JavaScriptCore/ChangeLog

   1 2010-10-15  Ilya Tikhonovsky  <loislo@chromium.org>
   2
   3         Unreviewed build fix for Debug Leopard which is failng to compile after r69842.
   4
   5         * yarr/RegexInterpreter.cpp:
   6         (JSC::Yarr::ByteCompiler::emitDisjunction):
   7
   8 2010-10-15  Peter Varga  <pvarga@inf.u-szeged.hu>
   9
  10         Reviewed by Gavin Barraclough.
  11
  12         The parenthetical assertion checking isn't working in some cases with YARR
  13         Interpreter
  14         https://bugs.webkit.org/show_bug.cgi?id=46893
  15
  16         Calculate the countToCheck value of a TypeParentheticalAssertion by
  17         subtracting the number of characters which follows
  18         a TypeParentheticalAssertion term with the number of characters which should
  19         be matched by terms which are contained
  20         in the TypeParentheticalAssertion term (minimumSize).
  21
  22         * yarr/RegexInterpreter.cpp:
  23         (JSC::Yarr::ByteCompiler::emitDisjunction):
  24
  25 2010-10-14  Nathan Vander Wilt  <nate@andyet.net>
  26
  27         Reviewed by Darin Adler.
  28
  29         Added parser for ECMAScript 5 standard date format, so Date.parse can handle RFC 3339 timestamps: https://bugs.webkit.org/show_bug.cgi?id=44632
  30
  31         * runtime/DateConversion.cpp:
  32         (JSC::parseDate):
  33         * wtf/DateMath.cpp:
  34         (WTF::ymdhmsToSeconds):
  35         (WTF::parseES5DateFromNullTerminatedCharacters):
  36         * wtf/DateMath.h:
  37
  38 2010-10-14  Nikolas Zimmermann  <nzimmermann@rim.com>
  39
  40         Reviewed by Gavin Barraclough.
  41
  42         Replace lots of String::format() usages by StringConcatenate
  43         https://bugs.webkit.org/show_bug.cgi?id=47664
  44
  45         Add StringTypeAdapter<char> to accept single characters for makeString().
  46
  47         * wtf/text/StringConcatenate.h:
  48         (WTF::makeString):
  49
  50 2010-10-14  David Goodwin  <david_goodwin@apple.com>
  51
  52         Reviewed by Darin Adler.
  53
  54         need way to measure size of JITed ARM code
  55         https://bugs.webkit.org/show_bug.cgi?id=47121
  56
  57         * assembler/LinkBuffer.h:
  58         (JSC::LinkBuffer::linkCode):
  59         (JSC::LinkBuffer::dumpLinkStats):
  60         (JSC::LinkBuffer::dumpCode):
  61
  62 2010-10-14  Peter Varga  <pvarga@inf.u-szeged.hu>
  63
  64         Reviewed by Gavin Barraclough.
  65
  66         The backreference checking isn't working in some cases with YARR Interpreter
  67         https://bugs.webkit.org/show_bug.cgi?id=46904
  68
  69         The Interpreter::matchBackReference() function returns true without matching
  70         when a backreference points to the same parentheses where it is.
  71
  72         * yarr/RegexInterpreter.cpp:
  73         (JSC::Yarr::Interpreter::matchBackReference):
  74
  75 2010-10-14  No'am Rosenthal  <noam.rosenthal@nokia.com>
  76
  77         Reviewed by Andreas Kling.
  78
  79         [Qt] Text breaking is slow: enable ICU as an opt-in
  80         https://bugs.webkit.org/show_bug.cgi?id=40332
  81
  82         Added a config flag that enables ICU as an opt-in instead of the Qt specific code.
  83         Because of the inclusion of ICU headers, some explicit casting was necessary in UnicodeQt4.h
  84
  85         * JavaScriptCore.pri:
  86         * wtf/unicode/qt4/UnicodeQt4.h:
  87         (WTF::Unicode::toLower):
  88         (WTF::Unicode::toUpper):
  89         (WTF::Unicode::toTitleCase):
  90         (WTF::Unicode::foldCase):
  91         (WTF::Unicode::isPrintableChar):
  92         (WTF::Unicode::isSeparatorSpace):
  93         (WTF::Unicode::isPunct):
  94         (WTF::Unicode::isLower):
  95         (WTF::Unicode::mirroredChar):
  96         (WTF::Unicode::combiningClass):
  97         (WTF::Unicode::direction):
  98         (WTF::Unicode::category):
  99
 100 2010-10-14  Anton Faern  <anton@bladehawke.com>
 101
 102         Reviewed by Csaba Osztrogonác.
 103
 104         https://bugs.webkit.org/show_bug.cgi?id=47658
 105         NetBSD was not included in the WTF_PLATFORM_FOO to WTF_OS_FOO
 106         change.  This means that OS(NETBSD) is also undefined.
 107
 108         * wtf/Platform.h: s/_PLATFORM_/_OS_/ for NetBSD
 109
 110 2010-10-13  David Goodwin  <david_goodwin@apple.com>
 111
 112         Reviewed by Oliver Hunt.
 113
 114         ARMv7 JIT should generated conditional branches when possible
 115         https://bugs.webkit.org/show_bug.cgi?id=47384
 116
 117         Use different jump padding sizes for conditional and unconditional
 118         jumps (12 bytes and 10 bytes respectively). This allows the JIT to
 119         include the IT instruction as part of the conditional jump sequence
 120         which in turn allows it to optimize away the IT using an ARMv7
 121         conditional branch instruction. Use 2-byte B(T1) and 4-byte B(T3) for
 122         conditional branches when displacement is in range. Also use IT/B(T4)
 123         for conditional branch when displacement does not fit in B(T3).
 124
 125         For unconditional jump, instruction selection options are:
 126         B(T2), B(T4), MOVW/MOVT/BX. For conditional jump, instruction selection
 127         options are: B(T1), B(T3), IT/B(T4), ITTT/MOVW/MOVT/BX.
 128
 129         * assembler/ARMv7Assembler.cpp:
 130         * assembler/ARMv7Assembler.h:
 131         (JSC::ARMv7Assembler::JmpSrc::JmpSrc):
 132         (JSC::ARMv7Assembler::ifThenElse):
 133         (JSC::ARMv7Assembler::jumpSizeDelta):
 134         (JSC::ARMv7Assembler::canCompact):
 135         (JSC::ARMv7Assembler::computeJumpType):
 136         (JSC::ARMv7Assembler::link):
 137         (JSC::ARMv7Assembler::canBeJumpT1):
 138         (JSC::ARMv7Assembler::canBeJumpT3):
 139         (JSC::ARMv7Assembler::canBeJumpT4):
 140         (JSC::ARMv7Assembler::linkJumpT1):
 141         (JSC::ARMv7Assembler::linkJumpT3):
 142         (JSC::ARMv7Assembler::linkJumpT4):
 143         (JSC::ARMv7Assembler::linkConditionalJumpT4):
 144         (JSC::ARMv7Assembler::linkBX):
 145         (JSC::ARMv7Assembler::linkConditionalBX):
 146         (JSC::ARMv7Assembler::linkJumpAbsolute):
 147         * assembler/LinkBuffer.h:
 148         (JSC::LinkBuffer::linkCode):
 149         * assembler/MacroAssemblerARMv7.h:
 150         (JSC::MacroAssemblerARMv7::canCompact):
 151         (JSC::MacroAssemblerARMv7::computeJumpType):
 152         (JSC::MacroAssemblerARMv7::jumpSizeDelta):
 153         (JSC::MacroAssemblerARMv7::jump):
 154         (JSC::MacroAssemblerARMv7::nearCall):
 155         (JSC::MacroAssemblerARMv7::call):
 156         (JSC::MacroAssemblerARMv7::ret):
 157         (JSC::MacroAssemblerARMv7::tailRecursiveCall):
 158         (JSC::MacroAssemblerARMv7::makeJump):
 159         (JSC::MacroAssemblerARMv7::makeBranch):
 160
 161 2010-10-13  Fridrich Strba  <fridrich.strba@bluewin.ch>
 162
 163         Reviewed by Darin Adler.
 164
 165         Don't depend on Windows on sched_yield and sched.h
 166         https://bugs.webkit.org/show_bug.cgi?id=45543
 167
 168         sched.h is part of pthreads and sched_yield is implemented
 169         in pthreads-win32 as Sleep(0). This patch avoids a gratuitous
 170         dependency on pthreads-win32 in this file.
 171
 172         * wtf/TCSpinLock.h:
 173         (TCMalloc_SlowLock):
 174
 175 2010-10-13  Kwang Yul Seo  <skyul@company100.net>
 176
 177         Reviewed by Kent Tamura.
 178
 179         [BREWMP] Port unicode
 180         https://bugs.webkit.org/show_bug.cgi?id=45716
 181
 182         Brew MP port uses only the subset of ICU library to reduce the binary size.
 183         Follow the WinCE's implementation.
 184
 185         * wtf/Platform.h:
 186         * wtf/unicode/Unicode.h:
 187         * wtf/unicode/brew/UnicodeBrew.cpp: Added.
 188         (WTF::Unicode::toLower):
 189         (WTF::Unicode::toUpper):
 190         (WTF::Unicode::foldCase):
 191         (WTF::Unicode::isPrintableChar):
 192         (WTF::Unicode::isUpper):
 193         (WTF::Unicode::isLower):
 194         (WTF::Unicode::isDigit):
 195         (WTF::Unicode::isPunct):
 196         (WTF::Unicode::isAlphanumeric):
 197         (WTF::Unicode::toTitleCase):
 198         (WTF::Unicode::direction):
 199         (WTF::Unicode::category):
 200         (WTF::Unicode::decompositionType):
 201         (WTF::Unicode::combiningClass):
 202         (WTF::Unicode::mirroredChar):
 203         (WTF::Unicode::digitValue):
 204         (WTF::Unicode::isSpace):
 205         (WTF::Unicode::isLetter):
 206         * wtf/unicode/brew/UnicodeBrew.h: Added.
 207         (WTF::Unicode::isArabicChar):
 208         (WTF::Unicode::isSeparatorSpace):
 209         (WTF::Unicode::hasLineBreakingPropertyComplexContext):
 210         (WTF::Unicode::hasLineBreakingPropertyComplexContextOrIdeographic):
 211         (WTF::Unicode::umemcasecmp):
 212
 213 2010-10-13  Gavin Barraclough  <barraclough@apple.com>
 214
 215         Windows build fix.
 216
 217         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def:
 218
 219 2010-10-13  Adam Barth  <abarth@webkit.org>
 220
 221         Reviewed by Maciej Stachowiak.
 222
 223         [WTFURL] Add URLQueryCanonicalizer
 224         https://bugs.webkit.org/show_bug.cgi?id=45088
 225
 226         This class canonicalizes the query component of URLs.  The main tricky
 227         bit there is the convertCharset function, which I've moved to a
 228         templated dependency.  There'll likely be more about that in future
 229         patches.
 230
 231         * JavaScriptCore.xcodeproj/project.pbxproj:
 232         * wtf/url/src/URLEscape.cpp: Added.
 233         * wtf/url/src/URLEscape.h: Added.
 234         (WTF::appendEscapedCharacter):
 235         * wtf/url/src/URLQueryCanonicalizer.h: Added.
 236         (WTF::URLQueryCanonicalizer::canonicalize):
 237         (WTF::URLQueryCanonicalizer::isAllASCII):
 238         (WTF::URLQueryCanonicalizer::appendRaw8BitQueryString):
 239         (WTF::URLQueryCanonicalizer::convertToQueryEncoding):
 240
 241 2010-10-13  Gavin Barraclough  <barraclough@apple.com>
 242
 243         Reviewed by Oliver Hunt.
 244
 245         Bug 43987 - Downloading using XHR is much slower than before
 246         Change StringBuilder to use overcapacity in a StringImpl, rather than a Vector.
 247         Fundamentally this should be the same (copies current contents to expand capacity,
 248         rather than using a rope), but this approach allows the intermadiate state of the
 249         String to be inspected in the buffer without copying to resolve.
 250
 251         * runtime/JSONObject.cpp:
 252         (JSC::Stringifier::appendQuotedString):
 253         (JSC::Stringifier::Holder::appendNextProperty):
 254             Renamed StringBuilder::size() -> length() (to match other String types).
 255
 256         * runtime/UStringBuilder.h:
 257         (JSC::UStringBuilder::append):
 258         (JSC::UStringBuilder::toUString):
 259             Update for changes in parent class, can just 'using' the append methods.
 260
 261         * wtf/text/StringBuilder.cpp: Added.
 262         (WTF::StringBuilder::reifyString):
 263         (WTF::StringBuilder::resize):
 264         (WTF::StringBuilder::reserveCapacity):
 265         (WTF::StringBuilder::allocateBuffer):
 266         (WTF::StringBuilder::appendUninitialized):
 267         (WTF::StringBuilder::append):
 268         (WTF::StringBuilder::shrinkToFit):
 269         * wtf/text/StringBuilder.h:
 270         (WTF::StringBuilder::StringBuilder):
 271         (WTF::StringBuilder::append):
 272         (WTF::StringBuilder::toString):
 273         (WTF::StringBuilder::toStringPreserveCapacity):
 274         (WTF::StringBuilder::length):
 275         (WTF::StringBuilder::isEmpty):
 276         (WTF::StringBuilder::operator[]):
 277         (WTF::StringBuilder::clear):
 278             Class updated to use overcapacity in a StringImpl, rather than a Vector.
 279
 280         * Android.mk:
 281         * Android.v8.wtf.mk:
 282         * GNUmakefile.am:
 283         * JavaScriptCore.exp:
 284         * JavaScriptCore.gypi:
 285         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCoreCommon.vsprops:
 286         * JavaScriptCore.vcproj/WTF/WTF.vcproj:
 287         * JavaScriptCore.vcproj/jsc/jsc.vcproj:
 288         * JavaScriptCore.xcodeproj/project.pbxproj:
 289         * wtf/CMakeLists.txt:
 290         * wtf/wtf.pri:
 291
 292 2010-10-13  Adam Roben  <aroben@apple.com>
 293
 294         Export tryFastRealloc for WebKit2's benefit
 295
 296         Rubber-stamped by Anders Carlsson.
 297
 298         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: Added
 299         tryFastRealloc. Removed RegExpObject::info, which is now exported via
 300         JS_EXPORTDATA.
 301
 302 2010-10-13  Adam Barth  <abarth@webkit.org>
 303
 304         Reviewed by Maciej Stachowiak.
 305
 306         [WTFURL] Add a mechanism for classifying types of characters
 307         https://bugs.webkit.org/show_bug.cgi?id=45085
 308
 309         Various characters have different escaping rules depending on where
 310         they are in URLs.  This patch adds a table containing that information.
 311
 312         * JavaScriptCore.xcodeproj/project.pbxproj:
 313         * wtf/url/src/URLCharacterTypes.cpp: Added.
 314         * wtf/url/src/URLCharacterTypes.h: Added.
 315         (WTF::URLCharacterTypes::isQueryChar):
 316         (WTF::URLCharacterTypes::isIPv4Char):
 317         (WTF::URLCharacterTypes::isHexChar):
 318         (WTF::URLCharacterTypes::isCharOfType):
 319
 320 2010-10-13  Xan Lopez  <xlopez@igalia.com>
 321
 322         Reviewed by Csaba Osztrogonác.
 323
 324         Missing parameters for bytecode dump of next_pname
 325         https://bugs.webkit.org/show_bug.cgi?id=47590
 326
 327         * bytecode/CodeBlock.cpp:
 328         (JSC::CodeBlock::dump): add missing parameters to the dump.
 329
 330 2010-10-13  Nikolas Zimmermann  <nzimmermann@rim.com>
 331
 332         Reviewed by Dirk Schulze.
 333
 334         Add wtf/text/StringConcatenate
 335         https://bugs.webkit.org/show_bug.cgi?id=47584
 336
 337         Move runtime/StringConcatenate.h to wtf/text, make it work for Strings too.
 338         Add a special runtime/UStringConcatenate.h class that inherits from StringConcatenate, and extends it for use with UString.
 339         Exactly the same design that has been followed while refactoring StringBuilder.
 340
 341         The UString variants can all be removed as soon as WTF::String & JSC::UString converge.
 342
 343         * GNUmakefile.am: Add wtf/text/StringConcatenate.h and runtime/UStringConcatenate.h.
 344         * JavaScriptCore.gypi: Ditto.
 345         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: Ditto.
 346         * JavaScriptCore.vcproj/WTF/WTF.vcproj: Ditto.
 347         * JavaScriptCore.xcodeproj/project.pbxproj: Ditto.
 348         * bytecode/CodeBlock.cpp: s/makeString/makeUString/
 349         (JSC::escapeQuotes):
 350         (JSC::valueToSourceString):
 351         (JSC::constantName):
 352         (JSC::idName):
 353         (JSC::CodeBlock::registerName):
 354         (JSC::regexpToSourceString):
 355         (JSC::regexpName):
 356         * bytecompiler/NodesCodegen.cpp: Ditto.
 357         (JSC::substitute):
 358         * profiler/Profiler.cpp: Ditto.
 359         (JSC::Profiler::createCallIdentifier):
 360         * runtime/ExceptionHelpers.cpp: Ditto.
 361         (JSC::createUndefinedVariableError):
 362         (JSC::createErrorMessage):
 363         (JSC::createInvalidParamError):
 364         * runtime/FunctionConstructor.cpp: Ditto.
 365         (JSC::constructFunction):
 366         * runtime/FunctionPrototype.cpp: Ditto.
 367         (JSC::insertSemicolonIfNeeded):
 368         * runtime/JSONObject.cpp: Ditto.
 369         (JSC::Stringifier::indent):
 370         * runtime/JSStringBuilder.h:
 371         (JSC::jsMakeNontrivialString):
 372         * runtime/RegExpConstructor.cpp: Ditto.
 373         (JSC::constructRegExp):
 374         * runtime/RegExpObject.cpp: Ditto.
 375         (JSC::RegExpObject::match):
 376         * runtime/RegExpPrototype.cpp: Ditto.
 377         (JSC::regExpProtoFuncCompile):
 378         * runtime/StringConcatenate.h: Removed.
 379         * runtime/UStringConcatenate.h: Added. Only contains the StringTypeAdapter<JSC::UString> code and the makeUString variants, the rest lives in wtf/text/StringConcatenate.h
 380         (JSC::makeUString):
 381         * wtf/text/StringConcatenate.h: Copied from runtime/StringConcatenate.h.
 382         (WTF::makeString):
 383
 384 2010-10-12  Gavin Barraclough  <barraclough@apple.com>
 385
 386         Windows build fix.
 387
 388         * wtf/text/StringBuilder.h:
 389         (WTF::StringBuilder::length):
 390
 391 2010-10-12  Nikolas Zimmermann  <nzimmermann@rim.com>
 392
 393         Reviewed by Gavin Barraclough.
 394
 395         Unify JSC::StringBuilder & WebCore::StringBuilder
 396         https://bugs.webkit.org/show_bug.cgi?id=47538
 397
 398         Move runtime/StringBuilder.h to wtf/text/StringBuilder.h. Rename build() to toString() and return a WTF::String().
 399         Move the append(const JSC::UString&) method into runtime/UStringBuilder.h.
 400         UStringBuilder inherits from StringBuilder.h and adds append(const JSC::UString&) and UString toUString() functionality.
 401
 402         No new code, just move code around.
 403
 404         * GNUmakefile.am: Add wtf/text/StringBuilder.h / runtime/UStringBuilder.h. Remove runtime/StringBuilder.h.
 405         * JavaScriptCore.gypi: Ditto.
 406         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: Ditto.
 407         * JavaScriptCore.vcproj/WTF/WTF.vcproj: Ditto.
 408         * JavaScriptCore.xcodeproj/project.pbxproj: Ditto.
 409         * runtime/Executable.cpp:
 410         (JSC::FunctionExecutable::paramString): Use UStringBuilder, instead of StringBuilder. Rename build() -> toUString().
 411         * runtime/FunctionConstructor.cpp:
 412         (JSC::constructFunction): Ditto.
 413         * runtime/JSGlobalObjectFunctions.cpp:
 414         (JSC::globalFuncUnescape): Ditto.
 415         * runtime/JSONObject.cpp:
 416         (JSC::Stringifier::stringify): Ditto.
 417         (JSC::Stringifier::appendQuotedString): Ditto.
 418         (JSC::Stringifier::appendStringifiedValue): Ditto.
 419         (JSC::Stringifier::startNewLine): Ditto.
 420         (JSC::Stringifier::Holder::appendNextProperty): Ditto.
 421         * runtime/LiteralParser.cpp:
 422         (JSC::LiteralParser::Lexer::lexString): Ditto.
 423         * runtime/NumberPrototype.cpp: Remove unneeded JSStringBuilder.h / StringBuilder.h include.
 424         * runtime/StringBuilder.h: Removed.
 425         * runtime/UStringBuilder.h: Added. Inherits from WTF::StringBuilder, extending it by two methods.
 426         (JSC::UStringBuilder::append): append(const JSC::UString&)
 427         (JSC::UStringBuilder::toUString):
 428         * wtf/text/StringBuilder.h: Copied from runtime/StringBuilder.h. Move JSC::UString parts into runtime/UStringBuilder.h
 429         (WTF::StringBuilder::append): Renamed m_buffer to buffer everywhere.
 430         (WTF::StringBuilder::isEmpty): Ditto (+ constify method).
 431         (WTF::StringBuilder::reserveCapacity): Ditto.
 432         (WTF::StringBuilder::resize): Ditto.
 433         (WTF::StringBuilder::size): Ditto.
 434         (WTF::StringBuilder::operator[]): Ditto.
 435         (WTF::StringBuilder::toString): Ditto (+ renamed from build()). Returns a String, not an UString. The old build() method is now named toUString() and lives in UStringBuilder.
 436
 437 2010-10-12  Michael Saboff  <msaboff@apple.com>
 438
 439         Reviewed by Oliver Hunt.
 440
 441         Cleaned up the processing of replacements after regular expression
 442         processing, especially the case where there wasn't a match.
 443         Changed to use empty strings instead of computing a zero length sub
 444         string.
 445         https://bugs.webkit.org/show_bug.cgi?id=47506
 446
 447         * runtime/StringPrototype.cpp:
 448         (JSC::jsSpliceSubstringsWithSeparators):
 449         (JSC::stringProtoFuncReplace):
 450
 451 2010-10-11  Patrick Gansterer  <paroga@webkit.org>
 452
 453         Unreviewed.
 454
 455         Clang build fix after r69472.
 456         https://bugs.webkit.org/show_bug.cgi?id=46523
 457
 458         * wtf/text/StringHash.h:
 459
 460 2010-10-11  Oliver Hunt  <oliver@apple.com>
 461
 462         Undo last minute change to 32bit build.
 463
 464         * jit/JITOpcodes32_64.cpp:
 465         (JSC::JIT::emit_op_convert_this_strict):
 466
 467 2010-10-11  Brian Weinstein  <bweinstein@apple.com>
 468
 469         Build fix for Windows. Add a necessary export from r69516.
 470
 471         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def:
 472
 473 2010-10-11  Oliver Hunt  <oliver@apple.com>
 474
 475         Fix interpreter build -- was broken by incorrect merge.
 476
 477         * interpreter/Interpreter.cpp:
 478         (JSC::Interpreter::privateExecute):
 479
 480 2010-10-01  Oliver Hunt  <oliver@apple.com>
 481
 482         Reviewed by Gavin Barraclough.
 483
 484         [ES5] Implement strict mode
 485         https://bugs.webkit.org/show_bug.cgi?id=10701
 486
 487         Initial strict mode implementation.  This is the simplest
 488         implementation that could possibly work and adds (hopefully)
 489         all of the restrictions required by strict mode.  There are
 490         a number of inefficiencies, especially in the handling of
 491         arguments and eval as smart implementations would make this
 492         patch more complicated.
 493
 494         The SyntaxChecker AST builder has become somewhat more complex
 495         as strict mode does require more parse tree information to
 496         validate the syntax.
 497
 498         Summary of major changes to the parser:
 499             * We track when we enter strict mode (this may come as a surprise)
 500             * Strict mode actually requires a degree of AST knowledge to validate
 501               so the SyntaxChecker now produces values that can be used to distinguish
 502               "node" types.
 503             * We now track variables that are written to.  We do this to
 504               statically identify writes to global properties that don't exist
 505               and abort at that point.  This should actually make it possible
 506               to optimise some other cases in the future but for now it's
 507               purely for validity checking.  Currently writes are only tracked
 508               in strict mode code.
 509             * Labels are now tracked as it is now a syntax error to jump to a label
 510               that does not exist (or to use break, continue, or return in a context
 511               where they would be invalid).
 512
 513         Runtime changes:
 514             * In order to get correct hanlding of the Arguments object all
 515               strict mode functions that reference arguments create and tearoff
 516               the arguments object on entry.  This is not strictly necessary
 517               but was the least work necessary to get the correct behaviour.
 518             * PutPropertySlot now tracks whether it is being used for a strict
 519               mode write, and if so Object::put will throw when a write can't be
 520               completed.
 521             * StrictEvalActivation was added as an "activation" object for strict
 522               mode eval (so that strict eval does not introduce new variables into
 523               the containing scope).
 524
 525         * CMakeLists.txt:
 526         * GNUmakefile.am:
 527         * JavaScriptCore.exp:
 528         * JavaScriptCore.pro:
 529         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
 530         * JavaScriptCore.xcodeproj/project.pbxproj:
 531         * bytecode/CodeBlock.cpp:
 532         (JSC::CodeBlock::dump):
 533         (JSC::CodeBlock::CodeBlock):
 534         (JSC::CodeBlock::reparseForExceptionInfoIfNecessary):
 535         * bytecode/CodeBlock.h:
 536         (JSC::CodeBlock::isStrictMode):
 537         * bytecode/EvalCodeCache.h:
 538         (JSC::EvalCodeCache::get):
 539         * bytecode/Opcode.h:
 540         * bytecompiler/BytecodeGenerator.cpp:
 541         (JSC::BytecodeGenerator::BytecodeGenerator):
 542         (JSC::BytecodeGenerator::createArgumentsIfNecessary):
 543         (JSC::BytecodeGenerator::emitReturn):
 544         * bytecompiler/BytecodeGenerator.h:
 545         (JSC::BytecodeGenerator::isStrictMode):
 546         (JSC::BytecodeGenerator::makeFunction):
 547         * debugger/Debugger.cpp:
 548         (JSC::evaluateInGlobalCallFrame):
 549         * debugger/DebuggerCallFrame.cpp:
 550         (JSC::DebuggerCallFrame::evaluate):
 551         * interpreter/Interpreter.cpp:
 552         (JSC::Interpreter::callEval):
 553         (JSC::Interpreter::unwindCallFrame):
 554         (JSC::Interpreter::execute):
 555         (JSC::Interpreter::privateExecute):
 556         * jit/JIT.cpp:
 557         (JSC::JIT::privateCompileMainPass):
 558         (JSC::JIT::privateCompileSlowCases):
 559         * jit/JIT.h:
 560         * jit/JITOpcodes.cpp:
 561         (JSC::JIT::emit_op_get_pnames):
 562         (JSC::JIT::emit_op_convert_this_strict):
 563         (JSC::JIT::emitSlow_op_convert_this_strict):
 564         * jit/JITOpcodes32_64.cpp:
 565         (JSC::JIT::emit_op_get_pnames):
 566         * jit/JITStubs.cpp:
 567         (JSC::DEFINE_STUB_FUNCTION):
 568         * jit/JITStubs.h:
 569         * parser/ASTBuilder.h:
 570         (JSC::ASTBuilder::createFunctionBody):
 571         (JSC::ASTBuilder::isResolve):
 572         * parser/JSParser.cpp:
 573         (JSC::JSParser::next):
 574         (JSC::JSParser::startLoop):
 575         (JSC::JSParser::endLoop):
 576         (JSC::JSParser::startSwitch):
 577         (JSC::JSParser::endSwitch):
 578         (JSC::JSParser::setStrictMode):
 579         (JSC::JSParser::strictMode):
 580         (JSC::JSParser::isValidStrictMode):
 581         (JSC::JSParser::declareParameter):
 582         (JSC::JSParser::breakIsValid):
 583         (JSC::JSParser::pushLabel):
 584         (JSC::JSParser::popLabel):
 585         (JSC::JSParser::hasLabel):
 586         (JSC::JSParser::DepthManager::DepthManager):
 587         (JSC::JSParser::DepthManager::~DepthManager):
 588         (JSC::JSParser::Scope::Scope):
 589         (JSC::JSParser::Scope::startSwitch):
 590         (JSC::JSParser::Scope::endSwitch):
 591         (JSC::JSParser::Scope::startLoop):
 592         (JSC::JSParser::Scope::endLoop):
 593         (JSC::JSParser::Scope::inLoop):
 594         (JSC::JSParser::Scope::breakIsValid):
 595         (JSC::JSParser::Scope::pushLabel):
 596         (JSC::JSParser::Scope::popLabel):
 597         (JSC::JSParser::Scope::hasLabel):
 598         (JSC::JSParser::Scope::isFunction):
 599         (JSC::JSParser::Scope::declareVariable):
 600         (JSC::JSParser::Scope::declareWrite):
 601         (JSC::JSParser::Scope::deleteProperty):
 602         (JSC::JSParser::Scope::declareParameter):
 603         (JSC::JSParser::Scope::setNeedsFullActivation):
 604         (JSC::JSParser::Scope::collectFreeVariables):
 605         (JSC::JSParser::Scope::getUncapturedWrittenVariables):
 606         (JSC::JSParser::Scope::getDeletedVariables):
 607         (JSC::JSParser::Scope::setStrictMode):
 608         (JSC::JSParser::Scope::strictMode):
 609         (JSC::JSParser::Scope::isValidStrictMode):
 610         (JSC::JSParser::pushScope):
 611         (JSC::JSParser::popScope):
 612         (JSC::JSParser::declareVariable):
 613         (JSC::JSParser::declareWrite):
 614         (JSC::JSParser::deleteProperty):
 615         (JSC::jsParse):
 616         (JSC::JSParser::JSParser):
 617         (JSC::JSParser::parseProgram):
 618         (JSC::JSParser::parseSourceElements):
 619         (JSC::JSParser::parseDoWhileStatement):
 620         (JSC::JSParser::parseWhileStatement):
 621         (JSC::JSParser::parseVarDeclarationList):
 622         (JSC::JSParser::parseConstDeclarationList):
 623         (JSC::JSParser::parseForStatement):
 624         (JSC::JSParser::parseBreakStatement):
 625         (JSC::JSParser::parseContinueStatement):
 626         (JSC::JSParser::parseReturnStatement):
 627         (JSC::JSParser::parseWithStatement):
 628         (JSC::JSParser::parseSwitchStatement):
 629         (JSC::JSParser::parseSwitchClauses):
 630         (JSC::JSParser::parseSwitchDefaultClause):
 631         (JSC::JSParser::parseTryStatement):
 632         (JSC::JSParser::parseBlockStatement):
 633         (JSC::JSParser::parseStatement):
 634         (JSC::JSParser::parseFormalParameters):
 635         (JSC::JSParser::parseFunctionBody):
 636         (JSC::JSParser::parseFunctionInfo):
 637         (JSC::JSParser::parseFunctionDeclaration):
 638         (JSC::JSParser::parseExpressionOrLabelStatement):
 639         (JSC::JSParser::parseIfStatement):
 640         (JSC::JSParser::parseExpression):
 641         (JSC::JSParser::parseAssignmentExpression):
 642         (JSC::JSParser::parseConditionalExpression):
 643         (JSC::JSParser::parseBinaryExpression):
 644         (JSC::JSParser::parseStrictObjectLiteral):
 645         (JSC::JSParser::parsePrimaryExpression):
 646         (JSC::JSParser::parseMemberExpression):
 647         (JSC::JSParser::parseUnaryExpression):
 648         * parser/JSParser.h:
 649         * parser/Lexer.cpp:
 650         (JSC::Lexer::parseString):
 651         (JSC::Lexer::lex):
 652         * parser/Lexer.h:
 653         (JSC::Lexer::isReparsing):
 654         * parser/Nodes.cpp:
 655         (JSC::ScopeNode::ScopeNode):
 656         (JSC::FunctionBodyNode::FunctionBodyNode):
 657         (JSC::FunctionBodyNode::create):
 658         * parser/Nodes.h:
 659         (JSC::ScopeNode::isStrictMode):
 660         * parser/Parser.cpp:
 661         (JSC::Parser::parse):
 662         * parser/Parser.h:
 663         (JSC::Parser::parse):
 664         * parser/SyntaxChecker.h:
 665         (JSC::SyntaxChecker::SyntaxChecker):
 666         (JSC::SyntaxChecker::makeFunctionCallNode):
 667         (JSC::SyntaxChecker::appendToComma):
 668         (JSC::SyntaxChecker::createCommaExpr):
 669         (JSC::SyntaxChecker::makeAssignNode):
 670         (JSC::SyntaxChecker::makePrefixNode):
 671         (JSC::SyntaxChecker::makePostfixNode):
 672         (JSC::SyntaxChecker::makeTypeOfNode):
 673         (JSC::SyntaxChecker::makeDeleteNode):
 674         (JSC::SyntaxChecker::makeNegateNode):
 675         (JSC::SyntaxChecker::makeBitwiseNotNode):
 676         (JSC::SyntaxChecker::createLogicalNot):
 677         (JSC::SyntaxChecker::createUnaryPlus):
 678         (JSC::SyntaxChecker::createVoid):
 679         (JSC::SyntaxChecker::thisExpr):
 680         (JSC::SyntaxChecker::createResolve):
 681         (JSC::SyntaxChecker::createObjectLiteral):
 682         (JSC::SyntaxChecker::createArray):
 683         (JSC::SyntaxChecker::createNumberExpr):
 684         (JSC::SyntaxChecker::createString):
 685         (JSC::SyntaxChecker::createBoolean):
 686         (JSC::SyntaxChecker::createNull):
 687         (JSC::SyntaxChecker::createBracketAccess):
 688         (JSC::SyntaxChecker::createDotAccess):
 689         (JSC::SyntaxChecker::createRegex):
 690         (JSC::SyntaxChecker::createNewExpr):
 691         (JSC::SyntaxChecker::createConditionalExpr):
 692         (JSC::SyntaxChecker::createAssignResolve):
 693         (JSC::SyntaxChecker::createFunctionExpr):
 694         (JSC::SyntaxChecker::createFunctionBody):
 695         (JSC::SyntaxChecker::appendBinaryExpressionInfo):
 696         (JSC::SyntaxChecker::operatorStackPop):
 697         * runtime/Arguments.cpp:
 698         (JSC::Arguments::createStrictModeCallerIfNecessary):
 699         (JSC::Arguments::createStrictModeCalleeIfNecessary):
 700         (JSC::Arguments::getOwnPropertySlot):
 701         (JSC::Arguments::getOwnPropertyDescriptor):
 702         (JSC::Arguments::put):
 703         (JSC::Arguments::deleteProperty):
 704         * runtime/Arguments.h:
 705         (JSC::Arguments::Arguments):
 706         * runtime/CommonIdentifiers.cpp:
 707         (JSC::CommonIdentifiers::CommonIdentifiers):
 708         * runtime/CommonIdentifiers.h:
 709         * runtime/Error.cpp:
 710         (JSC::StrictModeTypeErrorFunction::StrictModeTypeErrorFunction):
 711         (JSC::StrictModeTypeErrorFunction::constructThrowTypeError):
 712         (JSC::StrictModeTypeErrorFunction::getConstructData):
 713         (JSC::StrictModeTypeErrorFunction::callThrowTypeError):
 714         (JSC::StrictModeTypeErrorFunction::getCallData):
 715         (JSC::createTypeErrorFunction):
 716         * runtime/Error.h:
 717         * runtime/Executable.cpp:
 718         (JSC::EvalExecutable::EvalExecutable):
 719         (JSC::ProgramExecutable::ProgramExecutable):
 720         (JSC::FunctionExecutable::FunctionExecutable):
 721         (JSC::EvalExecutable::compileInternal):
 722         (JSC::ProgramExecutable::checkSyntax):
 723         (JSC::ProgramExecutable::compileInternal):
 724         (JSC::FunctionExecutable::compileForCallInternal):
 725         (JSC::FunctionExecutable::compileForConstructInternal):
 726         (JSC::FunctionExecutable::reparseExceptionInfo):
 727         (JSC::EvalExecutable::reparseExceptionInfo):
 728         (JSC::FunctionExecutable::fromGlobalCode):
 729         (JSC::ProgramExecutable::reparseExceptionInfo):
 730         * runtime/Executable.h:
 731         (JSC::ScriptExecutable::ScriptExecutable):
 732         (JSC::ScriptExecutable::isStrictMode):
 733         (JSC::EvalExecutable::create):
 734         (JSC::FunctionExecutable::create):
 735         * runtime/JSActivation.cpp:
 736         (JSC::JSActivation::toStrictThisObject):
 737         * runtime/JSActivation.h:
 738         * runtime/JSFunction.cpp:
 739         (JSC::createDescriptorForThrowingProperty):
 740         (JSC::JSFunction::getOwnPropertySlot):
 741         (JSC::JSFunction::getOwnPropertyDescriptor):
 742         (JSC::JSFunction::put):
 743         * runtime/JSGlobalData.cpp:
 744         (JSC::JSGlobalData::JSGlobalData):
 745         * runtime/JSGlobalData.h:
 746         * runtime/JSGlobalObject.cpp:
 747         (JSC::JSGlobalObject::reset):
 748         * runtime/JSGlobalObject.h:
 749         (JSC::JSGlobalObject::internalFunctionStructure):
 750         * runtime/JSGlobalObjectFunctions.cpp:
 751         (JSC::globalFuncEval):
 752         * runtime/JSObject.cpp:
 753         (JSC::JSObject::put):
 754         (JSC::JSObject::toStrictThisObject):
 755         (JSC::throwTypeError):
 756         * runtime/JSObject.h:
 757         (JSC::JSObject::isStrictModeFunction):
 758         (JSC::JSObject::putDirectInternal):
 759         (JSC::JSObject::putDirect):
 760         (JSC::JSValue::putDirect):
 761         (JSC::JSValue::toStrictThisObject):
 762         * runtime/JSStaticScopeObject.cpp:
 763         (JSC::JSStaticScopeObject::toStrictThisObject):
 764         * runtime/JSStaticScopeObject.h:
 765         * runtime/JSValue.h:
 766         * runtime/JSZombie.h:
 767         (JSC::JSZombie::toStrictThisObject):
 768         * runtime/PutPropertySlot.h:
 769         (JSC::PutPropertySlot::PutPropertySlot):
 770         (JSC::PutPropertySlot::isStrictMode):
 771         * runtime/StrictEvalActivation.cpp: Added.
 772         (JSC::StrictEvalActivation::StrictEvalActivation):
 773         (JSC::StrictEvalActivation::deleteProperty):
 774         (JSC::StrictEvalActivation::toThisObject):
 775         (JSC::StrictEvalActivation::toStrictThisObject):
 776         * runtime/StrictEvalActivation.h: Added.
 777
 778 2010-10-10  Patrick Gansterer  <paroga@webkit.org>
 779
 780         Unreviewed.
 781
 782         Windows build fix after r69472.
 783
 784         * wtf/text/StringHash.h:
 785         (WTF::CaseFoldingHash::hash):
 786
 787 2010-10-10  Patrick Gansterer  <paroga@webkit.org>
 788
 789         Reviewed by Adam Barth.
 790
 791         Use WTF::StringHasher in WTF::CaseFoldingHash
 792         https://bugs.webkit.org/show_bug.cgi?id=46523
 793
 794         * wtf/text/StringHash.h:
 795         (WTF::CaseFoldingHash::foldCase):
 796         (WTF::CaseFoldingHash::hash):
 797
 798 2010-10-09  Pratik Solanki  <psolanki@apple.com>
 799
 800         Reviewed by Xan Lopez.
 801
 802         https://bugs.webkit.org/show_bug.cgi?id=47445
 803         Remove unused function WTFThreadData::initializeIdentifierTable()
 804
 805         * wtf/WTFThreadData.h:
 806
 807 2010-10-08  Michael Saboff  <msaboff@apple.com>
 808
 809         Reviewed by Darin Adler.
 810
 811         Added check to start of subexpression being positive before using
 812         subexpression in replacement.
 813         https://bugs.webkit.org/show_bug.cgi?id=47324
 814
 815         * runtime/StringPrototype.cpp:
 816         (JSC::substituteBackreferencesSlow):
 817
 818 2010-10-08  Chris Evans  <cevans@google.com>
 819
 820         Reviewed by David Levin.
 821
 822         https://bugs.webkit.org/show_bug.cgi?id=47393
 823
 824         Use unsigned consistently to check for max StringImpl length.
 825         Add a few integer overflow checks.
 826         Uses the existing paradigm of CRASH() when we can't reasonably handle a crazily large request.
 827
 828         * wtf/text/WTFString.cpp:
 829         * wtf/text/StringImpl.h:
 830         * wtf/text/StringImpl.cpp:
 831         Better use of size_t vs. unsigned; check for integer overflows.
 832
 833 2010-10-07  David Goodwin  <david_goodwin@apple.com>
 834
 835         Reviewed by Oliver Hunt.
 836
 837         ARM JIT generates undefined operations due to partially uninitialized ShiftTypeAndAmount
 838         https://bugs.webkit.org/show_bug.cgi?id=47356
 839
 840         * assembler/ARMv7Assembler.h:
 841
 842 2010-10-06  Chris Evans  <cevans@google.com>
 843
 844         Reviewed by David Levin.
 845
 846         https://bugs.webkit.org/show_bug.cgi?id=47248
 847
 848         Use size_t consistently in CString, to prevent theoretical trouble
 849         with > 4GB strings on 64-bit platforms.
 850
 851         * wtf/text/CString.h:
 852         * wtf/text/CString.cpp:
 853         Use size_t for string lengths.
 854         * wtf/MD5.cpp:
 855         (WTF::expectMD5): use suitable format string + cast for size_t.
 856         * JavaScriptCore.exp:
 857         Update symbol name.
 858
 859 2010-10-06  Anders Carlsson  <andersca@apple.com>
 860
 861         Reviewed by Sam Weinig.
 862
 863         Start cleaning up Arguments.h
 864         https://bugs.webkit.org/show_bug.cgi?id=47304
 865
 866         * wtf/TypeTraits.h:
 867         * wtf/TypeTraits.cpp:
 868         Add RemoveReference type trait.
 869
 870 2010-10-06  Rafael Antognolli  <antognolli@profusion.mobi>
 871
 872         Unreviewed build fix.
 873
 874         [EFL] Build fix for glib support.
 875         https://bugs.webkit.org/show_bug.cgi?id=47221
 876
 877         If compiling with GLib support enabled, we also need to link wtf against
 878         glib library.
 879
 880         * wtf/CMakeListsEfl.txt:
 881
 882 2010-10-05  Kwang Yul Seo  <skyul@company100.net>
 883
 884         Reviewed by Gavin Barraclough.
 885
 886         [BREWMP] Port ExecutableAllocator::cacheFlush to enable ARM JIT
 887         https://bugs.webkit.org/show_bug.cgi?id=47117
 888
 889         Use IMemCache1 to flush data cache and invalidate instruction cache.
 890
 891         * jit/ExecutableAllocator.h:
 892         (JSC::ExecutableAllocator::cacheFlush):
 893
 894 2010-10-05  Leandro Pereira  <leandro@profusion.mobi>
 895
 896         Unreviewed. Build fix.
 897
 898         Moved "jsc" directory to "shell", so that the name does not clash with the
 899         JavaScriptCore shell in some build systems.
 900         http://webkit.org/b/47049
 901
 902         * CMakeLists.txt: Changed reference from "jsc" to "shell".
 903         * jsc: Removed.
 904         * jsc/CMakeLists.txt: Removed.
 905         * jsc/CMakeListsEfl.txt: Removed.
 906         * shell: Copied from JavaScriptCore/jsc.
 907
 908 2010-10-05  Kwang Yul Seo  <skyul@company100.net>
 909
 910         Reviewed by Kent Tamura.
 911
 912         [BREWMP] Use PlatformRefPtr in randomNumber
 913         https://bugs.webkit.org/show_bug.cgi?id=46989
 914
 915         Use PlatformRefPtr to free memory automatically.
 916
 917         * wtf/RandomNumber.cpp:
 918         (WTF::randomNumber):
 919
 920 2010-10-05  Oliver Hunt  <oliver@apple.com>
 921
 922         Reviewed by Darin Adler.
 923
 924         REGRESSION(r68338): JavaScript error on PowerPC only (crashes on Interpreter built for x86_64)
 925         https://bugs.webkit.org/show_bug.cgi?id=46690
 926
 927         Use the correct register value when initialising the arguments
 928         object in the interpreter.  This is covered by existing tests.
 929
 930         * interpreter/Interpreter.cpp:
 931         (JSC::Interpreter::privateExecute):
 932
 933 2010-10-04  David Goodwin  <david_goodwin@apple.com>
 934
 935         Reviewed by Oliver Hunt.
 936
 937         ARMv7 JIT should take advantage of 2-byte branches to reduce code size
 938         https://bugs.webkit.org/show_bug.cgi?id=47007
 939
 940         * assembler/ARMv7Assembler.cpp:
 941         * assembler/ARMv7Assembler.h:
 942         (JSC::ARMv7Assembler::computeJumpType):
 943         (JSC::ARMv7Assembler::link):
 944         (JSC::ARMv7Assembler::canBeJumpT2):
 945         (JSC::ARMv7Assembler::canBeJumpT4):
 946         (JSC::ARMv7Assembler::linkBX):
 947         (JSC::ARMv7Assembler::linkJumpT4):
 948         (JSC::ARMv7Assembler::linkJumpT2):
 949         (JSC::ARMv7Assembler::linkJumpAbsolute):
 950
 951 2010-10-04  Gyuyoung Kim  <gyuyoung.kim@samsung.com>
 952
 953         Reviewed by Antonio Gomes.
 954
 955         [EFL] Use fast malloc for WebKit EFL
 956         https://bugs.webkit.org/show_bug.cgi?id=46691
 957
 958         Use fast malloc for WebKit EFL because the fast malloc is to allocate
 959         memory quickly.
 960
 961         * wtf/CMakeListsEfl.txt:
 962
 963 2010-10-04  Oliver Hunt  <oliver@apple.com>
 964
 965         Reviewed by Geoff Garen.
 966
 967         Lazily create activation objects
 968         https://bugs.webkit.org/show_bug.cgi?id=47107
 969
 970         Make it possible to lazily create the activation object
 971         for a function that needs one.  This allows us to reduce
 972         the overhead of entering a function that may require
 973         an activation in some cases, but not always.
 974
 975         This does make exception handling a little more complex as
 976         it's now necessary to verify that a callframes activation
 977         has been created, and create it if not, in all of the
 978         paths used in exception handling.
 979
 980         We also need to add logic to check for the existence of
 981         the activation in the scoped_var opcodes, as well as
 982         op_ret, op_ret_object_or_this and op_tearoff_activation
 983         so that we can avoid creating an activation unnecesarily
 984         on function exit.
 985
 986         * bytecode/CodeBlock.cpp:
 987         (JSC::CodeBlock::dump):
 988         (JSC::CodeBlock::reparseForExceptionInfoIfNecessary):
 989         (JSC::CodeBlock::createActivation):
 990         * bytecode/CodeBlock.h:
 991         (JSC::CodeBlock::setActivationRegister):
 992         (JSC::CodeBlock::activationRegister):
 993         * bytecode/Opcode.h:
 994         * bytecompiler/BytecodeGenerator.cpp:
 995         (JSC::BytecodeGenerator::BytecodeGenerator):
 996         (JSC::BytecodeGenerator::emitNewFunctionInternal):
 997         (JSC::BytecodeGenerator::emitNewFunctionExpression):
 998         (JSC::BytecodeGenerator::createActivationIfNecessary):
 999         * bytecompiler/BytecodeGenerator.h:
1000         * interpreter/Interpreter.cpp:
1001         (JSC::Interpreter::resolveSkip):
1002         (JSC::Interpreter::resolveGlobalDynamic):
1003         (JSC::Interpreter::resolveBase):
1004         (JSC::Interpreter::unwindCallFrame):
1005         (JSC::Interpreter::throwException):
1006         (JSC::Interpreter::privateExecute):
1007         * jit/JIT.cpp:
1008         (JSC::JIT::privateCompileMainPass):
1009         * jit/JIT.h:
1010         * jit/JITCall32_64.cpp:
1011         (JSC::JIT::emit_op_ret):
1012         (JSC::JIT::emit_op_ret_object_or_this):
1013         * jit/JITOpcodes.cpp:
1014         (JSC::JIT::emit_op_end):
1015         (JSC::JIT::emit_op_get_scoped_var):
1016         (JSC::JIT::emit_op_put_scoped_var):
1017         (JSC::JIT::emit_op_tear_off_activation):
1018         (JSC::JIT::emit_op_ret):
1019         (JSC::JIT::emit_op_ret_object_or_this):
1020         (JSC::JIT::emit_op_create_activation):
1021         (JSC::JIT::emit_op_resolve_global_dynamic):
1022         * jit/JITOpcodes32_64.cpp:
1023         (JSC::JIT::emit_op_get_scoped_var):
1024         (JSC::JIT::emit_op_put_scoped_var):
1025         (JSC::JIT::emit_op_tear_off_activation):
1026         (JSC::JIT::emit_op_create_activation):
1027         * jit/JITStubs.cpp:
1028         (JSC::DEFINE_STUB_FUNCTION):
1029
1030 2010-10-04  Adam Barth  <abarth@webkit.org>
1031
1032         Reviewed by Sam Weinig.
1033
1034         Remove ENABLE_SANDBOX
1035         https://bugs.webkit.org/show_bug.cgi?id=47032
1036
1037         * Configurations/FeatureDefines.xcconfig:
1038
1039 2010-10-01  Pratik Solanki  <psolanki@apple.com>
1040
1041         Reviewed by Geoffrey Garen.
1042         Specify ALWAYS_INLINE at function declaration not function definition
1043         https://bugs.webkit.org/show_bug.cgi?id=46960
1044
1045         For functions defined with ALWAYS_INLINE, add the attribute to the declaration as well.
1046
1047         * bytecompiler/BytecodeGenerator.h:
1048         * wtf/FastMalloc.cpp:
1049
1050 2010-10-01  Kwang Yul Seo  <skyul@company100.net>
1051
1052         Unreviewed.
1053
1054         [BREWMP] Change Collector BLOCK_SIZE to 64KB
1055         https://bugs.webkit.org/show_bug.cgi?id=46436
1056
1057         Lower BLOCK_SIZE to 64KB because Brew MP runs on low end devices.
1058
1059         * runtime/Collector.h:
1060
1061 2010-10-01  Viatcheslav Ostapenko  <ostapenko.viatcheslav@nokia.com>
1062
1063         Reviewed by Andreas Kling.
1064
1065         [Qt] Stack overflow on symbian platform.
1066         https://bugs.webkit.org/show_bug.cgi?id=40598
1067
1068         Move big allocation in arrayProtoFuncToString from stack to heap.
1069         JSC::arrayProtoFuncToString function can be called recursivly and
1070         1K allocation on stack cahse stack overflow.
1071         Can be useful for other platforms with limited stack size.
1072
1073         * runtime/ArrayPrototype.cpp:
1074         (JSC::arrayProtoFuncToString):
1075
1076 2010-09-30  Kwang Yul Seo  <skyul@company100.net>
1077
1078         Reviewed by Kent Tamura.
1079
1080         [BREWMP] Add a factory function which returns an instance wrapped in PlatformRefPtr.
1081         https://bugs.webkit.org/show_bug.cgi?id=46373
1082
1083         A Brew MP instance has reference count 1 when it is created, so call adoptPlatformRef
1084         to wrap the instance in PlatformRefPtr.
1085
1086         * wtf/brew/ShellBrew.h:
1087         (WTF::createRefPtrInstance):
1088
1089 2010-09-30  Kwang Yul Seo  <skyul@company100.net>
1090
1091         Reviewed by Kent Tamura.
1092
1093         [BREWMP] Port PlatformRefPtr
1094         https://bugs.webkit.org/show_bug.cgi?id=46370
1095
1096         Implement refPlatformPtr and derefPlatformPtr to use PlatformRefPtr in Brew MP.
1097
1098         * wtf/brew/RefPtrBrew.h: Added.
1099         (WTF::refPlatformPtr):
1100         (WTF::derefPlatformPtr):
1101
1102 2010-09-29  Sam Weinig  <sam@webkit.org>
1103
1104         Reviewed by Darin Adler.
1105
1106         Add additional checks to StringBuffer.
1107         <rdar://problem/7756381>
1108
1109         * wtf/text/StringBuffer.h:
1110         (WTF::StringBuffer::StringBuffer):
1111         (WTF::StringBuffer::resize):
1112
1113 2010-09-30  Chris Marrin  <cmarrin@apple.com>
1114
1115         Reviewed by Simon Fraser.
1116
1117         Make 2D accelerated canvas rendering build on Mac
1118         https://bugs.webkit.org/show_bug.cgi?id=46007
1119
1120         Added ACCELERATED_2D_CANVAS to FeatureDefines
1121
1122         * Configurations/FeatureDefines.xcconfig:
1123
1124 2010-09-30  Kevin Ollivier  <kevino@theolliviers.com>
1125
1126         [wx] wxMSW build fix. Make sure we copy the compiler flags and remove exception handling from
1127         the copy so as not to alter global settings.
1128
1129         * wscript:
1130
1131 2010-09-30  Peter Varga  <pvarga@inf.u-szeged.hu>
1132
1133         Reviewed by Gavin Barraclough.
1134
1135         The case-insensitivity backreference checking isn't working with YARR
1136         Interpreter
1137         https://bugs.webkit.org/show_bug.cgi?id=46882
1138
1139         Add ignorecase checking to the Interpreter::tryConsumeBackReference() function.
1140
1141         * yarr/RegexInterpreter.cpp:
1142         (JSC::Yarr::Interpreter::tryConsumeBackReference):
1143
1144 2010-09-30  Kwang Yul Seo  <skyul@company100.net>
1145
1146         Reviewed by Andreas Kling.
1147
1148         [BREWMP] Leave initializeRandomNumberGenerator empty.
1149         https://bugs.webkit.org/show_bug.cgi?id=46851
1150
1151         On Brew MP, AEECLSID_RANDOM initializes itself.
1152
1153         * wtf/RandomNumberSeed.h:
1154         (WTF::initializeRandomNumberGenerator):
1155
1156 2010-09-30  Gabor Loki  <loki@webkit.org>
1157
1158         Reviewed by Csaba Osztrogonác.
1159
1160         Remove unnecessary cacheFlush calls from Thumb-2
1161         https://bugs.webkit.org/show_bug.cgi?id=46702
1162
1163         * assembler/ARMv7Assembler.h:
1164         (JSC::ARMv7Assembler::relinkCall):
1165         (JSC::ARMv7Assembler::repatchInt32):
1166         (JSC::ARMv7Assembler::repatchPointer):
1167
1168 2010-09-29  Patrick Gansterer  <paroga@webkit.org>
1169
1170         Unreviewed.
1171
1172         Next try to fix cygwin build.
1173
1174         * wtf/Assertions.cpp:
1175
1176 2010-09-29  Patrick Gansterer  <paroga@webkit.org>
1177
1178         Unreviewed.
1179
1180         Build fix for cygwin #2. It's OS(WINDOWS), not OS(WIN).
1181
1182         * wtf/Assertions.cpp:
1183
1184 2010-09-29  Patrick Gansterer  <paroga@webkit.org>
1185
1186         Unreviewed.
1187
1188         Build fix for cygwin.
1189
1190         * wtf/Assertions.cpp:
1191
1192 2010-09-29  Patrick Gansterer  <paroga@webkit.org>
1193
1194         Reviewed by Andreas Kling.
1195
1196         [WINCE] Buildfix for Assertions.cpp after r68511.
1197         https://bugs.webkit.org/show_bug.cgi?id=46807
1198
1199         Some, but not all WinCE environments have support for IsDebuggerPresent().
1200         Add HAVE(ISDEBUGGERPRESENT) to make this a build option.
1201         HAVE(ISDEBUGGERPRESENT) will be 1 for all OS(WIN) by default.
1202
1203         * wtf/Assertions.cpp:
1204         * wtf/Platform.h:
1205
1206 2010-09-29  Peter Varga  <pvarga@inf.u-szeged.hu>
1207
1208         Reviewed by Csaba Osztrogonác.
1209
1210         JSC compile fails on 32bit platform when Regexp Tracing is enabled
1211         https://bugs.webkit.org/show_bug.cgi?id=46713
1212
1213         Fix the cast of pointer in regexp tracing to avoid the warning.
1214
1215         * runtime/RegExp.cpp:
1216         (JSC::RegExp::match):
1217
1218 2010-09-28  Anders Carlsson  <andersca@apple.com>
1219
1220         Reviewed by Sam Weinig.
1221
1222         Begin hooking up painting in the plug-in process
1223         https://bugs.webkit.org/show_bug.cgi?id=46766
1224
1225         * JavaScriptCore.exp:
1226         Add tryFastRealloc, used by WebKit2.
1227
1228 2010-09-28  Philippe Normand  <pnormand@igalia.com>
1229
1230         Reviewed by Martin Robinson.
1231
1232         Guard GRefPtr/GOwnPtr files with ENABLE(GLIB_SUPPORT)
1233         https://bugs.webkit.org/show_bug.cgi?id=46721
1234
1235         Enable GOwnPtr/GRefPtr build only if glib support has been
1236         explicitly enabled using the WTF_ENABLE_GLIB_SUPPORT macro.
1237
1238         * wtf/gobject/GOwnPtr.cpp:
1239         * wtf/gobject/GOwnPtr.h:
1240         * wtf/gobject/GRefPtr.cpp:
1241         * wtf/gobject/GRefPtr.h:
1242
1243 2010-09-28  İsmail Dönmez  <ismail@namtrac.org>
1244
1245         Reviewed by Andreas Kling.
1246
1247         Test for WINCE instead of WINCEBASIC, compiler always defines WINCE.
1248         Remove reference to unexisting path JavaScriptCore/os-wince.
1249
1250         * JavaScriptCore.pri:
1251         * wtf/Assertions.cpp:
1252
1253 2010-09-27  Michael Saboff  <msaboff@apple.com>
1254
1255         Reviewed by Geoffrey Garen.
1256
1257         Changed the initialization of JSArray objects to have space for
1258         3 elements for the constructor that takes a ArgList argument.
1259         This improves v8-deltablue performance by about 2.8% by reducing
1260         the number of realloc() calls.
1261         https://bugs.webkit.org/show_bug.cgi?id=46664
1262
1263         * runtime/JSArray.cpp:
1264         (JSC::JSArray::JSArray):
1265
1266 2010-09-27  Gavin Barraclough  <barraclough@apple.com>
1267
1268         Reviewed by Darin Adler.
1269
1270         Bug 46680 - Inlining string concatenation can regress interpreter performance
1271         <rdar://problem/8362752> REGRESSION: ~6.4% sunspider regression in interpreter
1272         Do not inline calls to string concatenation in the interpret loop.
1273
1274         * interpreter/Interpreter.cpp:
1275         (JSC::concatenateStrings):
1276         (JSC::Interpreter::privateExecute):
1277
1278 2010-09-27  Anders Carlsson  <andersca@apple.com>
1279
1280         Fix thinko.
1281
1282         * runtime/JSCell.h:
1283
1284 2010-09-27  Anders Carlsson  <andersca@apple.com>
1285
1286         Reviewed by Adam Roben.
1287
1288         Try to fix Windows build.
1289
1290         * runtime/JSCell.h:
1291         (JSC::MSVCBugWorkaround::MSVCBugWorkaround):
1292         (JSC::MSVCBugWorkaround::~MSVCBugWorkaround):
1293
1294 2010-09-27  Erik Arvidsson  <arv@chromium.org>
1295
1296         Reviewed by Darin Adler.
1297
1298         Add operator == for AtomicString and Vector<Uchar>
1299         https://bugs.webkit.org/show_bug.cgi?id=46509
1300
1301         * JavaScriptCore.exp:
1302         * wtf/text/AtomicString.cpp:
1303         (WTF::operator==):
1304         * wtf/text/AtomicString.h:
1305         (WTF::operator==):
1306         (WTF::operator!=):
1307
1308 2010-09-27  Anders Carlsson  <andersca@apple.com>
1309
1310         Try to fix the Windows build.
1311
1312         * wtf/Noncopyable.h:
1313
1314 2010-09-26  Anders Carlsson  <andersca@apple.com>
1315
1316         Reviewed by Alexey Proskuryakov and Adam Barth.
1317
1318         Add WTF_MAKE_NONCOPYABLE macro
1319         https://bugs.webkit.org/show_bug.cgi?id=46589
1320
1321         Going forward, we'd like to get rid of the Noncopyable and FastAllocBase classes. The
1322         reason for this is that the Itanium C++ ABI states that no empty classes of the same type
1323         can be laid out at the same offset in the class. This can result in objects getting larger
1324         which leads to memory regressions. (One example of this is the String class which grew by
1325         sizeof(void*) when both its base class and its first member variable inherited indirectly
1326         from FastAllocBase).
1327
1328         * wtf/Noncopyable.h:
1329         Add a WTF_MAKE_NONCOPYABLE macro and get rid of NoncopyableCustomAllocated.
1330
1331         * runtime/JSCell.h:
1332         * wtf/RefCounted.h:
1333         Don't inherit from NoncopyableCustomAllocated. Instead, use WTF_MAKE_NONCOPYABLE.
1334
1335 2010-09-27  Philippe Normand  <pnormand@igalia.com>
1336
1337         Reviewed by Martin Robinson.
1338
1339         [GTK] use ENABLE(GLIB_SUPPORT)
1340         https://bugs.webkit.org/show_bug.cgi?id=46630
1341
1342         * wtf/Platform.h: Include GTypedefs.h only if glib support
1343         is explicitly enabled.
1344
1345 2010-09-25  Holger Hans Peter Freyther  <holger@moiji-mobile.com>
1346
1347         Reviewed by Adam Barth.
1348
1349         jsc: Document the strcat opcode.
1350         https://bugs.webkit.org/show_bug.cgi?id=46571
1351
1352         * interpreter/Interpreter.cpp:
1353         (JSC::Interpreter::privateExecute):
1354
1355 2010-09-21  Holger Hans Peter Freyther  <holger@moiji-mobile.com>
1356
1357         Reviewed by Adam Barth.
1358
1359         make-bytecode-docs.pl: Add a comment to the generated HTML
1360         https://bugs.webkit.org/show_bug.cgi?id=46570
1361
1362         Generate an HTML Comment that this file was generated from
1363         Interpreter.cpp with the make-bytecode-docs.pl script.
1364
1365         * docs/make-bytecode-docs.pl:
1366
1367 2010-09-27  Patrick Gansterer  <paroga@webkit.org>
1368
1369         Reviewed by Adam Barth.
1370
1371         Remove WTF::stringHash functions
1372         https://bugs.webkit.org/show_bug.cgi?id=46520
1373
1374         Since r68289 the stringHash functions are only wrappers around StringHasher::createHash.
1375         So use StringHasher::createHash directly and remove stringHash.
1376
1377         * wtf/StringHashFunctions.h:
1378         * wtf/text/StringImpl.h:
1379         (WTF::StringImpl::computeHash): Use WTF::StringHasher::createHash directly.
1380
1381 2010-09-26  Patrick Gansterer  <paroga@webkit.org>
1382
1383         Reviewed by Adam Barth.
1384
1385         Add WTF::StringHasher::createBlobHash
1386         https://bugs.webkit.org/show_bug.cgi?id=46514
1387
1388         Add this function for hashing FormElementKey and QualifiedNameComponents.
1389
1390         * wtf/StringHashFunctions.h:
1391         (WTF::StringHasher::createBlobHash):
1392
1393 2010-09-26  Patrick Gansterer  <paroga@webkit.org>
1394
1395         Reviewed by Adam Barth.
1396
1397         REGRESSION (r68289): Assertion failure in StringHasher::addCharacter() (ch != invalidCharacterValue)
1398         running websocket/tests/bad-sub-protocol-non-ascii.html
1399         https://bugs.webkit.org/show_bug.cgi?id=46553
1400
1401         Because we use StringHasher for binary data too, so the check for invalid unicode input is wrong.
1402         Add an additional member variable to indicate if we have an pending character
1403         instead of only using an invalid character for this purpose.
1404
1405         * wtf/StringHashFunctions.h:
1406         (WTF::StringHasher::StringHasher):
1407         (WTF::StringHasher::addCharacters):
1408         (WTF::StringHasher::addCharacter):
1409         (WTF::StringHasher::hash):
1410
1411 2010-09-26  Mark Hahnenberg  <mhahnenb@gmail.com>
1412
1413         Reviewed by Oliver Hunt.
1414
1415         valueOf called in wrong order in atan2 and date constructors.
1416         https://bugs.webkit.org/show_bug.cgi?id=26978
1417
1418         Fixed the bug where the arguments to atan2 were being evaluated
1419         out of order.
1420
1421         * runtime/MathObject.cpp:
1422         (JSC::mathProtoFuncATan2):
1423
1424 2010-09-26  Mark Hahnenberg  <mhahnenb@gmail.com>
1425
1426         Reviewed by Oliver Hunt.
1427
1428         valueOf called in wrong order in atan2 and date constructors.
1429         https://bugs.webkit.org/show_bug.cgi?id=26978
1430
1431         Fixed the issue where the parameters to the Date constructor
1432         were being evaluated to numbers more than once.
1433
1434         * runtime/DateConstructor.cpp:
1435         (JSC::constructDate):
1436         (JSC::dateUTC):
1437
1438 2010-09-25  Oliver Hunt  <oliver@apple.com>
1439
1440         Fix various builds
1441
1442         Relearning the lesson that last minute changes are bad.
1443
1444         * bytecode/CodeBlock.cpp:
1445         (JSC::CodeBlock::dump):
1446         * bytecompiler/BytecodeGenerator.cpp:
1447         (JSC::BytecodeGenerator::emitGetArgumentsLength):
1448         * jit/JITOpcodes.cpp:
1449         (JSC::JIT::emitSlow_op_get_argument_by_val):
1450
1451 2010-09-25  Oliver Hunt  <oliver@apple.com>
1452
1453         Reviewed by Cameron Zwarich.
1454
1455         Avoid constructing arguments object when accessing length and index properties
1456         https://bugs.webkit.org/show_bug.cgi?id=46572
1457
1458         Add opcodes to read argument length and properties, and then implement them.
1459         Much like other lazy opcodes these opcodes take a fast path when the arguments
1460         object has not been instantiated, and fall back on generic access mechanisms
1461         if they are acting on an instantiated object.
1462
1463         3% win on v8-earleyboyer, no change elsewhere.
1464
1465         * bytecode/CodeBlock.cpp:
1466         (JSC::CodeBlock::dump):
1467         * bytecode/Opcode.h:
1468         * bytecompiler/BytecodeGenerator.cpp:
1469         (JSC::BytecodeGenerator::emitGetArgumentsLength):
1470         (JSC::BytecodeGenerator::emitGetArgumentByVal):
1471         * bytecompiler/BytecodeGenerator.h:
1472         * bytecompiler/NodesCodegen.cpp:
1473         (JSC::BracketAccessorNode::emitBytecode):
1474         (JSC::DotAccessorNode::emitBytecode):
1475         * interpreter/Interpreter.cpp:
1476         (JSC::Interpreter::privateExecute):
1477         * jit/JIT.cpp:
1478         (JSC::JIT::privateCompileMainPass):
1479         (JSC::JIT::privateCompileSlowCases):
1480         * jit/JIT.h:
1481         * jit/JITOpcodes.cpp:
1482         (JSC::JIT::emit_op_get_arguments_length):
1483         (JSC::JIT::emitSlow_op_get_arguments_length):
1484         (JSC::JIT::emit_op_get_argument_by_val):
1485         (JSC::JIT::emitSlow_op_get_argument_by_val):
1486         * jit/JITOpcodes32_64.cpp:
1487         (JSC::JIT::emit_op_get_arguments_length):
1488         (JSC::JIT::emitSlow_op_get_arguments_length):
1489         (JSC::JIT::emit_op_get_argument_by_val):
1490         (JSC::JIT::emitSlow_op_get_argument_by_val):
1491
1492 2010-09-25  Patrick Gansterer  <paroga@webkit.org>
1493
1494         Unreviewed.
1495
1496         Fix typo in StringHasher class
1497         https://bugs.webkit.org/show_bug.cgi?id=45970
1498
1499         * wtf/StringHashFunctions.h:
1500         (WTF::StringHasher::createHash):
1501
1502 2010-09-24  Patrick Gansterer  <paroga@paroga.com>
1503
1504         Reviewed by Gavin Barraclough.
1505
1506         Add WTF::StringHasher
1507         https://bugs.webkit.org/show_bug.cgi?id=45970
1508
1509         StringHasher is a class for calculation stringHash out of character string.
1510         This class will unify the different usages of the same algorithm.
1511
1512         * wtf/StringHashFunctions.h:
1513         (WTF::StringHasher::StringHasher):
1514         (WTF::StringHasher::addCharacters):
1515         (WTF::StringHasher::addCharacter):
1516         (WTF::StringHasher::hash):
1517         (WTF::StringHasher::createHash):
1518         (WTF::StringHasher::defaultCoverter):
1519         (WTF::StringHasher::addCharactersToHash):
1520         (WTF::stringHash):
1521
1522 2010-09-24  Oliver Hunt  <oliver@apple.com>
1523
1524         Reviewed by Geoffrey Garen.
1525
1526         Variable declarations inside a catch scope don't get propogated to the parent scope
1527         https://bugs.webkit.org/show_bug.cgi?id=46501
1528
1529         Add logic to make variable declaration look for a scope for the
1530         new variable.  This allows us to create a scope (eg. for catch)
1531         and then seal it, so that additional variable declarations
1532         contained are propogated to the correct target.  Strangely this
1533         comes out as a performance win, but I think it's mostly cache
1534         effects.
1535
1536         * parser/JSParser.cpp:
1537         (JSC::JSParser::Scope::Scope):
1538         (JSC::JSParser::Scope::preventNewDecls):
1539         (JSC::JSParser::Scope::allowsNewDecls):
1540         (JSC::JSParser::declareVariable):
1541         (JSC::JSParser::parseVarDeclarationList):
1542         (JSC::JSParser::parseConstDeclarationList):
1543         (JSC::JSParser::parseTryStatement):
1544         (JSC::JSParser::parseFormalParameters):
1545         (JSC::JSParser::parseFunctionDeclaration):
1546
1547 2010-09-24  İsmail Dönmez  <ismail@namtrac.org>
1548
1549        Reviewed by Csaba Osztrogonác.
1550
1551        Add a Windows compatible inttypes.h header to fix WinCE build.
1552        https://bugs.webkit.org/show_bug.cgi?id=46463
1553
1554        * os-win32/inttypes.h: Added.
1555
1556 2010-09-24  Oliver Hunt  <oliver@apple.com>
1557
1558         Reviewed by Gavin Barraclough.
1559
1560         REGRESSION(r68223): It broke 2-3 tests on bots (Requested by Ossy on #webkit).
1561         https://bugs.webkit.org/show_bug.cgi?id=46448
1562
1563         Roll this back in, with additional logic to prevent us from delaying construction
1564         of functions named "arguments"
1565
1566         * bytecode/CodeBlock.cpp:
1567         (JSC::CodeBlock::dump):
1568         * bytecode/Opcode.h:
1569         * bytecompiler/BytecodeGenerator.cpp:
1570         (JSC::BytecodeGenerator::BytecodeGenerator):
1571         (JSC::BytecodeGenerator::emitInitLazyRegister):
1572         (JSC::BytecodeGenerator::registerFor):
1573         (JSC::BytecodeGenerator::createLazyRegisterIfNecessary):
1574         (JSC::BytecodeGenerator::constRegisterFor):
1575         (JSC::BytecodeGenerator::emitNewFunction):
1576         (JSC::BytecodeGenerator::emitLazyNewFunction):
1577         (JSC::BytecodeGenerator::emitNewFunctionInternal):
1578         * bytecompiler/BytecodeGenerator.h:
1579         * interpreter/Interpreter.cpp:
1580         (JSC::Interpreter::privateExecute):
1581         * jit/JIT.cpp:
1582         (JSC::JIT::privateCompileMainPass):
1583         * jit/JIT.h:
1584         * jit/JITOpcodes.cpp:
1585         (JSC::JIT::emit_op_init_lazy_reg):
1586         (JSC::JIT::emit_op_new_func):
1587         * jit/JITOpcodes32_64.cpp:
1588         (JSC::JIT::emit_op_init_lazy_reg):
1589         * parser/Nodes.h:
1590         (JSC::ScopeNode::needsActivationForMoreThanVariables):
1591
1592 2010-09-23  Sheriff Bot  <webkit.review.bot@gmail.com>
1593
1594         Unreviewed, rolling out r68223.
1595         http://trac.webkit.org/changeset/68223
1596         https://bugs.webkit.org/show_bug.cgi?id=46448
1597
1598         It broke 2-3 tests on bots (Requested by Ossy on #webkit).
1599
1600         * bytecode/CodeBlock.cpp:
1601         (JSC::CodeBlock::dump):
1602         * bytecode/Opcode.h:
1603         * bytecompiler/BytecodeGenerator.cpp:
1604         (JSC::BytecodeGenerator::BytecodeGenerator):
1605         (JSC::BytecodeGenerator::registerFor):
1606         (JSC::BytecodeGenerator::constRegisterFor):
1607         (JSC::BytecodeGenerator::emitNewFunction):
1608         * bytecompiler/BytecodeGenerator.h:
1609         * interpreter/Interpreter.cpp:
1610         (JSC::Interpreter::privateExecute):
1611         * jit/JIT.cpp:
1612         (JSC::JIT::privateCompileMainPass):
1613         * jit/JIT.h:
1614         * jit/JITOpcodes.cpp:
1615         (JSC::JIT::emit_op_new_func):
1616         (JSC::JIT::emit_op_init_arguments):
1617         * jit/JITOpcodes32_64.cpp:
1618         (JSC::JIT::emit_op_new_func):
1619         (JSC::JIT::emit_op_init_arguments):
1620         * parser/Nodes.h:
1621
1622 2010-09-23  Oliver Hunt  <oliver@apple.com>
1623
1624         Reviewed by Geoffrey Garen.
1625
1626         Delay construction of functions that aren't captured
1627         https://bugs.webkit.org/show_bug.cgi?id=46433
1628
1629         If a function isn't captured by an activation there's no
1630         way it can be accessed indirectly, so we can delay the
1631         construction until it's used (similar to what we do with
1632         arguments).  We rename the existing op_init_arguments to
1633         op_init_lazy_reg and removed its implicit handling of
1634         the anonymous argument register, and make op_new_function
1635         take a parameter to indicate whether it should null check
1636         the target slot before creating the function object.
1637
1638         * bytecode/CodeBlock.cpp:
1639         (JSC::CodeBlock::dump):
1640         * bytecode/Opcode.h:
1641         * bytecompiler/BytecodeGenerator.cpp:
1642         (JSC::BytecodeGenerator::BytecodeGenerator):
1643         (JSC::BytecodeGenerator::emitInitLazyRegister):
1644         (JSC::BytecodeGenerator::registerFor):
1645         (JSC::BytecodeGenerator::createLazyRegisterIfNecessary):
1646         (JSC::BytecodeGenerator::constRegisterFor):
1647         (JSC::BytecodeGenerator::emitNewFunction):
1648         (JSC::BytecodeGenerator::emitLazyNewFunction):
1649         (JSC::BytecodeGenerator::emitNewFunctionInternal):
1650         * bytecompiler/BytecodeGenerator.h:
1651         * interpreter/Interpreter.cpp:
1652         (JSC::Interpreter::privateExecute):
1653         * jit/JIT.cpp:
1654         (JSC::JIT::privateCompileMainPass):
1655         * jit/JIT.h:
1656         * jit/JITOpcodes.cpp:
1657         (JSC::JIT::emit_op_init_lazy_reg):
1658         (JSC::JIT::emit_op_new_func):
1659         * jit/JITOpcodes32_64.cpp:
1660         (JSC::JIT::emit_op_init_lazy_reg):
1661         * parser/Nodes.h:
1662         (JSC::ScopeNode::needsActivationForMoreThanVariables):
1663
1664 2010-09-23  David Kilzer  <ddkilzer@apple.com>
1665
1666         <rdar://problem/8460731> ~9.9% speedup when compiling interpreter with llvm-gcc-4.2
1667         https://bugs.webkit.org/show_bug.cgi?id=46423
1668
1669         Reviewed by Oliver Hunt.
1670
1671         * interpreter/Interpreter.cpp:
1672         (JSC::Interpreter::privateExecute): Disable the gcc computed
1673         goto hacks added in r55564 when compiling with llvm-gcc-4.2.
1674
1675 2010-09-23  Lucas De Marchi  <lucas.demarchi@profusion.mobi>
1676
1677         Reviewed by Darin Adler.
1678
1679         Fix usage of enum as if it was a define
1680         https://bugs.webkit.org/show_bug.cgi?id=46355
1681
1682         pthread.h defines PTHREAD_MUTEX_DEFAULT and PTHREAD_MUTEX_NORMAL as an
1683         enum.  Hence, it cannot be used by the preprocessor which always
1684         evaluates that condition as true. This was giving a warning when
1685         compiling with gcc and "-Wundef" flag.
1686
1687         The second path, when PTHREAD_MUTEX_DEFAULT is not the same of
1688         PTHREAD_MUTEX_NORMAL, is not slow. So, let's eliminate the first path
1689         and get rid of that #if.
1690
1691         * wtf/ThreadingPthreads.cpp: Always call pthread_mutexattr_init() to
1692         set mutex type to PTHREAD_MUTEX_NORMAL.
1693         (WTF::Mutex::Mutex):
1694
1695 2010-09-23  Michael Saboff  <msaboff@apple.com>
1696
1697         Reviewed by Geoffrey Garen.
1698
1699         Removed extraneous truncation of ovector on entry and error exit.
1700         Changed the initialization to -1 of vector to only initialize
1701         the start indecies, which is sufficient for the pattern/subpatterns.
1702         Changed the JIT code to not clear the end index for subpatterns
1703         as it isn't needed.  These changes are worth ~2.7% on v8-regexp.
1704         https://bugs.webkit.org/show_bug.cgi?id=46404
1705
1706         * runtime/RegExp.cpp:
1707         (JSC::RegExp::match):
1708         * yarr/RegexJIT.cpp:
1709         (JSC::Yarr::RegexGenerator::generateParenthesesSingle):
1710
1711 2010-09-22  Oliver Hunt  <oliver@apple.com>
1712
1713         Reviewed by Geoff Garen.
1714
1715         Only copy captured variables into activation
1716         https://bugs.webkit.org/show_bug.cgi?id=46330
1717
1718         We now track free variable information which means that
1719         we no longer need to copy every variable defined in a
1720         function.  With this patch activations only retain those
1721         variables needed for correctness.  In order to interact
1722         safely with the inspector this means that JSActivation
1723         now provides its own lookup functions so it can avoid
1724         trying to read or write to variables that have been
1725         optimised out.
1726
1727         * bytecode/CodeBlock.h:
1728         * bytecompiler/BytecodeGenerator.cpp:
1729         (JSC::BytecodeGenerator::BytecodeGenerator):
1730         * parser/Nodes.h:
1731         (JSC::ScopeNode::capturedVariableCount):
1732         (JSC::ScopeNode::captures):
1733         * runtime/Arguments.h:
1734         (JSC::JSActivation::copyRegisters):
1735         * runtime/Executable.cpp:
1736         (JSC::FunctionExecutable::FunctionExecutable):
1737         (JSC::FunctionExecutable::compileForCallInternal):
1738         (JSC::FunctionExecutable::compileForConstructInternal):
1739         * runtime/Executable.h:
1740         (JSC::FunctionExecutable::capturedVariableCount):
1741         * runtime/JSActivation.cpp:
1742         (JSC::JSActivation::markChildren):
1743         (JSC::JSActivation::symbolTableGet):
1744         (JSC::JSActivation::symbolTablePut):
1745         (JSC::JSActivation::getOwnPropertyNames):
1746         (JSC::JSActivation::symbolTablePutWithAttributes):
1747         * runtime/JSActivation.h:
1748
1749 2010-09-23  Ismail Donmez  <ismail@namtrac.org>
1750
1751         Reviewed by Andreas Kling.
1752
1753         Fix jsc.exe build for Windows CE
1754
1755         * jsc.pro: Add mmtimer.lib for Windows CE.
1756
1757 2010-09-23  Ismail Donmez  <ismail@namtrac.org>
1758
1759         Unreviewed.
1760
1761         JIT should be disabled on Windows CE. Broken in r64176.
1762
1763         * wtf/Platform.h:
1764
1765 2010-09-23  Peter Varga  <pvarga@inf.u-szeged.hu>
1766
1767         Reviewed by Gavin Barraclough.
1768
1769         Reduce the number of BOL checks in YARR Interpreter
1770         https://bugs.webkit.org/show_bug.cgi?id=46260
1771
1772         Extend the YARR Interpreter with an optimization which reduces the number of
1773         BOL assertion checks. If a "TypeBodyAlternative" byteTerm is followed by a
1774         "TypeAssertionBOL" byteTerm it will be checked just one time.
1775
1776         * yarr/RegexInterpreter.cpp:
1777         (JSC::Yarr::Interpreter::matchDisjunction):
1778         (JSC::Yarr::ByteCompiler::compile):
1779         (JSC::Yarr::ByteCompiler::regexBegin):
1780         (JSC::Yarr::ByteCompiler::alternativeBodyDisjunction):
1781         (JSC::Yarr::ByteCompiler::emitDisjunction):
1782         * yarr/RegexInterpreter.h:
1783         (JSC::Yarr::ByteTerm::BodyAlternativeBegin):
1784         (JSC::Yarr::ByteTerm::BodyAlternativeDisjunction):
1785         (JSC::Yarr::ByteTerm::BodyAlternativeEnd):
1786         (JSC::Yarr::ByteTerm::AlternativeBegin):
1787         (JSC::Yarr::ByteTerm::AlternativeDisjunction):
1788         (JSC::Yarr::ByteTerm::AlternativeEnd):
1789
1790 2010-09-22  Michael Saboff  <msaboff@apple.com>
1791
1792         Reviewed by Gavin Barraclough.
1793
1794         Fixed the cross over from alternatives executed once and
1795         those that loop.  This fixed the problem where the index
1796         was getting messed up for looping alternatives causing an
1797         infinite loop.
1798         https://bugs.webkit.org/show_bug.cgi?id=46189
1799
1800         * yarr/RegexJIT.cpp:
1801         (JSC::Yarr::RegexGenerator::generateDisjunction):
1802
1803 2010-09-22  Steve Falkenburg  <sfalken@apple.com>
1804
1805         Rubber stamped by Jon Honeycutt.
1806
1807         Allow jsc.exe to be run against unversioned ICU.
1808
1809         * JavaScriptCore.vcproj/jsc/jscCommon.vsprops:
1810
1811 2010-09-22  Kwang Yul Seo  <skyul@company100.net>
1812
1813         Reviewed by Laszlo Gombos.
1814
1815         Use "typedef wchar_t JSChar" when compiled with RVCT
1816         https://bugs.webkit.org/show_bug.cgi?id=40651
1817
1818         Use wchar_t for JSChar and UChar when compiled with RVCT.
1819         Linux is the exception for this rule.
1820
1821         * API/JSStringRef.h:
1822         * wtf/unicode/qt4/UnicodeQt4.h:
1823
1824 2010-09-22  Oliver Hunt  <oliver@apple.com>
1825
1826         Reviewed by Gavin Barraclough.
1827
1828         [INTERPRETER] Two tests fail with SputnikError: #1.1: if argArray is neither an array nor an arguments object (see 10.1.8), a TypeError exception is thrown
1829         https://bugs.webkit.org/show_bug.cgi?id=44245
1830
1831         Remove incorrect code from op_load_varargs in the interpreter.
1832
1833         * interpreter/Interpreter.cpp:
1834         (JSC::Interpreter::privateExecute):
1835
1836 2010-09-22  Oliver Hunt  <oliver@apple.com>
1837
1838         Reviewed by Gavin Barraclough.
1839
1840         [JIT] fast/js/sputnik/Conformance/15_Native_Objects/15.3_Function/15.3.5/S15.3.5.3_A2_T6.html fails
1841         https://bugs.webkit.org/show_bug.cgi?id=44246
1842
1843         JIT code generated for instanceof was not checking to ensure that the prototype property was
1844         an object, this patch ensures that it does.
1845
1846         * jit/JITOpcodes.cpp:
1847         (JSC::JIT::emit_op_instanceof):
1848         (JSC::JIT::emitSlow_op_instanceof):
1849         * jit/JITOpcodes32_64.cpp:
1850         (JSC::JIT::emit_op_instanceof):
1851         (JSC::JIT::emitSlow_op_instanceof):
1852
1853 2010-09-22  Patrick Gansterer  <paroga@webkit.org>
1854
1855         Reviewed by Darin Adler.
1856
1857         Inline UTF8SequenceLength
1858         https://bugs.webkit.org/show_bug.cgi?id=45589
1859
1860         * wtf/unicode/UTF8.cpp:
1861         (WTF::Unicode::convertUTF8ToUTF16): Use inline version of UTF8SequenceLength to improve performance.
1862
1863 2010-09-21  Oliver Hunt  <oliver@apple.com>
1864
1865         RS=Gavin Barraclough.
1866
1867         Fix codeblock dumping
1868
1869         * bytecode/CodeBlock.cpp:
1870         (JSC::CodeBlock::dump):
1871         * runtime/Executable.h:
1872         (JSC::ScriptExecutable::ScriptExecutable):
1873
1874 2010-09-21  Oliver Hunt  <oliver@apple.com>
1875
1876         Reviewed by Geoffrey Garen.
1877
1878         Speed up function.apply(..., arguments)
1879         https://bugs.webkit.org/show_bug.cgi?id=46207
1880
1881         Add code to do argument copying inline in the case
1882         where we're using Function.apply to forward our arguments
1883         directly.
1884
1885         * jit/JIT.cpp:
1886         (JSC::JIT::privateCompileSlowCases):
1887            Splitted op_load_varargs into fast and slow paths, so add the call
1888            to the slow path generator.
1889         * jit/JIT.h:
1890         * jit/JITCall32_64.cpp:
1891           Remove 32bit specific emit_op_load_varargs as the logic is the
1892           same for all value representations
1893         * jit/JITOpcodes.cpp:
1894         (JSC::JIT::emit_op_load_varargs):
1895           Copy arguments inline
1896         (JSC::JIT::emitSlow_op_load_varargs):
1897
1898 2010-09-21  Geoffrey Garen  <ggaren@apple.com>
1899
1900         Reviewed by Oliver Hunt.
1901
1902         <rdar://problem/8363003> REGRESSION: ~1.4% sunspider regression in
1903         interpreter due to 54724 and 54596
1904
1905         Fixed a typo (using "UNLIKELY" instead of "LIKELY").
1906
1907         * wtf/PassRefPtr.h:
1908         (WTF::refIfNotNull):
1909         (WTF::derefIfNotNull): It is likely that m_ptr != 0 because most RefPtrs
1910         hold real data. Also, in cases where they do not hold real data, the
1911         compiler usually sees a call to release() right before the call to the
1912         destructor, so it can probably optimize out the test completely.
1913
1914 2010-09-21  Fridrich Strba  <fridrich.strba@bluewin.ch>
1915
1916         Reviewed by Martin Robinson.
1917
1918         Build issues with Windows versions of the GTK+ port
1919         https://bugs.webkit.org/show_bug.cgi?id=45844
1920
1921         Link with winmm.dll when necessary and specify the executable extension
1922         explicitely so that the Programs/jsc-@WEBKITGTK_API_MAJOR_VERSION@
1923         rule actually works.
1924
1925         Don't try to build the ThreadSpecificWin.cpp since GTK+ port uses
1926         a section in ThreadSpecific.cpp
1927
1928         * GNUmakefile.am:
1929
1930 2010-09-21  Martin Robinson  <mrobinson@igalia.com>
1931
1932         Reviewed by Xan Lopez.
1933
1934         [GTK] 'make dist' should be fixed in preparation for the next release
1935         https://bugs.webkit.org/show_bug.cgi?id=46129
1936
1937         * GNUmakefile.am: Update the sources list to include missing headers.
1938
1939 2010-09-21  Dave Tapuska  <dtapuska@rim.com>
1940
1941         Reviewed by Csaba Osztrogonác.
1942
1943         https://bugs.webkit.org/show_bug.cgi?id=45673
1944
1945         r65596 caused ENABLE_PROFILER_REFERENCE_OFFSET to not be
1946         8 byte aligned. A non 8 byte divisible value for this will
1947         cause the sp to become non 8 byte aligned.
1948
1949         Verify and correct offset values that r65596 effected that
1950         weren't updated.
1951
1952         * jit/JITStubs.cpp:
1953         * jit/JITStubs.h:
1954
1955 2010-09-21  Xan Lopez  <xlopez@igalia.com>
1956
1957         Reviewed by Martin Robinson.
1958
1959         Fix Opcode stats compilation
1960         https://bugs.webkit.org/show_bug.cgi?id=46079
1961
1962         The FixedArray API had changed, and <stdio.h> was not included for
1963         printf.
1964
1965         * bytecode/Opcode.cpp:
1966         (JSC::OpcodeStats::~OpcodeStats):
1967
1968 2010-09-20  Michael Saboff  <msaboff@apple.com>
1969
1970         Reviewed by Gavin Barraclough.
1971
1972         Fixed detection of alternative smaller than the first alternative
1973         to only check looping alternatives.
1974         https://bugs.webkit.org/show_bug.cgi?id=46049
1975
1976         * yarr/RegexJIT.cpp:
1977         (JSC::Yarr::RegexGenerator::generateDisjunction):
1978
1979 2010-09-20  Peter Varga  <pvarga@inf.u-szeged.hu>
1980
1981         Reviewed by Geoffrey Garen.
1982
1983         REGRESSION(67790): jsc tests are failed with YARR interpreter
1984         https://bugs.webkit.org/show_bug.cgi?id=46083
1985
1986         Fix the initializing of the lastSubpatternId member of
1987         parentheses.
1988
1989         * yarr/RegexCompiler.cpp:
1990         (JSC::Yarr::RegexPatternConstructor::atomParenthesesEnd):
1991
1992 2010-09-20  Gavin Barraclough  <barraclough@apple.com>
1993
1994         Reviewed by Oliver Hunt.
1995
1996         Bug 46077 - ASSERT failure in YARR JIT
1997
1998         We will currently attempt to loop if there are multiple alternatives, they are all
1999         BOL predicated, and the last alternative is longer then the first - however if all
2000         alternatives are BOL predicated the head of loop label will not have been set, and
2001         we'll try to link a jump to an undefined label. Stop doing so.
2002
2003         * yarr/RegexJIT.cpp:
2004         (JSC::Yarr::RegexGenerator::generateDisjunction):
2005
2006 2010-09-20  Adam Roben  <aroben@apple.com>
2007
2008         Export RegExpObject::info from JavaScriptCore
2009
2010         This allows obj->inherits(&RegExpObject::info) to work correctly from
2011         outside JavaScriptCore.dll on Windows.
2012
2013         Fixes <http://webkit.org/b/46098>
2014         fast/loader/stateobjects/pushstate-object-types.html fails on Windows
2015
2016         Reviewed by John Sullivan.
2017
2018         * runtime/RegExpObject.h: Added JS_EXPORTDATA to the info member, as
2019         we already have for some other classes whose info members have to be
2020         used from outside the DLL.
2021
2022 2010-09-19  Gavin Barraclough  <barraclough@apple.com>
2023
2024         Windows build fix pt 2.
2025
2026         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def:
2027
2028 2010-09-19  Gavin Barraclough  <barraclough@apple.com>
2029
2030         Windows build fix pt 1.
2031
2032         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def:
2033
2034 2010-09-19  Gavin Barraclough  <barraclough@apple.com>
2035
2036         Build fix - implicit double-to-int conversion invalid on 32-bit.
2037
2038         * runtime/DatePrototype.cpp:
2039         (JSC::fillStructuresUsingDateArgs):
2040         (JSC::dateProtoFuncSetYear):
2041
2042 2010-09-19  Gavin Barraclough  <barraclough@apple.com>
2043
2044         Reviewed by Oliver Hunt.
2045
2046         Bug 46065 - Unify implementation of ToInt32 and ToUInt32, don't use fmod.
2047
2048         These methods implement the same conversion (see discussion in the notes
2049         of sections of 9.5 and 9.6 of the spec), only differing in how the result
2050         is interpretted.
2051
2052         Date prototype is incorrectly using toInt32, and this is causing us to
2053         provide an output value indicating whether the input to ToInt32 was finite
2054         (the corresponding methods on Date are actually spec'ed to use ToInteger,
2055         not ToInt32).  This patch partially fixes this in order to remove this
2056         bogus output value, hoewever more work will be require to bring Date
2057         fully up to spec compliance (the constructor is still performing ToInt32
2058         conversions).
2059
2060         * JavaScriptCore.exp:
2061         * runtime/DatePrototype.cpp:
2062         (JSC::fillStructuresUsingTimeArgs):
2063         (JSC::fillStructuresUsingDateArgs):
2064         (JSC::dateProtoFuncSetYear):
2065         * runtime/JSValue.cpp:
2066         (JSC::toInt32):
2067         * runtime/JSValue.h:
2068         (JSC::toUInt32):
2069         (JSC::JSValue::toInt32):
2070         (JSC::JSValue::toUInt32):
2071
2072 2010-09-18  Darin Adler  <darin@apple.com>
2073
2074         First step in fixing Windows build.
2075
2076         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def:
2077         Removed incorrect symbol. The build will probably still fail,
2078         but the failure will tell us what symbol to add.
2079
2080 2010-09-18  Michael Saboff  <msaboff@apple.com>
2081
2082         Reviewed by Gavin Barraclough.
2083
2084         Added code to unroll regular expressions containing ^.
2085         Alternatives that begin with ^ are tagged during parsing
2086         and rolled up in containing sub expression structs.
2087         After parsing, a regular expression flagged as containing
2088         a ^ (a.k.a. BOL) is processed further in optimizeBOL().
2089         A copy of the disjunction is made excluding alternatives that
2090         are rooted with BOL.  The original alternatives are flagged
2091         to only be executed once.  The copy of the other alternatives are
2092         added to the original expression.
2093         In the case that all original alternatives are flagged, there
2094         won't be any looping alternatives.
2095         The JIT generator will emit code accordingly, executing the
2096         original alternatives once and then looping over the
2097         alternatives that aren't anchored with a BOL (if any).
2098         https://bugs.webkit.org/show_bug.cgi?id=45787
2099
2100         * yarr/RegexCompiler.cpp:
2101         (JSC::Yarr::RegexPatternConstructor::assertionBOL):
2102         (JSC::Yarr::RegexPatternConstructor::atomParenthesesEnd):
2103         (JSC::Yarr::RegexPatternConstructor::copyDisjunction):
2104         (JSC::Yarr::RegexPatternConstructor::copyTerm):
2105         (JSC::Yarr::RegexPatternConstructor::optimizeBOL):
2106         (JSC::Yarr::compileRegex):
2107         * yarr/RegexJIT.cpp:
2108         (JSC::Yarr::RegexGenerator::generateDisjunction):
2109         * yarr/RegexPattern.h:
2110         (JSC::Yarr::PatternAlternative::PatternAlternative):
2111         (JSC::Yarr::PatternAlternative::setOnceThrough):
2112         (JSC::Yarr::PatternAlternative::onceThrough):
2113         (JSC::Yarr::PatternDisjunction::PatternDisjunction):
2114         (JSC::Yarr::RegexPattern::RegexPattern):
2115         (JSC::Yarr::RegexPattern::reset):
2116
2117 2010-09-18  Patrick Gansterer  <paroga@paroga.com>
2118
2119         Reviewed by Darin Adler.
2120
2121         Rename Wince files to WinCE
2122         https://bugs.webkit.org/show_bug.cgi?id=37287
2123
2124         * wtf/unicode/Unicode.h:
2125         * wtf/unicode/wince/UnicodeWinCE.cpp: Copied from JavaScriptCore/wtf/unicode/wince/UnicodeWince.cpp.
2126         * wtf/unicode/wince/UnicodeWinCE.h: Copied from JavaScriptCore/wtf/unicode/wince/UnicodeWince.h.
2127         * wtf/unicode/wince/UnicodeWince.cpp: Removed.
2128         * wtf/unicode/wince/UnicodeWince.h: Removed.
2129         * wtf/wince/FastMallocWinCE.h: Copied from JavaScriptCore/wtf/wince/FastMallocWince.h.
2130         * wtf/wince/FastMallocWince.h: Removed.
2131
2132 2010-09-18  Ademar de Souza Reis Jr  <ademar.reis@openbossa.org>
2133
2134         Reviewed by Kenneth Rohde Christiansen.
2135
2136         Enable Platform Strategies on Qt
2137
2138         [Qt] Turn on PLATFORM_STRATEGIES
2139         https://bugs.webkit.org/show_bug.cgi?id=45831
2140
2141         * wtf/Platform.h: Enable Platform Strategies when building QtWebkit
2142
2143 2010-09-17  Oliver Hunt  <oliver@apple.com>
2144
2145         Reviewed by Gavin Barraclough.
2146
2147         Imprecise tracking of variable capture leads to overly pessimistic creation of activations
2148         https://bugs.webkit.org/show_bug.cgi?id=46020
2149
2150         The old logic for track free and captured variables would cause us
2151         to decide we needed an activation in every function along the scope
2152         chain between a variable capture and its declaration.  We now track
2153         captured variables precisely which requires a bit of additional work
2154
2155         The most substantial change is that the parsing routine needs to
2156         be passed the list of function parameters when reparsing a function
2157         as when reparsing we don't parse the function declaration itself only
2158         its body.
2159
2160         * JavaScriptCore.exp:
2161         * parser/JSParser.cpp:
2162         (JSC::JSParser::Scope::Scope):
2163         (JSC::JSParser::Scope::needsFullActivation):
2164            We need to distinguish between use of a feature that requires
2165            an activation and eval so we now get this additional flag.
2166         (JSC::JSParser::Scope::collectFreeVariables):
2167         (JSC::JSParser::Scope::getCapturedVariables):
2168            We can't simply return the list of "capturedVariables" now as
2169            is insufficiently precise, so we compute them instead.
2170         (JSC::JSParser::popScope):
2171         (JSC::jsParse):
2172         (JSC::JSParser::JSParser):
2173         (JSC::JSParser::parseProgram):
2174         (JSC::JSParser::parseWithStatement):
2175         (JSC::JSParser::parseTryStatement):
2176         (JSC::JSParser::parseFunctionInfo):
2177         (JSC::JSParser::parseFunctionDeclaration):
2178         (JSC::JSParser::parseProperty):
2179         (JSC::JSParser::parseMemberExpression):
2180         * parser/JSParser.h:
2181         * parser/Parser.cpp:
2182         (JSC::Parser::parse):
2183         * parser/Parser.h:
2184         (JSC::Parser::parse):
2185         * runtime/Executable.cpp:
2186         (JSC::EvalExecutable::compileInternal):
2187         (JSC::ProgramExecutable::checkSyntax):
2188         (JSC::ProgramExecutable::compileInternal):
2189         (JSC::FunctionExecutable::compileForCallInternal):
2190         (JSC::FunctionExecutable::compileForConstructInternal):
2191         (JSC::FunctionExecutable::reparseExceptionInfo):
2192         (JSC::EvalExecutable::reparseExceptionInfo):
2193         (JSC::FunctionExecutable::fromGlobalCode):
2194            Pass function parameters (if available) to the parser.
2195
2196 2010-09-17  Anders Carlsson  <andersca@apple.com>
2197
2198         Reviewed by Sam Weinig.
2199
2200         Add IsFloatingPoint and IsArithmetic type traits
2201         https://bugs.webkit.org/show_bug.cgi?id=46018
2202
2203         * wtf/TypeTraits.h:
2204         * wtf/TypeTraits.cpp:
2205
2206 2010-09-17  Martin Robinson  <mrobinson@igalia.com>
2207
2208         Reviewed by Oliver Hunt.
2209
2210         [GTK] FontPlatformDataFreeType should use smart pointers to hold its members
2211         https://bugs.webkit.org/show_bug.cgi?id=45917
2212
2213         Added support to PlatformRefPtr for handling HashTableDeletedValue.
2214
2215         * wtf/PlatformRefPtr.h:
2216         (WTF::PlatformRefPtr::PlatformRefPtr): Added a constructor that takes HashTableDeletedValue.
2217         (WTF::PlatformRefPtr::isHashTableDeletedValue): Added.
2218
2219 2010-09-16  Oliver Hunt  <oliver@apple.com>
2220
2221         Reviewed by Geoffrey Garen.
2222
2223         Crash due to timer triggered GC on one heap while another heap is active
2224         https://bugs.webkit.org/show_bug.cgi?id=45932
2225         <rdar://problem/8318446>
2226
2227         The GC timer may trigger for one heap while another heap is active.  This
2228         is safe, but requires us to ensure that we have temporarily associated the
2229         thread's identifierTable with the heap we're collecting on.  Otherwise we
2230         may end up with the identifier tables in an inconsistent state leading to
2231         an eventual crash.
2232
2233         * runtime/Collector.cpp:
2234         (JSC::Heap::allocate):
2235         (JSC::Heap::reset):
2236         (JSC::Heap::collectAllGarbage):
2237            Add assertions to ensure we have the correct identifierTable active
2238            while collecting.
2239         * runtime/GCActivityCallbackCF.cpp:
2240         (JSC::DefaultGCActivityCallbackPlatformData::trigger):
2241            Temporarily make the expected IdentifierTable active
2242         * wtf/WTFThreadData.h:
2243         (JSC::IdentifierTable::remove):
2244            Make it possible to see when IdentifierTable::remove has succeeded
2245         * wtf/text/StringImpl.cpp:
2246         (WTF::StringImpl::~StringImpl):
2247            CRASH if an StringImpl is an Identifier but isn't present in the
2248            active IdentifierTable.  If we get to this state something has
2249            gone wrong and we should just crash immediately.
2250
2251 2010-09-16  Martin Robinson  <mrobinson@igalia.com>
2252
2253         Reviewed by Xan Lopez.
2254
2255         [GTK] Implement dissolveDragImageToFraction
2256         https://bugs.webkit.org/show_bug.cgi?id=45826
2257
2258         * wtf/gobject/GTypedefs.h: Added forward declarations for GtkWindow and GdkEventExpose.
2259
2260 2010-09-16  Eric Uhrhane  <ericu@chromium.org>
2261
2262         Reviewed by Jian Li.
2263
2264         Unify FILE_SYSTEM and FILE_WRITER enables under the name FILE_SYSTEM.
2265         https://bugs.webkit.org/show_bug.cgi?id=45798
2266
2267         * Configurations/FeatureDefines.xcconfig:
2268
2269 2010-09-15  Oliver Hunt  <oliver@apple.com>
2270
2271         Reviewed by Geoffrey Garen.
2272
2273         Use free variable analysis to improve activation performance
2274         https://bugs.webkit.org/show_bug.cgi?id=45837
2275
2276         Adds free and captured variable tracking to the JS parser.  This
2277         allows us to avoid construction of an activation object in some
2278         cases.  Future patches will make more use of this information to
2279         improve those cases where activations are still needed.
2280
2281         * parser/ASTBuilder.h:
2282         * parser/JSParser.cpp:
2283         (JSC::JSParser::Scope::Scope):
2284         (JSC::JSParser::Scope::declareVariable):
2285         (JSC::JSParser::Scope::useVariable):
2286         (JSC::JSParser::Scope::collectFreeVariables):
2287         (JSC::JSParser::Scope::capturedVariables):
2288         (JSC::JSParser::ScopeRef::ScopeRef):
2289         (JSC::JSParser::ScopeRef::operator->):
2290         (JSC::JSParser::ScopeRef::index):
2291         (JSC::JSParser::currentScope):
2292         (JSC::JSParser::pushScope):
2293         (JSC::JSParser::popScope):
2294         (JSC::JSParser::parseProgram):
2295         (JSC::JSParser::parseVarDeclarationList):
2296         (JSC::JSParser::parseConstDeclarationList):
2297         (JSC::JSParser::parseTryStatement):
2298         (JSC::JSParser::parseFormalParameters):
2299         (JSC::JSParser::parseFunctionInfo):
2300         (JSC::JSParser::parseFunctionDeclaration):
2301         (JSC::JSParser::parsePrimaryExpression):
2302         * parser/Nodes.cpp:
2303         (JSC::ScopeNodeData::ScopeNodeData):
2304         (JSC::ScopeNode::ScopeNode):
2305         (JSC::ProgramNode::ProgramNode):
2306         (JSC::ProgramNode::create):
2307         (JSC::EvalNode::EvalNode):
2308         (JSC::EvalNode::create):
2309         (JSC::FunctionBodyNode::FunctionBodyNode):
2310         (JSC::FunctionBodyNode::create):
2311         * parser/Nodes.h:
2312         (JSC::ScopeNode::needsActivation):
2313         (JSC::ScopeNode::hasCapturedVariables):
2314         * parser/Parser.cpp:
2315         (JSC::Parser::didFinishParsing):
2316         * parser/Parser.h:
2317         (JSC::Parser::parse):
2318         * parser/SyntaxChecker.h:
2319         * runtime/Executable.cpp:
2320         (JSC::EvalExecutable::compileInternal):
2321         (JSC::ProgramExecutable::compileInternal):
2322         (JSC::FunctionExecutable::compileForCallInternal):
2323         (JSC::FunctionExecutable::compileForConstructInternal):
2324         * runtime/Executable.h:
2325         (JSC::ScriptExecutable::needsActivation):
2326         (JSC::ScriptExecutable::recordParse):
2327
2328 2010-09-14  Hyung Song  <beergun@company100.net>
2329
2330         Reviewed by Kent Tamura.
2331
2332         [BREWMP] Add IMemGroup and IMemSpace to OwnPtr type.
2333         https://bugs.webkit.org/show_bug.cgi?id=44764
2334
2335         * wtf/OwnPtrCommon.h:
2336         * wtf/brew/OwnPtrBrew.cpp:
2337         (WTF::deleteOwnedPtr):
2338
2339 2010-09-14  Darin Adler  <darin@apple.com>
2340
2341         Reviewed by Geoffrey Garen.
2342
2343         Sort with non-numeric custom sort function fails on array with length but no values
2344         https://bugs.webkit.org/show_bug.cgi?id=45781
2345
2346         * runtime/JSArray.cpp:
2347         (JSC::JSArray::sort): Replaced early exit for an array of length zero to instead
2348         exit for any array without values, even if it has a non-0 length.
2349
2350 2010-09-14  Steve Falkenburg  <sfalken@apple.com>
2351
2352         Windows production build fix.
2353         Roll out r65143.
2354
2355         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCoreGenerated.make:
2356
2357 2010-09-14  Kwang Yul Seo  <skyul@company100.net>
2358
2359         Reviewed by Darin Adler.
2360
2361         Share UnicodeMacrosFromICU.h
2362         https://bugs.webkit.org/show_bug.cgi?id=45710
2363
2364         glib, qt4 and wince use the same macros from ICU.
2365         Remove the code duplication and use the same header file.
2366
2367         * wtf/unicode/UnicodeMacrosFromICU.h: Copied from JavaScriptCore/wtf/unicode/glib/UnicodeMacrosFromICU.h.
2368         * wtf/unicode/glib/UnicodeMacrosFromICU.h: Removed.
2369         * wtf/unicode/qt4/UnicodeQt4.h:
2370         * wtf/unicode/wince/UnicodeWince.h:
2371
2372 2010-09-13  Darin Adler  <darin@apple.com>
2373
2374         Reviewed by Adam Barth.
2375
2376         Preparation for eliminating deprecatedParseURL
2377         https://bugs.webkit.org/show_bug.cgi?id=45695
2378
2379         * wtf/text/WTFString.h: Added isAllSpecialCharacters, moved here from
2380         the HTML tree builder.
2381
2382 2010-09-13  Darin Fisher  <darin@chromium.org>
2383
2384         Reviewed by David Levin.
2385
2386         Add option to conditionally compile smooth scrolling support.
2387         https://bugs.webkit.org/show_bug.cgi?id=45689
2388
2389         ENABLE(SMOOTH_SCROLLING) is disabled by default for all platforms.
2390
2391         * wtf/Platform.h:
2392
2393 2010-09-13  Adam Roben  <aroben@apple.com>
2394
2395         Copy JavaScriptCore's generated sources to the right directory
2396
2397         * JavaScriptCore.vcproj/JavaScriptCore.make: Fixed typo.
2398
2399 2010-09-13  Kwang Yul Seo  <skyul@company100.net>
2400
2401         Reviewed by Kent Tamura.
2402
2403         [BREWMP] Don't call _msize
2404         https://bugs.webkit.org/show_bug.cgi?id=45556
2405
2406         Because Brew MP uses its own memory allocator, it is not correct to use
2407         _msize in fastMallocSize. Add !PLATFORM(BREWMP) guard.
2408
2409         * wtf/FastMalloc.cpp:
2410         (WTF::fastMallocSize):
2411
2412 2010-09-11  Simon Hausmann  <simon.hausmann@nokia.com>
2413
2414         Reviewed by Andreas Kling.
2415
2416         [Qt] V8 port: webcore project files changes
2417         https://bugs.webkit.org/show_bug.cgi?id=45141
2418
2419         * JavaScriptCore.pro: Moved wtf specific files to wtf.pri,
2420         so that they can also be used from WebCore.pro for v8 builds.
2421         * wtf/wtf.pri: Added.
2422
2423 2010-09-10  Fridrich Strba  <fridrich.strba@bluewin.ch>
2424
2425         Reviewed by Andreas Kling.
2426
2427         Add a define missing when building with glib unicode backend
2428         https://bugs.webkit.org/show_bug.cgi?id=45544
2429
2430         * wtf/unicode/glib/UnicodeMacrosFromICU.h:
2431
2432 2010-09-10  Stephanie Lewis  <slewis@apple.com>
2433
2434         Reviewed by Alexey Proskuryakov.
2435
2436         Refactor JavaScriptCore memory statistics so that WebKit doesn't need to know
2437         about the JIT and other implementation details of JavaScriptCore.  Necessary
2438         to fix PPC build.
2439
2440         https://bugs.webkit.org/show_bug.cgi?id=45528
2441
2442         * JavaScriptCore.exp:
2443         * JavaScriptCore.xcodeproj/project.pbxproj:
2444         * runtime/MemoryStatistics.cpp: Added.
2445         (JSC::memoryStatistics):
2446         * runtime/MemoryStatistics.h: Added.
2447
2448 2010-09-09  Michael Saboff  <msaboff@apple.com>
2449
2450         Reviewed by Gavin Barraclough.
2451
2452         Added a regular expression tracing facility.  This tracing is connected
2453         to jsc.  Every compiled regular expression object is added to a list.
2454         When the process exits, each regular expression dumps its pattern,
2455         JIT address, number of times it was executed and the number of matches.
2456         This tracing is controlled by the macro ENABLE_REGEXP_TRACING in
2457         wtf/Platform.h.
2458         https://bugs.webkit.org/show_bug.cgi?id=45401
2459
2460         * JavaScriptCore.exp:
2461         * jsc.cpp:
2462         (runWithScripts):
2463         * runtime/JSGlobalData.cpp:
2464         (JSC::JSGlobalData::JSGlobalData):
2465         (JSC::JSGlobalData::~JSGlobalData):
2466         (JSC::JSGlobalData::addRegExpToTrace):
2467         (JSC::JSGlobalData::dumpRegExpTrace):
2468         * runtime/JSGlobalData.h:
2469         * runtime/RegExp.cpp:
2470         (JSC::RegExp::RegExp):
2471         (JSC::RegExp::create):
2472         (JSC::RegExp::match):
2473         * runtime/RegExp.h:
2474         * wtf/Platform.h:
2475         * yarr/RegexJIT.h:
2476         (JSC::Yarr::RegexCodeBlock::getAddr):
2477
2478 2010-09-09  John Therrell  <jtherrell@apple.com>
2479
2480         32-bit build fix.
2481
2482         * jit/ExecutableAllocator.cpp:
2483         (JSC::ExecutableAllocator::committedByteCount):
2484
2485 2010-09-09  John Therrell  <jtherrell@apple.com>
2486
2487         Reviewed by Alexey Proskuryakov.
2488
2489         Added statistics sampling and reporting for JavaScriptCore's RegisterFile and ExecutableAllocator classes
2490         https://bugs.webkit.org/show_bug.cgi?id=45134
2491
2492         Added thread-safe committed byte counting and reporting functionality to RegisterFile and
2493         ExecutableAllocator.
2494
2495         * JavaScriptCore.exp:
2496         Exported new symbols to allow for WebKit to get statistics from JavaScriptCore classes.
2497
2498         * interpreter/RegisterFile.cpp:
2499         (JSC::registerFileStatisticsMutex):
2500         Added function which returns a static Mutex used for locking during read/write access to
2501         static committed byte count variable.
2502         (JSC::RegisterFile::~RegisterFile):
2503         Added call to addToStatistics since memory is decommitted here.
2504         (JSC::RegisterFile::releaseExcessCapacity):
2505         Added call to addToStatistics since memory is decommitted here.
2506         (JSC::RegisterFile::initializeThreading):
2507         Added function which calls registerFileStatisticsMutex().
2508         (JSC::RegisterFile::committedByteCount):
2509         Added function which returns the current committed byte count for RegisterFile.
2510         (JSC::RegisterFile::addToCommittedByteCount):
2511         Added function which updates committed byte count.
2512
2513         * interpreter/RegisterFile.h:
2514         (JSC::RegisterFile::RegisterFile):
2515         Added call to addToStatistics since memory is committed here.
2516         (JSC::RegisterFile::grow):
2517         Added call to addToStatistics since memory is committed here.
2518
2519         * jit/ExecutableAllocator.h:
2520         Added function prototype for public static function committedByteCount().
2521
2522         * jit/ExecutableAllocatorFixedVMPool.cpp:
2523         (JSC::FixedVMPoolAllocator::release):
2524         Added call to addToStatistics since memory is decommitted here.
2525         (JSC::FixedVMPoolAllocator::reuse):
2526         Added call to addToStatistics since memory is committed here.
2527         (JSC::FixedVMPoolAllocator::addToCommittedByteCount):
2528         Added function which updates committed byte count.
2529         (JSC::ExecutableAllocator::committedByteCount):
2530         Added function which returns the current committed byte count for ExecutableAllocator.
2531
2532         * runtime/InitializeThreading.cpp:
2533         (JSC::initializeThreadingOnce):
2534         Added call to RegisterFile::initializeThreading.
2535
2536 2010-09-09  Mark Rowe  <mrowe@apple.com>
2537
2538         Reviewed by Oliver Hunt.
2539
2540         <http://webkit.org/b/45502> JSObjectSetPrivateProperty does not handle NULL values as it claims
2541
2542         * API/JSObjectRef.cpp:
2543         (JSObjectSetPrivateProperty): Don't call toJS if we have a NULL value as that will cause an assertion
2544         failure. Instead map NULL directly to the null JSValue.
2545         * API/tests/testapi.c:
2546         (main): Add test coverage for the NULL value case.
2547
2548 2010-09-09  Csaba Osztrogonác  <ossy@webkit.org>
2549
2550         Reviewed by Gavin Barraclough.
2551
2552         [Qt] JSVALUE32_64 not works on Windows platform with MinGW compiler
2553         https://bugs.webkit.org/show_bug.cgi?id=29268
2554
2555         * wtf/Platform.h: Enable JSVALUE32_64 for Qt/Windows/MinGW, because it works now.
2556
2557 2010-09-08  Zoltan Herczeg  <zherczeg@webkit.org>
2558
2559         Reviewed by Darin Adler.
2560
2561         Removing doneSemicolon label in the lexer
2562         https://bugs.webkit.org/show_bug.cgi?id=45289
2563
2564         As a side effect of moving the multiline comment parsing
2565         to a separate function, an opportunity raised to simplify
2566         the single line comment parsing, and removing doneSemicolon
2567         label. Slight performance increase on --parse-only
2568         tests (from 32.8ms to 31.5ms)
2569
2570         * parser/Lexer.cpp:
2571         (JSC::Lexer::lex):
2572
2573 2010-09-08  Xan Lopez  <xlopez@igalia.com>
2574
2575         Reviewed by Alexey Proskuryakov.
2576
2577         Remove accessor for private member variable in JSParser
2578         https://bugs.webkit.org/show_bug.cgi?id=45378
2579
2580         m_token is private to JSParser, so it does not seem to be useful
2581         to have an accessor for it. On top of that, the file was both
2582         using the accessor and directly accessing the member variable,
2583         only one style should be used.
2584
2585 2010-09-08  Csaba Osztrogonác  <ossy@webkit.org>
2586
2587         Reviewed by Oliver Hunt.
2588
2589         [Qt] REGRESSION(63348): jsc is broken
2590         https://bugs.webkit.org/show_bug.cgi?id=42818
2591
2592         Need fastcall conventions on Qt/Win/MinGW.
2593         Based on patches of Gavin Barraclough: r63947 and r63948.
2594
2595         * jit/JITStubs.cpp:
2596         * jit/JITStubs.h:
2597
2598 2010-09-08  Robert Hogan  <robert@webkit.org>
2599
2600         Reviewed by Antonio Gomes.
2601
2602         Remove some unnecessary duplicate calls to string functions
2603
2604         https://bugs.webkit.org/show_bug.cgi?id=45314
2605
2606         * wtf/text/WTFString.cpp:
2607         (WTF::String::format):
2608
2609 2010-09-08  Jocelyn Turcotte  <jocelyn.turcotte@nokia.com>
2610
2611         Reviewed by Andreas Kling.
2612
2613         Re-Disable JIT for MSVC 64bit to fix the build on this compiler.
2614         https://bugs.webkit.org/show_bug.cgi?id=45382
2615
2616         It was enabled in the cleanup made in r64176, though it is still
2617         not implemented.
2618
2619         * wtf/Platform.h:
2620
2621 2010-09-08  Martin Robinson  <mrobinson@igalia.com>
2622
2623         Reviewed by Xan Lopez.
2624
2625         [GTK] Need a WebSocket implementation
2626         https://bugs.webkit.org/show_bug.cgi?id=45197
2627
2628         Add a GIO-based WebSocket implementation.
2629
2630         * wtf/gobject/GRefPtr.cpp: Added PlatformRefPtr support for GSource.
2631         (WTF::refPlatformPtr):
2632         (WTF::derefPlatformPtr):
2633         * wtf/gobject/GRefPtr.h: Added new template specialization declarations.
2634         * wtf/gobject/GTypedefs.h: Add some more GLib/GIO forward declarations.
2635
2636 2010-08-30  Maciej Stachowiak  <mjs@apple.com>
2637
2638         Reviewed by Darin Adler.
2639
2640         Handle MediaQueryExp memory management exclusively with smart pointers
2641         https://bugs.webkit.org/show_bug.cgi?id=44874
2642
2643         Implemented a non-copying sort function to make it possible to sort a Vector
2644         of OwnPtrs (which cannot be copied). This is required for the above.
2645
2646         * wtf/NonCopyingSort.h: Added.
2647         (WTF::nonCopyingSort): It's secretly heapsort.
2648         (WTF::heapSort): heapsort implementation.
2649         (WTF::siftDown): Helper function for heapsort.
2650         (WTF::heapify): ditto
2651
2652         Adjust build systems.
2653
2654         * GNUmakefile.am:
2655         * JavaScriptCore.gypi:
2656         * JavaScriptCore.vcproj/WTF/WTF.vcproj:
2657         * JavaScriptCore.xcodeproj/project.pbxproj:
2658
2659 2010-09-08  Zoltan Herczeg  <zherczeg@webkit.org>
2660
2661         Reviewed by Darin Adler.
2662
2663         Refactoring multiline comments in the lexer
2664         https://bugs.webkit.org/show_bug.cgi?id=45289
2665
2666         MultiLine comment parsing is moved to a separate function.
2667
2668         Slight performance increase on --parse-only tests (from 33.6ms to 32.8ms)
2669         SunSpider reports no change (from 523.1ms to 521.2ms).
2670
2671         * parser/Lexer.cpp:
2672         (JSC::Lexer::parseMultilineComment):
2673         (JSC::Lexer::lex):
2674         * parser/Lexer.h:
2675
2676 2010-09-07  James Robinson  <jamesr@chromium.org>
2677
2678         Compile fix attempt for windows.
2679
2680         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def:
2681
2682 2010-09-07  Mihai Parparita  <mihaip@chromium.org>
2683
2684         Reviewed by James Robinson.
2685
2686         Fix Windows build after r66936
2687         https://bugs.webkit.org/show_bug.cgi?id=45348
2688
2689         Add symbol names that were missing from r66936.
2690
2691         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def:
2692
2693 2010-09-07  Mihai Parparita  <mihaip@chromium.org>
2694
2695         Reviewed by Oliver Hunt.
2696
2697         pushState and replaceState do not clone RegExp objects correctly
2698         https://bugs.webkit.org/show_bug.cgi?id=44718
2699
2700         Move internal representation of JSC::RegExp (which depends on wether
2701         YARR and YARR_JIT is enabled) into RegExpRepresentation which can live
2702         in the implementation only. This makes it feasible to use RegExp in
2703         WebCore without bringing in all of YARR.
2704
2705         * JavaScriptCore.exp: Export RegExp and RegExpObject functions that are
2706         needed inside WebCore's JSC bindings.
2707         * runtime/RegExp.cpp:
2708         (JSC::RegExpRepresentation::~RegExpRepresentation):
2709         (JSC::RegExp::RegExp):
2710         (JSC::RegExp::~RegExp):
2711         (JSC::RegExp::compile):
2712         (JSC::RegExp::match):
2713         * runtime/RegExp.h:
2714
2715 2010-09-07  Anders Carlsson  <andersca@apple.com>
2716
2717         Reviewed by Darin Adler.
2718
2719         <rdar://problem/8381749> -Wcast-align warning emitted when building with clang
2720
2721         Remove the -Wcast-align-warning since it isn't really useful, and clang is more aggressive about warning than gcc.
2722
2723         * Configurations/Base.xcconfig:
2724
2725 2010-09-07  Zoltan Horvath  <zoltan@webkit.org>
2726
2727         Reviewed by Darin Adler.
2728
2729         REGRESSION(66741): Undefined pthread macros
2730         https://bugs.webkit.org/show_bug.cgi?id=45246
2731
2732         PTHREAD_MUTEX_NORMAL and PTHREAD_MUTEX_DEFAULT (introduced in r60487) are not defined on Linux,
2733         but used in a statement. Add an additional check to test this.
2734
2735         * wtf/FastMalloc.cpp:
2736         (WTF::TCMalloc_PageHeap::initializeScavenger):
2737
2738 2010-09-06  Oliver Hunt  <oliver@apple.com>
2739
2740         Windows build fix
2741
2742 2010-09-05  Oliver Hunt  <oliver@apple.com>
2743
2744         Reviewed by Sam Weinig.
2745
2746         SerializedScriptValue needs to use a flat storage mechanism
2747         https://bugs.webkit.org/show_bug.cgi?id=45244
2748
2749         Export JSArray::put
2750
2751         * JavaScriptCore.exp:
2752
2753 2010-09-06  Chao-ying Fu  <fu@mips.com>
2754
2755         Reviewed by Oliver Hunt.
2756
2757         Support JSVALUE32_64 on MIPS
2758         https://bugs.webkit.org/show_bug.cgi?id=43999
2759
2760         Add missing functions to support JSVALUE32_64 on MIPS.
2761         Remove JSVALUE32 as the default for MIPS.
2762
2763         * assembler/MIPSAssembler.h:
2764         (JSC::MIPSAssembler::divd):
2765         (JSC::MIPSAssembler::mthc1):
2766         (JSC::MIPSAssembler::cvtwd):
2767         * assembler/MacroAssemblerMIPS.h:
2768         (JSC::MacroAssemblerMIPS::neg32):
2769         (JSC::MacroAssemblerMIPS::branchOr32):
2770         (JSC::MacroAssemblerMIPS::set8):
2771         (JSC::MacroAssemblerMIPS::loadDouble):
2772         (JSC::MacroAssemblerMIPS::divDouble):
2773         (JSC::MacroAssemblerMIPS::convertInt32ToDouble):
2774         (JSC::MacroAssemblerMIPS::branchDouble):
2775         (JSC::MacroAssemblerMIPS::branchConvertDoubleToInt32):
2776         (JSC::MacroAssemblerMIPS::zeroDouble):
2777         * jit/JIT.h:
2778         * jit/JITOpcodes32_64.cpp:
2779         (JSC::JIT::privateCompileCTINativeCall):
2780         * jit/JITPropertyAccess32_64.cpp:
2781         (JSC::JIT::privateCompilePutByIdTransition):
2782         * jit/JITStubs.cpp:
2783         (JSC::JITThunks::JITThunks):
2784         * jit/JITStubs.h:
2785         * wtf/Platform.h:
2786
2787 2010-09-06  Robert Hogan  <robert@webkit.org>
2788
2789         Unreviewed, compile fix.
2790
2791         Fix compile failure in r66843
2792
2793         Revert to original patch in bugzilla. Leave bug open for
2794         discussion on potential removal of double utf8 conversion.
2795
2796         https://bugs.webkit.org/show_bug.cgi?id=45240
2797
2798         * wtf/text/WTFString.cpp:
2799         (WTF::String::format):
2800
2801 2010-09-06  Robert Hogan  <robert@webkit.org>
2802
2803         Reviewed by Andreas Kling.
2804
2805         [Qt] utf8 encoding of console() messages
2806
2807         Unskip:
2808         http/tests/security/xssAuditor/embed-tag-null-char.html
2809         http/tests/security/xssAuditor/object-embed-tag-null-char.html
2810
2811         Both tests failed because Qt's implementation of String::format()
2812         is casting a utf8 result to String, which assumes latin1 in
2813         its constructor. So instead of casting a QString to a String, use
2814         StringImpl::create() instead. Unfortunately, this involves a lot
2815         of extra casts but the end result is correct.
2816
2817         https://bugs.webkit.org/show_bug.cgi?id=45240
2818
2819         * wtf/text/WTFString.cpp:
2820         (WTF::String::format):
2821
2822 2010-09-03  Alexey Proskuryakov  <ap@apple.com>
2823
2824         Reviewed by Darin Adler.
2825
2826         https://bugs.webkit.org/show_bug.cgi?id=45135
2827         <rdar://problem/7823714> TCMalloc_PageHeap doesn't hold a mutex while manipulating shared data
2828
2829         * wtf/FastMalloc.cpp:
2830         (WTF::TCMalloc_PageHeap::initializeScavenger): Make sure to create a non-recursive mutex
2831         regardless of platform default, so that we can assert that it's held (this is for platforms
2832         that don't have libdispatch).
2833         (WTF::TCMalloc_PageHeap::signalScavenger): Assert that the mutex is held, so we can look
2834         at m_scavengeThreadActive. For platforms that have libdispatch, assert that pageheap_lock
2835         is held.
2836         (WTF::TCMalloc_PageHeap::periodicScavenge): Make sure that pageheap_lock is held before
2837         manipulating m_scavengeThreadActive. Otherwise, there is an obvious race condition, and we
2838         can make unbalanced calls to dispatch_resume().
2839
2840 2010-09-03  Lucas De Marchi  <lucas.demarchi@profusion.mobi>
2841
2842         Reviewed by Martin Robinson.
2843
2844         [EFL] Regression (66531) Build break with Glib Support
2845         https://bugs.webkit.org/show_bug.cgi?id=45011
2846
2847         Move GtkTypedefs.h to GTypedefs.h and let it inside gobject directory
2848         since when glib is enabled, EFL port needs it, too.
2849
2850         * CMakeListsEfl.txt: Include gobject directory to find new header
2851         file.
2852         * GNUmakefile.am: Ditto.
2853         * wtf/CMakeListsEfl.txt: Ditto.
2854         * wtf/Platform.h: Include header if port is EFL and glib support is
2855         enabled.
2856         * wtf/gtk/GtkTypedefs.h: Removed.
2857         * wtf/gobject/GTypedefs.h: Added. Sections specific to GTK are now
2858         guarded by PLATFORM(GTK).
2859
2860 2010-09-03  Csaba Osztrogonác  <ossy@webkit.org>
2861
2862         Reviewed by Simon Hausmann.
2863
2864         Fix warning in wtf/ByteArray.h
2865         https://bugs.webkit.org/show_bug.cgi?id=44672
2866
2867         * wtf/ByteArray.h: Use maximal sized array for MSVC and unsized array for other compilers.
2868
2869 2010-09-02  Adam Barth  <abarth@webkit.org>
2870
2871         Reviewed by Eric Seidel.
2872
2873         Actually parse a URL from ParsedURL
2874         https://bugs.webkit.org/show_bug.cgi?id=45080
2875
2876         This patch only handles standard URLs.  At some point we'll need to
2877         distinguish between standard URLs and other kinds of URLs.
2878
2879         * wtf/url/api/ParsedURL.cpp:
2880         (WTF::ParsedURL::ParsedURL):
2881
2882 2010-09-02  Adam Barth  <abarth@webkit.org>
2883
2884         Reviewed by Eric Seidel.
2885
2886         Add ParsedURL and URLString to WTFURL API
2887         https://bugs.webkit.org/show_bug.cgi?id=45078
2888
2889         Currently there's no actual URL parsing going on, but this patch is a
2890         start to sketching out the API.
2891
2892         * JavaScriptCore.xcodeproj/project.pbxproj:
2893         * wtf/url/api/ParsedURL.cpp: Added.
2894         (WTF::ParsedURL::ParsedURL):
2895         (WTF::ParsedURL::scheme):
2896         (WTF::ParsedURL::username):
2897         (WTF::ParsedURL::password):
2898         (WTF::ParsedURL::host):
2899         (WTF::ParsedURL::port):
2900         (WTF::ParsedURL::path):
2901         (WTF::ParsedURL::query):
2902         (WTF::ParsedURL::fragment):
2903         (WTF::ParsedURL::segment):
2904         * wtf/url/api/ParsedURL.h: Added.
2905         (WTF::ParsedURL::spec):
2906         * wtf/url/api/URLString.h: Added.
2907         (WTF::URLString::URLString):
2908         (WTF::URLString::string):
2909
2910 2010-09-02  Adam Barth  <abarth@webkit.org>
2911
2912         Reviewed by Eric Seidel.
2913
2914         Add WTFURL to the JavaScriptCore build on Mac
2915         https://bugs.webkit.org/show_bug.cgi?id=45075
2916
2917         Building code is good.
2918
2919         * JavaScriptCore.xcodeproj/project.pbxproj:
2920
2921 2010-09-02  Alexey Proskuryakov  <ap@apple.com>
2922
2923         Reviewed by Oliver Hunt.
2924
2925         https://bugs.webkit.org/show_bug.cgi?id=43230
2926         <rdar://problem/8254215> REGRESSION: Memory leak within JSParser::JSParser
2927
2928         One can't delete a ThreadSpecific object that has data in it. It's not even possible to
2929         enumerate data objects in all threads, much less destroy them from a thread that's destroying
2930         the ThreadSpecific.
2931
2932         * parser/JSParser.cpp:
2933         (JSC::JSParser::JSParser):
2934         * runtime/JSGlobalData.h:
2935         * wtf/WTFThreadData.cpp:
2936         (WTF::WTFThreadData::WTFThreadData):
2937         * wtf/WTFThreadData.h:
2938         (WTF::WTFThreadData::approximatedStackStart):
2939         Moved stack guard tracking from JSGlobalData to WTFThreadData.
2940
2941         * wtf/ThreadSpecific.h: Made destructor unimplemented. It's dangerous, and we probably won't
2942         ever face a situation where we'd want to delete a ThreadSpecific object.
2943
2944 2010-09-01  Gavin Barraclough  <barraclough@apple.com>
2945
2946         Rubber stamped by Oliver Hunt.
2947
2948         Ecma-262 15.11.1.1 states that if the argument is undefined then an
2949         Error object's message property should be set to the empty string.
2950
2951         * runtime/ErrorInstance.cpp:
2952         (JSC::ErrorInstance::ErrorInstance):
2953         (JSC::ErrorInstance::create):
2954         * runtime/ErrorInstance.h:
2955         * runtime/ErrorPrototype.cpp:
2956         (JSC::ErrorPrototype::ErrorPrototype):
2957
2958 2010-08-31  Darin Adler  <darin@apple.com>
2959
2960         Reviewed by Anders Carlsson.
2961
2962         * wtf/FastMalloc.cpp:
2963         (WTF::TCMalloc_PageHeap::scavenge): Replaced somewhat-quirky code that
2964         mixed types with code that uses size_t.
2965
2966         * wtf/TCPageMap.h: Removed names of unused arguments to avoid warning.
2967
2968 2010-08-31  Martin Robinson  <mrobinson@igalia.com>
2969
2970         Reviewed by Gustavo Noronha Silva.
2971
2972         [GTK] Isolate all GTK+ typedefs into one file
2973         https://bugs.webkit.org/show_bug.cgi?id=44900
2974
2975         * GNUmakefile.am: Add GtkTypedefs.h to the source lists.
2976         * wtf/Platform.h: #include GtkTypedefs.h for the GTK+ build.
2977         * wtf/ThreadingPrimitives.h: Remove GTK+ typedefs.
2978         * wtf/gobject/GOwnPtr.h: Ditto.
2979         * wtf/gobject/GRefPtr.h: Ditto.
2980         * wtf/gtk/GtkTypedefs.h: Added.
2981
2982 2010-08-31  Martin Robinson  <mrobinson@igalia.com>
2983
2984         Reviewed by Gustavo Noronha Silva.
2985
2986         [GTK] Fix 'make dist' in preparation of the 1.3.3 release
2987         https://bugs.webkit.org/show_bug.cgi?id=44978
2988
2989         * GNUmakefile.am: Adding missing headers to the sources list.
2990
2991 2010-08-31  Chao-ying Fu  <fu@mips.com>
2992
2993         Reviewed by Oliver Hunt.
2994
2995         Support emit_op_mod() for MIPS
2996         https://bugs.webkit.org/show_bug.cgi?id=42855
2997
2998         This patch uses MIPS div instructions for op_mod to improve performance.
2999
3000         * assembler/MIPSAssembler.h:
3001         (JSC::MIPSAssembler::div):
3002         * jit/JITArithmetic.cpp:
3003         (JSC::JIT::emit_op_mod):
3004         (JSC::JIT::emitSlow_op_mod):
3005
3006 2010-08-31  Csaba Osztrogonác  <ossy@webkit.org>
3007
3008         Reviewed by Darin Adler.
3009
3010         Modify ASSERT_UNUSED and UNUSED_PARAM similar to Qt's Q_UNUSED.
3011         https://bugs.webkit.org/show_bug.cgi?id=44870
3012
3013         * wtf/Assertions.h:
3014         * wtf/UnusedParam.h:
3015
3016 2010-08-31  Benjamin Poulain  <benjamin.poulain@nokia.com>
3017
3018         Reviewed by Kenneth Rohde Christiansen.
3019
3020         JSC TimeoutChecker::didTimeOut overflows on ARM
3021         https://bugs.webkit.org/show_bug.cgi?id=38538
3022
3023         Make getCPUTime() return values relative to the first call.
3024         The previous implementation relied on simply on currentTime(), which
3025         return a time since epoch and not a time since the thread started. This
3026         made the return value of getCPUTime() overflow on 32 bits.
3027
3028         * runtime/TimeoutChecker.cpp:
3029         (JSC::getCPUTime):
3030
3031 2010-08-30  Mihai Parparita  <mihaip@chromium.org>
3032
3033         Reviewed by Adam Barth.
3034
3035         HISTORY_ALWAYS_ASYNC should be removed (history should always be async)
3036         https://bugs.webkit.org/show_bug.cgi?id=44315
3037
3038         Remove ENABLE_HISTORY_ALWAYS_ASYNC #define.
3039
3040         * wtf/Platform.h:
3041
3042 2010-08-30  Chris Rogers  <crogers@google.com>
3043
3044         Reviewed by Kenneth Russell.
3045
3046         Fix namespace for wtf/Complex.h and wtf/Vector3.h
3047         https://bugs.webkit.org/show_bug.cgi?id=44892
3048
3049         * wtf/Complex.h:
3050         * wtf/Vector3.h:
3051
3052 2010-08-30  Andy Estes  <aestes@apple.com>
3053
3054         Reviewed by Eric Carlson.
3055
3056         Strings returned by asciiDebug() should be NULL-terminated.
3057         https://bugs.webkit.org/show_bug.cgi?id=44866
3058
3059         * wtf/text/WTFString.cpp:
3060         (asciiDebug):
3061
3062 2010-08-30  Zoltan Herczeg  <zherczeg@webkit.org>
3063
3064         Reviewed by Darin Adler.
3065
3066         Refactor number parsing in the lexer
3067         https://bugs.webkit.org/show_bug.cgi?id=44104
3068
3069         Number parsing was full of gotos, and needed a complete
3070         redesign to remove them (Only one remained). Furthermore
3071         integer arithmetic is empolyed for fast cases (= small
3072         integer numbers).
3073
3074         * parser/Lexer.cpp:
3075         (JSC::Lexer::parseHex):
3076         (JSC::Lexer::parseOctal):
3077         (JSC::Lexer::parseDecimal):
3078         (JSC::Lexer::parseNumberAfterDecimalPoint):
3079         (JSC::Lexer::parseNumberAfterExponentIndicator):
3080         (JSC::Lexer::lex):
3081         * parser/Lexer.h:
3082
3083 2010-08-29  Darin Adler  <darin@apple.com>
3084
3085         Fix Qt build.
3086
3087         * wtf/unicode/glib/UnicodeMacrosFromICU.h: Added U_IS_BMP.
3088         * wtf/unicode/qt4/UnicodeQt4.h: Ditto.
3089         * wtf/unicode/wince/UnicodeWince.h: Ditto.
3090
3091 2010-08-29  Kwang Yul Seo  <skyul@company100.net>
3092
3093         Reviewed by Kent Tamura.
3094
3095         [BREWMP] Port vprintf_stderr_common
3096         https://bugs.webkit.org/show_bug.cgi?id=33568
3097
3098         Use BREW's DBGPRINTF to output debug messages.
3099
3100         * wtf/Assertions.cpp:
3101
3102 2010-08-28  Gavin Barraclough  <barraclough@apple.com>
3103
3104         Reviewed by Oliver Hunt.
3105
3106         Bug 44830 - In Array's prototype functyions we're incorrectly handing large index values
3107
3108         We are in places casting doubles to unsigneds, and unsigneds to ints, without always check
3109         that the result is within bounds. This is problematic in the case of double-to-unsigned
3110         conversion because we should be saturating to array length.
3111
3112         Also, the error return value from Array.splice should be [], not undefined.
3113
3114         I don't see any security concerns here. These methods are spec'ed in such a way that they
3115         can be applied to non Array objects, so in all cases the (potentially bogus) indices are
3116         being passed to functions that will safely check accesses are within bounds.
3117
3118         * runtime/ArrayPrototype.cpp:
3119         (JSC::argumentClampedIndexFromStartOrEnd):
3120         (JSC::arrayProtoFuncJoin):
3121         (JSC::arrayProtoFuncConcat):
3122         (JSC::arrayProtoFuncReverse):
3123         (JSC::arrayProtoFuncShift):
3124         (JSC::arrayProtoFuncSlice):
3125         (JSC::arrayProtoFuncSort):
3126         (JSC::arrayProtoFuncSplice):
3127         (JSC::arrayProtoFuncUnShift):
3128         (JSC::arrayProtoFuncFilter):
3129         (JSC::arrayProtoFuncMap):
3130         (JSC::arrayProtoFuncEvery):
3131         (JSC::arrayProtoFuncForEach):
3132         (JSC::arrayProtoFuncSome):
3133         (JSC::arrayProtoFuncReduce):
3134         (JSC::arrayProtoFuncReduceRight):
3135         (JSC::arrayProtoFuncIndexOf):
3136         (JSC::arrayProtoFuncLastIndexOf):
3137         * runtime/JSValue.h:
3138         (JSC::JSValue::toUInt32):
3139
3140 2010-08-28  Pratik Solanki  <psolanki@apple.com>
3141
3142         Reviewed by Dan Bernstein.
3143
3144         Add an ENABLE define for purgeable memory support
3145         https://bugs.webkit.org/show_bug.cgi?id=44777
3146
3147         * wtf/Platform.h:
3148
3149 2010-08-27  Kimmo Kinnunen  <kimmo.t.kinnunen@nokia.com>
3150
3151         Reviewed by Kenneth Rohde Christiansen.
3152
3153         [Qt] NPAPI Plugin metadata should be cached, and loading a plugin should not require loading every plugin
3154         https://bugs.webkit.org/show_bug.cgi?id=43179
3155
3156         Add ENABLE_NETSCAPE_PLUGIN_METADATA_CACHE flag to enable persistent
3157         NPAPI Plugin Cache. The flag is enabled by default.
3158
3159         * wtf/Platform.h: Add ENABLE_NETSCAPE_PLUGIN_METADATA_CACHE
3160
3161 2010-07-27  Jer Noble  <jer.noble@apple.com>
3162
3163         Reviewed by Eric Carlson.
3164
3165         Add JavaScript API to allow a page to go fullscreen.
3166         rdar://problem/6867795
3167         https://bugs.webkit.org/show_bug.cgi?id=43099
3168
3169         * wtf/Platform.h: Enable FULLSCREEN_API mode for the Mac (except iOS).
3170
3171 2010-08-27  Gavin Barraclough  <barraclough@apple.com>
3172
3173         Windows build fix pt 2.
3174
3175         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def:
3176
3177 2010-08-27  Gavin Barraclough  <barraclough@apple.com>
3178
3179         Windows build fix pt 1.
3180
3181         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def:
3182
3183 2010-08-27  Gavin Barraclough  <barraclough@apple.com>
3184
3185         Reviewed by Oliver Hunt.
3186
3187         Bug 44745 - Number.toFixed/toExponential/toPrecision are inaccurate.
3188
3189         These methods should be using a version of dtoa that can generate results accurate
3190         to the requested precision, whereas our version of dtoa is only currently able to
3191         support producing results sufficiently accurate to distinguish the value from any
3192         other IEEE-754 double precision number.
3193
3194         This change has no impact on benchmarks we track.
3195
3196         On microbenchmarks for these functions, this is a slight regression where a high
3197         precision is requested (dtoa now need to iterate further to generate a a greater
3198         number of digits), but with smaller precision values (hopefully more common) this
3199         improves performance, since it reduced the accurate of result dtoa is required,
3200         to produce, and removes the need to pre-round values before calling dtoa.
3201
3202         * JavaScriptCore.exp:
3203             doubleToStringInJavaScriptFormat renamed to numberToString
3204
3205         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def:
3206             doubleToStringInJavaScriptFormat renamed to numberToString
3207
3208         * runtime/UString.cpp:
3209         (JSC::UString::number):
3210             doubleToStringInJavaScriptFormat renamed to numberToString
3211
3212         * wtf/DecimalNumber.h:
3213         (WTF::DecimalNumber::DecimalNumber):
3214         (WTF::DecimalNumber::toStringDecimal):
3215         (WTF::DecimalNumber::toStringExponential):
3216             Remove all pre-rounding of values, instead call dtoa correctly.
3217
3218         * wtf/dtoa.cpp:
3219         (WTF::dtoa):
3220         * wtf/dtoa.h:
3221             Reenable support for rounding to specific-figures/decimal-places in dtoa.
3222             Modify to remove unbiased rounding, provide ECMA required away-from-zero.
3223             Rewrite doubleToStringInJavaScriptFormat to use DecimalNumber, rename to
3224             numberToString.
3225
3226 2010-08-27  Chao-ying Fu  <fu@mips.com>
3227
3228         Reviewed by Oliver Hunt.
3229
3230         Byte alignment issue on MIPS
3231         https://bugs.webkit.org/show_bug.cgi?id=29415
3232
3233         MIPS accesses one byte at a time for now to avoid the help from the
3234         kernel to fix unaligned accesses.
3235
3236         * wtf/text/AtomicString.cpp:
3237         (WebCore::equal):
3238         * wtf/text/StringHash.h:
3239         (WebCore::StringHash::equal):
3240
3241 2010-08-27  Xan Lopez  <xlopez@igalia.com>
3242
3243         Reviewed by Tor Arne Vestbø.
3244
3245         Fix a couple of typos in comment.
3246
3247         * bytecode/CodeBlock.h:
3248
3249 2010-08-26  Gavin Barraclough  <barraclough@apple.com>
3250
3251         Windows build fix.
3252
3253         * wtf/dtoa.cpp:
3254
3255 2010-08-26  Gavin Barraclough  <baraclough@apple.com>
3256
3257         Reviewed by Sam Weinig.
3258
3259         Bug 44735 - Clean up dtoa.cpp
3260         Remove unused & unmaintained code paths, reformat code to match
3261         coding standard & use platform #defines from Platform.h directly.
3262
3263         * wtf/dtoa.cpp:
3264         (WTF::storeInc):
3265         (WTF::multadd):
3266         (WTF::s2b):
3267         (WTF::lo0bits):
3268         (WTF::mult):
3269         (WTF::pow5mult):
3270         (WTF::lshift):
3271         (WTF::diff):
3272         (WTF::ulp):
3273         (WTF::b2d):
3274         (WTF::d2b):
3275         (WTF::ratio):
3276         (WTF::):
3277         (WTF::strtod):
3278         (WTF::quorem):
3279         (WTF::dtoa):
3280
3281 2010-08-26  Gavin Barraclough  <barraclough@apple.com>
3282
3283         Rubber Stamped by Oliver Hunt.
3284
3285         Partially revert r65959. The toString changes regressed the v8 tests,
3286         but keep the toFixed/toExponential/toPrecision changes.
3287
3288         * JavaScriptCore.exp:
3289         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def:
3290         * runtime/NumberPrototype.cpp:
3291         * runtime/UString.cpp:
3292         (JSC::UString::number):
3293         * wtf/DecimalNumber.h:
3294         * wtf/dtoa.cpp:
3295         (WTF::append):
3296         (WTF::doubleToStringInJavaScriptFormat):
3297         * wtf/dtoa.h:
3298         * wtf/text/WTFString.cpp:
3299         * wtf/text/WTFString.h:
3300
3301 2010-08-26  James Robinson  <jamesr@chromium.org>
3302
3303         Reviewed by Darin Fisher.
3304
3305         [chromium] Remove the USE(GLES2_RENDERING) define and associated code
3306         https://bugs.webkit.org/show_bug.cgi?id=43761
3307
3308         Remove WTF_USE_GLES2_RENDERING from the list of defines in chromium, it's unused.
3309
3310         * wtf/Platform.h:
3311
3312 2010-08-26  Gavin Barraclough  <barraclough@apple.com>
3313
3314         Rolling out r64608, this regressed performance.
3315
3316         * JavaScriptCore.xcodeproj/project.pbxproj:
3317         * assembler/ARMAssembler.cpp:
3318         (JSC::ARMAssembler::executableCopy):
3319         * assembler/LinkBuffer.h:
3320         (JSC::LinkBuffer::LinkBuffer):
3321         (JSC::LinkBuffer::~LinkBuffer):
3322         (JSC::LinkBuffer::performFinalization):
3323         * assembler/MIPSAssembler.h:
3324         (JSC::MIPSAssembler::executableCopy):
3325         * assembler/X86Assembler.h:
3326         (JSC::X86Assembler::executableCopy):
3327         * bytecode/StructureStubInfo.h:
3328         (JSC::StructureStubInfo::initGetByIdProto):
3329         (JSC::StructureStubInfo::initGetByIdChain):
3330         (JSC::StructureStubInfo::initGetByIdSelfList):
3331         (JSC::StructureStubInfo::initGetByIdProtoList):
3332         (JSC::StructureStubInfo::initPutByIdTransition):
3333         * jit/ExecutableAllocator.cpp:
3334         (JSC::ExecutablePool::systemAlloc):
3335         * jit/ExecutableAllocator.h:
3336         (JSC::ExecutablePool::create):
3337         (JSC::ExecutableAllocator::ExecutableAllocator):
3338         (JSC::ExecutableAllocator::poolForSize):
3339         (JSC::ExecutablePool::ExecutablePool):
3340         (JSC::ExecutablePool::poolAllocate):
3341         * jit/ExecutableAllocatorFixedVMPool.cpp:
3342         (JSC::FixedVMPoolAllocator::allocInternal):
3343         * jit/JIT.cpp:
3344         (JSC::JIT::privateCompile):
3345         * jit/JIT.h:
3346         (JSC::JIT::compileGetByIdProto):
3347         (JSC::JIT::compileGetByIdSelfList):
3348         (JSC::JIT::compileGetByIdProtoList):
3349         (JSC::JIT::compileGetByIdChainList):
3350         (JSC::JIT::compileGetByIdChain):
3351         (JSC::JIT::compilePutByIdTransition):
3352         (JSC::JIT::compilePatchGetArrayLength):
3353         * jit/JITOpcodes.cpp:
3354         (JSC::JIT::privateCompileCTIMachineTrampolines):
3355         * jit/JITOpcodes32_64.cpp:
3356         (JSC::JIT::privateCompileCTIMachineTrampolines):
3357         (JSC::JIT::privateCompileCTINativeCall):
3358         * jit/JITPropertyAccess.cpp:
3359         (JSC::JIT::stringGetByValStubGenerator):
3360         (JSC::JIT::privateCompilePutByIdTransition):
3361         (JSC::JIT::privateCompilePatchGetArrayLength):
3362         (JSC::JIT::privateCompileGetByIdProto):
3363         (JSC::JIT::privateCompileGetByIdSelfList):
3364         (JSC::JIT::privateCompileGetByIdProtoList):
3365         (JSC::JIT::privateCompileGetByIdChainList):
3366         (JSC::JIT::privateCompileGetByIdChain):
3367         * jit/JITPropertyAccess32_64.cpp:
3368         (JSC::JIT::stringGetByValStubGenerator):
3369         (JSC::JIT::privateCompilePutByIdTransition):
3370         (JSC::JIT::privateCompilePatchGetArrayLength):
3371         (JSC::JIT::privateCompileGetByIdProto):
3372         (JSC::JIT::privateCompileGetByIdSelfList):
3373         (JSC::JIT::privateCompileGetByIdProtoList):
3374         (JSC::JIT::privateCompileGetByIdChainList):
3375         (JSC::JIT::privateCompileGetByIdChain):
3376         * jit/JITStubs.cpp:
3377         (JSC::JITThunks::tryCachePutByID):
3378         (JSC::JITThunks::tryCacheGetByID):
3379         (JSC::DEFINE_STUB_FUNCTION):
3380         (JSC::getPolymorphicAccessStructureListSlot):
3381         * jit/JITStubs.h:
3382         * jit/SpecializedThunkJIT.h:
3383         (JSC::SpecializedThunkJIT::finalize):
3384         * runtime/ExceptionHelpers.cpp:
3385         * runtime/ExceptionHelpers.h:
3386         * runtime/Executable.cpp:
3387         (JSC::EvalExecutable::compileInternal):
3388         (JSC::ProgramExecutable::compileInternal):
3389         (JSC::FunctionExecutable::compileForCallInternal):
3390         (JSC::FunctionExecutable::compileForConstructInternal):
3391         (JSC::FunctionExecutable::reparseExceptionInfo):
3392         (JSC::EvalExecutable::reparseExceptionInfo):
3393         * yarr/RegexJIT.cpp:
3394         (JSC::Yarr::RegexGenerator::compile):
3395
3396 2010-08-26  Gavin Barraclough  <barraclough@apple.com>
3397
3398         Reviewed by Brady Eidson.
3399
3400         Bug 44655 - Add debug only convenience methods to obtain a Vector<char> from a String/StringImpl.
3401
3402         * wtf/text/WTFString.cpp:
3403         (asciiDebug):
3404             Return a Vector<char> containing the contents of a string as ASCII.
3405
3406 2010-08-26  Sam Weinig  <sam@webkit.org>
3407
3408         Reviewed by Darin Adler.
3409
3410         Add PassOwnArrayPtr
3411         https://bugs.webkit.org/show_bug.cgi?id=44627
3412
3413         * GNUmakefile.am:
3414         * JavaScriptCore.vcproj/WTF/WTF.vcproj:
3415         * JavaScriptCore.xcodeproj/project.pbxproj:
3416         Add the new files.
3417
3418         * wtf/Forward.h:
3419         Forward declare PassOwnArrayPtr.
3420
3421         * wtf/OwnArrayPtr.h:
3422         Mimic the OwnPtr interface.
3423
3424         * wtf/OwnArrayPtrCommon.h: Added.
3425         (WTF::deleteOwnedArrayPtr):
3426         Move delete function here so it can be shared by OwnArrayPtr and
3427         PassOwnArrayPtr.
3428
3429         * wtf/PassOwnArrayPtr.h: Added.
3430         Mimic the PassOwnPtr interface.
3431
3432 2010-08-26  Oliver Hunt  <oliver@apple.com>
3433
3434         Reviewed by Gavin Barraclough.
3435
3436         [JSC] JavaScript parsing error when loading Equifax web page
3437         https://bugs.webkit.org/show_bug.cgi?id=42900
3438
3439         '-->' is ostensibly only meant to occur when there is only
3440         whitespace preceeding it on the line.  However firefox treats
3441         multiline comments as a space character, so they are allowed.
3442         One side effect of the firefox model is that any line terminators
3443         inside the multiline comment are ignored, so
3444
3445             foo/*
3446             */-->
3447
3448         is treated as
3449
3450             foo -->
3451
3452         and so '-->' will not be a comment in this case.  Happily this simply
3453         means that to fix this issue all we need to do is stop updating
3454         m_atLineStart when handling multiline comments.
3455
3456         * parser/Lexer.cpp:
3457         (JSC::Lexer::lex):
3458
3459 2010-08-25  Oliver Hunt  <oliver@apple.com>
3460
3461         Reviewed by Geoffrey Garen.
3462
3463         Improve overflow handling in StringImpl::Replace
3464         https://bugs.webkit.org/show_bug.cgi?id=42502
3465         <rdar://problem/8203794>
3466
3467         Harden StringImpl::replace against overflow -- I can't see how this
3468         could be abused, but it's better to be safe than sorry.
3469
3470         * wtf/text/StringImpl.cpp:
3471         (WTF::StringImpl::replace):
3472
3473 2010-08-26  Martin Robinson  <mrobinson@igalia.com>
3474
3475         Reviewed by Xan Lopez.
3476
3477         [GTK] The GNUmakefile.am files contain a myriad of confusing preprocessor and compiler flag definitions
3478         https://bugs.webkit.org/show_bug.cgi?id=44624
3479
3480         Clean up GNUmakefile.am.
3481
3482         * GNUmakefile.am: Alphabetize the include order in javascriptcore_cppflags. Move
3483         a couple include lines from the top-level GNUmakefile.am.
3484
3485 2010-08-25  Xan Lopez  <xlopez@igalia.com>
3486
3487         Reviewed by Kent Tamura.
3488
3489         Local variables 'k' and 'y' in s2b() in dtoa.cpp are computed but not used
3490         https://bugs.webkit.org/show_bug.cgi?id=29259
3491
3492         Remove unused code in dtoa.cpp, spotted by Wan-Teh Chang.
3493
3494         * wtf/dtoa.cpp:
3495         (WTF::s2b):
3496
3497 2010-08-25  Kwang Yul Seo  <skyul@company100.net>
3498
3499         Reviewed by Kevin Ollivier.
3500
3501         [BREWMP] Add build system
3502         https://bugs.webkit.org/show_bug.cgi?id=44645
3503
3504         Make waf script portable so that we can add more ports.
3505
3506         * wscript:
3507
3508 2010-08-25  Michael Saboff  <msaboff@apple.com>
3509
3510         Reviewed by Sam Weinig.
3511
3512         Remove the single entry regular expression cache introduced as part of
3513         the fix for https://bugs.webkit.org/show_bug.cgi?id=41238.
3514         The performance problem in Dromaeo that initiated that bug is no
3515         longer present. Dromaeo has been modified so that the regular
3516         expression tests are somewhat random and don't benefit from a
3517         single entry cache.
3518
3519         * runtime/RegExp.cpp:
3520         (JSC::RegExp::RegExp):
3521         (JSC::RegExp::match):
3522         * runtime/RegExp.h:
3523
3524 2010-08-25  Martin Robinson  <mrobinson@igalia.com>
3525
3526         Reviewed by Gustavo Noronha Silva.
3527
3528         Cairo and EFL port shouldn't depend on glib.
3529         https://bugs.webkit.org/show_bug.cgi?id=44354
3530
3531         Replace GRefPtr with PlatformRefPtr. Keep GLib specific bits in
3532         GRefPtr.h.
3533
3534         * GNUmakefile.am: Add PlatformRefPtr.h to the source list.
3535         * wtf/PlatformRefPtr.h: Migrated from GRefPtr.h.
3536         (WTF::PlatformRefPtr::PlatformRefPtr): Ditto.
3537         (WTF::PlatformRefPtr::~PlatformRefPtr): Ditto.
3538         (WTF::PlatformRefPtr::clear): Ditto.
3539         (WTF::PlatformRefPtr::get): Ditto.
3540         (WTF::PlatformRefPtr::operator*): Ditto.
3541         (WTF::PlatformRefPtr::operator->): Ditto.
3542         (WTF::PlatformRefPtr::operator!): Ditto.
3543         (WTF::PlatformRefPtr::operator UnspecifiedBoolType): Ditto.
3544         (WTF::PlatformRefPtr::hashTableDeletedValue): Ditto.
3545         (WTF::::operator): Ditto.
3546         (WTF::::swap): Ditto.
3547         (WTF::swap): Ditto.
3548         (WTF::operator==): Ditto.
3549         (WTF::operator!=): Ditto.
3550         (WTF::static_pointer_cast): Ditto.
3551         (WTF::const_pointer_cast): Ditto.
3552         (WTF::getPtr): Ditto.
3553         (WTF::adoptPlatformRef): Ditto.
3554         * wtf/gobject/GRefPtr.cpp: Changes to reflect new names.
3555         (WTF::refPlatformPtr):
3556         (WTF::derefPlatformPtr):
3557         * wtf/gobject/GRefPtr.h: Ditto.
3558         (WTF::refPlatformPtr):
3559         (WTF::derefPlatformPtr):
3560
3561 2010-08-25  Xan Lopez  <xlopez@igalia.com>
3562
3563         Reviewed by Alexey Proskuryakov.
3564
3565         Remove dead code in JSGlobalObject
3566         https://bugs.webkit.org/show_bug.cgi?id=44615
3567
3568         The recursion data member in the JSGlobalObject and its getter
3569         plus inc/dec methods seems to be unused, remove them.
3570
3571         * runtime/JSGlobalObject.cpp:
3572         (JSC::JSGlobalObject::init):
3573         * runtime/JSGlobalObject.h:
3574
3575 2010-08-25  Michael Saboff  <msaboff@apple.com>
3576
3577         Reviewed by Geoffrey Garen.
3578
3579         Changed the initial and subsequent allocation of vector storage to
3580         Array()s. The changes are to limit sparse arrays to 100000 entries
3581         and fixed the sparse map to vector storage conversion to use the
3582         minimum amount of memory needed to store the current number of entries.
3583         These changes address https://bugs.webkit.org/show_bug.cgi?id=43707
3584
3585         * runtime/JSArray.cpp:
3586         (JSC::JSArray::putSlowCase):
3587         (JSC::JSArray::getNewVectorLength):
3588
3589 2010-08-16  Gabor Loki  <loki@webkit.org>