runtime/MarkStack.h: 3383 (JSC::MarkStack::MarkStackArray::shrinkAllocation): 3384 * runtime/MarkStackPosix.cpp: 3385 * runtime/MarkStackSymbian.cpp: 3386 * runtime/MarkStackWin.cpp: 3387 * runtime/StringPrototype.cpp: 3388 (JSC::stringProtoFuncLastIndexOf): 3389 * runtime/TimeoutChecker.cpp: 3390 (JSC::getCPUTime): 3391 * runtime/UString.cpp: 3392 (JSC::UString::from): 3393 * wtf/Assertions.cpp: 3394 * wtf/Assertions.h: 3395 * wtf/CurrentTime.cpp: 3396 (WTF::lowResUTCTime): 3397 * wtf/CurrentTime.h: 3398 (WTF::getLocalTime): 3399 * wtf/DateMath.cpp: 3400 * wtf/FastMalloc.cpp: 3401 (WTF::TCMalloc_ThreadCache::InitModule): 3402 (WTF::TCMallocStats::): 3403 * wtf/FastMalloc.h: 3404 * wtf/MathExtras.h: 3405 * wtf/RandomNumber.cpp: 3406 (WTF::randomNumber): 3407 * wtf/RandomNumberSeed.h: 3408 (WTF::initializeRandomNumberGenerator): 3409 * wtf/StringExtras.h: 3410 * wtf/TCSpinLock.h: 3411 (TCMalloc_SpinLock::Unlock): 3412 (TCMalloc_SlowLock): 3413 * wtf/TCSystemAlloc.cpp: 3414 * wtf/ThreadSpecific.h: 3415 (WTF::::destroy): 3416 * wtf/Threading.h: 3417 * wtf/ThreadingPthreads.cpp: 3418 (WTF::initializeThreading): 3419 (WTF::isMainThread): 3420 * wtf/ThreadingWin.cpp: 3421 (WTF::wtfThreadEntryPoint): 3422 (WTF::createThreadInternal): 3423 * wtf/VMTags.h: 3424 * wtf/unicode/icu/CollatorICU.cpp: 3425 (WTF::Collator::userDefault): 3426 * wtf/win/MainThreadWin.cpp: 3427 (WTF::initializeMainThreadPlatform): 3428 34292010-01-04 Gustavo Noronha Silva <gns@gnome.org> 3430 3431 Add missing files to the build system - make distcheck build fix. 3432 3433 * GNUmakefile.am: 3434 34352010-01-04 Gavin Barraclough <barraclough@apple.com> 3436 3437 Reviewed by Sam Weinig, additional coding by Mark Rowe. 3438 3439 https://bugs.webkit.org/show_bug.cgi?id=33163 3440 Add string hashing functions to WTF. 3441 Use WTF's string hashing functions from UStringImpl. 3442 3443 * GNUmakefile.am: 3444 * JavaScriptCore.exp: 3445 * JavaScriptCore.gypi: 3446 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 3447 * JavaScriptCore.xcodeproj/project.pbxproj: 3448 * runtime/UStringImpl.cpp: 3449 * runtime/UStringImpl.h: 3450 (JSC::UStringImpl::computeHash): 3451 * wtf/HashFunctions.h: 3452 * wtf/StringHashFunctions.h: Added. 3453 (WTF::stringHash): 3454 34552010-01-04 Dmitry Titov <dimich@chromium.org> 3456 3457 Not reviewed, attempt to fix ARM bulid. 3458 3459 * wtf/Platform.h: 3460 34612010-01-04 Gavin Barraclough <barraclough@apple.com> 3462 3463 Rubber stamped by Geoff Garen. 3464 3465 Add an 'isIdentifier' to UStringImpl, use this where appropriate 3466 (where previously 'identifierTable' was being tested). 3467 3468 * API/JSClassRef.cpp: 3469 (OpaqueJSClass::~OpaqueJSClass): 3470 (OpaqueJSClassContextData::OpaqueJSClassContextData): 3471 * runtime/Identifier.cpp: 3472 (JSC::Identifier::addSlowCase): 3473 * runtime/Identifier.h: 3474 (JSC::Identifier::add): 3475 * runtime/PropertyNameArray.cpp: 3476 (JSC::PropertyNameArray::add): 3477 * runtime/UStringImpl.h: 3478 (JSC::UStringImpl::isIdentifier): 3479 34802010-01-04 Gavin Barraclough <barraclough@apple.com> 3481 3482 Reviewed by Sam "Shimmey Shimmey" Weinig. 3483 3484 https://bugs.webkit.org/show_bug.cgi?id=33158 3485 Refactor JSC API entry/exit to use RAII instead of copy/pasting code. 3486 Make it easier to change set of actions taken when passing across the API boundary. 3487 3488 * API/APIShims.h: Added. 3489 (JSC::APIEntryShimWithoutLock::APIEntryShimWithoutLock): 3490 (JSC::APIEntryShimWithoutLock::~APIEntryShimWithoutLock): 3491 (JSC::APIEntryShim::APIEntryShim): 3492 (JSC::APICallbackShim::APICallbackShim): 3493 (JSC::APICallbackShim::~APICallbackShim): 3494 * API/JSBase.cpp: 3495 (JSEvaluateScript): 3496 (JSCheckScriptSyntax): 3497 (JSGarbageCollect): 3498 (JSReportExtraMemoryCost): 3499 * API/JSCallbackConstructor.cpp: 3500 (JSC::constructJSCallback): 3501 * API/JSCallbackFunction.cpp: 3502 (JSC::JSCallbackFunction::call): 3503 * API/JSCallbackObjectFunctions.h: 3504 (JSC::::init): 3505 (JSC::::getOwnPropertySlot): 3506 (JSC::::put): 3507 (JSC::::deleteProperty): 3508 (JSC::::construct): 3509 (JSC::::hasInstance): 3510 (JSC::::call): 3511 (JSC::::getOwnPropertyNames): 3512 (JSC::::toNumber): 3513 (JSC::::toString): 3514 (JSC::::staticValueGetter): 3515 (JSC::::callbackGetter): 3516 * API/JSContextRef.cpp: 3517 * API/JSObjectRef.cpp: 3518 (JSObjectMake): 3519 (JSObjectMakeFunctionWithCallback): 3520 (JSObjectMakeConstructor): 3521 (JSObjectMakeFunction): 3522 (JSObjectMakeArray): 3523 (JSObjectMakeDate): 3524 (JSObjectMakeError): 3525 (JSObjectMakeRegExp): 3526 (JSObjectGetPrototype): 3527 (JSObjectSetPrototype): 3528 (JSObjectHasProperty): 3529 (JSObjectGetProperty): 3530 (JSObjectSetProperty): 3531 (JSObjectGetPropertyAtIndex): 3532 (JSObjectSetPropertyAtIndex): 3533 (JSObjectDeleteProperty): 3534 (JSObjectCallAsFunction): 3535 (JSObjectCallAsConstructor): 3536 (JSObjectCopyPropertyNames): 3537 (JSPropertyNameArrayRelease): 3538 (JSPropertyNameAccumulatorAddName): 3539 * API/JSValueRef.cpp: 3540 (JSValueGetType): 3541 (JSValueIsUndefined): 3542 (JSValueIsNull): 3543 (JSValueIsBoolean): 3544 (JSValueIsNumber): 3545 (JSValueIsString): 3546 (JSValueIsObject): 3547 (JSValueIsObjectOfClass): 3548 (JSValueIsEqual): 3549 (JSValueIsStrictEqual): 3550 (JSValueIsInstanceOfConstructor): 3551 (JSValueMakeUndefined): 3552 (JSValueMakeNull): 3553 (JSValueMakeBoolean): 3554 (JSValueMakeNumber): 3555 (JSValueMakeString): 3556 (JSValueToBoolean): 3557 (JSValueToNumber): 3558 (JSValueToStringCopy): 3559 (JSValueToObject): 3560 (JSValueProtect): 3561 (JSValueUnprotect): 3562 * JavaScriptCore.xcodeproj/project.pbxproj: 3563 35642010-01-04 Dan Bernstein <mitz@apple.com> 3565 3566 Reviewed by Ada Chan and Mark Rowe. 3567 3568 Updated copyright string 3569 3570 * Info.plist: 3571 * JavaScriptCore.vcproj/JavaScriptCore.resources/Info.plist: 3572 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.rc: 3573 35742010-01-04 Adam Roben <aroben@apple.com> 3575 3576 No review, rolling out r52741. 3577 http://trac.webkit.org/changeset/52741 3578 https://bugs.webkit.org/show_bug.cgi?id=33056 3579 3580 * wtf/AlwaysInline.h: 3581 35822010-01-04 Patrick Gansterer <paroga@paroga.com> 3583 3584 Reviewed by Darin Adler. 3585 3586 Add cacheFlush support for WinCE 3587 https://bugs.webkit.org/show_bug.cgi?id=33110 3588 3589 * jit/ExecutableAllocator.h: 3590 (JSC::ExecutableAllocator::cacheFlush): 3591 35922010-01-04 Patrick Gansterer <paroga@paroga.com> 3593 3594 Reviewed by Adam Roben. 3595 3596 Implement NO_RETURN for COMPILER(MSVC). 3597 https://bugs.webkit.org/show_bug.cgi?id=33056 3598 3599 * wtf/AlwaysInline.h: 3600 36012010-01-04 Maciej Stachowiak <mjs@apple.com> 3602 3603 Reviewed by Simon Hausmann. 3604 3605 Fix some PLATFORM(*_ENDIAN) uses to CPU() 3606 https://bugs.webkit.org/show_bug.cgi?id=33148 3607 3608 * runtime/JSCell.cpp: 3609 (JSC::): 3610 * runtime/JSValue.h: 3611 (JSC::JSValue::): 3612 36132010-01-04 Maciej Stachowiak <mjs@apple.com> 3614 3615 Reviewed by Adam Barth. 3616 3617 Document CPU() macros in comments. 3618 https://bugs.webkit.org/show_bug.cgi?id=33147 3619 3620 * wtf/Platform.h: 3621 36222010-01-04 Maciej Stachowiak <mjs@apple.com> 3623 3624 Reviewed by Adam Barth. 3625 3626 Reorganize, document and rename CPU() platform macros. 3627 https://bugs.webkit.org/show_bug.cgi?id=33145 3628 ExecutableAllocatorSymbian appears to have buggy ARM version check 3629 https://bugs.webkit.org/show_bug.cgi?id=33138 3630 3631 * wtf/Platform.h: 3632 Rename all macros related to detection of particular CPUs or 3633 classes of CPUs to CPU(), reorganize and document them. 3634 3635 All remaining changes are adapting to the renames, plus fixing the 3636 second bug cited above. 3637 3638 * assembler/ARMAssembler.cpp: 3639 * assembler/ARMAssembler.h: 3640 * assembler/ARMv7Assembler.h: 3641 * assembler/AbstractMacroAssembler.h: 3642 (JSC::AbstractMacroAssembler::Imm32::Imm32): 3643 * assembler/MacroAssembler.h: 3644 * assembler/MacroAssemblerARM.cpp: 3645 * assembler/MacroAssemblerARM.h: 3646 * assembler/MacroAssemblerCodeRef.h: 3647 (JSC::MacroAssemblerCodePtr::MacroAssemblerCodePtr): 3648 * assembler/MacroAssemblerX86.h: 3649 * assembler/MacroAssemblerX86Common.h: 3650 * assembler/MacroAssemblerX86_64.h: 3651 * assembler/X86Assembler.h: 3652 (JSC::X86Registers::): 3653 (JSC::X86Assembler::): 3654 (JSC::X86Assembler::movl_mEAX): 3655 (JSC::X86Assembler::movl_EAXm): 3656 (JSC::X86Assembler::repatchLoadPtrToLEA): 3657 (JSC::X86Assembler::X86InstructionFormatter::memoryModRM): 3658 * jit/ExecutableAllocator.h: 3659 * jit/ExecutableAllocatorFixedVMPool.cpp: 3660 * jit/ExecutableAllocatorPosix.cpp: 3661 * jit/ExecutableAllocatorSymbian.cpp: 3662 (JSC::ExecutableAllocator::intializePageSize): 3663 * jit/JIT.cpp: 3664 * jit/JIT.h: 3665 * jit/JITArithmetic.cpp: 3666 * jit/JITInlineMethods.h: 3667 (JSC::JIT::beginUninterruptedSequence): 3668 (JSC::JIT::restoreArgumentReferenceForTrampoline): 3669 (JSC::JIT::emitCount): 3670 * jit/JITOpcodes.cpp: 3671 (JSC::JIT::privateCompileCTIMachineTrampolines): 3672 * jit/JITPropertyAccess.cpp: 3673 (JSC::JIT::privateCompileGetByIdProto): 3674 (JSC::JIT::privateCompileGetByIdProtoList): 3675 (JSC::JIT::privateCompileGetByIdChainList): 3676 (JSC::JIT::privateCompileGetByIdChain): 3677 * jit/JITStubs.cpp: 3678 (JSC::JITThunks::JITThunks): 3679 * jit/JITStubs.h: 3680 * runtime/Collector.cpp: 3681 (JSC::currentThreadStackBase): 3682 (JSC::getPlatformThreadRegisters): 3683 (JSC::otherThreadStackPointer): 3684 * wrec/WREC.h: 3685 * wrec/WRECGenerator.cpp: 3686 (JSC::WREC::Generator::generateEnter): 3687 (JSC::WREC::Generator::generateReturnSuccess): 3688 (JSC::WREC::Generator::generateReturnFailure): 3689 * wrec/WRECGenerator.h: 3690 * wtf/FastMalloc.cpp: 3691 * wtf/TCSpinLock.h: 3692 (TCMalloc_SpinLock::Lock): 3693 (TCMalloc_SpinLock::Unlock): 3694 (TCMalloc_SlowLock): 3695 * wtf/Threading.h: 3696 * wtf/dtoa.cpp: 3697 * yarr/RegexJIT.cpp: 3698 (JSC::Yarr::RegexGenerator::generateEnter): 3699 (JSC::Yarr::RegexGenerator::generateReturn): 3700 * yarr/RegexJIT.h: 3701 37022010-01-04 Maciej Stachowiak <mjs@apple.com> 3703 3704 Reviewed by Adam Barth. 3705 3706 Clean up COMPILER macros and remove unused ones. 3707 https://bugs.webkit.org/show_bug.cgi?id=33132 3708 3709 Removed values are COMPILER(BORLAND) and COMPILER(CYGWIN) - they were 3710 not used anywhere. 3711 3712 * wtf/Platform.h: 3713 37142010-01-03 Maciej Stachowiak <mjs@apple.com> 3715 3716 Reviewed by Eric Seidel. 3717 3718 Update wtf/Platform.h to document the new system for porting macros. 3719 https://bugs.webkit.org/show_bug.cgi?id=33130 3720 3721 * wtf/Platform.h: 3722 37232009-12-29 Laszlo Gombos <laszlo.1.gombos@nokia.com> 3724 3725 Reviewed by Maciej Stachowiak. 3726 3727 PLATFORM(CAIRO) should be defined by WIN_CAIRO define 3728 https://bugs.webkit.org/show_bug.cgi?id=22250 3729 3730 * wtf/Platform.h: Define WTF_PLATFORM_CAIRO for GTK port only 3731 For the WinCairo port WTF_PLATFORM_CAIRO is already defined in config.h 3732 37332009-12-28 Shu Chang <Chang.Shu@nokia.com> 3734 3735 Reviewed by Laszlo Gombos. 3736 3737 [Qt] Delete ThreadPrivate instance after it is finished. 3738 https://bugs.webkit.org/show_bug.cgi?id=32614 3739 3740 * wtf/qt/ThreadingQt.cpp: 3741 (WTF::ThreadMonitor::instance): 3742 (WTF::ThreadMonitor::threadFinished): 3743 (WTF::createThreadInternal): 3744 (WTF::detachThread): 3745 37462009-12-28 Patrick Gansterer <paroga@paroga.com> 3747 3748 Reviewed by Maciej Stachowiak. 3749 3750 Cleanup of #define JS_EXPORT. 3751 3752 * API/JSBase.h: 3753 37542009-12-27 Patrick Gansterer <paroga@paroga.com> 3755 3756 Reviewed by Adam Barth. 3757 3758 WinCE buildfix (HWND_MESSAGE isn't supported there) 3759 3760 * wtf/win/MainThreadWin.cpp: 3761 (WTF::initializeMainThreadPlatform): 3762 37632009-12-27 Patrick Gansterer <paroga@paroga.com> 3764 3765 Reviewed by Adam Barth. 3766 3767 Added a file with WinMain function to link agains in WinCE. 3768 3769 * os-win32/WinMain.cpp: Added. 3770 (convertToUtf8): 3771 (WinMain): 3772 37732009-12-24 Laszlo Gombos <laszlo.1.gombos@nokia.com> 3774 3775 Unreviewed; revert of r52550. 3776 3777 The change regressed the following LayoutTests for QtWebKit. 3778 3779 fast/workers/worker-call.html -> crashed 3780 fast/workers/worker-close.html -> crashed 3781 3782 * wtf/qt/ThreadingQt.cpp: 3783 (WTF::waitForThreadCompletion): 3784 (WTF::detachThread): 3785 37862009-12-24 Shu Chang <Chang.Shu@nokia.com> 3787 3788 Reviewed by Laszlo Gombos. 3789 3790 [Qt] Fix memory leak by deleting instance of ThreadPrivate 3791 in function waitForThreadCompletion(), synchronously, or in 3792 detachThread(), asynchronously. 3793 https://bugs.webkit.org/show_bug.cgi?id=32614 3794 3795 * wtf/qt/ThreadingQt.cpp: 3796 (WTF::waitForThreadCompletion): 3797 (WTF::detachThread): 3798 37992009-12-23 Kwang Yul Seo <skyul@company100.net> 3800 3801 Reviewed by Laszlo Gombos. 3802 3803 Include stddef.h for ptrdiff_t 3804 https://bugs.webkit.org/show_bug.cgi?id=32891 3805 3806 ptrdiff_t is typedef-ed in stddef.h. 3807 Include stddef.h in jit/ExecutableAllocator.h. 3808 3809 * jit/ExecutableAllocator.h: 3810 38112009-12-23 Patrick Gansterer <paroga@paroga.com> 3812 3813 Reviewed by Eric Seidel. 3814 3815 Buildfix after r47092. 3816 3817 * wtf/wince/MemoryManager.cpp: 3818 (WTF::tryFastMalloc): 3819 (WTF::tryFastZeroedMalloc): 3820 (WTF::tryFastCalloc): 3821 (WTF::tryFastRealloc): 3822 38232009-12-23 Kent Tamura <tkent@chromium.org> 3824 3825 Reviewed by Darin Adler. 3826 3827 HTMLInputElement::valueAsDate getter support. 3828 https://bugs.webkit.org/show_bug.cgi?id=32876 3829 3830 Expose dateToDaysFrom1970(). 3831 3832 * JavaScriptCore.exp: 3833 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 3834 * wtf/DateMath.cpp: 3835 (WTF::dateToDaysFrom1970): 3836 * wtf/DateMath.h: 3837 38382009-12-22 Darin Adler <darin@apple.com> 3839 3840 Reviewed by Mark Rowe. 3841 3842 Turn off datagrid by default, at least for all platforms Apple ships. 3843 The datagrid implementation isn't ready for general web use yet. 3844 3845 * Configurations/FeatureDefines.xcconfig: Turn off datagrid by default. 3846 38472009-12-22 Steve Block <steveblock@google.com> 3848 3849 Reviewed by David Levin. 3850 3851 Updates Android's scheduleDispatchFunctionsOnMainThread() to use new 3852 AndroidThreading class, rather than using JavaSharedClient directly. 3853 This fixes the current layering violation. 3854 https://bugs.webkit.org/show_bug.cgi?id=32651 3855 3856 The pattern is copied from Chromium, which uses the ChromiumThreading 3857 class. This patch also fixes the style in ChromiumThreading.h. 3858 3859 * wtf/android/AndroidThreading.h: Added. Declares AndroidThreading. 3860 * wtf/android/MainThreadAndroid.cpp: Modified 3861 (WTF::scheduleDispatchFunctionsOnMainThread): Uses AndroidThreading. 3862 * wtf/chromium/ChromiumThreading.h: Modified. Fixes style. 3863 38642009-12-22 Gavin Barraclough <barraclough@apple.com> 3865 3866 Reviewed by Sam Weinig. 3867 3868 Fix a couple of problems with UntypedPtrAndBitfield. 3869 3870 Add a m_leaksPtr to reduce false positives from leaks in debug builds 3871 (this isn't perfect because we'd like a solution for release builds, 3872 but this is now at least as good as a PtrAndFlags would be). 3873 3874 Switch SmallStringsto use a regular string for the base, rather than 3875 a static one. UntypedPtrAndBitfield assumes all strings are at least 3876 8 byte aligned; this migt not be true of static strings. Shared buffers 3877 are heap allocated, as are all UStringImpls other than static strings. 3878 Static strings cannot end up being the owner string of substrings, 3879 since the only static strings are length 0. 3880 3881 * runtime/SmallStrings.cpp: 3882 (JSC::SmallStringsStorage::SmallStringsStorage): 3883 * runtime/UStringImpl.h: 3884 (JSC::UntypedPtrAndBitfield::UntypedPtrAndBitfield): 3885 (JSC::UStringImpl::UStringImpl): 3886 38872009-12-22 Kwang Yul Seo <skyul@company100.net> 3888 3889 Reviewed by Darin Adler. 3890 3891 RVCT (__ARMCC_VERSION < 400000) does not provide strcasecmp and strncasecmp 3892 https://bugs.webkit.org/show_bug.cgi?id=32857 3893 3894 Add implementation of strcasecmp and strncasecmp for RVCT < 4.0 3895 because earlier versions of RVCT 4.0 does not provide these functions. 3896 3897 * wtf/StringExtras.cpp: Added. 3898 (strcasecmp): 3899 (strncasecmp): 3900 * wtf/StringExtras.h: 3901 39022009-12-22 Kwang Yul Seo <skyul@company100.net> 3903 3904 Reviewed by Darin Adler. 3905 3906 Define ALWAYS_INLINE and WTF_PRIVATE_INLINE to __forceinline for RVCT 3907 https://bugs.webkit.org/show_bug.cgi?id=32853 3908 3909 Use __forceinline forces RVCT to compile a C or C++ function 3910 inline. The compiler attempts to inline the function, regardless of 3911 the characteristics of the function. 3912 3913 * wtf/AlwaysInline.h: 3914 * wtf/FastMalloc.h: 3915 39162009-12-21 Simon Hausmann <simon.hausmann@nokia.com> 3917 3918 Prospective GTK build fix: Add UStringImpl.cpp/h to the build. 3919 3920 * GNUmakefile.am: 3921 39222009-12-21 Simon Hausmann <simon.hausmann@nokia.com> 3923 3924 Fix the Qt build, add UStringImpl.cpp to the build. 3925 3926 * JavaScriptCore.pri: 3927 39282009-12-21 Gavin Barraclough <barraclough@apple.com> 3929 3930 Windows Build fix part 5. 3931 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: 3932 39332009-12-21 Gavin Barraclough <barraclough@apple.com> 3934 3935 Reviewed by NOBODY (build fix). 3936 Fix breakage of world introduced in build fix to r52463. 3937 3938 * runtime/UStringImpl.h: 3939 39402009-12-21 Gavin Barraclough <barraclough@apple.com> 3941 3942 Reviewed by Darin Adler. 3943 3944 https://bugs.webkit.org/show_bug.cgi?id=32831 3945 Replace UString::Rep implementation, following introduction of ropes to JSC. 3946 3947 * Remove redundant overcapacity mechanisms. 3948 * Reduce memory cost of Rep's. 3949 * Add an inline storage mechanism akin to that in WebCore's StringImpl. 3950 3951 ~1% Sunspider progression. 3952 3953 * JavaScriptCore.exp: 3954 * JavaScriptCore.xcodeproj/project.pbxproj: 3955 * runtime/JSString.cpp: 3956 (JSC::JSString::resolveRope): 3957 * runtime/SmallStrings.cpp: 3958 (JSC::SmallStringsStorage::SmallStringsStorage): 3959 * runtime/UString.cpp: 3960 (JSC::initializeUString): 3961 (JSC::createRep): 3962 (JSC::UString::createFromUTF8): 3963 (JSC::UString::createUninitialized): 3964 (JSC::UString::spliceSubstringsWithSeparators): 3965 (JSC::UString::replaceRange): 3966 (JSC::UString::ascii): 3967 (JSC::UString::operator=): 3968 (JSC::UString::toStrictUInt32): 3969 (JSC::equal): 3970 * runtime/UString.h: 3971 (JSC::UString::isEmpty): 3972 (JSC::UString::cost): 3973 (JSC::makeString): 3974 * runtime/UStringImpl.cpp: Added. 3975 (JSC::UStringImpl::baseSharedBuffer): 3976 (JSC::UStringImpl::sharedBuffer): 3977 (JSC::UStringImpl::destroy): 3978 (JSC::UStringImpl::computeHash): 3979 * runtime/UStringImpl.h: Added. 3980 (JSC::UntypedPtrAndBitfield::UntypedPtrAndBitfield): 3981 (JSC::UntypedPtrAndBitfield::asPtr): 3982 (JSC::UntypedPtrAndBitfield::operator&=): 3983 (JSC::UntypedPtrAndBitfield::operator|=): 3984 (JSC::UntypedPtrAndBitfield::operator&): 3985 (JSC::UStringImpl::create): 3986 (JSC::UStringImpl::createCopying): 3987 (JSC::UStringImpl::createUninitialized): 3988 (JSC::UStringImpl::data): 3989 (JSC::UStringImpl::size): 3990 (JSC::UStringImpl::cost): 3991 (JSC::UStringImpl::hash): 3992 (JSC::UStringImpl::computedHash): 3993 (JSC::UStringImpl::setHash): 3994 (JSC::UStringImpl::identifierTable): 3995 (JSC::UStringImpl::setIdentifierTable): 3996 (JSC::UStringImpl::ref): 3997 (JSC::UStringImpl::deref): 3998 (JSC::UStringImpl::allocChars): 3999 (JSC::UStringImpl::copyChars): 4000 (JSC::UStringImpl::computeHash): 4001 (JSC::UStringImpl::null): 4002 (JSC::UStringImpl::empty): 4003 (JSC::UStringImpl::checkConsistency): 4004 (JSC::UStringImpl::): 4005 (JSC::UStringImpl::UStringImpl): 4006 (JSC::UStringImpl::operator new): 4007 (JSC::UStringImpl::bufferOwnerString): 4008 (JSC::UStringImpl::bufferOwnership): 4009 (JSC::UStringImpl::isStatic): 4010 40112009-12-18 Laszlo Gombos <laszlo.1.gombos@nokia.com> 4012 4013 Reviewed by Kenneth Rohde Christiansen. 4014 4015 Move some build decisions from Qt build system into source files 4016 https://bugs.webkit.org/show_bug.cgi?id=31956 4017 4018 * JavaScriptCore.pri: Compile files unconditionally 4019 * jit/ExecutableAllocatorPosix.cpp: Guard with PLATFORM(UNIX) && !PLATFORM(SYMBIAN) 4020 * jit/ExecutableAllocatorWin.cpp: Guard with PLATFORM(WIN_OS) 4021 * runtime/MarkStackPosix.cpp: Guard with PLATFORM(UNIX) && !PLATFORM(SYMBIAN) 4022 * runtime/MarkStackSymbian.cpp: Guard with PLATFORM(SYMBIAN) 4023 * runtime/MarkStackWin.cpp: Guard with PLATFORM(WIN_OS) 4024 * wtf/Platform.h: Guard ENABLE_JSC_MULTIPLE_THREADS with ENABLE_SINGLE_THREADED for the Qt port 4025 * wtf/ThreadingNone.cpp: Guard with ENABLE(SINGLE_THREADED) 4026 * wtf/qt/ThreadingQt.cpp: Guard with !ENABLE(SINGLE_THREADED) 4027 40282009-12-18 Gavin Barraclough <barraclough@apple.com> 4029 4030 Reviewed by Sam Weinig. 4031 4032 Add createNonCopying method to UString to make replace constructor passed bool, 4033 to make behaviour more explicit. Add createFromUTF8 to UString (wrapping method 4034 on UString::Rep), since other cases of transliteration (e.g. from ascii) are 4035 performed in UString constructors. Add/use setHash & size() accessors on Rep, 4036 rather than accessing _hash/len directly. 4037 4038 * API/JSClassRef.cpp: 4039 (OpaqueJSClass::OpaqueJSClass): 4040 * API/OpaqueJSString.cpp: 4041 (OpaqueJSString::ustring): 4042 * JavaScriptCore.exp: 4043 * runtime/ArrayPrototype.cpp: 4044 (JSC::arrayProtoFuncToString): 4045 * runtime/Identifier.cpp: 4046 (JSC::Identifier::equal): 4047 (JSC::CStringTranslator::translate): 4048 (JSC::UCharBufferTranslator::translate): 4049 (JSC::Identifier::addSlowCase): 4050 * runtime/JSString.cpp: 4051 (JSC::JSString::resolveRope): 4052 * runtime/JSString.h: 4053 (JSC::JSString::Rope::Fiber::refAndGetLength): 4054 (JSC::JSString::Rope::append): 4055 * runtime/StringBuilder.h: 4056 (JSC::StringBuilder::release): 4057 * runtime/StringConstructor.cpp: 4058 (JSC::stringFromCharCodeSlowCase): 4059 * runtime/StringPrototype.cpp: 4060 (JSC::substituteBackreferencesSlow): 4061 (JSC::stringProtoFuncToLowerCase): 4062 (JSC::stringProtoFuncToUpperCase): 4063 (JSC::stringProtoFuncFontsize): 4064 (JSC::stringProtoFuncLink): 4065 * runtime/UString.cpp: 4066 (JSC::UString::UString): 4067 (JSC::UString::createNonCopying): 4068 (JSC::UString::createFromUTF8): 4069 * runtime/UString.h: 4070 (JSC::UString::Rep::setHash): 4071 (JSC::UString::~UString): 4072 (JSC::makeString): 4073 40742009-12-18 Geoffrey Garen <ggaren@apple.com> 4075 4076 Reviewed by Cameron Zwarich and Gavin Barraclough. 4077 4078 Changed Register constructors to assignment operators, to streamline 4079 moving values into registers. (In theory, there's no difference between 4080 the two, since the constructor should just inline away, but there seems 4081 to be a big difference in the addled mind of the GCC optimizer.) 4082 4083 In the interpreter, this is a 3.5% SunSpider speedup and a 1K-2K 4084 reduction in stack usage per privateExecute stack frame. 4085 4086 * interpreter/CallFrame.h: 4087 (JSC::ExecState::setCalleeArguments): 4088 (JSC::ExecState::setCallerFrame): 4089 (JSC::ExecState::setScopeChain): 4090 (JSC::ExecState::init): 4091 (JSC::ExecState::setArgumentCount): 4092 (JSC::ExecState::setCallee): 4093 (JSC::ExecState::setCodeBlock): Added a little bit of casting so these 4094 functions could use the new Register assignment operators. 4095 4096 * interpreter/Register.h: 4097 (JSC::Register::withInt): 4098 (JSC::Register::Register): 4099 (JSC::Register::operator=): Swapped in assignment operators for constructors. 4100 41012009-12-18 Yongjun Zhang <yongjun.zhang@nokia.com> 4102 4103 Reviewed by Simon Hausmann. 4104 4105 https://bugs.webkit.org/show_bug.cgi?id=32713 4106 [Qt] make wtf/Assertions.h compile in winscw compiler. 4107 4108 Add string arg before ellipsis to help winscw compiler resolve variadic 4109 macro definitions in wtf/Assertions.h. 4110 4111 * wtf/Assertions.h: 4112 41132009-12-18 Geoffrey Garen <ggaren@apple.com> 4114 4115 Reviewed by Adam Roben. 4116 4117 Fixed intermittent failure seen on Windows buildbot, and in other JSC 4118 API clients. 4119 4120 Added a WeakGCPtr class and changed OpaqueJSClass::cachedPrototype to 4121 use it, to avoid vending a stale object as a prototype. 4122 4123 * API/JSClassRef.cpp: 4124 (OpaqueJSClassContextData::OpaqueJSClassContextData): 4125 (OpaqueJSClass::prototype): 4126 * API/JSClassRef.h: Use WeakGCPtr. 4127 4128 * JavaScriptCore.xcodeproj/project.pbxproj: 4129 * runtime/WeakGCPtr.h: Added. 4130 (JSC::WeakGCPtr::WeakGCPtr): 4131 (JSC::WeakGCPtr::get): 4132 (JSC::WeakGCPtr::clear): 4133 (JSC::WeakGCPtr::operator*): 4134 (JSC::WeakGCPtr::operator->): 4135 (JSC::WeakGCPtr::operator!): 4136 (JSC::WeakGCPtr::operator bool): 4137 (JSC::WeakGCPtr::operator UnspecifiedBoolType): 4138 (JSC::WeakGCPtr::assign): 4139 (JSC::::operator): 4140 (JSC::operator==): 4141 (JSC::operator!=): 4142 (JSC::static_pointer_cast): 4143 (JSC::const_pointer_cast): 4144 (JSC::getPtr): Added WeakGCPtr to the project. 4145 41462009-12-18 Gavin Barraclough <barraclough@apple.com> 4147 4148 Reviewed by Sam Weinig. 4149 4150 https://bugs.webkit.org/show_bug.cgi?id=32720 4151 4152 * JavaScriptCore.exp: 4153 - Remove exports for UString::append 4154 * JavaScriptCore.xcodeproj/project.pbxproj: 4155 - Make StringBuilder a private header (was project). 4156 41572009-12-18 Martin Robinson <martin.james.robinson@gmail.com> 4158 4159 Reviewed by Gustavo Noronha Silva. 4160 4161 [GTK] GRefPtr does not take a reference when assigned a raw pointer 4162 https://bugs.webkit.org/show_bug.cgi?id=32709 4163 4164 Ensure that when assigning a raw pointer to a GRefPtr, the reference 4165 count is incremented. Also remove the GRefPtr conversion overload as 4166 GRefPtr types have necessarily incompatible reference counting. 4167 4168 * wtf/gtk/GRefPtr.h: 4169 (WTF::GRefPtr::operator=): 4170 41712009-12-18 Simon Hausmann <simon.hausmann@nokia.com> 4172 4173 Reviewed by Tor Arne Vestbø. 4174 4175 [Qt] Clean up the qmake build system to distinguish between trunk builds and package builds 4176 4177 https://bugs.webkit.org/show_bug.cgi?id=32716 4178 4179 * pcre/pcre.pri: Use standalone_package instead of QTDIR_build 4180 41812009-12-18 Martin Robinson <martin.james.robinson@gmail.com> 4182 4183 Reviewed by Gustavo Noronha Silva. 4184 4185 [GTK] Compile warning from line 29 of GRefPtr.cpp 4186 https://bugs.webkit.org/show_bug.cgi?id=32703 4187 4188 Fix memory leak and compiler warning in GRefPtr GHashTable template 4189 specialization. 4190 4191 * wtf/gtk/GRefPtr.cpp: 4192 (WTF::refGPtr): 4193 41942009-12-17 Sam Weinig <sam@webkit.org> 4195 4196 Reviewed by Mark Rowe. 4197 4198 Add BUILDING_ON_SNOW_LEOPARD and TARGETING_SNOW_LEOPARD #defines. 4199 4200 * wtf/Platform.h: 4201 42022009-12-17 Adam Roben <aroben@apple.com> 4203 4204 Sync JavaScriptCore.vcproj with JavaScriptCore.xcodeproj and the 4205 source tree 4206 4207 Fixes <http://webkit.org/b/32665>. 4208 4209 Reviewed by Ada Chan. 4210 4211 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: Moved 4212 around files and filters so that the structure matches 4213 JavaScriptCore.xcodeproj and the source tree. A few headers that were 4214 previously omitted have been added, as well as JSZombie.{cpp,h}. 4215 42162009-12-17 Adam Roben <aroben@apple.com> 4217 4218 Remove HeavyProfile and TreeProfile completely 4219 4220 These were mostly removed in r42808, but the empty files were left in 4221 place. 4222 4223 Fixes <http://webkit.org/b/32664>. 4224 4225 Reviewed by John Sullivan. 4226 4227 * Android.mk: 4228 * GNUmakefile.am: 4229 * JavaScriptCore.gypi: 4230 * JavaScriptCore.pri: 4231 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: 4232 * JavaScriptCoreSources.bkl: 4233 Removed HeavyProfile/TreeProfile source files. 4234 4235 * profiler/HeavyProfile.cpp: Removed. 4236 * profiler/HeavyProfile.h: Removed. 4237 * profiler/TreeProfile.cpp: Removed. 4238 * profiler/TreeProfile.h: Removed. 4239 42402009-12-17 Martin Robinson <martin.james.robinson@gmail.com> 4241 4242 Reviewed by Gustavo Noronha Silva. 4243 4244 [GTK] WebKit GTK needs a wrapper for ref counted glib/gobject structs 4245 https://bugs.webkit.org/show_bug.cgi?id=21599 4246 4247 Implement GRefPtr, a smart pointer for reference counted GObject types. 4248 4249 * GNUmakefile.am: 4250 * wtf/gtk/GOwnPtr.cpp: 4251 (WTF::GDir): 4252 * wtf/gtk/GRefPtr.h: Added. 4253 (WTF::): 4254 (WTF::GRefPtr::GRefPtr): 4255 (WTF::GRefPtr::~GRefPtr): 4256 (WTF::GRefPtr::clear): 4257 (WTF::GRefPtr::get): 4258 (WTF::GRefPtr::operator*): 4259 (WTF::GRefPtr::operator->): 4260 (WTF::GRefPtr::operator!): 4261 (WTF::GRefPtr::operator UnspecifiedBoolType): 4262 (WTF::GRefPtr::hashTableDeletedValue): 4263 (WTF::::operator): 4264 (WTF::::swap): 4265 (WTF::swap): 4266 (WTF::operator==): 4267 (WTF::operator!=): 4268 (WTF::static_pointer_cast): 4269 (WTF::const_pointer_cast): 4270 (WTF::getPtr): 4271 (WTF::adoptGRef): 4272 (WTF::refGPtr): 4273 (WTF::derefGPtr): 4274 42752009-12-17 Gustavo Noronha Silva <gustavo.noronha@collabora.co.uk> 4276 4277 Unreviewed. Build fixes for make distcheck. 4278 4279 * GNUmakefile.am: 4280 42812009-12-16 Geoffrey Garen <ggaren@apple.com> 4282 4283 Reviewed by Oliver Hunt. 4284 4285 Fixed <rdar://problem/7355025> Interpreter::privateExecute macro generates 4286 bloated code 4287 4288 This patch cuts Interpreter stack use by about a third. 4289 4290 * bytecode/Opcode.h: Changed Opcode to const void* to work with the 4291 const static initiliazation we want to do in Interpreter::privateExecute. 4292 4293 * interpreter/Interpreter.cpp: 4294 (JSC::Interpreter::Interpreter): Moved hashtable initialization here to 4295 avoid polluting Interpreter::privateExecute's stack, and changed it from a 4296 series of add() calls to one add() call in a loop, to cut down on code size. 4297 4298 (JSC::Interpreter::privateExecute): Changed a series of label computations 4299 to a copy of a compile-time constant array to cut down on code size. 4300 43012009-12-16 Mark Rowe <mrowe@apple.com> 4302 4303 Build fix. Disable debug variants of WebKit frameworks. 4304 4305 * JavaScriptCore.xcodeproj/project.pbxproj: 4306 43072009-12-15 Geoffrey Garen <ggaren@apple.com> 4308 4309 Reviewed by Sam "r=me" Weinig. 4310 4311 https://bugs.webkit.org/show_bug.cgi?id=32498 4312 <rdar://problem/7471495> 4313 REGRESSION(r51978-r52039): AJAX "Mark This Forum Read" function no longer 4314 works 4315 4316 Fixed a tyop. 4317 4318 * runtime/Operations.h: 4319 (JSC::jsAdd): Use the '&&' operator, not the ',' operator. 4320 43212009-12-15 Geoffrey Garen <ggaren@apple.com> 4322 4323 Try to fix the windows build: don't export this inlined function. 4324 4325 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 4326 43272009-12-15 Geoffrey Garen <ggaren@apple.com> 4328 4329 Reviewed by Beth Dakin. 4330 4331 Inlined JSCell's operator new. 4332 4333 3.7% speedup on bench-allocate-nonretained.js. 4334 4335 * JavaScriptCore.exp: 4336 * runtime/JSCell.cpp: 4337 * runtime/JSCell.h: 4338 (JSC::JSCell::operator new): 4339 43402009-12-15 Geoffrey Garen <ggaren@apple.com> 4341 4342 Reviewed by Oliver Hunt. 4343 4344 Removed the number heap, replacing it with a one-item free list for 4345 numbers, taking advantage of the fact that two number cells fit inside 4346 the space for one regular cell, and number cells don't require destruction. 4347 4348 SunSpider says 1.6% faster in JSVALUE32 mode (the only mode that 4349 heap-allocates numbers). 4350 4351 SunSpider says 1.1% faster in JSVALUE32_64 mode. v8 says 0.8% faster 4352 in JSVALUE32_64 mode. 10% speedup on bench-alloc-nonretained.js. 6% 4353 speedup on bench-alloc-retained.js. 4354 4355 There's a lot of formulaic change in this patch, but not much substance. 4356 4357 * JavaScriptCore.exp: 4358 * debugger/Debugger.cpp: 4359 (JSC::Debugger::recompileAllJSFunctions): 4360 * runtime/Collector.cpp: 4361 (JSC::Heap::Heap): 4362 (JSC::Heap::destroy): 4363 (JSC::Heap::allocateBlock): 4364 (JSC::Heap::freeBlock): 4365 (JSC::Heap::freeBlockPtr): 4366 (JSC::Heap::freeBlocks): 4367 (JSC::Heap::recordExtraCost): 4368 (JSC::Heap::allocate): 4369 (JSC::Heap::resizeBlocks): 4370 (JSC::Heap::growBlocks): 4371 (JSC::Heap::shrinkBlocks): 4372 (JSC::Heap::markConservatively): 4373 (JSC::Heap::clearMarkBits): 4374 (JSC::Heap::markedCells): 4375 (JSC::Heap::sweep): 4376 (JSC::Heap::markRoots): 4377 (JSC::Heap::objectCount): 4378 (JSC::Heap::addToStatistics): 4379 (JSC::Heap::statistics): 4380 (JSC::Heap::isBusy): 4381 (JSC::Heap::reset): 4382 (JSC::Heap::collectAllGarbage): 4383 (JSC::Heap::primaryHeapBegin): 4384 (JSC::Heap::primaryHeapEnd): 4385 * runtime/Collector.h: 4386 (JSC::): Removed all code pertaining to the number heap, and changed all 4387 heap template functions and classes to non-template functions and classes. 4388 4389 (JSC::Heap::allocateNumber): A new optimization to replace the number 4390 heap: allocate half-sized number cells in pairs, returning the first 4391 cell and caching the second cell for the next allocation. 4392 4393 * runtime/CollectorHeapIterator.h: 4394 (JSC::LiveObjectIterator::LiveObjectIterator): 4395 (JSC::LiveObjectIterator::operator++): 4396 (JSC::DeadObjectIterator::DeadObjectIterator): 4397 (JSC::DeadObjectIterator::operator++): 4398 (JSC::ObjectIterator::ObjectIterator): 4399 (JSC::ObjectIterator::operator++): 4400 * runtime/JSCell.h: 4401 (JSC::JSCell::isNumber): Removed all code pertaining to the number heap, 4402 and changed all heap template functions and classes to non-template functions 4403 and classes. 4404 44052009-12-15 Zoltan Horvath <zoltan@webkit.org> 4406 4407 Reviewed by Darin Adler. 4408 4409 Allow custom memory allocation control for WeakGCMap class 4410 https://bugs.webkit.org/show_bug.cgi?id=32547 4411 4412 Inherits WeakGCMap from FastAllocBase because it is instantiated by 4413 'new' at: WebCore/dom/Document.cpp:512. 4414 4415 * runtime/WeakGCMap.h: 4416 44172009-12-15 Zoltan Horvath <zoltan@webkit.org> 4418 4419 Reviewed by Darin Adler. 4420 4421 Allow custom memory allocation control for dtoa's P5Node struct 4422 https://bugs.webkit.org/show_bug.cgi?id=32544 4423 4424 Inherits P5Node struct from Noncopyable because it is instantiated by 4425 'new' at wtf/dtoa.cpp:588 and don't need to be copyable. 4426 4427 * wtf/dtoa.cpp: 4428 44292009-12-14 Geoffrey Garen <ggaren@apple.com> 4430 4431 Reviewed by Simon Fraser. 4432 4433 https://bugs.webkit.org/show_bug.cgi?id=32524 4434 REGRESSION(52084): fast/dom/prototypes.html failing two CSS tests 4435 4436 * wtf/StdLibExtras.h: 4437 (WTF::bitCount): The original patch put the parentheses in the wrong 4438 place, completely changing the calculation and making it almost always 4439 wrong. Moved the parentheses around the '+' operation, like the original 4440 compiler warning suggested. 4441 44422009-12-14 Gabor Loki <loki@inf.u-szeged.hu> 4443 4444 Unreviewed trivial buildfix. 4445 4446 Fix crosses initialization of usedPrimaryBlocks for JSValue32 4447 4448 * runtime/Collector.cpp: 4449 (JSC::Heap::markConservatively): 4450 44512009-12-14 Csaba Osztrogonác <ossy@webkit.org> 4452 4453 Reviewed by Simon Hausmann. 4454 4455 GCC 4.3.x warning fixed. Suggested parantheses added. 4456 warning: ../../../JavaScriptCore/wtf/StdLibExtras.h:77: warning: suggest parentheses around + or - in operand of & 4457 4458 * wtf/StdLibExtras.h: 4459 (WTF::bitCount): 4460 44612009-12-13 Geoffrey Garen <ggaren@apple.com> 4462 4463 Reviewed by Sam Weinig. 4464 4465 Changed GC from mark-sweep to mark-allocate. 4466 4467 Added WeakGCMap to keep WebCore blissfully ignorant about objects that 4468 have become garbage but haven't run their destructors yet. 4469 4470 1% SunSpider speedup. 4471 7.6% v8 speedup (37% splay speedup). 4472 17% speedup on bench-alloc-nonretained.js. 4473 18% speedup on bench-alloc-retained.js. 4474 4475 * API/JSBase.cpp: 4476 (JSGarbageCollect): 4477 * API/JSContextRef.cpp: 4478 * JavaScriptCore.exp: 4479 * JavaScriptCore.xcodeproj/project.pbxproj: Updated for renames and new 4480 files. 4481 4482 * debugger/Debugger.cpp: 4483 (JSC::Debugger::recompileAllJSFunctions): Updated to use the Collector 4484 iterator abstraction. 4485 4486 * jsc.cpp: 4487 (functionGC): Updated for rename. 4488 4489 * runtime/Collector.cpp: Slightly reduced the number of allocations per 4490 collection, so that small workloads only allocate on collector block, 4491 rather than two. 4492 4493 (JSC::Heap::Heap): Updated to use the new allocateBlock function. 4494 4495 (JSC::Heap::destroy): Updated to use the new freeBlocks function. 4496 4497 (JSC::Heap::allocateBlock): New function to initialize a block when 4498 allocating it. 4499 4500 (JSC::Heap::freeBlock): Consolidated the responsibility for running 4501 destructors into this function. 4502 4503 (JSC::Heap::freeBlocks): Updated to use freeBlock. 4504 4505 (JSC::Heap::recordExtraCost): Sweep the heap in this reporting function, 4506 so that allocation, which is more common, doesn't have to check extraCost. 4507 4508 (JSC::Heap::heapAllocate): Run destructors right before recycling a 4509 garbage cell. This has better cache utilization than a separate sweep phase. 4510 4511 (JSC::Heap::resizeBlocks): 4512 (JSC::Heap::growBlocks): 4513 (JSC::Heap::shrinkBlocks): New set of functions for managing the size of 4514 the heap, now that the heap doesn't maintain any information about its 4515 size. 4516 4517 (JSC::isPointerAligned): 4518 (JSC::isHalfCellAligned): 4519 (JSC::isPossibleCell): 4520 (JSC::isCellAligned): 4521 (JSC::Heap::markConservatively): Cleaned up this code a bit. 4522 4523 (JSC::Heap::clearMarkBits): 4524 (JSC::Heap::markedCells): Some helper functions for examining the the mark 4525 bitmap. 4526 4527 (JSC::Heap::sweep): Simplified this function by using a DeadObjectIterator. 4528 4529 (JSC::Heap::markRoots): Reordered some operations for clarity. 4530 4531 (JSC::Heap::objectCount): 4532 (JSC::Heap::addToStatistics): 4533 (JSC::Heap::statistics): Rewrote these functions to calculate an object 4534 count on demand, since the heap doesn't maintain this information by 4535 itself. 4536 4537 (JSC::Heap::reset): New function for resetting the heap once we've 4538 exhausted heap space. 4539 4540 (JSC::Heap::collectAllGarbage): This function matches the old collect() 4541 behavior, but it's now an uncommon function used only by API. 4542 4543 * runtime/Collector.h: 4544 (JSC::CollectorBitmap::count): 4545 (JSC::CollectorBitmap::isEmpty): Added some helper functions for managing 4546 the collector mark bitmap. 4547 4548 (JSC::Heap::reportExtraMemoryCost): Changed reporting from cell equivalents 4549 to bytes, so it's easier to understand. 4550 4551 * runtime/CollectorHeapIterator.h: 4552 (JSC::CollectorHeapIterator::CollectorHeapIterator): 4553 (JSC::CollectorHeapIterator::operator!=): 4554 (JSC::CollectorHeapIterator::operator*): 4555 (JSC::CollectorHeapIterator::advance): 4556 (JSC::::LiveObjectIterator): 4557 (JSC::::operator): 4558 (JSC::::DeadObjectIterator): 4559 (JSC::::ObjectIterator): New iterators for encapsulating details about 4560 heap layout, and what's live and dead on the heap. 4561 4562 * runtime/JSArray.cpp: 4563 (JSC::JSArray::putSlowCase): 4564 (JSC::JSArray::increaseVectorLength): Delay reporting extra cost until 4565 we're fully constructed, so the heap mark phase won't visit us in an 4566 invalid state. 4567 4568 * runtime/JSCell.h: 4569 (JSC::JSCell::): 4570 (JSC::JSCell::createDummyStructure): 4571 (JSC::JSCell::JSCell): 4572 * runtime/JSGlobalData.cpp: 4573 (JSC::JSGlobalData::JSGlobalData): 4574 * runtime/JSGlobalData.h: Added a dummy cell to simplify allocation logic. 4575 4576 * runtime/JSString.h: 4577 (JSC::jsSubstring): Don't report extra cost for substrings, since they 4578 share a buffer that's already reported extra cost. 4579 4580 * runtime/Tracing.d: 4581 * runtime/Tracing.h: Changed these dtrace hooks not to report object 4582 counts, since they're no longer cheap to compute. 4583 4584 * runtime/UString.h: Updated for renames. 4585 4586 * runtime/WeakGCMap.h: Added. 4587 (JSC::WeakGCMap::isEmpty): 4588 (JSC::WeakGCMap::uncheckedGet): 4589 (JSC::WeakGCMap::uncheckedBegin): 4590 (JSC::WeakGCMap::uncheckedEnd): 4591 (JSC::::get): 4592 (JSC::::take): 4593 (JSC::::set): 4594 (JSC::::uncheckedRemove): Mentioned above. 4595 4596 * wtf/StdLibExtras.h: 4597 (WTF::bitCount): Added a bit population count function, so the heap can 4598 count live objects to fulfill statistics questions. 4599 4600The very last cell in the block is not allocated -- should not be marked. 4601 46022009-12-13 Geoffrey Garen <ggaren@apple.com> 4603 4604 Windows build fix: Export some new symbols. 4605 4606 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 4607 46082009-12-13 Geoffrey Garen <ggaren@apple.com> 4609 4610 Windows build fix: Removed some old exports. 4611 4612 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 4613 46142009-12-13 Geoffrey Garen <ggaren@apple.com> 4615 4616 Windows build fix: Use unsigned instead of uint32_t to avoid dependencies. 4617 4618 * wtf/StdLibExtras.h: 4619 (WTF::bitCount): 4620 46212009-12-13 Gavin Barraclough <barraclough@apple.com> 4622 4623 Reviewed by NOBODY (speculative Windows build fix). 4624 4625 * runtime/JSGlobalObjectFunctions.cpp: 4626 46272009-12-13 Gavin Barraclough <barraclough@apple.com> 4628 4629 Reviewed by Sam Weinig. 4630 4631 https://bugs.webkit.org/show_bug.cgi?id=32496 4632 Switch remaining cases of string construction to use StringBuilder. 4633 Builds strings using a vector rather than using string append / addition. 4634 4635 * JavaScriptCore.exp: 4636 * JavaScriptCore.xcodeproj/project.pbxproj: 4637 * runtime/Executable.cpp: 4638 (JSC::FunctionExecutable::paramString): 4639 * runtime/FunctionConstructor.cpp: 4640 (JSC::constructFunction): 4641 * runtime/JSGlobalObjectFunctions.cpp: 4642 (JSC::encode): 4643 (JSC::decode): 4644 (JSC::globalFuncEscape): 4645 (JSC::globalFuncUnescape): 4646 * runtime/JSONObject.cpp: 4647 (JSC::Stringifier::stringify): 4648 (JSC::Stringifier::indent): 4649 * runtime/JSString.h: 4650 * runtime/LiteralParser.cpp: 4651 (JSC::LiteralParser::Lexer::lexString): 4652 * runtime/NumberPrototype.cpp: 4653 (JSC::integerPartNoExp): 4654 (JSC::numberProtoFuncToFixed): 4655 (JSC::numberProtoFuncToPrecision): 4656 * runtime/Operations.h: 4657 (JSC::jsString): 4658 * runtime/StringPrototype.cpp: 4659 (JSC::substituteBackreferencesSlow): 4660 (JSC::substituteBackreferences): 4661 (JSC::stringProtoFuncConcat): 4662 46632009-12-08 Jeremy Moskovich <jeremy@chromium.org> 4664 4665 Reviewed by Eric Seidel. 4666 4667 Add code to allow toggling ATSUI/Core Text rendering at runtime in ComplexTextController. 4668 https://bugs.webkit.org/show_bug.cgi?id=31802 4669 4670 The goal here is to allow for a zero runtime hit for ports that decide to select 4671 the API at compile time. 4672 When both USE(ATSUI) and USE(CORE_TEXT) are true, the API is toggled 4673 at runtime. Core Text is used for OS Versions >= 10.6. 4674 4675 * wtf/Platform.h: #define USE_CORE_TEXT and USE_ATSUI on Chrome/Mac. 4676 46772009-12-11 Maciej Stachowiak <mjs@apple.com> 4678 4679 Reviewed by Oliver Hunt. 4680 4681 Unify codegen for forward and backward variants of branches 4682 https://bugs.webkit.org/show_bug.cgi?id=32463 4683 4684 * jit/JIT.h: 4685 (JSC::JIT::emit_op_loop): Implemented in terms of forward variant. 4686 (JSC::JIT::emit_op_loop_if_true): ditto 4687 (JSC::JIT::emitSlow_op_loop_if_true): ditto 4688 (JSC::JIT::emit_op_loop_if_false): ditto 4689 (JSC::JIT::emitSlow_op_loop_if_false): ditto 4690 (JSC::JIT::emit_op_loop_if_less): ditto 4691 (JSC::JIT::emitSlow_op_loop_if_less): ditto 4692 * jit/JITOpcodes.cpp: 4693 46942009-12-11 Sam Weinig <sam@webkit.org> 4695 4696 Reviewed by Anders Carlsson. 4697 4698 Allow WTFs concept of the main thread to differ from pthreads when necessary. 4699 4700 * wtf/ThreadingPthreads.cpp: 4701 (WTF::initializeThreading): 4702 (WTF::isMainThread): 4703 * wtf/mac/MainThreadMac.mm: 4704 (WTF::initializeMainThreadPlatform): 4705 (WTF::scheduleDispatchFunctionsOnMainThread): 4706 47072009-12-11 Gavin Barraclough <barraclough@apple.com> 4708 4709 Reviewed by Oliver Hunt. 4710 4711 https://bugs.webkit.org/show_bug.cgi?id=32454 4712 Refactor construction of simple strings to avoid string concatenation. 4713 4714 Building strings through concatenation has a memory and performance cost - 4715 a memory cost since we must over-allocate the buffer to leave space to append 4716 into, and performance in that the string may still require reallocation (and 4717 thus copying during construction). Instead move the full construction to 4718 within a single function call (makeString), so that the arguments' lengths 4719 can be calculated and an appropriate sized buffer allocated before copying 4720 any characters. 4721 4722 ~No performance change (~2% progression on date tests). 4723 4724 * bytecode/CodeBlock.cpp: 4725 (JSC::escapeQuotes): 4726 (JSC::valueToSourceString): 4727 (JSC::constantName): 4728 (JSC::idName): 4729 (JSC::CodeBlock::registerName): 4730 (JSC::regexpToSourceString): 4731 (JSC::regexpName): 4732 * bytecompiler/NodesCodegen.cpp: 4733 (JSC::substitute): 4734 * profiler/Profiler.cpp: 4735 (JSC::Profiler::createCallIdentifier): 4736 * runtime/DateConstructor.cpp: 4737 (JSC::callDate): 4738 * runtime/DateConversion.cpp: 4739 (JSC::formatDate): 4740 (JSC::formatDateUTCVariant): 4741 (JSC::formatTime): 4742 (JSC::formatTimeUTC): 4743 * runtime/DateConversion.h: 4744 (JSC::): 4745 * runtime/DatePrototype.cpp: 4746 (JSC::dateProtoFuncToString): 4747 (JSC::dateProtoFuncToUTCString): 4748 (JSC::dateProtoFuncToDateString): 4749 (JSC::dateProtoFuncToTimeString): 4750 (JSC::dateProtoFuncToGMTString): 4751 * runtime/ErrorPrototype.cpp: 4752 (JSC::errorProtoFuncToString): 4753 * runtime/ExceptionHelpers.cpp: 4754 (JSC::createUndefinedVariableError): 4755 (JSC::createErrorMessage): 4756 (JSC::createInvalidParamError): 4757 * runtime/FunctionPrototype.cpp: 4758 (JSC::insertSemicolonIfNeeded): 4759 (JSC::functionProtoFuncToString): 4760 * runtime/ObjectPrototype.cpp: 4761 (JSC::objectProtoFuncToString): 4762 * runtime/RegExpConstructor.cpp: 4763 (JSC::constructRegExp): 4764 * runtime/RegExpObject.cpp: 4765 (JSC::RegExpObject::match): 4766 * runtime/RegExpPrototype.cpp: 4767 (JSC::regExpProtoFuncCompile): 4768 (JSC::regExpProtoFuncToString): 4769 * runtime/StringPrototype.cpp: 4770 (JSC::stringProtoFuncBig): 4771 (JSC::stringProtoFuncSmall): 4772 (JSC::stringProtoFuncBlink): 4773 (JSC::stringProtoFuncBold): 4774 (JSC::stringProtoFuncFixed): 4775 (JSC::stringProtoFuncItalics): 4776 (JSC::stringProtoFuncStrike): 4777 (JSC::stringProtoFuncSub): 4778 (JSC::stringProtoFuncSup): 4779 (JSC::stringProtoFuncFontcolor): 4780 (JSC::stringProtoFuncFontsize): 4781 (JSC::stringProtoFuncAnchor): 4782 * runtime/UString.h: 4783 (JSC::): 4784 (JSC::makeString): 4785 47862009-12-10 Gavin Barraclough <barraclough@apple.com> 4787 4788 Reviewed by Oliver Hunt. 4789 4790 https://bugs.webkit.org/show_bug.cgi?id=32400 4791 Switch remaining cases of string addition to use ropes. 4792 4793 Re-landing r51975 - added toPrimitiveString method, 4794 performs toPrimitive then subsequent toString operations. 4795 4796 ~1% progression on Sunspidey. 4797 4798 * jit/JITStubs.cpp: 4799 (JSC::DEFINE_STUB_FUNCTION): 4800 * runtime/JSString.h: 4801 (JSC::JSString::JSString): 4802 (JSC::JSString::appendStringInConstruct): 4803 * runtime/Operations.cpp: 4804 (JSC::jsAddSlowCase): 4805 * runtime/Operations.h: 4806 (JSC::jsString): 4807 (JSC::jsAdd): 4808 48092009-12-11 Adam Roben <aroben@apple.com> 4810 4811 Windows build fix 4812 4813 * JavaScriptCore.vcproj/jsc/jscCommon.vsprops: Added 4814 $(WebKitOutputDir)/include/private to the include path. 4815 48162009-12-11 Adam Roben <aroben@apple.com> 4817 4818 Move QuartzCorePresent.h to include/private 4819 4820 This fixes other projects that use wtf/Platform.h 4821 4822 Rubber-stamped by Steve Falkenburg. 4823 4824 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: Let VS do its thang. 4825 * JavaScriptCore.vcproj/JavaScriptCore/build-generated-files.sh: Write 4826 QuartzCorePresent.h to $(WebKitOutputDir)/include/private. 4827 4828 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCoreCommon.vsprops: 4829 * JavaScriptCore.vcproj/WTF/WTFCommon.vsprops: 4830 Added $(WebKitOutputDir)/include/private to the include path. 4831 48322009-12-11 Adam Roben <aroben@apple.com> 4833 4834 Fix clean builds and everything rebuilding on every build 4835 4836 Reviewed by Sam Weinig. 4837 4838 * JavaScriptCore.vcproj/JavaScriptCore/build-generated-files.sh: Don't 4839 write out QuartzCorePresent.h if it exists but is older than 4840 QuartzCore.h. Also, create the directory we write QuartzCorePresent.h 4841 into first. 4842 48432009-12-11 Adam Roben <aroben@apple.com> 4844 4845 Windows build fix for systems with spaces in their paths 4846 4847 * JavaScriptCore.vcproj/JavaScriptCore/build-generated-files.sh: Quote some paths. 4848 48492009-12-11 Chris Marrin <cmarrin@apple.com> 4850 4851 Reviewed by Adam Roben. 4852 4853 Add check for presence of QuartzCore headers 4854 https://bugs.webkit.org/show_bug.cgi?id=31856 4855 4856 The script now checks for the presence of QuartzCore.h. If present 4857 it will turn on ACCELERATED_COMPOSITING and 3D_RENDERING to enable 4858 HW compositing on Windows. The script writes QuartzCorePresent.h to 4859 the build directory which has a define telling whether QuartzCore is 4860 present. 4861 4862 * JavaScriptCore.vcproj/JavaScriptCore/build-generated-files.sh: 4863 * wtf/Platform.h: 4864 48652009-12-11 Kent Tamura <tkent@chromium.org> 4866 4867 Reviewed by Darin Adler. 4868 4869 Fix a problem that JSC::gregorianDateTimeToMS() returns a negative 4870 value for a huge year value. 4871 https://bugs.webkit.org/show_bug.cgi?id=32304 4872 4873 * wtf/DateMath.cpp: 4874 (WTF::dateToDaysFrom1970): Renamed from dateToDayInYear, and changed the return type to double. 4875 (WTF::calculateDSTOffset): Follow the dateToDaysFrom1970() change. 4876 (WTF::timeClip): Use maxECMAScriptTime. 4877 (JSC::gregorianDateTimeToMS): Follow the dateToDaysFrom1970() change. 4878 48792009-12-10 Adam Barth <abarth@webkit.org> 4880 4881 No review, rolling out r51975. 4882 http://trac.webkit.org/changeset/51975 4883 4884 * jit/JITStubs.cpp: 4885 (JSC::DEFINE_STUB_FUNCTION): 4886 * runtime/JSString.h: 4887 (JSC::JSString::JSString): 4888 (JSC::JSString::appendStringInConstruct): 4889 * runtime/Operations.cpp: 4890 (JSC::jsAddSlowCase): 4891 * runtime/Operations.h: 4892 (JSC::jsString): 4893 (JSC::jsAdd): 4894 48952009-12-10 Oliver Hunt <oliver@apple.com> 4896 4897 Reviewed by Gavin Barraclough. 4898 4899 Incorrect caching of prototype lookup with dictionary base 4900 https://bugs.webkit.org/show_bug.cgi?id=32402 4901 4902 Make sure we don't add cached prototype lookup to the proto_list 4903 lookup chain if the top level object is a dictionary. 4904 4905 * jit/JITStubs.cpp: 4906 (JSC::JITThunks::tryCacheGetByID): 4907 49082009-12-10 Gavin Barraclough <barraclough@apple.com> 4909 4910 Reviewed by Oliver Hunt. 4911 4912 https://bugs.webkit.org/show_bug.cgi?id=32400 4913 Switch remaining cases of string addition to use ropes. 4914 4915 ~1% progression on Sunspidey. 4916 4917 * jit/JITStubs.cpp: 4918 (JSC::DEFINE_STUB_FUNCTION): 4919 * runtime/JSString.h: 4920 (JSC::JSString::JSString): 4921 (JSC::JSString::appendStringInConstruct): 4922 * runtime/Operations.cpp: 4923 (JSC::jsAddSlowCase): 4924 * runtime/Operations.h: 4925 (JSC::jsString): 4926 (JSC::jsAdd): 4927 49282009-12-10 Kent Hansen <kent.hansen@nokia.com> 4929 4930 Reviewed by Geoffrey Garen. 4931 4932 Remove JSObject::getPropertyAttributes() and all usage of it. 4933 https://bugs.webkit.org/show_bug.cgi?id=31933 4934 4935 getOwnPropertyDescriptor() should be used instead. 4936 4937 * JavaScriptCore.exp: 4938 * JavaScriptCore.order: 4939 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 4940 * debugger/DebuggerActivation.cpp: 4941 (JSC::DebuggerActivation::getOwnPropertyDescriptor): 4942 * debugger/DebuggerActivation.h: 4943 * runtime/JSObject.cpp: 4944 (JSC::JSObject::propertyIsEnumerable): 4945 * runtime/JSObject.h: 4946 * runtime/JSVariableObject.cpp: 4947 * runtime/JSVariableObject.h: 4948 49492009-12-10 Gavin Barraclough <barraclough@apple.com> 4950 4951 Reviewed by Oliver Hunt & Mark Rowe. 4952 4953 https://bugs.webkit.org/show_bug.cgi?id=32367 4954 Add support for short Ropes (up to 3 entries) inline within JSString. 4955 (rather than externally allocating an object to hold the rope). 4956 Switch jsAdd of (JSString* + JSString*) to now make use of Ropes. 4957 4958 ~1% progression on Sunspidey. 4959 4960 * interpreter/Interpreter.cpp: 4961 (JSC::Interpreter::privateExecute): 4962 * jit/JITOpcodes.cpp: 4963 (JSC::JIT::privateCompileCTIMachineTrampolines): 4964 * jit/JITStubs.cpp: 4965 (JSC::DEFINE_STUB_FUNCTION): 4966 * runtime/JSString.cpp: 4967 (JSC::JSString::resolveRope): 4968 (JSC::JSString::toBoolean): 4969 (JSC::JSString::getStringPropertyDescriptor): 4970 * runtime/JSString.h: 4971 (JSC::JSString::Rope::Fiber::deref): 4972 (JSC::JSString::Rope::Fiber::ref): 4973 (JSC::JSString::Rope::Fiber::refAndGetLength): 4974 (JSC::JSString::Rope::append): 4975 (JSC::JSString::JSString): 4976 (JSC::JSString::~JSString): 4977 (JSC::JSString::value): 4978 (JSC::JSString::tryGetValue): 4979 (JSC::JSString::length): 4980 (JSC::JSString::canGetIndex): 4981 (JSC::JSString::appendStringInConstruct): 4982 (JSC::JSString::appendValueInConstructAndIncrementLength): 4983 (JSC::JSString::isRope): 4984 (JSC::JSString::string): 4985 (JSC::JSString::ropeLength): 4986 (JSC::JSString::getStringPropertySlot): 4987 * runtime/Operations.h: 4988 (JSC::jsString): 4989 (JSC::jsAdd): 4990 (JSC::resolveBase): 4991 49922009-12-09 Anders Carlsson <andersca@apple.com> 4993 4994 Reviewed by Geoffrey Garen. 4995 4996 Fix three more things found by compiling with clang++. 4997 4998 * runtime/Structure.h: 4999 (JSC::StructureTransitionTable::reifySingleTransition): 5000 Add the 'std' qualifier to the call to make_pair. 5001 5002 * wtf/DateMath.cpp: 5003 (WTF::initializeDates): 5004 Incrementing a bool is deprecated according to the C++ specification. 5005 5006 * wtf/PtrAndFlags.h: 5007 (WTF::PtrAndFlags::PtrAndFlags): 5008 Name lookup should not be done in dependent bases, so explicitly qualify the call to set. 5009 50102009-12-09 Maciej Stachowiak <mjs@apple.com> 5011 5012 Reviewed by Oliver Hunt. 5013 5014 Google reader gets stuck in the "Loading..." state and does not complete 5015 https://bugs.webkit.org/show_bug.cgi?id=32256 5016 <rdar://problem/7456388> 5017 5018 * jit/JITArithmetic.cpp: 5019 (JSC::JIT::emitSlow_op_jless): Fix some backward branches. 5020 50212009-12-09 Gavin Barraclough <barraclough@apple.com> 5022 5023 Reviewed by Oliver Hunt. 5024 5025 https://bugs.webkit.org/show_bug.cgi?id=32228 5026 Make destruction of ropes non-recursive to prevent stack exhaustion. 5027 Also, pass a UString& into initializeFiber rather than a Ustring::Rep*, 5028 since the Rep is not being ref counted this could result in usage of a 5029 Rep with refcount zero (where the Rep comes from a temporary UString 5030 returned from a function). 5031 5032 * runtime/JSString.cpp: 5033 (JSC::JSString::Rope::destructNonRecursive): 5034 (JSC::JSString::Rope::~Rope): 5035 * runtime/JSString.h: 5036 (JSC::JSString::Rope::initializeFiber): 5037 * runtime/Operations.h: 5038 (JSC::concatenateStrings): 5039 50402009-12-09 Zoltan Herczeg <zherczeg@inf.u-szeged.hu> 5041 5042 Reviewed by Eric Seidel. 5043 5044 https://bugs.webkit.org/show_bug.cgi?id=31930 5045 5046 Update to r51457. ASSERTs changed to COMPILE_ASSERTs. 5047 The speedup is 25%. 5048 5049 * runtime/JSGlobalData.cpp: 5050 (JSC::VPtrSet::VPtrSet): 5051 50522009-12-09 Steve Block <steveblock@google.com> 5053 5054 Reviewed by Adam Barth. 5055 5056 Updates Android Makefiles with latest additions. 5057 https://bugs.webkit.org/show_bug.cgi?id=32278 5058 5059 * Android.mk: Modified. 5060 * Android.v8.wtf.mk: Modified. 5061 50622009-12-09 Sam Weinig <sam@webkit.org> 5063 5064 Reviewed by Gavin Barraclough. 5065 5066 Fix a bug found while trying to compile JavaScriptCore with clang++. 5067 5068 * yarr/RegexPattern.h: 5069 (JSC::Yarr::PatternTerm::PatternTerm): Don't self assign here. Use false instead. 5070 50712009-12-09 Anders Carlsson <andersca@apple.com> 5072 5073 Reviewed by Sam Weinig. 5074 5075 Attempt to fix the Windows build. 5076 5077 * wtf/FastMalloc.h: 5078 50792009-12-09 Anders Carlsson <andersca@apple.com> 5080 5081 Reviewed by Sam Weinig. 5082 5083 Fix some things found while trying to compile JavaScriptCore with clang++. 5084 5085 * wtf/FastMalloc.h: 5086 Add correct exception specifications for the allocation/deallocation operators. 5087 5088 * wtf/Vector.h: 5089 * wtf/VectorTraits.h: 5090 Fix a bunch of struct/class mismatches. 5091 50922009-12-08 Maciej Stachowiak <mjs@apple.com> 5093 5094 Reviewed by Darin Adler. 5095 5096 move code generation portions of Nodes.cpp to bytecompiler directory 5097 https://bugs.webkit.org/show_bug.cgi?id=32284 5098 5099 * bytecompiler/NodesCodegen.cpp: Copied from parser/Nodes.cpp. Removed parts that 5100 are not about codegen. 5101 * parser/Nodes.cpp: Removed everything that is about codegen. 5102 5103 Update build systems: 5104 5105 * Android.mk: 5106 * GNUmakefile.am: 5107 * JavaScriptCore.gypi: 5108 * JavaScriptCore.pri: 5109 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: 5110 * JavaScriptCore.xcodeproj/project.pbxproj: 5111 * JavaScriptCoreSources.bkl: 5112 51132009-12-08 Kevin Watters <kevinwatters@gmail.com> 5114 5115 Reviewed by Kevin Ollivier. 5116 5117 [wx] Mac plugins support. 5118 5119 https://bugs.webkit.org/show_bug.cgi?id=32236 5120 5121 * wtf/Platform.h: 5122 51232009-12-08 Dmitry Titov <dimich@chromium.org> 5124 5125 Rubber-stamped by David Levin. 5126 5127 Revert and reopen "Add asserts to RefCounted to make sure ref/deref happens on the right thread." 5128 It may have caused massive increase of reported leaks on the bots. 5129 https://bugs.webkit.org/show_bug.cgi?id=31639 5130 5131 * GNUmakefile.am: 5132 * JavaScriptCore.gypi: 5133 * JavaScriptCore.vcproj/WTF/WTF.vcproj: 5134 * JavaScriptCore.xcodeproj/project.pbxproj: 5135 * runtime/Structure.cpp: 5136 (JSC::Structure::Structure): 5137 * wtf/RefCounted.h: 5138 (WTF::RefCountedBase::ref): 5139 (WTF::RefCountedBase::hasOneRef): 5140 (WTF::RefCountedBase::refCount): 5141 (WTF::RefCountedBase::derefBase): 5142 * wtf/ThreadVerifier.h: Removed. 5143 51442009-12-08 Gustavo Noronha Silva <gustavo.noronha@collabora.co.uk> 5145 5146 Reviewed by Darin Adler. 5147 5148 Make WebKit build correctly on FreeBSD, IA64, and Alpha. 5149 Based on work by Petr Salinger <Petr.Salinger@seznam.cz>, 5150 and Colin Watson <cjwatson@ubuntu.com>. 5151 5152 * wtf/Platform.h: 5153 51542009-12-08 Dmitry Titov <dimich@chromium.org> 5155 5156 Reviewed by Darin Adler. 5157 5158 Add asserts to RefCounted to make sure ref/deref happens on the right thread. 5159 https://bugs.webkit.org/show_bug.cgi?id=31639 5160 5161 * runtime/Structure.cpp: 5162 (JSC::Structure::Structure): Disable thread verification on this class since it uses addressOfCount(). 5163 * wtf/RefCounted.h: 5164 (WTF::RefCountedBase::ref): Add ASSERT. 5165 (WTF::RefCountedBase::hasOneRef): Ditto. 5166 (WTF::RefCountedBase::refCount): Ditto. 5167 (WTF::RefCountedBase::derefBase): Ditto. 5168 (WTF::RefCountedBase::disableThreadVerification): delegate to ThreadVerifier method. 5169 * wtf/ThreadVerifier.h: Added. 5170 (WTF::ThreadVerifier::ThreadVerifier): New Debug-only class to verify that ref/deref of RefCounted is done on the same thread. 5171 (WTF::ThreadVerifier::activate): Activates checks. Called when ref count becomes above 2. 5172 (WTF::ThreadVerifier::deactivate): Deactivates checks. Called when ref count drops below 2. 5173 (WTF::ThreadVerifier::disableThreadVerification): used on objects that should not be checked (StringImpl etc) 5174 (WTF::ThreadVerifier::verifyThread): 5175 * GNUmakefile.am: Add ThreadVerifier.h to the build file. 5176 * JavaScriptCore.gypi: Ditto. 5177 * JavaScriptCore.vcproj/WTF/WTF.vcproj: Ditto. 5178 * JavaScriptCore.xcodeproj/project.pbxproj: Ditto. 5179 51802009-12-08 Steve Block <steveblock@google.com> 5181 5182 Reviewed by Adam Barth. 5183 5184 [Android] Adds Makefiles for Android port. 5185 https://bugs.webkit.org/show_bug.cgi?id=31325 5186 5187 * Android.mk: Added. 5188 * Android.v8.wtf.mk: Added. 5189 51902009-12-07 Dmitry Titov <dimich@chromium.org> 5191 5192 Rubber-stamped by Darin Adler. 5193 5194 Remove ENABLE_SHARED_SCRIPT flags 5195 https://bugs.webkit.org/show_bug.cgi?id=32245 5196 This patch was obtained by "git revert" command and then un-reverting of ChangeLog files. 5197 5198 * Configurations/FeatureDefines.xcconfig: 5199 * wtf/Platform.h: 5200 52012009-12-07 Gavin Barraclough <barraclough@apple.com> 5202 5203 Reviewed by NOBODY (Windows build fixage part I). 5204 5205 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 5206 52072009-12-05 Gavin Barraclough <barraclough@apple.com> 5208 5209 Reviewed by Oliver Hunt. 5210 5211 https://bugs.webkit.org/show_bug.cgi?id=32184 5212 Handle out-of-memory conditions with JSC Ropes with a JS exception, rather than crashing. 5213 Switch from using fastMalloc to tryFastMalloc, pass an ExecState to record the exception on. 5214 5215 * API/JSCallbackObjectFunctions.h: 5216 (JSC::::toString): 5217 * API/JSValueRef.cpp: 5218 (JSValueIsStrictEqual): 5219 * JavaScriptCore.exp: 5220 * bytecompiler/BytecodeGenerator.cpp: 5221 (JSC::BytecodeGenerator::emitEqualityOp): 5222 * debugger/DebuggerCallFrame.cpp: 5223 (JSC::DebuggerCallFrame::functionName): 5224 (JSC::DebuggerCallFrame::calculatedFunctionName): 5225 * interpreter/Interpreter.cpp: 5226 (JSC::Interpreter::callEval): 5227 (JSC::Interpreter::privateExecute): 5228 * jit/JITStubs.cpp: 5229 (JSC::DEFINE_STUB_FUNCTION): 5230 * profiler/ProfileGenerator.cpp: 5231 (JSC::ProfileGenerator::addParentForConsoleStart): 5232 * profiler/Profiler.cpp: 5233 (JSC::Profiler::willExecute): 5234 (JSC::Profiler::didExecute): 5235 (JSC::Profiler::createCallIdentifier): 5236 (JSC::createCallIdentifierFromFunctionImp): 5237 * profiler/Profiler.h: 5238 * runtime/ArrayPrototype.cpp: 5239 (JSC::arrayProtoFuncIndexOf): 5240 (JSC::arrayProtoFuncLastIndexOf): 5241 * runtime/DateConstructor.cpp: 5242 (JSC::constructDate): 5243 * runtime/FunctionPrototype.cpp: 5244 (JSC::functionProtoFuncToString): 5245 * runtime/InternalFunction.cpp: 5246 (JSC::InternalFunction::name): 5247 (JSC::InternalFunction::displayName): 5248 (JSC::InternalFunction::calculatedDisplayName): 5249 * runtime/InternalFunction.h: 5250 * runtime/JSCell.cpp: 5251 (JSC::JSCell::getString): 5252 * runtime/JSCell.h: 5253 (JSC::JSValue::getString): 5254 * runtime/JSONObject.cpp: 5255 (JSC::gap): 5256 (JSC::Stringifier::Stringifier): 5257 (JSC::Stringifier::appendStringifiedValue): 5258 * runtime/JSObject.cpp: 5259 (JSC::JSObject::putDirectFunction): 5260 (JSC::JSObject::putDirectFunctionWithoutTransition): 5261 (JSC::JSObject::defineOwnProperty): 5262 * runtime/JSObject.h: 5263 * runtime/JSPropertyNameIterator.cpp: 5264 (JSC::JSPropertyNameIterator::get): 5265 * runtime/JSString.cpp: 5266 (JSC::JSString::Rope::~Rope): 5267 (JSC::JSString::resolveRope): 5268 (JSC::JSString::getPrimitiveNumber): 5269 (JSC::JSString::toNumber): 5270 (JSC::JSString::toString): 5271 (JSC::JSString::toThisString): 5272 (JSC::JSString::getStringPropertyDescriptor): 5273 * runtime/JSString.h: 5274 (JSC::JSString::Rope::createOrNull): 5275 (JSC::JSString::Rope::operator new): 5276 (JSC::JSString::value): 5277 (JSC::JSString::tryGetValue): 5278 (JSC::JSString::getIndex): 5279 (JSC::JSString::getStringPropertySlot): 5280 (JSC::JSValue::toString): 5281 * runtime/JSValue.h: 5282 * runtime/NativeErrorConstructor.cpp: 5283 (JSC::NativeErrorConstructor::NativeErrorConstructor): 5284 * runtime/Operations.cpp: 5285 (JSC::JSValue::strictEqualSlowCase): 5286 * runtime/Operations.h: 5287 (JSC::JSValue::equalSlowCaseInline): 5288 (JSC::JSValue::strictEqualSlowCaseInline): 5289 (JSC::JSValue::strictEqual): 5290 (JSC::jsLess): 5291 (JSC::jsLessEq): 5292 (JSC::jsAdd): 5293 (JSC::concatenateStrings): 5294 * runtime/PropertyDescriptor.cpp: 5295 (JSC::PropertyDescriptor::equalTo): 5296 * runtime/PropertyDescriptor.h: 5297 * runtime/StringPrototype.cpp: 5298 (JSC::stringProtoFuncReplace): 5299 (JSC::stringProtoFuncToLowerCase): 5300 (JSC::stringProtoFuncToUpperCase): 5301 53022009-12-07 Nikolas Zimmermann <nzimmermann@rim.com> 5303 5304 Reviewed by Holger Freyther. 5305 5306 Turn on (SVG) Filters support, by default. 5307 https://bugs.webkit.org/show_bug.cgi?id=32224 5308 5309 * Configurations/FeatureDefines.xcconfig: Enable FILTERS build flag. 5310 53112009-12-07 Steve Falkenburg <sfalken@apple.com> 5312 5313 Build fix. Be flexible about which version of ICU is used on Windows. 5314 5315 * JavaScriptCore.vcproj/jsc/jscCommon.vsprops: Add optional xcopy commands to copy ICU 4.2. 5316 53172009-12-07 Maciej Stachowiak <mjs@apple.com> 5318 5319 Reviewed by Oliver Hunt. 5320 5321 op_loop_if_less JIT codegen is broken for 64-bit 5322 https://bugs.webkit.org/show_bug.cgi?id=32221 5323 5324 * jit/JITOpcodes.cpp: 5325 (JSC::JIT::emit_op_loop_if_false): Fix codegen in this version - test was backwards. 5326 53272009-12-07 Oliver Hunt <oliver@apple.com> 5328 5329 Reviewed by Maciej Stachowiak. 5330 5331 Object.create fails if properties on the descriptor are getters 5332 https://bugs.webkit.org/show_bug.cgi?id=32219 5333 5334 Correctly initialise the PropertySlots with the descriptor object. 5335 5336 * runtime/ObjectConstructor.cpp: 5337 (JSC::toPropertyDescriptor): 5338 53392009-12-06 Maciej Stachowiak <mjs@apple.com> 5340 5341 Not reviewed, build fix. 5342 5343 Actually tested 64-bit *and* 32-bit build this time. 5344 5345 * jit/JITOpcodes.cpp: 5346 (JSC::JIT::emit_op_loop_if_false): 5347 53482009-12-06 Maciej Stachowiak <mjs@apple.com> 5349 5350 Not reviewed, build fix. 5351 5352 Really really fix 64-bit build for prior patch (actually tested this time). 5353 5354 * jit/JITOpcodes.cpp: 5355 (JSC::JIT::emit_op_loop_if_false): 5356 (JSC::JIT::emitSlow_op_loop_if_false): 5357 53582009-12-06 Maciej Stachowiak <mjs@apple.com> 5359 5360 Not reviewed, build fix. 5361 5362 Really fix 64-bit build for prior patch. 5363 5364 * jit/JITArithmetic.cpp: 5365 (JSC::JIT::emitSlow_op_jless): 5366 53672009-12-06 Maciej Stachowiak <mjs@apple.com> 5368 5369 Not reviewed, build fix. 5370 5371 Fix 64-bit build for prior patch. 5372 5373 * jit/JITOpcodes.cpp: 5374 (JSC::JIT::emitSlow_op_loop_if_less): 5375 53762009-12-05 Maciej Stachowiak <mjs@apple.com> 5377 5378 Reviewed by Oliver Hunt. 5379 5380 conway benchmark spends half it's time in op_less (jump fusion fails) 5381 https://bugs.webkit.org/show_bug.cgi?id=32190 5382 5383 <1% speedup on SunSpider and V8 5384 2x speedup on "conway" benchmark 5385 5386 Two optimizations: 5387 1) Improve codegen for logical operators &&, || and ! in a condition context 5388 5389 When generating code for combinations of &&, || and !, in a 5390 condition context (i.e. in an if statement or loop condition), we 5391 used to produce a value, and then separately jump based on its 5392 truthiness. Now we pass the false and true targets in, and let the 5393 logical operators generate jumps directly. This helps in four 5394 ways: 5395 5396 a) Individual clauses of a short-circuit logical operator can now 5397 jump directly to the then or else clause of an if statement (or to 5398 the top or exit of a loop) instead of jumping to a jump. 5399 5400 b) It used to be that jump fusion with the condition of the first 5401 clause of a logical operator was inhibited, because the register 5402 was ref'd to be used later, in the actual condition jump; this no 5403 longer happens since a jump straight to the final target is 5404 generated directly. 5405 5406 c) It used to be that jump fusion with the condition of the second 5407 clause of a logical operator was inhibited, because there was a 5408 jump target right after the second clause and before the actual 5409 condition jump. But now it's no longer necessary for the first 5410 clause to jump there so jump fusion is not blocked. 5411 5412 d) We avoid generating excess mov statements in some cases. 5413 5414 As a concrete example this source: 5415 5416 if (!((x < q && y < q) || (t < q && z < q))) { 5417 // ... 5418 } 5419 5420 Used to generate this bytecode: 5421 5422 [ 34] less r1, r-15, r-19 5423 [ 38] jfalse r1, 7(->45) 5424 [ 41] less r1, r-16, r-19 5425 [ 45] jtrue r1, 14(->59) 5426 [ 48] less r1, r-17, r-19 5427 [ 52] jfalse r1, 7(->59) 5428 [ 55] less r1, r-18, r-19 5429 [ 59] jtrue r1, 17(->76) 5430 5431 And now generates this bytecode (also taking advantage of the second optimization below): 5432 5433 [ 34] jnless r-15, r-19, 8(->42) 5434 [ 38] jless r-16, r-19, 26(->64) 5435 [ 42] jnless r-17, r-19, 8(->50) 5436 [ 46] jless r-18, r-19, 18(->64) 5437 5438 Note the jump fusion and the fact that there's less jump 5439 indirection - three of the four jumps go straight to the target 5440 clause instead of indirecting through another jump. 5441 5442 2) Implement jless opcode to take advantage of the above, since we'll now often generate 5443 a less followed by a jtrue where fusion is not forbidden. 5444 5445 * parser/Nodes.h: 5446 (JSC::ExpressionNode::hasConditionContextCodegen): Helper function to determine 5447 whether a node supports special conditional codegen. Return false as this is the default. 5448 (JSC::ExpressionNode::emitBytecodeInConditionContext): Assert not reached - only really 5449 defined for nodes that do have conditional codegen. 5450 (JSC::UnaryOpNode::expr): Add const version. 5451 (JSC::LogicalNotNode::hasConditionContextCodegen): Returne true only if subexpression 5452 supports it. 5453 (JSC::LogicalOpNode::hasConditionContextCodegen): Return true. 5454 * parser/Nodes.cpp: 5455 (JSC::LogicalNotNode::emitBytecodeInConditionContext): Implemented - just swap 5456 the true and false targets for the child node. 5457 (JSC::LogicalOpNode::emitBytecodeInConditionContext): Implemented - handle jumps 5458 directly, improving codegen quality. Also handles further nested conditional codegen. 5459 (JSC::ConditionalNode::emitBytecode): Use condition context codegen when available. 5460 (JSC::IfNode::emitBytecode): ditto 5461 (JSC::IfElseNode::emitBytecode): ditto 5462 (JSC::DoWhileNode::emitBytecode): ditto 5463 (JSC::WhileNode::emitBytecode): ditto 5464 (JSC::ForNode::emitBytecode): ditto 5465 5466 * bytecode/Opcode.h: 5467 - Added loop_if_false opcode - needed now that falsey jumps can be backwards. 5468 - Added jless opcode to take advantage of new fusion opportunities. 5469 * bytecode/CodeBlock.cpp: 5470 (JSC::CodeBlock::dump): Handle above. 5471 * bytecompiler/BytecodeGenerator.cpp: 5472 (JSC::BytecodeGenerator::emitJumpIfTrue): Add peephole for less + jtrue ==> jless. 5473 (JSC::BytecodeGenerator::emitJumpIfFalse): Add handling of backwrds falsey jumps. 5474 * bytecompiler/BytecodeGenerator.h: 5475 (JSC::BytecodeGenerator::emitNodeInConditionContext): Wrapper to handle tracking of 5476 overly deep expressions etc. 5477 * interpreter/Interpreter.cpp: 5478 (JSC::Interpreter::privateExecute): Implement the two new opcodes (loop_if_false, jless). 5479 * jit/JIT.cpp: 5480 (JSC::JIT::privateCompileMainPass): Implement JIT support for the two new opcodes. 5481 (JSC::JIT::privateCompileSlowCases): ditto 5482 * jit/JIT.h: 5483 * jit/JITArithmetic.cpp: 5484 (JSC::JIT::emit_op_jless): 5485 (JSC::JIT::emitSlow_op_jless): ditto 5486 (JSC::JIT::emitBinaryDoubleOp): ditto 5487 * jit/JITOpcodes.cpp: 5488 (JSC::JIT::emitSlow_op_loop_if_less): ditto 5489 (JSC::JIT::emit_op_loop_if_false): ditto 5490 (JSC::JIT::emitSlow_op_loop_if_false): ditto 5491 * jit/JITStubs.cpp: 5492 * jit/JITStubs.h: 5493 (JSC::): 5494 54952009-12-04 Kent Hansen <kent.hansen@nokia.com> 5496 5497 Reviewed by Darin Adler. 5498 5499 JavaScript delete operator should return false for string properties 5500 https://bugs.webkit.org/show_bug.cgi?id=32012 5501 5502 * runtime/StringObject.cpp: 5503 (JSC::StringObject::deleteProperty): 5504 55052009-12-03 Drew Wilson <atwilson@chromium.org> 5506 5507 Rolled back r51633 because it causes a perf regression in Chromium. 5508 5509 * wtf/Platform.h: 5510 55112009-12-03 Gavin Barraclough <barraclough@apple.com> 5512 5513 Try and fix the Windows build. 5514 5515 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: Export a symbol that should be exported. 5516 55172009-12-03 Mark Rowe <mrowe@apple.com> 5518 5519 Try and fix the Mac build. 5520 5521 * JavaScriptCore.exp: Export a symbol that should be exported. 5522 55232009-12-03 Oliver Hunt <oliver@apple.com> 5524 5525 Reviewed by Gavin Barraclough. 5526 5527 REGRESSION(4.0.3-48777): Crash in JSC::ExecState::propertyNames() (Debug-only?) 5528 https://bugs.webkit.org/show_bug.cgi?id=32133 5529 5530 Work around odd GCC-ism and correct the scopechain for use by 5531 calls made while a cachedcall is active on the callstack. 5532 5533 * interpreter/CachedCall.h: 5534 (JSC::CachedCall::newCallFrame): 5535 * runtime/JSArray.cpp: 5536 (JSC::AVLTreeAbstractorForArrayCompare::compare_key_key): 5537 * runtime/StringPrototype.cpp: 5538 (JSC::stringProtoFuncReplace): 5539 55402009-12-03 Gavin Barraclough <barraclough@apple.com> 5541 5542 Reviewed by Oliver "Brraaaaiiiinnnnnzzzzzzzz" Hunt. 5543 5544 https://bugs.webkit.org/show_bug.cgi?id=32136 5545 Add a rope representation to JSString. Presently JSString always holds its data in UString form. 5546 Instead, allow the result of a string concatenation to be represented in a tree form - with a 5547 variable sized, reference-counted rope node retaining a set of UString::Reps (or other rope nopes). 5548 5549 Strings must still currently be resolved down to a flat UString representation before being used, 5550 but by holding the string in a rope representation during construction we can avoid copying data 5551 until we know the final size of the string. 5552 5553 ~2% progression on SunSpider (~25% on date-format-xparb, ~20% on string-validate-input). 5554 5555 * JavaScriptCore.exp: 5556 5557 - Update exports. 5558 5559 * interpreter/Interpreter.cpp: 5560 (JSC::Interpreter::privateExecute): 5561 5562 - Make use of new JSString::length() method to avoid prematurely resolving ropes. 5563 5564 * jit/JITOpcodes.cpp: 5565 (JSC::JIT::privateCompileCTIMachineTrampolines): 5566 5567 - Switch the string length trampoline to read the length directly from JSString::m_length, 5568 rather than from the JSString's UString::Rep's 'len' property. 5569 5570 * jit/JITStubs.cpp: 5571 (JSC::DEFINE_STUB_FUNCTION): 5572 5573 - Modify op_add such that addition of two strings, where either or both strings are already 5574 in rope representation, produces a rope as a result. 5575 5576 * runtime/JSString.cpp: 5577 (JSC::JSString::Rope::~Rope): 5578 (JSC::copyChars): 5579 (JSC::JSString::resolveRope): 5580 (JSC::JSString::getPrimitiveNumber): 5581 (JSC::JSString::toBoolean): 5582 (JSC::JSString::toNumber): 5583 (JSC::JSString::toString): 5584 (JSC::JSString::toThisString): 5585 (JSC::JSString::getStringPropertyDescriptor): 5586 * runtime/JSString.h: 5587 (JSC::JSString::Rope::Fiber::Fiber): 5588 (JSC::JSString::Rope::Fiber::destroy): 5589 (JSC::JSString::Rope::Fiber::isRope): 5590 (JSC::JSString::Rope::Fiber::rope): 5591 (JSC::JSString::Rope::Fiber::string): 5592 (JSC::JSString::Rope::create): 5593 (JSC::JSString::Rope::initializeFiber): 5594 (JSC::JSString::Rope::ropeLength): 5595 (JSC::JSString::Rope::stringLength): 5596 (JSC::JSString::Rope::fibers): 5597 (JSC::JSString::Rope::Rope): 5598 (JSC::JSString::Rope::operator new): 5599 (JSC::JSString::JSString): 5600 (JSC::JSString::value): 5601 (JSC::JSString::length): 5602 (JSC::JSString::isRope): 5603 (JSC::JSString::rope): 5604 (JSC::JSString::string): 5605 (JSC::JSString::canGetIndex): 5606 (JSC::jsSingleCharacterSubstring): 5607 (JSC::JSString::getIndex): 5608 (JSC::jsSubstring): 5609 (JSC::JSString::getStringPropertySlot): 5610 5611 - Add rope form. 5612 5613 * runtime/Operations.h: 5614 (JSC::jsAdd): 5615 (JSC::concatenateStrings): 5616 5617 - Update string concatenation, and addition of ropes, to produce ropes. 5618 5619 * runtime/StringObject.cpp: 5620 (JSC::StringObject::getOwnPropertyNames): 5621 5622 - Make use of new JSString::length() method to avoid prematurely resolving ropes. 5623 56242009-11-23 Jeremy Moskovich <jeremy@chromium.org> 5625 5626 Reviewed by Eric Seidel. 5627 5628 Switch Chrome/Mac to use Core Text APIs rather than ATSUI APIs. 5629 https://bugs.webkit.org/show_bug.cgi?id=31802 5630 5631 No test since this is already covered by existing pixel tests. 5632 5633 * wtf/Platform.h: #define USE_CORE_TEXT for Chrome/Mac. 5634 56352009-12-02 Oliver Hunt <oliver@apple.com> 5636 5637 Reviewed by Gavin Barraclough. 5638 5639 Add files missed in prior patch. 5640 5641 * runtime/JSZombie.cpp: 5642 (JSC::): 5643 (JSC::JSZombie::leakedZombieStructure): 5644 * runtime/JSZombie.h: Added. 5645 (JSC::JSZombie::JSZombie): 5646 (JSC::JSZombie::isZombie): 5647 (JSC::JSZombie::classInfo): 5648 (JSC::JSZombie::isGetterSetter): 5649 (JSC::JSZombie::isAPIValueWrapper): 5650 (JSC::JSZombie::isPropertyNameIterator): 5651 (JSC::JSZombie::getCallData): 5652 (JSC::JSZombie::getConstructData): 5653 (JSC::JSZombie::getUInt32): 5654 (JSC::JSZombie::toPrimitive): 5655 (JSC::JSZombie::getPrimitiveNumber): 5656 (JSC::JSZombie::toBoolean): 5657 (JSC::JSZombie::toNumber): 5658 (JSC::JSZombie::toString): 5659 (JSC::JSZombie::toObject): 5660 (JSC::JSZombie::markChildren): 5661 (JSC::JSZombie::put): 5662 (JSC::JSZombie::deleteProperty): 5663 (JSC::JSZombie::toThisObject): 5664 (JSC::JSZombie::toThisString): 5665 (JSC::JSZombie::toThisJSString): 5666 (JSC::JSZombie::getJSNumber): 5667 (JSC::JSZombie::getOwnPropertySlot): 5668 56692009-12-02 Oliver Hunt <oliver@apple.com> 5670 5671 Reviewed by Gavin Barraclough. 5672 5673 Add zombies to JSC 5674 https://bugs.webkit.org/show_bug.cgi?id=32103 5675 5676 Add a compile time flag to make the JSC collector replace "unreachable" 5677 objects with zombie objects. The zombie object is a JSCell subclass that 5678 ASSERTs on any attempt to use the JSCell methods. In addition there are 5679 a number of additional assertions in bottleneck code to catch zombie usage 5680 as quickly as possible. 5681 5682 Grrr. Argh. Brains. 5683 5684 * JavaScriptCore.xcodeproj/project.pbxproj: 5685 * interpreter/Register.h: 5686 (JSC::Register::Register): 5687 * runtime/ArgList.h: 5688 (JSC::MarkedArgumentBuffer::append): 5689 (JSC::ArgList::ArgList): 5690 * runtime/Collector.cpp: 5691 (JSC::Heap::destroy): 5692 (JSC::Heap::sweep): 5693 * runtime/Collector.h: 5694 * runtime/JSCell.h: 5695 (JSC::JSCell::isZombie): 5696 (JSC::JSValue::isZombie): 5697 * runtime/JSValue.h: 5698 (JSC::JSValue::decode): 5699 (JSC::JSValue::JSValue): 5700 * wtf/Platform.h: 5701 57022009-12-01 Jens Alfke <snej@chromium.org> 5703 5704 Reviewed by Darin Adler. 5705 5706 Added variants of find/contains/add that allow a foreign key type to be used. 5707 This will allow AtomicString-keyed maps to be queried by C string without 5708 having to create a temporary AtomicString (see HTTPHeaderMap.) 5709 The code for this is adapted from the equivalent in HashSet.h. 5710 5711 * wtf/HashMap.h: 5712 (WTF::HashMap::find): 5713 (WTF::HashMap::contains): 5714 (WTF::HashMap::add): 5715 * wtf/HashSet.h: Changed "method" to "function member" in a comment. 5716 57172009-12-01 Gustavo Noronha Silva <gustavo.noronha@collabora.co.uk> 5718 5719 Revert 51551 because it broke GTK+. 5720 5721 * wtf/Platform.h: 5722 57232009-11-30 Gavin Barraclough <barraclough@apple.com> 5724 5725 Windows Build fix. Reviewed by NOBODY. 5726 5727 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 5728 57292009-11-24 Gavin Barraclough <barraclough@apple.com> 5730 5731 Reviewed by Geoff Garen. 5732 5733 Bug 31859 - Make world selection for JSC IsolatedWorlds automagical. 5734 5735 WebCore presently has to explicitly specify the world before entering into JSC, 5736 which is a little fragile (particularly since property access via a 5737 getter/setter might invoke execution). Instead derive the current world from 5738 the lexical global object. 5739 5740 Remove the temporary duct tape of willExecute/didExecute virtual hooks on the JSGlobalData::ClientData - these are no longer necessary. 5741 5742 * API/JSBase.cpp: 5743 (JSEvaluateScript): 5744 * API/JSObjectRef.cpp: 5745 (JSObjectCallAsFunction): 5746 * JavaScriptCore.exp: 5747 * runtime/JSGlobalData.cpp: 5748 * runtime/JSGlobalData.h: 5749 57502009-11-30 Laszlo Gombos <laszlo.1.gombos@nokia.com> 5751 5752 Reviewed by Kenneth Rohde Christiansen. 5753 5754 [Qt] Remove obsolete PLATFORM(KDE) code 5755 https://bugs.webkit.org/show_bug.cgi?id=31958 5756 5757 KDE is now using unpatched QtWebKit. 5758 5759 * parser/Lexer.cpp: Remove obsolete KDE_USE_FINAL guard 5760 * wtf/Platform.h: Remove PLATFORM(KDE) definition and code 5761 section that is guarded with it. 5762 57632009-11-30 Jan-Arve Sæther <jan-arve.saether@nokia.com> 5764 5765 Reviewed by Simon Hausmann. 5766 5767 [Qt] Fix compilation with win32-icc 5768 5769 The Intel compiler does not support the __has_trivial_constructor type 5770 trait. The Intel Compiler can report itself as _MSC_VER >= 1400. The 5771 reason for that is that the Intel Compiler depends on the Microsoft 5772 Platform SDK, and in order to try to be "fully" MS compatible it will 5773 "pretend" to be the same MS compiler as was shipped with the MS PSDK. 5774 (Thus, compiling with win32-icc with VC8 SDK will make the source code 5775 "think" the compiler at hand supports this type trait). 5776 5777 * wtf/TypeTraits.h: 5778 57792009-11-29 Laszlo Gombos <laszlo.1.gombos@nokia.com> 5780 5781 Reviewed by Eric Seidel. 5782 5783 [Qt] Mac build has JIT disabled 5784 https://bugs.webkit.org/show_bug.cgi?id=31828 5785 5786 * wtf/Platform.h: Enable JIT for Qt Mac builds 5787 57882009-11-28 Laszlo Gombos <laszlo.1.gombos@nokia.com> 5789 5790 Reviewed by Eric Seidel. 5791 5792 Apply workaround for the limitation of VirtualFree with MEM_RELEASE to all ports running on Windows 5793 https://bugs.webkit.org/show_bug.cgi?id=31943 5794 5795 * runtime/MarkStack.h: 5796 (JSC::MarkStack::MarkStackArray::shrinkAllocation): 5797 57982009-11-28 Zoltan Herczeg <zherczeg@inf.u-szeged.hu> 5799 5800 Reviewed by Gavin Barraclough. 5801 5802 https://bugs.webkit.org/show_bug.cgi?id=31930 5803 5804 Seems a typo. We don't need ~270k memory to determine the vptrs. 5805 5806 * runtime/JSGlobalData.cpp: 5807 (JSC::VPtrSet::VPtrSet): 5808 58092009-11-27 Shinichiro Hamaji <hamaji@chromium.org> 5810 5811 Unreviewed. 5812 5813 Move GOwnPtr* from wtf to wtf/gtk 5814 https://bugs.webkit.org/show_bug.cgi?id=31793 5815 5816 Build fix for chromium after r51423. 5817 Exclude gtk directory from chromium build. 5818 5819 * JavaScriptCore.gyp/JavaScriptCore.gyp: 5820 58212009-11-25 Oliver Hunt <oliver@apple.com> 5822 5823 Reviewed by Gavin Barraclough. 5824 5825 Incorrect behaviour of jneq_null in the interpreter 5826 https://bugs.webkit.org/show_bug.cgi?id=31901 5827 5828 Correct the logic of jneq_null. This is already covered by existing tests. 5829 5830 * interpreter/Interpreter.cpp: 5831 (JSC::Interpreter::privateExecute): 5832 58332009-11-26 Laszlo Gombos <laszlo.1.gombos@nokia.com> 5834 5835 Reviewed by Oliver Hunt. 5836 5837 Move GOwnPtr* from wtf to wtf/gtk 5838 https://bugs.webkit.org/show_bug.cgi?id=31793 5839 5840 * GNUmakefile.am: Change the path for GOwnPtr.*. 5841 * JavaScriptCore.gyp/JavaScriptCore.gyp: Remove 5842 GOwnPtr.cpp from the exclude list. 5843 * JavaScriptCore.gypi: Change the path for GOwnPtr.*. 5844 * wscript: Remove GOwnPtr.cpp from the exclude list. 5845 * wtf/GOwnPtr.cpp: Removed. 5846 * wtf/GOwnPtr.h: Removed. 5847 * wtf/Threading.h: Change the path for GOwnPtr.h. 5848 * wtf/gtk/GOwnPtr.cpp: Copied from JavaScriptCore/wtf/GOwnPtr.cpp. 5849 * wtf/gtk/GOwnPtr.h: Copied from JavaScriptCore/wtf/GOwnPtr.h. 5850 * wtf/unicode/glib/UnicodeGLib.h: Change the path for GOwnPtr.h. 5851 58522009-11-24 Dmitry Titov <dimich@chromium.org> 5853 5854 Reviewed by Eric Seidel. 5855 5856 Add ENABLE_SHARED_SCRIPT feature define and flag for build-webkit 5857 https://bugs.webkit.org/show_bug.cgi?id=31444 5858 5859 * Configurations/FeatureDefines.xcconfig: 5860 * wtf/Platform.h: 5861 58622009-11-24 Chris Marrin <cmarrin@apple.com> 5863 5864 Reviewed by Simon Fraser. 5865 5866 Add ability to enable ACCELERATED_COMPOSITING on Windows (currently disabled) 5867 https://bugs.webkit.org/show_bug.cgi?id=27314 5868 5869 * wtf/Platform.h: 5870 58712009-11-24 Jason Smith <dark.panda@gmail.com> 5872 5873 Reviewed by Alexey Proskuryakov. 5874 5875 RegExp#exec's returned Array-like object behaves differently from 5876 regular Arrays 5877 https://bugs.webkit.org/show_bug.cgi?id=31689 5878 5879 * JavaScriptCore/runtime/RegExpConstructor.cpp: ensure that undefined 5880 values are added to the returned RegExpMatchesArray 5881 58822009-11-24 Oliver Hunt <oliver@apple.com> 5883 5884 Reviewed by Alexey Proskuryakov. 5885 5886 JSON.stringify performance on undefined is very poor 5887 https://bugs.webkit.org/show_bug.cgi?id=31839 5888 5889 Switch from a UString to a Vector<UChar> when building 5890 the JSON string, allowing us to safely remove the substr-copy 5891 we otherwise did when unwinding an undefined property. 5892 5893 Also turns out to be a ~5% speedup on stringification. 5894 5895 * runtime/JSONObject.cpp: 5896 (JSC::Stringifier::StringBuilder::append): 5897 (JSC::Stringifier::stringify): 5898 (JSC::Stringifier::Holder::appendNextProperty): 5899 59002009-11-24 Mark Rowe <mrowe@apple.com> 5901 5902 Fix production builds where the source tree may be read-only. 5903 5904 * JavaScriptCore.xcodeproj/project.pbxproj: 5905 59062009-11-23 Laszlo Gombos <laszlo.1.gombos@nokia.com> 5907 5908 Reviewed by Kenneth Rohde Christiansen. 5909 5910 Include "config.h" to meet Coding Style Guidelines 5911 https://bugs.webkit.org/show_bug.cgi?id=31792 5912 5913 * wtf/unicode/UTF8.cpp: 5914 * wtf/unicode/glib/UnicodeGLib.cpp: 5915 * wtf/unicode/wince/UnicodeWince.cpp: 5916 59172009-11-23 Geoffrey Garen <ggaren@apple.com> 5918 5919 Reviewed by Oliver Hunt. 5920 5921 Streamlined some Math functions where we expect or know the result not 5922 to be representable as an int. 5923 5924 SunSpider says 0.6% faster. 5925 5926 * runtime/JSNumberCell.h: 5927 (JSC::JSValue::JSValue): 5928 * runtime/JSValue.h: 5929 (JSC::JSValue::): 5930 (JSC::jsDoubleNumber): 5931 (JSC::JSValue::JSValue): Added a function for making a numeric JSValue 5932 and skipping the "can I encode this as an int?" check, avoiding the 5933 overhead of int <-> double roundtripping and double <-> double comparison 5934 and branching. 5935 5936 * runtime/MathObject.cpp: 5937 (JSC::mathProtoFuncACos): 5938 (JSC::mathProtoFuncASin): 5939 (JSC::mathProtoFuncATan): 5940 (JSC::mathProtoFuncATan2): 5941 (JSC::mathProtoFuncCos): 5942 (JSC::mathProtoFuncExp): 5943 (JSC::mathProtoFuncLog): 5944 (JSC::mathProtoFuncRandom): 5945 (JSC::mathProtoFuncSin): 5946 (JSC::mathProtoFuncSqrt): 5947 (JSC::mathProtoFuncTan): For these functions, which we expect or know 5948 to produce results not representable as ints, call jsDoubleNumber instead 5949 of jsNumber. 5950 59512009-11-23 Mark Rowe <mrowe@apple.com> 5952 5953 Unreviewed. Unbreak the regression tests after r51329. 5954 5955 * API/JSBase.cpp: 5956 (JSEvaluateScript): Null-check clientData before dereferencing it. 5957 * API/JSObjectRef.cpp: 5958 (JSObjectCallAsFunction): Ditto. 5959 59602009-11-23 Gavin Barraclough <barraclough@apple.com> 5961 5962 Reviewed by Geoff Garen. 5963 5964 Part 1/3 of <rdar://problem/7377477> REGRESSION: Many web pages fail to render after interesting script runs in isolated world 5965 5966 Some clients of the JavaScriptCore API expect to be able to make callbacks over the JSC API, 5967 and for this to automagically cause execution to take place in the world associated with the 5968 global object associated with the ExecState (JSContextRef) passed. However this is not how 5969 things work - the world must be explicitly set within WebCore. 5970 5971 Making this work just for API calls to evaluate & call will be a far from perfect solution, 5972 since direct (non-API) use of JSC still relies on WebCore setting the current world correctly. 5973 A better solution would be to make this all work automagically all throughout WebCore, but this 5974 will require more refactoring. 5975 5976 Since the API is in JSC but worlds only exist in WebCore, add callbacks on the JSGlobalData::ClientData 5977 to allow it to update the current world on entry/exit via the JSC API. This is temporary duck 5978 tape, and should be removed once the current world no longer needs to be explicitly tracked. 5979 5980 * API/JSBase.cpp: 5981 (JSEvaluateScript): 5982 * API/JSObjectRef.cpp: 5983 (JSObjectCallAsFunction): 5984 * JavaScriptCore.exp: 5985 * runtime/JSGlobalData.cpp: 5986 (JSC::JSGlobalData::ClientData::beginningExecution): 5987 (JSC::JSGlobalData::ClientData::completedExecution): 5988 * runtime/JSGlobalData.h: 5989 59902009-11-23 Steve Block <steveblock@google.com> 5991 5992 Reviewed by Dmitry Titov. 5993 5994 Adds MainThreadAndroid.cpp with Android-specific WTF threading functions. 5995 https://bugs.webkit.org/show_bug.cgi?id=31807 5996 5997 * wtf/android: Added. 5998 * wtf/android/MainThreadAndroid.cpp: Added. 5999 (WTF::timeoutFired): 6000 (WTF::initializeMainThreadPlatform): 6001 (WTF::scheduleDispatchFunctionsOnMainThread): 6002 60032009-11-23 Alexey Proskuryakov <ap@apple.com> 6004 6005 Reviewed by Brady Eidson. 6006 6007 https://bugs.webkit.org/show_bug.cgi?id=31748 6008 Make WebSocketHandleCFNet respect proxy auto-configuration files via CFProxySupport 6009 6010 * JavaScriptCore.exp: Export callOnMainThreadAndWait. 6011 60122009-11-23 Laszlo Gombos <laszlo.1.gombos@nokia.com> 6013 6014 Reviewed by Kenneth Rohde Christiansen. 6015 6016 [Symbian] Fix lastIndexOf() for Symbian 6017 https://bugs.webkit.org/show_bug.cgi?id=31773 6018 6019 Symbian soft floating point library has problems with operators 6020 comparing NaN to numbers. Without a workaround lastIndexOf() 6021 function does not work. 6022 6023 Patch developed by David Leong. 6024 6025 * runtime/StringPrototype.cpp: 6026 (JSC::stringProtoFuncLastIndexOf):Add an extra test 6027 to check for NaN for Symbian. 6028 60292009-11-23 Steve Block <steveblock@google.com> 6030 6031 Reviewed by Eric Seidel. 6032 6033 Android port lacks implementation of atomicIncrement and atomicDecrement. 6034 https://bugs.webkit.org/show_bug.cgi?id=31715 6035 6036 * wtf/Threading.h: Modified. 6037 (WTF::atomicIncrement): Added Android implementation. 6038 (WTF::atomicDecrement): Added Android implementation. 6039 60402009-11-22 Laszlo Gombos <laszlo.1.gombos@nokia.com> 6041 6042 Unreviewed. 6043 6044 [Qt] Sort source lists and remove obsolete comments 6045 from the build system. 6046 6047 * JavaScriptCore.pri: 6048 60492009-11-21 Laszlo Gombos <laszlo.1.gombos@nokia.com> 6050 6051 Reviewed by Eric Seidel. 6052 6053 [Qt][Mac] Turn on multiple JavaScript threads for QtWebkit on Mac 6054 https://bugs.webkit.org/show_bug.cgi?id=31753 6055 6056 * wtf/Platform.h: 6057 60582009-11-19 Steve Block <steveblock@google.com> 6059 6060 Android port lacks configuration in Platform.h and config.h. 6061 https://bugs.webkit.org/show_bug.cgi?id=31671 6062 6063 * wtf/Platform.h: Modified. Added Android-specific configuration. 6064 60652009-11-19 Alexey Proskuryakov <ap@apple.com> 6066 6067 Reviewed by Darin Adler. 6068 6069 https://bugs.webkit.org/show_bug.cgi?id=31690 6070 Make SocketStreamHandleCFNet work on Windows 6071 6072 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 6073 * wtf/MainThread.cpp: 6074 (WTF::FunctionWithContext::FunctionWithContext): 6075 (WTF::dispatchFunctionsFromMainThread): 6076 (WTF::callOnMainThreadAndWait): 6077 * wtf/MainThread.h: 6078 Re-add callOnMainThreadAndWait(), which was removed in bug 23926. 6079 60802009-11-19 Dmitry Titov <dimich@chromium.org> 6081 6082 Reviewed by David Levin. 6083 6084 isMainThread() on Chromium (Mac and Linux) is so slow it timeouts LayoutTests.. 6085 https://bugs.webkit.org/show_bug.cgi?id=31693 6086 6087 * wtf/ThreadingPthreads.cpp: 6088 (WTF::initializeThreading): grab and use the pthread_t of the main thread instead of ThreadIdentifier. 6089 (WTF::isMainThread): Ditto. 6090 60912009-11-19 Laszlo Gombos <laszlo.1.gombos@nokia.com> 6092 6093 Reviewed by Darin Adler. 6094 6095 Remove HAVE(STRING_H) guard from JavaScriptCore 6096 https://bugs.webkit.org/show_bug.cgi?id=31668 6097 6098 * config.h: 6099 * runtime/UString.cpp: 6100 61012009-11-19 Dumitru Daniliuc <dumi@chromium.org> 6102 6103 Reviewed by Dmitry Titov. 6104 6105 Fixing a bug in MessageQueue::removeIf() that leads to an 6106 assertion failure. 6107 6108 https://bugs.webkit.org/show_bug.cgi?id=31657 6109 6110 * wtf/MessageQueue.h: 6111 (WTF::MessageQueue::removeIf): 6112 61132009-11-19 Laszlo Gombos <laszlo.1.gombos@nokia.com> 6114 6115 Reviewed by Darin Adler. 6116 6117 Remove HAVE(FLOAT_H) guard 6118 https://bugs.webkit.org/show_bug.cgi?id=31661 6119 6120 JavaScriptCore has a dependency on float.h, there is 6121 no need to guard float.h. 6122 6123 * runtime/DatePrototype.cpp: Remove include directive 6124 for float.h as it is included in MathExtras.h already. 6125 * runtime/Operations.cpp: Ditto. 6126 * runtime/UString.cpp: Ditto. 6127 * wtf/dtoa.cpp: Ditto. 6128 * wtf/MathExtras.h: Remove HAVE(FLOAT_H) guard. 6129 * wtf/Platform.h: Ditto. 6130 61312009-11-19 Thiago Macieira <thiago.macieira@nokia.com> 6132 6133 Reviewed by Simon Hausmann. 6134 6135 Build fix for 32-bit Sparc machines: these machines are big-endian. 6136 6137 * wtf/Platform.h: 6138 61392009-11-18 Laszlo Gombos <laszlo.1.gombos@nokia.com> 6140 6141 Reviewed by Kenneth Rohde Christiansen. 6142 6143 [Qt] Remove support for Qt v4.3 or older versions 6144 https://bugs.webkit.org/show_bug.cgi?id=29469 6145 6146 * JavaScriptCore.pro: 6147 * jsc.pro: 6148 * wtf/unicode/qt4/UnicodeQt4.h: 6149 61502009-11-18 Kent Tamura <tkent@chromium.org> 6151 6152 Reviewed by Darin Adler. 6153 6154 Move UString::from(double) implementation to new 6155 WTF::doubleToStringInJavaScriptFormat(), and expose it because WebCore 6156 code will use it. 6157 https://bugs.webkit.org/show_bug.cgi?id=31330 6158 6159 - Introduce new function createRep(const char*, unsigned) and 6160 UString::UString(const char*, unsigned) to reduce 2 calls to strlen(). 6161 - Fix a bug that dtoa() doesn't update *rve if the input value is NaN 6162 or Infinity. 6163 6164 No new tests because this doesn't change the behavior. 6165 6166 * JavaScriptCore.exp: 6167 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 6168 * runtime/UString.cpp: 6169 (JSC::createRep): 6170 (JSC::UString::UString): 6171 (JSC::UString::from): Move the code to doubleToStringInJavaScriptFormat(). 6172 * runtime/UString.h: 6173 * wtf/dtoa.cpp: 6174 (WTF::dtoa): Fix a bug about rve. 6175 (WTF::append): A helper for doubleToStringInJavaScriptFormat(). 6176 (WTF::doubleToStringInJavaScriptFormat): Move the code from UString::from(double). 6177 * wtf/dtoa.h: 6178 61792009-11-18 Laszlo Gombos <laszlo.1.gombos@nokia.com> 6180 6181 Reviewed by Kenneth Rohde Christiansen. 6182 6183 [Qt] Remove WTF_USE_JAVASCRIPTCORE_BINDINGS as it is no longer used 6184 https://bugs.webkit.org/show_bug.cgi?id=31643 6185 6186 * JavaScriptCore.pro: 6187 61882009-11-18 Nate Chapin <japhet@chromium.org> 6189 6190 Reviewed by Darin Fisher. 6191 6192 Remove Chromium's unnecessary dependency on wtf's tcmalloc files. 6193 6194 https://bugs.webkit.org/show_bug.cgi?id=31648 6195 6196 * JavaScriptCore.gyp/JavaScriptCore.gyp: 6197 61982009-11-18 Thiago Macieira <thiago.macieira@nokia.com> 6199 6200 Reviewed by Gavin Barraclough. 6201 6202 [Qt] Implement symbol hiding for JSC's JIT functions. 6203 6204 These functions are implemented directly in assembly, so they need the 6205 proper directives to enable/disable visibility. On ELF systems, it's 6206 .hidden, whereas on Mach-O systems (Mac) it's .private_extern. On 6207 Windows, it's not necessary since you have to explicitly export. I 6208 also implemented the AIX idiom, though it's unlikely anyone will 6209 implement AIX/POWER JIT. 6210 https://bugs.webkit.org/show_bug.cgi?id=30864 6211 6212 * jit/JITStubs.cpp: 6213 62142009-11-18 Oliver Hunt <oliver@apple.com> 6215 6216 Reviewed by Alexey Proskuryakov. 6217 6218 Interpreter may do an out of range access when throwing an exception in the profiler. 6219 https://bugs.webkit.org/show_bug.cgi?id=31635 6220 6221 Add bounds check. 6222 6223 * interpreter/Interpreter.cpp: 6224 (JSC::Interpreter::throwException): 6225 62262009-11-18 Gabor Loki <loki@inf.u-szeged.hu> 6227 6228 Reviewed by Darin Adler. 6229 6230 Fix the clobber list of cacheFlush for ARM and Thumb2 on Linux 6231 https://bugs.webkit.org/show_bug.cgi?id=31631 6232 6233 * jit/ExecutableAllocator.h: 6234 (JSC::ExecutableAllocator::cacheFlush): 6235 62362009-11-18 Harald Fernengel <harald.fernengel@nokia.com> 6237 6238 Reviewed by Simon Hausmann. 6239 6240 [Qt] Fix detection of linux-g++ 6241 6242 Never use "linux-g++*" to check for linux-g++, since this will break embedded 6243 builds which use linux-arm-g++ and friends. Use 'linux*-g++*' to check for any 6244 g++ on linux mkspec. 6245 6246 * JavaScriptCore.pri: 6247 62482009-11-17 Jon Honeycutt <jhoneycutt@apple.com> 6249 6250 Add JSContextRefPrivate.h to list of copied files. 6251 6252 Reviewed by Mark Rowe. 6253 6254 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCoreGenerated.make: 6255 62562009-11-17 Martin Robinson <martin.james.robinson@gmail.com> 6257 6258 Reviewed by Adam Barth. 6259 6260 [GTK] Style cleanup for GOwnPtr 6261 https://bugs.webkit.org/show_bug.cgi?id=31506 6262 6263 Remove forward declaration in GOwnPtr and do some style cleanup. 6264 6265 * wtf/GOwnPtr.cpp: 6266 * wtf/GOwnPtr.h: 6267 (WTF::GOwnPtr::GOwnPtr): 6268 (WTF::GOwnPtr::~GOwnPtr): 6269 (WTF::GOwnPtr::get): 6270 (WTF::GOwnPtr::release): 6271 (WTF::GOwnPtr::outPtr): 6272 (WTF::GOwnPtr::set): 6273 (WTF::GOwnPtr::clear): 6274 (WTF::GOwnPtr::operator*): 6275 (WTF::GOwnPtr::operator->): 6276 (WTF::GOwnPtr::operator!): 6277 (WTF::GOwnPtr::operator UnspecifiedBoolType): 6278 (WTF::GOwnPtr::swap): 6279 (WTF::swap): 6280 (WTF::operator==): 6281 (WTF::operator!=): 6282 (WTF::getPtr): 6283 (WTF::freeOwnedGPtr): 6284 62852009-11-17 Oliver Hunt <oliver@apple.com> 6286 6287 Reviewed by Maciej Stachowiak. 6288 6289 Incorrect use of JavaScriptCore API in DumpRenderTree 6290 https://bugs.webkit.org/show_bug.cgi?id=31577 6291 6292 Add assertions to the 'toJS' functions to catch mistakes like 6293 this early. Restructure existing code which blindly passed potentially 6294 null values to toJS when forwarding exceptions so that a null check is 6295 performed first. 6296 6297 * API/APICast.h: 6298 (toJS): 6299 (toJSForGC): 6300 * API/JSCallbackObjectFunctions.h: 6301 (JSC::::getOwnPropertySlot): 6302 (JSC::::put): 6303 (JSC::::deleteProperty): 6304 (JSC::::construct): 6305 (JSC::::hasInstance): 6306 (JSC::::call): 6307 (JSC::::toNumber): 6308 (JSC::::toString): 6309 (JSC::::staticValueGetter): 6310 (JSC::::callbackGetter): 6311 * API/tests/testapi.c: Fix errors in the API tester. 6312 (MyObject_getProperty): 6313 (MyObject_convertToType): 6314 (EvilExceptionObject_convertToType): 6315 63162009-11-16 Zoltan Herczeg <zherczeg@inf.u-szeged.hu> 6317 6318 Reviewed by Gavin Barraclough. 6319 6320 https://bugs.webkit.org/show_bug.cgi?id=31050 6321 6322 Minor fixes for JSVALUE32_64: branchConvertDoubleToInt32 6323 failed on a CortexA8 CPU, but not on a simulator; and 6324 JITCall.cpp modifications was somehow not committed to mainline. 6325 6326 * assembler/ARMAssembler.h: 6327 (JSC::ARMAssembler::fmrs_r): 6328 * assembler/MacroAssemblerARM.h: 6329 (JSC::MacroAssemblerARM::branchConvertDoubleToInt32): 6330 * jit/JITCall.cpp: 6331 (JSC::JIT::compileOpCall): 6332 63332009-11-16 Joerg Bornemann <joerg.bornemann@trolltech.com> 6334 6335 Reviewed by Simon Hausmann. 6336 6337 Fix Qt build on Windows CE 6. 6338 6339 * JavaScriptCore.pri: Add missing include path. 6340 * wtf/Platform.h: Include ce_time.h for Windows CE 6. 6341 63422009-11-13 Zoltan Herczeg <zherczeg@inf.u-szeged.hu> 6343 6344 Reviewed by Gavin Barraclough. 6345 6346 https://bugs.webkit.org/show_bug.cgi?id=31050 6347 6348 Adding optimization support for mode JSVALUE32_64 6349 on ARM systems. 6350 6351 * jit/JIT.h: 6352 * jit/JITCall.cpp: 6353 (JSC::JIT::compileOpCall): 6354 * jit/JITPropertyAccess.cpp: 6355 (JSC::JIT::emit_op_method_check): 6356 (JSC::JIT::compileGetByIdHotPath): 6357 (JSC::JIT::compileGetByIdSlowCase): 6358 (JSC::JIT::emit_op_put_by_id): 6359 63602009-11-14 Zoltan Herczeg <zherczeg@inf.u-szeged.hu> 6361 6362 Reviewed by Gavin Barraclough. 6363 6364 https://bugs.webkit.org/show_bug.cgi?id=31050 6365 6366 Adding JSVALUE32_64 support for ARM (but not turning it 6367 on by default). All optimizations must be disabled, since 6368 this patch is only the first of a series of patches. 6369 6370 During the work, a lot of x86 specific code revealed and 6371 made platform independent. 6372 See revisions: 50531 50541 50593 50594 50595 6373 6374 * assembler/ARMAssembler.h: 6375 (JSC::ARMAssembler::): 6376 (JSC::ARMAssembler::fdivd_r): 6377 * assembler/MacroAssemblerARM.h: 6378 (JSC::MacroAssemblerARM::lshift32): 6379 (JSC::MacroAssemblerARM::neg32): 6380 (JSC::MacroAssemblerARM::rshift32): 6381 (JSC::MacroAssemblerARM::branchOr32): 6382 (JSC::MacroAssemblerARM::set8): 6383 (JSC::MacroAssemblerARM::setTest8): 6384 (JSC::MacroAssemblerARM::loadDouble): 6385 (JSC::MacroAssemblerARM::divDouble): 6386 (JSC::MacroAssemblerARM::convertInt32ToDouble): 6387 (JSC::MacroAssemblerARM::zeroDouble): 6388 * jit/JIT.cpp: 6389 * jit/JIT.h: 6390 * jit/JITOpcodes.cpp: 6391 (JSC::JIT::privateCompileCTIMachineTrampolines): 6392 * jit/JITStubs.cpp: 6393 * wtf/StdLibExtras.h: 6394 63952009-11-13 Dominik Röttsches <dominik.roettsches@access-company.com> 6396 6397 Reviewed by Eric Seidel. 6398 6399 Unify TextBoundaries implementations by only relying on WTF Unicode abstractions 6400 https://bugs.webkit.org/show_bug.cgi?id=31468 6401 6402 Adding isAlphanumeric abstraction, required 6403 by TextBoundaries.cpp. 6404 6405 * wtf/unicode/glib/UnicodeGLib.h: 6406 (WTF::Unicode::isAlphanumeric): 6407 * wtf/unicode/icu/UnicodeIcu.h: 6408 (WTF::Unicode::isAlphanumeric): 6409 64102009-11-13 Norbert Leser <norbert.leser&nokia.com> 6411 6412 Reviewed by Eric Seidel. 6413 6414 Added macros for USERINCLUDE paths within symbian blocks 6415 to guarantee inclusion of respective header files from local path 6416 first (to avoid clashes with same names of header files in system include path). 6417 6418 * JavaScriptCore.pri: 6419 64202009-11-13 Oliver Hunt <oliver@apple.com> 6421 6422 Reviewed by Geoff Garen. 6423 6424 JSValueProtect and JSValueUnprotect don't protect API wrapper values 6425 https://bugs.webkit.org/show_bug.cgi?id=31485 6426 6427 Make JSValueProtect/Unprotect use a new 'toJS' function, 'toJSForGC' that 6428 does not attempt to to strip out API wrapper objects. 6429 6430 * API/APICast.h: 6431 (toJSForGC): 6432 * API/JSValueRef.cpp: 6433 (JSValueProtect): 6434 (JSValueUnprotect): 6435 * API/tests/testapi.c: 6436 (makeGlobalNumberValue): 6437 (main): 6438 64392009-11-13 İsmail Dönmez <ismail@namtrac.org> 6440 6441 Reviewed by Antti Koivisto. 6442 6443 Fix typo, ce_time.cpp should be ce_time.c 6444 6445 * JavaScriptCore.pri: 6446 64472009-11-12 Steve VanDeBogart <vandebo@chromium.org> 6448 6449 Reviewed by Adam Barth. 6450 6451 Calculate the time offset only if we were able to parse 6452 the date string. This saves an IPC in Chromium for 6453 invalid date strings. 6454 https://bugs.webkit.org/show_bug.cgi?id=31416 6455 6456 * wtf/DateMath.cpp: 6457 (WTF::parseDateFromNullTerminatedCharacters): 6458 (JSC::parseDateFromNullTerminatedCharacters): 6459 64602009-11-12 Oliver Hunt <oliver@apple.com> 6461 6462 Rollout r50896 until i can work out why it causes failures. 6463 6464 * bytecompiler/BytecodeGenerator.cpp: 6465 (JSC::BytecodeGenerator::emitReturn): 6466 * interpreter/Interpreter.cpp: 6467 (JSC::Interpreter::execute): 6468 * parser/Nodes.cpp: 6469 (JSC::EvalNode::emitBytecode): 6470 64712009-11-12 Steve Falkenburg <sfalken@apple.com> 6472 6473 Reviewed by Stephanie Lewis. 6474 6475 Remove LIBRARY directive from def file to fix Debug_All target. 6476 6477 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 6478 64792009-11-12 Gustavo Noronha Silva <gustavo.noronha@collabora.co.uk> 6480 6481 Rubber-stamped by Holger Freyther. 6482 6483 Revert r50204, since it makes DRT crash on 32 bits release builds 6484 for GTK+. 6485 6486 * wtf/FastMalloc.h: 6487 64882009-11-12 Oliver Hunt <oliver@apple.com> 6489 6490 Reviewed by Gavin Barraclough. 6491 6492 Start unifying entry logic for function and eval code. 6493 6494 Eval now uses a ret instruction to end execution, and sets up 6495 a callframe more in line with what we do for function entry. 6496 6497 * bytecompiler/BytecodeGenerator.cpp: 6498 (JSC::BytecodeGenerator::emitReturn): 6499 * interpreter/Interpreter.cpp: 6500 (JSC::Interpreter::execute): 6501 * parser/Nodes.cpp: 6502 (JSC::EvalNode::emitBytecode): 6503 65042009-11-12 Richard Moe Gustavsen <richard.gustavsen@nokia.com> 6505 6506 Reviewed by Kenneth Rohde Christiansen. 6507 6508 [Qt] Disable pthread_setname_np. 6509 6510 This allows Qt builds on Mac from 10.6 to run on earlier version 6511 where this symbol is not present. 6512 https://bugs.webkit.org/show_bug.cgi?id=31403 6513 6514 * wtf/Platform.h: 6515 65162009-11-12 Thiago Macieira <thiago.macieira@nokia.com> 6517 6518 Reviewed by Kenneth Rohde Christiansen. 6519 6520 [Qt] Fix linking on Linux 32-bit. 6521 6522 It was missing the ".text" directive at the top of the file, 6523 indicating that code would follow. Without it, the assembler created 6524 "NOTYPE" symbols, which would result in linker errors. 6525 https://bugs.webkit.org/show_bug.cgi?id=30863 6526 6527 * jit/JITStubs.cpp: 6528 65292009-11-11 Laszlo Gombos <laszlo.1.gombos@nokia.com> 6530 6531 Reviewed by Alexey Proskuryakov. 6532 6533 Refactor multiple JavaScriptCore threads 6534 https://bugs.webkit.org/show_bug.cgi?id=31328 6535 6536 Remove the id field from the PlatformThread structure 6537 as it is not used. 6538 6539 * runtime/Collector.cpp: 6540 (JSC::getCurrentPlatformThread): 6541 (JSC::suspendThread): 6542 (JSC::resumeThread): 6543 (JSC::getPlatformThreadRegisters): 6544 65452009-11-10 Geoffrey Garen <ggaren@apple.com> 6546 6547 Linux build fix: Added an #include for UINT_MAX. 6548 6549 * runtime/WeakRandom.h: 6550 65512009-11-10 Geoffrey Garen <ggaren@apple.com> 6552 6553 JavaScriptGlue build fix: Marked a file 'private' instead of 'project'. 6554 6555 * JavaScriptCore.xcodeproj/project.pbxproj: 6556 65572009-11-10 Geoffrey Garen <ggaren@apple.com> 6558 6559 Reviewed by Gavin "avGni arBalroguch" Barraclough. 6560 6561 Faster Math.random, based on GameRand. 6562 6563 SunSpider says 1.4% faster. 6564 6565 * GNUmakefile.am: 6566 * JavaScriptCore.gypi: 6567 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: 6568 * JavaScriptCore.xcodeproj/project.pbxproj: Added the header to the project. 6569 6570 * runtime/JSGlobalData.cpp: 6571 (JSC::JSGlobalData::JSGlobalData): 6572 * runtime/JSGlobalData.h: Use an object to track random number generation 6573 state, initialized to the current time. 6574 6575 * runtime/MathObject.cpp: 6576 (JSC::MathObject::MathObject): 6577 (JSC::mathProtoFuncRandom): Use the new hotness. 6578 6579 * runtime/WeakRandom.h: Added. 6580 (JSC::WeakRandom::WeakRandom): 6581 (JSC::WeakRandom::get): 6582 (JSC::WeakRandom::advance): The new hotness. 6583 65842009-11-09 Geoffrey Garen <ggaren@apple.com> 6585 6586 Reviewed by Oliver Hunt. 6587 6588 Imported the v8 DST cache. 6589 6590 SunSpider says 1.5% faster. 6591 6592 * runtime/JSGlobalData.cpp: 6593 (JSC::JSGlobalData::resetDateCache): Reset the DST cache when resetting 6594 other date data. 6595 6596 * runtime/JSGlobalData.h: 6597 (JSC::DSTOffsetCache::DSTOffsetCache): 6598 (JSC::DSTOffsetCache::reset): Added a struct for the DST cache. 6599 6600 * wtf/DateMath.cpp: 6601 (WTF::calculateDSTOffsetSimple): 6602 (WTF::calculateDSTOffset): 6603 (WTF::parseDateFromNullTerminatedCharacters): 6604 (JSC::getDSTOffset): 6605 (JSC::gregorianDateTimeToMS): 6606 (JSC::msToGregorianDateTime): 6607 (JSC::parseDateFromNullTerminatedCharacters): 6608 * wtf/DateMath.h: The imported code for probing and updating the cache. 6609 66102009-11-09 Geoffrey Garen <ggaren@apple.com> 6611 6612 Reviewed by Oliver Hunt. 6613 6614 Fixed an edge case that could cause the engine not to notice a timezone 6615 change. 6616 6617 No test because this case would require manual intervention to change 6618 the timezone during the test. 6619 6620 SunSpider reports no change. 6621 6622 * runtime/DateInstanceCache.h: 6623 (JSC::DateInstanceCache::DateInstanceCache): 6624 (JSC::DateInstanceCache::reset): Added a helper function for resetting 6625 this cache. Also, shrank the cache, since we'll be resetting it often. 6626 6627 * runtime/JSGlobalData.cpp: 6628 (JSC::JSGlobalData::resetDateCache): Include resetting the DateInstanceCache 6629 in resetting Date data. (Otherwise, a cache hit could bypass a necessary 6630 timezone update check.) 6631 66322009-11-09 Geoffrey Garen <ggaren@apple.com> 6633 6634 Reviewed by Sam Weinig. 6635 6636 Some manual inlining and constant propogation in Date code. 6637 6638 SunSpider reports a 0.4% speedup on date-*, no overall speedup. Shark 6639 says some previously evident stalls are now gone. 6640 6641 * runtime/DateConstructor.cpp: 6642 (JSC::callDate): 6643 * runtime/DateConversion.cpp: 6644 (JSC::formatTime): 6645 (JSC::formatTimeUTC): Split formatTime into UTC and non-UTC variants. 6646 6647 * runtime/DateConversion.h: 6648 * runtime/DateInstance.cpp: 6649 (JSC::DateInstance::calculateGregorianDateTime): 6650 (JSC::DateInstance::calculateGregorianDateTimeUTC): 6651 * runtime/DateInstance.h: 6652 (JSC::DateInstance::gregorianDateTime): 6653 (JSC::DateInstance::gregorianDateTimeUTC): Split gregorianDateTime into 6654 a UTC and non-UTC variant, and split each variant into a fast inline 6655 case and a slow out-of-line case. 6656 6657 * runtime/DatePrototype.cpp: 6658 (JSC::formatLocaleDate): 6659 (JSC::dateProtoFuncToString): 6660 (JSC::dateProtoFuncToUTCString): 6661 (JSC::dateProtoFuncToISOString): 6662 (JSC::dateProtoFuncToDateString): 6663 (JSC::dateProtoFuncToTimeString): 6664 (JSC::dateProtoFuncGetFullYear): 6665 (JSC::dateProtoFuncGetUTCFullYear): 6666 (JSC::dateProtoFuncToGMTString): 6667 (JSC::dateProtoFuncGetMonth): 6668 (JSC::dateProtoFuncGetUTCMonth): 6669 (JSC::dateProtoFuncGetDate): 6670 (JSC::dateProtoFuncGetUTCDate): 6671 (JSC::dateProtoFuncGetDay): 6672 (JSC::dateProtoFuncGetUTCDay): 6673 (JSC::dateProtoFuncGetHours): 6674 (JSC::dateProtoFuncGetUTCHours): 6675 (JSC::dateProtoFuncGetMinutes): 6676 (JSC::dateProtoFuncGetUTCMinutes): 6677 (JSC::dateProtoFuncGetSeconds): 6678 (JSC::dateProtoFuncGetUTCSeconds): 6679 (JSC::dateProtoFuncGetTimezoneOffset): 6680 (JSC::setNewValueFromTimeArgs): 6681 (JSC::setNewValueFromDateArgs): 6682 (JSC::dateProtoFuncSetYear): 6683 (JSC::dateProtoFuncGetYear): Updated for the gregorianDateTime change above. 6684 66852009-11-09 Geoffrey Garen <ggaren@apple.com> 6686 6687 Build fix: export a new symbol. 6688 6689 * JavaScriptCore.exp: 6690 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 6691 66922009-11-09 Geoffrey Garen <ggaren@apple.com> 6693 6694 Reviewed by Sam "Home Wrecker" Weinig. 6695 6696 Added a tiny cache for Date parsing. 6697 6698 SunSpider says 1.2% faster. 6699 6700 * runtime/DateConversion.cpp: 6701 (JSC::parseDate): Try to reuse the last parsed Date, if present. 6702 6703 * runtime/JSGlobalData.cpp: 6704 (JSC::JSGlobalData::resetDateCache): 6705 * runtime/JSGlobalData.h: Added storage for last parsed Date. Refactored 6706 this code to make resetting the date cache easier. 6707 6708 * runtime/JSGlobalObject.h: 6709 (JSC::DynamicGlobalObjectScope::DynamicGlobalObjectScope): Updated for 6710 refactoring. 6711 6712 * wtf/DateMath.cpp: 6713 (JSC::parseDateFromNullTerminatedCharacters): 6714 * wtf/DateMath.h: Changed ExecState to be first parameter, as is the JSC custom. 6715 67162009-11-09 Oliver Hunt <oliver@apple.com> 6717 6718 Reviewed by Gavin Barraclough. 6719 6720 Can cache prototype lookups on uncacheable dictionaries. 6721 https://bugs.webkit.org/show_bug.cgi?id=31198 6722 6723 Replace fromDictionaryTransition with flattenDictionaryObject and 6724 flattenDictionaryStructure. This change is necessary as we need to 6725 guarantee that our attempt to convert away from a dictionary structure 6726 will definitely succeed, and in some cases this requires mutating the 6727 object storage itself. 6728 6729 * interpreter/Interpreter.cpp: 6730 (JSC::Interpreter::tryCacheGetByID): 6731 * jit/JITStubs.cpp: 6732 (JSC::JITThunks::tryCacheGetByID): 6733 (JSC::DEFINE_STUB_FUNCTION): 6734 * runtime/BatchedTransitionOptimizer.h: 6735 (JSC::BatchedTransitionOptimizer::~BatchedTransitionOptimizer): 6736 * runtime/JSObject.h: 6737 (JSC::JSObject::flattenDictionaryObject): 6738 * runtime/Operations.h: 6739 (JSC::normalizePrototypeChain): 6740 * runtime/Structure.cpp: 6741 (JSC::Structure::flattenDictionaryStructure): 6742 (JSC::comparePropertyMapEntryIndices): 6743 * runtime/Structure.h: 6744 67452009-11-09 Laszlo Gombos <laszlo.1.gombos@nokia.com> 6746 6747 Not reviewed, build fix. 6748 6749 Remove extra character from r50701. 6750 6751 * JavaScriptCore.pri: 6752 67532009-11-09 Laszlo Gombos <laszlo.1.gombos@nokia.com> 6754 6755 Not reviewed, build fix. 6756 6757 Revert r50695 because it broke QtWebKit (clean builds). 6758 6759 * JavaScriptCore.pri: 6760 67612009-11-09 Norbert Leser <norbert.leser@nokia.com> 6762 6763 Reviewed by Kenneth Rohde Christiansen. 6764 6765 Prepended $$PWD to GENERATED_SOURCES_DIR to avoid potential ambiguities when included from WebCore.pro. 6766 Some preprocessors consider this GENERATED_SOURCES_DIR relative to current invoking dir (e.g., ./WebCore), 6767 and not the working dir of JavaCriptCore.pri (i.e., ../JavaScriptCore/). 6768 6769 * JavaScriptCore.pri: 6770 67712009-11-09 Laszlo Gombos <laszlo.1.gombos@nokia.com> 6772 6773 Reviewed by Kenneth Rohde Christiansen. 6774 6775 Use explicit parentheses to silence gcc 4.4 -Wparentheses warnings 6776 https://bugs.webkit.org/show_bug.cgi?id=31040 6777 6778 * interpreter/Interpreter.cpp: 6779 (JSC::Interpreter::privateExecute): 6780 67812009-11-08 David Levin <levin@chromium.org> 6782 6783 Reviewed by NOBODY (speculative snow leopard and windows build fixes). 6784 6785 * wtf/DateMath.cpp: 6786 (WTF::parseDateFromNullTerminatedCharacters): 6787 (JSC::gregorianDateTimeToMS): 6788 (JSC::msToGregorianDateTime): 6789 (JSC::parseDateFromNullTerminatedCharacters): 6790 * wtf/DateMath.h: 6791 (JSC::GregorianDateTime::GregorianDateTime): 6792 67932009-11-08 David Levin <levin@chromium.org> 6794 6795 Reviewed by NOBODY (chromium build fix). 6796 6797 Hopefully, the last build fix. 6798 6799 Create better separation in DateMath about the JSC 6800 and non-JSC portions. Also, only expose the non-JSC 6801 version in the exports. 6802 6803 * JavaScriptCore.exp: 6804 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 6805 * wtf/DateMath.cpp: 6806 (WTF::parseDateFromNullTerminatedCharacters): 6807 (JSC::getUTCOffset): 6808 (JSC::gregorianDateTimeToMS): 6809 (JSC::msToGregorianDateTime): 6810 (JSC::parseDateFromNullTerminatedCharacters): 6811 * wtf/DateMath.h: 6812 (JSC::gmtoffset): 6813 68142009-11-08 David Levin <levin@chromium.org> 6815 6816 Reviewed by NOBODY (chromium build fix). 6817 6818 For the change in DateMath. 6819 6820 * config.h: 6821 * wtf/DateMath.cpp: 6822 68232009-11-06 Geoffrey Garen <ggaren@apple.com> 6824 6825 Windows build fix: export some symbols. 6826 6827 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 6828 68292009-11-06 Geoffrey Garen <ggaren@apple.com> 6830 6831 Build fix: updated export file. 6832 6833 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 6834 68352009-11-06 Geoffrey Garen <ggaren@apple.com> 6836 6837 Build fix: added some #includes. 6838 6839 * wtf/CurrentTime.h: 6840 * wtf/DateMath.h: 6841 68422009-11-06 Geoffrey Garen <ggaren@apple.com> 6843 6844 Reviewed by Oliver Hunt. 6845 6846 https://bugs.webkit.org/show_bug.cgi?id=31197 6847 Implemented a timezone cache not based on Mac OS X's notify_check API. 6848 6849 If the VM calculates the local timezone offset from UTC, it caches the 6850 result until the end of the current VM invocation. (We don't want to cache 6851 forever, because the user's timezone may change over time.) 6852 6853 This removes notify_* overhead on Mac, and, more significantly, removes 6854 OS time and date call overhead on non-Mac platforms. 6855 6856 ~8% speedup on Date microbenchmark on Mac. SunSpider reports maybe a tiny 6857 speedup on Mac. (Speedup on non-Mac platforms should be even more noticeable.) 6858 6859 * JavaScriptCore.exp: 6860 6861 * interpreter/CachedCall.h: 6862 (JSC::CachedCall::CachedCall): 6863 * interpreter/Interpreter.cpp: 6864 (JSC::Interpreter::execute): 6865 * runtime/JSGlobalObject.h: 6866 (JSC::DynamicGlobalObjectScope::DynamicGlobalObjectScope): Made the 6867 DynamicGlobalObjectScope constructor responsible for checking whether a 6868 dynamicGlobalObject has already been set. This eliminated some duplicate 6869 client code, and allowed me to avoid adding even more duplicate client 6870 code. Made DynamicGlobalObjectScope responsible for resetting the 6871 local timezone cache upon first entry to the VM. 6872 6873 * runtime/DateConstructor.cpp: 6874 (JSC::constructDate): 6875 (JSC::callDate): 6876 (JSC::dateParse): 6877 (JSC::dateUTC): 6878 * runtime/DateConversion.cpp: 6879 (JSC::parseDate): 6880 * runtime/DateConversion.h: 6881 * runtime/DateInstance.cpp: 6882 (JSC::DateInstance::gregorianDateTime): 6883 * runtime/DateInstance.h: 6884 * runtime/DateInstanceCache.h: 6885 * runtime/DatePrototype.cpp: 6886 (JSC::setNewValueFromTimeArgs): 6887 (JSC::setNewValueFromDateArgs): 6888 (JSC::dateProtoFuncSetYear): 6889 * runtime/InitializeThreading.cpp: 6890 (JSC::initializeThreadingOnce): 6891 * runtime/JSGlobalData.cpp: 6892 (JSC::JSGlobalData::JSGlobalData): 6893 * runtime/JSGlobalData.h: 6894 * wtf/DateMath.cpp: 6895 (WTF::getCurrentUTCTime): 6896 (WTF::getCurrentUTCTimeWithMicroseconds): 6897 (WTF::getLocalTime): 6898 (JSC::getUTCOffset): Use the new cache. Also, see below. 6899 (JSC::gregorianDateTimeToMS): 6900 (JSC::msToGregorianDateTime): 6901 (JSC::initializeDates): 6902 (JSC::parseDateFromNullTerminatedCharacters): Simplified the way this function 6903 accounts for the local timezone offset, to accomodate our new caching API, 6904 and a (possibly misguided) caller in WebCore. Also, see below. 6905 * wtf/DateMath.h: 6906 (JSC::GregorianDateTime::GregorianDateTime): Moved most of the code in 6907 DateMath.* into the JSC namespace. The code needed to move so it could 6908 naturally interact with ExecState and JSGlobalData to support caching. 6909 Logically, it seemed right to move it, too, since this code is not really 6910 as low-level as the WTF namespace might imply -- it implements a set of 6911 date parsing and conversion quirks that are finely tuned to the JavaScript 6912 language. Also removed the Mac OS X notify_* infrastructure. 6913 6914 * wtf/CurrentTime.h: 6915 (WTF::currentTimeMS): 6916 (WTF::getLocalTime): Moved the rest of the DateMath code here, and renamed 6917 it to make it consistent with WTF's currentTime function. 6918 69192009-11-06 Gabor Loki <loki@inf.u-szeged.hu> 6920 6921 Unreviewed trivial buildfix after r50595. 6922 6923 Rename the remaining rshiftPtr calls to rshift32 6924 6925 * jit/JITArithmetic.cpp: 6926 (JSC::JIT::emit_op_rshift): 6927 * jit/JITInlineMethods.h: 6928 (JSC::JIT::emitFastArithImmToInt): 6929 69302009-11-06 Gavin Barraclough <barraclough@apple.com> 6931 6932 Reviewed by Oliver Hunt. 6933 6934 Tidy up the shift methods on the macro-assembler interface. 6935 6936 Currently behaviour of shifts of a magnitude > 0x1f is undefined. 6937 Instead defined that all shifts are masked to this range. This makes a lot of 6938 practical sense, both since having undefined behaviour is not particularly 6939 desirable, and because this behaviour is commonly required (particularly since 6940 it is required bt ECMA-262 for shifts). 6941 6942 Update the ARM assemblers to provide this behaviour. Remove (now) redundant 6943 masks from JITArithmetic, and remove rshiftPtr (this was used in case that 6944 could be rewritten in a simpler form using rshift32, only optimized JSVALUE32 6945 on x86-64, which uses JSVALUE64!) 6946 6947 * assembler/MacroAssembler.h: 6948 * assembler/MacroAssemblerARM.h: 6949 (JSC::MacroAssemblerARM::lshift32): 6950 (JSC::MacroAssemblerARM::rshift32): 6951 * assembler/MacroAssemblerARMv7.h: 6952 (JSC::MacroAssemblerARMv7::lshift32): 6953 (JSC::MacroAssemblerARMv7::rshift32): 6954 * assembler/MacroAssemblerX86_64.h: 6955 * jit/JITArithmetic.cpp: 6956 (JSC::JIT::emit_op_lshift): 6957 (JSC::JIT::emit_op_rshift): 6958 69592009-11-05 Gavin Barraclough <barraclough@apple.com> 6960 6961 Rubber Stamped by Oliver Hunt. 6962 6963 Remove a magic number (1) from the JIT, instead compute the value with OBJECT_OFFSET. 6964 6965 * jit/JITInlineMethods.h: 6966 (JSC::JIT::emitPutJITStubArg): 6967 (JSC::JIT::emitPutJITStubArgConstant): 6968 (JSC::JIT::emitGetJITStubArg): 6969 (JSC::JIT::emitPutJITStubArgFromVirtualRegister): 6970 * jit/JITStubCall.h: 6971 (JSC::JITStubCall::JITStubCall): 6972 (JSC::JITStubCall::getArgument): 6973 * jit/JITStubs.h: 6974 69752009-11-05 Zoltan Herczeg <zherczeg@inf.u-szeged.hu> 6976 6977 Reviewed by Gavin Barraclough. 6978 6979 https://bugs.webkit.org/show_bug.cgi?id=31159 6980 Fix branchDouble behaviour on ARM THUMB2 JIT. 6981 6982 The x86 branchDouble behaviour is reworked, and all JIT 6983 ports should follow the x86 port. See bug 31104 and 31151 6984 6985 This patch contains a fix for the traditional ARM port 6986 6987 * assembler/ARMAssembler.h: 6988 (JSC::ARMAssembler::): 6989 (JSC::ARMAssembler::fmrs_r): 6990 (JSC::ARMAssembler::ftosid_r): 6991 * assembler/MacroAssemblerARM.h: 6992 (JSC::MacroAssemblerARM::): 6993 (JSC::MacroAssemblerARM::branchDouble): 6994 (JSC::MacroAssemblerARM::branchConvertDoubleToInt32): 6995 69962009-11-05 Chris Jerdonek <chris.jerdonek@gmail.com> 6997 6998 Reviewed by Eric Seidel. 6999 7000 Removed the "this is part of the KDE project" comments from 7001 all *.h, *.cpp, *.idl, and *.pm files. 7002 7003 https://bugs.webkit.org/show_bug.cgi?id=31167 7004 7005 The maintenance and architecture page in the project wiki lists 7006 this as a task. 7007 7008 This change includes no changes or additions to test cases 7009 since the change affects only comments. 7010 7011 * wtf/wince/FastMallocWince.h: 7012 70132009-11-05 Gabor Loki <loki@inf.u-szeged.hu> 7014 7015 Reviewed by Gavin Barraclough. 7016 7017 Use ARMv7 specific encoding for immediate constants on ARMv7 target 7018 https://bugs.webkit.org/show_bug.cgi?id=31060 7019 7020 * assembler/ARMAssembler.cpp: 7021 (JSC::ARMAssembler::getOp2): Use INVALID_IMM 7022 (JSC::ARMAssembler::getImm): Use encodeComplexImm for complex immediate 7023 (JSC::ARMAssembler::moveImm): Ditto. 7024 (JSC::ARMAssembler::encodeComplexImm): Encode a constant by one or two 7025 instructions or a PC relative load. 7026 * assembler/ARMAssembler.h: Use INVALID_IMM if a constant cannot be 7027 encoded as an immediate constant. 7028 (JSC::ARMAssembler::): 7029 (JSC::ARMAssembler::movw_r): 16-bit immediate load 7030 (JSC::ARMAssembler::movt_r): High halfword 16-bit immediate load 7031 (JSC::ARMAssembler::getImm16Op2): Encode immediate constant for 7032 movw_r and mowt_r 7033 70342009-11-04 Mark Mentovai <mark@chromium.org> 7035 7036 Reviewed by Mark Rowe. 7037 7038 Provide TARGETING_TIGER and TARGETING_LEOPARD as analogues to 7039 BUILDING_ON_TIGER and BUILDING_ON_LEOPARD. The TARGETING_ macros 7040 consider the deployment target; the BUILDING_ON_ macros consider the 7041 headers being built against. 7042 7043 * wtf/Platform.h: 7044 70452009-11-04 Gavin Barraclough <barraclough@apple.com> 7046 7047 Reviewed by Oliver Hunt. 7048 7049 https://bugs.webkit.org/show_bug.cgi?id=31151 7050 Fix branchDouble behaviour on ARM THUMB2 JIT. 7051 7052 The ARMv7 JIT is currently using ARMv7Assembler::ConditionEQ to branch 7053 for DoubleEqualOrUnordered, however this is incorrect – ConditionEQ won't 7054 branch on unordered operands. Similarly, DoubleLessThanOrUnordered & 7055 DoubleLessThanOrEqualOrUnordered use ARMv7Assembler::ConditionLO & 7056 ARMv7Assembler::ConditionLS, whereas they should be using 7057 ARMv7Assembler::ConditionLT & ARMv7Assembler::ConditionLE. 7058 7059 Fix these, and fill out the missing DoubleConditions. 7060 7061 * assembler/MacroAssemblerARMv7.h: 7062 (JSC::MacroAssemblerARMv7::): 7063 (JSC::MacroAssemblerARMv7::branchDouble): 7064 70652009-11-04 Gavin Barraclough <barraclough@apple.com> 7066 7067 Rubber Stamped by Oliver Hunt. 7068 7069 Enable native call optimizations on ARMv7. (Existing ARM_TRADITIONAL 7070 implementation was generic, worked perfectly, just needed turning on). 7071 7072 * jit/JITOpcodes.cpp: 7073 * wtf/Platform.h: 7074 70752009-11-04 Gavin Barraclough <barraclough@apple.com> 7076 7077 Rubber Stamped by Mark Rowe, Oliver Hunt, and Sam Weinig. 7078 7079 Add a missing assert to the ARMv7 JIT. 7080 7081 * assembler/ARMv7Assembler.h: 7082 (JSC::ARMThumbImmediate::ARMThumbImmediate): 7083 70842009-11-04 Mark Rowe <mrowe@apple.com> 7085 7086 Rubber-stamped by Oliver Hunt. 7087 7088 Remove bogus op_ prefix on dumped version of three opcodes. 7089 7090 * bytecode/CodeBlock.cpp: 7091 (JSC::CodeBlock::dump): 7092 70932009-11-04 Mark Rowe <mrowe@apple.com> 7094 7095 Reviewed by Sam Weinig. 7096 7097 Fix dumping of constants in bytecode so that they aren't printed as large positive register numbers. 7098 7099 We do this by having the registerName function return information about the constant if the register 7100 number corresponds to a constant. This requires that registerName, and several functions that call it, 7101 be converted to member functions of CodeBlock so that the constant value can be retrieved. The 7102 ExecState also needs to be threaded down through these functions so that it can be passed on to 7103 constantName when needed. 7104 7105 * bytecode/CodeBlock.cpp: 7106 (JSC::constantName): 7107 (JSC::CodeBlock::registerName): 7108 (JSC::CodeBlock::printUnaryOp): 7109 (JSC::CodeBlock::printBinaryOp): 7110 (JSC::CodeBlock::printConditionalJump): 7111 (JSC::CodeBlock::printGetByIdOp): 7112 (JSC::CodeBlock::printPutByIdOp): 7113 (JSC::CodeBlock::dump): 7114 * bytecode/CodeBlock.h: 7115 (JSC::CodeBlock::isConstantRegisterIndex): 7116 71172009-11-04 Pavel Heimlich <tropikhajma@gmail.com> 7118 7119 Reviewed by Alexey Proskuryakov. 7120 7121 https://bugs.webkit.org/show_bug.cgi?id=30647 7122 Solaris build failure due to strnstr. 7123 7124 * wtf/StringExtras.h: Enable strnstr on Solaris, too. 7125 71262009-11-04 Gavin Barraclough <barraclough@apple.com> 7127 7128 Reviewed by Oliver Hunt. 7129 7130 https://bugs.webkit.org/show_bug.cgi?id=31104 7131 Refactor x86-specific behaviour out of the JIT. 7132 7133 - Add explicit double branch conditions for ordered and unordered comparisons (presently the brehaviour is a mix). 7134 - Refactor double to int conversion out into the MacroAssembler. 7135 - Remove broken double to int conversion for !JSVALUE32_64 builds - this code was broken and slowing us down, fixing it showed it not to be an improvement. 7136 - Remove exclusion of double to int conversion from (1 % X) cases in JSVALUE32_64 builds - if this was of benefit this is no longer the case; simplify. 7137 7138 * assembler/MacroAssemblerARM.h: 7139 (JSC::MacroAssemblerARM::): 7140 * assembler/MacroAssemblerARMv7.h: 7141 (JSC::MacroAssemblerARMv7::): 7142 * assembler/MacroAssemblerX86Common.h: 7143 (JSC::MacroAssemblerX86Common::): 7144 (JSC::MacroAssemblerX86Common::convertInt32ToDouble): 7145 (JSC::MacroAssemblerX86Common::branchDouble): 7146 (JSC::MacroAssemblerX86Common::branchConvertDoubleToInt32): 7147 * jit/JITArithmetic.cpp: 7148 (JSC::JIT::emitBinaryDoubleOp): 7149 (JSC::JIT::emit_op_div): 7150 (JSC::JIT::emitSlow_op_jnless): 7151 (JSC::JIT::emitSlow_op_jnlesseq): 7152 * jit/JITOpcodes.cpp: 7153 (JSC::JIT::emit_op_jfalse): 7154 71552009-11-04 Mark Mentovai <mark@chromium.org> 7156 7157 Reviewed by Eric Seidel. 7158 7159 Remove BUILDING_ON_LEOPARD from JavaScriptCore.gyp. This is supposed 7160 to be set as needed only in wtf/Platform.h. 7161 7162 * JavaScriptCore.gyp/JavaScriptCore.gyp: 7163 71642009-11-02 Oliver Hunt <oliver@apple.com> 7165 7166 Reviewed by Gavin Barraclough. 7167 7168 REGRESSION (r48573): JSC may incorrectly cache chain lookups with a dictionary at the head of the chain 7169 https://bugs.webkit.org/show_bug.cgi?id=31045 7170 7171 Add guards to prevent caching of prototype chain lookups with dictionaries at the 7172 head of the chain. Also add a few tighter assertions to cached prototype lookups 7173 to catch this in future. 7174 7175 * interpreter/Interpreter.cpp: 7176 (JSC::Interpreter::tryCacheGetByID): 7177 (JSC::Interpreter::privateExecute): 7178 * jit/JITStubs.cpp: 7179 (JSC::JITThunks::tryCacheGetByID): 7180 71812009-11-02 Laszlo Gombos <laszlo.1.gombos@nokia.com> 7182 7183 Reviewed by Darin Adler. 7184 7185 PLATFORM(CF) should be set when building for Qt on Darwin 7186 https://bugs.webkit.org/show_bug.cgi?id=23671 7187 7188 * wtf/Platform.h: Turn on CF support if both QT and DARWIN 7189 platforms are defined. 7190 71912009-11-02 Dmitry Titov <dimich@chromium.org> 7192 7193 Reviewed by David Levin. 7194 7195 Remove threadsafe refcounting from tasks used with WTF::MessageQueue. 7196 https://bugs.webkit.org/show_bug.cgi?id=30612 7197 7198 * wtf/MessageQueue.h: 7199 (WTF::MessageQueue::alwaysTruePredicate): 7200 (WTF::MessageQueue::~MessageQueue): 7201 (WTF::MessageQueue::append): 7202 (WTF::MessageQueue::appendAndCheckEmpty): 7203 (WTF::MessageQueue::prepend): 7204 (WTF::MessageQueue::waitForMessage): 7205 (WTF::MessageQueue::waitForMessageFilteredWithTimeout): 7206 (WTF::MessageQueue::tryGetMessage): 7207 (WTF::MessageQueue::removeIf): 7208 The MessageQueue is changed to act as a queue of OwnPtr<DataType>. It takes ownership 7209 of posted tasks and passes it to the new owner (in another thread) when the task is fetched. 7210 All methods have arguments of type PassOwnPtr<DataType> and return the same type. 7211 7212 * wtf/Threading.cpp: 7213 (WTF::createThread): 7214 Superficial change to trigger rebuild of JSC project on Windows, 7215 workaround for https://bugs.webkit.org/show_bug.cgi?id=30890 7216 72172009-10-30 Geoffrey Garen <ggaren@apple.com> 7218 7219 Reviewed by Oliver Hunt. 7220 7221 Fixed failing layout test: restore a special case I accidentally deleted. 7222 7223 * runtime/DatePrototype.cpp: 7224 (JSC::setNewValueFromDateArgs): In the case of applying a change to a date 7225 that is NaN, reset the date to 0 *and* then apply the change; don't just 7226 reset the date to 0. 7227 72282009-10-30 Geoffrey Garen <ggaren@apple.com> 7229 7230 Windows build fix: update for object-to-pointer change. 7231 7232 * runtime/DatePrototype.cpp: 7233 (JSC::formatLocaleDate): 7234 72352009-10-29 Geoffrey Garen <ggaren@apple.com> 7236 7237 Reviewed by Darin Adler. 7238 7239 https://bugs.webkit.org/show_bug.cgi?id=30942 7240 Use pointers instead of copies to pass GregorianDateTime objects around. 7241 7242 SunSpider reports a shocking 4.5% speedup on date-format-xparb, and 1.3% 7243 speedup on date-format-tofte. 7244 7245 * runtime/DateInstance.cpp: 7246 (JSC::DateInstance::gregorianDateTime): 7247 * runtime/DateInstance.h: 7248 * runtime/DatePrototype.cpp: 7249 (JSC::formatLocaleDate): 7250 (JSC::dateProtoFuncToString): 7251 (JSC::dateProtoFuncToUTCString): 7252 (JSC::dateProtoFuncToISOString): 7253 (JSC::dateProtoFuncToDateString): 7254 (JSC::dateProtoFuncToTimeString): 7255 (JSC::dateProtoFuncGetFullYear): 7256 (JSC::dateProtoFuncGetUTCFullYear): 7257 (JSC::dateProtoFuncToGMTString): 7258 (JSC::dateProtoFuncGetMonth): 7259 (JSC::dateProtoFuncGetUTCMonth): 7260 (JSC::dateProtoFuncGetDate): 7261 (JSC::dateProtoFuncGetUTCDate): 7262 (JSC::dateProtoFuncGetDay): 7263 (JSC::dateProtoFuncGetUTCDay): 7264 (JSC::dateProtoFuncGetHours): 7265 (JSC::dateProtoFuncGetUTCHours): 7266 (JSC::dateProtoFuncGetMinutes): 7267 (JSC::dateProtoFuncGetUTCMinutes): 7268 (JSC::dateProtoFuncGetSeconds): 7269 (JSC::dateProtoFuncGetUTCSeconds): 7270 (JSC::dateProtoFuncGetTimezoneOffset): 7271 (JSC::setNewValueFromTimeArgs): 7272 (JSC::setNewValueFromDateArgs): 7273 (JSC::dateProtoFuncSetYear): 7274 (JSC::dateProtoFuncGetYear): Renamed getGregorianDateTime to gregorianDateTime, 7275 since it no longer has an out parameter. Uses 0 to indicate invalid dates. 7276 72772009-10-30 Zoltan Horvath <zoltan@webkit.org> 7278 7279 Reviewed by Darin Adler. 7280 7281 Allow custom memory allocation control for JavaScriptCore's ListHashSet 7282 https://bugs.webkit.org/show_bug.cgi?id=30853 7283 7284 Inherits ListHashSet class from FastAllocBase because it is 7285 instantiated by 'new' in WebCore/rendering/RenderBlock.cpp:1813. 7286 7287 * wtf/ListHashSet.h: 7288 72892009-10-30 Oliver Hunt <oliver@apple.com> 7290 7291 Reviewed by Gavin Barraclough. 7292 7293 Regression: crash enumerating properties of an object with getters or setters 7294 https://bugs.webkit.org/show_bug.cgi?id=30948 7295 7296 Add a guard to prevent us trying to cache property enumeration on 7297 objects with getters or setters. 7298 7299 * runtime/JSPropertyNameIterator.cpp: 7300 (JSC::JSPropertyNameIterator::create): 7301 73022009-10-30 Roland Steiner <rolandsteiner@chromium.org> 7303 7304 Reviewed by Eric Seidel. 7305 7306 Remove ENABLE_RUBY guards as discussed with Dave Hyatt and Maciej Stachowiak. 7307 7308 Bug 28420 - Implement HTML5 <ruby> rendering 7309 (https://bugs.webkit.org/show_bug.cgi?id=28420) 7310 7311 No new tests (no functional change). 7312 7313 * Configurations/FeatureDefines.xcconfig: 7314 73152009-10-29 Oliver Hunt <oliver@apple.com> 7316 7317 Reviewed by Maciej Stachowiak. 7318 7319 REGRESSION (r50218-r50262): E*TRADE accounts page is missing content 7320 https://bugs.webkit.org/show_bug.cgi?id=30947 7321 <rdar://problem/7348833> 7322 7323 The logic for flagging that a structure has non-enumerable properties 7324 was in addPropertyWithoutTransition, rather than in the core Structure::put 7325 method. Despite this I was unable to produce a testcase that caused 7326 the failure that etrade was experiencing, but the new assertion in 7327 getEnumerablePropertyNames triggers on numerous layout tests without 7328 the fix, so in effecti all for..in enumeration in any test ends up 7329 doing the required consistency check. 7330 7331 * runtime/Structure.cpp: 7332 (JSC::Structure::addPropertyWithoutTransition): 7333 (JSC::Structure::put): 7334 (JSC::Structure::getEnumerablePropertyNames): 7335 (JSC::Structure::checkConsistency): 7336 73372009-10-29 Gabor Loki <loki@inf.u-szeged.hu> 7338 7339 Reviewed by Gavin Barraclough. 7340 7341 Add cacheFlush support for Thumb-2 on Linux 7342 https://bugs.webkit.org/show_bug.cgi?id=30865 7343 7344 * jit/ExecutableAllocator.h: 7345 (JSC::ExecutableAllocator::cacheFlush): 7346 73472009-10-28 Gavin Barraclough <barraclough@apple.com> 7348 7349 Reviewed by Oliver Hunt. 7350 7351 JSC JIT on ARMv7 cannot link jumps >16Mb range 7352 https://bugs.webkit.org/show_bug.cgi?id=30891 7353 7354 Start planing all relative jumps as move-32-bit-immediate-to-register-BX. 7355 In the cases where the jump would fall within a relative jump range, use a relative jump. 7356 7357 * JavaScriptCore.xcodeproj/project.pbxproj: 7358 * assembler/ARMv7Assembler.h: 7359 (JSC::ARMv7Assembler::~ARMv7Assembler): 7360 (JSC::ARMv7Assembler::LinkRecord::LinkRecord): 7361 (JSC::ARMv7Assembler::): 7362 (JSC::ARMv7Assembler::executableCopy): 7363 (JSC::ARMv7Assembler::linkJump): 7364 (JSC::ARMv7Assembler::relinkJump): 7365 (JSC::ARMv7Assembler::setInt32): 7366 (JSC::ARMv7Assembler::isB): 7367 (JSC::ARMv7Assembler::isBX): 7368 (JSC::ARMv7Assembler::isMOV_imm_T3): 7369 (JSC::ARMv7Assembler::isMOVT): 7370 (JSC::ARMv7Assembler::isNOP_T1): 7371 (JSC::ARMv7Assembler::isNOP_T2): 7372 (JSC::ARMv7Assembler::linkJumpAbsolute): 7373 (JSC::ARMv7Assembler::twoWordOp5i6Imm4Reg4EncodedImmFirst): 7374 (JSC::ARMv7Assembler::twoWordOp5i6Imm4Reg4EncodedImmSecond): 7375 (JSC::ARMv7Assembler::ARMInstructionFormatter::twoWordOp5i6Imm4Reg4EncodedImm): 7376 * assembler/MacroAssemblerARMv7.h: 7377 (JSC::MacroAssemblerARMv7::makeJump): 7378 (JSC::MacroAssemblerARMv7::makeBranch): 7379 * jit/JIT.h: 7380 * wtf/Platform.h: 7381 73822009-10-28 Oliver Hunt <oliver@apple.com> 7383 7384 Reviewed by Geoff Garen. 7385 7386 Improve for..in enumeration performance 7387 https://bugs.webkit.org/show_bug.cgi?id=30887 7388 7389 Improve indexing of an object with a for..in iterator by 7390 identifying cases where get_by_val is being used with a iterator 7391 as the subscript and replace it with a new get_by_pname 7392 bytecode. get_by_pname then optimizes lookups that directly access 7393 the base object. 7394 7395 * bytecode/CodeBlock.cpp: 7396 (JSC::CodeBlock::dump): 7397 * bytecode/Opcode.h: 7398 * bytecompiler/BytecodeGenerator.cpp: 7399 (JSC::BytecodeGenerator::emitGetByVal): 7400 * bytecompiler/BytecodeGenerator.h: 7401 (JSC::BytecodeGenerator::pushOptimisedForIn): 7402 (JSC::BytecodeGenerator::popOptimisedForIn): 7403 * interpreter/Interpreter.cpp: 7404 (JSC::Interpreter::privateExecute): 7405 * jit/JIT.cpp: 7406 (JSC::JIT::privateCompileMainPass): 7407 (JSC::JIT::privateCompileSlowCases): 7408 * jit/JIT.h: 7409 * jit/JITPropertyAccess.cpp: 7410 (JSC::JIT::compileGetDirectOffset): 7411 (JSC::JIT::emit_op_get_by_pname): 7412 (JSC::JIT::emitSlow_op_get_by_pname): 7413 * parser/Nodes.cpp: 7414 (JSC::ForInNode::emitBytecode): 7415 * runtime/JSObject.h: 7416 * runtime/JSPropertyNameIterator.cpp: 7417 (JSC::JSPropertyNameIterator::create): 7418 * runtime/JSPropertyNameIterator.h: 7419 (JSC::JSPropertyNameIterator::getOffset): 7420 (JSC::JSPropertyNameIterator::JSPropertyNameIterator): 7421 * runtime/JSValue.h: 7422 (JSC::JSValue::): 7423 * runtime/Structure.cpp: 7424 (JSC::Structure::addPropertyTransition): 7425 (JSC::Structure::changePrototypeTransition): 7426 (JSC::Structure::despecifyFunctionTransition): 7427 (JSC::Structure::addAnonymousSlotsTransition): 7428 (JSC::Structure::getterSetterTransition): 7429 (JSC::Structure::toDictionaryTransition): 7430 (JSC::Structure::addPropertyWithoutTransition): 7431 Track the existence (or not) of non-enumerable properties. 7432 * runtime/Structure.h: 7433 (JSC::Structure::propertyStorageCapacity): 7434 (JSC::Structure::propertyStorageSize): 7435 (JSC::Structure::hasNonEnumerableProperties): 7436 (JSC::Structure::hasAnonymousSlots): 7437 74382009-10-28 Dmitry Titov <dimich@chromium.org> 7439 7440 Not reviewed, attemp to fix Windows build. 7441 7442 Touch the cpp file to cause recompile. 7443 7444 * wtf/Threading.cpp: 7445 (WTF::threadEntryPoint): 7446 74472009-10-28 Dmitry Titov <dimich@chromium.org> 7448 7449 Reviewed by David Levin. 7450 7451 https://bugs.webkit.org/show_bug.cgi?id=30805 7452 Add MessageQueue::removeIf(Predicate&) to remove certain tasks without pulling them from the queue. 7453 Existing Database tests cover this since Database removes tasks when it is stopped. 7454 7455 * wtf/MessageQueue.h: 7456 (WTF::::removeIf): 7457 74582009-10-28 Afonso R. Costa Jr. <afonso.costa@openbossa.org> 7459 7460 Reviewed by Oliver Hunt. 7461 7462 [Qt] Enable YARR when YARR_JIT is enabled 7463 https://bugs.webkit.org/show_bug.cgi?id=30730 7464 7465 When enabling or disabling JIT using JAVASCRIPTCORE_JIT, the ENABLE_YARR should 7466 be toggled also. 7467 7468 * JavaScriptCore.pri: 7469 74702009-10-24 Martin Robinson <martin.james.robinson@gmail.com> 7471 7472 Reviewed by Oliver Hunt. 7473 7474 Fix strict aliasing warning by switching reinterpret_cast to bitwise_cast. 7475 7476 strict-aliasing warnings in JSFunction.h 7477 https://bugs.webkit.org/show_bug.cgi?id=27869 7478 7479 * runtime/JSFunction.h: 7480 (JSC::JSFunction::nativeFunction): 7481 (JSC::JSFunction::scopeChain): 7482 (JSC::JSFunction::setScopeChain): 7483 (JSC::JSFunction::setNativeFunction): 7484 74852009-10-28 Jan-Arve Sæther <jan-arve.saether@nokia.com> 7486 7487 Reviewed by Tor Arne Vestbø. 7488 7489 Build-fix for 64-bit Windows 7490 7491 * wtf/Platform.h: Make sure to use WTF_USE_JSVALUE64 7492 74932009-10-28 Gavin Barraclough <barraclough@apple.com> 7494 7495 Reviewed by NOBODY (build fix!). 7496 7497 * jit/JIT.h: 7498 74992009-10-26 Holger Hans Peter Freyther <zecke@selfish.org> 7500 7501 Rubber-stamped by Darin Adler. 7502 7503 Export fastMalloc, fastCalloc, fastRealloc and fastFree on GCC/Unix 7504 https://bugs.webkit.org/show_bug.cgi?id=30769 7505 7506 When using -fvisibility=hidden to hide all internal symbols by default 7507 the malloc symbols will be hidden as well. For memory instrumentation 7508 it is needed to provide an instrumented version of these symbols and 7509 override the normal routines and by changing the visibility back to 7510 default this becomes possible. 7511 7512 The only other solution would be to use system malloc instead of the 7513 TCmalloc implementation but this will not allow to analyze memory 7514 behavior with the default allocator. 7515 7516 * wtf/FastMalloc.h: Define WTF_FAST_MALLOC_EXPORT for GCC and !darwin 7517 75182009-10-27 Gavin Barraclough <barraclough@apple.com> 7519 7520 Rubber Stamped by Samuel Q. Weinig. 7521 7522 Make the asserts protecting the offsets in the JIT more descriptive. 7523 7524 * jit/JIT.h: 7525 * jit/JITCall.cpp: 7526 (JSC::JIT::compileOpCall): 7527 * jit/JITPropertyAccess.cpp: 7528 (JSC::JIT::emit_op_method_check): 7529 (JSC::JIT::compileGetByIdHotPath): 7530 (JSC::JIT::compileGetByIdSlowCase): 7531 (JSC::JIT::emit_op_put_by_id): 7532 75332009-10-27 Geoffrey Garen <ggaren@apple.com> 7534 7535 Reviewed by Sam Weinig. 7536 7537 A little bit of refactoring in the date code. 7538 7539 * JavaScriptCore.exp: Don't export this unused symbol. 7540 7541 * runtime/DateConstructor.cpp: 7542 (JSC::constructDate): 7543 7544 * runtime/DateInstance.cpp: 7545 (JSC::DateInstance::DateInstance): 7546 * runtime/DateInstance.h: Removed some unused functions. Changed the default 7547 constructor to ensure that a DateInstance is always initialized. 7548 7549 * runtime/DatePrototype.cpp: 7550 (JSC::DatePrototype::DatePrototype): Pass an initializer to our constructor, 7551 since it now requires one. 7552 7553 * wtf/DateMath.cpp: 7554 (WTF::msToGregorianDateTime): Only compute our offset from UTC if our 7555 output will require it. Otherwise, our offset is 0. 7556 75572009-10-27 Geoffrey Garen <ggaren@apple.com> 7558 7559 Build fix: Mark DateInstaceCache.h private, so other frameworks can see it. 7560 7561 * JavaScriptCore.xcodeproj/project.pbxproj: 7562 75632009-10-27 Geoffrey Garen <ggaren@apple.com> 7564 7565 Build fix: re-readded this file. 7566 7567 * runtime/DateInstanceCache.h: Added. 7568 (JSC::DateInstanceData::create): 7569 (JSC::DateInstanceData::DateInstanceData): 7570 (JSC::DateInstanceCache::DateInstanceCache): 7571 (JSC::DateInstanceCache::add): 7572 (JSC::DateInstanceCache::lookup): 7573 75742009-10-27 Geoffrey Garen <ggaren@apple.com> 7575 7576 Reviewed by Darin Adler and Oliver Hunt. 7577 7578 https://bugs.webkit.org/show_bug.cgi?id=30800 7579 Cache recently computed date data. 7580 7581 SunSpider reports a ~0.5% speedup, mostly from date-format-tofte.js. 7582 7583 * GNUmakefile.am: 7584 * JavaScriptCore.gypi: 7585 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: 7586 * JavaScriptCore.xcodeproj/project.pbxproj: Added new file. 7587 7588 * runtime/DateInstance.cpp: 7589 (JSC::DateInstance::DateInstance): 7590 (JSC::DateInstance::getGregorianDateTime): Use the shared cache. 7591 7592 * runtime/DateInstance.h: Renamed m_cache to m_data, to avoid the confusion 7593 of a "cache cache". 7594 7595 * runtime/DatePrototype.cpp: 7596 (JSC::formatLocaleDate): 7597 (JSC::dateProtoFuncToString): 7598 (JSC::dateProtoFuncToUTCString): 7599 (JSC::dateProtoFuncToISOString): 7600 (JSC::dateProtoFuncToDateString): 7601 (JSC::dateProtoFuncToTimeString): 7602 (JSC::dateProtoFuncGetFullYear): 7603 (JSC::dateProtoFuncGetUTCFullYear): 7604 (JSC::dateProtoFuncToGMTString): 7605 (JSC::dateProtoFuncGetMonth): 7606 (JSC::dateProtoFuncGetUTCMonth): 7607 (JSC::dateProtoFuncGetDate): 7608 (JSC::dateProtoFuncGetUTCDate): 7609 (JSC::dateProtoFuncGetDay): 7610 (JSC::dateProtoFuncGetUTCDay): 7611 (JSC::dateProtoFuncGetHours): 7612 (JSC::dateProtoFuncGetUTCHours): 7613 (JSC::dateProtoFuncGetMinutes): 7614 (JSC::dateProtoFuncGetUTCMinutes): 7615 (JSC::dateProtoFuncGetSeconds): 7616 (JSC::dateProtoFuncGetUTCSeconds): 7617 (JSC::dateProtoFuncGetTimezoneOffset): 7618 (JSC::setNewValueFromTimeArgs): 7619 (JSC::setNewValueFromDateArgs): 7620 (JSC::dateProtoFuncSetYear): 7621 (JSC::dateProtoFuncGetYear): Pass an ExecState to these functions, so they 7622 can access the DateInstanceCache. 7623 7624 * runtime/JSGlobalData.h: Keep a DateInstanceCache. 7625 76262009-10-27 James Robinson <jamesr@chromium.org> 7627 7628 Reviewed by Darin Fisher. 7629 7630 Ensures that JavaScriptCore/wtf/CurrentTime.cpp is not built in PLATFORM(CHROMIUM) builds. 7631 7632 Chromium uses a different method to calculate the current time than is used in 7633 JavaScriptCore/wtf/CurrentTime.cpp. This can lead to time skew when calls to currentTime() and Chromium's time 7634 function are mixed. In particular, timers can get scheduled in the past which leads to 100% CPU use. 7635 See http://code.google.com/p/chromium/issues/detail?id=25892 for an example. 7636 7637 https://bugs.webkit.org/show_bug.cgi?id=30833 7638 7639 * JavaScriptCore.gyp/JavaScriptCore.gyp: 7640 * wtf/CurrentTime.cpp: 7641 76422009-10-27 Peter Varga <pvarga@inf.u-szeged.hu> 7643 7644 Rubber-stamped by Tor Arne Vestbø. 7645 7646 Fix typo in RegexInterpreter.cpp and RegexJIT.cpp alterantive to 7647 alternative. 7648 7649 * yarr/RegexInterpreter.cpp: 7650 (JSC::Yarr::ByteCompiler::alternativeBodyDisjunction): 7651 (JSC::Yarr::ByteCompiler::alternativeDisjunction): 7652 (JSC::Yarr::ByteCompiler::emitDisjunction): 7653 * yarr/RegexJIT.cpp: 7654 (JSC::Yarr::RegexGenerator::generateDisjunction): 7655 76562009-10-26 Laszlo Gombos <laszlo.1.gombos@nokia.com> 7657 7658 Reviewed by Darin Adler. 7659 7660 Make .rc files compile on Windows without depending on MFC headers 7661 https://bugs.webkit.org/show_bug.cgi?id=30750 7662 7663 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.rc: Use 7664 winresrc.h because it exists even when MFC is not installed, and is 7665 all that's needed here. 7666 76672009-10-26 Gabor Loki <loki@inf.u-szeged.hu> 7668 7669 Reviewed by Gavin Barraclough. 7670 7671 The thunkReturnAddress is on JITStackFrame on ARM JIT as well 7672 https://bugs.webkit.org/show_bug.cgi?id=30782 7673 7674 Move the thunkReturnAddress from top of the stack into the JITStackFrame 7675 structure. This is a requirement for JSValue32_64 support on ARM. 7676 7677 * assembler/MacroAssemblerARM.h: 7678 (JSC::MacroAssemblerARM::ret): Return with link register 7679 (JSC::MacroAssemblerARM::prepareCall): Store the return address in link register 7680 * jit/JIT.h: Remove unused ctiReturnRegister 7681 * jit/JITInlineMethods.h: Same as ARMv7 7682 (JSC::JIT::restoreArgumentReference): Ditto. 7683 (JSC::JIT::restoreArgumentReferenceForTrampoline): Ditto. 7684 * jit/JITOpcodes.cpp: Remove ctiReturnRegister related instruction 7685 * jit/JITStubs.cpp: Store thunkReturnAddress on JITStackFrame. Use 7686 small trampoline functions which handle return addresses for each 7687 CTI_STUB_FUNCTION. 7688 * jit/JITStubs.h: Store thunkReturnAddress on JITStackFrame 7689 (JSC::JITStackFrame::returnAddressSlot): Return with the address of thunkReturnAddress 7690 * yarr/RegexJIT.cpp: 7691 (JSC::Yarr::RegexGenerator::generateEnter): Remove the unnecessary instruction 7692 76932009-10-26 Steve Block <steveblock@google.com> 7694 7695 Reviewed by Darin Adler. 7696 7697 Adds ability to disable ReadWriteLock on platforms (eg Android) that use pthreads but do not support pthread_rwlock. 7698 https://bugs.webkit.org/show_bug.cgi?id=30713 7699 7700 * wtf/Platform.h: Modified. Defines HAVE_PTHREAD_RWLOCK for all platforms currently using pthreads. 7701 * wtf/Threading.h: Modified. Use pthread_rwlock_t only when HAVE_PTHREAD_RWLOCK is defined. 7702 * wtf/ThreadingPthreads.cpp: Modified. Build ReadWriteLock methods only when HAVE_PTHREAD_RWLOCK is defined. 7703 77042009-10-24 Laszlo Gombos <laszlo.1.gombos@nokia.com> 7705 7706 Reviewed by Holger Freyther. 7707 7708 [Qt] [Symbian] Set the capability and memory required to run QtWebKit for Symbian 7709 https://bugs.webkit.org/show_bug.cgi?id=30476 7710 7711 Assign ReadUserData WriteUserData NetworkServices Symbian capabilities 7712 to jsc.exe. 7713 7714 * jsc.pro: 7715 77162009-10-23 Steve Block <steveblock@google.com> 7717 7718 Reviewed by Dmitry Titov. 7719 7720 Fixes a leak in createThreadInternal on Android. 7721 https://bugs.webkit.org/show_bug.cgi?id=30698 7722 7723 * wtf/ThreadingPthreads.cpp: Modified. 7724 (WTF::createThreadInternal): Avoid leaking a ThreadData object on failure. 7725 77262009-10-22 Geoffrey Garen <ggaren@apple.com> 7727 7728 Reviewed by Alexey Proskuryakov. 7729 7730 Fixed ASSERT when opening Safari's Caches window while the Web Inspector 7731 is open. 7732 7733 * runtime/Collector.cpp: 7734 (JSC::typeName): Added two new types to the type name list in the Collector. 7735 These types have been around for a while, but nobody remembered to consider them here. 7736 7737 * runtime/JSCell.h: 7738 (JSC::JSCell::isPropertyNameIterator): 7739 * runtime/JSPropertyNameIterator.h: 7740 (JSC::JSPropertyNameIterator::isPropertyNameIterator): Give the Collector 7741 a way to tell if a cell is a JSPropertyNameIterator. 7742 77432009-10-22 Steve Falkenburg <sfalken@apple.com> 7744 7745 Reviewed by Jon Honeycutt. 7746 7747 https://bugs.webkit.org/show_bug.cgi?id=30686 7748 Remove debug-specific def file. 7749 Only Debug_All target uses JavaScriptCore_debug.dll naming, and since 7750 that target is only used internally, maintaining two files just to 7751 suppress a single link warning isn't worthwhile. 7752 7753 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCoreCommon.vsprops: 7754 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: Removed. 7755 77562009-10-21 Jon Honeycutt <jhoneycutt@apple.com> 7757 7758 <rdar://problem/7270320> Screenshots of off-screen plug-ins are blank 7759 <rdar://problem/7270314> After halting a transparent PluginView on 7760 Windows, the transparency is applied twice 7761 7762 Reviewed by Dan Bernstein. 7763 7764 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 7765 Export WTF::deleteOwnedPtr(HDC). 7766 7767 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 7768 Ditto. 7769 77702009-10-20 Geoffrey Garen <ggaren@apple.com> 7771 7772 Windows build fix: updated variable name. 7773 7774 * runtime/DatePrototype.cpp: 7775 (JSC::formatLocaleDate): 7776 77772009-10-20 Geoffrey Garen <ggaren@apple.com> 7778 7779 Reviewed by Mark Rowe. 7780 7781 * jit/JITOpcodes.cpp: 7782 (JSC::JIT::emit_op_next_pname): Slightly tweaked this #ifdef to match the 7783 size of a JSValue because m_jsStrings is an array of JSValues. 7784 77852009-10-20 Geoffrey Garen <ggaren@apple.com> 7786 7787 Reviewed by Mark Rowe. 7788 7789 Fixed a 64-bit regression caused by the fix for 7790 https://bugs.webkit.org/show_bug.cgi?id=30570. 7791 7792 * jit/JITOpcodes.cpp: 7793 (JSC::JIT::emit_op_next_pname): Use TimesEight stepping on 64-bit, since 7794 64-bit pointers are eight bytes long. 7795 77962009-10-20 Geoffrey Garen <ggaren@apple.com> 7797 7798 Reviewed by Sam Weinig. 7799 7800 Refactored DateInstance::msToGregorianDateTime so that a DateInstance's 7801 caller doesn't need to supply the DateInstance's own internal value to 7802 the DateInstance. 7803 7804 * runtime/DateInstance.cpp: 7805 (JSC::DateInstance::getGregorianDateTime): Renamed from "msToGregorianDateTime". 7806 7807 * runtime/DateInstance.h: 7808 * runtime/DatePrototype.cpp: 7809 (JSC::formatLocaleDate): 7810 (JSC::dateProtoFuncToString): 7811 (JSC::dateProtoFuncToUTCString): 7812 (JSC::dateProtoFuncToISOString): 7813 (JSC::dateProtoFuncToDateString): 7814 (JSC::dateProtoFuncToTimeString): 7815 (JSC::dateProtoFuncToLocaleString): 7816 (JSC::dateProtoFuncToLocaleDateString): 7817 (JSC::dateProtoFuncToLocaleTimeString): 7818 (JSC::dateProtoFuncGetTime): 7819 (JSC::dateProtoFuncGetFullYear): 7820 (JSC::dateProtoFuncGetUTCFullYear): 7821 (JSC::dateProtoFuncToGMTString): 7822 (JSC::dateProtoFuncGetMonth): 7823 (JSC::dateProtoFuncGetUTCMonth): 7824 (JSC::dateProtoFuncGetDate): 7825 (JSC::dateProtoFuncGetUTCDate): 7826 (JSC::dateProtoFuncGetDay): 7827 (JSC::dateProtoFuncGetUTCDay): 7828 (JSC::dateProtoFuncGetHours): 7829 (JSC::dateProtoFuncGetUTCHours): 7830 (JSC::dateProtoFuncGetMinutes): 7831 (JSC::dateProtoFuncGetUTCMinutes): 7832 (JSC::dateProtoFuncGetSeconds): 7833 (JSC::dateProtoFuncGetUTCSeconds): 7834 (JSC::dateProtoFuncGetTimezoneOffset): 7835 (JSC::setNewValueFromTimeArgs): 7836 (JSC::setNewValueFromDateArgs): 7837 (JSC::dateProtoFuncSetYear): 7838 (JSC::dateProtoFuncGetYear): Also renamed "utc" to "outputIsUTC", for clarity. 7839 78402009-10-20 Gabor Loki <loki@inf.u-szeged.hu> 7841 7842 Reviewed by Geoffrey Garen. 7843 7844 The op_next_pname should use 4 bytes addressing mode in case of JSValue32 7845 https://bugs.webkit.org/show_bug.cgi?id=30570 7846 7847 * jit/JITOpcodes.cpp: 7848 (JSC::JIT::emit_op_next_pname): 7849 78502009-10-20 Gabor Loki <loki@inf.u-szeged.hu> 7851 7852 Reviewed by Oliver Hunt. 7853 7854 Move OverridesMarkChildren flag from DatePrototype to its parent class 7855 https://bugs.webkit.org/show_bug.cgi?id=30372 7856 7857 * runtime/DateInstance.h: 7858 (JSC::DateInstance::createStructure): 7859 * runtime/DatePrototype.h: 7860 78612009-10-19 Geoffrey Garen <ggaren@apple.com> 7862 7863 Reviewed by Oliver Hunt. 7864 7865 Tightened up some put_by_id_transition code generation. 7866 https://bugs.webkit.org/show_bug.cgi?id=30539 7867 7868 * jit/JIT.h: 7869 * jit/JITPropertyAccess.cpp: 7870 (JSC::JIT::testPrototype): 7871 (JSC::JIT::privateCompilePutByIdTransition): No need to do object type 7872 checks or read Structures and prototypes from objects: they're all known 7873 constants at compile time. 7874 78752009-10-19 Geoffrey Garen <ggaren@apple.com> 7876 7877 Reviewed by Sam Weinig. 7878 7879 Added a private API for getting a global context from a context, for 7880 clients who want to preserve a context for a later callback. 7881 7882 * API/APICast.h: 7883 (toGlobalRef): Added an ASSERT, since this function is used more often 7884 than before. 7885 7886 * API/JSContextRef.cpp: 7887 * API/JSContextRefPrivate.h: Added. The new API. 7888 7889 * API/tests/testapi.c: 7890 (print_callAsFunction): 7891 (main): Test the new API. 7892 7893 * JavaScriptCore.exp: 7894 * JavaScriptCore.xcodeproj/project.pbxproj: Build and export the new API. 7895 78962009-10-17 Geoffrey Garen <ggaren@apple.com> 7897 7898 Reviewed by Oliver Hunt. 7899 7900 Tightened up some instanceof code generation. 7901 https://bugs.webkit.org/show_bug.cgi?id=30488 7902 7903 * jit/JITOpcodes.cpp: 7904 (JSC::JIT::emit_op_instanceof): 7905 (JSC::JIT::emitSlow_op_instanceof): No need to do object type checks - 7906 cell type checks and ImplementsDefaultHasIntance checks implicitly 7907 supersede object type checks. 7908 79092009-10-18 Kwang Yul Seo <skyul@company100.net> 7910 7911 Reviewed by Darin Adler. 7912 7913 Use _stricmp and _strnicmp instead of deprecated stricmp and strnicmp. 7914 https://bugs.webkit.org/show_bug.cgi?id=30474 7915 7916 stricmp and strnicmp are deprecated beginning in Visual 7917 C++ 2005. Use _stricmp and _strnicmp instead in StringExtras.h. 7918 7919 * wtf/StringExtras.h: 7920 (strncasecmp): 7921 (strcasecmp): 7922 79232009-10-16 Geoffrey Garen <ggaren@apple.com> 7924 7925 Build fix: apparently we shouldn't export those symbols? 7926 7927 * JavaScriptCore.exp: 7928 79292009-10-16 Geoffrey Garen <ggaren@apple.com> 7930 7931 Build fix: export some symbols. 7932 7933 * JavaScriptCore.exp: 7934 79352009-10-16 Oliver Hunt <oliver@apple.com> 7936 7937 Reviewed by Gavin Barraclough. 7938 7939 structure typeinfo flags should be inherited. 7940 https://bugs.webkit.org/show_bug.cgi?id=30468 7941 7942 Add StructureFlag constant to the various JSC classes and use 7943 it for the TypeInfo construction. This allows us to simply 7944 accumulate flags by basing each classes StructureInfo on its parents. 7945 7946 * API/JSCallbackConstructor.h: 7947 (JSC::JSCallbackConstructor::createStructure): 7948 * API/JSCallbackFunction.h: 7949 (JSC::JSCallbackFunction::createStructure): 7950 * API/JSCallbackObject.h: 7951 (JSC::JSCallbackObject::createStructure): 7952 * debugger/DebuggerActivation.h: 7953 (JSC::DebuggerActivation::createStructure): 7954 * runtime/Arguments.h: 7955 (JSC::Arguments::createStructure): 7956 * runtime/BooleanObject.h: 7957 (JSC::BooleanObject::createStructure): 7958 * runtime/DatePrototype.h: 7959 (JSC::DatePrototype::createStructure): 7960 * runtime/FunctionPrototype.h: 7961 (JSC::FunctionPrototype::createStructure): 7962 * runtime/GlobalEvalFunction.h: 7963 (JSC::GlobalEvalFunction::createStructure): 7964 * runtime/InternalFunction.h: 7965 (JSC::InternalFunction::createStructure): 7966 * runtime/JSActivation.h: 7967 (JSC::JSActivation::createStructure): 7968 * runtime/JSArray.h: 7969 (JSC::JSArray::createStructure): 7970 * runtime/JSByteArray.cpp: 7971 (JSC::JSByteArray::createStructure): 7972 * runtime/JSByteArray.h: 7973 * runtime/JSFunction.h: 7974 (JSC::JSFunction::createStructure): 7975 * runtime/JSGlobalObject.h: 7976 (JSC::JSGlobalObject::createStructure): 7977 * runtime/JSNotAnObject.h: 7978 (JSC::JSNotAnObject::createStructure): 7979 * runtime/JSONObject.h: 7980 (JSC::JSONObject::createStructure): 7981 * runtime/JSObject.h: 7982 (JSC::JSObject::createStructure): 7983 * runtime/JSStaticScopeObject.h: 7984 (JSC::JSStaticScopeObject::createStructure): 7985 * runtime/JSVariableObject.h: 7986 (JSC::JSVariableObject::createStructure): 7987 * runtime/JSWrapperObject.h: 7988 (JSC::JSWrapperObject::createStructure): 7989 * runtime/MathObject.h: 7990 (JSC::MathObject::createStructure): 7991 * runtime/NumberConstructor.h: 7992 (JSC::NumberConstructor::createStructure): 7993 * runtime/NumberObject.h: 7994 (JSC::NumberObject::createStructure): 7995 * runtime/RegExpConstructor.h: 7996 (JSC::RegExpConstructor::createStructure): 7997 * runtime/RegExpObject.h: 7998 (JSC::RegExpObject::createStructure): 7999 * runtime/StringObject.h: 8000 (JSC::StringObject::createStructure): 8001 * runtime/StringObjectThatMasqueradesAsUndefined.h: 8002 (JSC::StringObjectThatMasqueradesAsUndefined::createStructure): 8003 80042009-10-16 Geoffrey Garen <ggaren@apple.com> 8005 8006 Reviewed by Oliver Hunt. 8007 8008 Fast for-in enumeration: Cache JSPropertyNameIterator; cache JSStrings 8009 in JSPropertyNameIterator; inline more code. 8010 8011 1.024x as fast on SunSpider (fasta: 1.43x as fast). 8012 8013 * bytecode/CodeBlock.cpp: 8014 (JSC::CodeBlock::dump): 8015 * bytecode/Opcode.h: 8016 * bytecompiler/BytecodeGenerator.cpp: 8017 (JSC::BytecodeGenerator::emitGetPropertyNames): 8018 (JSC::BytecodeGenerator::emitNextPropertyName): 8019 * bytecompiler/BytecodeGenerator.h: Added a few extra operands to 8020 op_get_pnames and op_next_pname so that we can track iteration state 8021 in the register file instead of in the JSPropertyNameIterator. (To be 8022 cacheable, the JSPropertyNameIterator must be stateless.) 8023 8024 * interpreter/Interpreter.cpp: 8025 (JSC::Interpreter::tryCachePutByID): 8026 (JSC::Interpreter::tryCacheGetByID): Updated for rename to 8027 "normalizePrototypeChain" and removal of "isCacheable". 8028 8029 (JSC::Interpreter::privateExecute): Updated for in-RegisterFile 8030 iteration state tracking. 8031 8032 * jit/JIT.cpp: 8033 (JSC::JIT::privateCompileMainPass): 8034 * jit/JIT.h: 8035 * jit/JITOpcodes.cpp: 8036 (JSC::JIT::emit_op_get_pnames): Updated for in-RegisterFile 8037 iteration state tracking. 8038 8039 (JSC::JIT::emit_op_next_pname): Inlined code generation for op_next_pname. 8040 8041 * jit/JITStubs.cpp: 8042 (JSC::JITThunks::tryCachePutByID): 8043 (JSC::JITThunks::tryCacheGetByID): Updated for rename to 8044 "normalizePrototypeChain" and removal of "isCacheable". 8045 8046 (JSC::DEFINE_STUB_FUNCTION): 8047 * jit/JITStubs.h: 8048 (JSC::): Added has_property and to_object stubs. Removed op_next_pname 8049 stub, since has_property is all we need anymore. 8050 8051 * parser/Nodes.cpp: 8052 (JSC::ForInNode::emitBytecode): Updated for in-RegisterFile 8053 iteration state tracking. 8054 8055 * runtime/JSCell.h: 8056 * runtime/JSObject.cpp: 8057 (JSC::JSObject::getPropertyNames): Don't do caching at this layer 8058 anymore, since we don't create a JSPropertyNameIterator at this layer. 8059 8060 * runtime/JSPropertyNameIterator.cpp: 8061 (JSC::JSPropertyNameIterator::create): Do do caching at this layer. 8062 (JSC::JSPropertyNameIterator::get): Updated for in-RegisterFile 8063 iteration state tracking. 8064 (JSC::JSPropertyNameIterator::markChildren): Mark our JSStrings. 8065 8066 * runtime/JSPropertyNameIterator.h: 8067 (JSC::JSPropertyNameIterator::size): 8068 (JSC::JSPropertyNameIterator::setCachedStructure): 8069 (JSC::JSPropertyNameIterator::cachedStructure): 8070 (JSC::JSPropertyNameIterator::setCachedPrototypeChain): 8071 (JSC::JSPropertyNameIterator::cachedPrototypeChain): 8072 (JSC::JSPropertyNameIterator::JSPropertyNameIterator): 8073 (JSC::Structure::setEnumerationCache): Don't store iteration state in 8074 a JSPropertyNameIterator. Do cache a JSPropertyNameIterator in a 8075 Structure. 8076 8077 * runtime/JSValue.h: 8078 (JSC::asCell): 8079 * runtime/MarkStack.h: Make those mischievous #include gods happy. 8080 8081 * runtime/ObjectConstructor.cpp: 8082 8083 * runtime/Operations.h: 8084 (JSC::normalizePrototypeChain): Renamed countPrototypeChainEntriesAndCheckForProxies 8085 to normalizePrototypeChain, since it changes dictionary prototypes to 8086 non-dictionary objects. 8087 8088 * runtime/PropertyNameArray.cpp: 8089 (JSC::PropertyNameArray::add): 8090 * runtime/PropertyNameArray.h: 8091 (JSC::PropertyNameArrayData::PropertyNameArrayData): 8092 (JSC::PropertyNameArray::data): 8093 (JSC::PropertyNameArray::size): 8094 (JSC::PropertyNameArray::begin): 8095 (JSC::PropertyNameArray::end): Simplified some code here to help with 8096 current and future refactoring. 8097 8098 * runtime/Protect.h: 8099 * runtime/Structure.cpp: 8100 (JSC::Structure::~Structure): 8101 (JSC::Structure::addPropertyWithoutTransition): 8102 (JSC::Structure::removePropertyWithoutTransition): No need to clear 8103 the enumeration cache with adding / removing properties without 8104 transition. It is an error to add / remove properties without transition 8105 once an object has been observed, and we can ASSERT to catch that. 8106 8107 * runtime/Structure.h: 8108 (JSC::Structure::enumerationCache): Changed the enumeration cache to 8109 hold a JSPropertyNameIterator. 8110 8111 * runtime/StructureChain.cpp: 8112 * runtime/StructureChain.h: 8113 (JSC::StructureChain::head): Removed StructureChain::isCacheable because 8114 it was wrong-headed in two ways: (1) It gave up when a prototype was a 8115 dictionary, but instead we want un-dictionary heavily accessed 8116 prototypes; (2) It folded a test for hasDefaultGetPropertyNames() into 8117 a generic test for "cacheable-ness", but hasDefaultGetPropertyNames() 8118 is only releavant to for-in caching. 8119 81202009-10-16 Steve Falkenburg <sfalken@apple.com> 8121 8122 Reviewed by Adam Roben. 8123 8124 Add a Debug_All configuration to build entire stack as debug. 8125 Change Debug_Internal to: 8126 - stop using _debug suffix for all WebKit/Safari binaries 8127 - not use _debug as a DLL naming suffix 8128 - use non-debug C runtime lib. 8129 8130 * JavaScriptCore.vcproj/JavaScriptCore.make: Debug build in makefile should build Debug_All. 8131 * JavaScriptCore.vcproj/JavaScriptCore.sln: Add Debug_All configuration. 8132 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: Add Debug_All configuration. 8133 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCoreGenerated.vcproj: Renamed single configuration from "Release" to "all". 8134 * JavaScriptCore.vcproj/JavaScriptCoreSubmit.sln: Add Debug_All configuration. 8135 * JavaScriptCore.vcproj/WTF/WTF.vcproj: Add Debug_All configuration. 8136 * JavaScriptCore.vcproj/jsc/jsc.vcproj: Add Debug_All configuration. 8137 * JavaScriptCore.vcproj/testapi/testapi.vcproj: Add Debug_All configuration. 8138 81392009-10-16 Oliver Hunt <oliver@apple.com> 8140 8141 Reviewed by Gavin Barraclough. 8142 8143 Make typeinfo flags default to false 8144 https://bugs.webkit.org/show_bug.cgi?id=30372 8145 8146 Last part -- replace HasDefaultGetPropertyNames with OverridesGetPropertyNames 8147 flag. 8148 8149 * API/JSCallbackConstructor.h: 8150 (JSC::JSCallbackConstructor::createStructure): 8151 * API/JSCallbackObject.h: 8152 (JSC::JSCallbackObject::createStructure): 8153 * debugger/DebuggerActivation.h: 8154 (JSC::DebuggerActivation::createStructure): 8155 * runtime/Arguments.h: 8156 (JSC::Arguments::createStructure): 8157 * runtime/BooleanObject.h: 8158 (JSC::BooleanObject::createStructure): 8159 * runtime/DatePrototype.h: 8160 (JSC::DatePrototype::createStructure): 8161 * runtime/FunctionPrototype.h: 8162 (JSC::FunctionPrototype::createStructure): 8163 * runtime/GlobalEvalFunction.h: 8164 (JSC::GlobalEvalFunction::createStructure): 8165 * runtime/JSAPIValueWrapper.h: 8166 (JSC::JSAPIValueWrapper::createStructure): 8167 * runtime/JSActivation.h: 8168 (JSC::JSActivation::createStructure): 8169 * runtime/JSArray.h: 8170 (JSC::JSArray::createStructure): 8171 * runtime/JSByteArray.cpp: 8172 (JSC::JSByteArray::createStructure): 8173 * runtime/JSFunction.h: 8174 (JSC::JSFunction::createStructure): 8175 * runtime/JSGlobalObject.h: 8176 (JSC::JSGlobalObject::createStructure): 8177 * runtime/JSNotAnObject.h: 8178 (JSC::JSNotAnObject::createStructure): 8179 * runtime/JSONObject.h: 8180 (JSC::JSONObject::createStructure): 8181 * runtime/JSObject.cpp: 8182 (JSC::JSObject::getPropertyNames): 8183 * runtime/JSObject.h: 8184 (JSC::JSObject::createStructure): 8185 * runtime/JSStaticScopeObject.h: 8186 (JSC::JSStaticScopeObject::createStructure): 8187 * runtime/JSTypeInfo.h: 8188 (JSC::TypeInfo::overridesGetPropertyNames): 8189 * runtime/JSVariableObject.h: 8190 (JSC::JSVariableObject::createStructure): 8191 * runtime/JSWrapperObject.h: 8192 (JSC::JSWrapperObject::createStructure): 8193 * runtime/MathObject.h: 8194 (JSC::MathObject::createStructure): 8195 * runtime/NumberConstructor.h: 8196 (JSC::NumberConstructor::createStructure): 8197 * runtime/NumberObject.h: 8198 (JSC::NumberObject::createStructure): 8199 * runtime/RegExpConstructor.h: 8200 (JSC::RegExpConstructor::createStructure): 8201 * runtime/RegExpObject.h: 8202 (JSC::RegExpObject::createStructure): 8203 * runtime/StringObject.h: 8204 (JSC::StringObject::createStructure): 8205 * runtime/StringObjectThatMasqueradesAsUndefined.h: 8206 (JSC::StringObjectThatMasqueradesAsUndefined::createStructure): 8207 * runtime/StructureChain.cpp: 8208 (JSC::StructureChain::isCacheable): 8209 82102009-10-16 Kevin Ollivier <kevino@theolliviers.com> 8211 8212 wxMSW build fix, we can't use the simple hash there because the PlatformModuleVersion 8213 structure differs. 8214 8215 * wtf/Platform.h: 8216 82172009-10-16 Laszlo Gombos <laszlo.1.gombos@nokia.com> 8218 8219 Reviewed by Simon Hausmann. 8220 8221 [Qt] Implement ExecutableAllocator for Symbian 8222 https://bugs.webkit.org/show_bug.cgi?id=29946 8223 8224 Tested with YARR JIT enabled for Symbian; 8225 This patch does not (yet) enable YARR JIT by default. 8226 8227 * JavaScriptCore.pri: 8228 * jit/ExecutableAllocator.h: 8229 * jit/ExecutableAllocatorSymbian.cpp: Added. 8230 (JSC::ExecutableAllocator::intializePageSize): 8231 (JSC::ExecutablePool::systemAlloc): 8232 (JSC::ExecutablePool::systemRelease): 8233 82342009-10-15 Oliver Hunt <oliver@apple.com> 8235 8236 Reviewed by Darin Adler. 8237 8238 Make typeinfo flags default to false 8239 https://bugs.webkit.org/show_bug.cgi?id=30372 8240 8241 Part 2 -- Reverse the TypeInfo HasDefaultMark flag to OverridesMarkChildren, etc 8242 8243 * API/JSCallbackConstructor.h: 8244 (JSC::JSCallbackConstructor::createStructure): 8245 * API/JSCallbackFunction.h: 8246 (JSC::JSCallbackFunction::createStructure): 8247 * API/JSCallbackObject.h: 8248 (JSC::JSCallbackObject::createStructure): 8249 * debugger/DebuggerActivation.h: 8250 (JSC::DebuggerActivation::createStructure): 8251 * runtime/Arguments.h: 8252 (JSC::Arguments::createStructure): 8253 * runtime/BooleanObject.h: 8254 (JSC::BooleanObject::createStructure): 8255 * runtime/DatePrototype.h: 8256 (JSC::DatePrototype::createStructure): 8257 * runtime/FunctionPrototype.h: 8258 (JSC::FunctionPrototype::createStructure): 8259 * runtime/GetterSetter.h: 8260 (JSC::GetterSetter::createStructure): 8261 * runtime/GlobalEvalFunction.h: 8262 (JSC::GlobalEvalFunction::createStructure): 8263 * runtime/InternalFunction.h: 8264 (JSC::InternalFunction::createStructure): 8265 * runtime/JSAPIValueWrapper.h: 8266 (JSC::JSAPIValueWrapper::createStructure): 8267 * runtime/JSActivation.h: 8268 (JSC::JSActivation::createStructure): 8269 * runtime/JSArray.h: 8270 (JSC::JSArray::createStructure): 8271 (JSC::MarkStack::markChildren): 8272 * runtime/JSByteArray.cpp: 8273 (JSC::JSByteArray::createStructure): 8274 * runtime/JSFunction.h: 8275 (JSC::JSFunction::createStructure): 8276 * runtime/JSGlobalObject.h: 8277 (JSC::JSGlobalObject::createStructure): 8278 * runtime/JSNotAnObject.h: 8279 (JSC::JSNotAnObject::createStructure): 8280 * runtime/JSNumberCell.h: 8281 (JSC::JSNumberCell::createStructure): 8282 * runtime/JSONObject.h: 8283 (JSC::JSONObject::createStructure): 8284 * runtime/JSObject.h: 8285 (JSC::JSObject::createStructure): 8286 * runtime/JSPropertyNameIterator.h: 8287 (JSC::JSPropertyNameIterator::createStructure): 8288 * runtime/JSStaticScopeObject.h: 8289 (JSC::JSStaticScopeObject::createStructure): 8290 * runtime/JSString.h: 8291 (JSC::JSString::createStructure): 8292 * runtime/JSTypeInfo.h: 8293 (JSC::TypeInfo::overridesMarkChildren): 8294 * runtime/JSVariableObject.h: 8295 (JSC::JSVariableObject::createStructure): 8296 * runtime/JSWrapperObject.h: 8297 (JSC::JSWrapperObject::createStructure): 8298 * runtime/MathObject.h: 8299 (JSC::MathObject::createStructure): 8300 * runtime/NumberConstructor.h: 8301 (JSC::NumberConstructor::createStructure): 8302 * runtime/NumberObject.h: 8303 (JSC::NumberObject::createStructure): 8304 * runtime/RegExpConstructor.h: 8305 (JSC::RegExpConstructor::createStructure): 8306 * runtime/RegExpObject.h: 8307 (JSC::RegExpObject::createStructure): 8308 * runtime/StringObject.h: 8309 (JSC::StringObject::createStructure): 8310 * runtime/StringObjectThatMasqueradesAsUndefined.h: 8311 (JSC::StringObjectThatMasqueradesAsUndefined::createStructure): 8312 83132009-10-14 Oliver Hunt <oliver@apple.com> 8314 8315 Reviewed by Geoff Garen. 8316 8317 Make typeinfo flags default to false 8318 https://bugs.webkit.org/show_bug.cgi?id=30372 8319 8320 Part 1. Reverse the HasStandardGetOwnPropertySlot flag. 8321 8322 * API/JSCallbackConstructor.h: 8323 (JSC::JSCallbackConstructor::createStructure): 8324 * API/JSCallbackFunction.h: 8325 (JSC::JSCallbackFunction::createStructure): 8326 * API/JSCallbackObject.h: 8327 (JSC::JSCallbackObject::createStructure): 8328 * debugger/DebuggerActivation.h: 8329 (JSC::DebuggerActivation::createStructure): 8330 * jit/JITStubs.cpp: 8331 (JSC::DEFINE_STUB_FUNCTION): 8332 * runtime/Arguments.h: 8333 (JSC::Arguments::createStructure): 8334 * runtime/BooleanObject.h: 8335 (JSC::BooleanObject::createStructure): 8336 * runtime/DatePrototype.h: 8337 (JSC::DatePrototype::createStructure): 8338 * runtime/FunctionPrototype.h: 8339 (JSC::FunctionPrototype::createStructure): 8340 * runtime/GlobalEvalFunction.h: 8341 (JSC::GlobalEvalFunction::createStructure): 8342 * runtime/InternalFunction.h: 8343 (JSC::InternalFunction::createStructure): 8344 * runtime/JSActivation.h: 8345 (JSC::JSActivation::createStructure): 8346 * runtime/JSArray.h: 8347 (JSC::JSArray::createStructure): 8348 * runtime/JSByteArray.cpp: 8349 (JSC::JSByteArray::createStructure): 8350 * runtime/JSFunction.h: 8351 (JSC::JSFunction::createStructure): 8352 * runtime/JSGlobalObject.h: 8353 (JSC::JSGlobalObject::createStructure): 8354 * runtime/JSNumberCell.h: 8355 (JSC::JSNumberCell::createStructure): 8356 * runtime/JSONObject.h: 8357 (JSC::JSONObject::createStructure): 8358 * runtime/JSObject.h: 8359 (JSC::JSObject::createStructure): 8360 (JSC::JSCell::fastGetOwnPropertySlot): 8361 * runtime/JSStaticScopeObject.h: 8362 (JSC::JSStaticScopeObject::createStructure): 8363 * runtime/JSString.h: 8364 (JSC::JSString::createStructure): 8365 * runtime/JSTypeInfo.h: 8366 (JSC::TypeInfo::overridesGetOwnPropertySlot): 8367 * runtime/JSVariableObject.h: 8368 (JSC::JSVariableObject::createStructure): 8369 * runtime/JSWrapperObject.h: 8370 (JSC::JSWrapperObject::createStructure): 8371 * runtime/MathObject.h: 8372 (JSC::MathObject::createStructure): 8373 * runtime/NumberConstructor.h: 8374 (JSC::NumberConstructor::createStructure): 8375 * runtime/NumberObject.h: 8376 (JSC::NumberObject::createStructure): 8377 * runtime/RegExpConstructor.h: 8378 (JSC::RegExpConstructor::createStructure): 8379 * runtime/RegExpObject.h: 8380 (JSC::RegExpObject::createStructure): 8381 * runtime/StringObject.h: 8382 (JSC::StringObject::createStructure): 8383 * runtime/StringObjectThatMasqueradesAsUndefined.h: 8384 (JSC::StringObjectThatMasqueradesAsUndefined::createStructure): 8385 83862009-10-14 Kevin Ollivier <kevino@theolliviers.com> 83872009-10-14 Darin Adler <darin@apple.com> 8388 8389 Additions so fix for https://bugs.webkit.org/show_bug.cgi?id=18994 8390 can build on Windows. 8391 8392 * wtf/MathExtras.h: Added llround and llroundf for Windows. 8393 83942009-10-14 Kevin Ollivier <kevino@theolliviers.com> 8395 8396 wx build fix. Set ENABLE_PLUGIN_PACKAGE_SIMPLE_HASH for plugins while we're still building stubs. 8397 8398 * wtf/Platform.h: 8399 84002009-10-13 Laszlo Gombos <laszlo.1.gombos@nokia.com> 8401 8402 Reviewed by Simon Hausmann. 8403 8404 Refactor ENABLE_PLUGIN_PACKAGE_SIMPLE_HASH 8405 https://bugs.webkit.org/show_bug.cgi?id=30278 8406 8407 Move the definition of ENABLE_PLUGIN_PACKAGE_SIMPLE_HASH 8408 from the make system into common code. 8409 8410 * wtf/Platform.h: 8411 84122009-10-13 Laszlo Gombos <laszlo.1.gombos@nokia.com> 8413 8414 Reviewed by Darin Adler. 8415 8416 ARM compiler does not understand reinterpret_cast<void*> 8417 https://bugs.webkit.org/show_bug.cgi?id=29034 8418 8419 Change reinterpret_cast<void*> to regular C style (void*) cast 8420 for the ARM RVCT compiler. 8421 8422 * assembler/MacroAssemblerCodeRef.h: 8423 (JSC::FunctionPtr::FunctionPtr): 8424 * jit/JITOpcodes.cpp: Cast to FunctionPtr first 8425 instead of directly casting to reinterpret_cast 8426 * jit/JITStubCall.h: Ditto + change the type of m_stub 8427 from void* to FunctionPtr. 8428 (JSC::JITStubCall::JITStubCall): 8429 (JSC::JITStubCall::call): 8430 * jit/JITStubs.cpp: Ditto. 8431 (JSC::DEFINE_STUB_FUNCTION(EncodedJSValue, op_throw)): 8432 84332009-10-11 Oliver Hunt <oliver@apple.com> 8434 8435 Re-enable the JIT. 8436 8437 * wtf/Platform.h: 8438 84392009-10-10 Oliver Hunt <oliver@apple.com> 8440 8441 Reviewed by Maciej Stachowiak. 8442 8443 Support for String.trim(), String.trimLeft() and String.trimRight() methods 8444 https://bugs.webkit.org/show_bug.cgi?id=26590 8445 8446 Implement trim, trimLeft, and trimRight 8447 8448 * runtime/StringPrototype.cpp: 8449 (JSC::isTrimWhitespace): 8450 Our normal string whitespace function does not include U+200B which 8451 is needed for compatibility with mozilla's implementation of trim. 8452 U+200B does not appear to be expected according to spec, however I am 8453 choosing to be lax, and match mozilla behavior so have added this 8454 exception. 8455 (JSC::trimString): 8456 84572009-10-09 Geoffrey Garen <ggaren@apple.com> 8458 8459 Reviewed by Oliver Hunt. 8460 8461 Eliminated some legacy bytecode weirdness. 8462 8463 Use vPC[x] subscripting instead of ++vPC to access instruction operands. 8464 This is simpler, and often more efficient. 8465 8466 To support this, and to remove use of hard-coded offsets in bytecode and 8467 JIT code generation and dumping, calculate jump offsets from the beginning 8468 of an instruction, rather than the middle or end. 8469 8470 Also, use OPCODE_LENGTH instead of hard-coded constants for the sizes of 8471 opcodes. 8472 8473 SunSpider reports no change in JIT mode, and a 1.01x speedup in Interpreter 8474 mode. 8475 8476 * bytecode/CodeBlock.cpp: 8477 (JSC::printConditionalJump): 8478 (JSC::CodeBlock::dump): 8479 * bytecompiler/BytecodeGenerator.cpp: 8480 (JSC::BytecodeGenerator::emitJump): 8481 (JSC::BytecodeGenerator::emitJumpIfTrue): 8482 (JSC::BytecodeGenerator::emitJumpIfFalse): 8483 (JSC::BytecodeGenerator::emitJumpIfNotFunctionCall): 8484 (JSC::BytecodeGenerator::emitJumpIfNotFunctionApply): 8485 (JSC::BytecodeGenerator::emitComplexJumpScopes): 8486 (JSC::BytecodeGenerator::emitJumpScopes): 8487 (JSC::BytecodeGenerator::emitNextPropertyName): 8488 (JSC::BytecodeGenerator::emitCatch): 8489 (JSC::BytecodeGenerator::emitJumpSubroutine): 8490 (JSC::prepareJumpTableForImmediateSwitch): 8491 (JSC::prepareJumpTableForCharacterSwitch): 8492 (JSC::prepareJumpTableForStringSwitch): 8493 (JSC::BytecodeGenerator::endSwitch): 8494 * bytecompiler/Label.h: 8495 (JSC::Label::setLocation): 8496 (JSC::Label::bind): 8497 * interpreter/Interpreter.cpp: 8498 (JSC::Interpreter::resolve): 8499 (JSC::Interpreter::resolveSkip): 8500 (JSC::Interpreter::resolveGlobal): 8501 (JSC::Interpreter::resolveBase): 8502 (JSC::Interpreter::resolveBaseAndProperty): 8503 (JSC::Interpreter::createExceptionScope): 8504 (JSC::Interpreter::privateExecute): 8505 * interpreter/Interpreter.h: 8506 * jit/JIT.cpp: 8507 (JSC::JIT::privateCompile): 8508 * jit/JITArithmetic.cpp: 8509 (JSC::JIT::emit_op_jnless): 8510 (JSC::JIT::emitSlow_op_jnless): 8511 (JSC::JIT::emit_op_jnlesseq): 8512 (JSC::JIT::emitSlow_op_jnlesseq): 8513 (JSC::JIT::emitBinaryDoubleOp): 8514 * jit/JITOpcodes.cpp: 8515 (JSC::JIT::emit_op_jmp): 8516 (JSC::JIT::emit_op_loop): 8517 (JSC::JIT::emit_op_loop_if_less): 8518 (JSC::JIT::emitSlow_op_loop_if_less): 8519 (JSC::JIT::emit_op_loop_if_lesseq): 8520 (JSC::JIT::emitSlow_op_loop_if_lesseq): 8521 (JSC::JIT::emit_op_loop_if_true): 8522 (JSC::JIT::emitSlow_op_loop_if_true): 8523 (JSC::JIT::emit_op_jfalse): 8524 (JSC::JIT::emitSlow_op_jfalse): 8525 (JSC::JIT::emit_op_jtrue): 8526 (JSC::JIT::emitSlow_op_jtrue): 8527 (JSC::JIT::emit_op_jeq_null): 8528 (JSC::JIT::emit_op_jneq_null): 8529 (JSC::JIT::emit_op_jneq_ptr): 8530 (JSC::JIT::emit_op_jsr): 8531 (JSC::JIT::emit_op_next_pname): 8532 (JSC::JIT::emit_op_jmp_scopes): 8533 85342009-10-09 Geoffrey Garen <ggaren@apple.com> 8535 8536 Reviewed by Sam Weinig. 8537 8538 Migrated some code that didn't belong out of Structure. 8539 8540 SunSpider says maybe 1.03x faster. 8541 8542 * runtime/JSCell.h: Nixed Structure::markAggregate, and made marking of 8543 a Structure's prototype the direct responsility of the object using it. 8544 (Giving Structure a mark function was misleading because it implied that 8545 all live structures get marked during GC, when they don't.) 8546 8547 * runtime/JSGlobalObject.cpp: 8548 (JSC::markIfNeeded): 8549 (JSC::JSGlobalObject::markChildren): Added code to mark prototypes stored 8550 on the global object. Maybe this wasn't necessary, but now we don't have 8551 to wonder. 8552 8553 * runtime/JSObject.cpp: 8554 (JSC::JSObject::getPropertyNames): 8555 (JSC::JSObject::getOwnPropertyNames): 8556 (JSC::JSObject::getEnumerableNamesFromClassInfoTable): 8557 * runtime/JSObject.h: 8558 (JSC::JSObject::markChildrenDirect): 8559 * runtime/PropertyNameArray.h: 8560 * runtime/Structure.cpp: 8561 * runtime/Structure.h: 8562 (JSC::Structure::setEnumerationCache): 8563 (JSC::Structure::enumerationCache): Moved property name gathering code 8564 from Structure to JSObject because having a Structure iterate its JSObject 8565 was a layering violation. A JSObject is implemented using a Structure; not 8566 the other way around. 8567 85682009-10-09 Mark Rowe <mrowe@apple.com> 8569 8570 Attempt to fix the GTK release build. 8571 8572 * GNUmakefile.am: Include Grammar.cpp in release builds now that 8573 AllInOneFile.cpp is gone. 8574 85752009-10-09 Gabor Loki <loki@inf.u-szeged.hu> 8576 8577 Rubber-stamped by Eric Seidel. 8578 8579 Add ARM JIT support for Gtk port (disabled by default) 8580 https://bugs.webkit.org/show_bug.cgi?id=30228 8581 8582 * GNUmakefile.am: 8583 85842009-10-08 Geoffrey Garen <ggaren@apple.com> 8585 8586 Tiger build fix: added a few more variable initializations. 8587 8588 * runtime/StringPrototype.cpp: 8589 (JSC::stringProtoFuncReplace): 8590 (JSC::stringProtoFuncSearch): 8591 85922009-10-08 Geoffrey Garen <ggaren@apple.com> 8593 8594 Qt build fix: added missing #include. 8595 8596 * jsc.cpp: 8597 85982009-10-08 Geoffrey Garen <ggaren@apple.com> 8599 8600 Tiger build fix: initialize variable whose initialization the compiler 8601 can't otherwise figure out. 8602 8603 * runtime/RegExpObject.cpp: 8604 (JSC::RegExpObject::match): 8605 86062009-10-08 Geoffrey Garen <ggaren@apple.com> 8607 8608 Windows build fix: updated exports. 8609 8610 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 8611 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 8612 86132009-10-08 Geoffrey Garen <ggaren@apple.com> 8614 8615 Tiger build fix: fixed file name case. 8616 8617 * JavaScriptCore.xcodeproj/project.pbxproj: 8618 86192009-10-08 Geoffrey Garen <ggaren@apple.com> 8620 8621 Reviewed by Maciej Stachowiak. 8622 8623 At long last, I pronounce the death of AllInOneFile.cpp. 8624 8625 SunSpider reports a 1.01x speedup. 8626 8627 * AllInOneFile.cpp: Removed. 8628 * GNUmakefile.am: 8629 * JavaScriptCore.exp: 8630 * JavaScriptCore.gypi: 8631 * JavaScriptCore.xcodeproj/project.pbxproj: Added missing project files 8632 to compilation stages. 8633 8634 * parser/Grammar.y: 8635 * parser/Lexer.cpp: 8636 * parser/Lexer.h: 8637 (JSC::jscyylex): 8638 * runtime/ArrayConstructor.cpp: 8639 (JSC::constructArrayWithSizeQuirk): 8640 * runtime/Collector.h: 8641 * runtime/JSCell.cpp: 8642 (JSC::JSCell::operator new): 8643 * runtime/JSCell.h: 8644 (JSC::JSCell::operator new): 8645 * runtime/JSGlobalObject.cpp: 8646 (JSC::JSGlobalObject::operator new): 8647 * runtime/JSNumberCell.h: 8648 (JSC::JSNumberCell::operator new): 8649 * runtime/JSString.cpp: 8650 * runtime/JSString.h: 8651 (JSC::jsString): 8652 (JSC::jsSubstring): 8653 (JSC::jsOwnedString): 8654 * runtime/RegExpConstructor.cpp: 8655 * runtime/RegExpConstructor.h: 8656 (JSC::RegExpConstructorPrivate::RegExpConstructorPrivate): 8657 (JSC::RegExpConstructorPrivate::lastOvector): 8658 (JSC::RegExpConstructorPrivate::tempOvector): 8659 (JSC::RegExpConstructorPrivate::changeLastOvector): 8660 (JSC::RegExpConstructor::performMatch): 8661 * runtime/StringPrototype.cpp: 8662 (JSC::stringProtoFuncMatch): 8663 * yarr/RegexJIT.cpp: 8664 * yarr/RegexJIT.h: 8665 (JSC::Yarr::executeRegex): Inlined a few things that Shark said 8666 were hot, on the presumption that AllInOneFile.cpp used to inline them 8667 automatically. 8668 86692009-10-08 Zoltan Herczeg <zherczeg@inf.u-szeged.hu> 8670 8671 Reviewed by Gavin Barraclough. 8672 8673 Fix for JIT'ed op_call instructions (evals, constructs, etc.) 8674 when !ENABLE(JIT_OPTIMIZE_CALL) && USE(JSVALUE32_64) 8675 8676 https://bugs.webkit.org/show_bug.cgi?id=30201 8677 8678 * jit/JITCall.cpp: 8679 (JSC::JIT::compileOpCall): 8680 86812009-10-07 Geoffrey Garen <ggaren@apple.com> 8682 8683 Windows build fix: removed no longer exported symbol. 8684 8685 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 8686 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 8687 86882009-10-07 Geoffrey Garen <ggaren@apple.com> 8689 8690 Reviewed by Oliver Hunt. 8691 8692 Fixed <rdar://problem/5751979> Database code takes JSLock on secondary 8693 thread, permanently slowing down JavaScript 8694 8695 Removed the optional lock from Heap::protect, Heap::unprotect, and friends, 8696 since WebCore no longer uses it. 8697 8698 * JavaScriptCore.exp: 8699 * runtime/Collector.cpp: 8700 (JSC::Heap::protect): 8701 (JSC::Heap::unprotect): 8702 (JSC::Heap::markProtectedObjects): 8703 (JSC::Heap::protectedGlobalObjectCount): 8704 (JSC::Heap::protectedObjectCount): 8705 (JSC::Heap::protectedObjectTypeCounts): 8706 * runtime/Collector.h: 8707 87082009-10-07 Zoltan Horvath <zoltan@webkit.org> 8709 8710 Reviewed by Darin Adler. 8711 8712 Allow custom memory allocation control for JavaScriptCore's IdentifierArena 8713 https://bugs.webkit.org/show_bug.cgi?id=30158 8714 8715 Inherits IdentifierArena class from FastAllocBase because it has been 8716 instantiated by 'new' in JavaScriptCore/parser/ParserArena.cpp:36. 8717 8718 * parser/ParserArena.h: 8719 87202009-10-07 Adam Roben <aroben@apple.com> 8721 8722 Export DateInstance::info in a way that works on Windows 8723 8724 Fixes <http://webkit.org/b/30171> 8725 fast/dom/Window/window-postmessage-clone.html fails on Windows 8726 8727 Reviewed by Anders Carlsson. 8728 8729 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 8730 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 8731 Removed the export of DateInstance::info from here. 8732 8733 * runtime/DateInstance.h: Use JS_EXPORTDATA to export 8734 DateInstance::info, which is the required way of exporting data on 8735 Windows. 8736 87372009-10-07 Jørgen Lind <jorgen.lind@nokia.com> 8738 8739 Reviewed by Simon Hausmann. 8740 8741 When enabling or disabling the JIT through .qmake.cache, make sure 8742 to also toggle ENABLE_YARR_JIT. 8743 8744 * JavaScriptCore.pri: 8745 87462009-10-06 Priit Laes <plaes@plaes.org> 8747 8748 Reviewed by Gavin Barraclough. 8749 8750 Linking fails with "relocation R_X86_64_PC32 against symbol 8751 `cti_vm_throw'" 8752 https://bugs.webkit.org/show_bug.cgi?id=28422 8753 8754 * jit/JITStubs.cpp: 8755 Mark cti_vm_throw symbol as PLT-indirect symbol, so it doesn't end up 8756 in text segment causing relocation errors on amd64 architecture. 8757 Introduced new define SYMBOL_STRING_RELOCATION for such symbols. 8758 87592009-10-06 Oliver Hunt <oliver@apple.com> 8760 8761 Windows linking fix 8762 8763 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 8764 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 8765 87662009-10-06 Oliver Hunt <oliver@apple.com> 8767 8768 Reviewed by NOBODY (build fix). 8769 8770 Windows build fix. 8771 8772 * runtime/DateInstance.cpp: 8773 87742009-10-05 Oliver Hunt <oliver@apple.com> 8775 8776 Reviewed by Gavin Barraclough. 8777 8778 It should be possible to post (clone) built-in JS objects to Workers 8779 https://bugs.webkit.org/show_bug.cgi?id=22878 8780 8781 Expose helpers to throw correct exceptions during object graph walk 8782 used for cloning and add a helper function to create Date instances 8783 without going through the JS Date constructor function. 8784 8785 * JavaScriptCore.exp: 8786 * JavaScriptCore.xcodeproj/project.pbxproj: 8787 * runtime/DateInstance.cpp: 8788 (JSC::DateInstance::DateInstance): 8789 * runtime/DateInstance.h: 8790 * runtime/ExceptionHelpers.cpp: 8791 (JSC::createTypeError): 8792 * runtime/ExceptionHelpers.h: 8793 87942009-10-06 David Levin <levin@chromium.org> 8795 8796 Reviewed by Oliver Hunt. 8797 8798 StringImpl needs a method to get an instance for another thread which doesn't copy the underlying buffer. 8799 https://bugs.webkit.org/show_bug.cgi?id=30095 8800 8801 * wtf/CrossThreadRefCounted.h: 8802 Removed an unused function and assert improvement. 8803 (WTF::CrossThreadRefCounted::isOwnedByCurrentThread): Moved out common code from asserts. 8804 (WTF::CrossThreadRefCounted::ref): Changed assert to use the common method. 8805 (WTF::CrossThreadRefCounted::deref): Changed assert to use the common method. 8806 (WTF::CrossThreadRefCounted::crossThreadCopy): Since this includes a potentially 8807 non-threadsafe operation, add an assert that the class is owned by the current thread. 8808 88092009-10-05 Kevin Ollivier <kevino@theolliviers.com> 8810 8811 wx build fix. Add Symbian files to the list of excludes. 8812 8813 * wscript: 8814 88152009-10-05 Jocelyn Turcotte <jocelyn.turcotte@nokia.com> 8816 8817 Reviewed by Simon Hausmann. 8818 8819 [Qt] Remove precompiled header from JavaScriptCore compilation to 8820 prevent qmake warning during autonomous compilation. 8821 https://bugs.webkit.org/show_bug.cgi?id=30069 8822 8823 * JavaScriptCore.pro: 8824 88252009-10-02 Geoffrey Garen <ggaren@apple.com> 8826 8827 Reviewed by Sam Weinig. 8828 8829 Removed the concept of a "fast access cutoff" in arrays, because it 8830 punished some patterns of array access too much, and made things too 8831 complex for inlining in some cases. 8832 8833 1.3% speedup on SunSpider. 8834 8835 * jit/JITOpcodes.cpp: 8836 (JSC::JIT::emitSlow_op_get_by_val): 8837 (JSC::JIT::emitSlow_op_put_by_val): 8838 * jit/JITPropertyAccess.cpp: 8839 (JSC::JIT::emit_op_get_by_val): 8840 (JSC::JIT::emitSlow_op_get_by_val): 8841 (JSC::JIT::emit_op_put_by_val): 8842 (JSC::JIT::emitSlow_op_put_by_val): 8843 * jit/JITStubs.cpp: 8844 * jit/JITStubs.h: 8845 (JSC::): Check m_vectorLength instead of m_fastAccessCutoff when 8846 getting / putting from / to an array. Inline putting past the end of 8847 the array. 8848 8849 * runtime/JSArray.cpp: 8850 (JSC::JSArray::JSArray): 8851 (JSC::JSArray::getOwnPropertySlot): 8852 (JSC::JSArray::getOwnPropertyDescriptor): 8853 (JSC::JSArray::put): 8854 (JSC::JSArray::putSlowCase): 8855 (JSC::JSArray::deleteProperty): 8856 (JSC::JSArray::getOwnPropertyNames): 8857 (JSC::JSArray::increaseVectorLength): 8858 (JSC::JSArray::setLength): 8859 (JSC::JSArray::pop): 8860 (JSC::JSArray::push): 8861 (JSC::JSArray::sort): 8862 (JSC::JSArray::fillArgList): 8863 (JSC::JSArray::copyToRegisters): 8864 (JSC::JSArray::compactForSorting): 8865 (JSC::JSArray::checkConsistency): 8866 * runtime/JSArray.h: 8867 (JSC::JSArray::canGetIndex): 8868 (JSC::JSArray::canSetIndex): 8869 (JSC::JSArray::setIndex): 8870 (JSC::JSArray::markChildrenDirect): Removed m_fastAccessCutoff, and 8871 replaced with checks for JSValue() to detect reads and writes from / to 8872 uninitialized parts of the array. 8873 88742009-10-02 Jonni Rainisto <jonni.rainisto@nokia.com> 8875 8876 Reviewed by Darin Adler. 8877 8878 Math.random() gives too low values on Win32 when _CRT_RAND_S is not defined 8879 https://bugs.webkit.org/show_bug.cgi?id=29956 8880 8881 * wtf/RandomNumber.cpp: 8882 (WTF::randomNumber): Added PLATFORM(WIN_OS) to handle 15bit rand() 8883 88842009-10-02 Geoffrey Garen <ggaren@apple.com> 8885 8886 Reviewed by Sam Weinig. 8887 8888 Take one branch instead of two to test for JSValue(). 8889 8890 1.1% SunSpider speedup. 8891 8892 * jit/JITCall.cpp: 8893 (JSC::JIT::compileOpCall): 8894 * jit/JITOpcodes.cpp: 8895 (JSC::JIT::emit_op_to_jsnumber): 8896 (JSC::JIT::emit_op_create_arguments): 8897 * jit/JITPropertyAccess.cpp: 8898 (JSC::JIT::emitSlow_op_get_by_val): 8899 (JSC::JIT::emit_op_put_by_val): Test for the empty value tag, instead 8900 of testing for the cell tag with a 0 payload. 8901 8902 * runtime/JSValue.cpp: 8903 (JSC::JSValue::description): Added support for dumping the new empty value, 8904 and deleted values, in debug builds. 8905 8906 * runtime/JSValue.h: 8907 (JSC::JSValue::JSValue()): Construct JSValue() with the empty value tag. 8908 8909 (JSC::JSValue::JSValue(JSCell*)): Convert null pointer to the empty value 8910 tag, to avoid having two different c++ versions of null / empty. 8911 8912 (JSC::JSValue::operator bool): Test for the empty value tag, instead 8913 of testing for the cell tag with a 0 payload. 8914 89152009-10-02 Steve Falkenburg <sfalken@apple.com> 8916 8917 Reviewed by Mark Rowe. 8918 8919 <https://bugs.webkit.org/show_bug.cgi?id=29989> 8920 Safari version number shouldn't be exposed in WebKit code 8921 8922 For a WebKit version of 532.3.4: 8923 Product version is: 5.32.3.4 (was 4.0.3.0) 8924 File version is: 5.32.3.4 (was 4.532.3.4) 8925 8926 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.rc: 8927 89282009-10-02 Tor Arne Vestbø <tor.arne.vestbo@nokia.com> 8929 8930 Rubber-stamped by Simon Hausmann. 8931 8932 Fix the Qt on Mac OS X build. 8933 8934 * wtf/FastMalloc.cpp: 8935 89362009-10-02 Jørgen Lind <jorgen.lind@nokia.com> 8937 8938 Reviewed by Simon Hausmann. 8939 8940 Allow enabling and disabling of the JIT through a qmake variable. 8941 8942 Qt's configure may set this variable through .qmake.cache if a 8943 commandline option is given and/or the compile test for hwcap.h 8944 failed/succeeded. 8945 8946 * JavaScriptCore.pri: 8947 89482009-10-01 Mark Rowe <mrowe@apple.com> 8949 8950 Fix the Tiger build. Don't unconditionally enable 3D canvas as it is not supported on Tiger. 8951 8952 * Configurations/FeatureDefines.xcconfig: 8953 89542009-10-01 Yongjun Zhang <yongjun.zhang@nokia.com> 8955 8956 Reviewed by Darin Adler. 8957 8958 https://bugs.webkit.org/show_bug.cgi?id=29187 8959 8960 Don't inline ~ListRefPtr() to work around winscw compiler forward declaration 8961 bug regarding templated classes. 8962 8963 The compiler bug is reported at: 8964 https://xdabug001.ext.nokia.com/bugzilla/show_bug.cgi?id=9812 8965 8966 The change will be reverted when the above bug is fixed in winscw compiler. 8967 8968 * wtf/ListRefPtr.h: 8969 (WTF::::~ListRefPtr): 8970 89712009-10-01 Zoltan Horvath <zoltan@webkit.org> 8972 8973 Reviewed by Simon Hausmann. 8974 8975 [Qt] Allow custom memory allocation control for the whole JavaScriptCore 8976 https://bugs.webkit.org/show_bug.cgi?id=27029 8977 8978 Since in JavaScriptCore almost every class which has been instantiated by operator new is 8979 inherited from FastAllocBase (bug #20422), we disable customizing global operator new for the Qt-port 8980 when USE_SYSTEM_MALLOC=0. 8981 8982 Add #include <unistd.h> to FastMalloc.cpp because it's used by TCMalloc_PageHeap::scavengerThread(). 8983 (It's needed for the functionality of TCmalloc.) 8984 8985 Add TCSystemAlloc.cpp to JavaScriptCore.pri if USE_SYSTEM_MALLOC is disabled. 8986 8987 * JavaScriptCore.pri: 8988 * wtf/FastMalloc.cpp: 8989 (WTF::sleep): 8990 * wtf/FastMalloc.h: 8991 89922009-09-30 Gabor Loki <loki@inf.u-szeged.hu> 8993 8994 Reviewed by George Staikos. 8995 8996 Defines two pseudo-platforms for ARM and Thumb-2 instruction set. 8997 https://bugs.webkit.org/show_bug.cgi?id=29122 8998 8999 Introduces WTF_PLATFORM_ARM_TRADITIONAL and WTF_PLATFORM_ARM_THUMB2 9000 macros on ARM platforms. The PLATFORM(ARM_THUMB2) should be used 9001 when Thumb-2 instruction set is the required target. The 9002 PLATFORM(ARM_TRADITIONAL) is for generic ARM instruction set. In 9003 case where the code is common the PLATFORM(ARM) have to be used. 9004 9005 Modified by George Wright <gwright@rim.com> to correctly work 9006 with the RVCT-defined __TARGET_ARCH_ARM and __TARGET_ARCH_THUMB 9007 compiler macros, as well as adding readability changes. 9008 9009 * wtf/Platform.h: 9010 90112009-09-30 Oliver Hunt <oliver@apple.com> 9012 9013 Reviewed by Geoff Garen. 9014 9015 Devirtualise array toString conversion 9016 9017 Tweak the implementation of Array.prototype.toString to have a fast path 9018 when acting on a true JSArray. 9019 9020 * runtime/ArrayPrototype.cpp: 9021 (JSC::arrayProtoFuncToString): 9022 90232009-09-30 Csaba Osztrogonac <oszi@inf.u-szeged.hu> 9024 9025 Reviewed by Geoffrey Garen. 9026 9027 Buildfix for platforms using JSVALUE32. 9028 https://bugs.webkit.org/show_bug.cgi?id=29915 9029 9030 After http://trac.webkit.org/changeset/48905 the build broke in JSVALUE32 case. 9031 Also removed unreachable code. 9032 9033 * jit/JITArithmetic.cpp: 9034 (JSC::JIT::emit_op_add): 9035 - Declaration of "OperandTypes types" moved before first use. 9036 - Typos fixed: dst modified to result, regT2 added. 9037 - Unreachable code removed. 9038 (JSC::JIT::emitSlow_op_add): 9039 - Missing declaration of "OperandTypes types" added. 9040 90412009-09-30 Janne Koskinen <janne.p.koskinen@digia.com> 9042 9043 Reviewed by Simon Hausmann. 9044 9045 Reduce heap size on Symbian from 64MB to 8MB. 9046 9047 This is not a perfect fix, it requires more fine tuning. 9048 But this makes it possible again to debug in the emulator, 9049 which is more important in order to be able to fix other 9050 run-time issues. 9051 9052 * runtime/Collector.h: 9053 90542009-09-30 Janne Koskinen <janne.p.koskinen@digia.com> 9055 9056 Reviewed by Simon Hausmann. 9057 9058 Fix CRASH() macro for Symbian build. 9059 9060 * wtf/Assertions.h: Added missing } 9061 90622009-09-29 Geoffrey Garen <ggaren@apple.com> 9063 9064 Reviewed by Gavin Barraclough. 9065 9066 Inlined a few math operations. 9067 9068 ~1% SunSpider speedup. 9069 9070 * jit/JIT.h: 9071 * jit/JITArithmetic.cpp: 9072 (JSC::JIT::compileBinaryArithOpSlowCase): 9073 (JSC::JIT::emitSlow_op_add): 9074 (JSC::JIT::emitSlow_op_mul): 9075 (JSC::JIT::emit_op_sub): 9076 (JSC::JIT::emitSlow_op_sub): Don't take a stub call when operating on 9077 a constant int and a double. 9078 90792009-09-28 Oliver Hunt <oliver@apple.com> 9080 9081 Reviewed by Gavin Barraclough. 9082 9083 Tidy up codeblock sampler 9084 https://bugs.webkit.org/show_bug.cgi?id=29836 9085 9086 Some rather simple refactoring of codeblock sampler so that 9087 it's easier for us to use it to find problems in non-jsc 9088 environments 9089 9090 * JavaScriptCore.exp: 9091 * bytecode/SamplingTool.h: 9092 * debugger/Debugger.cpp: 9093 (JSC::evaluateInGlobalCallFrame): 9094 * debugger/DebuggerCallFrame.cpp: 9095 (JSC::DebuggerCallFrame::evaluate): 9096 * interpreter/Interpreter.cpp: 9097 (JSC::Interpreter::Interpreter): 9098 (JSC::Interpreter::execute): 9099 (JSC::Interpreter::privateExecute): 9100 (JSC::Interpreter::enableSampler): 9101 (JSC::Interpreter::dumpSampleData): 9102 (JSC::Interpreter::startSampling): 9103 (JSC::Interpreter::stopSampling): 9104 * interpreter/Interpreter.h: 9105 (JSC::Interpreter::sampler): 9106 * jit/JIT.h: 9107 * jsc.cpp: 9108 (runWithScripts): 9109 * runtime/Completion.cpp: 9110 (JSC::checkSyntax): 9111 (JSC::evaluate): 9112 * runtime/Executable.h: 9113 (JSC::EvalExecutable::EvalExecutable): 9114 (JSC::ProgramExecutable::create): 9115 (JSC::ProgramExecutable::ProgramExecutable): 9116 * runtime/JSGlobalData.cpp: 9117 (JSC::JSGlobalData::startSampling): 9118 (JSC::JSGlobalData::stopSampling): 9119 (JSC::JSGlobalData::dumpSampleData): 9120 * runtime/JSGlobalData.h: 9121 * runtime/JSGlobalObjectFunctions.cpp: 9122 (JSC::globalFuncEval): 9123 91242009-09-29 Jeremy Orlow <jorlow@chromium.org> 9125 9126 Reviewed by Dimitri Glazkov. 9127 9128 Add GYP generated files to svn:ignore 9129 https://bugs.webkit.org/show_bug.cgi?id=29895 9130 9131 The following files are generated by JavaScriptCore's GYP file and should be ignored: 9132 9133 pcre.mk 9134 wtf.scons 9135 wtf.mk 9136 SConstruct 9137 wtf_config.scons 9138 wtf_config.mk 9139 pcre.scons 9140 9141 * JavaScriptCore.gyp: Changed property svn:ignore. 9142 91432009-09-29 Geoffrey Garen <ggaren@apple.com> 9144 9145 Reviewed by Sam Weinig. 9146 9147 Standardized an optimization for adding non-numbers. 9148 9149 SunSpider says maybe a tiny speedup. 9150 9151 * jit/JITArithmetic.cpp: 9152 (JSC::JIT::emit_op_add): 9153 (JSC::JIT::emitSlow_op_add): 9154 91552009-09-29 Geoffrey Garen <ggaren@apple.com> 9156 9157 Windows build fix: export a new symbol. 9158 9159 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 9160 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 9161 91622009-09-28 Geoffrey Garen <ggaren@apple.com> 9163 9164 Reviewed by Sam Weinig. 9165 9166 Removed virtual destructor from JSGlobalObjectData to eliminate pointer 9167 fix-ups when accessing JSGlobalObject::d. 9168 9169 Replaced with an explicit destructor function pointer. 9170 9171 6% speedup on bench-alloc-nonretained.js. 9172 9173 * JavaScriptCore.exp: 9174 * runtime/JSGlobalObject.cpp: 9175 (JSC::JSGlobalObject::~JSGlobalObject): 9176 (JSC::JSGlobalObject::destroyJSGlobalObjectData): 9177 * runtime/JSGlobalObject.h: 9178 (JSC::JSGlobalObject::JSGlobalObjectData::JSGlobalObjectData): 9179 (JSC::JSGlobalObject::JSGlobalObject): 9180 91812009-09-29 Janne Koskinen <janne.p.koskinen@digia.com> 9182 9183 Reviewed by David Kilzer. 9184 9185 [Qt] Assert messages prints visible in Symbian 9186 https://bugs.webkit.org/show_bug.cgi?id=29808 9187 9188 Asserts use vprintf to print the messages to stderr. 9189 In Symbian Open C it is not possible to see stderr so 9190 I routed the messages to stdout instead. 9191 9192 * wtf/Assertions.cpp: 9193 91942009-09-29 Janne Koskinen <janne.p.koskinen@digia.com> 9195 9196 Reviewed by Darin Adler. 9197 9198 [Qt] Symbian CRASH macro implementation 9199 9200 Added Symbian specific crash macro that 9201 stops to crash line if JIT debugging is used. 9202 Additional differentiation of access violation 9203 (KERN-EXEC 3) and CRASH panic. 9204 9205 * wtf/Assertions.h: 9206 92072009-09-28 Mark Rowe <mrowe@apple.com> 9208 9209 Fix the PowerPC build. 9210 9211 * JavaScriptCore.exp: 9212 92132009-09-28 Mark Rowe <mrowe@apple.com> 9214 9215 Reviewed by Gavin Barraclough. 9216 9217 <rdar://problem/7195704> JavaScriptCore fails to mark registers when built for x86_64 using LLVM GCC. 9218 9219 * runtime/Collector.cpp: 9220 (JSC::Heap::markCurrentThreadConservatively): Force jmp_buf to use the appropriate alignment for a pointer 9221 to ensure that we correctly interpret the contents of registers during marking. 9222 92232009-09-28 Geoffrey Garen <ggaren@apple.com> 9224 9225 Windows build fix: added new exports. 9226 9227 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 9228 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 9229 92302009-09-28 Geoffrey Garen <ggaren@apple.com> 9231 9232 Windows build fix: removed exports that no longer exist. 9233 9234 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 9235 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 9236 92372009-09-28 Geoffrey Garen <ggaren@apple.com> 9238 9239 Reviewed by Darin Adler. 9240 9241 NotNullPassRefPtr: smart pointer optimized for passing references that are not null 9242 https://bugs.webkit.org/show_bug.cgi?id=29822 9243 9244 Added NotNullPassRefPtr, and deployed it in all places that initialize 9245 JavaScript objects. 9246 9247 2.2% speedup on bench-allocate-nonretained.js. 9248 9249 * API/JSCallbackConstructor.cpp: 9250 (JSC::JSCallbackConstructor::JSCallbackConstructor): 9251 * API/JSCallbackConstructor.h: 9252 * API/JSCallbackObject.h: 9253 * API/JSCallbackObjectFunctions.h: 9254 (JSC::JSCallbackObject::JSCallbackObject): 9255 * JavaScriptCore.exp: 9256 * bytecode/CodeBlock.h: 9257 (JSC::CodeBlock::addFunctionDecl): 9258 (JSC::CodeBlock::addFunctionExpr): 9259 * runtime/ArrayConstructor.cpp: 9260 (JSC::ArrayConstructor::ArrayConstructor): 9261 * runtime/ArrayConstructor.h: 9262 * runtime/ArrayPrototype.cpp: 9263 (JSC::ArrayPrototype::ArrayPrototype): 9264 * runtime/ArrayPrototype.h: 9265 * runtime/BooleanConstructor.cpp: 9266 (JSC::BooleanConstructor::BooleanConstructor): 9267 * runtime/BooleanConstructor.h: 9268 * runtime/BooleanObject.cpp: 9269 (JSC::BooleanObject::BooleanObject): 9270 * runtime/BooleanObject.h: 9271 * runtime/BooleanPrototype.cpp: 9272 (JSC::BooleanPrototype::BooleanPrototype): 9273 * runtime/BooleanPrototype.h: 9274 * runtime/DateConstructor.cpp: 9275 (JSC::DateConstructor::DateConstructor): 9276 * runtime/DateConstructor.h: 9277 * runtime/DateInstance.cpp: 9278 (JSC::DateInstance::DateInstance): 9279 * runtime/DateInstance.h: 9280 * runtime/DatePrototype.cpp: 9281 (JSC::DatePrototype::DatePrototype): 9282 * runtime/DatePrototype.h: 9283 * runtime/ErrorConstructor.cpp: 9284 (JSC::ErrorConstructor::ErrorConstructor): 9285 * runtime/ErrorConstructor.h: 9286 * runtime/ErrorInstance.cpp: 9287 (JSC::ErrorInstance::ErrorInstance): 9288 * runtime/ErrorInstance.h: 9289 * runtime/ErrorPrototype.cpp: 9290 (JSC::ErrorPrototype::ErrorPrototype): 9291 * runtime/ErrorPrototype.h: 9292 * runtime/FunctionConstructor.cpp: 9293 (JSC::FunctionConstructor::FunctionConstructor): 9294 * runtime/FunctionConstructor.h: 9295 * runtime/FunctionPrototype.cpp: 9296 (JSC::FunctionPrototype::FunctionPrototype): 9297 * runtime/FunctionPrototype.h: 9298 * runtime/GlobalEvalFunction.cpp: 9299 (JSC::GlobalEvalFunction::GlobalEvalFunction): 9300 * runtime/GlobalEvalFunction.h: 9301 * runtime/InternalFunction.cpp: 9302 (JSC::InternalFunction::InternalFunction): 9303 * runtime/InternalFunction.h: 9304 (JSC::InternalFunction::InternalFunction): 9305 * runtime/JSActivation.cpp: 9306 (JSC::JSActivation::JSActivation): 9307 * runtime/JSActivation.h: 9308 (JSC::JSActivation::JSActivationData::JSActivationData): 9309 * runtime/JSArray.cpp: 9310 (JSC::JSArray::JSArray): 9311 * runtime/JSArray.h: 9312 * runtime/JSByteArray.cpp: 9313 (JSC::JSByteArray::JSByteArray): 9314 * runtime/JSByteArray.h: 9315 * runtime/JSFunction.cpp: 9316 (JSC::JSFunction::JSFunction): 9317 * runtime/JSFunction.h: 9318 * runtime/JSGlobalObject.h: 9319 (JSC::JSGlobalObject::JSGlobalObject): 9320 * runtime/JSONObject.h: 9321 (JSC::JSONObject::JSONObject): 9322 * runtime/JSObject.h: 9323 (JSC::JSObject::JSObject): 9324 (JSC::JSObject::setStructure): 9325 * runtime/JSVariableObject.h: 9326 (JSC::JSVariableObject::JSVariableObject): 9327 * runtime/JSWrapperObject.h: 9328 (JSC::JSWrapperObject::JSWrapperObject): 9329 * runtime/MathObject.cpp: 9330 (JSC::MathObject::MathObject): 9331 * runtime/MathObject.h: 9332 * runtime/NativeErrorConstructor.cpp: 9333 (JSC::NativeErrorConstructor::NativeErrorConstructor): 9334 * runtime/NativeErrorConstructor.h: 9335 * runtime/NativeErrorPrototype.cpp: 9336 (JSC::NativeErrorPrototype::NativeErrorPrototype): 9337 * runtime/NativeErrorPrototype.h: 9338 * runtime/NumberConstructor.cpp: 9339 (JSC::NumberConstructor::NumberConstructor): 9340 * runtime/NumberConstructor.h: 9341 * runtime/NumberObject.cpp: 9342 (JSC::NumberObject::NumberObject): 9343 * runtime/NumberObject.h: 9344 * runtime/NumberPrototype.cpp: 9345 (JSC::NumberPrototype::NumberPrototype): 9346 * runtime/NumberPrototype.h: 9347 * runtime/ObjectConstructor.cpp: 9348 (JSC::ObjectConstructor::ObjectConstructor): 9349 * runtime/ObjectConstructor.h: 9350 * runtime/ObjectPrototype.cpp: 9351 (JSC::ObjectPrototype::ObjectPrototype): 9352 * runtime/ObjectPrototype.h: 9353 * runtime/PropertyNameArray.h: 9354 (JSC::PropertyNameArrayData::setCachedPrototypeChain): 9355 * runtime/PrototypeFunction.cpp: 9356 (JSC::PrototypeFunction::PrototypeFunction): 9357 * runtime/PrototypeFunction.h: 9358 * runtime/RegExpConstructor.cpp: 9359 (JSC::RegExpConstructor::RegExpConstructor): 9360 * runtime/RegExpConstructor.h: 9361 * runtime/RegExpObject.cpp: 9362 (JSC::RegExpObject::RegExpObject): 9363 * runtime/RegExpObject.h: 9364 (JSC::RegExpObject::RegExpObjectData::RegExpObjectData): 9365 * runtime/RegExpPrototype.cpp: 9366 (JSC::RegExpPrototype::RegExpPrototype): 9367 * runtime/RegExpPrototype.h: 9368 * runtime/StringConstructor.cpp: 9369 (JSC::StringConstructor::StringConstructor): 9370 * runtime/StringConstructor.h: 9371 * runtime/StringObject.cpp: 9372 (JSC::StringObject::StringObject): 9373 * runtime/StringObject.h: 9374 * runtime/StringObjectThatMasqueradesAsUndefined.h: 9375 (JSC::StringObjectThatMasqueradesAsUndefined::StringObjectThatMasqueradesAsUndefined): 9376 * runtime/StringPrototype.cpp: 9377 (JSC::StringPrototype::StringPrototype): 9378 * runtime/StringPrototype.h: 9379 * wtf/PassRefPtr.h: 9380 (WTF::NotNullPassRefPtr::NotNullPassRefPtr): 9381 (WTF::NotNullPassRefPtr::~NotNullPassRefPtr): 9382 (WTF::NotNullPassRefPtr::get): 9383 (WTF::NotNullPassRefPtr::clear): 9384 (WTF::NotNullPassRefPtr::releaseRef): 9385 (WTF::NotNullPassRefPtr::operator*): 9386 (WTF::NotNullPassRefPtr::operator->): 9387 (WTF::NotNullPassRefPtr::operator!): 9388 (WTF::NotNullPassRefPtr::operator UnspecifiedBoolType): 9389 * wtf/RefPtr.h: 9390 (WTF::RefPtr::RefPtr): 9391 (WTF::operator==): 9392 93932009-09-28 Oliver Hunt <oliver@apple.com> 9394 9395 Reviewed by Geoff Garen. 9396 9397 Hard dependency on SSE2 instruction set with JIT 9398 https://bugs.webkit.org/show_bug.cgi?id=29779 9399 9400 Add floating point support checks to op_jfalse and op_jtrue, and 9401 fix the logic for the slow case of op_add 9402 9403 * jit/JITArithmetic.cpp: 9404 (JSC::JIT::emitSlow_op_add): 9405 * jit/JITOpcodes.cpp: 9406 (JSC::JIT::emit_op_jfalse): 9407 (JSC::JIT::emit_op_jtrue): 9408 94092009-09-28 Yaar Schnitman <yaar@chromium.org> 9410 9411 Reviewed by Dimitri Glazkov. 9412 9413 Chromium port - recognize we are being built independently 9414 of chromium and look for dependencies under webkit/chromium rather 9415 than chromium/src. 9416 9417 https://bugs.webkit.org/show_bug.cgi?id=29722 9418 9419 * JavaScriptCore.gyp/JavaScriptCore.gyp: 9420 94212009-09-28 Jakub Wieczorek <faw217@gmail.com> 9422 9423 Reviewed by Simon Hausmann. 9424 9425 [Qt] Implement XSLT support with QtXmlPatterns. 9426 https://bugs.webkit.org/show_bug.cgi?id=28303 9427 9428 * wtf/Platform.h: Add a WTF_USE_QXMLQUERY #define. 9429 94302009-09-28 Gabor Loki <loki@inf.u-szeged.hu> 9431 9432 Reviewed by Simon Hausmann. 9433 9434 Remove __clear_cache which is an internal function of GCC 9435 https://bugs.webkit.org/show_bug.cgi?id=28886 9436 9437 Although __clear_cache is exported from GCC, this is an internal 9438 function. GCC makes no promises about it. 9439 9440 * jit/ExecutableAllocator.h: 9441 (JSC::ExecutableAllocator::cacheFlush): 9442 94432009-09-28 Sam Weinig <sam@webkit.org> 9444 9445 Reviewed by Oliver Hunt. 9446 9447 Fix an absolute path to somewhere in Oliver's machine to a relative path 9448 for derived JSONObject.lut.h. 9449 9450 * JavaScriptCore.xcodeproj/project.pbxproj: 9451 94522009-09-28 Joerg Bornemann <joerg.bornemann@nokia.com> 9453 9454 Reviewed by Simon Hausmann. 9455 9456 Add ARM version detection for Windows CE. 9457 9458 * wtf/Platform.h: 9459 94602009-09-26 Yongjun Zhang <yongjun.zhang@nokia.com> 9461 9462 Reviewed by Simon Hausmann. 9463 9464 Add MarkStackSymbian.cpp to build JavascriptCore for Symbian. 9465 9466 Re-use Windows shrinkAllocation implementation because Symbian doesn't 9467 support releasing part of memory region. 9468 9469 Use fastMalloc and fastFree to implement allocateStack and releaseStack 9470 for Symbian port. 9471 9472 * JavaScriptCore.pri: 9473 * runtime/MarkStack.h: 9474 (JSC::MarkStack::MarkStackArray::shrinkAllocation): 9475 * runtime/MarkStackSymbian.cpp: Added. 9476 (JSC::MarkStack::initializePagesize): 9477 (JSC::MarkStack::allocateStack): 9478 (JSC::MarkStack::releaseStack): 9479 94802009-09-25 Gabor Loki <loki@inf.u-szeged.hu> 9481 9482 Reviewed by Gavin Barraclough. 9483 9484 Fix unaligned data access in YARR_JIT on ARMv5 and below. 9485 https://bugs.webkit.org/show_bug.cgi?id=29695 9486 9487 On ARMv5 and below all data access should be naturally aligned. 9488 In the YARR_JIT there is a case when character pairs are 9489 loaded from the input string, but this data access is not 9490 naturally aligned. This fix introduces load32WithUnalignedHalfWords 9491 and branch32WithUnalignedHalfWords functions which contain 9492 naturally aligned memory loads - half word loads - on ARMv5 and below. 9493 9494 * assembler/MacroAssemblerARM.cpp: 9495 (JSC::MacroAssemblerARM::load32WithUnalignedHalfWords): 9496 * assembler/MacroAssemblerARM.h: 9497 (JSC::MacroAssemblerARM::load32WithUnalignedHalfWords): 9498 (JSC::MacroAssemblerARM::branch32WithUnalignedHalfWords): 9499 * assembler/MacroAssemblerARMv7.h: 9500 (JSC::MacroAssemblerARMv7::load32WithUnalignedHalfWords): 9501 (JSC::MacroAssemblerARMv7::branch32): 9502 (JSC::MacroAssemblerARMv7::branch32WithUnalignedHalfWords): 9503 * assembler/MacroAssemblerX86Common.h: 9504 (JSC::MacroAssemblerX86Common::load32WithUnalignedHalfWords): 9505 (JSC::MacroAssemblerX86Common::branch32WithUnalignedHalfWords): 9506 * wtf/Platform.h: 9507 * yarr/RegexJIT.cpp: 9508 (JSC::Yarr::RegexGenerator::generatePatternCharacterPair): 9509 95102009-09-25 Jeremy Orlow <jorlow@chromium.org> 9511 9512 This is breaking Chromium try bots, so I'm counting this as a build fix. 9513 9514 Add more svn:ignore exceptions. On different platforms, these files are 9515 generated with different case for JavaScriptCore. Also there are some 9516 wtf project files that get built apparently. 9517 9518 * JavaScriptCore.gyp: Changed property svn:ignore. 9519 95202009-09-25 Ada Chan <adachan@apple.com> 9521 9522 Build fix. 9523 9524 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 9525 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 9526 95272009-09-25 Geoffrey Garen <ggaren@apple.com> 9528 9529 Reviewed by Darin Adler. 9530 9531 Inlined some object creation code, including lexicalGlobalObject access 9532 https://bugs.webkit.org/show_bug.cgi?id=29750 9533 9534 SunSpider says 0.5% faster. 9535 9536 0.8% speedup on bench-alloc-nonretained.js. 9537 2.5% speedup on v8-splay.js. 9538 9539 * interpreter/CachedCall.h: 9540 (JSC::CachedCall::CachedCall): 9541 * interpreter/CallFrame.h: 9542 (JSC::ExecState::lexicalGlobalObject): 9543 (JSC::ExecState::globalThisValue): 9544 * interpreter/Interpreter.cpp: 9545 (JSC::Interpreter::dumpRegisters): 9546 (JSC::Interpreter::execute): 9547 (JSC::Interpreter::privateExecute): 9548 * jit/JITStubs.cpp: 9549 (JSC::DEFINE_STUB_FUNCTION): 9550 * runtime/FunctionConstructor.cpp: 9551 (JSC::constructFunction): 9552 * runtime/ScopeChain.cpp: 9553 (JSC::ScopeChainNode::print): 9554 * runtime/ScopeChain.h: 9555 (JSC::ScopeChainNode::ScopeChainNode): 9556 (JSC::ScopeChainNode::~ScopeChainNode): 9557 (JSC::ScopeChainNode::push): 9558 (JSC::ScopeChain::ScopeChain): 9559 (JSC::ScopeChain::globalObject): Added a globalObject data member to ScopeChainNode. 9560 Replaced accessor function for globalObject() with data member. Replaced 9561 globalThisObject() accessor with direct access to globalThis, to match. 9562 9563 * runtime/JSGlobalObject.cpp: 9564 (JSC::JSGlobalObject::init): 9565 * runtime/JSGlobalObject.h: Inlined array and object construction. 9566 95672009-09-25 Laszlo Gombos <laszlo.1.gombos@nokia.com> 9568 9569 Reviewed by Gavin Barraclough. 9570 9571 Add ARM version detection rules for Symbian 9572 https://bugs.webkit.org/show_bug.cgi?id=29715 9573 9574 * wtf/Platform.h: 9575 95762009-09-24 Xan Lopez <xlopez@igalia.com> 9577 9578 Reviewed by Mark "Do It!" Rowe. 9579 9580 Some GCC versions don't like C++-style comments in preprocessor 9581 directives, change to C-style to shut them up. 9582 9583 * wtf/Platform.h: 9584 95852009-09-24 Oliver Hunt <oliver@apple.com> 9586 9587 Reviewed by Gavin Barraclough. 9588 9589 Division is needlessly slow in 64-bit 9590 https://bugs.webkit.org/show_bug.cgi?id=29723 9591 9592 Add codegen for op_div on x86-64 9593 9594 * jit/JIT.cpp: 9595 (JSC::JIT::privateCompileMainPass): 9596 (JSC::JIT::privateCompileSlowCases): 9597 * jit/JIT.h: 9598 * jit/JITArithmetic.cpp: 9599 (JSC::JIT::compileBinaryArithOpSlowCase): 9600 (JSC::JIT::emit_op_div): 9601 (JSC::JIT::emitSlow_op_div): 9602 * jit/JITInlineMethods.h: 9603 (JSC::JIT::isOperandConstantImmediateDouble): 9604 (JSC::JIT::addressFor): 9605 (JSC::JIT::emitLoadDouble): 9606 (JSC::JIT::emitLoadInt32ToDouble): 9607 (JSC::JIT::emitJumpSlowCaseIfNotImmediateNumber): 9608 96092009-09-24 Jeremy Orlow <jorlow@chromium.org> 9610 9611 Reviewed by Dimitri Glazkov. 9612 9613 Add GYP generated files to svn:ignore 9614 https://bugs.webkit.org/show_bug.cgi?id=29724 9615 9616 Adding the following files to the svn:ignore list (all in the 9617 JavaScriptCore/JavaScriptCore.gyp directory) 9618 9619 JavaScriptCore.xcodeproj 9620 JavaScriptCore.sln 9621 JavaScriptCore.vcproj 9622 JavaScriptCore_Debug.rules 9623 JavaScriptCore_Release.rules 9624 JavaScriptCore_Release - no tcmalloc.rules 9625 JavaScriptCore_Purify.rules 9626 JavaScriptCore.mk 9627 JavaScriptCore_Debug_rules.mk 9628 JavaScriptCore_Release_rules.mk 9629 JavaScriptCore_Release - no tcmalloc_rules.mk 9630 JavaScriptCore_Purify_rules.mk 9631 JavaScriptCore.scons 9632 JavaScriptCore_main.scons 9633 9634 * JavaScriptCore.gyp: Changed property svn:ignore. 9635 96362009-09-24 Yong Li <yong.li@torchmobile.com> 9637 9638 Reviewed by Adam Barth. 9639 9640 Replace platform-dependent code with WTF::currentTime() 9641 https://bugs.webkit.org/show_bug.cgi?id=29148 9642 9643 * jsc.cpp: 9644 (StopWatch::start): 9645 (StopWatch::stop): 9646 (StopWatch::getElapsedMS): 9647 * runtime/TimeoutChecker.cpp: 9648 (JSC::getCPUTime): 9649 96502009-09-24 Mark Rowe <mrowe@apple.com> 9651 9652 Reviewed by Sam Weinig. 9653 9654 <rdar://problem/7215058> FastMalloc scavenging thread should be named 9655 9656 * wtf/FastMalloc.cpp: 9657 (WTF::TCMalloc_PageHeap::scavengerThread): Set the thread name. 9658 * wtf/Platform.h: Move the knowledge of whether pthread_setname_np exists to here as HAVE(PTHREAD_SETNAME_NP). 9659 * wtf/ThreadingPthreads.cpp: 9660 (WTF::setThreadNameInternal): Use HAVE(PTHREAD_SETNAME_NP). 9661 96622009-09-24 Geoffrey Garen <ggaren@apple.com> 9663 9664 Reviewed by Sam Weinig. 9665 9666 Renamed clear to removeAll, as suggested by Darin Adler. 9667 9668 * wtf/HashCountedSet.h: 9669 (WTF::::removeAll): 9670 96712009-09-24 Mark Rowe <mrowe@apple.com> 9672 9673 Reviewed by Gavin Barraclough. 9674 9675 Fix FastMalloc to build with assertions enabled. 9676 9677 * wtf/FastMalloc.cpp: 9678 (WTF::TCMalloc_Central_FreeList::ReleaseToSpans): 9679 * wtf/TCSpinLock.h: 9680 (TCMalloc_SpinLock::IsHeld): 9681 96822009-09-24 Geoffrey Garen <ggaren@apple.com> 9683 9684 Suggested by Darin Adler. 9685 9686 Removed some unnecessary parameter names. 9687 9688 * wtf/HashCountedSet.h: 9689 96902009-09-24 Janne Koskinen <janne.p.koskinen@digia.com> 9691 9692 Reviewed by Simon Hausmann. 9693 9694 On Windows JSChar is typedef'ed to wchar_t. 9695 9696 When building with WINSCW for Symbian we need to do the 9697 same typedef. 9698 9699 * API/JSStringRef.h: 9700 97012009-09-23 Geoffrey Garen <ggaren@apple.com> 9702 9703 A piece of my last patch that I forgot. 9704 9705 * wtf/HashCountedSet.h: 9706 (WTF::::clear): Added HashCountedSet::clear. 9707 97082009-09-24 Gabor Loki <loki@inf.u-szeged.hu> 9709 9710 Reviewed by Gavin Barraclough. 9711 9712 Avoid __clear_cache built-in function if DISABLE_BUILTIN_CLEAR_CACHE define is set 9713 https://bugs.webkit.org/show_bug.cgi?id=28886 9714 9715 There are some GCC packages (for example GCC-2006q3 from CodeSourcery) 9716 which contain __clear_cache built-in function only for C while the C++ 9717 version of __clear_cache is missing on ARM architectures. 9718 9719 Fixed a small bug in the inline assembly of cacheFlush function on 9720 ARM_TRADITIONAL. 9721 9722 * jit/ExecutableAllocator.h: 9723 (JSC::ExecutableAllocator::cacheFlush): 9724 97252009-09-23 Geoffrey Garen <ggaren@apple.com> 9726 9727 Reviewed by Sam Weinig. 9728 9729 Added the ability to swap vectors with inline capacities, so you can 9730 store a vector with inline capacity in a hash table. 9731 9732 * wtf/Vector.h: 9733 (WTF::swap): 9734 (WTF::VectorBuffer::swap): 9735 97362009-09-23 David Kilzer <ddkilzer@apple.com> 9737 9738 Move definition of USE(PLUGIN_HOST_PROCESS) from WebKitPrefix.h to Platform.h 9739 9740 Reviewed by Mark Rowe. 9741 9742 * wtf/Platform.h: Define WTF_USE_PLUGIN_HOST_PROCESS to 1 when 9743 building on 64-bit SnowLeopard. Define to 0 elsewhere. 9744 97452009-09-22 Oliver Hunt <oliver@apple.com> 9746 9747 Reviewed by Geoff Garen. 9748 9749 Code sampling builds are broken. 9750 https://bugs.webkit.org/show_bug.cgi?id=29662 9751 9752 Fix build. 9753 9754 * bytecode/EvalCodeCache.h: 9755 (JSC::EvalCodeCache::get): 9756 * bytecode/SamplingTool.cpp: 9757 (JSC::ScriptSampleRecord::sample): 9758 (JSC::SamplingTool::doRun): 9759 (JSC::SamplingTool::notifyOfScope): 9760 (JSC::compareScriptSampleRecords): 9761 (JSC::SamplingTool::dump): 9762 * bytecode/SamplingTool.h: 9763 (JSC::ScriptSampleRecord::ScriptSampleRecord): 9764 (JSC::ScriptSampleRecord::~ScriptSampleRecord): 9765 (JSC::SamplingTool::SamplingTool): 9766 * bytecompiler/BytecodeGenerator.cpp: 9767 (JSC::BytecodeGenerator::BytecodeGenerator): 9768 (JSC::BytecodeGenerator::emitNewFunction): 9769 (JSC::BytecodeGenerator::emitNewFunctionExpression): 9770 * bytecompiler/BytecodeGenerator.h: 9771 (JSC::BytecodeGenerator::makeFunction): 9772 * debugger/Debugger.cpp: 9773 (JSC::evaluateInGlobalCallFrame): 9774 * debugger/DebuggerCallFrame.cpp: 9775 (JSC::DebuggerCallFrame::evaluate): 9776 * parser/Nodes.cpp: 9777 (JSC::ScopeNode::ScopeNode): 9778 * runtime/Completion.cpp: 9779 (JSC::checkSyntax): 9780 (JSC::evaluate): 9781 * runtime/Executable.cpp: 9782 (JSC::FunctionExecutable::fromGlobalCode): 9783 * runtime/Executable.h: 9784 (JSC::ScriptExecutable::ScriptExecutable): 9785 (JSC::EvalExecutable::EvalExecutable): 9786 (JSC::EvalExecutable::create): 9787 (JSC::ProgramExecutable::ProgramExecutable): 9788 (JSC::FunctionExecutable::create): 9789 (JSC::FunctionExecutable::FunctionExecutable): 9790 * runtime/JSGlobalObjectFunctions.cpp: 9791 (JSC::globalFuncEval): 9792 97932009-09-22 Darin Adler <darin@apple.com> 9794 9795 Reviewed by Sam Weinig. 9796 9797 * wtf/Forward.h: Added PassOwnPtr. 9798 97992009-09-22 Yaar Schnitman <yaar@chromium.org> 9800 9801 Reviewed by David Levin. 9802 9803 Ported chromium.org's javascriptcore.gyp for the webkit chromium port. 9804 9805 https://bugs.webkit.org/show_bug.cgi?id=29617 9806 9807 * JavaScriptCore.gyp/JavaScriptCore.gyp: Added. 9808 98092009-09-22 Thiago Macieira <thiago.macieira@nokia.com> 9810 9811 Reviewed by Simon Hausmann. 9812 9813 Fix compilation with WINSCW: no varargs macros 9814 9815 Disable variadic arguments for WINSCW just like we do 9816 for MSVC7. 9817 9818 * wtf/Assertions.h: 9819 98202009-09-22 Kent Hansen <khansen@trolltech.com> 9821 9822 Reviewed by Simon Hausmann. 9823 9824 Disable variadic macros on MSVC7. 9825 9826 This was originally added in r26589 but not extended 9827 when LOG_DISABLED/ASSERT_DISABLED was introduced. 9828 9829 * wtf/Assertions.h: 9830 98312009-09-22 Simon Hausmann <simon.hausmann@nokia.com> 9832 9833 Unreviewed build fix for Windows CE < 5 9834 9835 Define WINCEBASIC to disable the IsDebuggerPresent() code in 9836 wtf/Assertions.cpp. 9837 9838 * JavaScriptCore.pri: 9839 98402009-09-22 Joerg Bornemann <joerg.bornemann@nokia.com> 9841 9842 Reviewed by Simon Hausmann. 9843 9844 Fix major memory leak in JavaScriptCore RegisterFile on Windows CE 9845 9846 https://bugs.webkit.org/show_bug.cgi?id=29367 9847 9848 On Widows CE we must decommit all committed pages before we release 9849 them. See VirtualFree documentation. 9850 Desktop Windows behaves much smoother in this situation. 9851 9852 * interpreter/RegisterFile.cpp: 9853 (JSC::RegisterFile::~RegisterFile): 9854 98552009-09-21 Greg Bolsinga <bolsinga@apple.com> 9856 9857 Reviewed by Simon Fraser & Sam Weinig. 9858 9859 Add ENABLE(ORIENTATION_EVENTS) 9860 https://bugs.webkit.org/show_bug.cgi?id=29508 9861 9862 * wtf/Platform.h: Also sort PLATFORM(IPHONE) #defines. 9863 98642009-09-21 Jedrzej Nowacki <jedrzej.nowacki@nokia.com> 9865 9866 Reviewed by Eric Seidel. 9867 9868 [Fix] SourceCode's uninitialized member 9869 9870 Potential source of crashes and bugs was fixed. Default constructor 9871 didn't initialized m_provider member. 9872 9873 https://bugs.webkit.org/show_bug.cgi?id=29364 9874 9875 * parser/SourceCode.h: 9876 (JSC::SourceCode::SourceCode): 9877 98782009-09-21 Oliver Hunt <oliver@apple.com> 9879 9880 Reviewed by Geoff Garen. 9881 9882 REGRESSION (r48582): Crash in StructureStubInfo::initPutByIdTransition when reloading trac.webkit.org 9883 https://bugs.webkit.org/show_bug.cgi?id=29599 9884 9885 It is unsafe to attempt to cache new property transitions on 9886 dictionaries of any type. 9887 9888 * interpreter/Interpreter.cpp: 9889 (JSC::Interpreter::tryCachePutByID): 9890 * jit/JITStubs.cpp: 9891 (JSC::JITThunks::tryCachePutByID): 9892 98932009-09-21 Oliver Hunt <oliver@apple.com> 9894 9895 RS=Maciej Stachowiak. 9896 9897 Re-land SNES fix with corrected assertion. 9898 9899 * interpreter/Interpreter.cpp: 9900 (JSC::Interpreter::resolveGlobal): 9901 (JSC::Interpreter::tryCachePutByID): 9902 (JSC::Interpreter::tryCacheGetByID): 9903 * jit/JITStubs.cpp: 9904 (JSC::JITThunks::tryCachePutByID): 9905 (JSC::JITThunks::tryCacheGetByID): 9906 (JSC::DEFINE_STUB_FUNCTION): 9907 * runtime/BatchedTransitionOptimizer.h: 9908 (JSC::BatchedTransitionOptimizer::BatchedTransitionOptimizer): 9909 * runtime/JSObject.cpp: 9910 (JSC::JSObject::removeDirect): 9911 * runtime/Structure.cpp: 9912 (JSC::Structure::Structure): 9913 (JSC::Structure::getEnumerablePropertyNames): 9914 (JSC::Structure::despecifyDictionaryFunction): 9915 (JSC::Structure::addPropertyTransitionToExistingStructure): 9916 (JSC::Structure::addPropertyTransition): 9917 (JSC::Structure::removePropertyTransition): 9918 (JSC::Structure::toDictionaryTransition): 9919 (JSC::Structure::toCacheableDictionaryTransition): 9920 (JSC::Structure::toUncacheableDictionaryTransition): 9921 (JSC::Structure::fromDictionaryTransition): 9922 (JSC::Structure::removePropertyWithoutTransition): 9923 * runtime/Structure.h: 9924 (JSC::Structure::isDictionary): 9925 (JSC::Structure::isUncacheableDictionary): 9926 (JSC::Structure::): 9927 * runtime/StructureChain.cpp: 9928 (JSC::StructureChain::isCacheable): 9929 99302009-09-21 Adam Roben <aroben@apple.com> 9931 9932 Revert r48573, as it caused many assertion failures 9933 9934 * interpreter/Interpreter.cpp: 9935 * jit/JITStubs.cpp: 9936 * runtime/BatchedTransitionOptimizer.h: 9937 * runtime/JSObject.cpp: 9938 * runtime/Structure.cpp: 9939 * runtime/Structure.h: 9940 * runtime/StructureChain.cpp: 9941 99422009-09-21 Gustavo Noronha Silva <gustavo.noronha@collabora.co.uk> 9943 9944 Unreviewed make dist build fix. Missing files. 9945 9946 * GNUmakefile.am: 9947 99482009-09-19 Gavin Barraclough <barraclough@apple.com> 9949 9950 Reviewed by Sam 'Cabin Boy' Weinig. 9951 9952 Fix stack alignment with ARM THUMB2 JIT. 9953 https://bugs.webkit.org/show_bug.cgi?id=29526 9954 9955 Stack is currently being decremented by 0x3c, bump this to 0x40 to make this a 9956 multiple of 16 bytes. 9957 9958 * jit/JITStubs.cpp: 9959 (JSC::JITThunks::JITThunks): 9960 * jit/JITStubs.h: 9961 99622009-09-20 Oliver Hunt <oliver@apple.com> 9963 9964 Reviewed by Maciej Stachowiak. 9965 9966 SNES is too slow 9967 https://bugs.webkit.org/show_bug.cgi?id=29534 9968 9969 The problem was that the emulator used multiple classes with 9970 more properties than our dictionary cutoff allowed, this resulted 9971 in more or less all critical logic inside the emulator requiring 9972 uncached property access. 9973 9974 Rather than simply bumping the dictionary cutoff, this patch 9975 recognises that there are two ways to create a "dictionary" 9976 structure. Either by adding a large number of properties, or 9977 by removing a property. In the case of adding properties we 9978 know all the existing properties will maintain their existing 9979 offsets, so we could cache access to those properties, if we 9980 know they won't be removed. 9981 9982 To make this possible, this patch adds the logic required to 9983 distinguish a dictionary created by addition from one created 9984 by removal. With this logic in place we can now cache access 9985 to objects with large numbers of properties. 9986 9987 SNES performance improved by more than 6x. 9988 9989 * interpreter/Interpreter.cpp: 9990 (JSC::Interpreter::resolveGlobal): 9991 (JSC::Interpreter::tryCachePutByID): 9992 (JSC::Interpreter::tryCacheGetByID): 9993 * jit/JITStubs.cpp: 9994 (JSC::JITThunks::tryCachePutByID): 9995 (JSC::JITThunks::tryCacheGetByID): 9996 (JSC::DEFINE_STUB_FUNCTION): 9997 * runtime/BatchedTransitionOptimizer.h: 9998 (JSC::BatchedTransitionOptimizer::BatchedTransitionOptimizer): 9999 * runtime/JSObject.cpp: 10000 (JSC::JSObject::removeDirect): 10001 * runtime/Structure.cpp: 10002 (JSC::Structure::Structure): 10003 (JSC::Structure::getEnumerablePropertyNames): 10004 (JSC::Structure::despecifyDictionaryFunction): 10005 (JSC::Structure::addPropertyTransitionToExistingStructure): 10006 (JSC::Structure::addPropertyTransition): 10007 (JSC::Structure::removePropertyTransition): 10008 (JSC::Structure::toDictionaryTransition): 10009 (JSC::Structure::toCacheableDictionaryTransition): 10010 (JSC::Structure::toUncacheableDictionaryTransition): 10011 (JSC::Structure::fromDictionaryTransition): 10012 (JSC::Structure::removePropertyWithoutTransition): 10013 * runtime/Structure.h: 10014 (JSC::Structure::isDictionary): 10015 (JSC::Structure::isUncacheableDictionary): 10016 (JSC::Structure::): 10017 * runtime/StructureChain.cpp: 10018 (JSC::StructureChain::isCacheable): 10019 100202009-09-19 Oliver Hunt <oliver@apple.com> 10021 10022 Reviewed by Maciej Stachowiak. 10023 10024 Implement ES5 Object.create function 10025 https://bugs.webkit.org/show_bug.cgi?id=29524 10026 10027 Implement Object.create. Very simple patch, effectively Object.defineProperties 10028 only creating the target object itself. 10029 10030 * runtime/CommonIdentifiers.h: 10031 * runtime/ObjectConstructor.cpp: 10032 (JSC::ObjectConstructor::ObjectConstructor): 10033 (JSC::objectConstructorCreate): 10034 100352009-09-19 Dan Bernstein <mitz@apple.com> 10036 10037 Fix clean debug builds. 10038 10039 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 10040 100412009-09-19 Joerg Bornemann <joerg.bornemann@nokia.com> 10042 10043 Reviewed by George Staikos. 10044 10045 QtWebKit Windows CE compile fix 10046 10047 https://bugs.webkit.org/show_bug.cgi?id=29379 10048 10049 There is no _aligned_alloc or _aligned_free on Windows CE. 10050 We just use the Windows code that was there before and use VirtualAlloc. 10051 But that also means that the BLOCK_SIZE must be 64K as this function 10052 allocates on 64K boundaries. 10053 10054 * runtime/Collector.cpp: 10055 (JSC::Heap::allocateBlock): 10056 (JSC::Heap::freeBlock): 10057 * runtime/Collector.h: 10058 100592009-09-19 Oliver Hunt <oliver@apple.com> 10060 10061 Reviewed by Sam Weinig. 10062 10063 Implement ES5 Object.defineProperties function 10064 https://bugs.webkit.org/show_bug.cgi?id=29522 10065 10066 Implement Object.defineProperties. Fairly simple patch, simply makes use of 10067 existing functionality used for defineProperty. 10068 10069 * runtime/CommonIdentifiers.h: 10070 * runtime/ObjectConstructor.cpp: 10071 (JSC::ObjectConstructor::ObjectConstructor): 10072 (JSC::defineProperties): 10073 (JSC::objectConstructorDefineProperties): 10074 100752009-09-19 Oliver Hunt <oliver@apple.com> 10076 10077 Reviewed by NOBODY (Build fix). 10078 10079 Windows build fix part2 10080 10081 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 10082 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 10083 100842009-09-19 Oliver Hunt <oliver@apple.com> 10085 10086 Reviewed by NOBODY (Buildfix). 10087 10088 Windows build fix part 1. 10089 10090 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 10091 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 10092 100932009-09-18 Oliver Hunt <oliver@apple.com> 10094 10095 Reviewed by Geoff Garen. 10096 10097 Implement ES5 Object.defineProperty function 10098 https://bugs.webkit.org/show_bug.cgi?id=29503 10099 10100 Implement Object.defineProperty. This requires adding the API to 10101 ObjectConstructor, along with a helper function that implements the 10102 ES5 internal [[ToPropertyDescriptor]] function. It then adds 10103 JSObject::defineOwnProperty that implements the appropriate ES5 semantics. 10104 Currently defineOwnProperty uses a delete followed by a put to redefine 10105 attributes of a property, clearly this is less efficient than it could be 10106 but we can improve this if it needs to be possible in future. 10107 10108 * JavaScriptCore.exp: 10109 * debugger/DebuggerActivation.cpp: 10110 (JSC::DebuggerActivation::defineGetter): 10111 (JSC::DebuggerActivation::defineSetter): 10112 * debugger/DebuggerActivation.h: 10113 * interpreter/Interpreter.cpp: 10114 (JSC::Interpreter::privateExecute): 10115 * jit/JITStubs.cpp: 10116 Update defineGetter/Setter calls 10117 * runtime/CommonIdentifiers.h: 10118 * runtime/JSArray.cpp: 10119 (JSC::JSArray::getOwnPropertySlot): 10120 * runtime/JSGlobalObject.cpp: 10121 (JSC::JSGlobalObject::defineGetter): 10122 (JSC::JSGlobalObject::defineSetter): 10123 * runtime/JSGlobalObject.h: 10124 * runtime/JSObject.cpp: 10125 (JSC::JSObject::defineGetter): 10126 (JSC::JSObject::defineSetter): 10127 (JSC::putDescriptor): 10128 (JSC::JSObject::defineOwnProperty): 10129 * runtime/JSObject.h: 10130 * runtime/ObjectConstructor.cpp: 10131 (JSC::ObjectConstructor::ObjectConstructor): 10132 (JSC::objectConstructorGetOwnPropertyDescriptor): 10133 (JSC::toPropertyDescriptor): 10134 (JSC::objectConstructorDefineProperty): 10135 * runtime/ObjectPrototype.cpp: 10136 (JSC::objectProtoFuncDefineGetter): 10137 (JSC::objectProtoFuncDefineSetter): 10138 * runtime/PropertyDescriptor.cpp: 10139 (JSC::PropertyDescriptor::writable): 10140 (JSC::PropertyDescriptor::enumerable): 10141 (JSC::PropertyDescriptor::configurable): 10142 (JSC::PropertyDescriptor::isDataDescriptor): 10143 (JSC::PropertyDescriptor::isGenericDescriptor): 10144 (JSC::PropertyDescriptor::isAccessorDescriptor): 10145 (JSC::PropertyDescriptor::getter): 10146 (JSC::PropertyDescriptor::setter): 10147 (JSC::PropertyDescriptor::setDescriptor): 10148 (JSC::PropertyDescriptor::setAccessorDescriptor): 10149 (JSC::PropertyDescriptor::setWritable): 10150 (JSC::PropertyDescriptor::setEnumerable): 10151 (JSC::PropertyDescriptor::setConfigurable): 10152 (JSC::PropertyDescriptor::setSetter): 10153 (JSC::PropertyDescriptor::setGetter): 10154 (JSC::PropertyDescriptor::equalTo): 10155 (JSC::PropertyDescriptor::attributesEqual): 10156 (JSC::PropertyDescriptor::attributesWithOverride): 10157 * runtime/PropertyDescriptor.h: 10158 (JSC::PropertyDescriptor::PropertyDescriptor): 10159 (JSC::PropertyDescriptor::value): 10160 (JSC::PropertyDescriptor::setValue): 10161 (JSC::PropertyDescriptor::isEmpty): 10162 (JSC::PropertyDescriptor::writablePresent): 10163 (JSC::PropertyDescriptor::enumerablePresent): 10164 (JSC::PropertyDescriptor::configurablePresent): 10165 (JSC::PropertyDescriptor::setterPresent): 10166 (JSC::PropertyDescriptor::getterPresent): 10167 (JSC::PropertyDescriptor::operator==): 10168 (JSC::PropertyDescriptor::): 10169 101702009-09-18 Gabor Loki <loki@inf.u-szeged.hu> 10171 10172 Reviewed by Gavin Barraclough. 10173 10174 Build fix to enable ARM_THUMB2 on Linux 10175 https://bugs.webkit.org/show_bug.cgi?id= 10176 10177 * jit/ExecutableAllocator.h: 10178 (JSC::ExecutableAllocator::cacheFlush): 10179 * jit/JITStubs.cpp: 10180 * wtf/Platform.h: 10181 101822009-09-18 Gabor Loki <loki@inf.u-szeged.hu> 10183 10184 Reviewed by Gavin Barraclough. 10185 10186 Defines two pseudo-platforms for ARM and Thumb-2 instruction set. 10187 https://bugs.webkit.org/show_bug.cgi?id=29122 10188 10189 Introduces WTF_PLATFORM_ARM_TRADITIONAL and WTF_PLATFORM_ARM_THUMB2 10190 macros on ARM platforms. The PLATFORM(ARM_THUMB2) should be used 10191 when Thumb-2 instruction set is the required target. The 10192 PLATFORM(ARM_TRADITIONAL) is for generic ARM instruction set. In 10193 case where the code is common the PLATFORM(ARM) have to be used. 10194 10195 * assembler/ARMAssembler.cpp: 10196 * assembler/ARMAssembler.h: 10197 * assembler/ARMv7Assembler.h: 10198 * assembler/MacroAssembler.h: 10199 * assembler/MacroAssemblerARM.cpp: 10200 * assembler/MacroAssemblerARM.h: 10201 * assembler/MacroAssemblerCodeRef.h: 10202 (JSC::MacroAssemblerCodePtr::MacroAssemblerCodePtr): 10203 * jit/ExecutableAllocator.h: 10204 * jit/JIT.h: 10205 * jit/JITInlineMethods.h: 10206 (JSC::JIT::beginUninterruptedSequence): 10207 (JSC::JIT::preserveReturnAddressAfterCall): 10208 (JSC::JIT::restoreReturnAddressBeforeReturn): 10209 (JSC::JIT::restoreArgumentReference): 10210 (JSC::JIT::restoreArgumentReferenceForTrampoline): 10211 * jit/JITOpcodes.cpp: 10212 * jit/JITStubs.cpp: 10213 (JSC::JITThunks::JITThunks): 10214 * jit/JITStubs.h: 10215 * wtf/Platform.h: 10216 * yarr/RegexJIT.cpp: 10217 (JSC::Yarr::RegexGenerator::generateEnter): 10218 102192009-09-18 Joerg Bornemann <joerg.bornemann@nokia.com> 10220 10221 Reviewed by Simon Hausmann. 10222 10223 Fix the Qt/Windows CE build. 10224 10225 * JavaScriptCore.pri: Build the ce_time.cpp functions from 10226 within Qt externally. 10227 * wtf/DateMath.cpp: Removed unnecessary Qt #ifdef, for the 10228 Qt build these functions are no external, too. 10229 102302009-09-17 Janne Koskinen <janne.p.koskinen@digia.com> 10231 10232 Reviewed by Simon Hausmann. 10233 10234 Symbian/WINSCW build fox. 10235 10236 Repeat Q_OS_WIN wchar_t hack for WINSCW, similar to 10237 revision 24774. 10238 10239 WINSCW defines wchar_t, thus UChar has to be wchar_t 10240 10241 * wtf/unicode/qt4/UnicodeQt4.h: 10242 102432009-09-17 Janne Koskinen <janne.p.koskinen@digia.com> 10244 10245 Reviewed by Simon Hausmann. 10246 10247 Symbian/WINSCW build fix. 10248 10249 https://bugs.webkit.org/show_bug.cgi?id=29186 10250 10251 WINSCW Template specialisation name in declaration must the be the same as in implementation. 10252 10253 * runtime/LiteralParser.h: 10254 102552009-09-15 Norbert Leser <norbert.leser@nokia.com> 10256 10257 Reviewed by Darin Adler. 10258 10259 https://bugs.webkit.org/show_bug.cgi?id=27060 10260 10261 Symbian compiler for emulator target (WINSCW) fails with 10262 "illegal operand" for m_attributesInPrevious in structure.ccp 10263 (when calling make_pair functions). 10264 This error is apparently due to the compiler not properly 10265 resolving the unsigned type of the declared bitfield. 10266 10267 Initial patch explicitly casted m_attributesInPrevious 10268 to unsigned, but since bitfield optimization is not critical for 10269 the emulator target, this conditional change in header file 10270 appears to be least intrusive. 10271 10272 * runtime/Structure.h: 10273 102742009-09-16 Gabor Loki <loki@inf.u-szeged.hu> 10275 10276 Reviewed by Darin Adler. 10277 10278 Fix GCC warnings on ARM_THUMB2 platform 10279 10280 * assembler/ARMv7Assembler.h: 10281 (JSC::ARMThumbImmediate::countLeadingZerosPartial): 10282 * assembler/MacroAssemblerARMv7.h: 10283 (JSC::MacroAssemblerARMv7::branchTruncateDoubleToInt32): 10284 (JSC::MacroAssemblerARMv7::moveFixedWidthEncoding): 10285 102862009-09-16 Greg Bolsinga <bolsinga@apple.com> 10287 10288 Add ENABLE(INSPECTOR) 10289 https://bugs.webkit.org/show_bug.cgi?id=29260 10290 10291 Reviewed by David Kilzer. 10292 10293 * wtf/Platform.h: 10294 102952009-09-16 Greg Bolsinga <bolsinga@apple.com> 10296 10297 Add ENABLE(CONTEXT_MENUS) 10298 https://bugs.webkit.org/show_bug.cgi?id=29225 10299 10300 Reviewed by David Kilzer. 10301 10302 * wtf/Platform.h: 10303 103042009-09-16 Benjamin C Meyer <benjamin.meyer@torchmobile.com> 10305 10306 Reviewed by Eric Seidel. 10307 10308 The webkit stdint and stdbool headers exists because 10309 the compiler MSVC doesn't include them. The check 10310 should not check for PLATFORM(WIN_OS) but for MSVC. 10311 10312 * os-win32/stdbool.h: 10313 * os-win32/stdint.h: 10314 103152009-09-16 Greg Bolsinga <bolsinga@apple.com> 10316 10317 Add ENABLE(DRAG_SUPPORT) 10318 https://bugs.webkit.org/show_bug.cgi?id=29233 10319 10320 Reviewed by David Kilzer. 10321 10322 * wtf/Platform.h: 10323 103242009-09-16 Kevin Ollivier <kevino@theolliviers.com> 10325 10326 waf build fix after flag was moved to correct place. 10327 10328 * wscript: 10329 103302009-09-16 Tor Arne Vestbø <tor.arne.vestbo@nokia.com> 10331 10332 Reviewed by Simon Hausmann. 10333 10334 [Qt] Build fix for 64-bit Qt on Mac OS X 10335 10336 * wtf/Platform.h: Use JSVALUE64 on DARWIN, not only on MAC 10337 103382009-09-16 Zoltan Herczeg <zherczeg@inf.u-szeged.hu> 10339 10340 Reviewed by Simon Hausmann. 10341 10342 [Qt] Fix wtf/ThreadSpecific.h under Qt to free thread local objects. 10343 https://bugs.webkit.org/show_bug.cgi?id=29295 10344 10345 This is an important fix when JavaScript workers are in use, since 10346 unfreed ThreadGlobalDatas leak a big amount of memory (50-100k each). 10347 QThreadStorage calls the destructor of a given object, which is the 10348 ThreadSpecific::Data. Unlike pthread, Qt is object oriented, and does 10349 not support the calling of a static utility function when the thread 10350 is about to close. In this patch we call the ThreadSpecific::destroy() 10351 utility function from the destructor of ThreadSpecific::Data. Moreover, 10352 since Qt resets all thread local values to 0 before the calling of the 10353 appropriate destructors, we set back the pointer to its original value. 10354 This is necessary because the get() method of the ThreadSpecific 10355 object may be called during the exuction of the destructor. 10356 10357 * wtf/ThreadSpecific.h: 10358 (WTF::ThreadSpecific::Data::~Data): 10359 (WTF::::~ThreadSpecific): 10360 (WTF::::set): 10361 (WTF::::destroy): 10362 103632009-09-10 Oliver Hunt <oliver@apple.com> 10364 10365 Reviewed by Geoff Garen. 10366 10367 Allow anonymous storage inside JSObject 10368 https://bugs.webkit.org/show_bug.cgi?id=29168 10369 10370 Add the concept of anonymous slots to Structures so that it is 10371 possible to store references to values that need marking in the 10372 standard JSObject storage buffer. This allows us to reduce the 10373 malloc overhead of some objects (by allowing them to store JS 10374 values in the inline storage of the object) and reduce the 10375 dependence of custom mark functions (if all an objects children 10376 are in the standard object property storage there's no need to 10377 mark them manually). 10378 10379 * JavaScriptCore.exp: 10380 * runtime/JSObject.h: 10381 (JSC::JSObject::putAnonymousValue): 10382 (JSC::JSObject::getAnonymousValue): 10383 (JSC::JSObject::addAnonymousSlots): 10384 * runtime/JSWrapperObject.h: 10385 (JSC::JSWrapperObject::createStructure): 10386 (JSC::JSWrapperObject::JSWrapperObject): 10387 (JSC::JSWrapperObject::setInternalValue): 10388 * runtime/PropertyMapHashTable.h: 10389 * runtime/Structure.cpp: 10390 (JSC::Structure::~Structure): 10391 (JSC::Structure::materializePropertyMap): 10392 (JSC::Structure::addAnonymousSlotsTransition): 10393 (JSC::Structure::copyPropertyTable): 10394 (JSC::Structure::put): 10395 (JSC::Structure::rehashPropertyMapHashTable): 10396 * runtime/Structure.h: 10397 (JSC::Structure::propertyStorageSize): 10398 (JSC::StructureTransitionTable::reifySingleTransition): 10399 * runtime/StructureTransitionTable.h: 10400 (JSC::StructureTransitionTable::TransitionTable::addSlotTransition): 10401 (JSC::StructureTransitionTable::TransitionTable::removeSlotTransition): 10402 (JSC::StructureTransitionTable::TransitionTable::getSlotTransition): 10403 (JSC::StructureTransitionTable::getAnonymousSlotTransition): 10404 (JSC::StructureTransitionTable::addAnonymousSlotTransition): 10405 (JSC::StructureTransitionTable::removeAnonymousSlotTransition): 10406 104072009-09-15 Alex Milowski <alex@milowski.com> 10408 10409 Reviewed by Tor Arne Vestbø. 10410 10411 Added the ENABLE_MATHML define to the features 10412 10413 * Configurations/FeatureDefines.xcconfig: 10414 104152009-09-15 Csaba Osztrogonac <oszi@inf.u-szeged.hu> 10416 10417 Reviewed by Tor Arne Vestbø. 10418 10419 [Qt] Build fix for windows. 10420 10421 After http://trac.webkit.org/changeset/47795 the MinGW build broke, 10422 because MinGW has __mingw_aligned_malloc instead of _aligned_malloc. 10423 10424 * runtime/Collector.cpp: 10425 (JSC::Heap::allocateBlock): MinGW case added. 10426 (JSC::Heap::freeBlock): MinGW case added. 10427 104282009-09-15 Csaba Osztrogonac <oszi@inf.u-szeged.hu> 10429 10430 Reviewed by Tor Arne Vestbø. 10431 10432 [Qt] Build fix for Windows/MinGW 10433 10434 https://bugs.webkit.org/show_bug.cgi?id=29268 10435 10436 * wtf/Platform.h: JSVALUE32_64 temporarily disabled on PLATFORM(WIN_OS) with COMPILER(MINGW) 10437 104382009-09-14 Gabor Loki <loki@inf.u-szeged.hu> 10439 10440 Reviewed by Gavin Barraclough. 10441 10442 Detect VFP at runtime in generic ARM port on Linux platform. 10443 https://bugs.webkit.org/show_bug.cgi?id=29076 10444 10445 * JavaScriptCore.pri: 10446 * assembler/MacroAssemblerARM.cpp: Added. 10447 (JSC::isVFPPresent): 10448 * assembler/MacroAssemblerARM.h: 10449 (JSC::MacroAssemblerARM::supportsFloatingPoint): 10450 104512009-09-14 Csaba Osztrogonac <oszi@inf.u-szeged.hu> 10452 10453 Reviewed by Tor Arne Vestbø. 10454 10455 [Qt] Build fix for windows build. 10456 10457 * JavaScriptCore.pri: Correct a logic error. 10458 * pcre/dftables: Add missing paranthesis for tmpdir function. 10459 104602009-09-12 Oliver Hunt <oliver@apple.com> 10461 10462 Reviewed by NOBODY (Build fix). 10463 10464 Build fix for windows exports (again). 10465 10466 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 10467 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 10468 104692009-09-12 Oliver Hunt <oliver@apple.com> 10470 10471 Reviewed by NOBODY (Build fix). 10472 10473 Build fix for windows exports. 10474 10475 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 10476 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 10477 104782009-09-12 Oliver Hunt <oliver@apple.com> 10479 10480 Reviewed by NOBODY (Build fix). 10481 10482 Correct fix for non-allinonefile builds 10483 10484 * runtime/ObjectConstructor.cpp: 10485 104862009-09-12 Oliver Hunt <oliver@apple.com> 10487 10488 Reviewed by NOBODY (Build fix). 10489 10490 Fix non-allinonefile builds 10491 10492 * runtime/ObjectConstructor.cpp: 10493 104942009-09-12 Oliver Hunt <oliver@apple.com> 10495 10496 Reviewed by Maciej Stachowiak. 10497 10498 [ES5] Implement Object.keys 10499 https://bugs.webkit.org/show_bug.cgi?id=29170 10500 10501 This patch basically requires two separate steps, the first is to split getPropertyNames 10502 into two functions -- getOwnPropertyNames and getPropertyNames, basically making them behave 10503 in the same way as getOwnPropertySlot and getPropertySlot. In essence getOwnPropertyNames 10504 produces the list of properties on an object excluding its prototype chain and getPropertyNames 10505 just iterates the the object and its prototype chain calling getOwnPropertyNames at each level. 10506 10507 * API/JSCallbackObject.h: 10508 * API/JSCallbackObjectFunctions.h: 10509 (JSC::::getOwnPropertyNames): 10510 * JavaScriptCore.exp: 10511 * debugger/DebuggerActivation.cpp: 10512 (JSC::DebuggerActivation::getOwnPropertyNames): 10513 * debugger/DebuggerActivation.h: 10514 * runtime/CommonIdentifiers.h: 10515 * runtime/JSArray.cpp: 10516 (JSC::JSArray::getOwnPropertyNames): 10517 * runtime/JSArray.h: 10518 * runtime/JSByteArray.cpp: 10519 (JSC::JSByteArray::getOwnPropertyNames): 10520 * runtime/JSByteArray.h: 10521 * runtime/JSNotAnObject.cpp: 10522 (JSC::JSNotAnObject::getOwnPropertyNames): 10523 * runtime/JSNotAnObject.h: 10524 * runtime/JSObject.cpp: 10525 (JSC::JSObject::getOwnPropertyNames): 10526 * runtime/JSObject.h: 10527 * runtime/JSVariableObject.cpp: 10528 (JSC::JSVariableObject::getOwnPropertyNames): 10529 * runtime/JSVariableObject.h: 10530 * runtime/ObjectConstructor.cpp: 10531 (JSC::ObjectConstructor::ObjectConstructor): 10532 (JSC::objectConstructorKeys): 10533 * runtime/RegExpMatchesArray.h: 10534 (JSC::RegExpMatchesArray::getOwnPropertyNames): 10535 * runtime/StringObject.cpp: 10536 (JSC::StringObject::getOwnPropertyNames): 10537 * runtime/StringObject.h: 10538 * runtime/Structure.cpp: 10539 (JSC::Structure::getOwnEnumerablePropertyNames): 10540 (JSC::Structure::getEnumerablePropertyNames): 10541 * runtime/Structure.h: 10542 105432009-09-11 Oliver Hunt <oliver@apple.com> 10544 10545 Reviewed by Sam Weinig. 10546 10547 getPropertyNames caching is invalid when the prototype chain contains objects with custom getPropertyNames 10548 https://bugs.webkit.org/show_bug.cgi?id=29214 10549 10550 Add a flag to TypeInfo to indicate whether a type overrides getPropertyNames. 10551 This flag is used to make sure that caching of the property name data is safe. 10552 10553 * API/JSCallbackConstructor.h: 10554 (JSC::JSCallbackConstructor::createStructure): 10555 * debugger/DebuggerActivation.h: 10556 (JSC::DebuggerActivation::createStructure): 10557 * runtime/BooleanObject.h: 10558 (JSC::BooleanObject::createStructure): 10559 * runtime/DatePrototype.h: 10560 (JSC::DatePrototype::createStructure): 10561 * runtime/FunctionPrototype.h: 10562 (JSC::FunctionPrototype::createStructure): 10563 * runtime/JSONObject.h: 10564 (JSC::JSONObject::createStructure): 10565 * runtime/JSObject.h: 10566 (JSC::JSObject::createStructure): 10567 * runtime/JSTypeInfo.h: 10568 (JSC::TypeInfo::hasDefaultGetPropertyNames): 10569 * runtime/JSVariableObject.h: 10570 (JSC::JSVariableObject::createStructure): 10571 * runtime/JSWrapperObject.h: 10572 (JSC::JSWrapperObject::createStructure): 10573 * runtime/MathObject.h: 10574 (JSC::MathObject::createStructure): 10575 * runtime/NumberConstructor.h: 10576 (JSC::NumberConstructor::createStructure): 10577 * runtime/NumberObject.h: 10578 (JSC::NumberObject::createStructure): 10579 * runtime/RegExpConstructor.h: 10580 (JSC::RegExpConstructor::createStructure): 10581 * runtime/RegExpObject.h: 10582 (JSC::RegExpObject::createStructure): 10583 * runtime/StructureChain.cpp: 10584 (JSC::StructureChain::isCacheable): 10585 105862009-09-11 Alexey Proskuryakov <ap@webkit.org> 10587 10588 Reviewed by Geoff Garen. 10589 10590 https://bugs.webkit.org/show_bug.cgi?id=29207 10591 Add checks for using WebCore JS context on secondary threads 10592 10593 * runtime/JSGlobalData.cpp: (JSC::JSGlobalData::JSGlobalData): 10594 * runtime/JSGlobalData.h: 10595 Added a new mainThreadOnly flag that WebCore would set. 10596 10597 * runtime/Collector.cpp: (JSC::Heap::registerThread): JSC API methods always call this, 10598 so this is a good place to check that the API isn't used form a wrong thread. 10599 106002009-09-11 Jocelyn Turcotte <jocelyn.turcotte@nokia.com> 10601 10602 Reviewed by Simon Hausmann. 10603 10604 Compiling JavaScriptCore on sparc 64 with gcc fails. 10605 10606 ThreadSafeShared uses the atomic __gnu_cxx::__exchange_and_add with an int, 10607 however on sparc 64 the _Atomic_word argument is typedefed to long (8 bytes). 10608 10609 The patch disables WTF_USE_LOCKFREE_THREADSAFESHARED in ThreadSafeShared to use 10610 a mutex instead when compiling for sparc 64 with gcc. 10611 10612 https://bugs.webkit.org/show_bug.cgi?id=29175 10613 10614 * wtf/Platform.h: 10615 __sparc64__ is not defined on all OS. 10616 Uses instead: __sparc__ && __arch64__ || __sparcv9 10617 * wtf/Threading.h: 10618 106192009-09-11 Prasanth Ullattil <prasanth.ullattil@nokia.com> 10620 10621 Reviewed by Simon Hausmann. 10622 10623 Fix compile error on Windows7(64Bit) with latest SDK. 10624 10625 Added the missing include file. 10626 10627 * runtime/UString.cpp: 10628 106292009-09-11 Joerg Bornemann <joerg.bornemann@trolltech.com> 10630 10631 Reviewed by Simon Hausmann. 10632 10633 Qt/Windows CE compile fix, include the executable allocator and 10634 markstack implementation in the windows build. 10635 10636 * JavaScriptCore.pri: 10637 106382009-09-08 John Abd-El-Malek <jam@chromium.org> 10639 10640 Reviewed by Dimitri Glazkov. 10641 10642 Remove unneeded define for ActiveX. 10643 https://bugs.webkit.org/show_bug.cgi?id=29054 10644 10645 * wtf/Platform.h: 10646 106472009-09-10 Mark Rowe <mrowe@apple.com> 10648 10649 Rubber-stamped by Sam Weinig. 10650 10651 Update JavaScriptCore and WebKit's FeatureDefines.xcconfig so that they are in sync with WebCore as they need to be. 10652 10653 * Configurations/FeatureDefines.xcconfig: 10654 106552009-09-10 Fumitoshi Ukai <ukai@chromium.org> 10656 10657 Reviewed by Alexey Proskuryakov. 10658 10659 Export WTF::tryFastMalloc used in WebSocketChannel. 10660 https://bugs.webkit.org/show_bug.cgi?id=28038 10661 10662 * JavaScriptCore.exp: 10663 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 10664 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 10665 106662009-09-10 Oliver Hunt <oliver@apple.com> 10667 10668 Reviewed by NOBODY (Build fix). 10669 10670 Make StructureTransitionTable use an enum for the PtrAndFlags member 10671 used for the single transition slot optimisation. 10672 10673 * runtime/StructureTransitionTable.h: 10674 (JSC::StructureTransitionTable::StructureTransitionTable): 10675 (JSC::StructureTransitionTable::usingSingleTransitionSlot): 10676 (JSC::StructureTransitionTable::): 10677 106782009-09-10 Oliver Hunt <oliver@apple.com> 10679 10680 Reviewed by Geoff Garen. 10681 10682 Refactor StructureTransitionTable and Structure to unify handling of the single slot optimization 10683 https://bugs.webkit.org/show_bug.cgi?id=29141 10684 10685 Make StructureTransitionTable encapsulate the single transition slot optimization. 10686 10687 * runtime/Structure.cpp: 10688 (JSC::Structure::Structure): 10689 (JSC::Structure::~Structure): 10690 (JSC::Structure::addPropertyTransitionToExistingStructure): 10691 (JSC::Structure::addPropertyTransition): 10692 (JSC::Structure::addPropertyWithoutTransition): 10693 (JSC::Structure::removePropertyWithoutTransition): 10694 (JSC::Structure::hasTransition): 10695 * runtime/Structure.h: 10696 (JSC::StructureTransitionTable::contains): 10697 (JSC::StructureTransitionTable::get): 10698 (JSC::StructureTransitionTable::hasTransition): 10699 (JSC::StructureTransitionTable::reifySingleTransition): 10700 * runtime/StructureTransitionTable.h: 10701 (JSC::StructureTransitionTable::StructureTransitionTable): 10702 (JSC::StructureTransitionTable::~StructureTransitionTable): 10703 (JSC::StructureTransitionTable::remove): 10704 (JSC::StructureTransitionTable::add): 10705 (JSC::StructureTransitionTable::table): 10706 (JSC::StructureTransitionTable::singleTransition): 10707 (JSC::StructureTransitionTable::usingSingleTransitionSlot): 10708 (JSC::StructureTransitionTable::setSingleTransition): 10709 (JSC::StructureTransitionTable::setTransitionTable): 10710 (JSC::StructureTransitionTable::): 10711 * wtf/PtrAndFlags.h: 10712 (WTF::PtrAndFlags::PtrAndFlags): 10713 107142009-09-10 Zoltan Horvath <zoltan@webkit.org> 10715 10716 Reviewed by Darin Adler. 10717 10718 Implement fastDeleteSkippingDestructor for FastAllocBase and fastDeleteAllValues for HashSet 10719 https://bugs.webkit.org/show_bug.cgi?id=25930 10720 10721 FastAllocBase has been extended with fastDeleteSkippingDestructor function which 10722 releases memory without destructor call. fastDeleteAllValues has been implemented 10723 similar as deleteAllValues but it uses fastDelete function to release memory. 10724 10725 * wtf/FastAllocBase.h: 10726 (WTF::fastDeleteSkippingDestructor): 10727 * wtf/HashSet.h: 10728 (WTF::fastDeleteAllValues): 10729 107302009-09-10 Laszlo Gombos <laszlo.1.gombos@nokia.com> 10731 10732 Reviewed by Darin Adler. 10733 10734 ARM compiler does not understand GCC visibility attribute 10735 https://bugs.webkit.org/show_bug.cgi?id=29079 10736 10737 * API/JSBase.h: Make the test more specific to hit only 10738 the GCC compiler 10739 107402009-09-10 Adam Barth <abarth@webkit.org> 10741 10742 Unreviewed revert of the previous change. It broke the tests. 10743 10744 * wtf/dtoa.cpp: 10745 (WTF::dtoa): 10746 107472009-09-10 Ben Laurie <benl@google.com> 10748 10749 Reviewed by Adam Barth. 10750 10751 <https://bugs.webkit.org/show_bug.cgi?id=26836> 10752 10753 If dtoa was given a small buffer and the number was either infinite or 10754 NaN, then the buffer would be overflowed. 10755 10756 * wtf/dtoa.cpp: 10757 107582009-09-09 Laszlo Gombos <laszlo.1.gombos@nokia.com> 10759 10760 Reviewed by Darin Adler. 10761 10762 Change reinterpret_cast to static_cast in r48212. 10763 10764 * jit/ExecutableAllocator.h: 10765 (JSC::ExecutableAllocator::cacheFlush): 10766 107672009-09-09 Laszlo Gombos <laszlo.1.gombos@nokia.com> 10768 10769 Reviewed by Darin Adler. 10770 10771 Remove WTF_PLATFORM_FORCE_PACK as it is no longer used 10772 https://bugs.webkit.org/show_bug.cgi?id=29066 10773 10774 * wtf/Platform.h: 10775 107762009-09-09 Laszlo Gombos <laszlo.1.gombos@nokia.com> 10777 10778 Reviewed by Ariya Hidayat. 10779 10780 Implement flushing the instruction cache for Symbian 10781 https://bugs.webkit.org/show_bug.cgi?id=29075 10782 10783 * jit/ExecutableAllocator.h: 10784 (JSC::ExecutableAllocator::cacheFlush): Call IMB_Range to flush 10785 the instruction cache on Symbian 10786 107872009-09-09 Kent Hansen <khansen@trolltech.com> 10788 10789 Reviewed by Darin Adler. 10790 10791 https://bugs.webkit.org/show_bug.cgi?id=29024 10792 Make JavaScriptCore compile on platforms with case-insensitive file systems and typeinfo.h in STL 10793 10794 These platforms include Microsoft Visual Studio 2003, and Symbian with Metrowerks compiler. 10795 10796 * JavaScriptCore.gypi: 10797 * JavaScriptCore.xcodeproj/project.pbxproj: 10798 * runtime/JSTypeInfo.h: Copied from JavaScriptCore/runtime/TypeInfo.h. 10799 * runtime/Structure.h: 10800 * runtime/TypeInfo.h: Removed. 10801 108022009-09-08 Oliver Hunt <oliver@apple.com> 10803 10804 Reviewed by Maciej Stachowiak. 10805 10806 JSON.stringify(Date) loses the milliseconds information 10807 https://bugs.webkit.org/show_bug.cgi?id=29063 10808 10809 Make sure we include milliseconds in the output of toISOString. 10810 10811 * runtime/DatePrototype.cpp: 10812 (JSC::dateProtoFuncToISOString): 10813 108142009-09-08 Kevin Ollivier <kevino@theolliviers.com> 10815 10816 wx build fix, generate derived sources earlier in order to make sure 10817 they're found by the build system when generating the list of sources to build. 10818 10819 * wscript: 10820 108212009-09-08 Laszlo Gombos <laszlo.1.gombos@nokia.com> 10822 10823 Reviewed by Simon Hausmann. 10824 10825 Build fix when USE(LOCKFREE_THREADSAFESHARED) is not defined 10826 https://bugs.webkit.org/show_bug.cgi?id=29011 10827 10828 * wtf/Threading.h: Use LOCKFREE_THREADSAFESHARED guard for 10829 atomicIncrement and atomicDecrement 10830 108312009-09-07 Zoltan Horvath <zoltan@webkit.org> 10832 10833 Reviewed by Darin Adler. 10834 10835 Allow custom memory allocation control in Yarr's RegexInterpreter 10836 https://bugs.webkit.org/show_bug.cgi?id=29025 10837 10838 Inherits RegexInterpreter classes from FastAllocBase (bug #20422), which has 10839 been instantiated by 'new': 10840 10841 class ByteDisjunction 10842 -> instantiated in JavaScriptCore/yarr/RegexInterpreter.cpp:1462 10843 10844 struct BytecodePattern 10845 -> instantiated in JavaScriptCore/yarr/RegexInterpreter.cpp:1279 10846 10847 * yarr/RegexInterpreter.h: 10848 108492009-09-07 Drew Wilson <atwilson@google.com> 10850 10851 Reverting r48121 to fix Windows build errors. 10852 10853 * JavaScriptCore.exp: 10854 108552009-09-07 Drew Wilson <atwilson@google.com> 10856 10857 Reviewed by David Levin. 10858 10859 Enable SHARED_WORKERS by default 10860 https://bugs.webkit.org/show_bug.cgi?id=28959 10861 10862 * Configurations/FeatureDefines.xcconfig: 10863 108642009-09-07 Fumitoshi Ukai <ukai@chromium.org> 10865 10866 Reviewed by Alexey Proskuryakov. 10867 10868 Export WTF::tryFastMalloc used in WebSocketChannel. 10869 https://bugs.webkit.org/show_bug.cgi?id=28038 10870 10871 * JavaScriptCore.exp: 10872 108732009-09-04 Oliver Hunt <oliver@apple.com> 10874 10875 Reviewed by NOBODY (Build fix). 10876 10877 Fix windows export files 10878 10879 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 10880 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 10881 108822009-09-04 Oliver Hunt <oliver@apple.com> 10883 10884 Reviewed by Gavin Barraclough. 10885 10886 [[ToString]] conversion should use the actual toString function for String objects. 10887 10888 Remove incorrect specialisations of toString conversions on StringObject. 10889 10890 * JavaScriptCore.exp: 10891 * runtime/StringObject.cpp: 10892 * runtime/StringObject.h: 10893 108942009-09-04 Steve Falkenburg <sfalken@apple.com> 10895 10896 Windows build fix. 10897 10898 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: Add new export. 10899 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: Add new export. 10900 109012009-09-04 Steve Falkenburg <sfalken@apple.com> 10902 10903 Windows build fix. 10904 10905 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: Remove unneeded export. 10906 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: Remove unneeded export. 10907 109082009-09-04 Darin Adler <darin@apple.com> 10909 10910 Reviewed by Geoff Garen. 10911 10912 DateInstance object collected on ARM JIT (JSValue: WTF_USE_JSVALUE32) 10913 https://bugs.webkit.org/show_bug.cgi?id=28909 10914 10915 Part two. 10916 10917 Make some improvements to garbage collection code: 10918 10919 1) Create a runtime assertion that catches any classes that 10920 override markChildren but have the HasDefaultMark bit set. 10921 2) Remove checks of the mark bit outside the MarkStack::append 10922 function; they are redundant. 10923 3) Improve the efficiency of the asObject and asArray functions 10924 when called on JSCell* to avoid a round trip to JSValue. 10925 4) Make more callers use the checked asCell and asObject 10926 casting functions rather than unchecked casts. 10927 5) Removed the JSCell::marked function and other GC-related 10928 functions because these operations are no longer things that 10929 code other than the core GC code needs to do directly. Fixed 10930 callers that were calling them. 10931 10932 * runtime/Collector.cpp: 10933 (JSC::Heap::markConservatively): Removed unneeded call to MarkStack::drain. 10934 (JSC::Heap::markProtectedObjects): Removed unneeded check of the mark 10935 bit and call to MarkStack::drain. 10936 (JSC::Heap::collect): Removed unneeded checks of the mark bit and also 10937 changed call to SmallStrings::mark to call markChildren instead to match 10938 the rest of the objects. 10939 (JSC::typeName): Removed unneeded cast to JSObject*. 10940 10941 * runtime/JSArray.h: 10942 (JSC::asArray): Added an overload for JSCell* and changed the JSValue 10943 version to call it. Removed some unneeded casts. 10944 (JSC::JSArray::markChildrenDirect): Marked this function inline. It's in 10945 a header, and if not marked inline this could lead to linking problems. 10946 (JSC::MarkStack::markChildren): Added. This helper function is used by 10947 the drain function to avoid repating code. Also added the code here to 10948 check fro default mark violations in debug code. If a markChildren 10949 function adds something to the mark stack, but the type info claimed 10950 hasDefaultMark was true, then we will get an assertion now. Also fixed 10951 the assertion about the mark bit to use the Heap function directly 10952 because we don't have a JSCell::marked function any more. 10953 (JSC::MarkStack::drain): Changed a local variable from "v" to "value", 10954 and from "currentCell" to "cell". Changed to call markChildren in two 10955 places instead of repeating a chain of if statements twice. Changed 10956 code that reads and writes the mark bit to use Heap::isCellMarked and 10957 Heap::markCell so we can eliminate the JSCell::marked and 10958 JSCell::markCellDirect functions. 10959 10960 * runtime/JSCell.h: Removed JSCell's markCellDirect and marked member 10961 functions. Added a comment explaining that asCell should be deprecated 10962 in favor of the JSValue asCell member function. 10963 (JSC::MarkStack::append): Added the assertion that catches callers 10964 that have set the HasDefaultMark bit incorrectly. Changed 10965 code that reads and writes the mark bit to use Heap::isCellMarked and 10966 Heap::markCell so we can eliminate the JSCell::marked and 10967 JSCell::markCellDirect functions. Moved the overload of 10968 MarkStack::append for JSValue here so it can call through to the cell 10969 version. The old version had a copy of all the code instead, but that 10970 repeated the conversion from JSValue to JSCell* and the check for 10971 whether a value is a cell multiple times. 10972 (JSC::Structure::markAggregate): Moved this function here to avoid 10973 dependencies for Structure.h, since this calls MarkStack::append. 10974 10975 * runtime/JSObject.cpp: 10976 (JSC::JSObject::markChildren): Added code to clear 10977 m_isCheckingForDefaultMarkViolation so the marking done by JSObject 10978 doesn't trigger the assertion. 10979 10980 * runtime/JSValue.h: Moved some stray includes that were outside the 10981 header guard inside it. Not sure how that happened! Removed the 10982 GC-related member functions markChildren, hasChildren, marked, and 10983 markDirect. 10984 10985 * runtime/JSWrapperObject.h: Made markChildren private. 10986 (JSC::JSWrapperObject::createStructure): Added. Fixes a bug where the 10987 HasDefaultMark bit was set. 10988 10989 * runtime/MarkStack.h: Added m_isCheckingForDefaultMarkViolation and 10990 initialized it to false. Moved the append function body from here to 10991 JSCell.h. Added a declaration of a private markChildren function used 10992 inside the drain function. 10993 10994 * runtime/SmallStrings.cpp: 10995 (JSC::SmallStrings::markChildren): Changed the name and style of this 10996 function to match other functions. This allows us to share the normal 10997 mark stack code path. 10998 10999 * runtime/SmallStrings.h: Changed the name and interface of mark to 11000 the more-normal markChildren style. 11001 11002 * runtime/Structure.h: Moved the body of markAggregate into the 11003 JSCell.h to avoid a circular dependency with JSCell.h. 11004 110052009-09-04 Darin Adler <darin@apple.com> 11006 11007 Reviewed by Geoff Garen. 11008 11009 DateInstance object collected on ARM JIT (JSValue: WTF_USE_JSVALUE32) 11010 https://bugs.webkit.org/show_bug.cgi?id=28909 11011 11012 Part one. 11013 11014 Make some improvements to garbage collection code: 11015 11016 1) Fix the two classes that had the default mark bit set but 11017 should not. 11018 2) Remove checks of the mark bit outside the MarkStack::append 11019 function; they are redundant. 11020 3) Make more callers use the checked asCell and asObject 11021 casting functions rather than unchecked casts. 11022 4) Removed some GC-related functions because these operations are 11023 no longer things that code other than the core GC code needs 11024 to do directly. Fixed callers that were calling them. 11025 11026 * bytecode/CodeBlock.cpp: 11027 (JSC::CodeBlock::markAggregate): Removed unneeded check of the mark 11028 bit before calling MarkStack::append. 11029 11030 * interpreter/Register.h: Removed unneeded marked and markChildren 11031 functions. 11032 11033 * jit/JITStubs.cpp: 11034 (op_eq): Removed unneeded assertions, instead using checked casting 11035 functions such as asObject. 11036 11037 * runtime/ArgList.h: Added now-needed forward declaration of MarkStack. 11038 11039 * runtime/GetterSetter.cpp: 11040 (JSC::GetterSetter::markChildren): Remmoved unneeded check of the mark bit. 11041 11042 * runtime/GlobalEvalFunction.h: 11043 (JSC::GlobalEvalFunction::createStructure): Added. Fixes a bug where the 11044 HasDefaultMark bit was set. 11045 11046 * runtime/JSCell.cpp: 11047 (JSC::JSCell::getObject): Use asObject to avoid a direct static_cast. 11048 11049 * runtime/JSObject.h: 11050 (JSC::asObject): Added an overload for JSCell* and changed the JSValue 11051 version to call it. 11052 (JSC::JSValue::get): Use asObject to avoid a direct static_cast. 11053 11054 * runtime/JSWrapperObject.h: Made markChildren private. 11055 (JSC::JSWrapperObject::createStructure): Added. Fixes a bug where the 11056 HasDefaultMark bit was set. Later we may want to optimize this for 11057 wrapper types that never have cells in their internal values, but there 11058 is no measured performance regression in SunSpider or V8 doing this 11059 all the time. 11060 11061 * runtime/MarkStack.cpp: Tweaked formatting. 11062 110632009-09-04 Kevin Ollivier <kevino@theolliviers.com> 11064 11065 wx build fix. Switch USE_ defines over to the compiler so that they can be 11066 checked by files not including config.h (like WebCorePrefix.h). 11067 11068 * wtf/Platform.h: 11069 110702009-09-03 Yong Li <yong.li@torchmobile.com> 11071 11072 Reviewed by David Levin. 11073 11074 Remove unnecessary dependency on unistd.h 11075 https://bugs.webkit.org/show_bug.cgi?id=28962 11076 11077 * runtime/Completion.cpp: 11078 110792009-09-03 Fumitoshi Ukai <ukai@chromium.org> 11080 11081 Reviewed by Eric Seidel. 11082 11083 Add strnstr for Linux and Windows in StringExtras.h 11084 https://bugs.webkit.org/show_bug.cgi?id=28901 11085 11086 * wtf/StringExtras.h: 11087 (strnstr): 11088 110892009-09-03 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 11090 11091 Reviewed by Darin Adler. 11092 11093 Allow custom memory allocation control for JavaScriptCore's HashEntry class 11094 https://bugs.webkit.org/show_bug.cgi?id=27830 11095 11096 Inherits HashEntry class from FastAllocBase because it has been 11097 instantiated by 'new' JavaScriptCore/runtime/Lookup.cpp:32. 11098 11099 * runtime/Lookup.h: 11100 111012009-09-02 Gavin Barraclough <barraclough@apple.com> 11102 11103 Should crash if JIT code buffer allocation fails. 11104 11105 https://bugs.webkit.org/show_bug.cgi?id=28926 11106 <rdar://problem/7031922> 11107 11108 * jit/ExecutableAllocatorPosix.cpp: 11109 (JSC::ExecutablePool::systemAlloc): 11110 * jit/ExecutableAllocatorWin.cpp: 11111 (JSC::ExecutablePool::systemAlloc): 11112 111132009-09-02 Kevin Ollivier <kevino@theolliviers.com> 11114 11115 waf build fixes for Windows/MSVC. 11116 11117 * wscript: 11118 111192009-09-02 Kevin Ollivier <kevino@theolliviers.com> 11120 11121 Build fix for building on Windows. 11122 11123 * wtf/ThreadingPthreads.cpp: 11124 111252009-09-02 Norbert Leser <norbert.leser@nokia.com> 11126 11127 Reviewed by Eric Seidel. 11128 11129 Use fastMalloc when neither MMAP nor VIRTUALALLOC are enabled 11130 11131 RegisterFile constructor currently throws #error when both 11132 MMAP and VIRTUALALLOC conditions fail. 11133 On any platform that does not provide these features 11134 (for instance, Symbian), 11135 the fallback should be regular malloc (or fastMalloc). 11136 It is functionally equivalent in this case, even though it may 11137 have certain drawbacks such as lack of dynamic pre-allocation. 11138 11139 * interpreter/RegisterFile.cpp: 11140 (JSC::RegisterFile::~RegisterFile): 11141 * interpreter/RegisterFile.h: 11142 (JSC::RegisterFile::RegisterFile): 11143 111442009-08-31 Robert Agoston <Agoston.Robert@stud.u-szeged.hu> 11145 11146 Reviewed by Gavin Barraclough. 11147 11148 Fixed typo. 11149 https://bugs.webkit.org/show_bug.cgi?id=28691 11150 11151 * parser/Parser.h: 11152 (JSC::Parser::parse): 11153 111542009-08-27 Oliver Hunt <oliver@apple.com> 11155 11156 Reviewed by Maciej Stachowiak. 11157 11158 JSON Stringifier does not follow ES5 spec for handling of Number, String and Boolean objects 11159 https://bugs.webkit.org/show_bug.cgi?id=28797 11160 11161 Fixed unwrapBoxedPrimitive to do the right thing, which necessitated a couple of new exception 11162 checks, and corrected the logic in gap to correctly convert Number and String objects. 11163 11164 * runtime/JSONObject.cpp: 11165 (JSC::unwrapBoxedPrimitive): 11166 (JSC::gap): 11167 (JSC::Stringifier::Stringifier): 11168 (JSC::Stringifier::appendStringifiedValue): 11169 111702009-08-27 Oliver Hunt <oliver@apple.com> 11171 11172 Reviewed by Adam Roben. 11173 11174 JSON.stringify replacer array does not accept values that are not string primitives. 11175 https://bugs.webkit.org/show_bug.cgi?id=28788 11176 11177 Update the JSON stringifier to initialise its replacer array according to the most 11178 recent version of the spec. 11179 11180 * runtime/Identifier.h: 11181 (JSC::Identifier::from): 11182 * runtime/JSONObject.cpp: 11183 (JSC::Stringifier::Stringifier): 11184 111852009-08-27 Alexey Proskuryakov <ap@apple.com> 11186 11187 Reviewed by Oliver Hunt. 11188 11189 https://bugs.webkit.org/show_bug.cgi?id=28753 11190 <rdar://problem/7173448> Excessive number of threads (and a crash) 11191 11192 * wtf/Threading.h: (WTF::atomicIncrement): Changed atomicIncrement to match decrement 11193 and return the new value. Also added using directives for these functions, to match 11194 te rest of WTF. 11195 111962009-08-27 Brent Fulgham <bfulgham@webkit.org> 11197 11198 Reviewed by Adam Roben. 11199 11200 Link the testapi against CFLite when building the WinCairo port. 11201 11202 * JavaScriptCore.vcproj/testapi/testapi.vcproj: Add new Release_CFLite 11203 target. Update all targets to inherit from either the 11204 JavaScriptCF.vsprops (Apple target) or the JavaScriptCFLite.vsprops 11205 file (WinCairo target). 11206 * JavaScriptCore.vcproj/testapi/testapiCommon.vsprops: Remove 11207 input file CoreFoundation.lib. This is provided by either the 11208 JavaScriptCF.vsprops or JavaScriptCFLite.vsprops file. 11209 112102009-08-27 Steve Falkenburg <sfalken@apple.com> 11211 11212 Reviewed by Geoff Garen. 11213 11214 Fix Windows-specific crash due to missing memory clearing call. 11215 11216 * runtime/Collector.cpp: 11217 (JSC::Heap::allocateBlock): 11218 112192009-08-27 Brent Fulgham <bfulgham@webkit.org> 11220 11221 Build fix: JavaScriptCore_debug.def missing some exports. Apple 11222 Windows build does not use this file, so it was not noticed previously. 11223 11224 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 11225 112262009-08-27 Gavin Barraclough <barraclough@apple.com> 11227 11228 Reviewed by Oliver Hunt. 11229 11230 x86-64 GTK broken due to code offsets changing, pointers sometimes packed into immediates. 11231 https://bugs.webkit.org/show_bug.cgi?id=28317 11232 11233 Missed one, fix part II. 11234 11235 * assembler/MacroAssemblerX86Common.h: 11236 (JSC::MacroAssemblerX86Common::move): 11237 * assembler/X86Assembler.h: 11238 (JSC::CAN_SIGN_EXTEND_8_32): 11239 112402009-08-27 Oliver Hunt <oliver@apple.com> 11241 11242 Reviewed by Adam Roben. 11243 11244 JSON.stringify replacer array does not accept values that are not string primitives. 11245 https://bugs.webkit.org/show_bug.cgi?id=28788 11246 11247 Update the JSON stringifier to initialise its replacer array according to the most 11248 recent version of the spec. 11249 11250 * runtime/Identifier.h: 11251 (JSC::Identifier::from): 11252 * runtime/JSONObject.cpp: 11253 (JSC::Stringifier::Stringifier): 11254 112552009-08-27 Oliver Hunt <oliver@apple.com> 11256 11257 Reviewed by Alexey Proskuryakov. 11258 11259 JSON parser accepts trailing comma in array literals 11260 https://bugs.webkit.org/show_bug.cgi?id=28779 11261 11262 Update parser to correctly fail if there's a trailing comma. 11263 11264 * runtime/LiteralParser.cpp: 11265 (JSC::LiteralParser::parse): 11266 112672009-08-26 Oliver Hunt <oliver@apple.com> 11268 11269 Reviewed by Gavin Barraclough. 11270 11271 'this' in JSON.parse reviver is the global object 11272 https://bugs.webkit.org/show_bug.cgi?id=28752 11273 11274 This is a technically simple change, we merely update the code for calling 11275 the reviver function to pass the correct this object. Doing so however 11276 exposes the holder to arbitrary mutation by the reviver function so it is 11277 necessary for us to now guard all property accesses against the possibility 11278 of failure. 11279 11280 * runtime/JSArray.h: 11281 JSON needs to delete a property from the array, so we friend its 11282 Walker class so that we can make a non-virtual call to the arrays 11283 delete and getOwnPropertySlot methods. 11284 * runtime/JSONObject.cpp: 11285 (JSC::Walker::callReviver): 11286 We need to pass the correct this object 11287 (JSC::Walker::walk): 11288 Update calls to callReviver, and update property logic logic 11289 to correctly handle the holder being mutated by the reviver 11290 function. 11291 112922009-08-26 Alice Liu <alice.liu@apple.com> 11293 11294 Windows build fix: added some exported symbols 11295 11296 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 11297 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 11298 112992009-08-26 Geoffrey Garen <ggaren@apple.com> 11300 11301 Windows build fix: Removed some exported symbols that no longer exist. 11302 11303 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 11304 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 11305 113062009-08-26 Gavin Barraclough <barraclough@apple.com> 11307 11308 Reviewed by Olliejver Hunt. 11309 11310 x86-64 GTK broken due to code offsets changing, pointers sometimes packed into immediates. 11311 https://bugs.webkit.org/show_bug.cgi?id=28317 11312 11313 We rely on a slightly OS X specific behaviour, that x86-64 applications have a 4Gb zero page, 11314 so pointers are never representable as a 32-bit integer, and always have to be represented by 11315 a separate immediate load instruction, rather than within the immediate field of an arithmetic 11316 or memory operation. 11317 11318 We explicitly check for a couple of cases where a value might be representable in 32-bit, but 11319 these probably never kick in on Mac OS, and only kick in to hose GTK. Deleting these does not 11320 show a performance degradation on SunSpider. Remove. 11321 11322 * assembler/MacroAssemblerX86_64.h: 11323 (JSC::MacroAssemblerX86_64::storePtr): 11324 (JSC::MacroAssemblerX86_64::branchPtr): 11325 113262009-08-26 Geoffrey Garen <ggaren@apple.com> 11327 11328 Reviewed by Oliver Hunt. 11329 11330 A bit of Collector refatoring. 11331 11332 SunSpider says no change. v8 says 1.003x faster (1.02x faster on splay). 11333 11334 * JavaScriptCore.exp: 11335 11336 * runtime/JSCell.cpp: 11337 (JSC::JSCell::toPrimitive): 11338 (JSC::JSCell::getPrimitiveNumber): 11339 (JSC::JSCell::toBoolean): 11340 (JSC::JSCell::toNumber): 11341 (JSC::JSCell::toString): 11342 (JSC::JSCell::toObject): Removed pure virtual functions from 11343 JSCell, so the collector can construct one. This allowed 11344 me to remove a bunch of ASSERT_NOT_REACHED throughout the 11345 code, too. 11346 11347 * runtime/JSCell.h: 11348 (JSC::JSCell::JSCell): ditto 11349 (JSC::Heap::heap): Inlined this function because it's trivial. 11350 11351 * JavaScriptCore.exp: 11352 11353 * runtime/Collector.cpp: 11354 (JSC::Heap::destroy): 11355 (JSC::Heap::allocateBlock): 11356 (JSC::Heap::freeBlock): 11357 (JSC::Heap::freeBlocks): Renamed freeHeap to freeBlocks, since 11358 it doesn't actually free the Heap object. 11359 (JSC::Heap::heapAllocate): 11360 (JSC::Heap::sweep): 11361 * runtime/Collector.h: Refactored block allocation and destruction 11362 into helper functions. 11363 11364 * runtime/GetterSetter.cpp: 11365 * runtime/JSAPIValueWrapper.cpp: 11366 * runtime/JSPropertyNameIterator.cpp: Removed dummy implementations 11367 of pure virtual functions. (See above.) 11368 11369=== End re-roll-in of r47738:47740 with Windows crash fixed === 11370 113712009-08-26 Geoffrey Garen <ggaren@apple.com> 11372 11373 Build fix: start out with a 32-bit value to avoid a shortening warning. 11374 11375 * runtime/Collector.cpp: 11376 (JSC::Heap::sweep): 11377 113782009-08-24 Geoffrey Garen <ggaren@apple.com> 11379 11380 Reviewed by Oliver Hunt. 11381 11382 Substantially reduced VM thrash in the GC heap. 11383 11384 1.08x faster on v8 (1.60x faster on v8-splay). 11385 11386 1.40x faster on bench-alloc-nonretained. 11387 11388 1.90x faster on bench-alloc-retained. 11389 11390 SunSpider says no change. 11391 11392 * runtime/Collector.cpp: 11393 (JSC::Heap::heapAllocate): Fixed a long-standing bug: update a few local 11394 variables unconditionally after calling collect(), since they may be used 11395 even if we don't "goto scan". (In the bug I saw, usedBlocks got out of 11396 sync with heap.usedBlocks). 11397 (JSC::Heap::sweep): Keep enough free heap space to accomodate 11398 the number of objects we'll allocate before the next GC, plus 25%, for 11399 good measure. 11400 * runtime/Collector.h: Bumped the block size to 256k. This seems to give 11401 the best cache performance, and it prevents us from initiating lots of 11402 VM traffic to recover very small chunks of memory. 11403 11404=== Begin re-roll-in of r47738:47740 with Windows crash fixed === 11405 114062009-08-25 Drew Wilson <atwilson@google.com> 11407 11408 Reviewed by David Levin. 11409 11410 postMessage() spec now supports sending arrays of ports 11411 https://bugs.webkit.org/show_bug.cgi?id=26902 11412 11413 Added OwnPtr to VectorTraits so we can store OwnPtrs in Vectors. 11414 11415 * wtf/VectorTraits.h: 11416 114172009-08-26 Xan Lopez <xlopez@igalia.com> 11418 11419 Rubber-stamped by Gustavo Noronha. 11420 11421 Remove duplicated files from file list. 11422 11423 * GNUmakefile.am: 11424 114252009-08-26 Oliver Hunt <oliver@apple.com> 11426 11427 Reviewed by NOBODY (Build fix). 11428 11429 More export fixes. 11430 11431 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 11432 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 11433 114342009-08-26 Oliver Hunt <oliver@apple.com> 11435 11436 Reviewed by NOBODY (Build fix). 11437 11438 Hopefully fix all the exports from JSC on windows 11439 11440 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 11441 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 11442 114432009-08-26 Oliver Hunt <oliver@apple.com> 11444 11445 Reviewed by NOBODY (Build fixes). 11446 11447 Forgot I added files to JavaScriptCore. 11448 11449 * GNUmakefile.am: 11450 * JavaScriptCore.gypi: 11451 * JavaScriptCore.pri: 11452 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: 11453 * JavaScriptCoreSources.bkl: 11454 114552009-08-25 Oliver Hunt <oliver@apple.com> 11456 11457 Reviewed by Gavin Barraclough. 11458 11459 [ES5] Implement getOwnPropertyDescriptor 11460 https://bugs.webkit.org/show_bug.cgi?id=28724 11461 11462 Implement the core runtime support for getOwnPropertyDescriptor. 11463 This adds a virtual getOwnPropertyDescriptor method to every class 11464 that implements getOwnPropertySlot that shadows the behaviour of 11465 getOwnPropertySlot. The alternative would be to make getOwnPropertySlot 11466 (or PropertySlots in general) provide property attribute information, 11467 but quick testing showed this to be a regression. 11468 11469 * JavaScriptCore.exp: 11470 * JavaScriptCore.xcodeproj/project.pbxproj: 11471 * runtime/Arguments.cpp: 11472 (JSC::Arguments::getOwnPropertyDescriptor): 11473 * runtime/Arguments.h: 11474 * runtime/ArrayPrototype.cpp: 11475 (JSC::ArrayPrototype::getOwnPropertyDescriptor): 11476 * runtime/ArrayPrototype.h: 11477 * runtime/CommonIdentifiers.h: 11478 * runtime/DatePrototype.cpp: 11479 (JSC::DatePrototype::getOwnPropertyDescriptor): 11480 * runtime/DatePrototype.h: 11481 * runtime/JSArray.cpp: 11482 (JSC::JSArray::getOwnPropertyDescriptor): 11483 * runtime/JSArray.h: 11484 * runtime/JSByteArray.cpp: 11485 (JSC::JSByteArray::getOwnPropertyDescriptor): 11486 * runtime/JSByteArray.h: 11487 * runtime/JSFunction.cpp: 11488 (JSC::JSFunction::getOwnPropertyDescriptor): 11489 * runtime/JSFunction.h: 11490 * runtime/JSGlobalObject.h: 11491 (JSC::JSGlobalObject::getOwnPropertyDescriptor): 11492 * runtime/JSNotAnObject.cpp: 11493 (JSC::JSNotAnObject::getOwnPropertyDescriptor): 11494 * runtime/JSNotAnObject.h: 11495 * runtime/JSONObject.cpp: 11496 (JSC::JSONObject::getOwnPropertySlot): 11497 (JSC::JSONObject::getOwnPropertyDescriptor): 11498 * runtime/JSONObject.h: 11499 * runtime/JSObject.cpp: 11500 (JSC::JSObject::getOwnPropertyDescriptor): 11501 (JSC::JSObject::getPropertyDescriptor): 11502 * runtime/JSObject.h: 11503 * runtime/JSString.cpp: 11504 (JSC::JSString::getStringPropertyDescriptor): 11505 (JSC::JSString::getOwnPropertyDescriptor): 11506 * runtime/JSString.h: 11507 * runtime/JSVariableObject.cpp: 11508 (JSC::JSVariableObject::symbolTableGet): 11509 * runtime/JSVariableObject.h: 11510 * runtime/Lookup.h: 11511 (JSC::getStaticPropertyDescriptor): 11512 (JSC::getStaticFunctionDescriptor): 11513 (JSC::getStaticValueDescriptor): 11514 Add property descriptor equivalents of the lookup 11515 table access functions 11516 11517 * runtime/MathObject.cpp: 11518 (JSC::MathObject::getOwnPropertySlot): 11519 (JSC::MathObject::getOwnPropertyDescriptor): 11520 * runtime/MathObject.h: 11521 * runtime/NumberConstructor.cpp: 11522 (JSC::NumberConstructor::getOwnPropertyDescriptor): 11523 * runtime/NumberConstructor.h: 11524 * runtime/ObjectConstructor.cpp: 11525 (JSC::ObjectConstructor::ObjectConstructor): 11526 (JSC::objectConstructorGetOwnPropertyDescriptor): 11527 * runtime/PropertyDescriptor.cpp: Added. 11528 (JSC::PropertyDescriptor::writable): 11529 (JSC::PropertyDescriptor::enumerable): 11530 (JSC::PropertyDescriptor::configurable): 11531 (JSC::PropertyDescriptor::hasAccessors): 11532 (JSC::PropertyDescriptor::setUndefined): 11533 (JSC::PropertyDescriptor::getter): 11534 (JSC::PropertyDescriptor::setter): 11535 (JSC::PropertyDescriptor::setDescriptor): 11536 (JSC::PropertyDescriptor::setAccessorDescriptor): 11537 * runtime/PropertyDescriptor.h: Added. 11538 (JSC::PropertyDescriptor::PropertyDescriptor): 11539 (JSC::PropertyDescriptor::attributes): 11540 (JSC::PropertyDescriptor::isValid): 11541 (JSC::PropertyDescriptor::value): 11542 * runtime/RegExpConstructor.cpp: 11543 (JSC::RegExpConstructor::getOwnPropertyDescriptor): 11544 * runtime/RegExpConstructor.h: 11545 * runtime/RegExpMatchesArray.h: 11546 (JSC::RegExpMatchesArray::getOwnPropertyDescriptor): 11547 * runtime/RegExpObject.cpp: 11548 (JSC::RegExpObject::getOwnPropertyDescriptor): 11549 * runtime/RegExpObject.h: 11550 * runtime/StringObject.cpp: 11551 (JSC::StringObject::getOwnPropertyDescriptor): 11552 * runtime/StringObject.h: 11553 * runtime/StringPrototype.cpp: 11554 (JSC::StringPrototype::getOwnPropertyDescriptor): 11555 * runtime/StringPrototype.h: 11556 115572009-08-24 Gavin Barraclough <barraclough@apple.com> 11558 11559 Reviewed by Darin Adler. 11560 11561 How many copies of the parameters do you need? 11562 https://bugs.webkit.org/show_bug.cgi?id=28701 11563 11564 The function parameters in JSC get copied a lot - and unnecessarily so. 11565 11566 Originally this happened due to duplicating FunctionBodyNodes on recompilation, 11567 though the problem has been exacerbated by copying the parameters from the 11568 original function body onto the executable, then back onto the real body that 11569 will be generated (this happens on every function). And this is all made worse 11570 since the data structures in question are a little ugly - C style arrays of C++ 11571 objects containing ref counts, so they need a full copy-construct (rather than 11572 a simple memcpy). 11573 11574 This can all be greatly simplified by just punting the parameters off into 11575 their own ref-counted object, and forgoing all the copying. 11576 11577 ~no performance change, possible slight progression. 11578 11579 * bytecompiler/BytecodeGenerator.cpp: 11580 (JSC::BytecodeGenerator::BytecodeGenerator): 11581 * bytecompiler/BytecodeGenerator.h: 11582 (JSC::BytecodeGenerator::makeFunction): 11583 * parser/Nodes.cpp: 11584 (JSC::FunctionParameters::FunctionParameters): 11585 (JSC::FunctionBodyNode::FunctionBodyNode): 11586 (JSC::FunctionBodyNode::finishParsing): 11587 * parser/Nodes.h: 11588 (JSC::FunctionBodyNode::parameters): 11589 (JSC::FunctionBodyNode::parameterCount): 11590 * runtime/Executable.cpp: 11591 (JSC::FunctionExecutable::~FunctionExecutable): 11592 (JSC::FunctionExecutable::compile): 11593 (JSC::FunctionExecutable::reparseExceptionInfo): 11594 (JSC::FunctionExecutable::fromGlobalCode): 11595 (JSC::FunctionExecutable::paramString): 11596 * runtime/Executable.h: 11597 (JSC::FunctionExecutable::FunctionExecutable): 11598 (JSC::FunctionExecutable::parameterCount): 11599 116002009-08-25 Brent Fulgham <bfulgham@webkit.org> 11601 11602 Reviewed by NOBODY (Buildfix). 11603 11604 * JavaScriptCore.vcproj/jsc/jsc.vcproj: Add Debug_CFLite target 11605 that inherits from the debug_wincairo property sheet and therefore 11606 links to the proper debug library. 11607 * JavaScriptCore.vcproj/testapi/testapi.vcproj: Add Debug_CFLite target 11608 that inherits from the debug_wincairo property sheet and therefore 11609 links to the proper debug library. 11610 116112009-08-25 Chris Marrin <cmarrin@apple.com> 11612 11613 Reviewed by Simon Fraser. 11614 11615 Export tryFastMalloc for Canvas3D work 11616 https://bugs.webkit.org/show_bug.cgi?id=28018 11617 11618 * JavaScriptCore.exp: 11619 116202009-08-25 David Levin <levin@chromium.org> 11621 11622 Reviewed by Adam Roben. 11623 11624 PLATFORM(CFNETWORK) should be USE(CFNETWORK). 11625 https://bugs.webkit.org/show_bug.cgi?id=28713 11626 11627 * wtf/Platform.h: Added a #define to catch this issue in the 11628 future. The define would generate an error on gcc without the 11629 space in the expansion, but Visual C++ needs the space to cause an error. 11630 116312009-08-24 Brent Fulgham <bfulgham@webkit.org> 11632 11633 Reviewed by Steve Falkenburg. 11634 11635 Revise CFLite Debug build to emit DLL's with _debug label. 11636 https://bugs.webkit.org/show_bug.cgi?id=28695. 11637 11638 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: Modify 11639 Cairo debug build to inherit from new debug_cairo property sheet. 11640 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCoreCFLite.vsprops: 11641 Modify to look for debug CFLite when in debug build. 11642 116432009-08-24 Gavin Barraclough <barraclough@apple.com> 11644 11645 Reviewed by Oliver Adler & Darin Hunt. 11646 11647 https://bugs.webkit.org/show_bug.cgi?id=28691 11648 Do not retain ScopeNodes outside of parsing 11649 11650 There is now no need for these to exist outside of parsing - their use in the runtime is replaced by Executable types. 11651 11652 * bytecode/EvalCodeCache.h: 11653 (JSC::EvalCodeCache::get): 11654 * bytecompiler/BytecodeGenerator.cpp: 11655 (JSC::BytecodeGenerator::BytecodeGenerator): 11656 (JSC::BytecodeGenerator::emitNewFunction): 11657 (JSC::BytecodeGenerator::emitNewFunctionExpression): 11658 * bytecompiler/BytecodeGenerator.h: 11659 (JSC::BytecodeGenerator::makeFunction): 11660 * debugger/Debugger.cpp: 11661 (JSC::Debugger::recompileAllJSFunctions): 11662 (JSC::evaluateInGlobalCallFrame): 11663 * debugger/DebuggerCallFrame.cpp: 11664 (JSC::DebuggerCallFrame::evaluate): 11665 * interpreter/Interpreter.cpp: 11666 (JSC::Interpreter::execute): 11667 (JSC::Interpreter::prepareForRepeatCall): 11668 (JSC::Interpreter::privateExecute): 11669 * jit/JITStubs.cpp: 11670 (JSC::DEFINE_STUB_FUNCTION): 11671 * parser/Nodes.cpp: 11672 (JSC::ScopeNodeData::ScopeNodeData): 11673 (JSC::ProgramNode::create): 11674 (JSC::EvalNode::create): 11675 (JSC::FunctionBodyNode::create): 11676 * parser/Nodes.h: 11677 (JSC::ScopeNode::adoptData): 11678 (JSC::FunctionBodyNode::parameterCount): 11679 * parser/Parser.cpp: 11680 * parser/Parser.h: 11681 (JSC::Parser::arena): 11682 (JSC::Parser::Parser): 11683 (JSC::Parser::parse): 11684 * runtime/ArrayPrototype.cpp: 11685 (JSC::isNumericCompareFunction): 11686 (JSC::arrayProtoFuncSort): 11687 * runtime/Completion.cpp: 11688 (JSC::checkSyntax): 11689 (JSC::evaluate): 11690 * runtime/Executable.cpp: 11691 (JSC::FunctionExecutable::~FunctionExecutable): 11692 (JSC::EvalExecutable::compile): 11693 (JSC::ProgramExecutable::checkSyntax): 11694 (JSC::ProgramExecutable::compile): 11695 (JSC::FunctionExecutable::compile): 11696 (JSC::EvalExecutable::generateJITCode): 11697 (JSC::ProgramExecutable::generateJITCode): 11698 (JSC::FunctionExecutable::generateJITCode): 11699 (JSC::FunctionExecutable::reparseExceptionInfo): 11700 (JSC::EvalExecutable::reparseExceptionInfo): 11701 (JSC::FunctionExecutable::recompile): 11702 (JSC::FunctionExecutable::fromGlobalCode): 11703 (JSC::FunctionExecutable::copyParameters): 11704 (JSC::FunctionExecutable::paramString): 11705 * runtime/Executable.h: 11706 (JSC::ScriptExecutable::ScriptExecutable): 11707 (JSC::ScriptExecutable::sourceID): 11708 (JSC::ScriptExecutable::sourceURL): 11709 (JSC::ScriptExecutable::lineNo): 11710 (JSC::ScriptExecutable::lastLine): 11711 (JSC::ScriptExecutable::usesEval): 11712 (JSC::ScriptExecutable::usesArguments): 11713 (JSC::ScriptExecutable::needsActivation): 11714 (JSC::ScriptExecutable::recordParse): 11715 (JSC::EvalExecutable::bytecode): 11716 (JSC::EvalExecutable::jitCode): 11717 (JSC::ProgramExecutable::bytecode): 11718 (JSC::ProgramExecutable::reparseExceptionInfo): 11719 (JSC::ProgramExecutable::jitCode): 11720 (JSC::FunctionExecutable::FunctionExecutable): 11721 (JSC::FunctionExecutable::make): 11722 (JSC::FunctionExecutable::bytecode): 11723 (JSC::FunctionExecutable::isGenerated): 11724 (JSC::FunctionExecutable::name): 11725 (JSC::FunctionExecutable::parameterCount): 11726 (JSC::FunctionExecutable::jitCode): 11727 * runtime/FunctionConstructor.cpp: 11728 (JSC::constructFunction): 11729 * runtime/JSGlobalData.cpp: 11730 (JSC::JSGlobalData::numericCompareFunction): 11731 * runtime/JSGlobalObjectFunctions.cpp: 11732 (JSC::globalFuncEval): 11733 117342009-08-24 Darin Adler <darin@apple.com> 11735 11736 * runtime/ObjectPrototype.cpp: 11737 (JSC::ObjectPrototype::put): Landed revised version I had tested but forgot 11738 to land. Leave out the branch, since we don't need one. 11739 117402009-08-24 Darin Adler <darin@apple.com> 11741 11742 Reviewed by Geoff Garen. 11743 11744 Array index miss case creates a string every time 11745 https://bugs.webkit.org/show_bug.cgi?id=28664 11746 11747 SunSpider test results I saw: 11748 11749 0.5% faster overall 11750 1% faster on crypto-aes 11751 20% faster on crypto-md5 11752 13% faster on crypto-sha1 11753 11754 * runtime/ObjectPrototype.cpp: 11755 (JSC::ObjectPrototype::ObjectPrototype): Initialize m_hasNoPropertiesWithUInt32Names 11756 to true. 11757 (JSC::ObjectPrototype::put): Clearly m_hasNoPropertiesWithUInt32Names if the new 11758 property has a name that is the string form of a UInt32. 11759 (JSC::ObjectPrototype::getOwnPropertySlot): Don't call JSObject::getOwnPropertySlot 11760 if m_hasNoPropertiesWithUInt32Names is true, and it is highly likely to be true. 11761 11762 * runtime/ObjectPrototype.h: Added declarations for the above. 11763 117642009-08-24 Gustavo Noronha Silva <gustavo.noronha@collabora.co.uk> 11765 11766 Unreviewed. Fix a typo in my distcheck build fix. 11767 11768 * GNUmakefile.am: 11769 117702009-08-23 Gustavo Noronha Silva <gns@gnome.org> 11771 11772 Unreviewed build fix for make distcheck. 11773 11774 * GNUmakefile.am: Added files required for the build. 11775 117762009-08-22 Maciej Stachowiak <mjs@apple.com> 11777 11778 Reviewed by Mark Rowe. 11779 11780 REGRESSION(r47639-r47660): Webkit crashes on launch on PowerPC 11781 https://bugs.webkit.org/show_bug.cgi?id=28655 11782 11783 * runtime/JSFunction.cpp: 11784 (JSC::JSFunction::JSFunction): Initialize properly with a VPtrHackExecutable. 11785 * wtf/Platform.h: 11786 117872009-08-22 Darin Adler <darin@apple.com> 11788 11789 Fix storage leak from syntax tree arena allocation patch. 11790 11791 * parser/Nodes.h: CommaNode needs to inherit from ParserArenaDeletable 11792 because it has a vector. 11793 117942009-08-21 Darin Adler <darin@apple.com> 11795 11796 Fix Qt build. 11797 11798 * parser/Nodes.cpp: 11799 (JSC::ScopeNodeData::ScopeNodeData): Made non-inline again. 11800 This is used outside Nodes.cpp so can't be inline unless 11801 it is in the header. 11802 118032009-08-21 Darin Adler <darin@apple.com> 11804 11805 Two loose ends from the last commit. 11806 11807 * JavaScriptCore.xcodeproj/project.pbxproj: Made ParserArena.h 11808 and create_hash_table project-internal instead of "private". 11809 * runtime/Executable.h: Removed accidentally-added constructor. 11810 118112009-08-21 Darin Adler <darin@apple.com> 11812 11813 Reviewed by Gavin Barraclough. 11814 11815 Syntax tree nodes should use arena allocation 11816 https://bugs.webkit.org/show_bug.cgi?id=25674 11817 11818 Use an actual arena now. 0.6% speedup on SunSpider. 11819 11820 New and improved with 100% less leaking of the universe. 11821 11822 * JavaScriptCore.exp: 11823 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 11824 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 11825 Removed all exports involving the class FunctionBodyNode, which no 11826 longer needs to be used outside JavaScriptCore. 11827 11828 * JavaScriptCore.xcodeproj/project.pbxproj: Made Nodes.h and 11829 Executable.h project-internal instead of "private". 11830 11831 * bytecompiler/BytecodeGenerator.cpp: 11832 (JSC::BytecodeGenerator::BytecodeGenerator): Updated since VarStack 11833 contains const Identifier* now. 11834 11835 * parser/Grammar.y: Made identifiers from the lexer be const 11836 Identifier* and updated since VarStack contains const Identifier* now. 11837 11838 * parser/Lexer.cpp: 11839 (JSC::Lexer::setCode): Pass in ParserArena, used for identifiers. 11840 (JSC::Lexer::makeIdentifier): Changed return type to const Identifier* 11841 and changed to call ParserArena. 11842 (JSC::Lexer::clear): Removed the code to manage m_identifiers and 11843 added code to set m_arena to 0. 11844 * parser/Lexer.h: Updated for changes above. 11845 11846 * parser/NodeConstructors.h: 11847 (JSC::ParserArenaFreeable::operator new): Added. Calls allocateFreeable 11848 on the arena. 11849 (JSC::ParserArenaDeletable::operator new): Changed to call the 11850 allocateDeletable function on the arena instead of deleteWithArena. 11851 (JSC::PropertyNode::PropertyNode): Added new constructor that makes 11852 numeric identifiers. Some day we might want to optimize this for 11853 integers so it doesn't create a string for each one. 11854 (JSC::ContinueNode::ContinueNode): Initialize m_ident to nullIdentifier 11855 since it's now a const Identifier& so it can't be left uninitialized. 11856 (JSC::BreakNode::BreakNode): Ditto. 11857 (JSC::CaseClauseNode::CaseClauseNode): Updated to use SourceElements* 11858 to keep track of the statements rather than a separate statement vector. 11859 (JSC::BlockNode::BlockNode): Ditto. 11860 (JSC::ForInNode::ForInNode): Initialize m_ident to nullIdentifier. 11861 11862 * parser/Nodes.cpp: Moved the comment explaining emitBytecode in here. 11863 It seemed strangely out of place in the header. 11864 (JSC::ThrowableExpressionData::emitThrowError): Added an overload for 11865 UString as well as Identifier. 11866 (JSC::SourceElements::singleStatement): Added. 11867 (JSC::SourceElements::lastStatement): Added. 11868 (JSC::RegExpNode::emitBytecode): Changed the throwError code to use 11869 the substitution mechanism instead of doing a string append. 11870 (JSC::SourceElements::emitBytecode): Added. Replaces the old 11871 statementListEmitCode function, since we now keep the SourceElements 11872 objects around. 11873 (JSC::BlockNode::lastStatement): Added. 11874 (JSC::BlockNode::emitBytecode): Changed to use emitBytecode instead of 11875 statementListEmitCode. 11876 (JSC::CaseClauseNode::emitBytecode): Added. 11877 (JSC::CaseBlockNode::emitBytecodeForBlock): Changed to use emitBytecode 11878 instead of statementListEmitCode. 11879 (JSC::ScopeNodeData::ScopeNodeData): Changed to store the 11880 SourceElements* instead of using releaseContentsIntoVector. 11881 (JSC::ScopeNode::emitStatementsBytecode): Added. 11882 (JSC::ScopeNode::singleStatement): Added. 11883 (JSC::ProgramNode::emitBytecode): Call emitStatementsBytecode instead 11884 of statementListEmitCode. 11885 (JSC::EvalNode::emitBytecode): Ditto. 11886 (JSC::FunctionBodyNode::emitBytecode): Call emitStatementsBytecode 11887 insetad of statementListEmitCode and check for the return node using 11888 the new functions. 11889 11890 * parser/Nodes.h: Changed VarStack to store const Identifier* instead 11891 of Identifier and rely on the arena to control lifetime. Added a new 11892 ParserArenaFreeable class. Made ParserArenaDeletable inherit from 11893 FastAllocBase instead of having its own operator new. Base the Node 11894 class on ParserArenaFreeable. Changed the various Node classes 11895 to use const Identifier& instead of Identifier to avoid the need to 11896 call their destructors and allow them to function as "freeable" in the 11897 arena. Removed extraneous JSC_FAST_CALL on definitions of inline functions. 11898 Changed ElementNode, PropertyNode, ArgumentsNode, ParameterNode, 11899 CaseClauseNode, ClauseListNode, and CaseBlockNode to use ParserArenaFreeable 11900 as a base class since they do not descend from Node. Eliminated the 11901 StatementVector type and instead have various classes use SourceElements* 11902 instead of StatementVector. This prevents those classes from having to 11903 use ParserArenaDeletable to make sure the vector destructor is called. 11904 11905 * parser/Parser.cpp: 11906 (JSC::Parser::parse): Pass the arena to the lexer. 11907 11908 * parser/Parser.h: Added an include of ParserArena.h, which is no longer 11909 included by Nodes.h. 11910 (JSC::Parser::parseFunctionFromGlobalCode): Changed to use the 11911 singleStatement function, since there is no longer any children function. 11912 Removed some unneeded use of RefPtr. 11913 11914 * parser/ParserArena.cpp: 11915 (JSC::ParserArena::ParserArena): Added. Initializes the new members, 11916 m_freeableMemory, m_freeablePoolEnd, and m_identifiers. 11917 (JSC::ParserArena::freeablePool): Added. Computes the pool pointer, 11918 since we store only the current pointer and the end of pool pointer. 11919 (JSC::ParserArena::deallocateObjects): Added. Contains the common 11920 memory-deallocation logic used by both the destructor and the 11921 reset function. 11922 (JSC::ParserArena::~ParserArena): Changed to call deallocateObjects. 11923 (JSC::ParserArena::reset): Ditto. Also added code to zero out the 11924 new structures, and switched to use clear() instead of shrink(0) since 11925 we don't really reuse arenas. 11926 (JSC::ParserArena::makeNumericIdentifier): Added. 11927 (JSC::ParserArena::allocateFreeablePool): Added. Used when the pool 11928 is empty. 11929 (JSC::ParserArena::isEmpty): Added. No longer inline, which is fine 11930 since this is used only for assertions at the moment. 11931 (JSC::ParserArena::derefWithArena): Make non-inline. 11932 11933 * parser/ParserArena.h: Added an actual arena of "freeable" objects, 11934 ones that don't need destructors to be called. Also added a separate 11935 IdentifierArena object, a segmented vector of identifiers that used 11936 to be in the Lexer. 11937 11938 * runtime/Executable.h: Moved the definition of the 11939 FunctionExecutable::make function here. It can't go in JSFunction.h 11940 since that header has to be used outside JavaScriptCore and so can't 11941 include this, which includes Nodes.h. The function could be moved 11942 elswhere if we don't want to include JSFunction.h in this header, but 11943 for now this seems to be the best place. 11944 11945 * runtime/JSFunction.h: Removed the include of Executable.h and 11946 definition of the FunctionExecutable::make function. 11947 11948 * wtf/FastMalloc.cpp: Fixed an incorrect comment. 11949 119502009-08-21 Mark Rowe <mrowe@apple.com> 11951 11952 Fix the non-JIT build. 11953 11954 * runtime/Executable.cpp: 11955 * runtime/Executable.h: 11956 119572009-08-21 Gavin Barraclough <barraclough@apple.com> 11958 11959 Speculative QuickTime build fix. 11960 11961 * runtime/JSArray.cpp: 11962 119632009-08-21 Gavin Barraclough <barraclough@apple.com> 11964 11965 Speculative QT build fix. 11966 11967 * runtime/StringPrototype.cpp: 11968 119692009-08-21 Gavin Barraclough <barraclough@apple.com> 11970 11971 Reviewed by Oliver Hunt. 11972 11973 Restructure Executable types so that host functions do not hold a FunctionExecutable. 11974 https://bugs.webkit.org/show_bug.cgi?id=28621 11975 11976 All JSFunction objects have a pointer to an Executable*. This is currently always a 11977 FunctionExecutable, however this has a couple of drawbacks. Host functions do not 11978 store a range of information that the FunctionExecutable provides (source, name, 11979 CodeBlock & information presently held on the FunctionBodyNode). 11980 11981 [ * nearly all... see below! ] 11982 11983 Instead, make JSFunctions hold a pointer to an ExecutableBase, move fields specific 11984 to JS sourced executable types (source, node) into a new subclass (ScriptExecutable), 11985 and create a new NativeExecutable type. We now provide a new method in JSFunction 11986 to access & downcast to FunctionExecutable, but in doing so we can make an early 11987 check (with an ASSERT) to ensure that the Executable read from a function will only 11988 be treated as a FunctionExecutable (and thus the JS sepcific fields will only be 11989 accessed) if the JSFunction is not a host function. 11990 11991 There is one JSFunction that currently does not have an Executable, which is the 11992 object created to allow us to read out the vtable pointer. By making this change 11993 we can also add a new Executable type fror this object (VPtrHackExecutable). 11994 Since this means that really all JSFunctions have an Executable we no longer have 11995 to null-check m_executable before us it - particularly in isHostFunction(). 11996 11997 This patch removes CacheableEvalExecutable, since all subclasses of ExecutableBase 11998 can now be ref-counted - since both JSFunction holds (and ref-counts) an ExecutableBase 11999 that might be a FunctionExecutable or a NativeExecutable. This does now mean that all 12000 ProgramExecutables and EvalExecutables (unnecessarily) provide an interface to be 12001 ref-counted, however this seems less-bad than host functions unnecessarily providing 12002 interface to access non-host specific information. 12003 12004 The class hierarcy has changed from this: 12005 12006 - ExecutableBase 12007 - ProgramExecutable 12008 - EvalExecutable 12009 - CacheableEvalExecutable (also RefCounted by multiple-inheritance) 12010 - FunctionExecutable (also RefCounted by multiple-inheritance, 'special' FunctionExecutable also used for host functions) 12011 12012 To this: 12013 12014 - RefCounted 12015 - ExecutableBase 12016 - NativeExecutable 12017 - VPtrHackExecutable 12018 - ScriptExecutable 12019 - ProgramExecutable 12020 - EvalExecutable 12021 - FunctionExecutable 12022 12023 This patch speeds up sunspidey by a couple of ms (presumably due to the changes to isHostFunction()). 12024 12025 * bytecode/CodeBlock.cpp: 12026 (JSC::CodeBlock::CodeBlock): 12027 * bytecode/CodeBlock.h: 12028 (JSC::CodeBlock::ownerExecutable): 12029 (JSC::GlobalCodeBlock::GlobalCodeBlock): 12030 * bytecode/EvalCodeCache.h: 12031 (JSC::EvalCodeCache::get): 12032 * debugger/Debugger.cpp: 12033 (JSC::Debugger::recompileAllJSFunctions): 12034 * interpreter/CachedCall.h: 12035 (JSC::CachedCall::CachedCall): 12036 * interpreter/Interpreter.cpp: 12037 (JSC::Interpreter::callEval): 12038 (JSC::Interpreter::privateExecute): 12039 * jit/JITStubs.cpp: 12040 (JSC::DEFINE_STUB_FUNCTION): 12041 * profiler/Profiler.cpp: 12042 (JSC::createCallIdentifierFromFunctionImp): 12043 * runtime/Arguments.h: 12044 (JSC::Arguments::getArgumentsData): 12045 (JSC::Arguments::Arguments): 12046 * runtime/Executable.cpp: 12047 (JSC::NativeExecutable::~NativeExecutable): 12048 (JSC::VPtrHackExecutable::~VPtrHackExecutable): 12049 * runtime/Executable.h: 12050 (JSC::ExecutableBase::ExecutableBase): 12051 (JSC::ExecutableBase::~ExecutableBase): 12052 (JSC::ExecutableBase::isHostFunction): 12053 (JSC::NativeExecutable::NativeExecutable): 12054 (JSC::VPtrHackExecutable::VPtrHackExecutable): 12055 (JSC::ScriptExecutable::ScriptExecutable): 12056 (JSC::ScriptExecutable::source): 12057 (JSC::ScriptExecutable::sourceID): 12058 (JSC::ScriptExecutable::sourceURL): 12059 (JSC::ScriptExecutable::lineNo): 12060 (JSC::ScriptExecutable::lastLine): 12061 (JSC::ScriptExecutable::usesEval): 12062 (JSC::ScriptExecutable::usesArguments): 12063 (JSC::ScriptExecutable::needsActivation): 12064 (JSC::EvalExecutable::EvalExecutable): 12065 (JSC::EvalExecutable::create): 12066 (JSC::ProgramExecutable::ProgramExecutable): 12067 (JSC::FunctionExecutable::FunctionExecutable): 12068 * runtime/FunctionPrototype.cpp: 12069 (JSC::functionProtoFuncToString): 12070 * runtime/JSFunction.cpp: 12071 (JSC::JSFunction::JSFunction): 12072 (JSC::JSFunction::~JSFunction): 12073 (JSC::JSFunction::markChildren): 12074 (JSC::JSFunction::getCallData): 12075 (JSC::JSFunction::call): 12076 (JSC::JSFunction::lengthGetter): 12077 (JSC::JSFunction::getConstructData): 12078 (JSC::JSFunction::construct): 12079 * runtime/JSFunction.h: 12080 (JSC::JSFunction::executable): 12081 (JSC::JSFunction::jsExecutable): 12082 (JSC::JSFunction::isHostFunction): 12083 120842009-08-20 Oliver Hunt <oliver@apple.com> 12085 12086 Reviewed by Maciej Stachowiak. 12087 12088 Browser hangs on opening Web Inspector. 12089 https://bugs.webkit.org/show_bug.cgi?id=28438 12090 12091 Code generation needs to be able to walk the entire scopechain in some 12092 cases, however the symbol table used by activations was a member of the 12093 codeblock. Following recompilation this may no longer exist, leading 12094 to a crash or hang on lookup. 12095 12096 We fix this by introducing a refcounted SymbolTable subclass, SharedSymbolTable, 12097 for the CodeBlocks used by function code. This allows activations to 12098 maintain ownership of a copy of the symbol table even after recompilation so 12099 they can continue to work. 12100 12101 * bytecode/CodeBlock.cpp: 12102 (JSC::CodeBlock::CodeBlock): 12103 * bytecode/CodeBlock.h: 12104 (JSC::CodeBlock::symbolTable): 12105 (JSC::CodeBlock::sharedSymbolTable): 12106 (JSC::GlobalCodeBlock::GlobalCodeBlock): 12107 (JSC::FunctionCodeBlock::FunctionCodeBlock): 12108 (JSC::FunctionCodeBlock::~FunctionCodeBlock): 12109 * interpreter/Interpreter.cpp: 12110 (JSC::Interpreter::retrieveArguments): 12111 * runtime/Executable.cpp: 12112 (JSC::EvalExecutable::generateBytecode): 12113 (JSC::FunctionExecutable::generateBytecode): 12114 (JSC::FunctionExecutable::reparseExceptionInfo): 12115 (JSC::EvalExecutable::reparseExceptionInfo): 12116 * runtime/JSActivation.h: 12117 (JSC::JSActivation::JSActivationData::JSActivationData): 12118 (JSC::JSActivation::JSActivationData::~JSActivationData): 12119 * runtime/SymbolTable.h: 12120 121212009-08-20 Xan Lopez <xlopez@igalia.com> 12122 12123 Add new file to GTK+ build. 12124 12125 * GNUmakefile.am: 12126 121272009-08-20 Geoffrey Garen <ggaren@apple.com> 12128 12129 Reviewed by Maciej Stachowiak. 12130 12131 Added a number => string cache. 12132 12133 1.07x faster on v8 (1.7x faster on v8-splay). 12134 1.004x faster on SunSpider. 12135 12136 * runtime/JSCell.h: Moved JSValue::toString to JSString.h. 12137 * runtime/JSGlobalData.h: Holds the cache. 12138 * runtime/JSNumberCell.cpp: 12139 (JSC::JSNumberCell::toString): 12140 (JSC::JSNumberCell::toThisString): Removed -0 special case. 12141 UString handles this now, since too many clients were 12142 special-casing it. 12143 12144 * runtime/JSString.h: 12145 (JSC::JSValue::toString): Use the cache when converting 12146 an int or double to string. 12147 12148 * runtime/Operations.h: 12149 (JSC::concatenateStrings): Call toString to take advantage 12150 of the cache. 12151 12152 * runtime/SmallStrings.h: 12153 (JSC::NumericStrings::add): 12154 (JSC::NumericStrings::lookup): The cache. 12155 12156 * runtime/UString.cpp: 12157 (JSC::UString::from): Added -0 special case mentioned above. 12158 Removed appendNumeric because it's mutually exclusive with the 12159 cache. 12160 121612009-08-20 Oliver Hunt <oliver@apple.com> 12162 12163 Reviewed by Gavin Barraclough. 12164 12165 REGRESSION: fast/profiler/call.html is crashing occasionally 12166 https://bugs.webkit.org/show_bug.cgi?id=28476 12167 12168 Using the codeblock for information about how many parameters and 12169 locals a function has is unsafe in certain circumstances. The 12170 basic scenario is all function code being cleared in response to 12171 the debugger or profiler being enabled, and then an activation is 12172 marked before its associated function is re-executed. 12173 12174 To deal with this scenario we store the variable count of a function 12175 directly in the FunctionExecutable, and then use that information. 12176 12177 * runtime/Arguments.h: 12178 (JSC::Arguments::getArgumentsData): 12179 * runtime/Executable.cpp: 12180 (JSC::FunctionExecutable::generateBytecode): 12181 * runtime/Executable.h: 12182 (JSC::FunctionExecutable::FunctionExecutable): 12183 (JSC::FunctionExecutable::variableCount): 12184 * runtime/JSActivation.cpp: 12185 (JSC::JSActivation::markChildren): 12186 121872009-08-20 Gavin Barraclough <barraclough@apple.com> 12188 12189 Reviewed by Oliver Hunt. 12190 12191 Numbering of arguments to emitGetJITStubArg/emitPutJITStubArg incorrect 12192 <bug lost in the great bug disasteroony of 08/20/09!> 12193 12194 The argumentNumber argument to emitGetJITStubArg/emitPutJITStubArg should match 12195 the argument number used within the stub functions in JITStubs.cpp, but it doesn't. 12196 12197 Firstly, all the numbers changed when we added a void* 'reserved' as the first slot 12198 (rather than leaving argument 0 unused), and secondly in 32_64 builds the index to 12199 peek/poke needs to be multiplies by 2 (since the argument to peek/poke is a number 12200 of machine words, and on 32_64 build the argument slots to stub functions are two 12201 words wide). 12202 12203 * jit/JIT.h: 12204 * jit/JITCall.cpp: 12205 (JSC::JIT::compileOpCallSetupArgs): 12206 (JSC::JIT::compileOpConstructSetupArgs): 12207 (JSC::JIT::compileOpCallVarargsSetupArgs): 12208 (JSC::JIT::compileOpCall): 12209 * jit/JITInlineMethods.h: 12210 (JSC::JIT::emitPutJITStubArg): 12211 (JSC::JIT::emitPutJITStubArgConstant): 12212 (JSC::JIT::emitGetJITStubArg): 12213 (JSC::JIT::emitPutJITStubArgFromVirtualRegister): 12214 * jit/JITOpcodes.cpp: 12215 (JSC::JIT::privateCompileCTIMachineTrampolines): 12216 * jit/JITPropertyAccess.cpp: 12217 (JSC::JIT::privateCompilePutByIdTransition): 12218 122192009-08-20 Oliver Hunt <oliver@apple.com> 12220 12221 Reviewed by Geoff Garen. 12222 12223 REGRESSION: significant slowdown on Celtic Kane "AJAX declaration" subtest 12224 https://bugs.webkit.org/show_bug.cgi?id=28332 12225 12226 Follow up style fixes that were missed in review. 12227 12228 * runtime/Structure.cpp: 12229 (JSC::Structure::hasTransition): 12230 * runtime/Structure.h: 12231 (JSC::Structure::get): 12232 (JSC::StructureTransitionTable::contains): 12233 * runtime/StructureTransitionTable.h: 12234 (JSC::StructureTransitionTable::add): 12235 122362009-08-20 Oliver Hunt <oliver@apple.com> 12237 12238 Add new exports to windows jsc build 12239 12240 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 12241 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 12242 122432009-08-20 Oliver Hunt <oliver@apple.com> 12244 12245 Reviewed by Gavin Barraclough. 12246 12247 REGRESSION: significant slowdown on Celtic Kane "AJAX declaration" subtest 12248 https://bugs.webkit.org/show_bug.cgi?id=28332 12249 12250 The method check optimisation made transitions aware of the value being 12251 assigned when a transition was assigning a function. This had the side 12252 effect of making every assignment of a function expression result in a 12253 new transition, and thus a new Structure. The net result of this is that 12254 the common JS idiom of 12255 12256 function MyObject() { 12257 this.myFunction = function(...){...}; 12258 } 12259 new MyObject(); 12260 12261 Will produce a unique structure on every iteration, meaning that all 12262 caching is defeated and there is a significant amount of structure churn. 12263 12264 The fix is to return the transition to its original form where it is 12265 keyed off a property name + attributes tuple, but have each transition 12266 support an optional transition on a specific value. 12267 12268 * JavaScriptCore.exp: 12269 * runtime/JSObject.h: 12270 (JSC::JSObject::putDirectInternal): 12271 * runtime/Structure.cpp: 12272 (JSC::Structure::~Structure): 12273 (JSC::Structure::addPropertyTransitionToExistingStructure): 12274 (JSC::Structure::addPropertyTransition): 12275 (JSC::Structure::hasTransition): 12276 * runtime/Structure.h: 12277 (JSC::Structure::transitionedFor): 12278 (JSC::Structure::hasTransition): 12279 (JSC::Structure::): 12280 (JSC::StructureTransitionTable::contains): 12281 (JSC::StructureTransitionTable::get): 12282 * runtime/StructureTransitionTable.h: 12283 (JSC::StructureTransitionTableHashTraits::emptyValue): 12284 (JSC::StructureTransitionTable::hasTransition): 12285 (JSC::StructureTransitionTable::remove): 12286 (JSC::StructureTransitionTable::add): 12287 122882009-08-20 Gavin Barraclough <barraclough@apple.com> 12289 12290 Reviewed by Oliver Hunt. 12291 12292 Remove FunctionCodeBlock. 12293 https://bugs.webkit.org/show_bug.cgi?id=28502 12294 12295 These only exist to allow JIT code to dereference properties off the 12296 CodeBlock for any callee, regardless of whether it is a host function. 12297 12298 Instead just use the FunctionExecutable. Copy the m_parameters field 12299 from the CodeBlock into the Executable, and use this to distinguish 12300 between host functions, functions that have been bytecompiled, and 12301 functions that have not. 12302 12303 m_parameters is moved to ExecutableBase rather than FunctionExecutable 12304 so that (as a separate change) we can move make a separate class of 12305 executable for host code, which is not devived from FunctionExecutable 12306 (host code does not feature any of the properties that normal executable 12307 do and will provide, such as source, attributes, and a parsed name). 12308 12309 1% win on v8 tests, 0.5% on sunspider. 12310 12311 * bytecode/CodeBlock.cpp: 12312 (JSC::CodeBlock::derefStructures): 12313 (JSC::CodeBlock::refStructures): 12314 (JSC::CodeBlock::reparseForExceptionInfoIfNecessary): 12315 (JSC::CodeBlock::handlerForBytecodeOffset): 12316 (JSC::CodeBlock::lineNumberForBytecodeOffset): 12317 (JSC::CodeBlock::expressionRangeForBytecodeOffset): 12318 (JSC::CodeBlock::getByIdExceptionInfoForBytecodeOffset): 12319 (JSC::CodeBlock::functionRegisterForBytecodeOffset): 12320 (JSC::CodeBlock::hasGlobalResolveInstructionAtBytecodeOffset): 12321 (JSC::CodeBlock::hasGlobalResolveInfoAtBytecodeOffset): 12322 * bytecode/CodeBlock.h: 12323 (JSC::): 12324 (JSC::CodeBlock::source): 12325 (JSC::CodeBlock::sourceOffset): 12326 (JSC::CodeBlock::evalCodeCache): 12327 (JSC::CodeBlock::createRareDataIfNecessary): 12328 12329 remove NativeCodeBlocks and the NativeCode code type. 12330 12331 * jit/JIT.cpp: 12332 (JSC::JIT::linkCall): 12333 12334 Revert to previous behaviour (as currently still commented!) that Hhost functions have a null codeblock. 12335 12336 * jit/JITCall.cpp: 12337 (JSC::JIT::compileOpCallInitializeCallFrame): 12338 (JSC::JIT::compileOpCallSetupArgs): 12339 (JSC::JIT::compileOpCallVarargsSetupArgs): 12340 (JSC::JIT::compileOpConstructSetupArgs): 12341 (JSC::JIT::compileOpCallVarargs): 12342 (JSC::JIT::compileOpCall): 12343 (JSC::JIT::compileOpCallSlowCase): 12344 12345 Bring the 32_64 & non-32_64 JITs into line with each other, callee in regT0. 12346 12347 * jit/JITOpcodes.cpp: 12348 (JSC::JIT::privateCompileCTIMachineTrampolines): 12349 12350 Rewrite call trampolines to not use the CodeBlock. 12351 12352 * jit/JITStubs.cpp: 12353 (JSC::DEFINE_STUB_FUNCTION): 12354 12355 Make call_JSFunction & call_arityCheck return the callee, don't expect to be passed the CodeBlock. 12356 12357 * runtime/Executable.cpp: 12358 (JSC::FunctionExecutable::generateBytecode): 12359 (JSC::FunctionExecutable::recompile): 12360 (JSC::FunctionExecutable::FunctionExecutable): 12361 * runtime/Executable.h: 12362 (JSC::ExecutableBase::): 12363 (JSC::ExecutableBase::ExecutableBase): 12364 (JSC::FunctionExecutable::isHostFunction): 12365 12366 Add m_numParameters. 12367 12368 * runtime/JSFunction.cpp: 12369 (JSC::JSFunction::~JSFunction): 12370 12371 Only call generatedBytecode() on JSFunctions non-host FunctionExecutables. 12372 123732009-08-20 Yongjun Zhang <yongjun.zhang@nokia.com> 12374 12375 Reviewed by Eric Seidel. 12376 12377 https://bugs.webkit.org/show_bug.cgi?id=28054 12378 12379 Use a helper function to work around winscw compiler forward declaration bug 12380 regarding templated classes. 12381 12382 Add parenthesis around (PassRefPtr::*UnspecifiedBoolType) to make winscw compiler 12383 work with the default UnSpecifiedBoolType() operator, which removes the winscw 12384 specific bool cast hack. 12385 12386 * wtf/PassRefPtr.h: 12387 (WTF::derefIfNotNull): 12388 (WTF::PassRefPtr::~PassRefPtr): 12389 123902009-08-19 Yong Li <yong.li@torchmobile.com> 12391 12392 Reviewed by Gavin Barraclough. 12393 12394 Change namespace ARM to ARMRegisters 12395 X86 to X86Registers to avoid conflict with macros 12396 https://bugs.webkit.org/show_bug.cgi?id=28428 12397 12398 * assembler/ARMAssembler.cpp: 12399 * assembler/ARMAssembler.h: 12400 * assembler/ARMv7Assembler.h: 12401 * assembler/MacroAssemblerARM.h: 12402 * assembler/MacroAssemblerARMv7.h: 12403 * assembler/MacroAssemblerX86Common.h: 12404 * assembler/MacroAssemblerX86_64.h: 12405 * assembler/X86Assembler.h: 12406 * jit/JIT.h: 12407 * jit/JITArithmetic.cpp: 12408 * jit/JITInlineMethods.h: 12409 * jit/JITOpcodes.cpp: 12410 * wrec/WRECGenerator.cpp: 12411 * wrec/WRECGenerator.h: 12412 * yarr/RegexJIT.cpp: 12413 124142009-08-19 Oliver Hunt <oliver@apple.com> 12415 12416 Reviewed by Gavin Barraclough. 12417 12418 Devirtualise marking 12419 https://bugs.webkit.org/show_bug.cgi?id=28294 12420 12421 We actually need to mark the value in a number object if we're using the 12422 32bit number representation. 12423 12424 * runtime/NumberObject.h: 12425 (JSC::NumberObject::createStructure): 12426 124272009-08-19 Gavin Barraclough <barraclough@apple.com> 12428 12429 Reviewed by Darin Adler. 12430 12431 We probably shouldn't be keeping the AST for eval nodes around forevar. 12432 https://bugs.webkit.org/show_bug.cgi?id=28469 12433 12434 EvalNodes don't destroyData() (delete their parser data) since they need to hold onto 12435 their varStack. Copy a list of variable onto EvalCodeBlock, and this can go away. 12436 12437 * bytecode/CodeBlock.h: 12438 (JSC::EvalCodeBlock::variable): 12439 (JSC::EvalCodeBlock::numVariables): 12440 (JSC::EvalCodeBlock::adoptVariables): 12441 * bytecompiler/BytecodeGenerator.cpp: 12442 (JSC::BytecodeGenerator::BytecodeGenerator): 12443 * interpreter/Interpreter.cpp: 12444 (JSC::Interpreter::execute): 12445 * parser/Nodes.h: 12446 * runtime/Executable.cpp: 12447 (JSC::EvalExecutable::generateBytecode): 12448 * runtime/Executable.h: 12449 124502009-08-19 Jungshik Shin <jshin@chromium.org> 12451 12452 Reviewed by Darin Adler. 12453 12454 http://bugs.webkit.org/show_bug.cgi?id=28441 12455 12456 Fix a build issue with ICU 4.2 or later on Windows with Visual C++. 12457 Instead of defining all isXXX and toupper/tolower as 12458 WTF_Please_use_ASCIICType_instead_of_ctype_see_comment_in_ASCIICType_h, 12459 #define them to be different by prepending 'WTF_...ASCIIType_h' with 12460 the originial names like 'toupper_WTF_...ASCIIType_h'. 12461 12462 * wtf/DisallowCType.h: 12463 124642009-08-18 Oliver Hunt <oliver@apple.com> 12465 12466 Reviewed by Gavin Barraclough. 12467 12468 Assigning a function to an object should always use the existing transition, even if the transition is not specialized 12469 https://bugs.webkit.org/show_bug.cgi?id=28442 12470 12471 Check for an unspecialized transition as an alternative to always failing if specialisation does not match. 12472 12473 * runtime/Structure.cpp: 12474 (JSC::Structure::addPropertyTransitionToExistingStructure): 12475 124762009-08-18 Dirk Schulze <krit@webkit.org> 12477 12478 Reviewed by Oliver Hunt. 12479 12480 Added additional getter to ByteArray with an unsigned char as return. 12481 ByteArray can take unsigned char directly now. 12482 12483 * wtf/ByteArray.h: 12484 (WTF::ByteArray::set): 12485 (WTF::ByteArray::get): 12486 124872009-08-18 Peter Kasting <pkasting@google.com> 12488 12489 Reviewed by Eric Seidel. 12490 12491 https://bugs.webkit.org/show_bug.cgi?id=28415 12492 Set svn:eol-style CRLF on all .sln and .vcproj files that don't already 12493 have it. 12494 12495 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCoreGenerated.vcproj: 12496 * JavaScriptCore.vcproj/testapi/testapi.vcproj: 12497 124982009-08-18 Xan Lopez <xlopez@igalia.com> 12499 12500 Try to fix the GTK+ build. 12501 12502 * GNUmakefile.am: 12503 125042009-08-17 Gavin Barraclough <barraclough@apple.com> 12505 12506 Reviewed by Sam Weinig. 12507 12508 No, silly runtime, AST nodes are not for you. 12509 12510 We still use AST nodes (ScopeNodes, particularly FunctionBodyNodes) within 12511 the runtime, which means that these nodes must be persisted outside of the 12512 arena, contain both parser & runtime data, etc. This is all a bit of a mess. 12513 12514 Move functionality into a new FunctionExecutable class. 12515 12516 * API/JSCallbackFunction.cpp: 12517 * API/JSObjectRef.cpp: 12518 * JavaScriptCore.exp: 12519 * JavaScriptCore.xcodeproj/project.pbxproj: 12520 * bytecode/CodeBlock.cpp: 12521 (JSC::CodeBlock::CodeBlock): 12522 (JSC::CodeBlock::markAggregate): 12523 (JSC::CodeBlock::reparseForExceptionInfoIfNecessary): 12524 (JSC::CodeBlock::lineNumberForBytecodeOffset): 12525 (JSC::CodeBlock::shrinkToFit): 12526 * bytecode/CodeBlock.h: 12527 (JSC::CodeBlock::getBytecodeIndex): 12528 (JSC::CodeBlock::discardBytecode): 12529 (JSC::CodeBlock::instructionCount): 12530 (JSC::CodeBlock::getJITCode): 12531 (JSC::CodeBlock::executablePool): 12532 (JSC::CodeBlock::ownerExecutable): 12533 (JSC::CodeBlock::extractExceptionInfo): 12534 (JSC::CodeBlock::addFunctionDecl): 12535 (JSC::CodeBlock::functionDecl): 12536 (JSC::CodeBlock::numberOfFunctionDecls): 12537 (JSC::CodeBlock::addFunctionExpr): 12538 (JSC::CodeBlock::functionExpr): 12539 (JSC::GlobalCodeBlock::GlobalCodeBlock): 12540 (JSC::ProgramCodeBlock::ProgramCodeBlock): 12541 (JSC::EvalCodeBlock::EvalCodeBlock): 12542 (JSC::FunctionCodeBlock::FunctionCodeBlock): 12543 (JSC::NativeCodeBlock::NativeCodeBlock): 12544 * bytecode/EvalCodeCache.h: 12545 * bytecode/SamplingTool.cpp: 12546 (JSC::SamplingTool::doRun): 12547 * bytecompiler/BytecodeGenerator.cpp: 12548 (JSC::BytecodeGenerator::BytecodeGenerator): 12549 (JSC::BytecodeGenerator::emitNewFunction): 12550 (JSC::BytecodeGenerator::emitNewFunctionExpression): 12551 * bytecompiler/BytecodeGenerator.h: 12552 * debugger/Debugger.cpp: 12553 (JSC::Debugger::recompileAllJSFunctions): 12554 * interpreter/CachedCall.h: 12555 (JSC::CachedCall::CachedCall): 12556 * interpreter/CallFrameClosure.h: 12557 * interpreter/Interpreter.cpp: 12558 (JSC::Interpreter::unwindCallFrame): 12559 (JSC::Interpreter::throwException): 12560 (JSC::Interpreter::execute): 12561 (JSC::Interpreter::prepareForRepeatCall): 12562 (JSC::Interpreter::debug): 12563 (JSC::Interpreter::privateExecute): 12564 (JSC::Interpreter::retrieveLastCaller): 12565 * interpreter/Interpreter.h: 12566 * jit/JIT.cpp: 12567 (JSC::JIT::privateCompile): 12568 * jit/JIT.h: 12569 (JSC::JIT::compile): 12570 * jit/JITOpcodes.cpp: 12571 (JSC::JIT::privateCompileCTIMachineTrampolines): 12572 (JSC::JIT::emit_op_new_func): 12573 (JSC::JIT::emit_op_new_func_exp): 12574 * jit/JITStubs.cpp: 12575 (JSC::DEFINE_STUB_FUNCTION): 12576 * jit/JITStubs.h: 12577 (JSC::): 12578 * parser/Nodes.cpp: 12579 (JSC::FunctionBodyNode::reparseDataIfNecessary): 12580 * parser/Nodes.h: 12581 (JSC::EvalNode::partialDestroyData): 12582 * parser/Parser.h: 12583 * profiler/ProfileGenerator.cpp: 12584 * profiler/Profiler.cpp: 12585 (JSC::Profiler::createCallIdentifier): 12586 (JSC::createCallIdentifierFromFunctionImp): 12587 * runtime/Arguments.h: 12588 (JSC::Arguments::getArgumentsData): 12589 (JSC::Arguments::Arguments): 12590 (JSC::JSActivation::copyRegisters): 12591 * runtime/ArrayPrototype.cpp: 12592 (JSC::isNumericCompareFunction): 12593 * runtime/CallData.h: 12594 (JSC::): 12595 * runtime/Collector.cpp: 12596 (JSC::Heap::collect): 12597 * runtime/ConstructData.h: 12598 (JSC::): 12599 * runtime/ExceptionHelpers.cpp: 12600 (JSC::createUndefinedVariableError): 12601 (JSC::createInvalidParamError): 12602 (JSC::createNotAConstructorError): 12603 (JSC::createNotAFunctionError): 12604 (JSC::createNotAnObjectError): 12605 * runtime/Executable.cpp: Added. 12606 (JSC::EvalExecutable::generateBytecode): 12607 (JSC::ProgramExecutable::generateBytecode): 12608 (JSC::FunctionExecutable::generateBytecode): 12609 (JSC::EvalExecutable::generateJITCode): 12610 (JSC::ProgramExecutable::generateJITCode): 12611 (JSC::FunctionExecutable::generateJITCode): 12612 (JSC::FunctionExecutable::isHostFunction): 12613 (JSC::FunctionExecutable::markAggregate): 12614 (JSC::FunctionExecutable::reparseExceptionInfo): 12615 (JSC::EvalExecutable::reparseExceptionInfo): 12616 (JSC::FunctionExecutable::recompile): 12617 (JSC::FunctionExecutable::FunctionExecutable): 12618 * runtime/Executable.h: 12619 (JSC::ExecutableBase::~ExecutableBase): 12620 (JSC::ExecutableBase::ExecutableBase): 12621 (JSC::ExecutableBase::source): 12622 (JSC::ExecutableBase::sourceID): 12623 (JSC::ExecutableBase::lastLine): 12624 (JSC::ExecutableBase::usesEval): 12625 (JSC::ExecutableBase::usesArguments): 12626 (JSC::ExecutableBase::needsActivation): 12627 (JSC::ExecutableBase::astNode): 12628 (JSC::ExecutableBase::generatedJITCode): 12629 (JSC::ExecutableBase::getExecutablePool): 12630 (JSC::EvalExecutable::EvalExecutable): 12631 (JSC::EvalExecutable::bytecode): 12632 (JSC::EvalExecutable::varStack): 12633 (JSC::EvalExecutable::evalNode): 12634 (JSC::EvalExecutable::jitCode): 12635 (JSC::ProgramExecutable::ProgramExecutable): 12636 (JSC::ProgramExecutable::reparseExceptionInfo): 12637 (JSC::ProgramExecutable::bytecode): 12638 (JSC::ProgramExecutable::programNode): 12639 (JSC::ProgramExecutable::jitCode): 12640 (JSC::FunctionExecutable::FunctionExecutable): 12641 (JSC::FunctionExecutable::name): 12642 (JSC::FunctionExecutable::bytecode): 12643 (JSC::FunctionExecutable::generatedBytecode): 12644 (JSC::FunctionExecutable::usesEval): 12645 (JSC::FunctionExecutable::usesArguments): 12646 (JSC::FunctionExecutable::parameterCount): 12647 (JSC::FunctionExecutable::paramString): 12648 (JSC::FunctionExecutable::isGenerated): 12649 (JSC::FunctionExecutable::body): 12650 (JSC::FunctionExecutable::jitCode): 12651 (JSC::FunctionExecutable::createNativeThunk): 12652 * runtime/FunctionConstructor.cpp: 12653 (JSC::constructFunction): 12654 * runtime/FunctionPrototype.cpp: 12655 (JSC::functionProtoFuncToString): 12656 * runtime/JSActivation.cpp: 12657 (JSC::JSActivation::JSActivation): 12658 (JSC::JSActivation::markChildren): 12659 (JSC::JSActivation::isDynamicScope): 12660 (JSC::JSActivation::argumentsGetter): 12661 * runtime/JSActivation.h: 12662 (JSC::JSActivation::JSActivationData::JSActivationData): 12663 * runtime/JSFunction.cpp: 12664 (JSC::JSFunction::isHostFunction): 12665 (JSC::JSFunction::JSFunction): 12666 (JSC::JSFunction::~JSFunction): 12667 (JSC::JSFunction::markChildren): 12668 (JSC::JSFunction::getCallData): 12669 (JSC::JSFunction::call): 12670 (JSC::JSFunction::lengthGetter): 12671 (JSC::JSFunction::getConstructData): 12672 (JSC::JSFunction::construct): 12673 * runtime/JSFunction.h: 12674 (JSC::JSFunction::executable): 12675 (JSC::FunctionExecutable::make): 12676 * runtime/JSGlobalData.cpp: 12677 (JSC::JSGlobalData::JSGlobalData): 12678 (JSC::JSGlobalData::numericCompareFunction): 12679 * runtime/JSGlobalData.h: 12680 126812009-08-17 Mark Rowe <mrowe@apple.com> 12682 12683 Reviewed by Darin Adler. 12684 12685 Fix 300,000+ leaks seen during the regression tests. 12686 12687 EvalCodeCache::get was heap-allocating an EvalExecutable instance without adopting the initial reference. 12688 While fixing this we noticed that EvalExecutable was a RefCounted type that was sometimes stack allocated. 12689 To make this cleaner and to prevent clients from attempting to ref a stack-allocated instance, we move the 12690 refcounting down to a new CacheableEvalExecutable class that derives from EvalExecutable. EvalCodeCache::get 12691 now uses CacheableEvalExecutable::create and avoids the leak. 12692 12693 * bytecode/EvalCodeCache.h: 12694 (JSC::EvalCodeCache::get): 12695 * interpreter/Interpreter.cpp: 12696 (JSC::Interpreter::callEval): 12697 * runtime/Executable.h: 12698 (JSC::CacheableEvalExecutable::create): 12699 (JSC::CacheableEvalExecutable::CacheableEvalExecutable): 12700 127012009-08-17 Oliver Hunt <oliver@apple.com> 12702 12703 RS=Mark Rowe. 12704 12705 REGRESSION (r47292): Prototype.js is broken by ES5 Arguments changes 12706 https://bugs.webkit.org/show_bug.cgi?id=28341 12707 <rdar://problem/7145615> 12708 12709 Reverting r47292. Alas Prototype.js breaks with Arguments inheriting 12710 from Array as ES5 attempted. Prototype.js defines $A in terms of a 12711 function it places on (among other global objects) the Array prototype, 12712 thus breaking $A for arrays. 12713 12714 * runtime/Arguments.h: 12715 (JSC::Arguments::Arguments): 12716 * runtime/JSGlobalObject.cpp: 12717 (JSC::JSGlobalObject::reset): 12718 (JSC::JSGlobalObject::markChildren): 12719 * runtime/JSGlobalObject.h: 12720 (JSC::JSGlobalObject::JSGlobalObjectData::JSGlobalObjectData): 12721 * runtime/ObjectPrototype.cpp: 12722 (JSC::ObjectPrototype::ObjectPrototype): 12723 * runtime/ObjectPrototype.h: 12724 * tests/mozilla/ecma_3/Function/arguments-001.js: 12725 127262009-08-17 Peter Kasting <pkasting@google.com> 12727 12728 Reviewed by Steve Falkenburg. 12729 12730 https://bugs.webkit.org/show_bug.cgi?id=27323 12731 Only add Cygwin to the path when it isn't already there. This avoids 12732 causing problems for people who purposefully have non-Cygwin versions of 12733 executables like svn in front of the Cygwin ones in their paths. 12734 12735 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCoreCommon.vsprops: 12736 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCoreGenerated.vcproj: 12737 * JavaScriptCore.vcproj/WTF/WTFCommon.vsprops: 12738 * JavaScriptCore.vcproj/jsc/jscCommon.vsprops: 12739 * JavaScriptCore.vcproj/testapi/testapiCommon.vsprops: 12740 127412009-08-17 Xan Lopez <xlopez@igalia.com> 12742 12743 Reviewed by Mark Rowe. 12744 12745 Fix build with FAST_MALLOC_MATCH_VALIDATION enabled. 12746 12747 * wtf/FastMalloc.cpp: 12748 (WTF::fastMalloc): 12749 (WTF::fastCalloc): 12750 (WTF::fastRealloc): 12751 127522009-08-16 Holger Hans Peter Freyther <zecke@selfish.org> 12753 12754 Reviewed by Mark Rowe. 12755 12756 Fix crash on ./ecma_2/RegExp/exec-002.js. 12757 https://bugs.webkit.org/show_bug.cgi?id=28353 12758 12759 Change the order of freeParenthesesDisjunctionContext and 12760 popParenthesesDisjunctionContext on all call sites as the pop 12761 method is accessing backTrack->lastContext which is the context 12762 that is about to be freed. 12763 12764 * yarr/RegexInterpreter.cpp: 12765 (JSC::Yarr::Interpreter::parenthesesDoBacktrack): 12766 (JSC::Yarr::Interpreter::backtrackParentheses): 12767 127682009-08-16 Holger Hans Peter Freyther <zecke@selfish.org> 12769 12770 Reviewed by Mark Rowe. 12771 12772 https://bugs.webkit.org/show_bug.cgi?id=28352 12773 12774 Fix coding style violations. Use m_ for C++ class members. Remove 12775 trailing whitespace on empty lines. 12776 12777 * yarr/RegexInterpreter.cpp: 12778 (JSC::Yarr::Interpreter::ParenthesesDisjunctionContext::ParenthesesDisjunctionContext): 12779 (JSC::Yarr::Interpreter::tryConsumeCharacter): 12780 (JSC::Yarr::Interpreter::tryConsumeBackReference): 12781 (JSC::Yarr::Interpreter::parenthesesDoBacktrack): 12782 (JSC::Yarr::Interpreter::backtrackParentheses): 12783 (JSC::Yarr::ByteCompiler::ByteCompiler): 12784 (JSC::Yarr::ByteCompiler::compile): 12785 (JSC::Yarr::ByteCompiler::checkInput): 12786 (JSC::Yarr::ByteCompiler::assertionBOL): 12787 (JSC::Yarr::ByteCompiler::assertionEOL): 12788 (JSC::Yarr::ByteCompiler::assertionWordBoundary): 12789 (JSC::Yarr::ByteCompiler::atomPatternCharacter): 12790 (JSC::Yarr::ByteCompiler::atomCharacterClass): 12791 (JSC::Yarr::ByteCompiler::atomBackReference): 12792 (JSC::Yarr::ByteCompiler::atomParenthesesSubpatternBegin): 12793 (JSC::Yarr::ByteCompiler::atomParentheticalAssertionBegin): 12794 (JSC::Yarr::ByteCompiler::popParenthesesStack): 12795 (JSC::Yarr::ByteCompiler::closeAlternative): 12796 (JSC::Yarr::ByteCompiler::closeBodyAlternative): 12797 (JSC::Yarr::ByteCompiler::atomParenthesesEnd): 12798 (JSC::Yarr::ByteCompiler::regexBegin): 12799 (JSC::Yarr::ByteCompiler::alterantiveBodyDisjunction): 12800 (JSC::Yarr::ByteCompiler::alterantiveDisjunction): 12801 (JSC::Yarr::ByteCompiler::emitDisjunction): 12802 128032009-08-15 Mark Rowe <mrowe@apple.com> 12804 12805 Fix the build with JIT disabled. 12806 12807 * runtime/Arguments.h: Only compile the jitCode method when the JIT is enabled. 12808 * runtime/Executable.h: Include PrototypeFunction.h so the compiler knows what 12809 NativeFunctionWrapper is when the JIT is disabled. 12810 128112009-08-15 Adam Bergkvist <adam.bergkvist@ericsson.com> 12812 12813 Reviewed by Sam Weinig. 12814 12815 Added ENABLE_EVENTSOURCE flag. 12816 https://bugs.webkit.org/show_bug.cgi?id=14997 12817 12818 * Configurations/FeatureDefines.xcconfig: 12819 128202009-08-14 Gavin Barraclough <barraclough@apple.com> 12821 12822 * parser/Parser.h: 12823 (JSC::EvalExecutable::parse): 12824 (JSC::ProgramExecutable::parse): 12825 * runtime/Executable.h: 12826 128272009-08-14 Gavin Barraclough <barraclough@apple.com> 12828 12829 Reviewed by Oliver Hunt. 12830 12831 Remove AST nodes from use within the Runtime (outside of parsing), stage 1 12832 https://bugs.webkit.org/show_bug.cgi?id=28330 12833 12834 Remove the EvalNode and ProgramNode from use in the runtime. They still exist 12835 after this patch, but are hidden behind EvalExecutable and FunctionExecutable, 12836 and are also still reachable behind CodeBlock::m_ownerNode. 12837 12838 The next step will be to beat back FunctionBodyNode in the same fashion. 12839 Then remove the usage via CodeBlock, then only construct these nodes only on 12840 demand during bytecode generation. 12841 12842 * JavaScriptCore.xcodeproj/project.pbxproj: 12843 * bytecode/CodeBlock.h: 12844 (JSC::GlobalCodeBlock::GlobalCodeBlock): 12845 (JSC::GlobalCodeBlock::~GlobalCodeBlock): 12846 (JSC::ProgramCodeBlock::ProgramCodeBlock): 12847 (JSC::EvalCodeBlock::EvalCodeBlock): 12848 (JSC::FunctionCodeBlock::FunctionCodeBlock): 12849 (JSC::NativeCodeBlock::NativeCodeBlock): 12850 * bytecode/EvalCodeCache.h: 12851 (JSC::EvalCodeCache::get): 12852 * debugger/Debugger.cpp: 12853 (JSC::evaluateInGlobalCallFrame): 12854 * debugger/DebuggerCallFrame.cpp: 12855 (JSC::DebuggerCallFrame::evaluate): 12856 * interpreter/Interpreter.cpp: 12857 (JSC::Interpreter::callEval): 12858 (JSC::Interpreter::execute): 12859 * interpreter/Interpreter.h: 12860 * parser/Nodes.cpp: 12861 (JSC::FunctionBodyNode::createNativeThunk): 12862 (JSC::FunctionBodyNode::generateBytecode): 12863 (JSC::FunctionBodyNode::bytecodeForExceptionInfoReparse): 12864 * parser/Parser.h: 12865 (JSC::Parser::parse): 12866 (JSC::Parser::reparse): 12867 (JSC::Parser::parseFunctionFromGlobalCode): 12868 (JSC::::parse): 12869 * runtime/Completion.cpp: 12870 (JSC::checkSyntax): 12871 (JSC::evaluate): 12872 * runtime/Error.cpp: 12873 (JSC::throwError): 12874 * runtime/Error.h: 12875 * runtime/Executable.h: Added. 12876 (JSC::TemplateExecutable::TemplateExecutable): 12877 (JSC::TemplateExecutable::markAggregate): 12878 (JSC::TemplateExecutable::sourceURL): 12879 (JSC::TemplateExecutable::lineNo): 12880 (JSC::TemplateExecutable::bytecode): 12881 (JSC::TemplateExecutable::jitCode): 12882 (JSC::EvalExecutable::EvalExecutable): 12883 (JSC::ProgramExecutable::ProgramExecutable): 12884 * runtime/FunctionConstructor.cpp: 12885 (JSC::constructFunction): 12886 * runtime/FunctionConstructor.h: 12887 * runtime/JSGlobalData.cpp: 12888 (JSC::JSGlobalData::numericCompareFunction): 12889 * runtime/JSGlobalObject.cpp: 12890 (JSC::JSGlobalObject::~JSGlobalObject): 12891 (JSC::JSGlobalObject::markChildren): 12892 * runtime/JSGlobalObject.h: 12893 (JSC::JSGlobalObject::codeBlocks): 12894 * runtime/JSGlobalObjectFunctions.cpp: 12895 (JSC::globalFuncEval): 12896 128972009-08-14 Darin Adler <darin@apple.com> 12898 12899 Reviewed by Sam Weinig. 12900 12901 Rename the confusing isObject(<class>) to inherits(<class>). 12902 It still works on non-objects, returning false. 12903 12904 * runtime/ArrayConstructor.cpp: 12905 (JSC::arrayConstructorIsArray): Removed unneeded isObject call 12906 and updated remaining isObject call to new name, inherits. 12907 12908 * runtime/JSCell.h: Renamed isObject(<class>) to inherits(<class>) 12909 but more importantly, made it non-virtual (it was already inline) 12910 so it is now as fast as JSObject::inherits was. 12911 12912 * runtime/JSObject.h: Removed inherits function since the one 12913 in the base class is fine as-is. Also made various JSCell functions 12914 that should not be called on JSObject uncallable by making them 12915 both private and not implemented. 12916 (JSC::JSCell::inherits): Updated name. 12917 (JSC::JSValue::inherits): Ditto. 12918 12919 * debugger/Debugger.cpp: 12920 (JSC::Debugger::recompileAllJSFunctions): 12921 * interpreter/Interpreter.cpp: 12922 (JSC::Interpreter::unwindCallFrame): 12923 * runtime/ArrayPrototype.cpp: 12924 (JSC::arrayProtoFuncToString): 12925 (JSC::arrayProtoFuncToLocaleString): 12926 (JSC::arrayProtoFuncConcat): 12927 * runtime/BooleanPrototype.cpp: 12928 (JSC::booleanProtoFuncToString): 12929 (JSC::booleanProtoFuncValueOf): 12930 * runtime/DateConstructor.cpp: 12931 (JSC::constructDate): 12932 * runtime/DatePrototype.cpp: 12933 (JSC::dateProtoFuncToString): 12934 (JSC::dateProtoFuncToUTCString): 12935 (JSC::dateProtoFuncToISOString): 12936 (JSC::dateProtoFuncToDateString): 12937 (JSC::dateProtoFuncToTimeString): 12938 (JSC::dateProtoFuncToLocaleString): 12939 (JSC::dateProtoFuncToLocaleDateString): 12940 (JSC::dateProtoFuncToLocaleTimeString): 12941 (JSC::dateProtoFuncGetTime): 12942 (JSC::dateProtoFuncGetFullYear): 12943 (JSC::dateProtoFuncGetUTCFullYear): 12944 (JSC::dateProtoFuncToGMTString): 12945 (JSC::dateProtoFuncGetMonth): 12946 (JSC::dateProtoFuncGetUTCMonth): 12947 (JSC::dateProtoFuncGetDate): 12948 (JSC::dateProtoFuncGetUTCDate): 12949 (JSC::dateProtoFuncGetDay): 12950 (JSC::dateProtoFuncGetUTCDay): 12951 (JSC::dateProtoFuncGetHours): 12952 (JSC::dateProtoFuncGetUTCHours): 12953 (JSC::dateProtoFuncGetMinutes): 12954 (JSC::dateProtoFuncGetUTCMinutes): 12955 (JSC::dateProtoFuncGetSeconds): 12956 (JSC::dateProtoFuncGetUTCSeconds): 12957 (JSC::dateProtoFuncGetMilliSeconds): 12958 (JSC::dateProtoFuncGetUTCMilliseconds): 12959 (JSC::dateProtoFuncGetTimezoneOffset): 12960 (JSC::dateProtoFuncSetTime): 12961 (JSC::setNewValueFromTimeArgs): 12962 (JSC::setNewValueFromDateArgs): 12963 (JSC::dateProtoFuncSetYear): 12964 (JSC::dateProtoFuncGetYear): 12965 * runtime/FunctionPrototype.cpp: 12966 (JSC::functionProtoFuncToString): 12967 * runtime/JSActivation.cpp: 12968 (JSC::JSActivation::argumentsGetter): 12969 * runtime/JSValue.h: 12970 * runtime/RegExpConstructor.cpp: 12971 (JSC::constructRegExp): 12972 * runtime/RegExpPrototype.cpp: 12973 (JSC::regExpProtoFuncTest): 12974 (JSC::regExpProtoFuncExec): 12975 (JSC::regExpProtoFuncCompile): 12976 (JSC::regExpProtoFuncToString): 12977 * runtime/ScopeChain.cpp: 12978 (JSC::ScopeChain::localDepth): 12979 * runtime/StringPrototype.cpp: 12980 (JSC::stringProtoFuncReplace): 12981 (JSC::stringProtoFuncToString): 12982 (JSC::stringProtoFuncMatch): 12983 (JSC::stringProtoFuncSearch): 12984 (JSC::stringProtoFuncSplit): 12985 Updated to new name, inherits, from old name, isObject. 12986 129872009-07-31 Harald Fernengel <harald.fernengel@nokia.com> 12988 12989 Reviewed by Simon Hausmann. 12990 12991 Adding QNX as a platform. Currently only tested with Qt. 12992 12993 https://bugs.webkit.org/show_bug.cgi?id=27885 12994 12995 * JavaScriptCore/runtime/Collector.cpp: Added retrieving of stack base 12996 since QNX doesn't have the pthread _nt functions 12997 * JavaScriptCore/wtf/Platform.h: Added WTF_PLATFORM_QNX and corresponding 12998 defines 12999 * WebCore/bridge/npapi.h: Build fix for missing typedefs on QNX 13000 130012009-08-14 Gabor Loki <loki@inf.u-szeged.hu> 13002 13003 Reviewed by Simon Hausmann. 13004 13005 Currently generic ARM and ARMv7 platforms work only with JSVALUE32 13006 https://bugs.webkit.org/show_bug.cgi?id=28300 13007 13008 * wtf/Platform.h: 13009 130102009-08-14 Gabor Loki <loki@inf.u-szeged.hu> 13011 13012 Reviewed by Simon Hausmann. 13013 13014 Enable JIT on ARM for QT by default 13015 https://bugs.webkit.org/show_bug.cgi?id=28259 13016 13017 * wtf/Platform.h: 13018 130192009-08-14 Gabor Loki <loki@inf.u-szeged.hu> 13020 13021 Reviewed by Simon Hausmann. 13022 13023 Enable YARR_JIT on ARM for QT by default 13024 https://bugs.webkit.org/show_bug.cgi?id=28259 13025 13026 * wtf/Platform.h: 13027 130282009-08-14 Oliver Hunt <oliver@apple.com> 13029 13030 Reviewed by Gavin Barraclough. 13031 13032 [ES5] Arguments object should inherit from Array 13033 https://bugs.webkit.org/show_bug.cgi?id=28298 13034 13035 Make the Arguments object conform to the behaviour specified in ES5. 13036 The simple portion of this is to make Arguments use Array.prototype 13037 as its prototype rather than Object.prototype. 13038 13039 The spec then requires us to set instance.constructor to the pristine 13040 Object constructor, and instance.toString and instance.toLocaleString 13041 to the pristine versions from Object.prototype. To do this we now 13042 make the ObjectPrototype constructor return its toString and 13043 toLocaleString functions (similar to the call and apply functions 13044 from FunctionPrototype). 13045 13046 Oddly enough this reports itself as a slight win, but given the code 13047 isn't hit in the tests that claim to have improved I put this down to 13048 code motion. 13049 13050 * runtime/Arguments.h: 13051 (JSC::Arguments::Arguments): 13052 (JSC::Arguments::initializeStandardProperties): 13053 * runtime/JSGlobalObject.cpp: 13054 (JSC::JSGlobalObject::reset): 13055 (JSC::JSGlobalObject::markChildren): 13056 * runtime/JSGlobalObject.h: 13057 (JSC::JSGlobalObject::JSGlobalObjectData::JSGlobalObjectData): 13058 (JSC::JSGlobalObject::objectConstructor): 13059 (JSC::JSGlobalObject::objectToStringFunction): 13060 (JSC::JSGlobalObject::objectToLocaleStringFunction): 13061 * runtime/ObjectPrototype.cpp: 13062 (JSC::ObjectPrototype::ObjectPrototype): 13063 * runtime/ObjectPrototype.h: 13064 * tests/mozilla/ecma_3/Function/arguments-001.js: 13065 Update test to new es5 behaviour 13066 130672009-08-14 Oliver Hunt <oliver@apple.com> 13068 13069 Remove MarkStack::drain from the JSC exports file 13070 13071 MarkStack::drain is now marked inline, the including it in the exports file 13072 produces an ld warning 13073 13074 * JavaScriptCore.exp: 13075 130762009-08-13 Sam Weinig <sam@webkit.org> 13077 13078 Reviewed by Oliver Hunt. 13079 13080 Remove accidentally left in debugging statement. 13081 13082 * runtime/JSArray.h: 13083 (JSC::MarkStack::drain): 13084 130852009-08-13 Oliver Hunt <oliver@apple.com> 13086 13087 Reviewed by Maciej Stachowiak. 13088 13089 [ES5] Implement Array.isArray 13090 https://bugs.webkit.org/show_bug.cgi?id=28296 13091 13092 Add support for Array.isArray to the Array constructor 13093 13094 * runtime/ArrayConstructor.cpp: 13095 (JSC::ArrayConstructor::ArrayConstructor): 13096 (JSC::arrayConstructorIsArray): 13097 * runtime/ArrayConstructor.h: 13098 * runtime/CommonIdentifiers.h: 13099 * runtime/JSArray.h: 13100 (JSC::MarkStack::drain): 13101 * runtime/JSGlobalObject.cpp: 13102 (JSC::JSGlobalObject::reset): 13103 131042009-08-13 Oliver Hunt <oliver@apple.com> 13105 13106 Reviewed by NOBODY (Buildfix). 13107 13108 Attempt to fix windows build 13109 13110 * runtime/Collector.cpp: 13111 131122009-08-13 Oliver Hunt <oliver@apple.com> 13113 13114 Reviewed by Maciej Stachowiak. 13115 13116 Devirtualise marking 13117 https://bugs.webkit.org/show_bug.cgi?id=28294 13118 13119 Add a bit to TypeInfo to indicate that an object uses the standard 13120 JSObject::markChildren method. This allows us to devirtualise marking 13121 of most objects (though a branch is still needed). We also add a branch 13122 to identify arrays thus devirtualising marking in that case as well. 13123 13124 In order to make the best use of this devirtualisation I've also reworked 13125 the MarkStack::drain() logic to make the iteration more efficient. 13126 13127 * API/JSCallbackConstructor.h: 13128 (JSC::JSCallbackConstructor::createStructure): 13129 * API/JSCallbackFunction.h: 13130 (JSC::JSCallbackFunction::createStructure): 13131 * JavaScriptCore.exp: 13132 * runtime/BooleanObject.h: 13133 (JSC::BooleanObject::createStructure): 13134 * runtime/FunctionPrototype.h: 13135 (JSC::FunctionPrototype::createStructure): 13136 * runtime/InternalFunction.h: 13137 (JSC::InternalFunction::createStructure): 13138 * runtime/JSAPIValueWrapper.h: 13139 (JSC::JSAPIValueWrapper::JSAPIValueWrapper): 13140 * runtime/JSArray.cpp: 13141 (JSC::JSArray::markChildren): 13142 * runtime/JSArray.h: 13143 (JSC::JSArray::markChildrenDirect): 13144 (JSC::MarkStack::drain): 13145 * runtime/JSByteArray.cpp: 13146 (JSC::JSByteArray::createStructure): 13147 * runtime/JSCell.h: 13148 (JSC::MarkStack::append): 13149 * runtime/JSGlobalData.cpp: 13150 (JSC::JSGlobalData::JSGlobalData): 13151 * runtime/JSNumberCell.h: 13152 (JSC::JSNumberCell::createStructure): 13153 * runtime/JSONObject.h: 13154 (JSC::JSONObject::createStructure): 13155 * runtime/JSObject.cpp: 13156 (JSC::JSObject::markChildren): 13157 * runtime/JSObject.h: 13158 (JSC::JSObject::markChildrenDirect): 13159 (JSC::JSObject::createStructure): 13160 * runtime/JSString.h: 13161 (JSC::JSString::createStructure): 13162 * runtime/JSType.h: 13163 (JSC::): 13164 * runtime/MarkStack.h: 13165 (JSC::MarkStack::MarkStack): 13166 (JSC::MarkStack::MarkSet::MarkSet): 13167 (JSC::MarkStack::MarkStackArray::last): 13168 * runtime/MathObject.h: 13169 (JSC::MathObject::createStructure): 13170 * runtime/NumberConstructor.h: 13171 (JSC::NumberConstructor::createStructure): 13172 * runtime/NumberObject.h: 13173 (JSC::NumberObject::createStructure): 13174 * runtime/RegExpConstructor.h: 13175 (JSC::RegExpConstructor::createStructure): 13176 * runtime/RegExpObject.h: 13177 (JSC::RegExpObject::createStructure): 13178 * runtime/StringObjectThatMasqueradesAsUndefined.h: 13179 (JSC::StringObjectThatMasqueradesAsUndefined::createStructure): 13180 * runtime/TypeInfo.h: 13181 (JSC::TypeInfo::hasDefaultMark): 13182 131832009-08-13 Darin Adler <darin@apple.com> 13184 13185 Reviewed by Mark Rowe. 13186 13187 Some small bits of housekeeping. 13188 13189 * JavaScriptCore.xcodeproj/project.pbxproj: Make Parser.h 13190 project instead of private. Remove JSONObject.lut.h. 13191 13192 * assembler/ARMAssembler.h: Remove unneeded WTF prefix. 13193 * assembler/AssemblerBufferWithConstantPool.h: Ditto. 13194 * bytecompiler/BytecodeGenerator.h: Ditto. 13195 13196 * wtf/SegmentedVector.h: Add a "using" statement as we do 13197 with the other WTF headers. 13198 131992009-08-13 Darin Adler <darin@apple.com> 13200 13201 Fix Tiger build. 13202 13203 * parser/Grammar.y: Use a template function so we can compile 13204 setStatementLocation even if it comes before YYLTYPE is defined. 13205 132062009-08-13 Darin Adler <darin@apple.com> 13207 13208 Reviewed by George Staikos. 13209 13210 Too much use of void* in Grammar.y 13211 https://bugs.webkit.org/show_bug.cgi?id=28287 13212 13213 * parser/Grammar.y: Changed all the helper functions to 13214 take a JSGlobalData* instead of a void*. A couple formatting 13215 tweaks that I missed when breaking this into pieces. 13216 132172009-08-13 Darin Adler <darin@apple.com> 13218 13219 Reviewed by George Staikos. 13220 13221 Another part of https://bugs.webkit.org/show_bug.cgi?id=28287 13222 13223 * parser/Grammar.y: Reduced and sorted includes. Tweaked comment 13224 format. Marked a few more functions inline. 13225 132262009-08-13 Darin Adler <darin@apple.com> 13227 13228 Reviewed by George Staikos. 13229 13230 Another part of https://bugs.webkit.org/show_bug.cgi?id=28287 13231 13232 * parser/Grammar.y: Pass the number to the PropertyNode instead of 13233 first turning it into an Identifier. 13234 13235 * parser/NodeConstructors.h: 13236 (JSC::PropertyNode::PropertyNode): Add an overload that takes a double 13237 so the code to convert to a string can be here instead of Grammar.y. 13238 * parser/Nodes.h: Ditto. 13239 132402009-08-13 Darin Adler <darin@apple.com> 13241 13242 Reviewed by George Staikos. 13243 13244 Another part of https://bugs.webkit.org/show_bug.cgi?id=28287 13245 13246 * parser/Grammar.y: Eliminate the DBG macro. 13247 132482009-08-13 Darin Adler <darin@apple.com> 13249 13250 Reviewed by George Staikos. 13251 13252 Another part of https://bugs.webkit.org/show_bug.cgi?id=28287 13253 13254 * parser/Grammar.y: Eliminate the SET_EXCEPTION_LOCATION macro. 13255 132562009-08-13 Darin Adler <darin@apple.com> 13257 13258 Reviewed by George Staikos. 13259 13260 George asked me to break the patch from 13261 https://bugs.webkit.org/show_bug.cgi?id=28287 13262 into smaller pieces and land it in stages. 13263 13264 * parser/Grammar.y: Eliminate the LEXER macro. 13265 132662009-08-13 Mark Rowe <mrowe@apple.com> 13267 13268 Try some more to fix the Windows build. 13269 13270 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: Export a new symbol. 13271 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: Ditto. 13272 132732009-08-13 Mark Rowe <mrowe@apple.com> 13274 13275 Try and fix the Windows build. 13276 13277 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: Export a new symbol. 13278 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: Ditto. 13279 132802009-08-13 Darin Adler <darin@apple.com> 13281 13282 Reviewed by David Levin. 13283 13284 JavaScriptCore tweaks to get ready for the parser arena 13285 https://bugs.webkit.org/show_bug.cgi?id=28243 13286 13287 Eliminate dependencies on Nodes.h outside JavaScriptCore, 13288 and cut down on them inside JavaScriptCore. 13289 13290 Change regular expression parsing to use identifiers as 13291 with other strings we parse. 13292 13293 Fix a couple things that are needed to use const Identifier 13294 more, which will be part of the parser arena work. 13295 13296 * JavaScriptCore.exp: Resorted and updated. 13297 13298 * JavaScriptCore.xcodeproj/project.pbxproj: Changed 13299 CollectorHeapIterator.h to be project-internal. 13300 13301 * bytecompiler/BytecodeGenerator.cpp: 13302 (JSC::BytecodeGenerator::emitPushNewScope): Added const. 13303 * bytecompiler/BytecodeGenerator.h: Ditto. 13304 13305 * debugger/Debugger.cpp: 13306 (JSC::Debugger::recompileAllJSFunctions): Moved this function 13307 here from WebCore. Here is better since it uses so many internals. 13308 Removed unimportant optimization for the no listener case. 13309 * debugger/Debugger.h: Ditto. Also removed unneeded include 13310 and tweaked formatting and comments. 13311 13312 * debugger/DebuggerCallFrame.cpp: 13313 (JSC::DebuggerCallFrame::functionName): Call asFunction instead 13314 of doing the unchecked static_cast. 13315 (JSC::DebuggerCallFrame::calculatedFunctionName): Ditto. 13316 13317 * jit/JITStubs.cpp: 13318 (JSC::op_call_JSFunction): Call isHostFunction on the body rather 13319 than on the JSFunction. 13320 (JSC::vm_lazyLinkCall): Ditto. 13321 (JSC::op_construct_JSConstruct): Ditto. 13322 13323 * parser/Grammar.y: Changed callers to use new scanRegExp with 13324 out arguments instead of relying on state in the Lexer. And 13325 callers that just want to skip a regular expression to use 13326 skipRegExp. 13327 13328 * parser/Lexer.cpp: 13329 (JSC::Lexer::scanRegExp): Changed to use out arguments, and to 13330 add a prefix argument so we can add in the "=" character as needed. 13331 Also rewrote to streamline the logic a bit inspired by suggestions 13332 by David Levin. 13333 (JSC::Lexer::skipRegExp): Added. Version of the function above that 13334 does not actually put the regular expression into a string. 13335 (JSC::Lexer::clear): Removed code to clear m_pattern and m_flags. 13336 * parser/Lexer.h: Changed scanRegExp to have out arguments. Added 13337 skipRegExp. Eliminated pattern, flags, m_pattern, and m_flags. 13338 13339 * parser/NodeConstructors.h: 13340 (JSC::RegExpNode::RegExpNode): Changed to take const Identifier&. 13341 * parser/Nodes.cpp: 13342 (JSC::RegExpNode::emitBytecode): Changed since m_pattern and 13343 m_flags are now Identifier instead of UString. 13344 (JSC::FunctionBodyNode::make): Moved this function here instead 13345 of putting it in the JSFunction.h header. 13346 * parser/Nodes.h: Changed RegExpNode to use Identifier. 13347 13348 * profiler/Profiler.cpp: 13349 (JSC::Profiler::createCallIdentifier): Changed to use isHostFunction 13350 on the body instead of on the JSFunction object. 13351 * runtime/FunctionPrototype.cpp: 13352 (JSC::functionProtoFuncToString): Ditto. 13353 13354 * runtime/JSFunction.cpp: 13355 (JSC::JSFunction::isHostFunction): Moved here from header. 13356 (JSC::JSFunction::isHostFunctionNonInline): Added. 13357 (JSC::JSFunction::JSFunction): Removed unneeded initialization of 13358 m_body to 0. 13359 (JSC::JSFunction::setBody): Moved here from header. 13360 13361 * runtime/JSFunction.h: Removed unneeded includes. Moved private 13362 constructor down to the private section. Made virtual functions 13363 private. Removed unneeded overload of setBody and moved the body 13364 of the function into the .cpp file. Changed assertions to use 13365 the non-inline version of isHostFunction. 13366 13367 * runtime/PropertySlot.cpp: 13368 (JSC::PropertySlot::functionGetter): Use asFunction instead 13369 of doing the unchecked static_cast. 13370 13371 * wtf/SegmentedVector.h: 13372 (WTF::SegmentedVector::isEmpty): Added. 13373 133742009-08-13 Mark Rowe <mrowe@apple.com> 13375 13376 Rubber-stamped by Darin Adler. 13377 13378 Use the version of operator new that takes a JSGlobalData when allocating FuncDeclNode and FuncExprNode 13379 from within the grammar to prevent these nodes from being leaked. 13380 13381 * parser/Grammar.y: 13382 133832009-08-13 Simon Hausmann <simon.hausmann@nokia.com> 13384 13385 Reviewed by Ariya Hidayat. 13386 13387 Remove the special-case for Qt wrt JSVALUE_32 introduced in 13388 r46709. It must've been a dependency issue on the bot, as 13389 after a manual build all the tests pass on amd64 and ia32. 13390 13391 * wtf/Platform.h: 13392 133932009-08-12 Gabor Loki <loki@inf.u-szeged.hu> 13394 13395 Reviewed by Gavin Barraclough. 13396 13397 Add optimize call and property access support for ARM JIT. 13398 https://bugs.webkit.org/show_bug.cgi?id=24986 13399 13400 For tightly coupled sequences the BEGIN_UNINTERRUPTED_SEQUENCE and 13401 END_UNINTERRUPTED_SEQUENCE macros have been introduced which ensure 13402 space for instructions and constants of the named sequence. This 13403 method is vital for those architecture which are using constant pool. 13404 13405 The 'latePatch' method - which was linked to JmpSrc - is replaced with 13406 a port specific solution (each calls are marked to place their address 13407 on the constant pool). 13408 13409 * assembler/ARMAssembler.cpp: 13410 (JSC::ARMAssembler::linkBranch): 13411 (JSC::ARMAssembler::executableCopy): Add extra align for constant pool. 13412 * assembler/ARMAssembler.h: 13413 (JSC::ARMAssembler::JmpSrc::JmpSrc): 13414 (JSC::ARMAssembler::sizeOfConstantPool): 13415 (JSC::ARMAssembler::jmp): 13416 (JSC::ARMAssembler::linkCall): 13417 * assembler/ARMv7Assembler.h: 13418 * assembler/AbstractMacroAssembler.h: 13419 * assembler/AssemblerBufferWithConstantPool.h: 13420 (JSC::AssemblerBufferWithConstantPool::flushIfNoSpaceFor): Fix the 13421 computation of the remaining space. 13422 * assembler/MacroAssemblerARM.h: 13423 (JSC::MacroAssemblerARM::branch32): 13424 (JSC::MacroAssemblerARM::nearCall): 13425 (JSC::MacroAssemblerARM::call): 13426 (JSC::MacroAssemblerARM::branchPtrWithPatch): 13427 (JSC::MacroAssemblerARM::ensureSpace): 13428 (JSC::MacroAssemblerARM::sizeOfConstantPool): 13429 (JSC::MacroAssemblerARM::prepareCall): 13430 * assembler/X86Assembler.h: 13431 * jit/JIT.h: 13432 * jit/JITCall.cpp: 13433 (JSC::JIT::compileOpCall): 13434 * jit/JITInlineMethods.h: 13435 (JSC::JIT::beginUninterruptedSequence): 13436 (JSC::JIT::endUninterruptedSequence): 13437 * jit/JITPropertyAccess.cpp: 13438 (JSC::JIT::emit_op_method_check): 13439 (JSC::JIT::compileGetByIdHotPath): 13440 (JSC::JIT::compileGetByIdSlowCase): 13441 (JSC::JIT::emit_op_put_by_id): 13442 134432009-08-12 Gavin Barraclough <barraclough@apple.com> 13444 13445 Rubber Stamped by Dave Kilzer. 13446 13447 Disable WTF_USE_JSVALUE32_64 on iPhone for now (support not yet added for ARMv7). 13448 13449 * wtf/Platform.h: 13450 134512009-08-12 Gavin Barraclough <barraclough@apple.com> 13452 13453 Reviewed by Maciej Stachoviak. 13454 13455 Ooops - moved code that had been accidentally added to op_new_func instead of 13456 op_new_func_exp, to where it shoulds be. 13457 13458 * interpreter/Interpreter.cpp: 13459 (JSC::Interpreter::privateExecute): 13460 * wtf/Platform.h: 13461 134622009-08-12 Ada Chan <adachan@apple.com> 13463 13464 Added workaround for the limitation that VirtualFree with MEM_RELEASE 13465 can only accept the base address returned by VirtualAlloc when the region 13466 was reserved and it can only free the entire region, and not a part of it. 13467 13468 Reviewed by Oliver Hunt. 13469 13470 * runtime/MarkStack.h: 13471 (JSC::MarkStack::MarkStackArray::shrinkAllocation): 13472 * runtime/MarkStackWin.cpp: 13473 (JSC::MarkStack::releaseStack): 13474 134752009-08-12 Balazs Kelemen <kelemen.balazs.3@stud.u-szeged.hu> 13476 13477 Reviewed by Ariya Hidayat. 13478 13479 Build fix: use std::numeric_limits<long long>::min() instead of LLONG_MIN 13480 since LLONG_MIN is not defined in standard c++. 13481 13482 * runtime/UString.cpp: 13483 (JSC::UString::from): 13484 134852009-08-12 Benjamin Otte <otte@gnome.org> 13486 13487 Reviewed by Jan Alonzo. 13488 13489 Buildfix for Gtk platforms debug builds. 13490 13491 * GNUmakefile.am: Choose MarkStackPosix.cpp or MarkStackWin.cpp 13492 depending on platform. 13493 134942009-08-12 Simon Hausmann <simon.hausmann@nokia.com> 13495 13496 Prospective build fix for Mac and 32-bit Windows. 13497 13498 * runtime/UString.cpp: Include wtf/StringExtras.h for snprintf. 13499 (JSC::UString::from): Use %lld instead of %I64d for snprintf 13500 on non-windows platforms. 13501 135022009-08-12 Prasanth Ullattil <prasanth.ullattil@nokia.com> 13503 13504 Reviewed by Simon Hausmann. 13505 13506 Fix compile error on 64Bit Windows, when UString::from 13507 is called with an intptr_t. 13508 13509 Added new UString::From overload with long long parameter. 13510 13511 Thanks to Holger for the long long idea. 13512 13513 * runtime/UString.cpp: 13514 (JSC::UString::from): 13515 * runtime/UString.h: 13516 135172009-08-11 Oliver Hunt <oliver@apple.com> 13518 13519 Reviewed by Mark Rowe. 13520 13521 Minor style fixes. 13522 13523 * runtime/UString.h: 13524 (JSC::UString::Rep::createEmptyBuffer): 13525 * wtf/FastMalloc.h: 13526 (WTF::TryMallocReturnValue::getValue): 13527 135282009-08-11 Oliver Hunt <oliver@apple.com> 13529 13530 Reviewed by Gavin Barraclough. 13531 13532 Make it harder to misuse try* allocation routines 13533 https://bugs.webkit.org/show_bug.cgi?id=27469 13534 13535 Jump through a few hoops to make it much harder to accidentally 13536 miss null-checking of values returned by the try-* allocation 13537 routines. 13538 13539 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 13540 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 13541 * JavaScriptCore.xcodeproj/project.pbxproj: 13542 * runtime/JSArray.cpp: 13543 (JSC::JSArray::putSlowCase): 13544 (JSC::JSArray::increaseVectorLength): 13545 * runtime/StringPrototype.cpp: 13546 (JSC::stringProtoFuncFontsize): 13547 (JSC::stringProtoFuncLink): 13548 * runtime/UString.cpp: 13549 (JSC::allocChars): 13550 (JSC::reallocChars): 13551 (JSC::expandCapacity): 13552 (JSC::UString::Rep::reserveCapacity): 13553 (JSC::UString::expandPreCapacity): 13554 (JSC::createRep): 13555 (JSC::concatenate): 13556 (JSC::UString::spliceSubstringsWithSeparators): 13557 (JSC::UString::replaceRange): 13558 (JSC::UString::append): 13559 (JSC::UString::operator=): 13560 * runtime/UString.h: 13561 (JSC::UString::Rep::createEmptyBuffer): 13562 * wtf/FastMalloc.cpp: 13563 (WTF::tryFastZeroedMalloc): 13564 (WTF::tryFastMalloc): 13565 (WTF::tryFastCalloc): 13566 (WTF::tryFastRealloc): 13567 (WTF::TCMallocStats::tryFastMalloc): 13568 (WTF::TCMallocStats::tryFastCalloc): 13569 (WTF::TCMallocStats::tryFastRealloc): 13570 * wtf/FastMalloc.h: 13571 (WTF::TryMallocReturnValue::TryMallocReturnValue): 13572 (WTF::TryMallocReturnValue::~TryMallocReturnValue): 13573 (WTF::TryMallocReturnValue::operator PossiblyNull<T>): 13574 (WTF::TryMallocReturnValue::getValue): 13575 * wtf/Platform.h: 13576 * wtf/PossiblyNull.h: Added. 13577 (WTF::PossiblyNull::PossiblyNull): 13578 (WTF::PossiblyNull::~PossiblyNull): 13579 (WTF::::getValue): 13580 135812009-08-11 Gavin Barraclough <barraclough@apple.com> 13582 13583 Reviewed by NOBODY (build fix part deux). 13584 13585 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 13586 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 13587 135882009-08-11 Gavin Barraclough <barraclough@apple.com> 13589 13590 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 13591 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 13592 135932009-08-11 Gavin Barraclough <barraclough@apple.com> 13594 13595 Reviewed by Oliver Hunt. 13596 13597 Restrict use of FuncDeclNode & FuncExprNode to the parser. 13598 https://bugs.webkit.org/show_bug.cgi?id=28209 13599 13600 These objects were also being referenced from the CodeBlock. By changing this 13601 to just retain pointers to FunctionBodyNodes these classes can be restricted to 13602 use during parsing. 13603 13604 No performance impact (or sub-percent progression). 13605 13606 * JavaScriptCore.exp: 13607 Update symbols. 13608 13609 * bytecode/CodeBlock.cpp: 13610 (JSC::CodeBlock::mark): 13611 (JSC::CodeBlock::reparseForExceptionInfoIfNecessary): 13612 (JSC::CodeBlock::shrinkToFit): 13613 * bytecode/CodeBlock.h: 13614 (JSC::CodeBlock::addFunction): 13615 (JSC::CodeBlock::function): 13616 Unify m_functions & m_functionExpressions into a single Vector<RefPtr<FuncExprNode> >. 13617 13618 * bytecompiler/BytecodeGenerator.cpp: 13619 (JSC::BytecodeGenerator::BytecodeGenerator): 13620 (JSC::BytecodeGenerator::addConstant): 13621 (JSC::BytecodeGenerator::emitNewFunction): 13622 (JSC::BytecodeGenerator::emitNewFunctionExpression): 13623 * bytecompiler/BytecodeGenerator.h: 13624 FunctionStacks now contain FunctionBodyNodes not FuncDeclNodes. 13625 13626 * interpreter/Interpreter.cpp: 13627 (JSC::Interpreter::execute): 13628 (JSC::Interpreter::privateExecute): 13629 Update to reflect chnages in CodeBlock. 13630 13631 * jit/JITOpcodes.cpp: 13632 (JSC::JIT::emit_op_new_func_exp): 13633 * jit/JITStubs.cpp: 13634 (JSC::DEFINE_STUB_FUNCTION): 13635 * jit/JITStubs.h: 13636 (JSC::): 13637 Update to reflect chnages in CodeBlock. 13638 13639 * parser/Grammar.y: 13640 FunctionStacks now contain FunctionBodyNodes not FuncDeclNodes. 13641 13642 * parser/NodeConstructors.h: 13643 (JSC::FuncExprNode::FuncExprNode): 13644 (JSC::FuncDeclNode::FuncDeclNode): 13645 * parser/Nodes.cpp: 13646 (JSC::ScopeNodeData::mark): 13647 (JSC::FunctionBodyNode::finishParsing): 13648 * parser/Nodes.h: 13649 (JSC::FunctionBodyNode::ident): 13650 Move m_ident & make methods from FuncDeclNode & FuncExprNode to FunctionBodyNode. 13651 13652 * runtime/JSFunction.h: 13653 (JSC::FunctionBodyNode::make): 13654 Make this method inline (was FuncDeclNode::makeFunction). 13655 136562009-08-11 Oliver Hunt <oliver@apple.com> 13657 13658 Reviewed by Gavin Barraclough. 13659 13660 Native JSON.stringify does not omit functions 13661 https://bugs.webkit.org/show_bug.cgi?id=28117 13662 13663 Objects that are callable should be treated as undefined when 13664 serialising to JSON. 13665 13666 * runtime/JSONObject.cpp: 13667 (JSC::Stringifier::appendStringifiedValue): 13668 136692009-08-11 Oliver Hunt <oliver@apple.com> 13670 13671 Reviewed by Geoff Garen. 13672 13673 REGRESSION: Hang/crash in BytecodeGenerator::constRegisterFor loading simple page 13674 https://bugs.webkit.org/show_bug.cgi?id=28169 13675 13676 Handle the case where someone has attempted to shadow a property 13677 on the global object with a constant. 13678 13679 * bytecompiler/BytecodeGenerator.cpp: 13680 (JSC::BytecodeGenerator::constRegisterFor): 13681 * parser/Nodes.cpp: 13682 (JSC::ConstDeclNode::emitCodeSingle): 13683 136842009-08-11 John Gregg <johnnyg@google.com> 13685 13686 Reviewed by Maciej Stachowiak. 13687 13688 Desktop Notifications API 13689 https://bugs.webkit.org/show_bug.cgi?id=25463 13690 13691 Adds ENABLE_NOTIFICATION flag. 13692 13693 * Configurations/FeatureDefines.xcconfig: 13694 * wtf/Platform.h: 13695 136962009-08-11 Maxime Simon <simon.maxime@gmail.com> 13697 13698 Reviewed by Eric Seidel. 13699 13700 Modifications on JavaScriptCore to allow Haiku port. 13701 https://bugs.webkit.org/show_bug.cgi?id=28121 13702 13703 * runtime/Collector.cpp: Haiku doesn't have sys/mman.h, using OS.h instead. 13704 (JSC::currentThreadStackBase): Haiku uses its own threading system. 13705 * wtf/Platform.h: Defining all Haiku platform values. 13706 * wtf/haiku/MainThreadHaiku.cpp: Adding a missing header (NotImplemented.h). 13707 137082009-08-11 Jessie Berlin <jberlin@apple.com> 13709 13710 Reviewed by Adam Roben. 13711 13712 Fix windows build. 13713 13714 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 13715 137162009-08-11 Csaba Osztrogonac <oszi@inf.u-szeged.hu> 13717 13718 Reviewed by Tor Arne Vestbø. 13719 13720 Buildfix for Qt-win platforms. 13721 13722 * JavaScriptCore.pri: Choose MarkStackPosix.cpp or MarkStackWin.cpp depend on platform. 13723 137242009-08-10 Oliver Hunt <oliver@apple.com> 13725 13726 Reviewed by NOBODY (And another build fix). 13727 13728 Add new exports for MSVC 13729 13730 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 13731 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 13732 * JavaScriptCore.xcodeproj/project.pbxproj: 13733 137342009-08-10 Oliver Hunt <oliver@apple.com> 13735 13736 Reviewed by NOBODY (yet another build fix). 13737 13738 Remove obsolete entries from MSVC exports file 13739 13740 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 13741 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 13742 137432009-08-10 Oliver Hunt <oliver@apple.com> 13744 13745 Add includes needed for non-allinonefile builds 13746 13747 * runtime/GetterSetter.h: 13748 * runtime/ScopeChain.h: 13749 137502009-08-10 Oliver Hunt <oliver@apple.com> 13751 13752 Fix export file for last build fix 13753 13754 * JavaScriptCore.exp: 13755 137562009-08-10 Oliver Hunt <oliver@apple.com> 13757 13758 Hoist page size initialization into platform specific code. 13759 13760 * jit/ExecutableAllocatorPosix.cpp: 13761 * jit/ExecutableAllocatorWin.cpp: 13762 * runtime/MarkStack.h: 13763 (JSC::MarkStack::pageSize): 13764 * runtime/MarkStackPosix.cpp: 13765 (JSC::MarkStack::initializePagesize): 13766 * runtime/MarkStackWin.cpp: 13767 (JSC::MarkStack::initializePagesize): 13768 137692009-08-07 Oliver Hunt <oliver@apple.com> 13770 13771 Reviewed by Gavin Barraclough. 13772 13773 Stack overflow crash in JavaScript garbage collector mark pass 13774 https://bugs.webkit.org/show_bug.cgi?id=12216 13775 13776 Make the GC mark phase iterative by using an explicit mark stack. 13777 To do this marking any single object is performed in multiple stages 13778 * The object is appended to the MarkStack, this sets the marked 13779 bit for the object using the new markDirect() function, and then 13780 returns 13781 * When the MarkStack is drain()ed the object is popped off the stack 13782 and markChildren(MarkStack&) is called on the object to collect 13783 all of its children. drain() then repeats until the stack is empty. 13784 13785 Additionally I renamed a number of methods from 'mark' to 'markAggregate' 13786 in order to make it more clear that marking of those object was not 13787 going to result in an actual recursive mark. 13788 13789 * GNUmakefile.am 13790 * JavaScriptCore.exp: 13791 * JavaScriptCore.gypi: 13792 * JavaScriptCore.pri: 13793 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: 13794 * JavaScriptCore.xcodeproj/project.pbxproj: 13795 * bytecode/CodeBlock.cpp: 13796 (JSC::CodeBlock::markAggregate): 13797 * bytecode/CodeBlock.h: 13798 * bytecode/EvalCodeCache.h: 13799 (JSC::EvalCodeCache::markAggregate): 13800 * debugger/DebuggerActivation.cpp: 13801 (JSC::DebuggerActivation::markChildren): 13802 * debugger/DebuggerActivation.h: 13803 * interpreter/Register.h: 13804 * interpreter/RegisterFile.h: 13805 (JSC::RegisterFile::markGlobals): 13806 (JSC::RegisterFile::markCallFrames): 13807 * parser/Nodes.cpp: 13808 (JSC::ScopeNodeData::markAggregate): 13809 (JSC::EvalNode::markAggregate): 13810 (JSC::FunctionBodyNode::markAggregate): 13811 * parser/Nodes.h: 13812 (JSC::ScopeNode::markAggregate): 13813 * runtime/ArgList.cpp: 13814 (JSC::MarkedArgumentBuffer::markLists): 13815 * runtime/ArgList.h: 13816 * runtime/Arguments.cpp: 13817 (JSC::Arguments::markChildren): 13818 * runtime/Arguments.h: 13819 * runtime/Collector.cpp: 13820 (JSC::Heap::markConservatively): 13821 (JSC::Heap::markCurrentThreadConservativelyInternal): 13822 (JSC::Heap::markCurrentThreadConservatively): 13823 (JSC::Heap::markOtherThreadConservatively): 13824 (JSC::Heap::markStackObjectsConservatively): 13825 (JSC::Heap::markProtectedObjects): 13826 (JSC::Heap::collect): 13827 * runtime/Collector.h: 13828 * runtime/GetterSetter.cpp: 13829 (JSC::GetterSetter::markChildren): 13830 * runtime/GetterSetter.h: 13831 (JSC::GetterSetter::GetterSetter): 13832 (JSC::GetterSetter::createStructure): 13833 * runtime/GlobalEvalFunction.cpp: 13834 (JSC::GlobalEvalFunction::markChildren): 13835 * runtime/GlobalEvalFunction.h: 13836 * runtime/JSActivation.cpp: 13837 (JSC::JSActivation::markChildren): 13838 * runtime/JSActivation.h: 13839 * runtime/JSArray.cpp: 13840 (JSC::JSArray::markChildren): 13841 * runtime/JSArray.h: 13842 * runtime/JSCell.h: 13843 (JSC::JSCell::markCellDirect): 13844 (JSC::JSCell::markChildren): 13845 (JSC::JSValue::markDirect): 13846 (JSC::JSValue::markChildren): 13847 (JSC::JSValue::hasChildren): 13848 (JSC::MarkStack::append): 13849 (JSC::MarkStack::drain): 13850 * runtime/JSFunction.cpp: 13851 (JSC::JSFunction::markChildren): 13852 * runtime/JSFunction.h: 13853 * runtime/JSGlobalData.cpp: 13854 (JSC::JSGlobalData::JSGlobalData): 13855 * runtime/JSGlobalData.h: 13856 * runtime/JSGlobalObject.cpp: 13857 (JSC::markIfNeeded): 13858 (JSC::JSGlobalObject::markChildren): 13859 * runtime/JSGlobalObject.h: 13860 * runtime/JSNotAnObject.cpp: 13861 (JSC::JSNotAnObject::markChildren): 13862 * runtime/JSNotAnObject.h: 13863 * runtime/JSONObject.cpp: 13864 (JSC::Stringifier::markAggregate): 13865 (JSC::JSONObject::markStringifiers): 13866 * runtime/JSONObject.h: 13867 * runtime/JSObject.cpp: 13868 (JSC::JSObject::markChildren): 13869 (JSC::JSObject::defineGetter): 13870 (JSC::JSObject::defineSetter): 13871 * runtime/JSObject.h: 13872 * runtime/JSPropertyNameIterator.cpp: 13873 (JSC::JSPropertyNameIterator::markChildren): 13874 * runtime/JSPropertyNameIterator.h: 13875 (JSC::JSPropertyNameIterator::createStructure): 13876 (JSC::JSPropertyNameIterator::JSPropertyNameIterator): 13877 (JSC::JSPropertyNameIterator::create): 13878 * runtime/JSStaticScopeObject.cpp: 13879 (JSC::JSStaticScopeObject::markChildren): 13880 * runtime/JSStaticScopeObject.h: 13881 * runtime/JSType.h: 13882 (JSC::): 13883 * runtime/JSValue.h: 13884 * runtime/JSWrapperObject.cpp: 13885 (JSC::JSWrapperObject::markChildren): 13886 * runtime/JSWrapperObject.h: 13887 * runtime/MarkStack.cpp: Added. 13888 (JSC::MarkStack::compact): 13889 * runtime/MarkStack.h: Added. 13890 (JSC::): 13891 (JSC::MarkStack::MarkStack): 13892 (JSC::MarkStack::append): 13893 (JSC::MarkStack::appendValues): 13894 (JSC::MarkStack::~MarkStack): 13895 (JSC::MarkStack::MarkSet::MarkSet): 13896 (JSC::MarkStack::pageSize): 13897 13898 MarkStackArray is a non-shrinking, mmap-based vector type 13899 used for storing objects to be marked. 13900 (JSC::MarkStack::MarkStackArray::MarkStackArray): 13901 (JSC::MarkStack::MarkStackArray::~MarkStackArray): 13902 (JSC::MarkStack::MarkStackArray::expand): 13903 (JSC::MarkStack::MarkStackArray::append): 13904 (JSC::MarkStack::MarkStackArray::removeLast): 13905 (JSC::MarkStack::MarkStackArray::isEmpty): 13906 (JSC::MarkStack::MarkStackArray::size): 13907 (JSC::MarkStack::MarkStackArray::shrinkAllocation): 13908 * runtime/MarkStackPosix.cpp: Added. 13909 (JSC::MarkStack::allocateStack): 13910 (JSC::MarkStack::releaseStack): 13911 * runtime/MarkStackWin.cpp: Added. 13912 (JSC::MarkStack::allocateStack): 13913 (JSC::MarkStack::releaseStack): 13914 13915 * runtime/ScopeChain.h: 13916 * runtime/ScopeChainMark.h: 13917 (JSC::ScopeChain::markAggregate): 13918 * runtime/SmallStrings.cpp: 13919 (JSC::SmallStrings::mark): 13920 * runtime/Structure.h: 13921 (JSC::Structure::markAggregate): 13922 139232009-08-10 Mark Rowe <mrowe@apple.com> 13924 13925 Reviewed by Darin Adler. 13926 13927 Fix hundreds of "pointer being freed was not allocated" errors seen on the build bot. 13928 13929 * wtf/FastMalloc.h: Implement nothrow variants of the delete and delete[] operators since 13930 we implement the nothrow variants of new and new[]. The nothrow variant of delete is called 13931 explicitly in the implementation of std::sort which was resulting in FastMalloc-allocated 13932 memory being passed to the system allocator to free. 13933 139342009-08-10 Jan Michael Alonzo <jmalonzo@webkit.org> 13935 13936 [Gtk] Unreviewed build fix. Move JSAPIValueWrapper.cpp/.h in the debug 13937 section. This file is already part of AllInOneFile in Release builds. 13938 13939 * GNUmakefile.am: 13940 139412009-08-10 Darin Adler <darin@apple.com> 13942 13943 * wtf/FastMalloc.h: Fix build. 13944 139452009-08-10 Darin Adler <darin@apple.com> 13946 13947 Reviewed by Mark Rowe. 13948 13949 FastMalloc.h has cross-platform code but marked as WinCE-only 13950 https://bugs.webkit.org/show_bug.cgi?id=28160 13951 13952 1) The support for nothrow was inside #if PLATFORM(WINCE) even though it is 13953 not platform-specific. 13954 2) The code tried to override operator delete nothrow, which does not exist. 13955 3) The code in the header checks the value of USE_SYSTEM_MALLOC, but the code 13956 in FastMalloc.cpp checks only if the macro is defined. 13957 13958 * wtf/FastMalloc.h: See above. 13959 * wtf/FastMalloc.cpp: Ditto. 13960 139612009-08-10 Sam Weinig <sam@webkit.org> 13962 13963 Reviewed by Anders Carlsson. 13964 13965 Fix an annoying indentation issue. 13966 13967 * runtime/DateConstructor.cpp: 13968 (JSC::constructDate): 13969 139702009-08-10 Xan Lopez <xlopez@igalia.com> 13971 13972 Unreviewed build fix. 13973 13974 Add new files to makefile. 13975 13976 * GNUmakefile.am: 13977 139782009-08-10 Simon Hausmann <simon.hausmann@nokia.com> 13979 13980 Fix compilation with the interpreter instead of the JIT by including 13981 PrototypeFunction.h as forward-declared through NativeFunctionWrapper.h. 13982 13983 * runtime/ObjectConstructor.cpp: 13984 139852009-08-09 Oliver Hunt <oliver@apple.com> 13986 13987 Reviewed by George Staikos. 13988 13989 JSON.stringify replacer returning undefined does not omit object properties 13990 https://bugs.webkit.org/show_bug.cgi?id=28118 13991 13992 Correct behaviour of stringify when using a replacer function that returns 13993 undefined. This is a simple change to move the undefined value check to 13994 after the replacer function is called. This means that the replacer function 13995 is now called for properties with the value undefined, however i've confirmed 13996 that this behaviour is correct. 13997 13998 In addition I've made the cyclic object exception have a more useful error 13999 message. 14000 14001 * runtime/JSONObject.cpp: 14002 (JSC::Stringifier::appendStringifiedValue): 14003 140042009-08-08 Oliver Hunt <oliver@apple.com> 14005 14006 Reviewed by Eric Seidel and Sam Weinig. 14007 14008 [ES5] Implement Object.getPrototypeOf 14009 https://bugs.webkit.org/show_bug.cgi?id=28114 14010 14011 Implement getPrototypeOf 14012 14013 * runtime/CommonIdentifiers.h: 14014 * runtime/JSGlobalObject.cpp: 14015 (JSC::JSGlobalObject::reset): 14016 * runtime/ObjectConstructor.cpp: 14017 (JSC::ObjectConstructor::ObjectConstructor): 14018 (JSC::objectConsGetPrototypeOf): 14019 * runtime/ObjectConstructor.h: 14020 140212009-08-07 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 14022 14023 Reviewed by Eric Seidel. 14024 14025 Allow custom memory allocation control for Noncopyable class 14026 https://bugs.webkit.org/show_bug.cgi?id=27879 14027 14028 Several classes which are inherited from Noncopyable are instantiated by 14029 operator new, so Noncopyable class has been inherited from FastAllocBase. 14030 14031 * wtf/Noncopyable.h: 14032 140332009-08-07 George Staikos <george.staikos@torchmobile.com> 14034 14035 Reviewed by Eric Seidel. 14036 14037 https://bugs.webkit.org/show_bug.cgi?id=27305 14038 Implement WinCE-specific unicode layer. 14039 Written by George Staikos <george.staikos@torchmobile.com> 14040 with bug fixes by Yong Li <yong.li@torchmobile.com> 14041 refactored by Joe Mason <joe.mason@torchmobile.com> 14042 14043 * wtf/Platform.h: 14044 * wtf/unicode/Unicode.h: 14045 * wtf/unicode/wince/UnicodeWince.cpp: Added. 14046 (WTF::Unicode::toLower): 14047 (WTF::Unicode::toUpper): 14048 (WTF::Unicode::foldCase): 14049 (WTF::Unicode::isPrintableChar): 14050 (WTF::Unicode::isSpace): 14051 (WTF::Unicode::isLetter): 14052 (WTF::Unicode::isUpper): 14053 (WTF::Unicode::isLower): 14054 (WTF::Unicode::isDigit): 14055 (WTF::Unicode::isPunct): 14056 (WTF::Unicode::toTitleCase): 14057 (WTF::Unicode::direction): 14058 (WTF::Unicode::category): 14059 (WTF::Unicode::decompositionType): 14060 (WTF::Unicode::combiningClass): 14061 (WTF::Unicode::mirroredChar): 14062 (WTF::Unicode::digitValue): 14063 * wtf/unicode/wince/UnicodeWince.h: Added. 14064 (WTF::Unicode::): 14065 (WTF::Unicode::isSeparatorSpace): 14066 (WTF::Unicode::isHighSurrogate): 14067 (WTF::Unicode::isLowSurrogate): 14068 (WTF::Unicode::isArabicChar): 14069 (WTF::Unicode::hasLineBreakingPropertyComplexContext): 14070 (WTF::Unicode::umemcasecmp): 14071 (WTF::Unicode::surrogateToUcs4): 14072 140732009-08-07 Yongjun Zhang <yongjun.zhang@nokia.com> 14074 14075 Reviewed by Eric Seidel. 14076 14077 https://bugs.webkit.org/show_bug.cgi?id=28069 14078 14079 Add inline to help winscw compiler resolve specialized argument in 14080 templated functions. 14081 14082 * runtime/LiteralParser.cpp: 14083 (JSC::LiteralParser::Lexer::lexString): 14084 140852009-08-07 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 14086 14087 Reviewed by Eric Seidel. 14088 14089 Allow custom memory allocation control for RegExpObjectData struct 14090 http://bugs.webkit.org/show_bug.cgi?id=26750 14091 14092 Inherits RegExpObjectData struct from FastAllocBase because 14093 it has been instantiated by 'new' in JavaScriptCore/runtime/RegExpObject.cpp:62 14094 14095 * runtime/RegExpObject.h: 14096 140972009-08-06 Norbert Leser <norbert.leser@nokia.com> 14098 14099 Reviewed by Darin Adler. 14100 14101 Updated patch for bug #27059: 14102 Symbian platform always uses little endian encoding, 14103 regardless of compiler. 14104 We need to make sure that we correctly detect EABI architecture 14105 for armv5 targets on Symbian, 14106 where __EABI__ is set but not __ARM_EABI__ 14107 14108 * wtf/Platform.h: 14109 141102009-08-06 Adam Barth <abarth@webkit.org> 14111 14112 Unreviewed revert. 14113 14114 http://bugs.webkit.org/show_bug.cgi?id=27879 14115 14116 Revert 46877 because it broke GTK. 14117 14118 * wtf/Noncopyable.h: 14119 141202009-08-06 Gavin Barraclough <barraclough@apple.com> 14121 14122 Reviewed by Oliver Hunt. 14123 14124 Make get_by_id/put_by_id/method_check/call defer optimization using a data flag rather than a code modification. 14125 ( https://bugs.webkit.org/show_bug.cgi?id=27635 ) 14126 14127 This improves performance of ENABLE(ASSEMBLER_WX_EXCLUSIVE) builds by 2-2.5%, reducing the overhead to about 2.5%. 14128 (No performance impact with ASSEMBLER_WX_EXCLUSIVE disabled). 14129 14130 * bytecode/CodeBlock.cpp: 14131 (JSC::printStructureStubInfo): 14132 - Make StructureStubInfo store the type as an integer, rather than an OpcodeID. 14133 14134 * bytecode/CodeBlock.h: 14135 (JSC::): 14136 (JSC::CallLinkInfo::seenOnce): 14137 (JSC::CallLinkInfo::setSeen): 14138 (JSC::MethodCallLinkInfo::seenOnce): 14139 (JSC::MethodCallLinkInfo::setSeen): 14140 - Change a pointer in CallLinkInfo/MethodCallLinkInfo to use a PtrAndFlags, use a flag to track when an op has been executed once. 14141 14142 * bytecode/StructureStubInfo.cpp: 14143 (JSC::StructureStubInfo::deref): 14144 - Make StructureStubInfo store the type as an integer, rather than an OpcodeID. 14145 14146 * bytecode/StructureStubInfo.h: 14147 (JSC::StructureStubInfo::StructureStubInfo): 14148 (JSC::StructureStubInfo::initGetByIdSelf): 14149 (JSC::StructureStubInfo::initGetByIdProto): 14150 (JSC::StructureStubInfo::initGetByIdChain): 14151 (JSC::StructureStubInfo::initGetByIdSelfList): 14152 (JSC::StructureStubInfo::initGetByIdProtoList): 14153 (JSC::StructureStubInfo::initPutByIdTransition): 14154 (JSC::StructureStubInfo::initPutByIdReplace): 14155 (JSC::StructureStubInfo::seenOnce): 14156 (JSC::StructureStubInfo::setSeen): 14157 - Make StructureStubInfo store the type as an integer, rather than an OpcodeID, add a flag to track when an op has been executed once. 14158 14159 * bytecompiler/BytecodeGenerator.cpp: 14160 (JSC::BytecodeGenerator::emitGetById): 14161 (JSC::BytecodeGenerator::emitPutById): 14162 - Make StructureStubInfo store the type as an integer, rather than an OpcodeID. 14163 14164 * jit/JIT.cpp: 14165 (JSC::JIT::privateCompileCTIMachineTrampolines): 14166 (JSC::JIT::unlinkCall): 14167 - Remove the "don't lazy link" stage of calls. 14168 14169 * jit/JIT.h: 14170 (JSC::JIT::compileCTIMachineTrampolines): 14171 - Remove the "don't lazy link" stage of calls. 14172 14173 * jit/JITCall.cpp: 14174 (JSC::JIT::compileOpCallSlowCase): 14175 - Remove the "don't lazy link" stage of calls. 14176 14177 * jit/JITStubs.cpp: 14178 (JSC::JITThunks::JITThunks): 14179 (JSC::JITThunks::tryCachePutByID): 14180 (JSC::JITThunks::tryCacheGetByID): 14181 (JSC::JITStubs::DEFINE_STUB_FUNCTION): 14182 (JSC::JITStubs::getPolymorphicAccessStructureListSlot): 14183 - Remove the "don't lazy link" stage of calls, and the "_second" stage of get_by_id/put_by_id/method_check. 14184 14185 * jit/JITStubs.h: 14186 (JSC::JITThunks::ctiStringLengthTrampoline): 14187 (JSC::JITStubs::): 14188 - Remove the "don't lazy link" stage of calls, and the "_second" stage of get_by_id/put_by_id/method_check. 14189 14190 * wtf/PtrAndFlags.h: 14191 (WTF::PtrAndFlags::PtrAndFlags): 14192 (WTF::PtrAndFlags::operator!): 14193 (WTF::PtrAndFlags::operator->): 14194 - Add ! and -> operators, add constuctor with pointer argument. 14195 141962009-08-06 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 14197 14198 Reviewed by Adam Barth. 14199 14200 Allow custom memory allocation control for Noncopyable class 14201 https://bugs.webkit.org/show_bug.cgi?id=27879 14202 14203 Several classes which inherited from Noncopyable are instantiated by 14204 operator new, so Noncopyable class has been inherited from FastAllocBase. 14205 14206 * wtf/Noncopyable.h: 14207 142082009-08-06 Mark Rowe <mrowe@apple.com> 14209 14210 Rubber-stamped by Sam Weinig. 14211 14212 Add explicit dependencies for our build verification scripts to ensure that they always run after linking has completed. 14213 14214 * JavaScriptCore.xcodeproj/project.pbxproj: 14215 142162009-08-06 Mark Rowe <mrowe@apple.com> 14217 14218 Bring a little order to our otherwise out of control lives. 14219 14220 * JavaScriptCore.xcodeproj/project.pbxproj: 14221 142222009-08-06 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 14223 14224 Reviewed by Darin Adler. 14225 14226 Allow custom memory allocation control for JavaScriptCore's PolymorphicAccessStructureList struct 14227 https://bugs.webkit.org/show_bug.cgi?id=27877 14228 14229 Inherits PolymorphicAccessStructureList struct from FastAllocBase because it has been instantiated by 14230 'new' in JavaScriptCore/jit/JITStubs.cpp:1229. 14231 14232 * bytecode/Instruction.h: 14233 142342009-08-05 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 14235 14236 Reviewed by Darin Adler. 14237 14238 Allow custom memory allocation control for JavaScriptCore's ScopeNodeData struct 14239 https://bugs.webkit.org/show_bug.cgi?id=27875 14240 14241 Inherits ScopeNodeData struct from FastAllocBase because it has been instantiated by 14242 'new' in JavaScriptCore/parser/Nodes.cpp:1848. 14243 14244 * parser/Nodes.h: 14245 142462009-08-05 Zoltan Herczeg <zherczeg@inf.u-szeged.hu> 14247 14248 Reviewed by Gavin Barraclough. 14249 14250 Add floating point support for generic ARM port. 14251 https://bugs.webkit.org/show_bug.cgi?id=24986 14252 14253 * assembler/ARMAssembler.cpp: 14254 (JSC::ARMAssembler::doubleTransfer): 14255 * assembler/ARMAssembler.h: 14256 (JSC::ARM::): 14257 (JSC::ARMAssembler::): 14258 (JSC::ARMAssembler::faddd_r): 14259 (JSC::ARMAssembler::fsubd_r): 14260 (JSC::ARMAssembler::fmuld_r): 14261 (JSC::ARMAssembler::fcmpd_r): 14262 (JSC::ARMAssembler::fdtr_u): 14263 (JSC::ARMAssembler::fdtr_d): 14264 (JSC::ARMAssembler::fmsr_r): 14265 (JSC::ARMAssembler::fsitod_r): 14266 (JSC::ARMAssembler::fmstat): 14267 * assembler/MacroAssemblerARM.h: 14268 (JSC::MacroAssemblerARM::): 14269 (JSC::MacroAssemblerARM::supportsFloatingPoint): 14270 (JSC::MacroAssemblerARM::loadDouble): 14271 (JSC::MacroAssemblerARM::storeDouble): 14272 (JSC::MacroAssemblerARM::addDouble): 14273 (JSC::MacroAssemblerARM::subDouble): 14274 (JSC::MacroAssemblerARM::mulDouble): 14275 (JSC::MacroAssemblerARM::convertInt32ToDouble): 14276 (JSC::MacroAssemblerARM::branchDouble): 14277 * jit/JIT.h: 14278 142792009-08-05 Zoltan Herczeg <zherczeg@inf.u-szeged.hu> 14280 14281 Reviewed by Gavin Barraclough. 14282 14283 Add JIT support for generic ARM port without optimizations. 14284 https://bugs.webkit.org/show_bug.cgi?id=24986 14285 14286 All JIT optimizations are disabled. 14287 14288 Signed off by Zoltan Herczeg <zherczeg@inf.u-szeged.hu> 14289 Signed off by Gabor Loki <loki@inf.u-szeged.hu> 14290 14291 * assembler/ARMAssembler.cpp: 14292 (JSC::ARMAssembler::baseIndexTransfer32): 14293 * assembler/AbstractMacroAssembler.h: 14294 (JSC::AbstractMacroAssembler::Imm32::Imm32): 14295 * assembler/MacroAssemblerARM.h: 14296 (JSC::MacroAssemblerARM::store32): 14297 (JSC::MacroAssemblerARM::move): 14298 (JSC::MacroAssemblerARM::branch32): 14299 (JSC::MacroAssemblerARM::add32): 14300 (JSC::MacroAssemblerARM::sub32): 14301 (JSC::MacroAssemblerARM::load32): 14302 * bytecode/CodeBlock.h: 14303 (JSC::CodeBlock::getBytecodeIndex): 14304 * jit/JIT.h: 14305 * jit/JITInlineMethods.h: 14306 (JSC::JIT::restoreArgumentReference): 14307 * jit/JITOpcodes.cpp: 14308 * jit/JITStubs.cpp: 14309 * jit/JITStubs.h: 14310 (JSC::JITStackFrame::returnAddressSlot): 14311 * wtf/Platform.h: 14312 143132009-08-04 Gavin Barraclough <barraclough@apple.com> 14314 14315 Rubber Stamped by Oiver Hunt. 14316 14317 Revert r46643 since this breaks the Yarr::Interpreter running the v8 tests. 14318 https://bugs.webkit.org/show_bug.cgi?id=27874 14319 14320 * yarr/RegexInterpreter.cpp: 14321 (JSC::Yarr::Interpreter::allocDisjunctionContext): 14322 (JSC::Yarr::Interpreter::freeDisjunctionContext): 14323 (JSC::Yarr::Interpreter::allocParenthesesDisjunctionContext): 14324 (JSC::Yarr::Interpreter::freeParenthesesDisjunctionContext): 14325 143262009-08-04 Oliver Hunt <oliver@apple.com> 14327 14328 PPC64 Build fix 14329 14330 * wtf/Platform.h: 14331 143322009-08-04 Benjamin C Meyer <benjamin.meyer@torchmobile.com> 14333 14334 Reviewed by Adam Treat 14335 14336 Explicitly include limits.h header when using INT_MAX and INT_MIN 14337 14338 * interpreter/Interpreter.cpp 14339 143402009-08-03 Harald Fernengel <harald.fernengel@nokia.com> 14341 14342 Reviewed by Darin Adler. 14343 14344 Fix compile error for ambigous call to abs() 14345 https://bugs.webkit.org/show_bug.cgi?id=27873 14346 14347 Fix ambiguity in abs(long int) call by calling labs() instead 14348 14349 * wtf/DateMath.cpp: replace call to abs() with labs() 14350 143512009-08-03 Laszlo Gombos <laszlo.1.gombos@nokia.com> 14352 14353 Reviewed by Eric Seidel. 14354 14355 [Qt] Consolidate common gcc flags to WebKit.pri 14356 https://bugs.webkit.org/show_bug.cgi?id=27934 14357 14358 * JavaScriptCore.pro: 14359 143602009-08-03 Ada Chan <adachan@apple.com> 14361 14362 Fixed the Tiger build. 14363 14364 * wtf/FastMalloc.cpp: 14365 143662009-08-03 Ada Chan <adachan@apple.com> 14367 14368 Reviewed by Darin Adler. 14369 14370 Don't use background thread to scavenge memory on Tiger until we figure out why it causes a crash. 14371 https://bugs.webkit.org/show_bug.cgi?id=27900 14372 14373 * wtf/FastMalloc.cpp: 14374 143752009-08-03 Fumitoshi Ukai <ukai@chromium.org> 14376 14377 Reviewed by Jan Alonzo. 14378 14379 Fix build break on Gtk/x86_64. 14380 https://bugs.webkit.org/show_bug.cgi?id=27936 14381 14382 Use JSVALUE64 for X86_64 LINUX, except Qt. 14383 14384 * wtf/Platform.h: 14385 143862009-08-02 Xan Lopez <xlopez@igalia.com> 14387 14388 Fix the GTK+ build. 14389 14390 * wtf/Platform.h: 14391 143922009-08-02 Geoffrey Garen <ggaren@apple.com> 14393 14394 Reviewed by Sam Weinig. 14395 14396 Disabled JSVALUE32_64 on Qt builds, since all layout tests mysteriously 14397 crash with it enabled. 14398 14399 * wtf/Platform.h: 14400 144012009-08-02 Geoffrey Garen <ggaren@apple.com> 14402 14403 Qt build fix. 14404 14405 Added JSAPIValueWrapper.cpp to the build. 14406 14407 * JavaScriptCore.pri: 14408 144092009-08-02 Geoffrey Garen <ggaren@apple.com> 14410 14411 Windows build fix. 14412 14413 Exported symbols for JSAPIValueWrapper. 14414 14415 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 14416 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 14417 144182009-08-02 Geoffrey Garen <ggaren@apple.com> 14419 14420 GTK build fix. 14421 14422 * jit/JITStubs.cpp: #include <stdarg.h>, for a definition of va_start. 14423 144242009-08-02 Geoffrey Garen <ggaren@apple.com> 14425 14426 Qt build fix. 14427 14428 * runtime/Collector.cpp: #include <limits.h>, for a definition of ULONG_MAX. 14429 144302009-08-02 Geoffrey Garen <ggaren@apple.com> 14431 14432 Windows build fix: Nixed JSImmediate::prototype, JSImmediate::toObject, 14433 and JSImmediate::toThisObject, and removed their exported symbols. 14434 14435 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 14436 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 14437 * runtime/JSImmediate.cpp: 14438 * runtime/JSImmediate.h: 14439 144402009-08-02 Geoffrey Garen <ggaren@apple.com> 14441 14442 Reviewed by Mark Rowe. 14443 14444 Enabled JSVALUE32_64 by default on all platforms other than x86_64 (which uses JSVALUE64). 14445 14446 * wtf/Platform.h: 14447 144482009-08-02 Kevin Ollivier <kevino@theolliviers.com> 14449 14450 Reviewed by Jan Alonzo. 14451 14452 Script for building the JavaScriptCore library for wx. 14453 https://bugs.webkit.org/show_bug.cgi?id=27619 14454 14455 * wscript: Added. 14456 144572009-08-02 Yong Li <yong.li@torchmobile.com> 14458 14459 Reviewed by George Staikos. 14460 14461 DateMath depends on strftime and localtime, which need to be imported manually on WinCE 14462 https://bugs.webkit.org/show_bug.cgi?id=26558 14463 14464 * wtf/DateMath.cpp: 14465 144662009-08-01 David Kilzer <ddkilzer@apple.com> 14467 14468 wtf/Threading.h: added include of Platform.h 14469 14470 Reviewed by Mark Rowe. 14471 14472 * wtf/Threading.h: Added #include "Platform.h" since this header 14473 uses PLATFORM() and other macros. 14474 144752009-08-01 Mark Rowe <mrowe@apple.com> 14476 14477 Rubber-stamped by Oliver Hunt. 14478 14479 Roll out r46668 as it was misinformed. ScopeChain is only used with placement new. 14480 14481 * runtime/ScopeChain.h: 14482 144832009-08-01 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 14484 14485 Allow custom memory allocation control for JavaScriptCore's HashMap class 14486 http://bugs.webkit.org/show_bug.cgi?id=27871 14487 14488 Inherits HashMap class from FastAllocBase because it has been 14489 instantiated by 'new' in JavaScriptCore/API/JSClassRef.cpp:148. 14490 14491 * wtf/RefPtrHashMap.h: 14492 (WTF::): 14493 144942009-08-01 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 14495 14496 Allow custom memory allocation control for JavaScriptCore's ScopeChain class 14497 https://bugs.webkit.org/show_bug.cgi?id=27834 14498 14499 Inherits ScopeChain class from FastAllocBase because it has been 14500 instantiated by 'new' in JavaScriptCore/runtime/JSFunction.h:109. 14501 14502 * runtime/ScopeChain.h: 14503 145042009-08-01 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 14505 14506 Reviewed by Darin Adler. 14507 14508 Allow custom memory allocation control for JavaScriptCore's RegExpConstructorPrivate struct 14509 https://bugs.webkit.org/show_bug.cgi?id=27833 14510 14511 Inherits RegExpConstructorPrivate class from FastAllocBase because it has been 14512 instantiated by 'new' in JavaScriptCore/runtime/RegExpConstructor.cpp:152. 14513 14514 * runtime/RegExpConstructor.cpp: 14515 145162009-07-31 Yong Li <yong.li@torchmobile.com> 14517 14518 Reviewed by George Staikos. 14519 14520 Resurrect the old GetTickCount implementation of currentTime, controlled by WTF_USE_QUERY_PERFORMANCE_COUNTER 14521 currentSystemTime taken from older WebKit; currentTime written by Yong Li <yong.li@torchmobile.com>; cleanup by Joe Mason <joe.mason@torchmobile.com> 14522 https://bugs.webkit.org/show_bug.cgi?id=27848 14523 14524 * wtf/CurrentTime.cpp: 14525 (WTF::currentSystemTime): get current time with GetCurrentFT 14526 (WTF::currentTime): track msec elapsed since first currentSystemTime call using GetTickCount 14527 * wtf/Platform.h: 14528 145292009-07-31 Ada Chan <adachan@apple.com> 14530 14531 Fixes the Windows release-PGO build. 14532 14533 Reviewed by Jon Honeycutt. 14534 14535 * JavaScriptCore.vcproj/WTF/WTF.vcproj: Suppresses the warning about unreachable code that we get by adding "return 0" to WTF::TCMalloc_PageHeap::runScavengerThread(). 14536 * wtf/FastMalloc.cpp: 14537 (WTF::TCMalloc_PageHeap::runScavengerThread): Fixes the error about the method not returning a value in the release-PGO build. 14538 145392009-07-31 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 14540 14541 Change malloc to fastMalloc and free to fastFree in Yarr's RegexInterpreter.cpp 14542 https://bugs.webkit.org/show_bug.cgi?id=27874 14543 14544 Use fastMalloc and fastFree instead of malloc and free in RegexInterpreter.cpp's methods. 14545 14546 * yarr/RegexInterpreter.cpp: 14547 (JSC::Yarr::Interpreter::allocDisjunctionContext): 14548 (JSC::Yarr::Interpreter::freeDisjunctionContext): 14549 (JSC::Yarr::Interpreter::allocParenthesesDisjunctionContext): 14550 (JSC::Yarr::Interpreter::freeParenthesesDisjunctionContext): 14551 145522009-07-30 Xan Lopez <xlopez@igalia.com> 14553 14554 Reviewed by Jan Alonzo. 14555 14556 Fix compiler warning. 14557 14558 GCC does not like C++-style comments in preprocessor directives. 14559 14560 * wtf/Platform.h: 14561 145622009-07-30 John McCall <rjmccall@apple.com> 14563 14564 Reviewed by Gavin Barraclough. 14565 14566 Optimize the X86_64 trampolines: avoid the need for filler arguments 14567 and move the stub-args area closer to the stack pointer. 14568 14569 * jit/JIT.h: adjust patch offsets because of slight code-size change 14570 * jit/JITCode.h: 14571 (JSC::JITCode::execute): don't pass filler args 14572 * jit/JITStubs.cpp: 14573 (ctiTrampoline): (X86_64): push args onto stack, use args directly 14574 (ctiVMThrowTrampoline): (X86_64): adjust %rsp by correct displacement 14575 (ctiOpThrowNotCaught): (X86_64): adjust %rsp by correct displacement 14576 * jit/JITStubs.h: 14577 (JITStackFrame): (X86_64): move args area earlier 14578 (ctiTrampoline): remove filler args from prototype 14579 145802009-07-30 Gavin Barraclough <barraclough@apple.com> 14581 14582 Temporarily revert r46618 since this is b0rking on Linux. 14583 145842009-07-23 Gavin Barraclough <barraclough@apple.com> 14585 14586 Reviewed by Oliver Hunt. 14587 14588 Make get_by_id/put_by_id/method_check/call defer optimization using a data flag rather than a code modification. 14589 ( https://bugs.webkit.org/show_bug.cgi?id=27635 ) 14590 14591 This improves performance of ENABLE(ASSEMBLER_WX_EXCLUSIVE) builds by 2-2.5%, reducing the overhead to about 2.5%. 14592 (No performance impact with ASSEMBLER_WX_EXCLUSIVE disabled). 14593 14594 * bytecode/CodeBlock.cpp: 14595 (JSC::printStructureStubInfo): 14596 - Make StructureStubInfo store the type as an integer, rather than an OpcodeID. 14597 14598 * bytecode/CodeBlock.h: 14599 (JSC::): 14600 (JSC::CallLinkInfo::seenOnce): 14601 (JSC::CallLinkInfo::setSeen): 14602 (JSC::MethodCallLinkInfo::seenOnce): 14603 (JSC::MethodCallLinkInfo::setSeen): 14604 - Change a pointer in CallLinkInfo/MethodCallLinkInfo to use a PtrAndFlags, use a flag to track when an op has been executed once. 14605 14606 * bytecode/StructureStubInfo.cpp: 14607 (JSC::StructureStubInfo::deref): 14608 - Make StructureStubInfo store the type as an integer, rather than an OpcodeID. 14609 14610 * bytecode/StructureStubInfo.h: 14611 (JSC::StructureStubInfo::StructureStubInfo): 14612 (JSC::StructureStubInfo::initGetByIdSelf): 14613 (JSC::StructureStubInfo::initGetByIdProto): 14614 (JSC::StructureStubInfo::initGetByIdChain): 14615 (JSC::StructureStubInfo::initGetByIdSelfList): 14616 (JSC::StructureStubInfo::initGetByIdProtoList): 14617 (JSC::StructureStubInfo::initPutByIdTransition): 14618 (JSC::StructureStubInfo::initPutByIdReplace): 14619 (JSC::StructureStubInfo::seenOnce): 14620 (JSC::StructureStubInfo::setSeen): 14621 - Make StructureStubInfo store the type as an integer, rather than an OpcodeID, add a flag to track when an op has been executed once. 14622 14623 * bytecompiler/BytecodeGenerator.cpp: 14624 (JSC::BytecodeGenerator::emitGetById): 14625 (JSC::BytecodeGenerator::emitPutById): 14626 - Make StructureStubInfo store the type as an integer, rather than an OpcodeID. 14627 14628 * jit/JIT.cpp: 14629 (JSC::JIT::privateCompileCTIMachineTrampolines): 14630 (JSC::JIT::unlinkCall): 14631 - Remove the "don't lazy link" stage of calls. 14632 14633 * jit/JIT.h: 14634 (JSC::JIT::compileCTIMachineTrampolines): 14635 - Remove the "don't lazy link" stage of calls. 14636 14637 * jit/JITCall.cpp: 14638 (JSC::JIT::compileOpCallSlowCase): 14639 - Remove the "don't lazy link" stage of calls. 14640 14641 * jit/JITStubs.cpp: 14642 (JSC::JITThunks::JITThunks): 14643 (JSC::JITThunks::tryCachePutByID): 14644 (JSC::JITThunks::tryCacheGetByID): 14645 (JSC::JITStubs::DEFINE_STUB_FUNCTION): 14646 (JSC::JITStubs::getPolymorphicAccessStructureListSlot): 14647 - Remove the "don't lazy link" stage of calls, and the "_second" stage of get_by_id/put_by_id/method_check. 14648 14649 * jit/JITStubs.h: 14650 (JSC::JITThunks::ctiStringLengthTrampoline): 14651 (JSC::JITStubs::): 14652 - Remove the "don't lazy link" stage of calls, and the "_second" stage of get_by_id/put_by_id/method_check. 14653 14654 * wtf/PtrAndFlags.h: 14655 (WTF::PtrAndFlags::PtrAndFlags): 14656 (WTF::PtrAndFlags::operator!): 14657 (WTF::PtrAndFlags::operator->): 14658 - Add ! and -> operators, add constuctor with pointer argument. 14659 146602009-07-30 Geoffrey Garen <ggaren@apple.com> 14661 14662 Reviewed by Gavin Barraclough. 14663 14664 Fixed failing tests seen on Windows buildbot. 14665 14666 * jit/JITStubs.cpp: 14667 (JSC::DEFINE_STUB_FUNCTION): 14668 * jit/JITStubs.h: 14669 (JSC::): Use "int" instead of "bool" to guarantee a 32-bit result, 14670 regardless of compiler. gcc on mac uses 32-bit values for bool, 14671 but gcc on linux and MSVC on Windows use 8-bit values. 14672 146732009-07-30 Geoffrey Garen <ggaren@apple.com> 14674 14675 Windows build fix: added missing symbols on Windows. 14676 14677 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 14678 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 14679 146802009-07-30 Geoffrey Garen <ggaren@apple.com> 14681 14682 Windows build fix: removed stale symbols on Windows. 14683 14684 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 14685 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 14686 14687=== End merge of nitro-extreme branch 2009-07-30 === 14688 146892009-07-20 Geoffrey Garen <ggaren@apple.com> 14690 14691 Fixed a post-review typo in r46066 that caused tons of test failures. 14692 14693 SunSpider reports no change. 14694 14695 * runtime/JSArray.cpp: 14696 (JSC::JSArray::JSArray): Initialize the full vector capacity, to avoid 14697 uninitialized members at the end. 14698 146992009-07-20 Geoffrey Garen <ggaren@apple.com> 14700 14701 Windows WebKit build fix: Added some missing exports. 14702 14703 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 14704 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 14705 147062009-07-17 Geoffrey Garen <ggaren@apple.com> 14707 14708 Reviewed by Sam Weinig. 14709 14710 Get the branch working on windows. 14711 https://bugs.webkit.org/show_bug.cgi?id=27391 14712 14713 SunSpider says 0.3% faster. 14714 14715 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 14716 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: Updated 14717 MSVC export lists to fix linker errors. 14718 14719 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: Added / removed 14720 new / old project files. 14721 14722 * jit/JIT.cpp: 14723 (JSC::JIT::privateCompileCTIMachineTrampolines): Used #pragma pack to tell 14724 MSVC that these structures represent actual memory layout, and should not be 14725 automatically aligned. Changed the return value load to load a 64bit quantity 14726 into the canonical registers. 14727 14728 * jit/JIT.h: Moved OBJECT_OFFSETOF definition to StdLibExtras.h because 14729 it's needed by more than just the JIT, and it supplements a standard library 14730 macro (offsetof). 14731 14732 * jit/JITCall.cpp: 14733 (JSC::JIT::compileOpCallInitializeCallFrame): Fixed an incorrectly signed 14734 cast to resolve an MSVC warning. 14735 14736 * jit/JITStubs.h: Used #pragma pack to tell MSVC that these structures 14737 represent actual memory layout, and should not be automatically aligned. 14738 14739 * runtime/JSArray.cpp: 14740 (JSC::JSArray::JSArray): Replaced memset_pattern8 with a for loop, since 14741 memset_pattern8 is not portable. (I verified that this version of the loop 14742 gives the best performance / generated code in GCC.) 14743 14744 * runtime/JSObject.h: 14745 (JSC::JSObject::JSObject): Removed accidental usage of FIELD_OFFSET -- 14746 OBJECT_OFFSETOF is our new macro name. (FIELD_OFFSET conflicts with a 14747 definition in winnt.h.) 14748 14749 * runtime/JSValue.cpp: Added some headers needed by non-all-in-one builds. 14750 14751 * runtime/JSValue.h: 14752 (JSC::JSValue::): Made the tag signed, to match MSVC's signed enum values. 14753 (GCC doesn't seem to care one way or the other.) 14754 14755 * wtf/MainThread.cpp: Moved the StdLibExtras.h #include -- I did this a 14756 while ago to resolve a conflict with winnt.h. I can't remember if it's truly 14757 still needed, but what the heck. 14758 14759 * wtf/StdLibExtras.h: Moved OBJECT_OFFSETOF definition here. 14760 147612009-07-06 Geoffrey Garen <ggaren@apple.com> 14762 14763 Reviewed by Sam Weinig (?). 14764 14765 Fixed an assertion seen during the stress test. 14766 14767 Don't assume that, if op1 is constant, op2 is not, and vice versa. Sadly, 14768 not all constants get folded. 14769 14770 * jit/JITArithmetic.cpp: 14771 (JSC::JIT::emit_op_jnless): 14772 (JSC::JIT::emitSlow_op_jnless): 14773 (JSC::JIT::emit_op_jnlesseq): 14774 (JSC::JIT::emitSlow_op_jnlesseq): 14775 147762009-07-06 Geoffrey Garen <ggaren@apple.com> 14777 14778 Reviewed by Sam Weinig. 14779 14780 Include op_convert_this in result caching. 14781 14782 No change on SunSpider or v8. 14783 14784 * jit/JITOpcodes.cpp: 14785 (JSC::JIT::emit_op_convert_this): 14786 14787 * jit/JITStubs.cpp: 14788 (JSC::DEFINE_STUB_FUNCTION): 14789 * jit/JITStubs.h: 14790 (JSC::): Made the op_convert_this JIT stub return an EncodedJSValue, so 14791 to maintain the result caching contract that { tag, payload } can be 14792 found in { regT1, regT0 }. 14793 147942009-07-06 Geoffrey Garen <ggaren@apple.com> 14795 14796 Reviewed by Sam Weinig. 14797 14798 Implemented result chaining. 14799 14800 1% faster on SunSpider. 4%-5% faster on v8. 14801 14802 * assembler/MacroAssemblerX86Common.h: 14803 (JSC::MacroAssemblerX86Common::move): 14804 * assembler/X86Assembler.h: 14805 (JSC::X86Assembler::movl_rr): Added an optimization to eliminate 14806 no-op mov instructions, to simplify chaining. 14807 14808 * jit/JIT.cpp: 14809 (JSC::JIT::JIT): 14810 * jit/JIT.h: Added data members and helper functions for recording 14811 chained results. We record both a mapping from virtual to machine register 14812 and the opcode for which the mapping is valid, to help ensure that the 14813 mapping isn't used after the mapped register has been stomped by other 14814 instructions. 14815 14816 * jit/JITCall.cpp: 14817 (JSC::JIT::compileOpCallVarargs): 14818 (JSC::JIT::compileOpCallVarargsSlowCase): 14819 (JSC::JIT::emit_op_ret): 14820 (JSC::JIT::emit_op_construct_verify): 14821 (JSC::JIT::compileOpCall): 14822 (JSC::JIT::compileOpCallSlowCase): Chain function call results. 14823 14824 * jit/JITInlineMethods.h: 14825 (JSC::JIT::emitLoadTag): 14826 (JSC::JIT::emitLoadPayload): 14827 (JSC::JIT::emitLoad): 14828 (JSC::JIT::emitLoad2): 14829 (JSC::JIT::isLabeled): 14830 (JSC::JIT::map): 14831 (JSC::JIT::unmap): 14832 (JSC::JIT::isMapped): 14833 (JSC::JIT::getMappedPayload): 14834 (JSC::JIT::getMappedTag): Use helper functions when loading virtual 14835 registers into machine registers, in case the loads can be eliminated 14836 by chaining. 14837 14838 * jit/JITOpcodes.cpp: 14839 (JSC::JIT::emit_op_mov): 14840 (JSC::JIT::emit_op_end): 14841 (JSC::JIT::emit_op_instanceof): 14842 (JSC::JIT::emit_op_get_global_var): 14843 (JSC::JIT::emit_op_put_global_var): 14844 (JSC::JIT::emit_op_get_scoped_var): 14845 (JSC::JIT::emit_op_put_scoped_var): 14846 (JSC::JIT::emit_op_to_primitive): 14847 (JSC::JIT::emit_op_resolve_global): 14848 (JSC::JIT::emit_op_jneq_ptr): 14849 (JSC::JIT::emit_op_next_pname): 14850 (JSC::JIT::emit_op_to_jsnumber): 14851 (JSC::JIT::emit_op_catch): Chain results from these opcodes. 14852 14853 (JSC::JIT::emit_op_profile_will_call): 14854 (JSC::JIT::emit_op_profile_did_call): Load the profiler into regT2 to 14855 avoid stomping a chained result. 14856 14857 * jit/JITPropertyAccess.cpp: 14858 (JSC::JIT::emit_op_method_check): 14859 (JSC::JIT::emit_op_get_by_val): 14860 (JSC::JIT::emit_op_get_by_id): Chain results from these opcodes. 14861 14862 * jit/JITStubCall.h: 14863 (JSC::JITStubCall::addArgument): Always use { regT1, regT0 }, to facilitate 14864 chaining. 14865 14866 (JSC::JITStubCall::call): Unmap all mapped registers, since our callee 14867 stub might stomp them. 14868 148692009-07-01 Sam Weinig <sam@webkit.org> 14870 14871 Reviewed by Gavin Barraclough. 14872 14873 Don't reload values in emitBinaryDoubleOp. 14874 14875 SunSpider reports a 0.6% progression. 14876 14877 * jit/JIT.h: 14878 * jit/JITArithmetic.cpp: 14879 (JSC::JIT::emit_op_jnless): 14880 (JSC::JIT::emit_op_jnlesseq): 14881 (JSC::JIT::emitBinaryDoubleOp): 14882 148832009-07-01 Sam Weinig <sam@webkit.org> 14884 14885 Reviewed by Geoffrey Garen. 14886 14887 Convert op_div to load op1 and op2 up front. 14888 14889 * jit/JITArithmetic.cpp: 14890 (JSC::JIT::emit_op_div): 14891 148922009-07-01 Sam Weinig <sam@webkit.org> 14893 14894 Reviewed by Geoffrey Garen. 14895 14896 Don't emit code in emitBinaryDoubleOp if code is unreachable, observable 14897 via an empty (unlinked) jumplist passed in. This only effects op_jnless 14898 and op_jnlesseq at present. 14899 14900 * jit/JITArithmetic.cpp: 14901 (JSC::JIT::emitSlow_op_jnless): 14902 (JSC::JIT::emitSlow_op_jnlesseq): 14903 (JSC::JIT::emitBinaryDoubleOp): 14904 149052009-07-01 Geoffrey Garen <ggaren@apple.com> 14906 14907 Reviewed by Sam Weinig. 14908 14909 Converted op_mod to put { tag, payload } in { regT1, regT0 }, and 14910 tidied up its constant case. 14911 14912 SunSpider reports a 0.2% regression, but a micro-benchmark of op_mod 14913 shows a 12% speedup, and the SunSpider test that uses op_mod most should 14914 benefit a lot from result caching in the end, since it almost always 14915 performs (expression) % constant. 14916 14917 * jit/JITArithmetic.cpp: 14918 (JSC::JIT::emit_op_mod): 14919 (JSC::JIT::emitSlow_op_mod): 14920 149212009-06-30 Sam Weinig <sam@webkit.org> 14922 14923 Reviewed by Geoffrey Garen. 14924 14925 Converted some more arithmetic ops to put { tag, payload } in 14926 { regT1, regT0 }. 14927 14928 * jit/JITArithmetic.cpp: 14929 (JSC::JIT::emit_op_mul): 14930 (JSC::JIT::emitSlow_op_mul): 14931 149322009-06-30 Geoffrey Garen <ggaren@apple.com> 14933 14934 Reviewed by Sam Weinig. 14935 14936 Converted some more arithmetic ops to put { tag, payload } in 14937 { regT1, regT0 }, and added a case for subtract constant. 14938 14939 SunSpider says no change. v8 says 0.3% slower. 14940 14941 * jit/JIT.h: 14942 * jit/JITArithmetic.cpp: 14943 (JSC::JIT::emit_op_add): 14944 (JSC::JIT::emitAdd32Constant): 14945 (JSC::JIT::emitSlow_op_add): 14946 (JSC::JIT::emit_op_sub): 14947 (JSC::JIT::emitSub32Constant): 14948 (JSC::JIT::emitSlow_op_sub): 14949 149502009-06-30 Gavin Barraclough <barraclough@apple.com> 14951 14952 Reviewed by Sam Weinig. 14953 14954 Remove more uses of addressFor(), load double constants directly from 14955 the constantpool in the CodeBlock, rather than from the register file. 14956 14957 * jit/JITArithmetic.cpp: 14958 (JSC::JIT::emitAdd32Constant): 14959 (JSC::JIT::emitBinaryDoubleOp): 14960 149612009-06-30 Geoffrey Garen <ggaren@apple.com> 14962 14963 Reviewed by Sam Weinig. 14964 14965 Fixed a bug in postfix ops, where we would treat x = x++ and x = x-- 14966 as a no-op, even if x were not an int, and the ++/-- could have side-effects. 14967 14968 * jit/JITArithmetic.cpp: 14969 (JSC::JIT::emit_op_post_inc): 14970 (JSC::JIT::emitSlow_op_post_inc): 14971 (JSC::JIT::emit_op_post_dec): 14972 (JSC::JIT::emitSlow_op_post_dec): 14973 149742009-06-30 Geoffrey Garen <ggaren@apple.com> 14975 14976 Reviewed by Sam Weinig. 14977 14978 Converted some arithmetic ops to put { tag, payload } in 14979 { regT1, regT0 }. 14980 14981 SunSpider says 0.7% faster. v8 says no change. 14982 14983 * jit/JIT.h: 14984 * jit/JITArithmetic.cpp: 14985 (JSC::JIT::emit_op_jnless): 14986 (JSC::JIT::emit_op_jnlesseq): 14987 (JSC::JIT::emit_op_lshift): 14988 (JSC::JIT::emit_op_rshift): 14989 (JSC::JIT::emit_op_bitand): 14990 (JSC::JIT::emit_op_bitor): 14991 (JSC::JIT::emit_op_bitxor): 14992 * jit/JITInlineMethods.h: 14993 (JSC::JIT::isOperandConstantImmediateInt): 14994 (JSC::JIT::getOperandConstantImmediateInt): 14995 149962009-06-30 Gavin Barraclough <barraclough@apple.com> 14997 14998 Reviewed by Sam Weinig. 14999 15000 Start removing cases of addressFor(). 15001 15002 * jit/JIT.h: 15003 * jit/JITArithmetic.cpp: 15004 (JSC::JIT::emitAdd32Constant): 15005 (JSC::JIT::emitBinaryDoubleOp): 15006 (JSC::JIT::emit_op_div): 15007 * jit/JITInlineMethods.h: 15008 (JSC::JIT::emitLoadDouble): 15009 (JSC::JIT::emitLoadInt32ToDouble): 15010 (JSC::JIT::emitStoreDouble): 15011 * jit/JITOpcodes.cpp: 15012 (JSC::JIT::emit_op_jfalse): 15013 (JSC::JIT::emit_op_jtrue): 15014 150152009-06-30 Geoffrey Garen <ggaren@apple.com> 15016 15017 Rolled back in my last patch with regression fixed. 15018 15019 * jit/JIT.cpp: 15020 (JSC::JIT::privateCompileSlowCases): 15021 * jit/JIT.h: 15022 * jit/JITOpcodes.cpp: 15023 (JSC::JIT::emit_op_loop_if_less): 15024 (JSC::JIT::emit_op_loop_if_lesseq): 15025 (JSC::JIT::emit_op_resolve_global): 15026 (JSC::JIT::emitSlow_op_resolve_global): 15027 (JSC::JIT::emit_op_eq): 15028 (JSC::JIT::emitSlow_op_eq): 15029 (JSC::JIT::emit_op_neq): 15030 (JSC::JIT::emitSlow_op_neq): 15031 150322009-06-30 Geoffrey Garen <ggaren@apple.com> 15033 15034 Rolled out my last patch because it was a 2% SunSpider regression. 15035 15036 * jit/JIT.cpp: 15037 (JSC::JIT::privateCompileSlowCases): 15038 * jit/JIT.h: 15039 * jit/JITOpcodes.cpp: 15040 (JSC::JIT::emit_op_loop_if_less): 15041 (JSC::JIT::emit_op_loop_if_lesseq): 15042 (JSC::JIT::emit_op_resolve_global): 15043 (JSC::JIT::emit_op_eq): 15044 (JSC::JIT::emitSlow_op_eq): 15045 (JSC::JIT::emit_op_neq): 15046 (JSC::JIT::emitSlow_op_neq): 15047 150482009-06-30 Geoffrey Garen <ggaren@apple.com> 15049 15050 Reviewed by Gavin "Sam Weinig" Barraclough. 15051 15052 Standardized the rest of our opcodes to put { tag, payload } in 15053 { regT1, regT0 } where possible. 15054 15055 * jit/JIT.cpp: 15056 (JSC::JIT::privateCompileSlowCases): 15057 * jit/JIT.h: 15058 * jit/JITOpcodes.cpp: 15059 (JSC::JIT::emit_op_loop_if_less): 15060 (JSC::JIT::emit_op_loop_if_lesseq): 15061 (JSC::JIT::emit_op_resolve_global): 15062 (JSC::JIT::emitSlow_op_resolve_global): 15063 (JSC::JIT::emit_op_eq): 15064 (JSC::JIT::emitSlow_op_eq): 15065 (JSC::JIT::emit_op_neq): 15066 (JSC::JIT::emitSlow_op_neq): 15067 150682009-06-30 Gavin Barraclough <barraclough@apple.com> 15069 15070 Reviewed by Geoffrey Garen. 15071 15072 Replace calls to store32(tagFor()) and store32(payloadFor()) 15073 with emitStoreInt32(), emitStoreBool(), and emitStoreCell(). 15074 15075 * jit/JIT.h: 15076 * jit/JITArithmetic.cpp: 15077 (JSC::JIT::emit_op_negate): 15078 (JSC::JIT::emit_op_lshift): 15079 (JSC::JIT::emit_op_rshift): 15080 (JSC::JIT::emit_op_bitand): 15081 (JSC::JIT::emitBitAnd32Constant): 15082 (JSC::JIT::emit_op_bitor): 15083 (JSC::JIT::emitBitOr32Constant): 15084 (JSC::JIT::emit_op_bitxor): 15085 (JSC::JIT::emitBitXor32Constant): 15086 (JSC::JIT::emit_op_bitnot): 15087 (JSC::JIT::emit_op_post_inc): 15088 (JSC::JIT::emit_op_post_dec): 15089 (JSC::JIT::emit_op_pre_inc): 15090 (JSC::JIT::emit_op_pre_dec): 15091 (JSC::JIT::emit_op_add): 15092 (JSC::JIT::emitAdd32Constant): 15093 (JSC::JIT::emit_op_sub): 15094 (JSC::JIT::emitSub32ConstantLeft): 15095 (JSC::JIT::emitSub32ConstantRight): 15096 (JSC::JIT::emit_op_mul): 15097 (JSC::JIT::emitSlow_op_mul): 15098 (JSC::JIT::emit_op_div): 15099 (JSC::JIT::emit_op_mod): 15100 * jit/JITCall.cpp: 15101 (JSC::JIT::emit_op_load_varargs): 15102 * jit/JITInlineMethods.h: 15103 (JSC::JIT::emitStoreInt32): 15104 (JSC::JIT::emitStoreCell): 15105 (JSC::JIT::emitStoreBool): 15106 (JSC::JIT::emitStore): 15107 * jit/JITOpcodes.cpp: 15108 (JSC::JIT::emit_op_instanceof): 15109 (JSC::JIT::emit_op_not): 15110 (JSC::JIT::emit_op_eq): 15111 (JSC::JIT::emitSlow_op_eq): 15112 (JSC::JIT::emit_op_neq): 15113 (JSC::JIT::emitSlow_op_neq): 15114 (JSC::JIT::compileOpStrictEq): 15115 (JSC::JIT::emit_op_eq_null): 15116 (JSC::JIT::emit_op_neq_null): 15117 * jit/JITStubCall.h: 15118 (JSC::JITStubCall::call): 15119 151202009-06-30 Geoffrey Garen <ggaren@apple.com> 15121 15122 Reviewed by Sam Weinig. 15123 15124 Standardized the rest of the property access instructions to put { tag, 15125 payload } in { regT1, regT0 }. 15126 15127 Small v8 speedup, 0.2% SunSpider slowdown. 15128 15129 * jit/JIT.h: 15130 * jit/JITInlineMethods.h: 15131 (JSC::JIT::emitLoad): 15132 (JSC::JIT::emitLoad2): 15133 * jit/JITPropertyAccess.cpp: 15134 (JSC::JIT::emit_op_get_by_val): 15135 (JSC::JIT::emitSlow_op_get_by_val): 15136 (JSC::JIT::emit_op_put_by_val): 15137 (JSC::JIT::emitSlow_op_put_by_val): 15138 (JSC::JIT::emit_op_put_by_id): 15139 (JSC::JIT::emitSlow_op_put_by_id): 15140 (JSC::JIT::patchPutByIdReplace): 15141 151422009-06-29 Sam Weinig <sam@webkit.org> 15143 15144 Reviewed by Gavin Barraclough. 15145 15146 Various cleanups. 15147 - Use fpRegT* instead of X86::xmm*. 15148 - Use a switch statement in emitBinaryDoubleOp instead of a bunch of 15149 if/elses. 15150 15151 * jit/JITArithmetic.cpp: 15152 (JSC::JIT::emitAdd32Constant): 15153 (JSC::JIT::emitBinaryDoubleOp): 15154 (JSC::JIT::emit_op_div): 15155 151562009-06-29 Sam Weinig <sam@webkit.org> 15157 15158 Reviewed by Geoffrey Garen. 15159 15160 Add inline code dealing with doubles for op_jfalse and op_jtrue. 15161 15162 * assembler/MacroAssemblerX86Common.h: 15163 (JSC::MacroAssemblerX86Common::): 15164 (JSC::MacroAssemblerX86Common::zeroDouble): 15165 * jit/JITOpcodes.cpp: 15166 (JSC::JIT::emit_op_jfalse): 15167 (JSC::JIT::emit_op_jtrue): 15168 151692009-06-28 Geoffrey Garen <ggaren@apple.com> 15170 15171 Reviewed by Sam Weinig. 15172 15173 Standardized op_get_by_id to put { tag, payload } in { regT1, regT0 }. 15174 15175 SunSpider and v8 report maybe 0.2%-0.4% regressions, but the optimization 15176 this enables will win much more than that back. 15177 15178 * jit/JIT.cpp: 15179 (JSC::JIT::privateCompileCTIMachineTrampolines): 15180 * jit/JIT.h: 15181 * jit/JITPropertyAccess.cpp: 15182 (JSC::JIT::emit_op_method_check): 15183 (JSC::JIT::emit_op_get_by_id): 15184 (JSC::JIT::compileGetByIdHotPath): 15185 (JSC::JIT::compileGetByIdSlowCase): 15186 (JSC::JIT::patchGetByIdSelf): 15187 (JSC::JIT::privateCompilePatchGetArrayLength): 15188 (JSC::JIT::privateCompileGetByIdProto): 15189 (JSC::JIT::privateCompileGetByIdSelfList): 15190 (JSC::JIT::privateCompileGetByIdProtoList): 15191 (JSC::JIT::privateCompileGetByIdChainList): 15192 (JSC::JIT::privateCompileGetByIdChain): 15193 151942009-06-26 Geoffrey Garen <ggaren@apple.com> 15195 15196 Reviewed by Maciej Stachowiak. 15197 15198 Standardized op_call to put { tag, payload } in { regT1, regT0 }. 15199 15200 SunSpider and v8 report no change. 15201 15202 * jit/JIT.cpp: 15203 (JSC::JIT::privateCompileCTIMachineTrampolines): 15204 * jit/JITCall.cpp: 15205 (JSC::JIT::compileOpCallInitializeCallFrame): 15206 (JSC::JIT::compileOpCallSetupArgs): 15207 (JSC::JIT::compileOpConstructSetupArgs): 15208 (JSC::JIT::compileOpCallVarargsSetupArgs): 15209 (JSC::JIT::compileOpCallVarargs): 15210 (JSC::JIT::compileOpCall): 15211 (JSC::JIT::compileOpCallSlowCase): 15212 152132009-06-26 Sam Weinig <sam@webkit.org> 15214 15215 Reviewed by Geoffrey Garen. 15216 15217 Handle multiplying by zero a little better by 15218 inlining the case that both operands are non-negative 15219 into the slowpath. 15220 15221 * assembler/MacroAssemblerX86Common.h: 15222 (JSC::MacroAssemblerX86Common::branchOr32): 15223 * jit/JITArithmetic.cpp: 15224 (JSC::JIT::emit_op_mul): 15225 (JSC::JIT::emitSlow_op_mul): 15226 152272009-06-25 Geoffrey Garen <ggaren@apple.com> 15228 15229 Reviewed by Sam Weinig. 15230 15231 Optimize x++ to ++x inside for loops. 15232 15233 Sadly, no measurable speedup, but this should help with result chaining. 15234 15235 * parser/Nodes.cpp: 15236 (JSC::ForNode::emitBytecode): 15237 152382009-06-25 Geoffrey Garen <ggaren@apple.com> 15239 15240 Reviewed by Sam Weinig. 15241 15242 Standardized some more opcodes to put { tag, payload } in { regT1, regT0 }. 15243 15244 * jit/JITArithmetic.cpp: 15245 (JSC::JIT::emitSlow_op_bitnot): 15246 (JSC::JIT::emit_op_post_inc): 15247 152482009-06-25 Geoffrey Garen <ggaren@apple.com> 15249 15250 Reviewed by Sam Weinig. 15251 15252 Standardized some more opcodes to put { tag, payload } in { regT1, regT0 }. 15253 15254 * jit/JITArithmetic.cpp: 15255 (JSC::JIT::emit_op_bitnot): 15256 (JSC::JIT::emit_op_post_dec): 15257 (JSC::JIT::emit_op_pre_inc): 15258 (JSC::JIT::emitSlow_op_pre_inc): 15259 (JSC::JIT::emit_op_pre_dec): 15260 (JSC::JIT::emitSlow_op_pre_dec): 15261 152622009-06-25 Geoffrey Garen <ggaren@apple.com> 15263 15264 Reviewed by Sam Weinig. 15265 15266 Standardized some more opcodes to put { tag, payload } in { regT1, regT0 }. 15267 15268 * jit/JITArithmetic.cpp: 15269 (JSC::JIT::emit_op_negate): 15270 (JSC::JIT::emitSlow_op_negate): 15271 * jit/JITCall.cpp: 15272 (JSC::JIT::emit_op_construct_verify): 15273 (JSC::JIT::emitSlow_op_construct_verify): 15274 152752009-06-25 Geoffrey Garen <ggaren@apple.com> 15276 15277 Reviewed by Sam Weinig. 15278 15279 Standardized some more opcodes to put { tag, payload } in { regT1, regT0 }. 15280 15281 * jit/JITOpcodes.cpp: 15282 (JSC::JIT::emit_op_loop_if_true): 15283 (JSC::JIT::emit_op_jfalse): 15284 (JSC::JIT::emit_op_jtrue): 15285 (JSC::JIT::emit_op_jeq_null): 15286 (JSC::JIT::emit_op_jneq_null): 15287 (JSC::JIT::emit_op_eq_null): 15288 (JSC::JIT::emit_op_neq_null): 15289 152902009-06-25 Geoffrey Garen <ggaren@apple.com> 15291 15292 Reviewed by Sam Weinig (sort of, maybe). 15293 15294 Fixed some ASSERTs in http/tests/security. 15295 15296 These ASSERTs were introduced by http://trac.webkit.org/changeset/45057, 15297 but the underlying problem was actually older. http://trac.webkit.org/changeset/45057 15298 just exposed the problem by enabling optimization in more cases. 15299 15300 The ASSERTs fired because we tested PropertySlot::slotBase() for validity, 15301 but slotBase() ASSERTs if it's invalid, so we would ASSERT before 15302 the test could happen. Solution: Remove the ASSERT. Maybe it was valid 15303 once, but it clearly goes against a pattern we've deployed of late. 15304 15305 The underlying problem was that WebCore would re-use a PropertySlot in 15306 the case of a forwarding access, and the second use would not completely 15307 overwrite the first use. Solution: Make sure to overwrite m_offset when 15308 setting a value on a PropertySlot. (Other values already get implicitly 15309 overwritten during reuse.) 15310 15311 * runtime/PropertySlot.h: 15312 (JSC::PropertySlot::PropertySlot): 15313 (JSC::PropertySlot::setValueSlot): 15314 (JSC::PropertySlot::setValue): 15315 (JSC::PropertySlot::setRegisterSlot): 15316 (JSC::PropertySlot::setUndefined): 15317 (JSC::PropertySlot::slotBase): 15318 (JSC::PropertySlot::clearOffset): 15319 153202009-06-24 Gavin Barraclough <barraclough@apple.com> 15321 15322 Reviewed by Geoff Garen. 15323 15324 Enable JIT_OPTIMIZE_METHOD_CALLS on the branch, implementation matches current implemenatation in ToT. 15325 15326 * jit/JIT.h: 15327 * jit/JITPropertyAccess.cpp: 15328 (JSC::JIT::emit_op_method_check): 15329 (JSC::JIT::emitSlow_op_method_check): 15330 (JSC::JIT::emit_op_get_by_id): 15331 (JSC::JIT::compileGetByIdHotPath): 15332 (JSC::JIT::emitSlow_op_get_by_id): 15333 (JSC::JIT::compileGetByIdSlowCase): 15334 153352009-06-23 Geoffrey Garen <ggaren@apple.com> 15336 15337 Reviewed by Sam Weinig. 15338 15339 Bit off a tiny bit more of standardizing opcode behavior to help with result 15340 caching. 15341 15342 SunSpider reports no change, v8 maybe a tiny speedup. 15343 15344 * jit/JITOpcodes.cpp: 15345 (JSC::JIT::emit_op_to_jsnumber): 15346 (JSC::JIT::emitSlow_op_to_jsnumber): 15347 (JSC::JIT::emit_op_convert_this): 15348 (JSC::JIT::emitSlow_op_convert_this): 15349 153502009-06-23 Geoffrey Garen <ggaren@apple.com> 15351 15352 Reviewed by Sam Weinig. 15353 15354 Bit off a tiny bit more of standardizing opcode behavior to help with result 15355 caching -- including removing my old enemy, op_resolve_function, because 15356 it was non-standard, and removing it felt better than helping it limp along. 15357 15358 SunSpider reports no change, v8 maybe a tiny speedup. 15359 15360 * bytecode/CodeBlock.cpp: 15361 (JSC::CodeBlock::dump): 15362 * bytecode/Opcode.h: 15363 * bytecompiler/BytecodeGenerator.cpp: 15364 * bytecompiler/BytecodeGenerator.h: 15365 * interpreter/Interpreter.cpp: 15366 (JSC::Interpreter::privateExecute): 15367 * jit/JIT.cpp: 15368 (JSC::JIT::privateCompileMainPass): 15369 * jit/JIT.h: 15370 * jit/JITOpcodes.cpp: 15371 (JSC::JIT::emit_op_get_scoped_var): 15372 (JSC::JIT::emit_op_put_scoped_var): 15373 (JSC::JIT::emit_op_to_primitive): 15374 (JSC::JIT::emitSlow_op_to_primitive): 15375 * jit/JITStubs.cpp: 15376 * jit/JITStubs.h: 15377 * parser/Nodes.cpp: 15378 (JSC::FunctionCallResolveNode::emitBytecode): 15379 153802009-06-23 Geoffrey Garen <ggaren@apple.com> 15381 15382 Reviewed by Sam Weinig. 15383 15384 Bit off a tiny bit of standardizing opcode behavior to help with result 15385 caching. 15386 15387 0.6% SunSpider speedup. 0.3% v8 speedup. 15388 15389 * jit/JITInlineMethods.h: 15390 (JSC::JIT::emitLoad): Accomodate a base register that overlaps with payload 15391 by loading tag before payload, to avoid stomping base/payload. 15392 15393 * jit/JITOpcodes.cpp: 15394 (JSC::JIT::emit_op_mov): Abide by the standard "tag in regT1, payload in 15395 regT0" semantics. 15396 15397 (JSC::JIT::emit_op_get_global_var): 15398 (JSC::JIT::emit_op_put_global_var): Ditto. Also, removed some irrelevent 15399 loads while I was at it. The global object's "d" pointer never changes 15400 after construction. 15401 154022009-06-23 Gavin Barraclough <barraclough@apple.com> 15403 15404 Reviewed by Sam Weinig. 15405 15406 Remove 'arguments' field from Register union (again). 15407 This time do so without breaking tests (radical, I know). 15408 15409 * interpreter/CallFrame.h: 15410 (JSC::ExecState::optionalCalleeArguments): 15411 (JSC::ExecState::setArgumentCount): 15412 (JSC::ExecState::init): 15413 * interpreter/Interpreter.cpp: 15414 (JSC::Interpreter::dumpRegisters): 15415 (JSC::Interpreter::unwindCallFrame): 15416 (JSC::Interpreter::privateExecute): 15417 (JSC::Interpreter::retrieveArguments): 15418 * interpreter/Register.h: 15419 (JSC::Register::withInt): 15420 (JSC::Register::): 15421 (JSC::Register::Register): 15422 (JSC::Register::i): 15423 * jit/JITStubs.cpp: 15424 (JSC::JITStubs::cti_op_tear_off_arguments): 15425 * runtime/Arguments.h: 15426 (JSC::JSActivation::copyRegisters): 15427 (JSC::Register::arguments): 15428 * runtime/JSActivation.cpp: 15429 (JSC::JSActivation::argumentsGetter): 15430 * runtime/JSActivation.h: 15431 154322009-06-23 Geoffrey Garen <ggaren@apple.com> 15433 15434 Reviewed by Sam Weinig. 15435 15436 Removed some result register tracking cruft in preparation for a new 15437 result tracking mechanism. 15438 15439 SunSpider reports no change. 15440 15441 * assembler/AbstractMacroAssembler.h: 15442 * assembler/X86Assembler.h: 15443 (JSC::X86Assembler::JmpDst::JmpDst): No need to track jump targets in 15444 machine code; we already do this in bytecode. 15445 15446 * jit/JIT.cpp: 15447 (JSC::JIT::JIT): 15448 (JSC::JIT::emitTimeoutCheck): Make sure to save and restore the result 15449 registers, so an opcode with a timeout check can still benefit from result 15450 register caching. 15451 15452 (JSC::JIT::privateCompileMainPass): 15453 (JSC::JIT::privateCompileSlowCases): Removed calls to killLastResultRegister() 15454 in preparation for something new. 15455 15456 * jit/JIT.h: 15457 * jit/JITArithmetic.cpp: 15458 (JSC::JIT::emit_op_jnless): 15459 (JSC::JIT::emit_op_jnlesseq): 15460 * jit/JITInlineMethods.h: 15461 (JSC::JIT::emitGetFromCallFrameHeaderPtr): 15462 (JSC::JIT::emitGetFromCallFrameHeader32): 15463 * jit/JITOpcodes.cpp: 15464 (JSC::JIT::emit_op_jmp): 15465 (JSC::JIT::emit_op_jfalse): 15466 (JSC::JIT::emit_op_jtrue): 15467 (JSC::JIT::emit_op_jeq_null): 15468 (JSC::JIT::emit_op_jneq_null): 15469 (JSC::JIT::emit_op_jneq_ptr): 15470 (JSC::JIT::emit_op_jsr): 15471 (JSC::JIT::emit_op_sret): 15472 (JSC::JIT::emit_op_jmp_scopes): ditto 15473 15474 * jit/JITStubCall.h: 15475 (JSC::JITStubCall::JITStubCall): 15476 (JSC::JITStubCall::getArgument): added a mechanism for reloading an argument 15477 you passed to a JIT stub, for use in emitTimeoutCheck. 15478 154792009-06-23 Sam Weinig <sam@webkit.org> 15480 15481 Reviewed by Geoffrey Garen. 15482 15483 Remove now-useless inplace variants of binary ops. 15484 15485 * jit/JIT.h: 15486 * jit/JITArithmetic.cpp: 15487 (JSC::JIT::emit_op_bitand): 15488 (JSC::JIT::emit_op_bitor): 15489 (JSC::JIT::emit_op_bitxor): 15490 (JSC::JIT::emit_op_add): 15491 (JSC::JIT::emit_op_sub): 15492 (JSC::JIT::emit_op_mul): 15493 154942009-06-23 Sam Weinig <sam@webkit.org> 15495 15496 Reviewed by Geoffrey Garen. 15497 15498 Move off memory operands to aid in re-enabling result caching. 15499 15500 - No regression measured. 15501 15502 * jit/JIT.h: 15503 * jit/JITArithmetic.cpp: 15504 (JSC::JIT::emit_op_negate): 15505 (JSC::JIT::emit_op_jnless): 15506 (JSC::JIT::emit_op_jnlesseq): 15507 (JSC::JIT::emit_op_lshift): 15508 (JSC::JIT::emit_op_rshift): 15509 (JSC::JIT::emit_op_bitand): 15510 (JSC::JIT::emitBitAnd32Constant): 15511 (JSC::JIT::emitBitAnd32InPlace): 15512 (JSC::JIT::emit_op_bitor): 15513 (JSC::JIT::emitBitOr32Constant): 15514 (JSC::JIT::emitBitOr32InPlace): 15515 (JSC::JIT::emit_op_bitxor): 15516 (JSC::JIT::emitBitXor32Constant): 15517 (JSC::JIT::emitBitXor32InPlace): 15518 (JSC::JIT::emit_op_bitnot): 15519 (JSC::JIT::emit_op_post_inc): 15520 (JSC::JIT::emit_op_post_dec): 15521 (JSC::JIT::emit_op_pre_inc): 15522 (JSC::JIT::emitSlow_op_pre_inc): 15523 (JSC::JIT::emit_op_pre_dec): 15524 (JSC::JIT::emitSlow_op_pre_dec): 15525 (JSC::JIT::emit_op_add): 15526 (JSC::JIT::emitAdd32Constant): 15527 (JSC::JIT::emitAdd32InPlace): 15528 (JSC::JIT::emitSlow_op_add): 15529 (JSC::JIT::emitSlowAdd32Constant): 15530 (JSC::JIT::emit_op_sub): 15531 (JSC::JIT::emitSlow_op_sub): 15532 (JSC::JIT::emitSub32ConstantLeft): 15533 (JSC::JIT::emitSub32ConstantRight): 15534 (JSC::JIT::emitSub32InPlaceLeft): 15535 (JSC::JIT::emitSub32InPlaceRight): 15536 (JSC::JIT::emitBinaryDoubleOp): 15537 (JSC::JIT::emit_op_mul): 15538 (JSC::JIT::emitMul32InPlace): 15539 (JSC::JIT::emit_op_div): 15540 (JSC::JIT::emit_op_mod): 15541 * jit/JITCall.cpp: 15542 (JSC::JIT::compileOpCallVarargs): 15543 * jit/JITOpcodes.cpp: 15544 (JSC::JIT::emit_op_loop_if_less): 15545 (JSC::JIT::emit_op_loop_if_lesseq): 15546 (JSC::JIT::emit_op_instanceof): 15547 (JSC::JIT::emit_op_to_primitive): 15548 (JSC::JIT::emit_op_not): 15549 (JSC::JIT::emit_op_jneq_ptr): 15550 (JSC::JIT::emit_op_eq): 15551 (JSC::JIT::emit_op_neq): 15552 (JSC::JIT::emit_op_to_jsnumber): 15553 * jit/JITPropertyAccess.cpp: 15554 (JSC::JIT::emit_op_get_by_val): 15555 (JSC::JIT::emit_op_put_by_val): 15556 155572009-06-23 Geoffrey Garen <ggaren@apple.com> 15558 15559 Reviewed by Sam Weinig. 15560 15561 Fixed some missing and/or misplaced labels in bytecode generation, so 15562 we don't have to work around them in JIT code generation. 15563 15564 * bytecompiler/BytecodeGenerator.cpp: 15565 (JSC::BytecodeGenerator::emitJumpSubroutine): 15566 * parser/Nodes.cpp: 15567 (JSC::TryNode::emitBytecode): 15568 155692009-06-22 Geoffrey Garen <ggaren@apple.com> 15570 15571 Reviewed by Sam Weinig. 15572 15573 For member function calls, emit "this" directly into the "this" slot 15574 for the function call, instead of moving it there later. This reduces 15575 time spent in op_mov during certain calls, like "a.b.c()". 15576 15577 1%-2% speedup on v8, mostly richards and delta-blue. 15578 15579 * parser/Nodes.cpp: 15580 (JSC::FunctionCallDotNode::emitBytecode): 15581 155822009-06-22 Gavin Barraclough <barraclough@apple.com> 15583 15584 Reviewed by Sam Weinig. 15585 15586 Remove 'arguments' field from Register union. Having JSCell derived types in the union is 15587 dangerous since it opens the possibility for the field to be written as a raw pointer but 15588 then read as a JSValue. This will lead to statle data being read for the tag, which may 15589 be dangerous. Having removed Arguments* types form Register, all arguments objects must 15590 always explicitly be stored in the register file as JSValues. 15591 15592 * interpreter/CallFrame.h: 15593 (JSC::ExecState::optionalCalleeArguments): 15594 * interpreter/Interpreter.cpp: 15595 (JSC::Interpreter::unwindCallFrame): 15596 (JSC::Interpreter::privateExecute): 15597 (JSC::Interpreter::retrieveArguments): 15598 * interpreter/Register.h: 15599 (JSC::Register::): 15600 * jit/JITStubs.cpp: 15601 (JSC::JITStubs::cti_op_tear_off_arguments): 15602 * runtime/Arguments.h: 15603 (JSC::JSActivation::copyRegisters): 15604 * runtime/JSActivation.cpp: 15605 (JSC::JSActivation::argumentsGetter): 15606 * runtime/JSActivation.h: 15607 156082009-06-03 Sam Weinig <sam@webkit.org> 15609 15610 Reviewed by Geoffrey Garen. 15611 15612 Add back known this value optimization by abstracting 15613 slow case if not JSCell jumps. 15614 15615 * jit/JIT.h: 15616 * jit/JITCall.cpp: 15617 (JSC::JIT::compileOpCallVarargs): 15618 (JSC::JIT::compileOpCallVarargsSlowCase): 15619 (JSC::JIT::compileOpCall): 15620 (JSC::JIT::compileOpCallSlowCase): 15621 * jit/JITInlineMethods.h: 15622 (JSC::JIT::emitJumpSlowCaseIfNotJSCell): 15623 (JSC::JIT::linkSlowCaseIfNotJSCell): 15624 * jit/JITOpcodes.cpp: 15625 (JSC::JIT::emit_op_instanceof): 15626 (JSC::JIT::emitSlow_op_instanceof): 15627 * jit/JITPropertyAccess.cpp: 15628 (JSC::JIT::emit_op_get_by_val): 15629 (JSC::JIT::emitSlow_op_get_by_val): 15630 (JSC::JIT::emit_op_put_by_val): 15631 (JSC::JIT::emitSlow_op_put_by_val): 15632 (JSC::JIT::emit_op_get_by_id): 15633 (JSC::JIT::emitSlow_op_get_by_id): 15634 (JSC::JIT::emit_op_put_by_id): 15635 (JSC::JIT::emitSlow_op_put_by_id): 15636 156372009-06-01 Geoffrey Garen <ggaren@apple.com> 15638 15639 Reviewed by Sam Weinig. 15640 15641 Fixed some of the regression in crypto-aes.js. (8.5% speedup in 15642 crypto-aes.js.) 15643 15644 SunSpider reports no change overall. 15645 15646 Division was producing double results, which took the slow path through 15647 array access code. 15648 15649 Strangely, all my attempts at versions of this patch that modified array 15650 access code to accept ints encoded as doubles along the fast or slow paths 15651 were regressions. So I did this instead. 15652 15653 * jit/JITArithmetic.cpp: 15654 (JSC::JIT::emit_op_div): When dividing an int by an int, go ahead and try 15655 to turn the result into an int. Don't just do int division, though, because 15656 testing shows it to be slower than SSE double division, and the corner 15657 cases are pretty complicated / lengthy on top of that. Also, don't try 15658 to canonicalize division of known tiny numerators into ints, since that's a 15659 waste of time. 15660 156612009-05-26 Geoffrey Garen <ggaren@apple.com> 15662 15663 Reviewed by Oliver Hunt. 15664 15665 Fixed a regression caused by my recent fix for NaN. 15666 15667 * jit/JITArithmetic.cpp: 15668 (JSC::JIT::emitBinaryDoubleOp): Actually do the comparison in reverse 15669 order, like the ChangeLog said we would, bokay? 15670 156712009-05-26 Geoffrey Garen <ggaren@apple.com> 15672 15673 Reviewed by Sam Weinig and Oliver Hunt. 15674 15675 Fixed two edge cases in %: 15676 15677 - Don't do -2147483648 % x as a fast case, since you might do -2147483648 % -1, 15678 which will signal a hardware exception due to overflow. 15679 15680 - In the case of a zero remainder, be sure to store negative zero if the 15681 dividend was zero. 15682 15683 SunSpider reports no change. 15684 15685 * jit/JITArithmetic.cpp: 15686 (JSC::JIT::emit_op_mod): 15687 (JSC::JIT::emitSlow_op_mod): 15688 156892009-05-25 Geoffrey Garen <ggaren@apple.com> 15690 15691 Reviewed by Maciej Stachowiak. 15692 15693 Fixed a regression when comparing to NaN. 15694 15695 * jit/JITArithmetic.cpp: 15696 (JSC::JIT::emitBinaryDoubleOp): For op_jnless and op_jnless_eq, do the 15697 comparison in reverse order, and jump if the result is below or 15698 below-or-equal. This ensures that we do jump in the case of NaN. 15699 157002009-05-25 Geoffrey Garen <ggaren@apple.com> 15701 15702 Reviewed by Oliver Hunt. 15703 15704 SunSpider says no change. 15705 15706 Fixed regressions in fast/js/var-declarations-shadowing.html and 15707 fast/js/equality.html, caused by recent == and != optimizations. 15708 15709 * jit/JITStubs.cpp: 15710 (JSC::JITStubs::cti_op_eq): Don't treat "compare to string" as always 15711 numeric or string comparison. If the second operand is an object, you 15712 need to ToPrimitive it, and start all over again. Also, I wrote out each 15713 of the possible cases explicitly, to cut down on redundant branching. 15714 157152009-05-25 Sam Weinig <sam@webkit.org> 15716 15717 Reviewed by Mark Rowe. 15718 15719 Fix bug in fast/js/constant-folding.html where we were not negating 15720 -0 properly. 15721 15722 * jit/JITArithmetic.cpp: 15723 (JSC::JIT::emit_op_negate): 15724 157252009-05-23 Geoffrey Garen <ggaren@apple.com> 15726 15727 Reviewed by Oliver Hunt. 15728 15729 Refactored new slow case codegen for == and !=. 15730 15731 SunSpider reports no change, maybe a tiny speedup. 15732 15733 * jit/JITOpcodes.cpp: 15734 (JSC::JIT::emitSlow_op_eq): 15735 (JSC::JIT::emitSlow_op_neq): Made a vptr comparison a *Ptr operation, 15736 instead of *32, to make it portable to 64bit. Reorganized the string 15737 and generic cases to make their control flow a little clearer. 15738 157392009-05-23 Geoffrey Garen <ggaren@apple.com> 15740 15741 Reviewed by Maciej Stachowiak. 15742 15743 Optimized == and != for our new value representation -- especially for strings. 15744 15745 14% speedup on date-format-tofte. 15746 15747 * jit/JITOpcodes.cpp: 15748 (JSC::JIT::emit_op_eq): 15749 (JSC::JIT::emitSlow_op_eq): 15750 (JSC::JIT::emit_op_neq): 15751 (JSC::JIT::emitSlow_op_neq): 15752 * jit/JITStubCall.h: 15753 (JSC::JITStubCall::JITStubCall): 15754 * jit/JITStubs.cpp: 15755 (JSC::JITStubs::cti_op_eq): 15756 (JSC::JITStubs::cti_op_eq_strings): 15757 (JSC::JITStubs::cti_op_call_eval): 15758 * jit/JITStubs.h: 15759 (JSC::): 15760 * runtime/JSValue.h: 15761 157622009-05-22 Sam Weinig <sam@webkit.org> 15763 15764 Reviewed by Gavin Barraclough. 15765 15766 Fix non-SSE enabled builds. 15767 15768 * jit/JITArithmetic.cpp: 15769 (JSC::JIT::emitSlow_op_add): Don't early return here, we still need to call the JIT stub. 15770 (JSC::JIT::emitSlow_op_sub): Ditto. 15771 157722009-05-22 Geoffrey Garen <ggaren@apple.com> 15773 15774 Reviewed by Sam Weinig. 15775 15776 Here's a thought: let's not take a jit stub call just to multiply by 1, 15777 bokay? 15778 15779 imul doesn't set the zero flag, so to test for a zero result, we need 15780 an explicit instruction. (Luckily, it does set the overflow flag, so 15781 we can still use that.) 15782 15783 * jit/JIT.h: 15784 * jit/JITArithmetic.cpp: 15785 (JSC::JIT::emit_op_mul): 15786 (JSC::JIT::emitSlow_op_mul): 15787 (JSC::JIT::emitMul32InPlace): 15788 157892009-05-22 Sam Weinig <sam@webkit.org> 15790 15791 Reviewed by Geoffrey "Premature Commit" Garen. 15792 15793 Add back constant integer cases for op_add. 15794 15795 * jit/JIT.h: 15796 * jit/JITArithmetic.cpp: 15797 (JSC::JIT::emit_op_add): 15798 (JSC::JIT::emitAdd32Constant): 15799 (JSC::JIT::emitSlow_op_add): 15800 (JSC::JIT::emitSlowAdd32Constant): 15801 * jit/JITInlineMethods.h: 15802 (JSC::JIT::getConstantOperandImmediateDouble): 15803 (JSC::JIT::isOperandConstantImmediateDouble): 15804 158052009-05-22 Geoffrey Garen <ggaren@apple.com> 15806 15807 Reviewed by Sam Weinig. 15808 15809 Added fast double cases for op_jnless and op_jnlesseq. 15810 15811 * assembler/AbstractMacroAssembler.h: 15812 (JSC::AbstractMacroAssembler::JumpList::jumps): New accesor, used by 15813 addSlowCase. 15814 15815 * assembler/X86Assembler.h: 15816 (JSC::X86Assembler::ucomisd_rm): New method for comparing register to 15817 memory. 15818 15819 * jit/JIT.h: 15820 * jit/JITArithmetic.cpp: 15821 (JSC::JIT::emit_op_jnless): 15822 (JSC::JIT::emitSlow_op_jnless): 15823 (JSC::JIT::emit_op_jnlesseq): 15824 (JSC::JIT::emitSlow_op_jnlesseq): 15825 (JSC::JIT::emit_op_add): 15826 (JSC::JIT::emit_op_sub): 15827 (JSC::JIT::emitBinaryDoubleOp): 15828 (JSC::JIT::emit_op_mul): 15829 (JSC::JIT::emit_op_div): Modified emitBinaryDoubleOp to accept comparison/jump 15830 operations in addition to operations with explicit result registers. 15831 15832 * jit/JITInlineMethods.h: 15833 (JSC::JIT::addSlowCase): Added an "addSlowCase" for JumpLists, so clients 15834 can track multiple jumps to the same slow case condition together. 15835 158362009-05-21 Sam Weinig <sam@webkit.org> 15837 15838 Reviewed by Gavin Barraclough. 15839 15840 Implement op_negate inline fast cases. 15841 15842 * assembler/MacroAssemblerX86Common.h: 15843 (JSC::MacroAssemblerX86Common::neg32): 15844 * assembler/X86Assembler.h: 15845 (JSC::X86Assembler::): 15846 (JSC::X86Assembler::negl_m): 15847 (JSC::X86Assembler::xorpd_rr): 15848 * jit/JIT.cpp: 15849 (JSC::JIT::privateCompileMainPass): 15850 (JSC::JIT::privateCompileSlowCases): 15851 * jit/JIT.h: 15852 * jit/JITArithmetic.cpp: 15853 (JSC::JIT::emit_op_negate): 15854 (JSC::JIT::emitSlow_op_negate): 15855 158562009-05-20 Sam Weinig <sam@webkit.org> 15857 15858 Reviewed by Gavin Barraclough. 15859 15860 Update the patchOffsetGetByIdSlowCaseCall constant for the 15861 case that OPCODE_SAMPLING is enabled. 15862 15863 * jit/JIT.h: 15864 158652009-05-20 Geoffrey Garen <ggaren@apple.com> 15866 15867 Reviewed by Sam Weinig. 15868 15869 Added support for inline subtraction of doubles. 15870 15871 * jit/JITArithmetic.cpp: 15872 (JSC::JIT::emit_op_sub): 15873 (JSC::JIT::emitSlow_op_sub): 15874 (JSC::JIT::emitSlowSub32InPlaceLeft): 15875 (JSC::JIT::emitBinaryDoubleOp): 15876 158772009-05-20 Sam Weinig <sam@webkit.org> 15878 15879 Reviewed by Geoffrey Garen. 15880 15881 Added support for inline division. 15882 15883 * assembler/X86Assembler.h: 15884 (JSC::X86Assembler::): 15885 (JSC::X86Assembler::divsd_rr): 15886 (JSC::X86Assembler::divsd_mr): 15887 * bytecode/CodeBlock.cpp: 15888 (JSC::CodeBlock::dump): 15889 * bytecode/Opcode.h: 15890 * bytecompiler/BytecodeGenerator.cpp: 15891 (JSC::BytecodeGenerator::emitBinaryOp): 15892 * interpreter/Interpreter.cpp: 15893 (JSC::Interpreter::privateExecute): 15894 * jit/JIT.cpp: 15895 (JSC::JIT::privateCompileMainPass): 15896 (JSC::JIT::privateCompileSlowCases): 15897 * jit/JIT.h: 15898 * jit/JITArithmetic.cpp: 15899 (JSC::JIT::emitBinaryDoubleOp): 15900 (JSC::JIT::emit_op_div): 15901 (JSC::JIT::emitSlow_op_div): 15902 159032009-05-20 Geoffrey Garen <ggaren@apple.com> 15904 15905 Reviewed by Sam Weinig. 15906 15907 Added support for inline addition of doubles. 15908 15909 * jit/JITArithmetic.cpp: 15910 (JSC::JIT::emit_op_add): 15911 (JSC::JIT::emitSlow_op_add): 15912 (JSC::JIT::emitSlowAdd32InPlace): 15913 (JSC::JIT::emitBinaryDoubleOp): 15914 (JSC::JIT::emit_op_mul): 15915 (JSC::JIT::emitSlow_op_mul): 15916 159172009-05-20 Geoffrey Garen <ggaren@apple.com> 15918 15919 Reviewed by Sam Weinig. 15920 15921 Factored inline double operations into a helper function, so that we 15922 can reuse this code for other math operations. 15923 15924 * jit/JIT.h: 15925 * jit/JITArithmetic.cpp: 15926 (JSC::JIT::emitBinaryDoubleOp): 15927 (JSC::JIT::emit_op_mul): 15928 * jit/JITCall.cpp: 15929 (JSC::JIT::compileOpCallInitializeCallFrame): 15930 159312009-05-20 Geoffrey Garen <ggaren@apple.com> 15932 15933 Reviewed by Sam Weinig. 15934 15935 Added support for inline multiplication of doubles. 15936 15937 * assembler/X86Assembler.h: 15938 (JSC::X86Assembler::cvtsi2sd_mr): New function, useful for loading an 15939 int32 into a double register. 15940 15941 * jit/JITArithmetic.cpp: 15942 (JSC::JIT::emit_op_mul): 15943 (JSC::JIT::emitSlow_op_mul): Filled out these cases for double arithmetic. 15944 15945 * jit/JIT.h: 15946 * jit/JITInlineMethods.h: 15947 (JSC::JIT::addressFor): New function, useful for addressing a JSValue's 15948 full 64bits as a double. 15949 159502009-05-19 Sam Weinig <sam@webkit.org> 15951 15952 Reviewed by Geoffrey Garen. 15953 15954 Implement and enable optimized calls. 15955 15956 * jit/JIT.cpp: 15957 (JSC::JIT::privateCompileCTIMachineTrampolines): Add ENABLE(JIT_OPTIMIZE_CALL) guards 15958 around the the optimize call only trampolines (virtualCallPreLink and virtualCallLink). 15959 Update the trampolines to account for the new JSValue representation. 15960 (JSC::JIT::unlinkCall): Use NULL instead of JSValue noValue. 15961 15962 * jit/JITCall.cpp: 15963 (JSC::JIT::compileOpCall): Update to account for the new JSValue representation 15964 (JSC::JIT::compileOpCallSlowCase): Ditto. 15965 15966 * jit/JITStubs.h: Remove incorrect !ENABLE(JIT_OPTIMIZE_CALL) guard. 15967 15968 * wtf/Platform.h: Enable ENABLE_JIT_OPTIMIZE_CALL. 15969 159702009-05-19 Sam Weinig <sam@webkit.org> 15971 15972 Reviewed by Geoffrey Garen. 15973 15974 Implement and enable optimized property access. 15975 15976 * assembler/AbstractMacroAssembler.h: Fix comment. 15977 * jit/JIT.cpp: 15978 (JSC::JIT::privateCompileCTIMachineTrampolines): Remove array length trampoline 15979 and implement the string length trampoline. 15980 * jit/JIT.h: Add new constants for patch offsets. 15981 * jit/JITInlineMethods.h: Remove FIELD_OFFSET which is now in StdLibExtras.h. 15982 * jit/JITPropertyAccess.cpp: 15983 (JSC::JIT::emit_op_get_by_id): 15984 (JSC::JIT::emitSlow_op_get_by_id): 15985 (JSC::JIT::emit_op_put_by_id): 15986 (JSC::JIT::emitSlow_op_put_by_id): 15987 (JSC::JIT::compilePutDirectOffset): 15988 (JSC::JIT::compileGetDirectOffset): 15989 (JSC::JIT::privateCompilePutByIdTransition): 15990 (JSC::JIT::patchGetByIdSelf): 15991 (JSC::JIT::patchPutByIdReplace): 15992 (JSC::JIT::privateCompilePatchGetArrayLength): 15993 (JSC::JIT::privateCompileGetByIdProto): 15994 (JSC::JIT::privateCompileGetByIdSelfList): 15995 (JSC::JIT::privateCompileGetByIdProtoList): 15996 (JSC::JIT::privateCompileGetByIdChainList): 15997 (JSC::JIT::privateCompileGetByIdChain): 15998 * jit/JITStubCall.h: 15999 (JSC::JITStubCall::addArgument): Add version of addArgument that takes 16000 two registers for the tag and payload. 16001 * jit/JITStubs.cpp: 16002 (JSC::JITStubs::JITStubs): Remove array length trampoline pointer. 16003 (JSC::JITStubs::cti_op_get_by_id_self_fail): 16004 * jit/JITStubs.h: 16005 * runtime/JSObject.h: 16006 (JSC::JSObject::JSObject): Move m_inheritorID below the property storage 16007 to align it to a 16 byte boundary. 16008 * wtf/Platform.h: Enable ENABLE_JIT_OPTIMIZE_PROPERTY_ACCESS 16009 * wtf/StdLibExtras.h: Move FIELD_OFFSET here. 16010 160112009-05-17 Sam Weinig <sam@webkit.org> 16012 16013 Reviewed by Geoffrey Garen. 16014 16015 Remove unneeded ExecState parameter from the number JSValue constructors. 16016 16017 * runtime/JSValue.h: 16018 (JSC::jsNumber): 16019 (JSC::jsNaN): 16020 (JSC::JSValue::JSValue): 16021 160222009-05-15 Sam Weinig <sam@webkit.org> 16023 16024 Reviewed by Geoffrey Garen. 16025 16026 Implemented fast path for op_put_by_val when putting to arrays. 16027 16028 * jit/JITPropertyAccess.cpp: 16029 (JSC::JIT::emit_op_put_by_val): 16030 (JSC::JIT::emitSlow_op_put_by_val): 16031 160322009-05-15 Geoffrey Garen <ggaren@apple.com> (Mostly by Sam) 16033 16034 Reviewed by Sam Weinig. 16035 16036 Implemented fast path for op_get_by_val when accessing array. 16037 16038 * jit/JIT.cpp: 16039 * jit/JITPropertyAccess.cpp: 16040 (JSC::JIT::emit_op_get_by_val): 16041 (JSC::JIT::emitSlow_op_get_by_val): 16042 160432009-05-14 Geoffrey Garen <ggaren@apple.com> 16044 16045 Reviewed by Sam Weinig. 16046 16047 Fixed a failure in fast/js/math-transforms.html caused by failing to 16048 preserve -0 in multiplication. 16049 16050 * assembler/X86Assembler.h: 16051 (JSC::X86Assembler::jz): 16052 * jit/JITArithmetic.cpp: 16053 (JSC::JIT::emit_op_mul): 16054 (JSC::JIT::emitSlow_op_mul): 16055 (JSC::JIT::emitMul32Constant): 16056 (JSC::JIT::emitMul32InPlace): Check both for overflow and for zero when 16057 doing multiplication. Use a slow case to get these right. 16058 160592009-05-14 Geoffrey Garen <ggaren@apple.com> 16060 16061 Reviewed by Sam Weinig. 16062 16063 Fixed a bug in the varargs calling convention. 16064 16065 * jit/JITCall.cpp: 16066 (JSC::JIT::compileOpCallVarargs): Move the argument count into regT1, 16067 since that's where ctiVirtualCall expects it to be. 16068 160692009-05-14 Geoffrey Garen <ggaren@apple.com> 16070 16071 Reviewed by Sam Weinig. 16072 16073 Fixed a small bug in instanceof's looping code. 16074 16075 * jit/JITOpcodes.cpp: 16076 (JSC::JIT::emit_op_instanceof): NULL means the object has no prototype, 16077 so only loop when *not* equal to NULL. 16078 160792009-05-14 Geoffrey Garen <ggaren@apple.com> 16080 16081 Reviewed by Sam Weinig. 16082 16083 Fixed a small bug in instanceof's result writing code. 16084 16085 * jit/JITOpcodes.cpp: 16086 (JSC::JIT::emit_op_instanceof): Make sure to fill out the payload bits 16087 in all cases. 16088 160892009-05-14 Sam Weinig <sam@webkit.org> 16090 16091 Reviewed by Geoffrey Garen. 16092 16093 Removed an invalid assertion in cti_op_urshift which 16094 depended on a fast path for op_urshift which has 16095 never existed. 16096 16097 * jit/JITStubs.cpp: 16098 (JSC::JITStubs::cti_op_urshift): 16099 161002009-05-14 Geoffrey Garen <ggaren@apple.com> 16101 16102 Reviewed by Sam Weinig. 16103 16104 Fixed loop_if_true, which had the same reversed test that jtrue had. 16105 16106 * jit/JITOpcodes.cpp: 16107 (JSC::JIT::emit_op_loop_if_true): 16108 161092009-05-14 Sam Weinig <sam@webkit.org> 16110 16111 Reviewed by Geoffrey Garen. 16112 16113 In op_neq, we apparently want to check that one value 16114 does *not* equal another. Go figure. 16115 16116 * jit/JITOpcodes.cpp: 16117 (JSC::JIT::emit_op_neq): 16118 161192009-05-14 Sam Weinig <sam@webkit.org> 16120 16121 Reviewed by Geoffrey Garen. 16122 16123 The slow case of op_mod should call op_mod's jit stub, 16124 not op_mul. That would be dumb. 16125 16126 * jit/JITArithmetic.cpp: 16127 (JSC::JIT::emitSlow_op_mod): 16128 161292009-05-14 Geoffrey Garen <ggaren@apple.com> 16130 16131 Reviewed by Sam Weinig. 16132 16133 Fixed problems when using 'arguments' due to a half-initialized register. 16134 16135 * interpreter/CallFrame.h: 16136 (JSC::ExecState::setCalleeArguments): 16137 (JSC::ExecState::init): Require a full JSValue when setting up the 16138 'arguments' virtual register, since this register is accessible from JIT 16139 code and bytecode, and needs to be a true JSValue. 16140 16141 * interpreter/CallFrameClosure.h: 16142 (JSC::CallFrameClosure::resetCallFrame): ditto 16143 16144 * interpreter/Interpreter.cpp: 16145 (JSC::Interpreter::privateExecute): ditto 16146 16147 * interpreter/Register.h: Removed the constructor that allowed assignment 16148 of a JSArguments* to a register. That is not safe. See above. 16149 16150 * jit/JITStubs.cpp: 16151 (JSC::JITStubs::cti_op_create_arguments): 16152 (JSC::JITStubs::cti_op_create_arguments_no_params): ditto 16153 161542009-05-14 Sam Weinig <sam@webkit.org> 16155 16156 Reviewed by Geoffrey Garen. 16157 16158 We really want to go to the slow case in op_jfalse and 16159 op_jtrue if the value is *not* boolean. 16160 16161 * jit/JITOpcodes.cpp: 16162 (JSC::JIT::emit_op_jfalse): 16163 (JSC::JIT::emit_op_jtrue): 16164 161652009-05-14 Sam Weinig <sam@webkit.org> 16166 16167 Reviewed by Geoffrey Garen. 16168 16169 Flipped the condition when emitting a an op_loop_if_less or op_loop_if_lesseq 16170 if the first operand is a constant. 16171 16172 * jit/JITOpcodes.cpp: 16173 (JSC::JIT::emit_op_loop_if_less): 16174 (JSC::JIT::emit_op_loop_if_lesseq): 16175 161762009-05-14 Sam Weinig <sam@webkit.org> 16177 16178 Reviewed by Geoffrey Garen. 16179 16180 Added missing return in op_jnless and op_jnlesseq. 16181 16182 * jit/JITArithmetic.cpp: 16183 (JSC::JIT::emit_op_jnless): 16184 (JSC::JIT::emit_op_jnlesseq): 16185 161862009-05-14 Sam Weinig <sam@webkit.org> 16187 16188 Reviewed by Geoffrey Garen. 16189 16190 Load constants into the the register file as a temporary measure to 16191 aid bring up. This allows us to use to treat constants like any 16192 other virtual register. 16193 16194 * jit/JITOpcodes.cpp: 16195 (JSC::JIT::emit_op_enter): 16196 (JSC::JIT::emit_op_enter_with_activation): 16197 161982009-05-14 Geoffrey Garen <ggaren@apple.com> 16199 16200 Reviewed by Sam Weinig. 16201 16202 Implemented op_strict_eq. Original patch by Snowy, by way of Sam and Gavin. 16203 16204 * assembler/MacroAssemblerX86Common.h: 16205 (JSC::MacroAssemblerX86Common::set8): Added set8, since it's slightly 16206 faster than set32, and the new value representation usually doesn't 16207 need set32. 16208 16209 * jit/JIT.cpp: 16210 * jit/JIT.h: 16211 * jit/JITInlineMethods.h: 16212 (JSC::JIT::emitLoadTag): 16213 (JSC::JIT::emitLoadPayload): Added helper functions for dealing with 16214 constants. Eventually, we should write special cases for all constants, 16215 but these are helpful in the short term. 16216 16217 * jit/JITOpcodes.cpp: 16218 (JSC::JIT::compileOpStrictEq): 16219 (JSC::JIT::emitSlow_op_stricteq): 16220 (JSC::JIT::emitSlow_op_nstricteq): teh opcodez. 16221 16222 * runtime/JSValue.h: 16223 (JSC::JSValue::): 16224 (JSC::JSValue::isDouble): Added a LowestTag for clarity. 16225 162262009-05-13 Geoffrey Garen <ggaren@apple.com> 16227 16228 Reviewed by Sam Weinig. 16229 16230 Fixed some bugs in host function calls. 16231 16232 testapi now passes! 16233 16234 * jit/JIT.cpp: Changed some registers around to avoid overwriting edx:eax, 16235 which is how JSValues are now returned. Also changed the code that 16236 passes thisValue to pass the full 64bits of the value. Also added 16237 an #error compiler directive to other platform builds, since the JSValue 16238 return signature probably won't return in edx:eax on those platforms, 16239 and we'll have to investigate a solution. 16240 162412009-05-13 Geoffrey Garen <ggaren@apple.com> 16242 16243 Reviewed by Sam Weinig. 16244 16245 Removed parameters from functions that are intended never to use their 16246 parameters. 16247 16248 * jit/JITPropertyAccess.cpp: 16249 (JSC::JIT::emitSlow_op_get_by_val): 16250 (JSC::JIT::emitSlow_op_put_by_val): 16251 162522009-05-13 Geoffrey Garen <ggaren@apple.com> 16253 16254 Reviewed by Sam Weinig. 16255 16256 Ported op_instance_of from TOT. It's basically the same, but some register 16257 stuff changed to memory stuff. 16258 16259 * jit/JITInlineMethods.h: 16260 (JSC::JIT::emitPutJITStubArgFromVirtualRegister): 16261 (JSC::JIT::emitStore): Changed to use helper functions. 16262 16263 * jit/JITOpcodes.cpp: 16264 (JSC::JIT::emit_op_instanceof): 16265 (JSC::JIT::emitSlow_op_instanceof): Ported from TOT. 16266 162672009-05-13 Geoffrey Garen <ggaren@apple.com> 16268 16269 Reviewed by Gavin Barraclough. 16270 16271 Added a comment to explain an exception-handling subtelty that we found 16272 hard to remember when reviewing my last patch. 16273 16274 * jit/JITOpcodes.cpp: 16275 (JSC::JIT::emit_op_catch): 16276 162772009-05-13 Geoffrey Garen <ggaren@apple.com> 16278 16279 Reviewed by Sam Weinig. 16280 16281 Implemented try/catch. 16282 16283 * jit/JITOpcodes.cpp: 16284 (JSC::JIT::emit_op_throw): Updated to use JITStackFrame abstraction. 16285 (JSC::JIT::emit_op_catch): Filled out. 16286 162872009-05-13 Sam Weinig <sam@webkit.org> 16288 16289 Reviewed by Geoffrey Garen. 16290 16291 Implemented op_loop_if_true, op_jfalse, op_jtrue, op_jeq_null and op_jneq_null 16292 16293 * jit/JITOpcodes.cpp: 16294 (JSC::JIT::emitSlow_op_instanceof): Moved from below to be next to its 16295 fast brother. 16296 16297 (JSC::JIT::emit_op_loop_if_true): Similar to the old version 16298 in that it tries to do the integer case first and reduce the 16299 number of jumps you might need to take. 16300 (JSC::JIT::emitSlow_op_loop_if_true): 16301 16302 (JSC::JIT::emit_op_jfalse): Very similar to op_loop_if_true, only 16303 the inverse and without a timeout check. 16304 (JSC::JIT::emitSlow_op_jfalse): 16305 16306 (JSC::JIT::emit_op_jtrue): Very similar to op_loop_if_true except 16307 without the timeout check. 16308 (JSC::JIT::emitSlow_op_jtrue): 16309 16310 (JSC::JIT::emit_op_jeq_null): Very similar to the implementation 16311 of op_eq, except it takes jumps instead of copying the condition 16312 to a dst. 16313 (JSC::JIT::emit_op_jneq_null): Ditto but for op_neq. 16314 163152009-05-13 Geoffrey Garen <ggaren@apple.com> 16316 16317 Reviewed by Sam Weinig. 16318 16319 Implemented op_call_varargs. 16320 16321 * jit/JITCall.cpp: 16322 (JSC::JIT::compileOpCallVarargsSetupArgs): 16323 (JSC::JIT::compileOpCallVarargs): 16324 (JSC::JIT::emit_op_call): 16325 (JSC::JIT::emit_op_call_eval): 16326 (JSC::JIT::emit_op_load_varargs): 16327 (JSC::JIT::emit_op_call_varargs): 16328 (JSC::JIT::emit_op_construct): 16329 * jit/JITOpcodes.cpp: 16330 (JSC::JIT::emit_op_jneq_ptr): 16331 163322009-05-13 Geoffrey Garen <ggaren@apple.com> 16333 16334 Reviewed by Sam Weinig. 16335 16336 Implemented op_call_eval. 16337 16338 * jit/JITCall.cpp: 16339 (JSC::JIT::compileOpCallVarargsSetupArgs): 16340 (JSC::JIT::compileOpCall): 16341 * jit/JITStubCall.h: 16342 (JSC::CallEvalJITStub::CallEvalJITStub): 16343 163442009-05-13 Sam Weinig <sam@webkit.org> 16345 16346 Reviewed by Gavin Barraclough. 16347 16348 Implemented op_not. (Gavin did most of the work!) 16349 16350 * jit/JITOpcodes.cpp: 16351 (JSC::JIT::emit_op_not): 16352 (JSC::JIT::emitSlow_op_not): 16353 163542009-05-13 Geoffrey Garen <ggaren@apple.com> 16355 16356 Reviewed by Sam Weinig. 16357 16358 Implemented op_global_resolve. 16359 16360 * jit/JITOpcodes.cpp: 16361 (JSC::JIT::emit_op_loop_if_less): 16362 (JSC::JIT::emit_op_loop_if_lesseq): Added back accidentally removed 16363 early returns. 16364 16365 (JSC::JIT::emit_op_resolve_global): 16366 * jit/JITStubs.cpp: 16367 (JSC::JITStubs::cti_op_resolve_global): Pretty similar to the old code, 16368 but we need two reads and a TimesEight step in order to account for the 16369 64bit value size. 16370 16371 * jit/JITStubs.h: 16372 (JSC::): Slightly tweaked this code to specialize for a JSGlobalObject*, 16373 to avoid having to pass an irrelevant tag pointer to the stub. 16374 163752009-05-13 Sam Weinig <sam@webkit.org> 16376 16377 Reviewed by Geoffrey Garen. 16378 16379 Implemented op_to_jsnumber. 16380 16381 * jit/JITOpcodes.cpp: 16382 (JSC::JIT::emit_op_to_jsnumber): 16383 (JSC::JIT::emitSlow_op_to_jsnumber): 16384 163852009-05-13 Sam Weinig <sam@webkit.org> 16386 16387 Reviewed by Geoffrey Garen. 16388 16389 Implemented op_convert_this. 16390 16391 * jit/JITOpcodes.cpp: 16392 (JSC::JIT::emit_op_convert_this): 16393 (JSC::JIT::emitSlow_op_convert_this): 16394 163952009-05-13 Geoffrey Garen <ggaren@apple.com> 16396 16397 Reviewed by Sam Weinig. 16398 16399 Got basic JS function and constructor calls working. 16400 16401 * jit/JIT.cpp: 16402 (JSC::JIT::privateCompileCTIMachineTrampolines): 16403 * jit/JIT.h: 16404 * jit/JITCall.cpp: 16405 (JSC::JIT::compileOpCallSetupArgs): 16406 (JSC::JIT::compileOpCallVarargsSetupArgs): 16407 (JSC::JIT::compileOpConstructSetupArgs): 16408 (JSC::JIT::emit_op_ret): 16409 (JSC::JIT::emit_op_construct_verify): 16410 (JSC::JIT::emitSlow_op_construct_verify): 16411 (JSC::JIT::emitSlow_op_call): 16412 (JSC::JIT::emitSlow_op_call_eval): 16413 (JSC::JIT::emitSlow_op_call_varargs): 16414 (JSC::JIT::emitSlow_op_construct): 16415 (JSC::JIT::compileOpCall): Filled out these cases, with call_eval #if'd out. 16416 16417 * jit/JITInlineMethods.h: 16418 (JSC::JIT::emitPutJITStubArgFromVirtualRegister): 16419 (JSC::JIT::emitLoad): Restored some legacy "*CTIArg*" functions, 16420 since I wanted to avoid the complexity of revamping the API here while 16421 trying to bring it up. Eventually, we should re-remove all of these functions. 16422 16423 (JSC::JIT::recordJumpTarget): Removed unnecessary macro cruft. You will 16424 not silence me, Sam Weinig! The world will know that you are a crufty, 16425 crufty, crufty programmer!!! 16426 16427 * jit/JITOpcodes.cpp: 16428 * jit/JITStubs.cpp: 16429 (JSC::): 16430 * jit/JITStubs.h: Changed up some offsets in the JITStackFrame class, since 16431 and off-by-one error was causing stack misalignment. 16432 164332009-05-13 Sam Weinig <sam@webkit.org> 16434 16435 Reviewed by Geoffrey Garen. 16436 16437 Implement op_eq_null and op_neq_null. 16438 16439 * assembler/MacroAssemblerX86Common.h: 16440 (JSC::MacroAssemblerX86Common::set8): 16441 (JSC::MacroAssemblerX86Common::setTest8): 16442 * jit/JITOpcodes.cpp: 16443 (JSC::JIT::emit_op_stricteq): 16444 (JSC::JIT::emitSlow_op_stricteq): 16445 (JSC::JIT::emit_op_nstricteq): 16446 (JSC::JIT::emitSlow_op_nstricteq): 16447 (JSC::JIT::emit_op_eq_null): 16448 (JSC::JIT::emit_op_neq_null): 16449 * jsc.cpp: 16450 164512009-05-12 Sam Weinig <sam@webkit.org> 16452 16453 Reviewed by Geoffrey Garen. 16454 16455 Implement op_new_error. 16456 16457 * jit/JITOpcodes.cpp: 16458 (JSC::JIT::emit_op_new_error): 16459 * jit/JITStubCall.h: 16460 (JSC::JITStubCall::addArgument): Add a version of addArgument 16461 that takes a constant JSValue. 16462 164632009-05-12 Sam Weinig <sam@webkit.org> 16464 16465 Reviewed by Geoffrey Garen. 16466 16467 Remove now unused emitGetVariableObjectRegister and emitPutVariableObjectRegister. 16468 16469 * jit/JIT.cpp: 16470 * jit/JIT.h: 16471 164722009-05-12 Sam Weinig <sam@webkit.org> 16473 16474 Reviewed by Geoffrey Garen. 16475 16476 Implement op_to_primitive and op_next_pname. 16477 16478 * jit/JITOpcodes.cpp: 16479 (JSC::JIT::emitSlow_op_construct_verify): 16480 (JSC::JIT::emit_op_to_primitive): 16481 (JSC::JIT::emitSlow_op_to_primitive): 16482 (JSC::JIT::emitSlow_op_loop_if_true): 16483 (JSC::JIT::emit_op_jtrue): 16484 (JSC::JIT::emit_op_next_pname): 16485 164862009-05-12 Sam Weinig <sam@webkit.org> 16487 16488 Reviewed by Geoffrey Garen. 16489 16490 Add op_get_global_var, op_put_global_var, emit_op_get_scoped_var, emit_op_put_scoped_var and 16491 op_unexpected_load. 16492 16493 * jit/JIT.h: 16494 * jit/JITInlineMethods.h: 16495 (JSC::JIT::tagFor): 16496 (JSC::JIT::payloadFor): 16497 (JSC::JIT::emitLoad): 16498 (JSC::JIT::emitStore): 16499 (JSC::JIT::emitLoadReturnValue): 16500 * jit/JITOpcodes.cpp: 16501 (JSC::JIT::emit_op_get_global_var): 16502 (JSC::JIT::emit_op_put_global_var): 16503 (JSC::JIT::emit_op_get_scoped_var): 16504 (JSC::JIT::emit_op_put_scoped_var): 16505 (JSC::JIT::emit_op_unexpected_load): 16506 165072009-05-12 Geoffrey Garen <ggaren@apple.com> 16508 16509 Reviewed by Sam Weinig. 16510 16511 Added overflow handling to op_sub. 16512 16513 * jit/JIT.h: 16514 * jit/JITArithmetic.cpp: 16515 (JSC::JIT::emitSlow_op_sub): 16516 (JSC::JIT::emitSlowSub32InPlaceLeft): 16517 165182009-05-12 Sam Weinig <sam@webkit.org> 16519 16520 Reviewed by Geoffrey Garen. 16521 16522 Remove a function call by folding op_get_by_id and op_put_by_id into 16523 their respective compile functions. 16524 16525 * jit/JIT.h: 16526 * jit/JITPropertyAccess.cpp: 16527 (JSC::JIT::emit_op_get_by_id): 16528 (JSC::JIT::emitSlow_op_get_by_id): 16529 (JSC::JIT::emit_op_put_by_id): 16530 (JSC::JIT::emitSlow_op_put_by_id): 16531 165322009-05-12 Sam Weinig <sam@webkit.org> 16533 16534 Reviewed by Geoffrey Garen. 16535 16536 Make JITStubCall work in 64bit by making the stack index 16537 step dependent on the size of void*. 16538 16539 * jit/JITStubCall.h: 16540 (JSC::JITStubCall::JITStubCall): 16541 (JSC::JITStubCall::addArgument): 16542 165432009-05-12 Sam Weinig <sam@webkit.org> 16544 16545 Reviewed by Geoffrey Garen. 16546 16547 Implement simple version of property access opcodes 16548 which just call a stub functions. 16549 16550 * jit/JITOpcodes.cpp: 16551 * jit/JITPropertyAccess.cpp: 16552 (JSC::JIT::emitSlow_op_put_by_id): 16553 (JSC::JIT::emitSlow_op_get_by_id): 16554 (JSC::JIT::emit_op_get_by_val): 16555 (JSC::JIT::emitSlow_op_get_by_val): 16556 (JSC::JIT::emit_op_put_by_val): 16557 (JSC::JIT::emitSlow_op_put_by_val): 16558 (JSC::JIT::emit_op_put_by_index): 16559 (JSC::JIT::emit_op_put_getter): 16560 (JSC::JIT::emit_op_put_setter): 16561 (JSC::JIT::emit_op_del_by_id): 16562 (JSC::JIT::compileGetByIdHotPath): 16563 (JSC::JIT::compilePutByIdHotPath): 16564 * jit/JITStubCall.h: 16565 (JSC::JITStubCall::addArgument): 16566 * jsc.cpp: 16567 165682009-05-12 Geoffrey Garen <ggaren@apple.com> 16569 16570 Reviewed by Sam Weinig. 16571 16572 Added work-around for XCode debugging echo problem. 16573 16574 * jsc.cpp: 16575 (runInteractive): 16576 165772009-05-12 Geoffrey Garen <ggaren@apple.com> 16578 16579 Reviewed by Sam Weinig. 16580 16581 Added overflow handling to op_add. 16582 16583 * jit/JIT.h: 16584 * jit/JITArithmetic.cpp: 16585 (JSC::JIT::emitSlow_op_add): 16586 (JSC::JIT::emitSlowAdd32InPlace): 16587 165882009-05-12 Sam Weinig <sam@webkit.org> 16589 16590 Reviewed by Geoffrey Garen. 16591 16592 Add slow cases for op_jnless or emit_op_jnlesseq. 16593 16594 * jit/JITArithmetic.cpp: 16595 (JSC::JIT::emitSlow_op_jnless): 16596 (JSC::JIT::emitSlow_op_jnlesseq): 16597 165982009-05-12 Sam Weinig <sam@webkit.org> 16599 16600 Reviewed by Geoffrey Garen. 16601 16602 Add implementations for op_jnless, emit_op_jnlesseq, op_loop_if_less and op_loop_if_lesseq. 16603 No slow cases for op_jnless or emit_op_jnlesseq yet. 16604 16605 * jit/JITArithmetic.cpp: 16606 (JSC::JIT::emit_op_jnless): 16607 (JSC::JIT::emitSlow_op_jnless): 16608 (JSC::JIT::emit_op_jnlesseq): 16609 (JSC::JIT::emitSlow_op_jnlesseq): 16610 * jit/JITOpcodes.cpp: 16611 (JSC::JIT::emit_op_loop_if_less): 16612 (JSC::JIT::emitSlow_op_loop_if_less): 16613 (JSC::JIT::emit_op_loop_if_lesseq): 16614 (JSC::JIT::emitSlow_op_loop_if_lesseq): 16615 166162009-05-12 Sam Weinig <sam@webkit.org> 16617 16618 Reviewed by Geoffrey Garen. 16619 16620 Turn the RECORD_JUMP_TARGET macro into an inline function. 16621 16622 * jit/JIT.h: 16623 * jit/JITInlineMethods.h: 16624 (JSC::JIT::recordJumpTarget): 16625 * jit/JITOpcodes.cpp: 16626 (JSC::JIT::emit_op_jmp): 16627 (JSC::JIT::emit_op_jsr): 16628 (JSC::JIT::emit_op_jmp_scopes): 16629 166302009-05-12 Sam Weinig <sam@webkit.org> 16631 16632 Add MacroAssemblerX86Common::set8 to fix the build. 16633 16634 * assembler/MacroAssemblerX86Common.h: 16635 (JSC::MacroAssemblerX86Common::set8): 16636 166372009-05-12 Geoffrey Garen <ggaren@apple.com> 16638 16639 Reviewed by Sam Weinig. 16640 16641 Added overflow recovery for pre_inc and pre_dec. 16642 16643 Turned some short-circuit code into early returns, as is the WebKit style. 16644 16645 * jit/JITArithmetic.cpp: 16646 (JSC::JIT::emit_op_post_inc): 16647 (JSC::JIT::emitSlow_op_post_inc): 16648 (JSC::JIT::emit_op_post_dec): 16649 (JSC::JIT::emitSlow_op_post_dec): 16650 (JSC::JIT::emitSlow_op_pre_inc): 16651 (JSC::JIT::emitSlow_op_pre_dec): 16652 166532009-05-12 Sam Weinig <sam@webkit.org> 16654 16655 Reviewed by Geoffrey Garen. 16656 16657 Implement op_jmp, op_loop, op_eq and op_neq. 16658 16659 * jit/JITOpcodes.cpp: 16660 (JSC::JIT::emit_op_jmp): 16661 (JSC::JIT::emit_op_loop): 16662 (JSC::JIT::emit_op_eq): 16663 (JSC::JIT::emitSlow_op_eq): 16664 (JSC::JIT::emit_op_neq): 16665 (JSC::JIT::emitSlow_op_neq): 16666 (JSC::JIT::emit_op_enter): 16667 (JSC::JIT::emit_op_enter_with_activation): 16668 166692009-05-12 Sam Weinig <sam@webkit.org> 16670 16671 Reviewed by Geoffrey Garen. 16672 16673 Implement the slow cases for arithmetic opcodes. 16674 16675 * jit/JITArithmetic.cpp: 16676 (JSC::JIT::emitSlow_op_lshift): 16677 (JSC::JIT::emitSlow_op_rshift): 16678 (JSC::JIT::emitSlow_op_bitand): 16679 (JSC::JIT::emitSlow_op_bitor): 16680 (JSC::JIT::emitSlow_op_bitxor): 16681 (JSC::JIT::emitSlow_op_bitnot): 16682 (JSC::JIT::emitSlow_op_sub): 16683 (JSC::JIT::emitSlow_op_mul): 16684 (JSC::JIT::emitSlow_op_mod): 16685 (JSC::JIT::emit_op_mod): 16686 166872009-05-12 Sam Weinig <sam@webkit.org> 16688 16689 Reviewed by Geoffrey Garen. 16690 16691 Implement op_bitnot. 16692 16693 * assembler/MacroAssemblerX86Common.h: 16694 (JSC::MacroAssemblerX86Common::not32): 16695 * assembler/X86Assembler.h: 16696 (JSC::X86Assembler::notl_m): 16697 * jit/JITArithmetic.cpp: 16698 (JSC::JIT::emit_op_bitnot): 16699 167002009-05-12 Sam Weinig <sam@webkit.org> 16701 16702 Reviewed by Geoffrey Garen. 16703 16704 Add arithmetic opcode implementations from the old nitro-extreme branch. 16705 16706 * jit/JIT.h: 16707 * jit/JITArithmetic.cpp: 16708 (JSC::JIT::emit_op_jnless): 16709 (JSC::JIT::emitSlow_op_jnless): 16710 (JSC::JIT::emit_op_jnlesseq): 16711 (JSC::JIT::emitSlow_op_jnlesseq): 16712 (JSC::JIT::emit_op_lshift): 16713 (JSC::JIT::emitSlow_op_lshift): 16714 (JSC::JIT::emit_op_rshift): 16715 (JSC::JIT::emitSlow_op_rshift): 16716 (JSC::JIT::emit_op_bitand): 16717 (JSC::JIT::emitBitAnd32Constant): 16718 (JSC::JIT::emitBitAnd32InPlace): 16719 (JSC::JIT::emit_op_bitor): 16720 (JSC::JIT::emitSlow_op_bitor): 16721 (JSC::JIT::emitBitOr32Constant): 16722 (JSC::JIT::emitBitOr32InPlace): 16723 (JSC::JIT::emit_op_bitxor): 16724 (JSC::JIT::emitSlow_op_bitxor): 16725 (JSC::JIT::emitBitXor32Constant): 16726 (JSC::JIT::emitBitXor32InPlace): 16727 (JSC::JIT::emit_op_bitnot): 16728 (JSC::JIT::emitSlow_op_bitnot): 16729 (JSC::JIT::emit_op_post_inc): 16730 (JSC::JIT::emitSlow_op_post_inc): 16731 (JSC::JIT::emit_op_post_dec): 16732 (JSC::JIT::emitSlow_op_post_dec): 16733 (JSC::JIT::emit_op_pre_inc): 16734 (JSC::JIT::emitSlow_op_pre_inc): 16735 (JSC::JIT::emit_op_pre_dec): 16736 (JSC::JIT::emitSlow_op_pre_dec): 16737 (JSC::JIT::emit_op_add): 16738 (JSC::JIT::emitAdd32Constant): 16739 (JSC::JIT::emitAdd32InPlace): 16740 (JSC::JIT::emitSlow_op_add): 16741 (JSC::JIT::emit_op_sub): 16742 (JSC::JIT::emitSlow_op_sub): 16743 (JSC::JIT::emitSub32ConstantLeft): 16744 (JSC::JIT::emitSub32ConstantRight): 16745 (JSC::JIT::emitSub32InPlaceLeft): 16746 (JSC::JIT::emitSub32InPlaceRight): 16747 (JSC::JIT::emit_op_mul): 16748 (JSC::JIT::emitSlow_op_mul): 16749 (JSC::JIT::emitMul32Constant): 16750 (JSC::JIT::emitMul32InPlace): 16751 (JSC::JIT::emit_op_mod): 16752 (JSC::JIT::emitSlow_op_mod): 16753 * jit/JITOpcodes.cpp: 16754 167552009-05-12 Geoffrey Garen <ggaren@apple.com> 16756 16757 Removed JIT_OPTIMIZE_ARITHMETIC setting, since it was all about 32bit 16758 value representations. 16759 16760 Added JSAPIValueWrapper to the repository. 16761 16762 * jit/JIT.h: 16763 * jit/JITArithmetic.cpp: 16764 * runtime/JSAPIValueWrapper.cpp: Added. 16765 (JSC::JSAPIValueWrapper::toPrimitive): 16766 (JSC::JSAPIValueWrapper::getPrimitiveNumber): 16767 (JSC::JSAPIValueWrapper::toBoolean): 16768 (JSC::JSAPIValueWrapper::toNumber): 16769 (JSC::JSAPIValueWrapper::toString): 16770 (JSC::JSAPIValueWrapper::toObject): 16771 * runtime/JSAPIValueWrapper.h: Added. 16772 (JSC::JSAPIValueWrapper::value): 16773 (JSC::JSAPIValueWrapper::isAPIValueWrapper): 16774 (JSC::JSAPIValueWrapper::JSAPIValueWrapper): 16775 (JSC::jsAPIValueWrapper): 16776 * wtf/Platform.h: 16777 167782009-05-12 Geoffrey Garen <ggaren@apple.com> 16779 16780 Turned on the JIT and got it building and running the most trivial of 16781 programs. 16782 16783 All configurable optimizations are turned off, and a few opcodes are ad 16784 hoc #if'd out. 16785 16786 So far, I've only merged op_mov and op_end, but some stub-reliant 16787 opcodes work as-is from TOT. 16788 16789 * bytecode/CodeBlock.cpp: 16790 (JSC::CodeBlock::~CodeBlock): 16791 * bytecode/CodeBlock.h: 16792 * jit/JIT.cpp: 16793 (JSC::JIT::compileOpStrictEq): 16794 * jit/JIT.h: 16795 * jit/JITArithmetic.cpp: 16796 (JSC::JIT::emit_op_lshift): 16797 (JSC::JIT::emitSlow_op_lshift): 16798 (JSC::JIT::emit_op_rshift): 16799 (JSC::JIT::emitSlow_op_rshift): 16800 (JSC::JIT::emit_op_jnless): 16801 (JSC::JIT::emitSlow_op_jnless): 16802 (JSC::JIT::emit_op_jnlesseq): 16803 (JSC::JIT::emitSlow_op_jnlesseq): 16804 (JSC::JIT::emit_op_bitand): 16805 (JSC::JIT::emitSlow_op_bitand): 16806 (JSC::JIT::emit_op_post_inc): 16807 (JSC::JIT::emitSlow_op_post_inc): 16808 (JSC::JIT::emit_op_post_dec): 16809 (JSC::JIT::emitSlow_op_post_dec): 16810 (JSC::JIT::emit_op_pre_inc): 16811 (JSC::JIT::emitSlow_op_pre_inc): 16812 (JSC::JIT::emit_op_pre_dec): 16813 (JSC::JIT::emitSlow_op_pre_dec): 16814 (JSC::JIT::emit_op_mod): 16815 (JSC::JIT::emitSlow_op_mod): 16816 (JSC::JIT::emit_op_add): 16817 (JSC::JIT::emit_op_mul): 16818 (JSC::JIT::emit_op_sub): 16819 (JSC::JIT::compileBinaryArithOpSlowCase): 16820 (JSC::JIT::emitSlow_op_add): 16821 (JSC::JIT::emitSlow_op_mul): 16822 * jit/JITCall.cpp: 16823 (JSC::JIT::compileOpCallInitializeCallFrame): 16824 (JSC::JIT::compileOpConstructSetupArgs): 16825 (JSC::JIT::compileOpCallVarargs): 16826 (JSC::JIT::compileOpCall): 16827 (JSC::JIT::compileOpCallSlowCase): 16828 * jit/JITInlineMethods.h: 16829 (JSC::JIT::getConstantOperandImmediateInt): 16830 (JSC::JIT::isOperandConstantImmediateInt): 16831 (JSC::JIT::emitInitRegister): 16832 (JSC::JIT::addSlowCase): 16833 (JSC::JIT::addJump): 16834 (JSC::JIT::emitJumpSlowToHot): 16835 (JSC::JIT::tagFor): 16836 (JSC::JIT::payloadFor): 16837 (JSC::JIT::emitLoad): 16838 (JSC::JIT::emitLoadReturnValue): 16839 (JSC::JIT::emitStore): 16840 (JSC::JIT::emitStoreReturnValue): 16841 * jit/JITOpcodes.cpp: 16842 (JSC::JIT::emit_op_mov): 16843 (JSC::JIT::emit_op_end): 16844 (JSC::JIT::emit_op_jmp): 16845 (JSC::JIT::emit_op_loop): 16846 (JSC::JIT::emit_op_loop_if_less): 16847 (JSC::JIT::emit_op_loop_if_lesseq): 16848 (JSC::JIT::emit_op_instanceof): 16849 (JSC::JIT::emit_op_get_global_var): 16850 (JSC::JIT::emit_op_put_global_var): 16851 (JSC::JIT::emit_op_get_scoped_var): 16852 (JSC::JIT::emit_op_put_scoped_var): 16853 (JSC::JIT::emit_op_tear_off_activation): 16854 (JSC::JIT::emit_op_ret): 16855 (JSC::JIT::emit_op_construct_verify): 16856 (JSC::JIT::emit_op_to_primitive): 16857 (JSC::JIT::emit_op_loop_if_true): 16858 (JSC::JIT::emit_op_resolve_global): 16859 (JSC::JIT::emit_op_not): 16860 (JSC::JIT::emit_op_jfalse): 16861 (JSC::JIT::emit_op_jeq_null): 16862 (JSC::JIT::emit_op_jneq_null): 16863 (JSC::JIT::emit_op_jneq_ptr): 16864 (JSC::JIT::emit_op_unexpected_load): 16865 (JSC::JIT::emit_op_eq): 16866 (JSC::JIT::emit_op_bitnot): 16867 (JSC::JIT::emit_op_jtrue): 16868 (JSC::JIT::emit_op_neq): 16869 (JSC::JIT::emit_op_bitxor): 16870 (JSC::JIT::emit_op_bitor): 16871 (JSC::JIT::emit_op_throw): 16872 (JSC::JIT::emit_op_next_pname): 16873 (JSC::JIT::emit_op_push_scope): 16874 (JSC::JIT::emit_op_to_jsnumber): 16875 (JSC::JIT::emit_op_push_new_scope): 16876 (JSC::JIT::emit_op_catch): 16877 (JSC::JIT::emit_op_switch_imm): 16878 (JSC::JIT::emit_op_switch_char): 16879 (JSC::JIT::emit_op_switch_string): 16880 (JSC::JIT::emit_op_new_error): 16881 (JSC::JIT::emit_op_eq_null): 16882 (JSC::JIT::emit_op_neq_null): 16883 (JSC::JIT::emit_op_convert_this): 16884 (JSC::JIT::emit_op_profile_will_call): 16885 (JSC::JIT::emit_op_profile_did_call): 16886 (JSC::JIT::emitSlow_op_construct_verify): 16887 (JSC::JIT::emitSlow_op_get_by_val): 16888 (JSC::JIT::emitSlow_op_loop_if_less): 16889 (JSC::JIT::emitSlow_op_loop_if_lesseq): 16890 (JSC::JIT::emitSlow_op_put_by_val): 16891 (JSC::JIT::emitSlow_op_not): 16892 (JSC::JIT::emitSlow_op_instanceof): 16893 * jit/JITPropertyAccess.cpp: 16894 (JSC::JIT::emit_op_get_by_val): 16895 (JSC::JIT::emit_op_put_by_val): 16896 (JSC::JIT::emit_op_put_by_index): 16897 (JSC::JIT::emit_op_put_getter): 16898 (JSC::JIT::emit_op_put_setter): 16899 (JSC::JIT::emit_op_del_by_id): 16900 (JSC::JIT::compileGetByIdHotPath): 16901 (JSC::JIT::compilePutByIdHotPath): 16902 * jit/JITStubCall.h: 16903 (JSC::JITStubCall::JITStubCall): 16904 (JSC::JITStubCall::addArgument): 16905 (JSC::JITStubCall::call): 16906 (JSC::JITStubCall::): 16907 (JSC::CallEvalJITStub::CallEvalJITStub): 16908 * jit/JITStubs.cpp: 16909 (JSC::): 16910 (JSC::JITStubs::cti_op_add): 16911 (JSC::JITStubs::cti_op_pre_inc): 16912 (JSC::JITStubs::cti_op_mul): 16913 (JSC::JITStubs::cti_op_get_by_val): 16914 (JSC::JITStubs::cti_op_get_by_val_string): 16915 (JSC::JITStubs::cti_op_get_by_val_byte_array): 16916 (JSC::JITStubs::cti_op_sub): 16917 (JSC::JITStubs::cti_op_put_by_val): 16918 (JSC::JITStubs::cti_op_put_by_val_array): 16919 (JSC::JITStubs::cti_op_put_by_val_byte_array): 16920 (JSC::JITStubs::cti_op_negate): 16921 (JSC::JITStubs::cti_op_div): 16922 (JSC::JITStubs::cti_op_pre_dec): 16923 (JSC::JITStubs::cti_op_post_inc): 16924 (JSC::JITStubs::cti_op_eq): 16925 (JSC::JITStubs::cti_op_lshift): 16926 (JSC::JITStubs::cti_op_bitand): 16927 (JSC::JITStubs::cti_op_rshift): 16928 (JSC::JITStubs::cti_op_bitnot): 16929 (JSC::JITStubs::cti_op_mod): 16930 (JSC::JITStubs::cti_op_neq): 16931 (JSC::JITStubs::cti_op_post_dec): 16932 (JSC::JITStubs::cti_op_urshift): 16933 (JSC::JITStubs::cti_op_bitxor): 16934 (JSC::JITStubs::cti_op_bitor): 16935 (JSC::JITStubs::cti_op_switch_imm): 16936 * jit/JITStubs.h: 16937 * runtime/JSArray.cpp: 16938 (JSC::JSArray::JSArray): 16939 * runtime/JSFunction.cpp: 16940 (JSC::JSFunction::~JSFunction): 16941 * runtime/JSValue.h: 16942 (JSC::JSValue::payload): 16943 * wtf/Platform.h: 16944 169452009-05-07 Sam Weinig <sam@webkit.org> 16946 16947 Reviewed by Geoffrey Garen. 16948 16949 Add some new MacroAssembler and assembler functions that will be needed shortly. 16950 16951 * assembler/MacroAssemblerX86Common.h: 16952 (JSC::MacroAssemblerX86Common::add32): 16953 (JSC::MacroAssemblerX86Common::and32): 16954 (JSC::MacroAssemblerX86Common::mul32): 16955 (JSC::MacroAssemblerX86Common::neg32): 16956 (JSC::MacroAssemblerX86Common::or32): 16957 (JSC::MacroAssemblerX86Common::sub32): 16958 (JSC::MacroAssemblerX86Common::xor32): 16959 (JSC::MacroAssemblerX86Common::branchAdd32): 16960 (JSC::MacroAssemblerX86Common::branchMul32): 16961 (JSC::MacroAssemblerX86Common::branchSub32): 16962 * assembler/X86Assembler.h: 16963 (JSC::X86Assembler::): 16964 (JSC::X86Assembler::addl_rm): 16965 (JSC::X86Assembler::andl_mr): 16966 (JSC::X86Assembler::andl_rm): 16967 (JSC::X86Assembler::andl_im): 16968 (JSC::X86Assembler::negl_r): 16969 (JSC::X86Assembler::notl_r): 16970 (JSC::X86Assembler::orl_rm): 16971 (JSC::X86Assembler::orl_im): 16972 (JSC::X86Assembler::subl_rm): 16973 (JSC::X86Assembler::xorl_mr): 16974 (JSC::X86Assembler::xorl_rm): 16975 (JSC::X86Assembler::xorl_im): 16976 (JSC::X86Assembler::imull_mr): 16977 169782009-05-11 Sam Weinig <sam@webkit.org> 16979 16980 Reviewed by Cameron Zwarich. 16981 16982 Remove the NumberHeap. 16983 16984 * JavaScriptCore.exp: 16985 * runtime/Collector.cpp: 16986 (JSC::Heap::Heap): 16987 (JSC::Heap::destroy): 16988 (JSC::Heap::recordExtraCost): 16989 (JSC::Heap::heapAllocate): 16990 (JSC::Heap::markConservatively): 16991 (JSC::Heap::sweep): 16992 (JSC::Heap::collect): 16993 (JSC::Heap::objectCount): 16994 (JSC::Heap::statistics): 16995 (JSC::typeName): 16996 (JSC::Heap::isBusy): 16997 * runtime/Collector.h: 16998 (JSC::Heap::globalData): 16999 * runtime/JSCell.h: 17000 170012009-05-11 Geoffrey Garen <ggaren@apple.com> 17002 17003 Reviewed by Sam Weinig. 17004 17005 Land initial commit of new number representation for 32 bit platforms, 17006 with JIT disabled. 17007 17008 * API/APICast.h: 17009 (toJS): 17010 (toRef): 17011 * API/JSCallbackObjectFunctions.h: 17012 (JSC::::hasInstance): 17013 (JSC::::toNumber): 17014 (JSC::::toString): 17015 * API/tests/testapi.c: 17016 (EvilExceptionObject_convertToType): 17017 * AllInOneFile.cpp: 17018 * JavaScriptCore.exp: 17019 * JavaScriptCore.xcodeproj/project.pbxproj: 17020 * bytecode/CodeBlock.cpp: 17021 (JSC::valueToSourceString): 17022 * bytecompiler/BytecodeGenerator.cpp: 17023 (JSC::BytecodeGenerator::emitLoad): 17024 (JSC::BytecodeGenerator::emitUnexpectedLoad): 17025 (JSC::keyForImmediateSwitch): 17026 * bytecompiler/BytecodeGenerator.h: 17027 * interpreter/Interpreter.cpp: 17028 (JSC::Interpreter::dumpRegisters): 17029 (JSC::Interpreter::privateExecute): 17030 * parser/Nodes.cpp: 17031 (JSC::ArrayNode::emitBytecode): 17032 (JSC::processClauseList): 17033 * runtime/ArgList.h: 17034 * runtime/Collector.h: 17035 (JSC::sizeof): 17036 * runtime/DateMath.cpp: 17037 * runtime/ExceptionHelpers.h: 17038 * runtime/InitializeThreading.cpp: 17039 * runtime/JSArray.cpp: 17040 (JSC::JSArray::JSArray): 17041 * runtime/JSCell.cpp: 17042 * runtime/JSCell.h: 17043 (JSC::JSCell::isAPIValueWrapper): 17044 (JSC::JSValue::isString): 17045 (JSC::JSValue::isGetterSetter): 17046 (JSC::JSValue::isObject): 17047 (JSC::JSValue::getString): 17048 (JSC::JSValue::getObject): 17049 (JSC::JSValue::getCallData): 17050 (JSC::JSValue::getConstructData): 17051 (JSC::JSValue::getUInt32): 17052 (JSC::JSValue::marked): 17053 (JSC::JSValue::toPrimitive): 17054 (JSC::JSValue::getPrimitiveNumber): 17055 (JSC::JSValue::toBoolean): 17056 (JSC::JSValue::toNumber): 17057 (JSC::JSValue::toString): 17058 (JSC::JSValue::needsThisConversion): 17059 (JSC::JSValue::toThisString): 17060 (JSC::JSValue::getJSNumber): 17061 (JSC::JSValue::toObject): 17062 (JSC::JSValue::toThisObject): 17063 * runtime/JSGlobalData.cpp: 17064 (JSC::JSGlobalData::JSGlobalData): 17065 * runtime/JSGlobalData.h: 17066 * runtime/JSGlobalObject.h: 17067 (JSC::Structure::prototypeForLookup): 17068 * runtime/JSGlobalObjectFunctions.cpp: 17069 (JSC::globalFuncParseInt): 17070 * runtime/JSImmediate.h: 17071 * runtime/JSNumberCell.cpp: Removed. 17072 * runtime/JSNumberCell.h: Removed. 17073 * runtime/JSObject.h: 17074 (JSC::JSValue::get): 17075 (JSC::JSValue::put): 17076 * runtime/JSString.h: 17077 (JSC::JSValue::toThisJSString): 17078 * runtime/JSValue.cpp: 17079 (JSC::JSValue::toInteger): 17080 (JSC::JSValue::toIntegerPreserveNaN): 17081 (JSC::JSValue::toObjectSlowCase): 17082 (JSC::JSValue::toThisObjectSlowCase): 17083 (JSC::JSValue::synthesizeObject): 17084 (JSC::JSValue::synthesizePrototype): 17085 (JSC::JSValue::description): 17086 (JSC::nonInlineNaN): 17087 * runtime/JSValue.h: 17088 (JSC::JSValue::): 17089 (JSC::EncodedJSValueHashTraits::emptyValue): 17090 (JSC::jsNaN): 17091 (JSC::operator==): 17092 (JSC::operator!=): 17093 (JSC::toInt32): 17094 (JSC::toUInt32): 17095 (JSC::JSValue::encode): 17096 (JSC::JSValue::decode): 17097 (JSC::JSValue::JSValue): 17098 (JSC::JSValue::operator bool): 17099 (JSC::JSValue::operator==): 17100 (JSC::JSValue::operator!=): 17101 (JSC::JSValue::isUndefined): 17102 (JSC::JSValue::isNull): 17103 (JSC::JSValue::isUndefinedOrNull): 17104 (JSC::JSValue::isCell): 17105 (JSC::JSValue::isInt32): 17106 (JSC::JSValue::isUInt32): 17107 (JSC::JSValue::isDouble): 17108 (JSC::JSValue::isTrue): 17109 (JSC::JSValue::isFalse): 17110 (JSC::JSValue::tag): 17111 (JSC::JSValue::asInt32): 17112 (JSC::JSValue::asUInt32): 17113 (JSC::JSValue::asDouble): 17114 (JSC::JSValue::asCell): 17115 (JSC::JSValue::isNumber): 17116 (JSC::JSValue::isBoolean): 17117 (JSC::JSValue::getBoolean): 17118 (JSC::JSValue::uncheckedGetNumber): 17119 (JSC::JSValue::toJSNumber): 17120 (JSC::JSValue::getNumber): 17121 (JSC::JSValue::toInt32): 17122 (JSC::JSValue::toUInt32): 17123 * runtime/Operations.h: 17124 (JSC::JSValue::equal): 17125 (JSC::JSValue::equalSlowCaseInline): 17126 (JSC::JSValue::strictEqual): 17127 (JSC::JSValue::strictEqualSlowCaseInline): 17128 (JSC::jsLess): 17129 (JSC::jsLessEq): 17130 (JSC::jsAdd): 17131 * runtime/PropertySlot.h: 17132 * runtime/StringPrototype.cpp: 17133 (JSC::stringProtoFuncCharAt): 17134 (JSC::stringProtoFuncCharCodeAt): 17135 (JSC::stringProtoFuncIndexOf): 17136 * wtf/Platform.h: 17137 17138=== Start merge of nitro-extreme branch 2009-07-30 === 17139 171402009-07-29 Laszlo Gombos <laszlo.1.gombos@nokia.com> 17141 17142 Reviewed by George Staikos. 17143 17144 Resolve class/struct mixup in forward declarations 17145 https://bugs.webkit.org/show_bug.cgi?id=27708 17146 17147 * API/JSClassRef.h: 17148 * bytecode/SamplingTool.h: 17149 * interpreter/Interpreter.h: 17150 * jit/JIT.h: 17151 * profiler/ProfileGenerator.h: 17152 * profiler/Profiler.h: 17153 * runtime/ClassInfo.h: 17154 * runtime/ExceptionHelpers.h: 17155 * runtime/JSByteArray.h: 17156 * runtime/JSCell.h: 17157 * runtime/JSFunction.h: 17158 * runtime/JSGlobalData.h: 17159 * runtime/JSObject.h: 17160 * runtime/JSString.h: 17161 171622009-07-28 Ada Chan <adachan@apple.com> 17163 17164 Reviewed by Darin Adler. 17165 17166 https://bugs.webkit.org/show_bug.cgi?id=27236 17167 - Implement TCMalloc_SystemRelease and TCMalloc_SystemCommit for Windows. 17168 - Use a background thread to periodically scavenge memory to release back to the system. 17169 17170 * wtf/FastMalloc.cpp: 17171 (WTF::TCMalloc_PageHeap::init): 17172 (WTF::TCMalloc_PageHeap::runScavengerThread): 17173 (WTF::TCMalloc_PageHeap::scavenge): 17174 (WTF::TCMalloc_PageHeap::shouldContinueScavenging): 17175 (WTF::TCMalloc_PageHeap::New): 17176 (WTF::TCMalloc_PageHeap::AllocLarge): 17177 (WTF::TCMalloc_PageHeap::Delete): 17178 (WTF::TCMalloc_PageHeap::GrowHeap): 17179 (WTF::sleep): 17180 (WTF::TCMalloc_PageHeap::scavengerThread): 17181 * wtf/TCSystemAlloc.cpp: 17182 (TCMalloc_SystemRelease): 17183 (TCMalloc_SystemCommit): 17184 * wtf/TCSystemAlloc.h: 17185 171862009-07-28 Xan Lopez <xlopez@igalia.com> 17187 17188 Add new files, fixes distcheck. 17189 17190 * GNUmakefile.am: 17191 171922009-07-28 Csaba Osztrogonac <oszi@inf.u-szeged.hu> 17193 17194 Reviewed by Simon Hausmann. 17195 17196 [Qt] Determining whether to use JIT or interpreter 17197 moved from JavaScriptCore.pri to Platform.h 17198 17199 * JavaScriptCore.pri: 17200 * wtf/Platform.h: 17201 172022009-07-27 Brian Weinstein <bweinstein@apple.com> 17203 17204 Fix of misuse of sort command. 17205 17206 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 17207 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 17208 172092009-07-27 Brian Weinstein <bweinstein@apple.com> 17210 17211 Build fix for Windows. 17212 17213 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 17214 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 17215 172162009-07-27 Gavin Barraclough <barraclough@apple.com> 17217 17218 Rubber stamped by Oliver Hunt. 17219 17220 Fix tyop in JIT, renamed preverveReturnAddressAfterCall -> preserveReturnAddressAfterCall. 17221 17222 * jit/JIT.cpp: 17223 (JSC::JIT::privateCompile): 17224 (JSC::JIT::privateCompileCTIMachineTrampolines): 17225 * jit/JIT.h: 17226 * jit/JITInlineMethods.h: 17227 (JSC::JIT::preserveReturnAddressAfterCall): 17228 * jit/JITPropertyAccess.cpp: 17229 (JSC::JIT::privateCompilePutByIdTransition): 17230 172312009-07-27 Alexey Proskuryakov <ap@webkit.org> 17232 17233 Gtk build fix. 17234 17235 * runtime/JSLock.cpp: (JSC::JSLock::JSLock): Fix "no threading" case. 17236 172372009-07-27 Alexey Proskuryakov <ap@webkit.org> 17238 17239 Release build fix. 17240 17241 * runtime/JSLock.h: (JSC::JSLock::~JSLock): 17242 172432009-07-27 Alexey Proskuryakov <ap@webkit.org> 17244 17245 Reviewed by Darin Adler. 17246 17247 https://bugs.webkit.org/show_bug.cgi?id=27735 17248 Give a helpful name to JSLock constructor argument 17249 17250 * API/JSBase.cpp: 17251 (JSGarbageCollect): 17252 * API/JSContextRef.cpp: 17253 * API/JSObjectRef.cpp: 17254 (JSPropertyNameArrayRelease): 17255 (JSPropertyNameAccumulatorAddName): 17256 * JavaScriptCore.exp: 17257 * jsc.cpp: 17258 (functionGC): 17259 (cleanupGlobalData): 17260 (jscmain): 17261 * runtime/Collector.cpp: 17262 (JSC::Heap::destroy): 17263 * runtime/JSLock.cpp: 17264 (JSC::JSLock::JSLock): 17265 (JSC::JSLock::lock): 17266 (JSC::JSLock::unlock): 17267 (JSC::JSLock::DropAllLocks::DropAllLocks): 17268 (JSC::JSLock::DropAllLocks::~DropAllLocks): 17269 * runtime/JSLock.h: 17270 (JSC::): 17271 (JSC::JSLock::JSLock): 17272 (JSC::JSLock::~JSLock): 17273 172742009-07-25 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 17275 17276 Reviewed by Eric Seidel. 17277 17278 Allow custom memory allocation control for OpaqueJSPropertyNameArray struct 17279 https://bugs.webkit.org/show_bug.cgi?id=27342 17280 17281 Inherits OpaqueJSPropertyNameArray struct from FastAllocBase because it has been 17282 instantiated by 'new' JavaScriptCore/API/JSObjectRef.cpp:473. 17283 17284 * API/JSObjectRef.cpp: 17285 172862009-07-24 Ada Chan <adachan@apple.com> 17287 17288 In preparation for https://bugs.webkit.org/show_bug.cgi?id=27236: 17289 Remove TCMALLOC_TRACK_DECOMMITED_SPANS. We'll always track decommitted spans. 17290 We have tested this and show it has little impact on performance. 17291 17292 Reviewed by Mark Rowe. 17293 17294 * wtf/FastMalloc.cpp: 17295 (WTF::TCMalloc_PageHeap::New): 17296 (WTF::TCMalloc_PageHeap::AllocLarge): 17297 (WTF::propagateDecommittedState): 17298 (WTF::mergeDecommittedStates): 17299 (WTF::TCMalloc_PageHeap::Delete): 17300 (WTF::TCMalloc_PageHeap::IncrementalScavenge): 17301 173022009-07-24 Csaba Osztrogonac <oszi@inf.u-szeged.hu> 17303 17304 Reviewed by Darin Adler and Adam Barth. 17305 17306 Build fix for x86 platforms. 17307 https://bugs.webkit.org/show_bug.cgi?id=27602 17308 17309 * jit/JIT.cpp: 17310 173112009-07-23 Kevin Ollivier <kevino@theolliviers.com> 17312 17313 wx build fix, adding missing header. 17314 17315 * jit/JIT.cpp: 17316 173172009-07-22 Yong Li <yong.li@torchmobile.com> 17318 17319 Reviewed by George Staikos. 17320 17321 Add wince specific memory files into wtf/wince 17322 https://bugs.webkit.org/show_bug.cgi?id=27550 17323 17324 * wtf/wince/FastMallocWince.h: Added. 17325 * wtf/wince/MemoryManager.cpp: Added. 17326 * wtf/wince/MemoryManager.h: Added. 17327 173282009-07-23 Norbert Leser <norbert.leser@nokia.com> 17329 17330 Reviewed by Simon Hausmann. 17331 17332 Fix for missing mmap features in Symbian 17333 https://bugs.webkit.org/show_bug.cgi?id=24540 17334 17335 Fix, conditionally for PLATFORM(SYMBIAN), as an alternative 17336 to missing support for the MAP_ANON property flag in mmap. 17337 It utilizes Symbian specific memory allocation features. 17338 17339 * runtime/Collector.cpp 17340 173412009-07-22 Gavin Barraclough <barraclough@apple.com> 17342 17343 Reviewed by Sam Weinig. 17344 17345 With ENABLE(ASSEMBLER_WX_EXCLUSIVE), only change permissions once per repatch event. 17346 ( https://bugs.webkit.org/show_bug.cgi?id=27564 ) 17347 17348 Currently we change permissions forwards and backwards for each instruction modified, 17349 instead we should only change permissions once per complete repatching event. 17350 17351 2.5% progression running with ENABLE(ASSEMBLER_WX_EXCLUSIVE) enabled, 17352 which recoups 1/3 of the penalty of running with this mode enabled. 17353 17354 * assembler/ARMAssembler.cpp: 17355 (JSC::ARMAssembler::linkBranch): 17356 - Replace usage of MakeWritable with cacheFlush. 17357 17358 * assembler/ARMAssembler.h: 17359 (JSC::ARMAssembler::patchPointerInternal): 17360 (JSC::ARMAssembler::repatchLoadPtrToLEA): 17361 - Replace usage of MakeWritable with cacheFlush. 17362 17363 * assembler/ARMv7Assembler.h: 17364 (JSC::ARMv7Assembler::relinkJump): 17365 (JSC::ARMv7Assembler::relinkCall): 17366 (JSC::ARMv7Assembler::repatchInt32): 17367 (JSC::ARMv7Assembler::repatchPointer): 17368 (JSC::ARMv7Assembler::repatchLoadPtrToLEA): 17369 (JSC::ARMv7Assembler::setInt32): 17370 - Replace usage of MakeWritable with cacheFlush. 17371 17372 * assembler/LinkBuffer.h: 17373 (JSC::LinkBuffer::performFinalization): 17374 - Make explicit call to cacheFlush. 17375 17376 * assembler/MacroAssemblerCodeRef.h: 17377 (JSC::MacroAssemblerCodeRef::MacroAssemblerCodeRef): 17378 - Make size always available. 17379 17380 * assembler/RepatchBuffer.h: 17381 (JSC::RepatchBuffer::RepatchBuffer): 17382 (JSC::RepatchBuffer::~RepatchBuffer): 17383 - Add calls to MakeWritable & makeExecutable. 17384 17385 * assembler/X86Assembler.h: 17386 (JSC::X86Assembler::relinkJump): 17387 (JSC::X86Assembler::relinkCall): 17388 (JSC::X86Assembler::repatchInt32): 17389 (JSC::X86Assembler::repatchPointer): 17390 (JSC::X86Assembler::repatchLoadPtrToLEA): 17391 - Remove usage of MakeWritable. 17392 17393 * bytecode/CodeBlock.h: 17394 (JSC::CodeBlock::getJITCode): 17395 - Provide access to CodeBlock's JITCode. 17396 17397 * jit/ExecutableAllocator.h: 17398 (JSC::ExecutableAllocator::makeExecutable): 17399 (JSC::ExecutableAllocator::cacheFlush): 17400 - Remove MakeWritable, make cacheFlush public. 17401 17402 * jit/JIT.cpp: 17403 (JSC::ctiPatchNearCallByReturnAddress): 17404 (JSC::ctiPatchCallByReturnAddress): 17405 (JSC::JIT::privateCompile): 17406 (JSC::JIT::unlinkCall): 17407 (JSC::JIT::linkCall): 17408 - Add CodeBlock argument to RepatchBuffer. 17409 17410 * jit/JIT.h: 17411 - Pass CodeBlock argument for use by RepatchBuffer. 17412 17413 * jit/JITCode.h: 17414 (JSC::JITCode::start): 17415 (JSC::JITCode::size): 17416 - Provide access to code start & size. 17417 17418 * jit/JITPropertyAccess.cpp: 17419 (JSC::JIT::privateCompilePutByIdTransition): 17420 (JSC::JIT::patchGetByIdSelf): 17421 (JSC::JIT::patchMethodCallProto): 17422 (JSC::JIT::patchPutByIdReplace): 17423 (JSC::JIT::privateCompilePatchGetArrayLength): 17424 (JSC::JIT::privateCompileGetByIdProto): 17425 (JSC::JIT::privateCompileGetByIdSelfList): 17426 (JSC::JIT::privateCompileGetByIdProtoList): 17427 (JSC::JIT::privateCompileGetByIdChainList): 17428 (JSC::JIT::privateCompileGetByIdChain): 17429 - Add CodeBlock argument to RepatchBuffer. 17430 17431 * jit/JITStubs.cpp: 17432 (JSC::JITThunks::tryCachePutByID): 17433 (JSC::JITThunks::tryCacheGetByID): 17434 (JSC::JITStubs::DEFINE_STUB_FUNCTION): 17435 - Pass CodeBlock argument for use by RepatchBuffer. 17436 174372009-07-21 Zoltan Herczeg <zherczeg@inf.u-szeged.hu> 17438 17439 Reviewed by Gavin Barraclough. 17440 17441 Cache not only the structure of the method, but the 17442 structure of its prototype as well. 17443 https://bugs.webkit.org/show_bug.cgi?id=27077 17444 17445 * bytecode/CodeBlock.cpp: 17446 (JSC::CodeBlock::~CodeBlock): 17447 * bytecode/CodeBlock.h: 17448 (JSC::MethodCallLinkInfo::MethodCallLinkInfo): 17449 * jit/JITPropertyAccess.cpp: 17450 (JSC::JIT::patchMethodCallProto): 17451 174522009-07-21 Gavin Barraclough <barraclough@apple.com> 17453 17454 Reviewed by Sam Weinig. 17455 17456 Move call linking / repatching down from AbstractMacroAssembler into MacroAssemblerARCH classes. 17457 ( https://bugs.webkit.org/show_bug.cgi?id=27527 ) 17458 17459 This allows the implementation to be defined per architecture. Specifically this addresses the 17460 fact that x86-64 MacroAssembler implements far calls as a load to register, followed by a call 17461 to register. Patching the call actually requires the pointer load to be patched, rather than 17462 the call to be patched. This is implementation detail specific to MacroAssemblerX86_64, and as 17463 such is best handled there. 17464 17465 * assembler/AbstractMacroAssembler.h: 17466 * assembler/MacroAssemblerARM.h: 17467 (JSC::MacroAssemblerARM::linkCall): 17468 (JSC::MacroAssemblerARM::repatchCall): 17469 * assembler/MacroAssemblerARMv7.h: 17470 (JSC::MacroAssemblerARMv7::linkCall): 17471 (JSC::MacroAssemblerARMv7::repatchCall): 17472 * assembler/MacroAssemblerX86.h: 17473 (JSC::MacroAssemblerX86::linkCall): 17474 (JSC::MacroAssemblerX86::repatchCall): 17475 * assembler/MacroAssemblerX86_64.h: 17476 (JSC::MacroAssemblerX86_64::linkCall): 17477 (JSC::MacroAssemblerX86_64::repatchCall): 17478 174792009-07-21 Adam Treat <adam.treat@torchmobile.com> 17480 17481 Reviewed by George Staikos. 17482 17483 Every wtf file includes other wtf files with <> style includes 17484 except this one. Fix the exception. 17485 17486 * wtf/ByteArray.h: 17487 174882009-07-21 Gavin Barraclough <barraclough@apple.com> 17489 17490 Reviewed by Oliver Hunt. 17491 17492 Move LinkBuffer/RepatchBuffer out of AbstractMacroAssembler. 17493 ( https://bugs.webkit.org/show_bug.cgi?id=27485 ) 17494 17495 This change is the first step in a process to move code that should be in 17496 the architecture-specific MacroAssembler classes up out of Assmbler and 17497 AbstractMacroAssembler. 17498 17499 * JavaScriptCore.xcodeproj/project.pbxproj: 17500 - added new files 17501 17502 * assembler/ARMAssembler.h: 17503 (JSC::ARMAssembler::linkPointer): 17504 - rename patchPointer to bring it in line with the current link/repatch naming scheme 17505 17506 * assembler/ARMv7Assembler.h: 17507 (JSC::ARMv7Assembler::linkCall): 17508 (JSC::ARMv7Assembler::linkPointer): 17509 (JSC::ARMv7Assembler::relinkCall): 17510 (JSC::ARMv7Assembler::repatchInt32): 17511 (JSC::ARMv7Assembler::repatchPointer): 17512 (JSC::ARMv7Assembler::setInt32): 17513 (JSC::ARMv7Assembler::setPointer): 17514 - rename patchPointer to bring it in line with the current link/repatch naming scheme 17515 17516 * assembler/AbstractMacroAssembler.h: 17517 (JSC::AbstractMacroAssembler::linkJump): 17518 (JSC::AbstractMacroAssembler::linkCall): 17519 (JSC::AbstractMacroAssembler::linkPointer): 17520 (JSC::AbstractMacroAssembler::getLinkerAddress): 17521 (JSC::AbstractMacroAssembler::getLinkerCallReturnOffset): 17522 (JSC::AbstractMacroAssembler::repatchJump): 17523 (JSC::AbstractMacroAssembler::repatchCall): 17524 (JSC::AbstractMacroAssembler::repatchNearCall): 17525 (JSC::AbstractMacroAssembler::repatchInt32): 17526 (JSC::AbstractMacroAssembler::repatchPointer): 17527 (JSC::AbstractMacroAssembler::repatchLoadPtrToLEA): 17528 - remove the LinkBuffer/RepatchBuffer classes, but leave a set of (private, friended) methods to interface to the Assembler 17529 17530 * assembler/LinkBuffer.h: Added. 17531 (JSC::LinkBuffer::LinkBuffer): 17532 (JSC::LinkBuffer::~LinkBuffer): 17533 (JSC::LinkBuffer::link): 17534 (JSC::LinkBuffer::patch): 17535 (JSC::LinkBuffer::locationOf): 17536 (JSC::LinkBuffer::locationOfNearCall): 17537 (JSC::LinkBuffer::returnAddressOffset): 17538 (JSC::LinkBuffer::finalizeCode): 17539 (JSC::LinkBuffer::finalizeCodeAddendum): 17540 (JSC::LinkBuffer::code): 17541 (JSC::LinkBuffer::performFinalization): 17542 - new file containing the LinkBuffer class, previously a member of AbstractMacroAssembler 17543 17544 * assembler/RepatchBuffer.h: Added. 17545 (JSC::RepatchBuffer::RepatchBuffer): 17546 (JSC::RepatchBuffer::relink): 17547 (JSC::RepatchBuffer::repatch): 17548 (JSC::RepatchBuffer::repatchLoadPtrToLEA): 17549 (JSC::RepatchBuffer::relinkCallerToTrampoline): 17550 (JSC::RepatchBuffer::relinkCallerToFunction): 17551 (JSC::RepatchBuffer::relinkNearCallerToTrampoline): 17552 - new file containing the RepatchBuffer class, previously a member of AbstractMacroAssembler 17553 17554 * assembler/X86Assembler.h: 17555 (JSC::X86Assembler::linkJump): 17556 (JSC::X86Assembler::linkCall): 17557 (JSC::X86Assembler::linkPointerForCall): 17558 (JSC::X86Assembler::linkPointer): 17559 (JSC::X86Assembler::relinkJump): 17560 (JSC::X86Assembler::relinkCall): 17561 (JSC::X86Assembler::repatchInt32): 17562 (JSC::X86Assembler::repatchPointer): 17563 (JSC::X86Assembler::setPointer): 17564 (JSC::X86Assembler::setInt32): 17565 (JSC::X86Assembler::setRel32): 17566 - rename patchPointer to bring it in line with the current link/repatch naming scheme 17567 17568 * jit/JIT.cpp: 17569 (JSC::ctiPatchNearCallByReturnAddress): 17570 (JSC::ctiPatchCallByReturnAddress): 17571 - include new headers 17572 - remove MacroAssembler:: specification from RepatchBuffer usage 17573 17574 * jit/JITPropertyAccess.cpp: 17575 * yarr/RegexJIT.cpp: 17576 - include new headers 17577 175782009-07-21 Robert Agoston <Agoston.Robert@stud.u-szeged.hu> 17579 17580 Reviewed by David Levin. 17581 17582 Fixed #undef typo. 17583 https://bugs.webkit.org/show_bug.cgi?id=27506 17584 17585 * bytecode/Opcode.h: 17586 175872009-07-21 Adam Roben <aroben@apple.com> 17588 17589 Roll out r46153, r46154, and r46155 17590 17591 These changes were causing build failures and assertion failures on 17592 Windows. 17593 17594 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 17595 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 17596 * JavaScriptCore.xcodeproj/project.pbxproj: 17597 * runtime/JSArray.cpp: 17598 * runtime/StringPrototype.cpp: 17599 * runtime/UString.cpp: 17600 * runtime/UString.h: 17601 * wtf/FastMalloc.cpp: 17602 * wtf/FastMalloc.h: 17603 * wtf/Platform.h: 17604 * wtf/PossiblyNull.h: Removed. 17605 176062009-07-21 Roland Steiner <rolandsteiner@google.com> 17607 17608 Reviewed by David Levin. 17609 17610 Add ENABLE_RUBY to list of build options 17611 https://bugs.webkit.org/show_bug.cgi?id=27324 17612 17613 * Configurations/FeatureDefines.xcconfig: Added flag ENABLE_RUBY. 17614 176152009-07-20 Oliver Hunt <oliver@apple.com> 17616 17617 Build fix attempt #2 17618 17619 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 17620 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 17621 176222009-07-20 Oliver Hunt <oliver@apple.com> 17623 17624 Build fix attempt #1 17625 17626 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 17627 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 17628 176292009-07-20 Oliver Hunt <oliver@apple.com> 17630 17631 Reviewed by Gavin Barraclough. 17632 17633 Make it harder to misuse try* allocation routines 17634 https://bugs.webkit.org/show_bug.cgi?id=27469 17635 17636 Jump through a few hoops to make it much harder to accidentally 17637 miss null-checking of values returned by the try-* allocation 17638 routines. 17639 17640 * JavaScriptCore.xcodeproj/project.pbxproj: 17641 * runtime/JSArray.cpp: 17642 (JSC::JSArray::putSlowCase): 17643 (JSC::JSArray::increaseVectorLength): 17644 * runtime/StringPrototype.cpp: 17645 (JSC::stringProtoFuncFontsize): 17646 (JSC::stringProtoFuncLink): 17647 * runtime/UString.cpp: 17648 (JSC::allocChars): 17649 (JSC::reallocChars): 17650 (JSC::expandCapacity): 17651 (JSC::UString::Rep::reserveCapacity): 17652 (JSC::UString::expandPreCapacity): 17653 (JSC::createRep): 17654 (JSC::concatenate): 17655 (JSC::UString::spliceSubstringsWithSeparators): 17656 (JSC::UString::replaceRange): 17657 (JSC::UString::append): 17658 (JSC::UString::operator=): 17659 * runtime/UString.h: 17660 (JSC::UString::Rep::createEmptyBuffer): 17661 * wtf/FastMalloc.cpp: 17662 (WTF::tryFastZeroedMalloc): 17663 (WTF::tryFastMalloc): 17664 (WTF::tryFastCalloc): 17665 (WTF::tryFastRealloc): 17666 (WTF::TCMallocStats::tryFastMalloc): 17667 (WTF::TCMallocStats::tryFastCalloc): 17668 (WTF::TCMallocStats::tryFastRealloc): 17669 * wtf/FastMalloc.h: 17670 (WTF::TryMallocReturnValue::TryMallocReturnValue): 17671 (WTF::TryMallocReturnValue::~TryMallocReturnValue): 17672 (WTF::TryMallocReturnValue::operator Maybe<T>): 17673 (WTF::TryMallocReturnValue::getValue): 17674 * wtf/PossiblyNull.h: 17675 (WTF::PossiblyNull::PossiblyNull): 17676 (WTF::PossiblyNull::~PossiblyNull): 17677 (WTF::PossiblyNull::getValue): 17678 * wtf/Platform.h: 17679 176802009-07-20 Gavin Barraclough <barraclough@apple.com> 17681 17682 RS Oliver Hunt. 17683 17684 Add ARM assembler files to xcodeproj, for convenience editing. 17685 17686 * JavaScriptCore.xcodeproj/project.pbxproj: 17687 176882009-07-20 Jessie Berlin <jberlin@apple.com> 17689 17690 Reviewed by David Levin. 17691 17692 Fix an incorrect assertion in Vector::remove. 17693 17694 https://bugs.webkit.org/show_bug.cgi?id=27477 17695 17696 * wtf/Vector.h: 17697 (WTF::::remove): 17698 Assert that the position at which to start removing elements + the 17699 length (the number of elements to remove) is less than or equal to the 17700 size of the entire Vector. 17701 177022009-07-20 Peter Kasting <pkasting@google.com> 17703 17704 Reviewed by Mark Rowe. 17705 17706 https://bugs.webkit.org/show_bug.cgi?id=27468 17707 Back out r46060, which caused problems for some Apple developers. 17708 17709 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCoreCommon.vsprops: 17710 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCoreGenerated.vcproj: 17711 * JavaScriptCore.vcproj/WTF/WTFCommon.vsprops: 17712 * JavaScriptCore.vcproj/jsc/jscCommon.vsprops: 17713 * JavaScriptCore.vcproj/testapi/testapiCommon.vsprops: 17714 177152009-07-20 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 17716 17717 Reviewed by Oliver Hunt. 17718 17719 Allow custom memory allocation control in NewThreadContext 17720 https://bugs.webkit.org/show_bug.cgi?id=27338 17721 17722 Inherits NewThreadContext struct from FastAllocBase because it 17723 has been instantiated by 'new' JavaScriptCore/wtf/Threading.cpp:76. 17724 17725 * wtf/Threading.cpp: 17726 177272009-07-20 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 17728 17729 Reviewed by Oliver Hunt. 17730 17731 Allow custom memory allocation control in JavaScriptCore's JSClassRef.h 17732 https://bugs.webkit.org/show_bug.cgi?id=27340 17733 17734 Inherit StaticValueEntry and StaticFunctionEntry struct from FastAllocBase because these 17735 have been instantiated by 'new' in JavaScriptCore/API/JSClassRef.cpp:153 17736 and in JavaScriptCore/API/JSClassRef.cpp:166. 17737 17738 * API/JSClassRef.h: 17739 177402009-07-20 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 17741 17742 Reviewed by Darin Adler. 17743 17744 Allow custom memory allocation control in JavaScriptCore's RegexPattern.h 17745 https://bugs.webkit.org/show_bug.cgi?id=27343 17746 17747 Inherits RegexPattern.h's structs (which have been instantiated by operator new) from FastAllocBase: 17748 17749 CharacterClass (new call: JavaScriptCore/yarr/RegexCompiler.cpp:144) 17750 PatternAlternative (new call: JavaScriptCore/yarr/RegexPattern.h:221) 17751 PatternDisjunction (new call: JavaScriptCore/yarr/RegexCompiler.cpp:446) 17752 17753 * yarr/RegexPattern.h: 17754 177552009-07-20 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 17756 17757 Reviewed by Darin Adler. 17758 17759 Allow custom memory allocation control for JavaScriptCore's MatchFrame struct 17760 https://bugs.webkit.org/show_bug.cgi?id=27344 17761 17762 Inherits MatchFrame struct from FastAllocBase because it has 17763 been instantiated by 'new' JavaScriptCore/pcre/pcre_exec.cpp:359. 17764 17765 * pcre/pcre_exec.cpp: 17766 177672009-07-20 Laszlo Gombos <laszlo.1.gombos@nokia.com> 17768 17769 Reviewed by Holger Freyther. 17770 17771 Remove some outdated S60 platform specific code 17772 https://bugs.webkit.org/show_bug.cgi?id=27423 17773 17774 * wtf/Platform.h: 17775 177762009-07-20 Csaba Osztrogonac <oszi@inf.u-szeged.hu> 17777 17778 Reviewed by Simon Hausmann. 17779 17780 Qt build fix with MSVC and MinGW. 17781 17782 * jsc.pro: Make sure jsc is a console application, and turn off 17783 exceptions and stl support to fix the build. 17784 177852009-07-20 Xan Lopez <xlopez@igalia.com> 17786 17787 Reviewed by Gustavo Noronha. 17788 17789 Do not use C++-style comments in preprocessor directives. 17790 17791 GCC does not like this in some configurations, using C-style 17792 comments is safer. 17793 17794 * wtf/Platform.h: 17795 177962009-07-17 Peter Kasting <pkasting@google.com> 17797 17798 Reviewed by Steve Falkenburg. 17799 17800 https://bugs.webkit.org/show_bug.cgi?id=27323 17801 Only add Cygwin to the path when it isn't already there. This avoids 17802 causing problems for people who purposefully have non-Cygwin versions of 17803 executables like svn in front of the Cygwin ones in their paths. 17804 17805 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCoreCommon.vsprops: 17806 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCoreGenerated.vcproj: 17807 * JavaScriptCore.vcproj/WTF/WTFCommon.vsprops: 17808 * JavaScriptCore.vcproj/jsc/jscCommon.vsprops: 17809 * JavaScriptCore.vcproj/testapi/testapiCommon.vsprops: 17810 178112009-07-17 Gabor Loki <loki@inf.u-szeged.hu> 17812 17813 Reviewed by Gavin Barraclough. 17814 17815 Add YARR support for generic ARM platforms (disabled by default). 17816 https://bugs.webkit.org/show_bug.cgi?id=24986 17817 17818 Add generic ARM port for MacroAssembler. It supports the whole 17819 MacroAssembler functionality except floating point. 17820 17821 The class JmpSrc is extended with a flag which enables to patch 17822 the jump destination offset during execution. This feature is 17823 required for generic ARM port. 17824 17825 Signed off by Zoltan Herczeg <zherczeg@inf.u-szeged.hu> 17826 Signed off by Gabor Loki <loki@inf.u-szeged.hu> 17827 17828 * JavaScriptCore.pri: 17829 * assembler/ARMAssembler.cpp: Added. 17830 (JSC::ARMAssembler::getLdrImmAddress): 17831 (JSC::ARMAssembler::linkBranch): 17832 (JSC::ARMAssembler::patchConstantPoolLoad): 17833 (JSC::ARMAssembler::getOp2): 17834 (JSC::ARMAssembler::genInt): 17835 (JSC::ARMAssembler::getImm): 17836 (JSC::ARMAssembler::moveImm): 17837 (JSC::ARMAssembler::dataTransfer32): 17838 (JSC::ARMAssembler::baseIndexTransfer32): 17839 (JSC::ARMAssembler::executableCopy): 17840 * assembler/ARMAssembler.h: Added. 17841 (JSC::ARM::): 17842 (JSC::ARMAssembler::ARMAssembler): 17843 (JSC::ARMAssembler::): 17844 (JSC::ARMAssembler::JmpSrc::JmpSrc): 17845 (JSC::ARMAssembler::JmpSrc::enableLatePatch): 17846 (JSC::ARMAssembler::JmpDst::JmpDst): 17847 (JSC::ARMAssembler::JmpDst::isUsed): 17848 (JSC::ARMAssembler::JmpDst::used): 17849 (JSC::ARMAssembler::emitInst): 17850 (JSC::ARMAssembler::and_r): 17851 (JSC::ARMAssembler::ands_r): 17852 (JSC::ARMAssembler::eor_r): 17853 (JSC::ARMAssembler::eors_r): 17854 (JSC::ARMAssembler::sub_r): 17855 (JSC::ARMAssembler::subs_r): 17856 (JSC::ARMAssembler::rsb_r): 17857 (JSC::ARMAssembler::rsbs_r): 17858 (JSC::ARMAssembler::add_r): 17859 (JSC::ARMAssembler::adds_r): 17860 (JSC::ARMAssembler::adc_r): 17861 (JSC::ARMAssembler::adcs_r): 17862 (JSC::ARMAssembler::sbc_r): 17863 (JSC::ARMAssembler::sbcs_r): 17864 (JSC::ARMAssembler::rsc_r): 17865 (JSC::ARMAssembler::rscs_r): 17866 (JSC::ARMAssembler::tst_r): 17867 (JSC::ARMAssembler::teq_r): 17868 (JSC::ARMAssembler::cmp_r): 17869 (JSC::ARMAssembler::orr_r): 17870 (JSC::ARMAssembler::orrs_r): 17871 (JSC::ARMAssembler::mov_r): 17872 (JSC::ARMAssembler::movs_r): 17873 (JSC::ARMAssembler::bic_r): 17874 (JSC::ARMAssembler::bics_r): 17875 (JSC::ARMAssembler::mvn_r): 17876 (JSC::ARMAssembler::mvns_r): 17877 (JSC::ARMAssembler::mul_r): 17878 (JSC::ARMAssembler::muls_r): 17879 (JSC::ARMAssembler::mull_r): 17880 (JSC::ARMAssembler::ldr_imm): 17881 (JSC::ARMAssembler::ldr_un_imm): 17882 (JSC::ARMAssembler::dtr_u): 17883 (JSC::ARMAssembler::dtr_ur): 17884 (JSC::ARMAssembler::dtr_d): 17885 (JSC::ARMAssembler::dtr_dr): 17886 (JSC::ARMAssembler::ldrh_r): 17887 (JSC::ARMAssembler::ldrh_d): 17888 (JSC::ARMAssembler::ldrh_u): 17889 (JSC::ARMAssembler::strh_r): 17890 (JSC::ARMAssembler::push_r): 17891 (JSC::ARMAssembler::pop_r): 17892 (JSC::ARMAssembler::poke_r): 17893 (JSC::ARMAssembler::peek_r): 17894 (JSC::ARMAssembler::clz_r): 17895 (JSC::ARMAssembler::bkpt): 17896 (JSC::ARMAssembler::lsl): 17897 (JSC::ARMAssembler::lsr): 17898 (JSC::ARMAssembler::asr): 17899 (JSC::ARMAssembler::lsl_r): 17900 (JSC::ARMAssembler::lsr_r): 17901 (JSC::ARMAssembler::asr_r): 17902 (JSC::ARMAssembler::size): 17903 (JSC::ARMAssembler::ensureSpace): 17904 (JSC::ARMAssembler::label): 17905 (JSC::ARMAssembler::align): 17906 (JSC::ARMAssembler::jmp): 17907 (JSC::ARMAssembler::patchPointerInternal): 17908 (JSC::ARMAssembler::patchConstantPoolLoad): 17909 (JSC::ARMAssembler::patchPointer): 17910 (JSC::ARMAssembler::repatchInt32): 17911 (JSC::ARMAssembler::repatchPointer): 17912 (JSC::ARMAssembler::repatchLoadPtrToLEA): 17913 (JSC::ARMAssembler::linkJump): 17914 (JSC::ARMAssembler::relinkJump): 17915 (JSC::ARMAssembler::linkCall): 17916 (JSC::ARMAssembler::relinkCall): 17917 (JSC::ARMAssembler::getRelocatedAddress): 17918 (JSC::ARMAssembler::getDifferenceBetweenLabels): 17919 (JSC::ARMAssembler::getCallReturnOffset): 17920 (JSC::ARMAssembler::getOp2Byte): 17921 (JSC::ARMAssembler::placeConstantPoolBarrier): 17922 (JSC::ARMAssembler::RM): 17923 (JSC::ARMAssembler::RS): 17924 (JSC::ARMAssembler::RD): 17925 (JSC::ARMAssembler::RN): 17926 (JSC::ARMAssembler::getConditionalField): 17927 * assembler/ARMv7Assembler.h: 17928 (JSC::ARMv7Assembler::JmpSrc::enableLatePatch): 17929 * assembler/AbstractMacroAssembler.h: 17930 (JSC::AbstractMacroAssembler::Call::enableLatePatch): 17931 (JSC::AbstractMacroAssembler::Jump::enableLatePatch): 17932 * assembler/MacroAssembler.h: 17933 * assembler/MacroAssemblerARM.h: Added. 17934 (JSC::MacroAssemblerARM::): 17935 (JSC::MacroAssemblerARM::add32): 17936 (JSC::MacroAssemblerARM::and32): 17937 (JSC::MacroAssemblerARM::lshift32): 17938 (JSC::MacroAssemblerARM::mul32): 17939 (JSC::MacroAssemblerARM::not32): 17940 (JSC::MacroAssemblerARM::or32): 17941 (JSC::MacroAssemblerARM::rshift32): 17942 (JSC::MacroAssemblerARM::sub32): 17943 (JSC::MacroAssemblerARM::xor32): 17944 (JSC::MacroAssemblerARM::load32): 17945 (JSC::MacroAssemblerARM::load32WithAddressOffsetPatch): 17946 (JSC::MacroAssemblerARM::loadPtrWithPatchToLEA): 17947 (JSC::MacroAssemblerARM::load16): 17948 (JSC::MacroAssemblerARM::store32WithAddressOffsetPatch): 17949 (JSC::MacroAssemblerARM::store32): 17950 (JSC::MacroAssemblerARM::pop): 17951 (JSC::MacroAssemblerARM::push): 17952 (JSC::MacroAssemblerARM::move): 17953 (JSC::MacroAssemblerARM::swap): 17954 (JSC::MacroAssemblerARM::signExtend32ToPtr): 17955 (JSC::MacroAssemblerARM::zeroExtend32ToPtr): 17956 (JSC::MacroAssemblerARM::branch32): 17957 (JSC::MacroAssemblerARM::branch16): 17958 (JSC::MacroAssemblerARM::branchTest32): 17959 (JSC::MacroAssemblerARM::jump): 17960 (JSC::MacroAssemblerARM::branchAdd32): 17961 (JSC::MacroAssemblerARM::mull32): 17962 (JSC::MacroAssemblerARM::branchMul32): 17963 (JSC::MacroAssemblerARM::branchSub32): 17964 (JSC::MacroAssemblerARM::breakpoint): 17965 (JSC::MacroAssemblerARM::nearCall): 17966 (JSC::MacroAssemblerARM::call): 17967 (JSC::MacroAssemblerARM::ret): 17968 (JSC::MacroAssemblerARM::set32): 17969 (JSC::MacroAssemblerARM::setTest32): 17970 (JSC::MacroAssemblerARM::tailRecursiveCall): 17971 (JSC::MacroAssemblerARM::makeTailRecursiveCall): 17972 (JSC::MacroAssemblerARM::moveWithPatch): 17973 (JSC::MacroAssemblerARM::branchPtrWithPatch): 17974 (JSC::MacroAssemblerARM::storePtrWithPatch): 17975 (JSC::MacroAssemblerARM::supportsFloatingPoint): 17976 (JSC::MacroAssemblerARM::supportsFloatingPointTruncate): 17977 (JSC::MacroAssemblerARM::loadDouble): 17978 (JSC::MacroAssemblerARM::storeDouble): 17979 (JSC::MacroAssemblerARM::addDouble): 17980 (JSC::MacroAssemblerARM::subDouble): 17981 (JSC::MacroAssemblerARM::mulDouble): 17982 (JSC::MacroAssemblerARM::convertInt32ToDouble): 17983 (JSC::MacroAssemblerARM::branchDouble): 17984 (JSC::MacroAssemblerARM::branchTruncateDoubleToInt32): 17985 (JSC::MacroAssemblerARM::ARMCondition): 17986 (JSC::MacroAssemblerARM::prepareCall): 17987 (JSC::MacroAssemblerARM::call32): 17988 * assembler/X86Assembler.h: 17989 (JSC::X86Assembler::JmpSrc::enableLatePatch): 17990 * jit/ExecutableAllocator.h: 17991 (JSC::ExecutableAllocator::cacheFlush): 17992 * wtf/Platform.h: 17993 * yarr/RegexJIT.cpp: 17994 (JSC::Yarr::RegexGenerator::generateEnter): 17995 (JSC::Yarr::RegexGenerator::generateReturn): 17996 179972009-07-17 Gabor Loki <loki@inf.u-szeged.hu> 17998 17999 Reviewed by Gavin Barraclough. 18000 18001 Extend AssemblerBuffer with constant pool handling mechanism. 18002 https://bugs.webkit.org/show_bug.cgi?id=24986 18003 18004 Add a platform independed constant pool framework. 18005 This pool can store 32 or 64 bits values which is enough to hold 18006 any integer, pointer or double constant. 18007 18008 * assembler/AssemblerBuffer.h: 18009 (JSC::AssemblerBuffer::putIntUnchecked): 18010 (JSC::AssemblerBuffer::putInt64Unchecked): 18011 (JSC::AssemblerBuffer::append): 18012 (JSC::AssemblerBuffer::grow): 18013 * assembler/AssemblerBufferWithConstantPool.h: Added. 18014 (JSC::): 18015 180162009-07-17 Eric Roman <eroman@chromium.org> 18017 18018 Reviewed by Darin Adler. 18019 18020 Build fix for non-Darwin. 18021 Add a guard for inclusion of RetainPtr.h which includes CoreFoundation.h 18022 18023 https://bugs.webkit.org/show_bug.cgi?id=27382 18024 18025 * wtf/unicode/icu/CollatorICU.cpp: 18026 180272009-07-17 Alexey Proskuryakov <ap@webkit.org> 18028 18029 Reviewed by John Sullivan. 18030 18031 Get user default collation order via a CFLocale API when available. 18032 18033 * wtf/unicode/icu/CollatorICU.cpp: (WTF::Collator::userDefault): 18034 180352009-07-17 Laszlo Gombos <laszlo.1.gombos@nokia.com> 18036 18037 Reviewed by Simon Hausmann. 18038 18039 [Qt] Fix the include path for the Symbian port 18040 https://bugs.webkit.org/show_bug.cgi?id=27358 18041 18042 * JavaScriptCore.pri: 18043 180442009-07-17 Csaba Osztrogonac <oszi@inf.u-szeged.hu> 18045 18046 Reviewed by David Levin. 18047 18048 Build fix on platforms don't have MMAP. 18049 https://bugs.webkit.org/show_bug.cgi?id=27365 18050 18051 * interpreter/RegisterFile.h: Including stdio.h irrespectively of HAVE(MMAP) 18052 180532009-07-16 Fumitoshi Ukai <ukai@chromium.org> 18054 18055 Reviewed by David Levin. 18056 18057 Add --web-sockets flag and ENABLE_WEB_SOCKETS define. 18058 https://bugs.webkit.org/show_bug.cgi?id=27206 18059 18060 Add ENABLE_WEB_SOCKETS 18061 18062 * Configurations/FeatureDefines.xcconfig: add ENABLE_WEB_SOCKETS 18063 180642009-07-16 Maxime Simon <simon.maxime@gmail.com> 18065 18066 Reviewed by Eric Seidel. 18067 18068 Added Haiku-specific files for JavaScriptCore. 18069 https://bugs.webkit.org/show_bug.cgi?id=26620 18070 18071 * wtf/haiku/MainThreadHaiku.cpp: Added. 18072 (WTF::initializeMainThreadPlatform): 18073 (WTF::scheduleDispatchFunctionsOnMainThread): 18074 180752009-07-16 Gavin Barraclough <barraclough@apple.com> 18076 18077 RS by Oliver Hunt. 18078 18079 Revert r45969, this fix does not appear to be valid. 18080 https://bugs.webkit.org/show_bug.cgi?id=27077 18081 18082 * bytecode/CodeBlock.cpp: 18083 (JSC::CodeBlock::~CodeBlock): 18084 (JSC::CodeBlock::unlinkCallers): 18085 * jit/JIT.cpp: 18086 * jit/JIT.h: 18087 180882009-07-16 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 18089 18090 Reviewed by Oliver Hunt. 18091 18092 Allow custom memory allocation control in ExceptionInfo and RareData struct 18093 https://bugs.webkit.org/show_bug.cgi?id=27336 18094 18095 Inherits ExceptionInfo and RareData struct from FastAllocBase because these 18096 have been instantiated by 'new' in JavaScriptCore/bytecode/CodeBlock.cpp:1289 and 18097 in JavaScriptCore/bytecode/CodeBlock.h:453. 18098 18099 Remove unnecessary WTF:: namespace from CodeBlock inheritance. 18100 18101 * bytecode/CodeBlock.h: 18102 181032009-07-16 Mark Rowe <mrowe@apple.com> 18104 18105 Rubber-stamped by Geoff Garen. 18106 18107 Fix FeatureDefines.xcconfig to not be out of sync with the rest of the world. 18108 18109 * Configurations/FeatureDefines.xcconfig: 18110 181112009-07-16 Yong Li <yong.li@torchmobile.com> 18112 18113 Reviewed by George Staikos. 18114 18115 https://bugs.webkit.org/show_bug.cgi?id=27320 18116 _countof is only included in CE6; for CE5 we need to define it ourself 18117 18118 * wtf/Platform.h: 18119 181202009-07-16 Zoltan Herczeg <zherczeg@inf.u-szeged.hu> 18121 18122 Reviewed by Oliver Hunt. 18123 18124 Workers + garbage collector: weird crashes 18125 https://bugs.webkit.org/show_bug.cgi?id=27077 18126 18127 We need to unlink cached method call sites when a function is destroyed. 18128 18129 * JavaScriptCore.xcodeproj/project.pbxproj: 18130 * bytecode/CodeBlock.cpp: 18131 (JSC::CodeBlock::~CodeBlock): 18132 (JSC::CodeBlock::unlinkCallers): 18133 * jit/JIT.cpp: 18134 (JSC::JIT::unlinkMethodCall): 18135 * jit/JIT.h: 18136 181372009-07-15 Steve Falkenburg <sfalken@apple.com> 18138 18139 Windows Build fix. 18140 18141 Visual Studio reset our intermediate directory on us. 18142 This sets it back. 18143 18144 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: 18145 * JavaScriptCore.vcproj/testapi/testapi.vcproj: 18146 181472009-07-15 Kwang Yul Seo <skyul@company100.net> 18148 18149 Reviewed by Eric Seidel. 18150 18151 https://bugs.webkit.org/show_bug.cgi?id=26794 18152 Make Yacc-generated parsers to use fastMalloc/fastFree. 18153 18154 Define YYMALLOC and YYFREE to fastMalloc and fastFree 18155 respectively. 18156 18157 * parser/Grammar.y: 18158 181592009-07-15 Darin Adler <darin@apple.com> 18160 18161 Fix a build for a particular Apple configuration. 18162 18163 * wtf/FastAllocBase.h: Change include to use "" style for 18164 including another wtf header. This is the style we use for 18165 including other public headers in the same directory. 18166 181672009-07-15 George Staikos <george.staikos@torchmobile.com> 18168 18169 Reviewed by Adam Treat. 18170 18171 https://bugs.webkit.org/show_bug.cgi?id=27303 18172 Implement createThreadInternal for WinCE. 18173 Contains changes by George Staikos <george.staikos@torchmobile.com> and Joe Mason <joe.mason@torchmobile.com> 18174 18175 * wtf/ThreadingWin.cpp: 18176 (WTF::createThreadInternal): 18177 181782009-07-15 Joe Mason <joe.mason@torchmobile.com> 18179 18180 Reviewed by George Staikos. 18181 18182 https://bugs.webkit.org/show_bug.cgi?id=27298 18183 Platform defines for WINCE. 18184 Contains changes by Yong Li <yong.li@torchmobile.com>, 18185 George Staikos <george.staikos@torchmobile.com> and Joe Mason <joe.mason@torchmobile.com> 18186 18187 * wtf/Platform.h: 18188 181892009-07-15 Yong Li <yong.li@torchmobile.com> 18190 18191 Reviewed by Adam Treat. 18192 18193 https://bugs.webkit.org/show_bug.cgi?id=27306 18194 Use RegisterClass instead of RegisterClassEx on WinCE. 18195 18196 * wtf/win/MainThreadWin.cpp: 18197 (WTF::initializeMainThreadPlatform): 18198 181992009-07-15 Yong Li <yong.li@torchmobile.com> 18200 18201 Reviewed by George Staikos. 18202 18203 https://bugs.webkit.org/show_bug.cgi?id=27301 18204 Use OutputDebugStringW on WinCE since OutputDebugStringA is not supported 18205 Originally written by Yong Li <yong.li@torchmobile.com> and refactored by 18206 Joe Mason <joe.mason@torchmobile.com> 18207 18208 * wtf/Assertions.cpp: vprintf_stderr_common 18209 182102009-07-15 Yong Li <yong.li@torchmobile.com> 18211 18212 Reviewed by George Staikos. 18213 18214 https://bugs.webkit.org/show_bug.cgi?id=27020 18215 msToGregorianDateTime should set utcOffset to 0 when outputIsUTC is false 18216 18217 * wtf/DateMath.cpp: 18218 (WTF::gregorianDateTimeToMS): 18219 182202009-07-15 Laszlo Gombos <laszlo.1.gombos@nokia.com> 18221 18222 Reviewed by Simon Hausmann. 18223 18224 [Qt] Cleanup - Remove obsolete code from the make system 18225 https://bugs.webkit.org/show_bug.cgi?id=27299 18226 18227 * JavaScriptCore.pro: 18228 * jsc.pro: 18229 182302009-07-07 Norbert Leser <norbert.leser@nokia.com> 18231 18232 Reviewed by Simon Hausmann. 18233 18234 https://bugs.webkit.org/show_bug.cgi?id=27056 18235 18236 Alternate bool operator for codewarrior compiler (WINSCW). 18237 Compiler (latest b482) reports error for UnspecifiedBoolType construct: 18238 "illegal explicit conversion from 'WTF::OwnArrayPtr<JSC::Register>' to 'bool'" 18239 18240 Same fix as in r38391. 18241 18242 * JavaScriptCore/wtf/OwnArrayPtr.h: 18243 182442009-07-15 Norbert Leser <norbert.leser@nokia.com> 18245 18246 Reviewed by Darin Adler. 18247 18248 Qualify include path with wtf to fix compilation 18249 on Symbian. 18250 https://bugs.webkit.org/show_bug.cgi?id=27055 18251 18252 * interpreter/Interpreter.h: 18253 182542009-07-15 Laszlo Gombos <laszlo.1.gombos@nokia.com> 18255 18256 Reviewed by Dave Kilzer. 18257 18258 Turn off non-portable date manipulations for SYMBIAN 18259 https://bugs.webkit.org/show_bug.cgi?id=27064 18260 18261 Introduce HAVE(TM_GMTOFF), HAVE(TM_ZONE) and HAVE(TIMEGM) guards 18262 and place the rules for controlling the guards in Platform.h. 18263 Turn off these newly introduced guards for SYMBIAN. 18264 18265 * wtf/DateMath.cpp: 18266 (WTF::calculateUTCOffset): 18267 * wtf/DateMath.h: 18268 (WTF::GregorianDateTime::GregorianDateTime): 18269 (WTF::GregorianDateTime::operator tm): 18270 * wtf/Platform.h: 18271 182722009-07-15 Norbert Leser <norbert.leser@nokia.com> 18273 18274 Reviewed by Simon Hausmann. 18275 18276 Undef ASSERT on Symbian, to avoid excessive warnings 18277 https://bugs.webkit.org/show_bug.cgi?id=27052 18278 18279 * wtf/Assertions.h: 18280 182812009-07-15 Oliver Hunt <oliver@apple.com> 18282 18283 Reviewed by Simon Hausmann. 18284 18285 REGRESSION: fast/js/postfix-syntax.html fails with interpreter 18286 https://bugs.webkit.org/show_bug.cgi?id=27294 18287 18288 When postfix operators operating on locals assign to the same local 18289 the order of operations has to be to store the incremented value, then 18290 store the unmodified number. Rather than implementing this subtle 18291 semantic in the interpreter I've just made the logic explicit in the 18292 bytecode generator, so x=x++ effectively becomes x=ToNumber(x) (for a 18293 local var x). 18294 18295 * parser/Nodes.cpp: 18296 (JSC::emitPostIncOrDec): 18297 182982009-07-15 Oliver Hunt <oliver@apple.com> 18299 18300 Reviewed by Simon Hausmann. 18301 18302 REGRESSION(43559): fast/js/kde/arguments-scope.html fails with interpreter 18303 https://bugs.webkit.org/show_bug.cgi?id=27259 18304 18305 The interpreter was incorrectly basing its need to create the arguments object 18306 based on the presence of the callframe's argument reference rather than the local 18307 arguments reference. Based on this it then overrode the local variable reference. 18308 18309 * interpreter/Interpreter.cpp: 18310 (JSC::Interpreter::privateExecute): 18311 183122009-07-14 Steve Falkenburg <sfalken@apple.com> 18313 18314 Reorganize JavaScriptCore headers into: 18315 API: include/JavaScriptCore/ 18316 Private: include/private/JavaScriptCore/ 18317 18318 Reviewed by Darin Adler. 18319 18320 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: 18321 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCoreCommon.vsprops: 18322 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCoreGenerated.make: 18323 * JavaScriptCore.vcproj/testapi/testapi.vcproj: 18324 * JavaScriptCore.vcproj/testapi/testapiCommon.vsprops: 18325 183262009-07-14 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 18327 18328 Reviewed by Darin Adler. 18329 18330 Change JSCell's superclass to NoncopyableCustomAllocated 18331 https://bugs.webkit.org/show_bug.cgi?id=27248 18332 18333 JSCell class customizes operator new, since Noncopyable will be 18334 inherited from FastAllocBase, NoncopyableCustomAllocated has 18335 to be used. 18336 18337 * runtime/JSCell.h: 18338 183392009-07-14 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 18340 18341 Reviewed by Darin Adler. 18342 18343 Change all Noncopyable inheriting visibility to public. 18344 https://bugs.webkit.org/show_bug.cgi?id=27225 18345 18346 Change all Noncopyable inheriting visibility to public because 18347 it is needed to the custom allocation framework (bug #20422). 18348 18349 * bytecode/SamplingTool.h: 18350 * bytecompiler/RegisterID.h: 18351 * interpreter/CachedCall.h: 18352 * interpreter/RegisterFile.h: 18353 * parser/Lexer.h: 18354 * parser/Parser.h: 18355 * runtime/ArgList.h: 18356 * runtime/BatchedTransitionOptimizer.h: 18357 * runtime/Collector.h: 18358 * runtime/CommonIdentifiers.h: 18359 * runtime/JSCell.h: 18360 * runtime/JSGlobalObject.h: 18361 * runtime/JSLock.h: 18362 * runtime/JSONObject.cpp: 18363 * runtime/SmallStrings.cpp: 18364 * runtime/SmallStrings.h: 18365 * wtf/CrossThreadRefCounted.h: 18366 * wtf/GOwnPtr.h: 18367 * wtf/Locker.h: 18368 * wtf/MessageQueue.h: 18369 * wtf/OwnArrayPtr.h: 18370 * wtf/OwnFastMallocPtr.h: 18371 * wtf/OwnPtr.h: 18372 * wtf/RefCounted.h: 18373 * wtf/ThreadSpecific.h: 18374 * wtf/Threading.h: 18375 * wtf/Vector.h: 18376 * wtf/unicode/Collator.h: 18377 183782009-07-14 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 18379 18380 Reviewed by Darin Adler. 18381 18382 Change ParserArenaRefCounted's superclass to RefCountedCustomAllocated 18383 https://bugs.webkit.org/show_bug.cgi?id=27249 18384 18385 ParserArenaDeletable customizes operator new, to avoid double inheritance 18386 ParserArenaDeletable's superclass has been changed to RefCountedCustomAllocated. 18387 18388 * parser/Nodes.h: 18389 183902009-07-14 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 18391 18392 Reviewed by Darin Adler. 18393 18394 Add RefCountedCustomAllocated to RefCounted.h 18395 https://bugs.webkit.org/show_bug.cgi?id=27232 18396 18397 Some class which are inherited from RefCounted customize 18398 operator new, but RefCounted is inherited from Noncopyable 18399 which will be inherited from FastAllocBase. To avoid 18400 conflicts Noncopyable inheriting was moved down to RefCounted 18401 and to avoid double inheritance this class has been added. 18402 18403 * wtf/RefCounted.h: 18404 (WTF::RefCountedCustomAllocated::deref): 18405 (WTF::RefCountedCustomAllocated::~RefCountedCustomAllocated): 18406 184072009-07-14 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 18408 18409 Reviewed by Darin Adler. 18410 18411 Add NoncopyableCustomAllocated to Noncopyable.h. 18412 https://bugs.webkit.org/show_bug.cgi?id=27228 18413 18414 Some classes which inherited from Noncopyable overrides operator new 18415 since Noncopyable'll be inherited from FastAllocBase, Noncopyable.h 18416 needs to be extended with this new class to support the overriding. 18417 18418 * wtf/Noncopyable.h: 18419 (WTFNoncopyable::NoncopyableCustomAllocated::NoncopyableCustomAllocated): 18420 (WTFNoncopyable::NoncopyableCustomAllocated::~NoncopyableCustomAllocated): 18421 184222009-07-14 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 18423 18424 Reviewed by Darin Adler. 18425 18426 Allow custom memory allocation control for JavaScriptCore's IdentifierTable class 18427 https://bugs.webkit.org/show_bug.cgi?id=27260 18428 18429 Inherits IdentifierTable class from FastAllocBase because it has been 18430 instantiated by 'new' in JavaScriptCore/runtime/Identifier.cpp:70. 18431 18432 * runtime/Identifier.cpp: 18433 184342009-07-14 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 18435 18436 Reviewed by Darin Adler. 18437 18438 Allow custom memory allocation control for JavaScriptCore's Profiler class 18439 https://bugs.webkit.org/show_bug.cgi?id=27253 18440 18441 Inherits Profiler class from FastAllocBase because it has been instantiated by 18442 'new' in JavaScriptCore/profiler/Profiler.cpp:56. 18443 18444 * profiler/Profiler.h: 18445 184462009-07-06 George Staikos <george.staikos@torchmobile.com> 18447 18448 Reviewed by Adam Treat. 18449 18450 Authors: George Staikos <george.staikos@torchmobile.com>, Joe Mason <joe.mason@torchmobile.com>, Makoto Matsumoto <matumoto@math.keio.ac.jp>, Takuji Nishimura 18451 18452 https://bugs.webkit.org/show_bug.cgi?id=27030 18453 Implement custom RNG for WinCE using Mersenne Twister 18454 18455 * wtf/RandomNumber.cpp: 18456 (WTF::randomNumber): 18457 * wtf/RandomNumberSeed.h: 18458 (WTF::initializeRandomNumberGenerator): 18459 * wtf/wince/mt19937ar.c: Added. 18460 (init_genrand): 18461 (init_by_array): 18462 (genrand_int32): 18463 (genrand_int31): 18464 (genrand_real1): 18465 (genrand_real2): 18466 (genrand_real3): 18467 (genrand_res53): 18468 184692009-07-13 Gustavo Noronha Silva <gustavo.noronha@collabora.co.uk> 18470 18471 Unreviewed make dist build fix. 18472 18473 * GNUmakefile.am: 18474 184752009-07-13 Drew Wilson <atwilson@google.com> 18476 18477 Reviewed by David Levin. 18478 18479 Add ENABLE(SHARED_WORKERS) flag and define SharedWorker APIs 18480 https://bugs.webkit.org/show_bug.cgi?id=26932 18481 18482 Added ENABLE(SHARED_WORKERS) flag (off by default). 18483 18484 * Configurations/FeatureDefines.xcconfig: 18485 184862009-07-07 Norbert Leser <norbert.leser@nokia.com> 18487 18488 Reviewed by Maciej Stachoviak. 18489 18490 https://bugs.webkit.org/show_bug.cgi?id=27058 18491 18492 Removed superfluous parenthesis around single expression. 18493 Compilers on Symbian platform fail to properly parse and compile. 18494 18495 * JavaScriptCore/wtf/Platform.h: 18496 184972009-07-13 Norbert Leser <norbert.leser@nokia.com> 18498 18499 Reviewed by Maciej Stachoviak. 18500 18501 https://bugs.webkit.org/show_bug.cgi?id=27054 18502 18503 Renamed Translator to HashTranslator 18504 18505 Codewarrior compiler (WINSCW) latest b482 cannot resolve typename 18506 mismatch between template declaration and definition 18507 (HashTranslator / Translator) 18508 18509 * wtf/HashSet.h: 18510 185112009-07-13 Norbert Leser <norbert.leser@nokia.com> 18512 18513 Reviewed by Eric Seidel. 18514 18515 https://bugs.webkit.org/show_bug.cgi?id=27053 18516 18517 Ambiguity in LabelScope initialization 18518 18519 Codewarrior compiler (WINSCW) latest b482 on Symbian cannot resolve 18520 type of "0" unambiguously. Set expression explicitly to 18521 PassRefPtr<Label>::PassRefPtr() 18522 18523 * bytecompiler/BytecodeGenerator.cpp 18524 185252009-07-11 Simon Fraser <simon.fraser@apple.com> 18526 18527 Enable support for accelerated compositing and 3d transforms on Leopard. 18528 <https://bugs.webkit.org/show_bug.cgi?id=20166> 18529 <rdar://problem/6120614> 18530 18531 Reviewed by Oliver Hunt. 18532 18533 * Configurations/FeatureDefines.xcconfig: 18534 * wtf/Platform.h: 18535 185362009-07-10 Mark Rowe <mrowe@apple.com> 18537 18538 Second part of the "make Windows happier" dance. 18539 18540 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 18541 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 18542 185432009-07-10 Mark Rowe <mrowe@apple.com> 18544 18545 Try and make the Windows build happy. 18546 18547 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 18548 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 18549 185502009-07-10 Kevin McCullough <kmccullough@apple.com> 18551 18552 Reviewed by Geoffrey Garen. 18553 18554 * debugger/Debugger.h: Made this function virtual for use in WebCore's 18555 WebInspector. 18556 185572009-07-10 Kwang Yul Seo <skyul@company100.net> 18558 18559 Reviewed by Darin Adler. 18560 18561 ParserArenaDeletable should override delete 18562 https://bugs.webkit.org/show_bug.cgi?id=26790 18563 18564 ParserArenaDeletable overrides new, but it does not override delete. 18565 ParserArenaDeletable must be freed by fastFree 18566 because it is allocated by fastMalloc. 18567 18568 * parser/NodeConstructors.h: 18569 (JSC::ParserArenaDeletable::operator delete): 18570 * parser/Nodes.h: 18571 185722009-07-10 Adam Roben <aroben@apple.com> 18573 18574 Sort all our Xcode projects 18575 18576 Accomplished using sort-Xcode-project-file. 18577 18578 Requested by Dave Kilzer. 18579 18580 * JavaScriptCore.xcodeproj/project.pbxproj: 18581 185822009-07-09 Maciej Stachowiak <mjs@apple.com> 18583 18584 Not reviewed, build fix. 18585 18586 Windows build fix for the last change. 18587 18588 * wtf/dtoa.cpp: Forgot to include Vector.h 18589 185902009-07-09 Maciej Stachowiak <mjs@apple.com> 18591 18592 Reviewed by Darin Adler. 18593 18594 REGRESSION: crash in edge cases of floating point parsing. 18595 https://bugs.webkit.org/show_bug.cgi?id=27110 18596 <rdar://problem/7044458> 18597 18598 Tests: fast/css/number-parsing-crash.html 18599 fast/css/number-parsing-crash.html 18600 fast/js/number-parsing-crash.html 18601 18602 * wtf/dtoa.cpp: 18603 (WTF::BigInt::BigInt): Converted this to more a proper class, using a Vector 18604 with inline capacity 18605 18606 (WTF::lshift): Rearranged logic somewhat nontrivially to deal with the new way of sizing BigInts. 18607 Added an assertion to verify that invariants are maintained. 18608 18609 All other functions are adapted fairly mechanically to the above changes. 18610 (WTF::BigInt::clear): 18611 (WTF::BigInt::size): 18612 (WTF::BigInt::resize): 18613 (WTF::BigInt::words): 18614 (WTF::BigInt::append): 18615 (WTF::multadd): 18616 (WTF::s2b): 18617 (WTF::i2b): 18618 (WTF::mult): 18619 (WTF::cmp): 18620 (WTF::diff): 18621 (WTF::b2d): 18622 (WTF::d2b): 18623 (WTF::ratio): 18624 (WTF::strtod): 18625 (WTF::quorem): 18626 (WTF::dtoa): 18627 186282009-07-09 Drew Wilson <atwilson@google.com> 18629 18630 Reviewed by Alexey Proskuryakov. 18631 18632 Turned on CHANNEL_MESSAGING by default because the MessageChannel API 18633 can now be implemented for Web Workers and is reasonably stable. 18634 18635 * Configurations/FeatureDefines.xcconfig: 18636 186372009-07-09 Oliver Hunt <oliver@apple.com> 18638 18639 * interpreter/Interpreter.cpp: 18640 (JSC::Interpreter::privateExecute): 18641 186422009-07-09 Oliver Hunt <oliver@apple.com> 18643 18644 Reviewed by Darin Adler. 18645 18646 Bug 27016 - Interpreter crashes due to invalid array indexes 18647 <https://bugs.webkit.org/show_bug.cgi?id=27016> 18648 18649 Unsigned vs signed conversions results in incorrect behaviour in 18650 64bit interpreter builds. 18651 18652 * interpreter/Interpreter.cpp: 18653 (JSC::Interpreter::privateExecute): 18654 186552009-07-09 Dimitri Glazkov <dglazkov@chromium.org> 18656 18657 Reviewed by Darin Fisher. 18658 18659 [Chromium] Upstream JavaScriptCore.gypi, the project file for Chromium build. 18660 https://bugs.webkit.org/show_bug.cgi?id=27135 18661 18662 * JavaScriptCore.gypi: Added. 18663 186642009-07-09 Joe Mason <joe.mason@torchmobile.com> 18665 18666 Reviewed by George Staikos. 18667 18668 Authors: Yong Li <yong.li@torchmobile.com>, Joe Mason <joe.mason@torchmobile.com> 18669 18670 https://bugs.webkit.org/show_bug.cgi?id=27031 18671 Add an override for deleteOwnedPtr(HDC) on Windows 18672 18673 * wtf/OwnPtrCommon.h: 18674 * wtf/OwnPtrWin.cpp: 18675 (WTF::deleteOwnedPtr): 18676 186772009-07-09 Laszlo Gombos <laszlo.1.gombos@nokia.com> 18678 18679 Reviewed by Darin Adler. 18680 18681 Guard singal.h dependency with HAVE(SIGNAL_H) to enable building jsc 18682 on SYMBIAN. 18683 18684 https://bugs.webkit.org/show_bug.cgi?id=27026 18685 18686 Based on Norbert Leser's work. 18687 18688 * jsc.cpp: 18689 (printUsageStatement): 18690 (parseArguments): 18691 * wtf/Platform.h: 18692 186932009-07-07 Gavin Barraclough <barraclough@apple.com> 18694 18695 Reviewed by Sam Weinig. 18696 18697 Stop loading constants into the register file. 18698 18699 Instead, use high register values (highest bit bar the sign bit set) to indicate 18700 constants in the instruction stream, and when we encounter such a value load it 18701 directly from the CodeBlock. 18702 18703 Since constants are no longer copied into the register file, this patch renders 18704 the 'unexpected constant' mechanism redundant, and removes it. 18705 18706 2% improvement, thanks to Sam Weinig. 18707 18708 * bytecode/CodeBlock.cpp: 18709 (JSC::CodeBlock::dump): 18710 (JSC::CodeBlock::CodeBlock): 18711 (JSC::CodeBlock::mark): 18712 (JSC::CodeBlock::shrinkToFit): 18713 * bytecode/CodeBlock.h: 18714 (JSC::CodeBlock::isTemporaryRegisterIndex): 18715 (JSC::CodeBlock::constantRegister): 18716 (JSC::CodeBlock::isConstantRegisterIndex): 18717 (JSC::CodeBlock::getConstant): 18718 (JSC::ExecState::r): 18719 * bytecode/Opcode.h: 18720 * bytecompiler/BytecodeGenerator.cpp: 18721 (JSC::BytecodeGenerator::preserveLastVar): 18722 (JSC::BytecodeGenerator::BytecodeGenerator): 18723 (JSC::BytecodeGenerator::addConstantValue): 18724 (JSC::BytecodeGenerator::emitEqualityOp): 18725 (JSC::BytecodeGenerator::emitLoad): 18726 (JSC::BytecodeGenerator::emitResolveBase): 18727 (JSC::BytecodeGenerator::emitResolveWithBase): 18728 (JSC::BytecodeGenerator::emitNewError): 18729 * bytecompiler/BytecodeGenerator.h: 18730 (JSC::BytecodeGenerator::emitNode): 18731 * interpreter/CallFrame.h: 18732 (JSC::ExecState::noCaller): 18733 (JSC::ExecState::hasHostCallFrameFlag): 18734 (JSC::ExecState::addHostCallFrameFlag): 18735 (JSC::ExecState::removeHostCallFrameFlag): 18736 * interpreter/Interpreter.cpp: 18737 (JSC::Interpreter::resolve): 18738 (JSC::Interpreter::resolveSkip): 18739 (JSC::Interpreter::resolveGlobal): 18740 (JSC::Interpreter::resolveBase): 18741 (JSC::Interpreter::resolveBaseAndProperty): 18742 (JSC::Interpreter::resolveBaseAndFunc): 18743 (JSC::Interpreter::dumpRegisters): 18744 (JSC::Interpreter::throwException): 18745 (JSC::Interpreter::createExceptionScope): 18746 (JSC::Interpreter::privateExecute): 18747 (JSC::Interpreter::retrieveArguments): 18748 * jit/JIT.cpp: 18749 (JSC::JIT::privateCompileMainPass): 18750 * jit/JITInlineMethods.h: 18751 (JSC::JIT::emitLoadDouble): 18752 (JSC::JIT::emitLoadInt32ToDouble): 18753 * jit/JITOpcodes.cpp: 18754 (JSC::JIT::emit_op_new_error): 18755 (JSC::JIT::emit_op_enter): 18756 (JSC::JIT::emit_op_enter_with_activation): 18757 * parser/Nodes.cpp: 18758 (JSC::DeleteResolveNode::emitBytecode): 18759 (JSC::DeleteValueNode::emitBytecode): 18760 (JSC::PrefixResolveNode::emitBytecode): 18761 * runtime/JSActivation.cpp: 18762 (JSC::JSActivation::JSActivation): 18763 * wtf/Platform.h: 18764 187652009-07-07 Mark Rowe <mrowe@apple.com> 18766 18767 Reviewed by Darin Adler. 18768 18769 Fix <https://bugs.webkit.org/show_bug.cgi?id=27025> / <rdar://problem/7033448>. 18770 Bug 27025: Crashes and regression test failures related to regexps in 64-bit 18771 18772 For x86_64 RegexGenerator uses rbx, a callee-save register, as a scratch register but 18773 neglects to save and restore it. The change in handling of the output vector in r45545 18774 altered code generation so that the RegExp::match was now storing important data in rbx, 18775 which caused crashes and bogus results when it was clobbered. 18776 18777 * yarr/RegexJIT.cpp: 18778 (JSC::Yarr::RegexGenerator::generateEnter): Save rbx. 18779 (JSC::Yarr::RegexGenerator::generateReturn): Restore rbx. 18780 187812009-07-06 Ada Chan <adachan@apple.com> 18782 18783 Reviewed by Darin Adler and Mark Rowe. 18784 18785 Decommitted spans are added to the list of normal spans rather than 18786 the returned spans in TCMalloc_PageHeap::Delete(). 18787 https://bugs.webkit.org/show_bug.cgi?id=26998 18788 18789 In TCMalloc_PageHeap::Delete(), the deleted span can be decommitted in 18790 the process of merging with neighboring spans that are also decommitted. 18791 The merged span needs to be placed in the list of returned spans (spans 18792 whose memory has been returned to the system). Right now it's always added 18793 to the list of the normal spans which can theoretically cause thrashing. 18794 18795 * wtf/FastMalloc.cpp: 18796 (WTF::TCMalloc_PageHeap::Delete): 18797 187982009-07-05 Lars Knoll <lars.knoll@nokia.com> 18799 18800 Reviewed by Maciej Stachowiak. 18801 18802 https://bugs.webkit.org/show_bug.cgi?id=26843 18803 18804 Fix run-time crashes in JavaScriptCore with the Metrowerks compiler on Symbian. 18805 18806 The Metrowerks compiler on the Symbian platform moves the globally 18807 defined Hashtables into read-only memory, despite one of the members 18808 being mutable. This causes crashes at run-time due to write access to 18809 read-only memory. 18810 18811 Avoid the use of const with this compiler by introducing the 18812 JSC_CONST_HASHTABLE macro. 18813 18814 Based on idea by Norbert Leser. 18815 18816 * runtime/Lookup.h: Define JSC_CONST_HASHTABLE as const for !WINSCW. 18817 * create_hash_table: Use JSC_CONST_HASHTABLE for hashtables. 18818 * runtime/JSGlobalData.cpp: Import various global hashtables via the macro. 18819 188202009-07-04 Dan Bernstein <mitz@apple.com> 18821 18822 - debug build fix 18823 18824 * runtime/RegExpConstructor.cpp: 18825 (JSC::RegExpConstructor::getLastParen): 18826 188272009-07-03 Yong Li <yong.li@torchmobile.com> 18828 18829 Reviewed by Maciej Stachowiak (and revised slightly) 18830 18831 RegExp::match to be optimized 18832 https://bugs.webkit.org/show_bug.cgi?id=26957 18833 18834 Allow regexp matching to use Vectors with inline capacity instead of 18835 allocating a new ovector buffer every time. 18836 18837 ~5% speedup on SunSpider string-unpack-code test, 0.3% on SunSpider overall. 18838 18839 * runtime/RegExp.cpp: 18840 (JSC::RegExp::match): 18841 * runtime/RegExp.h: 18842 * runtime/RegExpConstructor.cpp: 18843 (JSC::RegExpConstructorPrivate::RegExpConstructorPrivate): 18844 (JSC::RegExpConstructorPrivate::lastOvector): 18845 (JSC::RegExpConstructorPrivate::tempOvector): 18846 (JSC::RegExpConstructorPrivate::changeLastOvector): 18847 (JSC::RegExpConstructor::performMatch): 18848 (JSC::RegExpMatchesArray::RegExpMatchesArray): 18849 (JSC::RegExpMatchesArray::fillArrayInstance): 18850 (JSC::RegExpConstructor::getBackref): 18851 (JSC::RegExpConstructor::getLastParen): 18852 (JSC::RegExpConstructor::getLeftContext): 18853 (JSC::RegExpConstructor::getRightContext): 18854 * runtime/StringPrototype.cpp: 18855 (JSC::stringProtoFuncSplit): 18856 188572009-06-30 Kwang Yul Seo <skyul@company100.net> 18858 18859 Reviewed by Eric Seidel. 18860 18861 Override operator new/delete with const std::nothrow_t& as the second 18862 argument. 18863 https://bugs.webkit.org/show_bug.cgi?id=26792 18864 18865 On Windows CE, operator new/delete, new[]/delete[] with const 18866 std::nothrow_t& must be overrided because some standard template 18867 libraries use these operators. 18868 18869 The problem occurs when memory allocated by new(size_t s, const 18870 std::nothrow_t&) is freed by delete(void* p). This causes the umatched 18871 malloc/free problem. 18872 18873 The patch overrides all new, delete, new[] and delete[] to use 18874 fastMaloc and fastFree consistently. 18875 18876 * wtf/FastMalloc.h: 18877 (throw): 18878 188792009-06-30 Gabor Loki <loki@inf.u-szeged.hu> 18880 18881 Reviewed by Sam Weinig. 18882 18883 <https://bugs.webkit.org/show_bug.cgi?id=24986> 18884 18885 Remove unnecessary references to AssemblerBuffer. 18886 18887 * interpreter/Interpreter.cpp: 18888 * interpreter/Interpreter.h: 18889 188902009-06-29 David Levin <levin@chromium.org> 18891 18892 Reviewed by Oliver Hunt. 18893 18894 Still seeing occasional leaks from UString::sharedBuffer code 18895 https://bugs.webkit.org/show_bug.cgi?id=26420 18896 18897 The problem is that the pointer to the memory allocation isn't visible 18898 by "leaks" due to the lower bits being used as flags. The fix is to 18899 make the pointer visible in memory (in debug only). The downside of 18900 this fix that the memory allocated by sharedBuffer will still look like 18901 a leak in non-debug builds when any flags are set. 18902 18903 * wtf/PtrAndFlags.h: 18904 (WTF::PtrAndFlags::set): 18905 189062009-06-29 Sam Weinig <sam@webkit.org> 18907 18908 Reviewed by Mark Rowe. 18909 18910 Remove more unused scons support. 18911 18912 * SConstruct: Removed. 18913 189142009-06-29 Oliver Hunt <oliver@apple.com> 18915 18916 Reviewed by Gavin Barraclough. 18917 18918 <rdar://problem/7016214> JSON.parse fails to parse valid JSON with most Unicode characters 18919 <https://bugs.webkit.org/show_bug.cgi?id=26802> 18920 18921 In the original JSON.parse patch unicode was handled correctly, however in some last 18922 minute "clean up" I oversimplified isSafeStringCharacter. This patch corrects this bug. 18923 18924 * runtime/LiteralParser.cpp: 18925 (JSC::isSafeStringCharacter): 18926 (JSC::LiteralParser::Lexer::lexString): 18927 189282009-06-26 Oliver Hunt <oliver@apple.com> 18929 18930 Reviewed by Dan Bernstein. 18931 18932 <rdar://problem/7009684> REGRESSION(r45039): Crashes inside JSEvent::put on PowerPC (26746) 18933 <https://bugs.webkit.org/show_bug.cgi?id=26746> 18934 18935 Fix for r45039 incorrectly uncached a get_by_id by converting it to put_by_id. Clearly this 18936 is less than correct. This patch corrects that error. 18937 18938 * interpreter/Interpreter.cpp: 18939 (JSC::Interpreter::tryCacheGetByID): 18940 189412009-06-26 Eric Seidel <eric@webkit.org> 18942 18943 No review, only rolling out r45259. 18944 18945 Roll out r45259 after crash appeared on the bots: 18946 plugins/undefined-property-crash.html 18947 ASSERTION FAILED: s <= HeapConstants<heapType>::cellSize 18948 (leopard-intel-debug-tests/build/JavaScriptCore/runtime/Collector.cpp:278 18949 void* JSC::Heap::heapAllocate(size_t) [with JSC::HeapType heapType = PrimaryHeap]) 18950 18951 * runtime/DateInstance.cpp: 18952 * runtime/Identifier.cpp: 18953 * runtime/Lookup.h: 18954 * runtime/RegExpConstructor.cpp: 18955 * runtime/RegExpObject.h: 18956 * runtime/ScopeChain.h: 18957 * runtime/UString.h: 18958 189592009-06-26 Jedrzej Nowacki <jedrzej.nowacki@nokia.com> 18960 18961 Reviewed by Simon Hausmann. 18962 18963 Add support for QDataStream operators to Vector. 18964 18965 * wtf/Vector.h: 18966 (WTF::operator<<): 18967 (WTF::operator>>): 18968 189692009-06-24 Sam Weinig <sam@webkit.org> 18970 18971 Reviewed by Gavin Barraclough. 18972 18973 Make the opcode sampler work once again. 18974 18975 * jit/JIT.h: 18976 (JSC::JIT::compileGetByIdProto): 18977 (JSC::JIT::compileGetByIdSelfList): 18978 (JSC::JIT::compileGetByIdProtoList): 18979 (JSC::JIT::compileGetByIdChainList): 18980 (JSC::JIT::compileGetByIdChain): 18981 (JSC::JIT::compilePutByIdTransition): 18982 (JSC::JIT::compileCTIMachineTrampolines): 18983 (JSC::JIT::compilePatchGetArrayLength): 18984 * jit/JITStubCall.h: 18985 (JSC::JITStubCall::call): 18986 189872009-06-24 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 18988 18989 Reviewed by Maciej Stachowiak. 18990 18991 Extend FastAllocBase.h with 'using WTF::FastAllocBase' to avoid 18992 unnecessary WTF:: usings. 18993 Remove existing unnecessary WTF:: usings. 18994 18995 * interpreter/Interpreter.h: 18996 * profiler/CallIdentifier.h: 18997 * runtime/ScopeChain.h: 18998 * wtf/FastAllocBase.h: 18999 190002009-06-24 David Levin <levin@chromium.org> 19001 19002 Fix all builds. 19003 19004 * bytecode/CodeBlock.h: 19005 * bytecompiler/BytecodeGenerator.h: 19006 * interpreter/Register.h: 19007 190082009-06-24 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 19009 19010 Reviewed by Maciej Stachowiak. 19011 19012 https://bugs.webkit.org/show_bug.cgi?id=26677 19013 19014 Inherits CodeBlock class from FastAllocBase because it 19015 has been instantiated by 'new' in JavaScriptCore/bytecode/CodeBlock.h:217. 19016 19017 * bytecode/CodeBlock.h: 19018 190192009-06-24 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 19020 19021 Reviewed by Maciej Stachowiak. 19022 19023 https://bugs.webkit.org/show_bug.cgi?id=26676 19024 19025 Inherits BytecodeGenerator class from FastAllocBase because it has been 19026 instantiated by 'new' in JavaScriptCore/parser/Nodes.cpp:1892. 19027 19028 * bytecompiler/BytecodeGenerator.h: 19029 190302009-06-24 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 19031 19032 Reviewed by Maciej Stachowiak. 19033 19034 https://bugs.webkit.org/show_bug.cgi?id=26675 19035 19036 Inherits Register class from FastAllocBase because it has been 19037 instantiated by 'new' in JavaScriptCore/runtime/JSVariableObject.h:149. 19038 19039 * interpreter/Register.h: 19040 190412009-06-24 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 19042 19043 Reviewed by Darin Adler. 19044 19045 https://bugs.webkit.org/show_bug.cgi?id=26674 19046 19047 Inherits HashMap class from FastAllocBase because it has been 19048 instantiated by 'new' in JavaScriptCore/runtime/Structure.cpp:458. 19049 19050 * wtf/HashMap.h: 19051 190522009-06-24 Oliver Hunt <oliver@apple.com> 19053 19054 Reviewed by Darin Adler. 19055 19056 <rdar://problem/6940519> REGRESSION (Safari 4 Public Beta - TOT): google.com/adplanner shows blank page instead of site details in "basic research' 19057 19058 The problem was caused by the page returned with a function using a 19059 var declaration list containing around ~3000 variables. The solution 19060 to this is to flatten the comma expression representation and make 19061 codegen comma expressions and initializer lists iterative rather than 19062 recursive. 19063 19064 * parser/Grammar.y: 19065 * parser/NodeConstructors.h: 19066 (JSC::CommaNode::CommaNode): 19067 * parser/Nodes.cpp: 19068 (JSC::CommaNode::emitBytecode): 19069 * parser/Nodes.h: 19070 (JSC::ExpressionNode::isCommaNode): 19071 (JSC::CommaNode::isCommaNode): 19072 (JSC::CommaNode::append): 19073 190742009-06-24 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 19075 19076 Reviewed by Maciej Stachowiak. 19077 19078 https://bugs.webkit.org/show_bug.cgi?id=26645 19079 19080 Inherits ScopeChainNode class from FastAllocBase because it has been 19081 instantiated by 'new' in JavaScriptCore/runtime/ScopeChain.h:95. 19082 19083 * wtf/RefPtr.h: 19084 190852009-06-24 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 19086 19087 Reviewed by Darin Adler. 19088 19089 https://bugs.webkit.org/show_bug.cgi?id=26648 19090 19091 Inherits Deque class from FastAllocBase because it has been 19092 instantiated by 'new' with DEFINE_STATIC_LOCAL macro in 19093 JavaScriptCore/wtf/MainThread.cpp:62. 19094 19095 * wtf/Deque.h: 19096 190972009-06-24 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 19098 19099 Reviewed by Darin Adler. 19100 19101 https://bugs.webkit.org/show_bug.cgi?id=26644 19102 19103 Inherits RefPtr class from FastAllocBase because it has been 19104 instantiated by 'new' in JavaScriptCore/runtime/StructureChain.cpp:41. 19105 19106 * wtf/RefPtr.h: 19107 191082009-06-24 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 19109 19110 Reviewed by Darin Adler. 19111 19112 Inherits HashSet class from FastAllocBase, because it has been 19113 instantiated by 'new' in JavaScriptCore/runtime/Collector.h:116. 19114 19115 * wtf/HashSet.h: 19116 191172009-06-24 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 19118 19119 Reviewed by Darin Adler. 19120 19121 Inherits Vector class from FastAllocBase because it has been 19122 instantiated by 'new' in JavaScriptCore/runtime/Structure.cpp:633. 19123 19124 * wtf/Vector.h: 19125 191262009-06-24 Norbert Leser <norbert.leser@nokia.com> 19127 19128 Reviewed by Maciej Stachoviak. 19129 19130 The BytecodeGenerator objects were instantiated on stack, which takes up ~38kB per instance 19131 (each instance includes copy of JSC::CodeBlock with large SymbolTable, etc.). 19132 Specifically, since there is nested invocation (e.g., GlobalCode --> FunctionCode), 19133 the stack overflows immediately on Symbian hardware (max. 80 kB). 19134 Proposed change allocates generator objects on heap. 19135 Performance impact (if any) should be negligible and change is proposed as general fix, 19136 rather than ifdef'd for SYMBIAN. 19137 19138 * parser/Nodes.cpp: 19139 (JSC::ProgramNode::generateBytecode): 19140 (JSC::EvalNode::generateBytecode): 19141 (JSC::EvalNode::bytecodeForExceptionInfoReparse): 19142 (JSC::FunctionBodyNode::generateBytecode): 19143 (JSC::FunctionBodyNode::bytecodeForExceptionInfoReparse): 19144 191452009-06-23 Oliver Hunt <oliver@apple.com> 19146 19147 Reviewed by Gavin Barraclough. 19148 19149 <rdar://problem/6992806> REGRESSION: Enumeration can skip new properties in cases of prototypes that have more than 64 (26593) 19150 <https://bugs.webkit.org/show_bug.cgi?id=26593> 19151 19152 Do not attempt to cache structure chains if they contain a dictionary at any level. 19153 19154 * interpreter/Interpreter.cpp: 19155 (JSC::Interpreter::tryCachePutByID): 19156 (JSC::Interpreter::tryCacheGetByID): 19157 * jit/JITStubs.cpp: 19158 (JSC::JITThunks::tryCachePutByID): 19159 * runtime/Structure.cpp: 19160 (JSC::Structure::getEnumerablePropertyNames): 19161 (JSC::Structure::addPropertyTransition): 19162 * runtime/StructureChain.cpp: 19163 (JSC::StructureChain::isCacheable): 19164 * runtime/StructureChain.h: 19165 191662009-06-23 Yong Li <yong.li@torchmobile.com> 19167 19168 Reviewed by George Staikos. 19169 19170 https://bugs.webkit.org/show_bug.cgi?id=26654 19171 Add the proper export define for the JavaScriptCore API when building for WINCE. 19172 19173 * API/JSBase.h: 19174 191752009-06-23 Joe Mason <joe.mason@torchmobile.com> 19176 19177 Reviewed by Adam Treat. 19178 19179 Authors: Yong Li <yong.li@torchmobile.com>, Joe Mason <joe.mason@torchmobile.com> 19180 19181 https://bugs.webkit.org/show_bug.cgi?id=26611 19182 Implement currentThreadStackBase on WINCE by adding a global, 19183 g_stackBase, which must be set to the address of a local variable 19184 by the caller before calling any WebKit function that invokes JSC. 19185 19186 * runtime/Collector.cpp: 19187 (JSC::isPageWritable): 19188 (JSC::getStackBase): 19189 Starts at the top of the stack and returns the entire range of 19190 consecutive writable pages as an estimate of the actual stack. 19191 This will be much bigger than the actual stack range, so some 19192 dead objects can't be collected, but it guarantees live objects 19193 aren't collected prematurely. 19194 19195 (JSC::currentThreadStackBase): 19196 On WinCE, returns g_stackBase if set or call getStackBase as a 19197 fallback if not. 19198 191992009-06-23 Oliver Hunt <oliver@apple.com> 19200 19201 Reviewed by Alexey Proskuryakov. 19202 19203 Fix stupid performance problem in the LiteralParser 19204 19205 The LiteralParser was making a new UString in order to use 19206 toDouble, however UString's toDouble allows a much wider range 19207 of numberic strings than the LiteralParser accepts, and requires 19208 an additional heap allocation or two for the construciton of the 19209 UString. To rectify this we just call WTF::dtoa directly using 19210 a stack allocated buffer to hold the validated numeric literal. 19211 19212 * runtime/LiteralParser.cpp: 19213 (JSC::LiteralParser::Lexer::lexNumber): 19214 (JSC::LiteralParser::parse): 19215 * runtime/LiteralParser.h: 19216 192172009-06-22 Oliver Hunt <oliver@apple.com> 19218 19219 Reviewed by Alexey Proskuryakov. 19220 19221 Bug 26640: JSON.stringify needs to special case Boolean objects 19222 <https://bugs.webkit.org/show_bug.cgi?id=26640> 19223 19224 Add special case handling of the Boolean object so we match current 19225 ES5 errata. 19226 19227 * runtime/JSONObject.cpp: 19228 (JSC::unwrapBoxedPrimitive): renamed from unwrapNumberOrString 19229 (JSC::gap): 19230 (JSC::Stringifier::appendStringifiedValue): 19231 192322009-06-22 Oliver Hunt <oliver@apple.com> 19233 19234 Reviewed by Darin Adler. 19235 19236 Bug 26591: Support revivers in JSON.parse 19237 <https://bugs.webkit.org/show_bug.cgi?id=26591> 19238 19239 Add reviver support to JSON.parse. This completes the JSON object. 19240 19241 * runtime/JSONObject.cpp: 19242 (JSC::Walker::Walker): 19243 (JSC::Walker::callReviver): 19244 (JSC::Walker::walk): 19245 (JSC::JSONProtoFuncParse): 19246 192472009-06-21 Oliver Hunt <oliver@apple.com> 19248 19249 Reviewed by Darin Adler. 19250 19251 Bug 26592: Support standard toJSON functions 19252 <https://bugs.webkit.org/show_bug.cgi?id=26592> 19253 19254 Add support for the standard Date.toJSON function. 19255 19256 * runtime/DatePrototype.cpp: 19257 (JSC::dateProtoFuncToJSON): 19258 192592009-06-21 Oliver Hunt <oliver@apple.com> 19260 19261 Reviewed by Sam Weinig. 19262 19263 Bug 26594: JSC needs to support Date.toISOString 19264 <https://bugs.webkit.org/show_bug.cgi?id=26594> 19265 19266 Add support for Date.toISOString. 19267 19268 * runtime/DatePrototype.cpp: 19269 (JSC::dateProtoFuncToISOString): 19270 192712009-06-21 Oliver Hunt <oliver@apple.com> 19272 19273 Remove dead code. 19274 19275 * runtime/LiteralParser.cpp: 19276 (JSC::LiteralParser::parse): 19277 192782009-06-21 Oliver Hunt <oliver@apple.com> 19279 19280 Reviewed by Darin Adler and Cameron Zwarich. 19281 19282 Bug 26587: Support JSON.parse 19283 <https://bugs.webkit.org/show_bug.cgi?id=26587> 19284 19285 Extend the LiteralParser to support the full strict JSON 19286 grammar, fix a few places where the grammar was incorrectly 19287 lenient. Doesn't yet support the JSON.parse reviver function 19288 but that does not block the JSON.parse functionality itself. 19289 19290 * interpreter/Interpreter.cpp: 19291 (JSC::Interpreter::callEval): 19292 * runtime/JSGlobalObjectFunctions.cpp: 19293 (JSC::globalFuncEval): 19294 * runtime/JSONObject.cpp: 19295 (JSC::JSONProtoFuncParse): 19296 * runtime/LiteralParser.cpp: 19297 (JSC::LiteralParser::Lexer::lex): 19298 (JSC::isSafeStringCharacter): 19299 (JSC::LiteralParser::Lexer::lexString): 19300 (JSC::LiteralParser::parse): 19301 * runtime/LiteralParser.h: 19302 (JSC::LiteralParser::LiteralParser): 19303 (JSC::LiteralParser::tryJSONParse): 19304 (JSC::LiteralParser::): 19305 (JSC::LiteralParser::Lexer::Lexer): 19306 193072009-06-21 David Levin <levin@chromium.org> 19308 19309 Reviewed by NOBODY (speculative build fix for windows). 19310 19311 Simply removed some whitespace form this file to make windows build wtf and 19312 hopefully copy the new MessageQueque.h so that WebCore picks it up. 19313 19314 * wtf/Assertions.cpp: 19315 193162009-06-21 Drew Wilson <atwilson@google.com> 19317 19318 Reviewed by David Levin. 19319 19320 <https://bugs.webkit.org/show_bug.cgi?id=25043> 19321 Added support for multi-threaded MessagePorts. 19322 19323 * wtf/MessageQueue.h: 19324 (WTF::::appendAndCheckEmpty): 19325 Added API to test whether the queue was empty before adding an element. 19326 193272009-06-20 David D. Kilzer <ddkilzer@webkit.org> 19328 19329 Fix namespace comment in SegmentedVector.h 19330 19331 * wtf/SegmentedVector.h: Updated namespace comment to reflect 19332 new namespace after r44897. 19333 193342009-06-20 Zoltan Herczeg <zherczeg@inf.u-szeged.hu> 19335 19336 Bug 24986: ARM JIT port 19337 <https://bugs.webkit.org/show_bug.cgi?id=24986> 19338 19339 Reviewed by Oliver Hunt. 19340 19341 An Iterator added for SegmentedVector. Currently 19342 only the pre ++ operator is supported. 19343 19344 * wtf/SegmentedVector.h: 19345 (WTF::SegmentedVectorIterator::~SegmentedVectorIterator): 19346 (WTF::SegmentedVectorIterator::operator*): 19347 (WTF::SegmentedVectorIterator::operator->): 19348 (WTF::SegmentedVectorIterator::operator++): 19349 (WTF::SegmentedVectorIterator::operator==): 19350 (WTF::SegmentedVectorIterator::operator!=): 19351 (WTF::SegmentedVectorIterator::operator=): 19352 (WTF::SegmentedVectorIterator::SegmentedVectorIterator): 19353 (WTF::SegmentedVector::alloc): 19354 (WTF::SegmentedVector::begin): 19355 (WTF::SegmentedVector::end): 19356 193572009-06-20 Zoltan Herczeg <zherczeg@inf.u-szeged.hu> 19358 19359 Bug 24986: ARM JIT port 19360 <https://bugs.webkit.org/show_bug.cgi?id=24986> 19361 19362 Reviewed by Oliver Hunt. 19363 19364 Move SegmentedVector to /wtf subdirectory 19365 and change "namespace JSC" to "namespace WTF" 19366 19367 Additional build file updates by David Kilzer. 19368 19369 * GNUmakefile.am: Updated path to SegmentedVector.h. 19370 * JavaScriptCore.order: Updated SegmentedVector namespace from 19371 JSC to WTF in mangled C++ method name. 19372 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: 19373 Removed reference to bytecompiler\SegmentedVector.h. 19374 * JavaScriptCore.vcproj/WTF/WTF.vcproj: Added reference to 19375 wtf\SegmentedVector.h. 19376 * JavaScriptCore.xcodeproj/project.pbxproj: Moved 19377 SegmentedVector.h definition from bytecompiler subdirectory to 19378 wtf subdirectory. 19379 * bytecompiler/BytecodeGenerator.h: Updated #include path to 19380 SegmentedVector.h and prepended WTF:: namespace to its use. 19381 * parser/Lexer.h: Ditto. 19382 * wtf/SegmentedVector.h: Renamed from JavaScriptCore/bytecompiler/SegmentedVector.h. 19383 (WTF::SegmentedVector::SegmentedVector): 19384 (WTF::SegmentedVector::~SegmentedVector): 19385 (WTF::SegmentedVector::size): 19386 (WTF::SegmentedVector::at): 19387 (WTF::SegmentedVector::operator[]): 19388 (WTF::SegmentedVector::last): 19389 (WTF::SegmentedVector::append): 19390 (WTF::SegmentedVector::removeLast): 19391 (WTF::SegmentedVector::grow): 19392 (WTF::SegmentedVector::clear): 19393 (WTF::SegmentedVector::deleteAllSegments): 19394 (WTF::SegmentedVector::segmentExistsFor): 19395 (WTF::SegmentedVector::segmentFor): 19396 (WTF::SegmentedVector::subscriptFor): 19397 (WTF::SegmentedVector::ensureSegmentsFor): 19398 (WTF::SegmentedVector::ensureSegment): 19399 194002009-06-19 Gavin Barraclough <barraclough@apple.com> 19401 19402 Reviewed by NOBODY (build fix take 2 - rename FIELD_OFFSET to something that doesn't conflict with winnt.h). 19403 19404 * jit/JIT.cpp: 19405 (JSC::JIT::privateCompile): 19406 (JSC::JIT::privateCompileCTIMachineTrampolines): 19407 (JSC::JIT::emitGetVariableObjectRegister): 19408 (JSC::JIT::emitPutVariableObjectRegister): 19409 * jit/JIT.h: 19410 * jit/JITArithmetic.cpp: 19411 (JSC::JIT::emit_op_rshift): 19412 (JSC::JIT::emitSlow_op_jnless): 19413 (JSC::JIT::emitSlow_op_jnlesseq): 19414 (JSC::JIT::compileBinaryArithOp): 19415 * jit/JITCall.cpp: 19416 (JSC::JIT::compileOpCallInitializeCallFrame): 19417 (JSC::JIT::compileOpCall): 19418 * jit/JITInlineMethods.h: 19419 (JSC::JIT::restoreArgumentReference): 19420 (JSC::JIT::checkStructure): 19421 * jit/JITOpcodes.cpp: 19422 (JSC::JIT::emit_op_instanceof): 19423 (JSC::JIT::emit_op_get_scoped_var): 19424 (JSC::JIT::emit_op_put_scoped_var): 19425 (JSC::JIT::emit_op_construct_verify): 19426 (JSC::JIT::emit_op_resolve_global): 19427 (JSC::JIT::emit_op_jeq_null): 19428 (JSC::JIT::emit_op_jneq_null): 19429 (JSC::JIT::emit_op_to_jsnumber): 19430 (JSC::JIT::emit_op_catch): 19431 (JSC::JIT::emit_op_eq_null): 19432 (JSC::JIT::emit_op_neq_null): 19433 (JSC::JIT::emit_op_convert_this): 19434 (JSC::JIT::emit_op_profile_will_call): 19435 (JSC::JIT::emit_op_profile_did_call): 19436 (JSC::JIT::emitSlow_op_get_by_val): 19437 * jit/JITPropertyAccess.cpp: 19438 (JSC::JIT::emit_op_get_by_val): 19439 (JSC::JIT::emit_op_put_by_val): 19440 (JSC::JIT::emit_op_method_check): 19441 (JSC::JIT::compileGetByIdHotPath): 19442 (JSC::JIT::emit_op_put_by_id): 19443 (JSC::JIT::compilePutDirectOffset): 19444 (JSC::JIT::compileGetDirectOffset): 19445 (JSC::JIT::privateCompilePutByIdTransition): 19446 (JSC::JIT::privateCompilePatchGetArrayLength): 19447 * jit/JITStubs.cpp: 19448 (JSC::JITThunks::JITThunks): 19449 194502009-06-19 Gavin Barraclough <barraclough@apple.com> 19451 19452 Reviewed by NOBODY (Windows build fix). 19453 19454 * jit/JIT.h: 19455 * jit/JITInlineMethods.h: 19456 194572009-06-19 Gabor Loki <loki@inf.u-szeged.hu> 19458 19459 Reviewed by Gavin Barraclough. 19460 19461 Reorganize ARM architecture specific macros. 19462 Use PLATFORM_ARM_ARCH(7) instead of PLATFORM(ARM_V7). 19463 19464 Bug 24986: ARM JIT port 19465 <https://bugs.webkit.org/show_bug.cgi?id=24986> 19466 19467 * assembler/ARMv7Assembler.h: 19468 * assembler/AbstractMacroAssembler.h: 19469 (JSC::AbstractMacroAssembler::Imm32::Imm32): 19470 * assembler/MacroAssembler.h: 19471 * assembler/MacroAssemblerCodeRef.h: 19472 (JSC::MacroAssemblerCodePtr::MacroAssemblerCodePtr): 19473 * jit/ExecutableAllocator.h: 19474 (JSC::ExecutableAllocator::cacheFlush): 19475 * jit/JIT.h: 19476 * jit/JITInlineMethods.h: 19477 (JSC::JIT::restoreArgumentReferenceForTrampoline): 19478 * jit/JITStubs.cpp: 19479 * jit/JITStubs.h: 19480 * wtf/Platform.h: 19481 * yarr/RegexJIT.cpp: 19482 (JSC::Yarr::RegexGenerator::generateEnter): 19483 (JSC::Yarr::RegexGenerator::generateReturn): 19484 194852009-06-19 Gavin Barraclough <barraclough@apple.com> 19486 19487 Reviewed by Oliver Hunt. 19488 19489 Fix armv7 JIT build issues. 19490 19491 Unfortunate the arm compiler does not like the use of offsetof on JITStackFrame (since it now contains non POD types), 19492 and the FIELD_OFFSET macro does not appear constantish enough for it to be happy with its use in COMPILE_ASSERT macros. 19493 19494 * Replace offsetofs with FIELD_OFFSETs (safe on C++ objects). 19495 * Move COMPILE_ASSERTs defending layout of JITStackFrame structure on armv7 into JITThunks constructor. 19496 19497 * jit/JIT.cpp: 19498 * jit/JIT.h: 19499 * jit/JITInlineMethods.h: 19500 (JSC::JIT::restoreArgumentReference): 19501 * jit/JITOpcodes.cpp: 19502 (JSC::JIT::emit_op_catch): 19503 * jit/JITStubs.cpp: 19504 (JSC::JITThunks::JITThunks): 19505 195062009-06-19 Adam Treat <adam.treat@torchmobile.com> 19507 19508 Blind attempt at build fix. 19509 19510 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 19511 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 19512 195132009-06-19 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 19514 19515 Reviewed by Oliver Hunt. 19516 19517 Inherits CallIdentifier struct from FastAllocBase because it has been 19518 instantiated by 'new' in JavaScriptCore/profiler/CallIdentifier.h:86. 19519 19520 * wtf/HashCountedSet.h: 19521 195222009-06-19 Adam Treat <adam.treat@torchmobile.com> 19523 19524 Reviewed by Oliver Hunt. 19525 19526 https://bugs.webkit.org/show_bug.cgi?id=26540 19527 Modify the test shell to add a new function 'checkSyntax' that will 19528 only parse the source instead of executing it. In this way we can test 19529 pure parsing performance against some of the larger scripts in the wild. 19530 19531 * jsc.cpp: 19532 (GlobalObject::GlobalObject): 19533 (functionCheckSyntax): 19534 195352009-06-19 Zoltan Horvath <hzoltan@inf.u-szeged.hu> 19536 19537 Reviewed by Darin Adler. 19538 19539 Inherits HashCountedSet class from FastAllocBase because it has been 19540 instantiated by 'new' in JavaScriptCore/runtime/Collector.cpp:1095. 19541 19542 * wtf/HashCountedSet.h: 19543 195442009-06-19 Yong Li <yong.li@torchmobile.com> 19545 19546 Reviewed by George Staikos. 19547 19548 https://bugs.webkit.org/show_bug.cgi?id=26558 19549 Declare these symbols extern for WINCE as they are provided by libce. 19550 19551 * runtime/DateConstructor.cpp: 19552 * runtime/DatePrototype.cpp: 19553 (JSC::formatLocaleDate): 19554 195552009-06-19 Oliver Hunt <oliver@apple.com> 19556 19557 Reviewed by Maciej Stachowiak. 19558 19559 <rdar://problem/6988973> ScopeChain leak in interpreter builds 19560 19561 Move the Scopechain destruction code in JSFunction outside of the ENABLE(JIT) 19562 path. 19563 19564 * runtime/JSFunction.cpp: 19565 (JSC::JSFunction::~JSFunction): 19566 * wtf/Platform.h: 19567 195682009-06-19 Yong Li <yong.li@torchmobile.com> 19569 19570 Reviewed by George Staikos. 19571 19572 https://bugs.webkit.org/show_bug.cgi?id=26543 19573 Windows CE uses 'GetLastError' instead of 'errno.' 19574 19575 * interpreter/RegisterFile.h: 19576 (JSC::RegisterFile::RegisterFile): 19577 (JSC::RegisterFile::grow): 19578 195792009-06-19 David Levin <levin@chromium.org> 19580 19581 Reviewed by NOBODY (Windows build fix). 19582 19583 Add export for Windows corresponding to OSX export done in r44844. 19584 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: 19585 * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def: 19586 195872009-06-18 Oliver Hunt <oliver@apple.com> 19588 19589 Reviewed by Gavin "Viceroy of Venezuela" Barraclough. 19590 19591 Bug 26532: Native functions do not correctly unlink from optimised callsites when they're collected 19592 <https://bugs.webkit.org/show_bug.cgi?id=26532> <rdar://problem/6625385> 19593 19594 We need to make sure that each native function instance correctly unlinks any references to it 19595 when it is collected. Allowing this to happen required a few changes: 19596 * Every native function needs a codeblock to track the link information 19597 * To have this codeblock, every function now also needs its own functionbodynode 19598 so we no longer get to have a single shared instance. 19599 * Identifying a host function is now done by looking for CodeBlock::codeType() == NativeCode 19600 19601 * JavaScriptCore.exp: 19602 * bytecode/CodeBlock.cpp: 19603 (JSC::CodeBlock::CodeBlock): 19604 Constructor for NativeCode CodeBlock 19605 (JSC::CodeBlock::derefStructures): 19606 (JSC::CodeBlock::refStructures): 19607 (JSC::CodeBlock::reparseForExceptionInfoIfNecessary): 19608 (JSC::CodeBlock::handlerForBytecodeOffset): 19609 (JSC::CodeBlock::lineNumberForBytecodeOffset): 19610 (JSC::CodeBlock::expressionRangeForBytecodeOffset): 19611 (JSC::CodeBlock::getByIdExceptionInfoForBytecodeOffset): 19612 (JSC::CodeBlock::functionRegisterForBytecodeOffset): 19613 (JSC::CodeBlock::hasGlobalResolveInstructionAtBytecodeOffset): 19614 (JSC::CodeBlock::hasGlobalResolveInfoAtBytecodeOffset): 19615 (JSC::CodeBlock::setJITCode): 19616 Add assertions to ensure we don't try and use NativeCode CodeBlocks as 19617 a normal codeblock. 19618 19619 * bytecode/CodeBlock.h: 19620 (JSC::): 19621 (JSC::CodeBlock::source): 19622 (JSC::CodeBlock::sourceOffset): 19623 (JSC::CodeBlock::evalCodeCache): 19624 (JSC::CodeBlock::createRareDataIfNecessary): 19625 More assertions. 19626 19627 * jit/JIT.cpp: 19628 (JSC::JIT::privateCompileCTIMachineTrampolines): 19629 (JSC::JIT::linkCall): 19630 Update logic to allow native function caching 19631 19632 * jit/JITStubs.cpp: 19633 * parser/Nodes.cpp: 19634 (JSC::FunctionBodyNode::createNativeThunk): 19635 (JSC::FunctionBodyNode::isHostFunction): 19636 * parser/Nodes.h: 19637 * runtime/JSFunction.cpp: 19638 (JSC::JSFunction::JSFunction): 19639 (JSC::JSFunction::~JSFunction): 19640 (JSC::JSFunction::mark): 19641 * runtime/JSGlobalData.cpp: 19642 (JSC::JSGlobalData::~JSGlobalData): 19643 * runtime/JSGlobalData.h: 19644 196452009-06-18 Gavin Barraclough <barraclough@apple.com> 19646 19647 Reviewed by NOBODY (Windows build fix). 19648 19649 * wtf/DateMath.cpp: 19650 (WTF::calculateUTCOffset): 19651 196522009-06-18 Gavin Barraclough <barraclough@apple.com> 19653 19654 Reviewed by Geoff Garen. 19655 19656 Timezone calculation incorrect in Venezuela. 19657 19658 https://bugs.webkit.org/show_bug.cgi?id=26531 19659 <rdar://problem/6646169> Time is incorrectly reported to JavaScript in both Safari 3 and Firefox 3 19660 19661 The problem is that we're calculating the timezone relative to 01/01/2000, 19662 but the VET timezone changed from -4 hours to -4:30 hours on 12/09/2007. 19663 According to the spec, section 15.9.1.9 states "the time since the beginning 19664 of the year", presumably meaning the *current* year. Change the calculation 19665 to be based on whatever the current year is, rather than a canned date. 19666 19667 No performance impact. 19668 19669 * wtf/DateMath.cpp: 19670 (WTF::calculateUTCOffset): 19671 196722009-06-18 Gavin Barraclough <barraclough@apple.com> 19673 19674 Rubber Stamped by Mark Rowe (originally reviewed by Sam Weinig). 19675 19676 (Reintroducing patch added in r44492, and reverted in r44796.) 19677 19678 Change the implementation of op_throw so the stub function always modifies its 19679 return address - if it doesn't find a 'catch' it will switch to a trampoline 19680 to force a return from JIT execution. This saves memory, by avoiding the need 19681 for a unique return for every op_throw. 19682 19683 * jit/JITOpcodes.cpp: 19684 (JSC::JIT::emit_op_throw): 19685 JITStubs::cti_op_throw now always changes its return address, 19686 remove return code generated after the stub call (this is now 19687 handled by ctiOpThrowNotCaught). 19688 * jit/JITStubs.cpp: 19689 (JSC::): 19690 Add ctiOpThrowNotCaught definitions. 19691 (JSC::JITStubs::DEFINE_STUB_FUNCTION): 19692 Change cti_op_throw to always change its return address. 19693 * jit/JITStubs.h: 19694 Add ctiOpThrowNotCaught declaration. 19695 196962009-06-18 Kevin McCullough <kmccullough@apple.com> 19697 19698 Reviewed by Oliver Hunt. 19699 19700 <rdar://problem/6940880> REGRESSION: Breakpoints don't break in 64-bit 19701 19702 - Exposed functions now needed by WebCore. 19703 19704 * JavaScriptCore.exp: 19705 197062009-06-17 Darin Adler <darin@apple.com> 19707 19708 Reviewed by Oliver Hunt. 19709 19710 Bug 26429: Make JSON.stringify non-recursive so it can handle objects 19711 of arbitrary complexity 19712 https://bugs.webkit.org/show_bug.cgi?id=26429 19713 19714 For marking I decided not to use gcProtect, because this is inside the engine 19715 so it's easy enough to just do marking. And that darned gcProtect does locking! 19716 Oliver tried to convince me to used MarkedArgumentBuffer, but the constructor 19717 for that class says "FIXME: Remove all clients of this API, then remove this API." 19718 19719 * runtime/Collector.cpp: 19720 (JSC::Heap::collect): Add a call to JSONObject::markStringifiers. 19721 19722 * runtime/CommonIdentifiers.cpp: 19723 (JSC::CommonIdentifiers::CommonIdentifiers): Added emptyIdentifier. 19724 * runtime/CommonIdentifiers.h: Ditto. 19725 19726 * runtime/JSGlobalData.cpp: 19727 (JSC::JSGlobalData::JSGlobalData): Initialize firstStringifierToMark to 0. 19728 * runtime/JSGlobalData.h: Added firstStringifierToMark. 19729 19730 * runtime/JSONObject.cpp: Cut down the includes to the needed ones only. 19731 (JSC::unwrapNumberOrString): Added. Helper for unwrapping number and string 19732 objects to get their number and string values. 19733 (JSC::ReplacerPropertyName::ReplacerPropertyName): Added. The class is used 19734 to wrap an identifier or integer so we don't have to do any work unless we 19735 actually call a replacer. 19736 (JSC::ReplacerPropertyName::value): Added. 19737 (JSC::gap): Added. Helper function for the Stringifier constructor. 19738 (JSC::PropertyNameForFunctionCall::PropertyNameForFunctionCall): Added. 19739 The class is used to wrap an identifier or integer so we don't have to 19740 allocate a number or string until we actually call toJSON or a replacer. 19741 (JSC::PropertyNameForFunctionCall::asJSValue): Added. 19742 (JSC::Stringifier::Stringifier): Updated and moved out of the class 19743 definition. Added code to hook this into a singly linked list for marking. 19744 (JSC::Stringifier::~Stringifier): Remove from the singly linked list. 19745 (JSC::Stringifier::mark): Mark all the objects in the holder stacks. 19746 (JSC::Stringifier::stringify): Updated. 19747 (JSC::Stringifier::appendQuotedString): Tweaked and streamlined a bit. 19748 (JSC::Stringifier::toJSON): Renamed from toJSONValue. 19749 (JSC::Stringifier::appendStringifiedValue): Renamed from stringify. 19750 Added code to use the m_holderStack to do non-recursive stringify of 19751 objects and arrays. This code also uses the timeout checker since in 19752 pathological cases it could be slow even without calling into the 19753 JavaScript virtual machine. 19754 (JSC::Stringifier::willIndent): Added. 19755 (JSC::Stringifier::indent): Added. 19756 (JSC::Stringifier::unindent): Added. 19757 (JSC::Stringifier::startNewLine): Added. 19758 (JSC::Stringifier::Holder::Holder): Added. 19759 (JSC::Stringifier::Holder::appendNextProperty): Added. This is the 19760 function that handles the format of arrays and objects. 19761 (JSC::JSONObject::getOwnPropertySlot): Moved this down to the bottom 19762 of the file so the JSONObject class is not interleaved with the 19763 Stringifier class. 19764 (JSC::JSONObject::markStringifiers): Added. Calls mark. 19765 (JSC::JSONProtoFuncStringify): Streamlined the code here. The code 19766 to compute the gap string is now a separate function. 19767 19768 * runtime/JSONObject.h: Made everything private. Added markStringifiers. 19769 197702009-06-17 Oliver Hunt <oliver@apple.com> 19771 19772 Reviewed by Gavin Barraclough. 19773 19774 <rdar://problem/6974140> REGRESSION(r43849): Crash in cti_op_call_NotJSFunction when getting directions on maps.google.com 19775 19776 Roll out r43849 as it appears that we cannot rely on the address of 19777 an objects property storage being constant even if the structure is 19778 unchanged. 19779 19780 * jit/JIT.h: 19781 * jit/JITPropertyAccess.cpp: 19782 (JSC::JIT::compileGetDirectOffset): 19783 (JSC::JIT::privateCompileGetByIdProto): 19784 (JSC::JIT::privateCompileGetByIdProtoList): 19785 (JSC::JIT::privateCompileGetByIdChainList): 19786 (JSC::JIT::privateCompileGetByIdChain): 19787 197882009-06-17 Gavin Barraclough <barraclough@apple.com> 19789 19790 Rubber Stamped by Mark Rowe. 19791 19792 Fully revert r44492 & r44748 while we fix a bug they cause on internal builds <rdar://problem/6955963>. 19793 19794 * jit/JITOpcodes.cpp: 19795 (JSC::JIT::emit_op_throw): 19796 * jit/JITStubs.cpp: 19797 (JSC::): 19798 (JSC::JITStubs::DEFINE_STUB_FUNCTION): 19799 * jit/JITStubs.h: 19800 198012009-06-17 Gavin Barraclough <barraclough@apple.com> 19802 19803 Reviewed by Mark Rowe. 19804 19805 <rdar://problem/6947426> sunspider math-cordic.js exhibits different intermediate results running 32-bit vs. 64-bit 19806 19807 On 64-bit, NaN-encoded values must be detagged before they can be used in rshift. 19808 19809 No performance impact. 19810 19811 * jit/JITArithmetic.cpp: 19812 (JSC::JIT::emit_op_rshift): 19813 198142009-06-17 Adam Treat <adam.treat@torchmobile.com> 19815 19816 Reviewed by George Staikos. 19817 19818 https://bugs.webkit.org/show_bug.cgi?id=23155 19819 Move WIN_CE -> WINCE as previously discussed with Qt WINCE folks. 19820 19821 * jsc.cpp: 19822 (main): 19823 198242009-06-17 George Staikos <george.staikos@torchmobile.com> 19825 19826 Reviewed by Adam Treat. 19827 19828 https://bugs.webkit.org/show_bug.cgi?id=23155 19829 Move WIN_CE -> WINCE as previously discussed with Qt WINCE folks. 19830 19831 * config.h: 19832 * jsc.cpp: 19833 * wtf/Assertions.cpp: 19834 * wtf/Assertions.h: 19835 * wtf/CurrentTime.cpp: 19836 (WTF::lowResUTCTime): 19837 * wtf/DateMath.cpp: 19838 (WTF::getLocalTime): 19839 * wtf/MathExtras.h: 19840 * wtf/Platform.h: 19841 * wtf/StringExtras.h: 19842 * wtf/Threading.h: 19843 * wtf/win/MainThreadWin.cpp: 19844 198452009-06-17 Gavin Barraclough <barraclough@apple.com> 19846 19847 Reviewed by Oliver Hunt. 19848 19849 <rdar://problem/6974175> ASSERT in JITStubs.cpp at appsaccess.apple.com 19850 19851 Remove PropertySlot::putValue - PropertySlots should only be used for getting, 19852 not putting. Rename JSGlobalObject::getOwnPropertySlot to hasOwnPropertyForWrite, 19853 which is what it really was being used to ask, and remove some other getOwnPropertySlot 19854 & getOwnPropertySlotForWrite methods, which were unused and likely to lead to confusion. 19855 19856 * runtime/JSGlobalObject.h: 19857 (JSC::JSGlobalObject::hasOwnPropertyForWrite): 19858 * runtime/JSObject.h: 19859 * runtime/JSStaticScopeObject.cpp: 19860 * runtime/JSStaticScopeObject.h: 19861 * runtime/PropertySlot.h: 19862 198632009-06-16 Gavin Barraclough <barraclough@apple.com> 19864 19865 Reviewed by Oliver hunt. 19866 19867 Temporarily partially disable r44492, since this is causing some problems on internal builds. 19868 19869 * jit/JITOpcodes.cpp: 19870 (JSC::JIT::emit_op_throw): 19871 * jit/JITStubs.cpp: 19872 (JSC::JITStubs::DEFINE_STUB_FUNCTION): 19873 198742009-06-16 Sam Weinig <sam@webkit.org> 19875 19876 Fix windows build. 19877 19878 * jit/JIT.cpp: 19879 (JSC::JIT::JIT): 19880 198812009-06-16 Sam Weinig <sam@webkit.org> 19882 19883 Reviewed by Oliver Hunt. 19884 19885 Initialize m_bytecodeIndex to -1 in JIT, and correctly initialize 19886 it for each type of stub using the return address to find the correct 19887 offset. 19888 19889 * jit/JIT.cpp: 19890 (JSC::JIT::JIT): 19891 * jit/JIT.h: 19892 (JSC::JIT::compileGetByIdProto): 19893 (JSC::JIT::compileGetByIdSelfList): 19894 (JSC::JIT::compileGetByIdProtoList): 19895 (JSC::JIT::compileGetByIdChainList): 19896 (JSC::JIT::compileGetByIdChain): 19897 (JSC::JIT::compilePutByIdTransition): 19898 (JSC::JIT::compileCTIMachineTrampolines): 19899 (JSC::JIT::compilePatchGetArrayLength): 19900 * jit/JITStubCall.h: 19901 (JSC::JITStubCall::call): 19902 19903== Rolled over to ChangeLog-2009-06-16 == 19904