JavaScriptCore/ChangeLog

   1 2006-11-20  Samuel Weinig  <sam@webkit.org>
   2
   3         Reviewed by Maciej.
   4
   5         Fix for http://bugs.webkit.org/show_bug.cgi?id=11647
   6         Fix Win32 build
   7
   8         * kjs/config.h: define NOMINMAX instead of min/max
   9         as themselves.
  10         * wtf/Vector.h: put back hack to ensure that min/max
  11         are not defined as macros.
  12
  13 2006-11-19  Simon Hausmann  <hausmann@kde.org>
  14
  15         Reviewed by Zack.
  16
  17         http://bugs.webkit.org/show_bug.cgi?id=11649
  18         Fix CMake Qt-only build without KDE CMake files
  19
  20         * CMakeLists.txt:
  21         * pcre/CMakeLists.txt:
  22
  23 2006-11-17  Anders Carlsson  <acarlsson@apple.com>
  24
  25         Reviewed by Adam.
  26
  27         Make sure that we always use std::min and std::max instead of macros.
  28
  29         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
  30         * kjs/config.h:
  31         * wtf/Vector.h:
  32
  33 === Safari-521.31 ===
  34
  35 2006-11-12  Geoffrey Garen  <ggaren@apple.com>
  36
  37         Reviewed by Beth Dakin.
  38
  39         Added project-wide setting to disable Microsoft's made-up deprecation
  40         warnings related to std:: functions. (Doesn't have any affect yet,
  41         since we currently disable all deprecation warnings.)
  42
  43         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
  44
  45 2006-11-12  Mark Rowe  <bdash@webkit.org>
  46
  47         Reviewed by Mitz.
  48
  49         Clean up of JavaScriptCore bakefiles.
  50
  51         * JavaScriptCoreSources.bkl:
  52         * jscore.bkl:
  53
  54 2006-11-11  Alexey Proskuryakov  <ap@webkit.org>
  55
  56         Reviewed by Maciej.
  57
  58         http://bugs.webkit.org/show_bug.cgi?id=11508
  59         Undisable some warnings for JSImmediate.h
  60
  61         Fix suggested by Don Gibson.
  62
  63         * kjs/JSImmediate.h:
  64         Re-enable all MSVC warnings, move the remaining runtime checks
  65         to compile-time.
  66
  67 2006-11-10  Zalan Bujtas  <zalan.bujtas@nokia.com>
  68
  69         Reviewed by Maciej.
  70
  71         Added s60/symbian platform defines.
  72         http://bugs.webkit.org/show_bug.cgi?id=11540
  73
  74         * wtf/Platform.h:
  75
  76 === Safari-521.30 ===
  77
  78 2006-11-08  Ada Chan  <adachan@apple.com>
  79
  80         Reviewed by darin.
  81
  82         Added a method to delete all the keys in a HashMap.
  83
  84         * wtf/HashMap.h:
  85         (WTF::deleteAllPairFirsts):
  86         (WTF::deleteAllKeys):
  87
  88 2006-11-07  Anders Carlsson  <acarlsson@apple.com>
  89
  90         Reviewed by Geoff.
  91
  92         * API/JSClassRef.cpp:
  93         (OpaqueJSClass::OpaqueJSClass):
  94         Initialize cachedPrototype to 0.
  95
  96 2006-11-06  Krzysztof Kowalczyk  <kkowalczyk@gmail.com>
  97
  98         Reviewed by Maciej.
  99
 100         Remove warning about garbage after #else. #else clause applies for all
 101         non-mac platforms, not only win.
 102
 103         * kjs/date_object.cpp:
 104
 105 2006-11-06  Mark Rowe  <bdash@webkit.org>
 106
 107         Reviewed by the wonderful Mitz Pettel.
 108
 109         http://bugs.webkit.org/show_bug.cgi?id=11524
 110         Bug 11524: REGRESSION(r9842): Array.prototype.join should use ToString operator rather than calling toString on each element
 111
 112         * kjs/array_object.cpp:
 113         (ArrayProtoFunc::callAsFunction): Use ToString operator on each element rather than calling their toString method.
 114
 115 2006-11-03  Steve Falkenburg  <sfalken@apple.com>
 116
 117         Fix build
 118
 119         * kjs/JSImmediate.h:
 120
 121 2006-11-03  Alexey Proskuryakov  <ap@nypop.com>
 122
 123         Reviewed by Darin.
 124
 125         http://bugs.webkit.org/show_bug.cgi?id=11504
 126         Fix warnings on non 32 bit platforms
 127
 128         * kjs/JSImmediate.h:
 129         (KJS::JSImmediate::NanAsBits):
 130         (KJS::JSImmediate::oneAsBits):
 131         Rewrite in a way that moves runtime checks to compile-time.
 132
 133         (KJS::):
 134         (KJS::JSImmediate::fromDouble):
 135         (KJS::JSImmediate::toDouble):
 136
 137 2006-11-02  George Staikos <staikos@kde.org>
 138
 139         Reviewed by Maciej.
 140
 141         * collector.cpp:
 142         Remove a deprecated pthreads call.
 143
 144 2006-11-02  Anders Carlsson  <acarlsson@apple.com>
 145
 146         Reviewed by Maciej, landed by Anders.
 147
 148         * CMakeLists.txt:
 149         Make KDE support optional.
 150
 151 2006-11-01  Kevin McCullough  <KMcCullough@apple.com>
 152
 153         Reviewed by Brady.
 154
 155         - Fixes many JavaScriptCore tests in other timezones.  The root problem is that on mac localtime() returns historically accurate information for DST, but the JavaScript spec explicitly states to not take into account historical information but rather to interpolate from valid years.
 156
 157         * kjs/DateMath.cpp:
 158         (KJS::equivalentYearForDST):
 159         (KJS::getDSTOffsetSimple):
 160         (KJS::getDSTOffset):
 161
 162 2006-10-31  Geoffrey Garen  <ggaren@apple.com>
 163
 164         Reviewed by Beth.
 165
 166         Fixed http://bugs.webkit.org/show_bug.cgi?id=11477
 167         REGRESSION: GMail crashes in KJS::FunctionImp::callerGetter
 168
 169         * kjs/function.cpp:
 170         (KJS::FunctionImp::argumentsGetter): Removed unnecessary braces.
 171         (KJS::FunctionImp::callerGetter): More logical NULL checking.
 172
 173 2006-10-31  Oliver Hunt  <oliver@apple.com>
 174
 175         Reviewed by Geoff.
 176
 177         Adding definition for PLATFORM(CI)
 178
 179         * wtf/Platform.h:
 180
 181 2006-10-31  Vladimir Olexa  <vladimir.olexa@gmail.com>
 182
 183         Reviewed by Geoff.
 184
 185         http://bugs.webkit.org/show_bug.cgi?id=4166
 186         Function object does not support caller property
 187
 188         Test: fast/js/caller-property.html
 189
 190         * kjs/function.cpp:
 191         (KJS::FunctionImp::callerGetter): added
 192         (KJS::FunctionImp::getOwnPropertySlot): added if statement to handle callerGetter()
 193         * kjs/function.h: added callerGetter() declaration
 194         * kjs/identifier.h: added caller property macro
 195         * tests/mozilla/expected.html:
 196
 197 2006-10-30  Kevin McCullough  <KMcCullough@apple.com>
 198
 199         Reviewed by Adam.
 200
 201         - Fix some timezone issues and JavaScriptCore date tests.  Addresses bugzilla 4930.
 202
 203         * kjs/DateMath.h:
 204         (KJS::GregorianDateTime::GregorianDateTime): Here's the fix, to add parenthesis for order of precedence.
 205         * kjs/date_object.cpp:
 206         (KJS::DateProtoFunc::callAsFunction):
 207         (KJS::DateObjectImp::construct): memset not needed as GregorianDateTime initializes itself.
 208
 209 2006-10-30  Darin Adler  <darin@apple.com>
 210
 211         Reviewed by John Sullivan.
 212
 213         * kjs/SavedBuiltins.h: Added needed include.
 214         * wtf/OwnPtr.h: (WTF::OwnPtr::set): Fixed mistake in assertion.
 215
 216 2006-10-28  Darin Adler  <darin@apple.com>
 217
 218         Reviewed by Maciej.
 219
 220         - renamed PassRefPtr::release to releaseRef to make it clearer that
 221           it's the counterpart of adoptRef, and to make it harder to confuse
 222           it with the safer-to-use RefPtr::release
 223
 224         * kjs/identifier.cpp:
 225         (KJS::CStringTranslator::translate):
 226         (KJS::UCharBufferTranslator::translate):
 227         * kjs/ustring.cpp:
 228         (KJS::UString::Rep::create):
 229         * wtf/PassRefPtr.h:
 230         (WTF::PassRefPtr::PassRefPtr):
 231         (WTF::PassRefPtr::~PassRefPtr):
 232         (WTF::PassRefPtr::get):
 233         (WTF::PassRefPtr::releaseRef):
 234         (WTF::PassRefPtr::operator->):
 235         (WTF::PassRefPtr::operator=):
 236         (WTF::adoptRef):
 237         (WTF::static_pointer_cast):
 238         (WTF::const_pointer_cast):
 239         * wtf/RefPtr.h:
 240         (WTF::RefPtr::RefPtr):
 241         (WTF::RefPtr::operator=):
 242
 243 2006-10-28  Darin Adler  <darin@apple.com>
 244
 245         Reviewed by Steve.
 246
 247         * kjs/grammar.y: Add definitions of YYMALLOC and YYFREE to fix
 248         a warning some people see (not sure why others don't see it).
 249
 250         * JavaScriptCore.vcproj/JavaScriptCore/grammarWrapper.cpp: Touch
 251         this file to force it to re-build grammar.cpp.
 252
 253 2006-10-28  Darin Adler  <darin@apple.com>
 254
 255         Reviewed by Geoff.
 256
 257         - made changes so the code compiles with the highest warning level
 258           under MSVC (disabling some warnings, making some code fixes)
 259
 260         * API/JSCallbackConstructor.cpp:
 261         (KJS::JSCallbackConstructor::construct):
 262         * API/JSCallbackFunction.cpp:
 263         (KJS::JSCallbackFunction::callAsFunction):
 264         * API/JSCallbackObject.cpp:
 265         (KJS::JSCallbackObject::init):
 266         (KJS::JSCallbackObject::construct):
 267         (KJS::JSCallbackObject::callAsFunction):
 268         * API/JSObjectRef.cpp:
 269         (JSPropertyNameArrayGetNameAtIndex):
 270         * API/JSStringRef.cpp:
 271         (JSStringCreateWithCharacters):
 272         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
 273         * bindings/c/c_utility.cpp:
 274         (KJS::Bindings::convertUTF8ToUTF16):
 275         (KJS::Bindings::coerceValueToNPVariantStringType):
 276         (KJS::Bindings::convertValueToNPVariant):
 277         * kjs/DateMath.h:
 278         (KJS::GregorianDateTime::GregorianDateTime):
 279         * kjs/ExecState.h:
 280         (KJS::ExecState::hadException):
 281         * kjs/JSImmediate.h:
 282         (KJS::JSImmediate::fromDouble):
 283         (KJS::JSImmediate::toDouble):
 284         (KJS::JSImmediate::NanAsBits):
 285         (KJS::JSImmediate::oneAsBits):
 286         * kjs/Parser.h:
 287         * kjs/PropertyNameArray.h:
 288         (KJS::PropertyNameArray::size):
 289         * kjs/array_object.cpp:
 290         (ArrayObjectImp::callAsFunction):
 291         * kjs/bool_object.cpp:
 292         (BooleanObjectImp::callAsFunction):
 293         * kjs/collector.cpp:
 294         (KJS::Collector::allocate):
 295         (KJS::Collector::markCurrentThreadConservatively):
 296         (KJS::Collector::collect):
 297         * kjs/completion.h:
 298         (KJS::Completion::isValueCompletion):
 299         * kjs/date_object.cpp:
 300         (KJS::findMonth):
 301         * kjs/debugger.cpp:
 302         (Debugger::sourceParsed):
 303         (Debugger::sourceUnused):
 304         (Debugger::exception):
 305         (Debugger::atStatement):
 306         (Debugger::callEvent):
 307         (Debugger::returnEvent):
 308         * kjs/dtoa.cpp:
 309         * kjs/error_object.cpp:
 310         (ErrorObjectImp::callAsFunction):
 311         (NativeErrorImp::callAsFunction):
 312         * kjs/function.cpp:
 313         (KJS::FunctionImp::processVarDecls):
 314         (KJS::GlobalFuncImp::callAsFunction):
 315         * kjs/function_object.cpp:
 316         (FunctionPrototype::callAsFunction):
 317         * kjs/grammar.y:
 318         * kjs/identifier.cpp:
 319         (KJS::CStringTranslator::translate):
 320         (KJS::Identifier::add):
 321         * kjs/internal.h:
 322         * kjs/lexer.cpp:
 323         (Lexer::lex):
 324         (Lexer::isIdentStart):
 325         (Lexer::isIdentPart):
 326         (isDecimalDigit):
 327         (Lexer::isHexDigit):
 328         (Lexer::isOctalDigit):
 329         (Lexer::matchPunctuator):
 330         (Lexer::singleEscape):
 331         (Lexer::convertOctal):
 332         (Lexer::convertHex):
 333         (Lexer::convertUnicode):
 334         (Lexer::record8):
 335         * kjs/lexer.h:
 336         * kjs/math_object.cpp:
 337         (MathFuncImp::callAsFunction):
 338         * kjs/number_object.cpp:
 339         (integer_part_noexp):
 340         (intPow10):
 341         (NumberProtoFunc::callAsFunction):
 342         (NumberObjectImp::callAsFunction):
 343         * kjs/object.cpp:
 344         (KJS::JSObject::deleteProperty):
 345         (KJS::JSObject::callAsFunction):
 346         (KJS::JSObject::toBoolean):
 347         (KJS::JSObject::toObject):
 348         * kjs/object.h:
 349         (KJS::JSObject::getPropertySlot):
 350         * kjs/property_map.cpp:
 351         (KJS::isValid):
 352         (KJS::PropertyMap::put):
 353         (KJS::PropertyMap::insert):
 354         (KJS::PropertyMap::containsGettersOrSetters):
 355         * kjs/property_map.h:
 356         (KJS::PropertyMap::hasGetterSetterProperties):
 357         * kjs/property_slot.h:
 358         * kjs/string_object.cpp:
 359         (StringInstance::getPropertyNames):
 360         (StringObjectImp::callAsFunction):
 361         (StringObjectFuncImp::callAsFunction):
 362         * kjs/ustring.cpp:
 363         (KJS::UString::Rep::computeHash):
 364         (KJS::UString::UString):
 365         (KJS::UString::from):
 366         (KJS::UString::append):
 367         (KJS::UString::ascii):
 368         (KJS::UString::operator=):
 369         (KJS::UString::find):
 370         (KJS::UString::rfind):
 371         * kjs/ustring.h:
 372         (KJS::UChar::high):
 373         (KJS::UChar::low):
 374         (KJS::UCharReference::low):
 375         (KJS::UCharReference::high):
 376         * kjs/value.cpp:
 377         (KJS::JSValue::toUInt16):
 378         * kjs/value.h:
 379         * pcre/pcre_compile.c:
 380         (get_othercase_range):
 381         * pcre/pcre_exec.c:
 382         (match):
 383         * pcre/pcre_internal.h:
 384         * wtf/HashFunctions.h:
 385         (WTF::intHash):
 386         (WTF::PtrHash::hash):
 387         * wtf/MathExtras.h:
 388         (isnan):
 389         (lround):
 390         (lroundf):
 391         * wtf/StringExtras.h:
 392         (strncasecmp):
 393         * wtf/unicode/icu/UnicodeIcu.h:
 394         (WTF::Unicode::isPrintableChar):
 395
 396 2006-10-26  W. Andy Carrel  <wac@google.com>
 397
 398         Reviewed by Maciej.
 399
 400         - Fix http://bugs.webkit.org/show_bug.cgi?id=7445 /
 401           <rdar://problem/4614195> (and 7253 / <rdar://4694011>) by changing
 402           inline regexps so that they can have \u escaped Unicode sequences and
 403           still work properly.
 404
 405         * kjs/lexer.cpp:
 406         (Lexer::Lexer):
 407         (Lexer::setCode):
 408         (Lexer::shift): Looking ahead one additional character for the benefit
 409         of scanRegExp
 410         (Lexer::scanRegExp): Change code to support unicode escapes in inline
 411         regexps.
 412         * kjs/lexer.h: Extra lookahead added.
 413
 414 === Safari-521.29 ===
 415
 416 2006-10-26  Nikolas Zimmermann  <zimmermann@kde.org>
 417
 418         Reviewed by Darin.
 419
 420         Fix build with older gcc 3.3.4.
 421
 422         * kjs/DateMath.cpp: Remove inline prefix.
 423         (KJS::equivalentYearForDST):
 424
 425 2006-10-26  Darin Adler  <darin@apple.com>
 426
 427         Reviewed by John.
 428
 429         - fix iteration of properties of string objects (found because of a warning
 430           emitted by the MSVC compiler)
 431
 432         * kjs/string_object.cpp: (StringInstance::getPropertyNames): Change code that
 433         wants to format a number as a string to use UString::from. Before it was using
 434         the UString constructor that makes a string from a character!
 435
 436         * kjs/ustring.h:
 437         * kjs/ustring.cpp: Remove the dangerous and not all that helpful UString(char)
 438         constructor.
 439
 440         * kjs/grammar.y: Change code to not depend on the UString(char) constructor.
 441         This is potentially more efficient anyway because we could overload the + operator
 442         some day to handle char* directly instead of creating a UString.
 443
 444         * kjs/nodes2string.cpp: (SourceStream::operator<<): Change code to not depend on
 445         the UString(char) constructor.
 446
 447 2006-10-25  Kevin McCullough  <KMcCullough@apple.com>
 448
 449         Reviewed by Steve (rubber stamp).
 450
 451         - Link against your local build of JavaScriptCore.lib first, this fixes some errors on release builds of testkjs.
 452
 453         * JavaScriptCore.vcproj/testkjs/testkjs.vcproj:
 454
 455 2006-10-25  Geoffrey Garen  <ggaren@apple.com>
 456
 457         Reviewed by Lou.
 458
 459         Removed duplicate symbol declaration.
 460
 461         * JavaScriptCore.xcodeproj/project.pbxproj:
 462         * kjs/grammar.y:
 463
 464 2006-10-24  Steve Falkenburg  <sfalken@apple.com>
 465
 466         Build config change
 467
 468         * JavaScriptCore.vcproj/testkjs/testkjs.vcproj:
 469
 470 2006-10-24  Kevin McCullough  <KMcCullough@apple.com>
 471
 472         Reviewed by Brady.
 473
 474         - Fixes a date formatting issue on win.  Specifically strftime cannot handle some ranges of time so we shift time call strftime and then manipulate the returned string, if needed.
 475
 476         * kjs/date_object.cpp:
 477         (KJS::):
 478         (KJS::formatLocaleDate):
 479         (KJS::DateProtoFunc::callAsFunction):
 480
 481 2006-10-23  Kevin McCullough  <KMcCullough@apple.com>
 482
 483         Reviewed by
 484
 485        - Build fix
 486
 487         * JavaScriptCore.xcodeproj/project.pbxproj:
 488         * kjs/grammar.y:
 489
 490 2006-10-23  Kevin McCullough  <KMcCullough@apple.com>
 491
 492         Reviewed by Maciej.
 493
 494         - Makes the toTM function an operator.  Was going to piggy back on a patch but the patch needs more work.
 495
 496         * kjs/DateMath.cpp:
 497         (KJS::equivalentYearForDST):
 498         * kjs/DateMath.h:
 499         (KJS::GregorianDateTime::operator tm):
 500         * kjs/date_object.cpp:
 501         (KJS::formatTime):
 502         (KJS::DateProtoFunc::callAsFunction):
 503
 504 2006-10-23  Kevin McCullough  <KMcCullough@apple.com>
 505
 506         Reviewed by Maciej.
 507
 508        - Fixes two regressions on win.  Both are stack overflows. For one the number of recursions is capped at 100, and for the other, nested parenthesis pairs are not evaluated (since they would evaluate to whatever is in them anyway).
 509
 510         * kjs/grammar.y:
 511         * kjs/object.cpp:
 512
 513 2006-10-21  Steve Falkenburg  <sfalken@apple.com>
 514
 515         Reviewed by Adam.
 516
 517         Add minimal compatibility with MSVCRT leak checker
 518
 519         * wtf/FastMalloc.h:
 520
 521 2006-10-23  Kevin McCullough  <KMcCullough@apple.com>
 522
 523         Reviewed by Geof.
 524
 525         - Sets the lowercase range correctly in the test and consolidates a variable to make the test more readable.
 526
 527        * tests/mozilla/ecma/String/15.5.4.11-2.js:
 528
 529 2006-10-21  Darin Adler  <darin@apple.com>
 530
 531         Reviewed by Anders.
 532
 533         - http://bugs.webkit.org/show_bug.cgi?id=11377
 534           swap(Vector, Vector) should be O(1) instead of O(n)
 535
 536         * wtf/Vector.h:
 537         (WTF::VectorBuffer::swap): Added.
 538         (WTF::Vector::swap): Added.
 539         (WTF::swap): Added overload that takes two Vector objects.
 540
 541 2006-10-21  Darin Adler  <darin@apple.com>
 542
 543         Reviewed by Adam.
 544
 545         - http://bugs.webkit.org/show_bug.cgi?id=11376
 546           build scripts should invoke make with "-j" option for multiple processors
 547
 548         * JavaScriptCore.xcodeproj/project.pbxproj: Pass -j `sysctl -n hw.ncpu` to make.
 549
 550 2006-10-19  Kevin McCullough  <KMcCullough@apple.com>
 551
 552         Reviewed by Geof.
 553
 554         Changed test to make us pass Georgian case changing for Unicode 4.0 and 5.0.  This incorporates changes from the 1.4 revision of the same mozilla test.
 555         On Tiger we are still using Unicode 4.0 but on win and Leopard we are using Unicode 5.0, so this test currently allows for either answer.
 556
 557         * tests/mozilla/ecma/String/15.5.4.11-2.js:
 558
 559 2006-10-18  Maciej Stachowiak  <mjs@apple.com>
 560
 561         Reviewed by Geoff.
 562
 563         - remove vestiges of KXMLCore name (former name of WTF).
 564
 565         * wtf/Assertions.h:
 566         * wtf/FastMalloc.h:
 567         (operator new):
 568         (operator delete):
 569         (operator new[]):
 570         (operator delete[]):
 571         * wtf/FastMallocInternal.h:
 572         * wtf/Forward.h:
 573         * wtf/GetPtr.h:
 574         * wtf/HashCountedSet.h:
 575         * wtf/HashFunctions.h:
 576         * wtf/HashMap.h:
 577         * wtf/HashSet.h:
 578         * wtf/HashTable.h:
 579         * wtf/HashTraits.h:
 580         * wtf/ListRefPtr.h:
 581         * wtf/MathExtras.h:
 582         * wtf/Noncopyable.h:
 583         * wtf/OwnArrayPtr.h:
 584         * wtf/OwnPtr.h:
 585         * wtf/PassRefPtr.h:
 586         * wtf/Platform.h:
 587         * wtf/RefPtr.h:
 588         * wtf/StringExtras.h:
 589         (snprintf):
 590         * wtf/UnusedParam.h:
 591         * wtf/Vector.h:
 592         * wtf/VectorTraits.h:
 593
 594 2006-10-17  Steve Falkenburg  <sfalken@apple.com>
 595
 596         Reviewed by Maciej.
 597
 598         Adjust include paths
 599
 600         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
 601
 602 2006-10-17  Kevin McCullough  <KMcCullough@apple.com>
 603
 604         Reviewed by Darin.
 605
 606         Fixed a date issue where the UTC offset was not set in win.
 607
 608         * kjs/DateMath.cpp:
 609         (KJS::getDSTOffsetSimple):
 610         (KJS::getDSTOffset):
 611         (KJS::msToGregorianDateTime):
 612         * kjs/DateMath.h:
 613         (KJS::):
 614         (KJS::GregorianDateTime::GregorianDateTime):
 615
 616 2006-10-17  Kevin McCullough  <KMcCullough@apple.com>
 617
 618         Reviewed by Brady.
 619
 620         Fixes a JavaScriptCore math issue on win.
 621
 622         * kjs/math_object.cpp:
 623         (MathFuncImp::callAsFunction):
 624         * wtf/MathExtras.h:
 625         (wtf_atan2):
 626
 627 2006-10-16  Kevin McCullough  <kmccullough@apple.com>
 628
 629         Reviewed by Geof.
 630
 631         Removed unecessary global specifiers.
 632
 633         * kjs/math_object.cpp:
 634         (MathFuncImp::callAsFunction):
 635
 636 2006-10-16  Kevin McCullough  <KMcCullough@apple.com>
 637
 638         Reviewed by John.
 639
 640         Fixes a compile order issue for testkjs on win.
 641
 642         * JavaScriptCore.vcproj/testkjs/testkjs.vcproj:
 643
 644 2006-10-15  Krzysztof Kowalczyk  <kkowalczyk@gmail.com>
 645
 646         Reviewed by Anders.
 647
 648         Remove junk (as gcc calls it) after #else clause.
 649
 650         * wtf/FastMalloc.cpp:
 651         (WTF::do_free):
 652
 653 2006-10-14  Krzysztof Kowalczyk  <kkowalczyk@gmail.com>
 654
 655         Reviewed by Maciej.
 656
 657         Define KXMLCORE_USE_CURL for platforms that wish to use CURL as
 658         networking, and set it for GDK build
 659
 660         * wtf/Platform.h:
 661
 662 2006-10-13  Marvin Decker  <marv.decker@gmail.com>
 663
 664         Reviewed by Kevin McCullough.
 665
 666         Fixes http://bugs.webkit.org/show_bug.cgi?id=11283
 667         Fixes Qt/Linux and Windows build
 668
 669         * kjs/DateMath.cpp:
 670         * kjs/DateMath.h:
 671         * kjs/date_object.cpp:
 672         (KJS::DateProtoFunc::callAsFunction):
 673
 674 2006-10-13  Kevin McCullough  <KMcCullough@apple.com>
 675
 676         Reviewed by Adam, Geoff, Darin.
 677
 678        Fixed displaying the UTC offset and time zone string, as well as renamed the GregorianDateTime structure and clean up.
 679
 680         * ChangeLog:
 681         * kjs/DateMath.cpp:
 682         (KJS::getUTCOffset):
 683         (KJS::getDSTOffsetSimple):
 684         (KJS::gregorianDateTimeToMS):
 685         (KJS::msToGregorianDateTime):
 686         * kjs/DateMath.h:
 687         (KJS::GregorianDateTime::GregorianDateTime):
 688         (KJS::GregorianDateTime::~GregorianDateTime):
 689         (KJS::GregorianDateTime::toTM):
 690         * kjs/date_object.cpp:
 691         (KJS::gmtoffset):
 692         (KJS::formatDate):
 693         (KJS::formatDateUTCVariant):
 694         (KJS::formatTime):
 695         (KJS::fillStructuresUsingTimeArgs):
 696         (KJS::fillStructuresUsingDateArgs):
 697         (KJS::DateInstance::getTime):
 698         (KJS::DateInstance::getUTCTime):
 699         (KJS::DateProtoFunc::callAsFunction):
 700         (KJS::DateObjectImp::construct):
 701         (KJS::DateObjectImp::callAsFunction):
 702         (KJS::DateObjectFuncImp::callAsFunction):
 703         (KJS::parseDate):
 704         * kjs/date_object.h:
 705
 706 2006-10-13  Kevin McCullough  <KMcCullough@apple.com>
 707
 708         Reviewed by Adam.
 709
 710         Gets JavaScripCore tests running on windows.
 711
 712         * Scripts/run-javascriptcore-tests:
 713         * Scripts/webkitdirs.pm:
 714
 715 2006-10-12  Geoffrey Garen  <ggaren@apple.com>
 716
 717         Reviewed by Maciej.
 718
 719         Removed JSObjectMakeWithPrototype, clarified some comments. We really
 720         don't want people to manage their own prototypes, so we don't want an
 721         extra function in the API devoted to just that. People can still manage
 722         their own prototypes if they really want by using JSObjectSetPrototype.
 723
 724         * API/JSClassRef.cpp:
 725         (OpaqueJSClass::createNoAutomaticPrototype):
 726         (OpaqueJSClass::create):
 727         * API/JSClassRef.h:
 728         * API/JSObjectRef.cpp:
 729         (JSClassCreate):
 730         (JSObjectMake):
 731         * API/JSObjectRef.h:
 732         * API/testapi.c:
 733         (main):
 734         * JavaScriptCore.exp:
 735
 736 2006-10-12  Kevin McCullough  <KMcCullough@apple.com>
 737
 738         Reviewed by Adam.
 739
 740         Build breakage fix
 741
 742         * kjs/DateMath.cpp:
 743         (KJS::msToTM):
 744         * kjs/date_object.cpp:
 745         (KJS::gmtoffset):
 746
 747 2006-10-11  Kevin McCullough  <KMcCullough@apple.com>
 748
 749         Reviewed by Geoff.
 750
 751         Added our own tm struct to have a consistent set of fields, which lets us display the DST offset and timezone strings correctly.  Also there is some code cleanup.
 752
 753         * kjs/DateMath.cpp:
 754         (KJS::timeToMS):
 755         (KJS::getUTCOffset):
 756         (KJS::getDSTOffsetSimple):
 757         (KJS::dateToMS):
 758         (KJS::msToTM):
 759         (KJS::tmToKJStm):
 760         (KJS::KJStmToTm):
 761         * kjs/DateMath.h:
 762         * kjs/date_object.cpp:
 763         (KJS::gmtoffset):
 764         (KJS::formatTime):
 765         (KJS::DateProtoFunc::callAsFunction):
 766         (KJS::DateObjectImp::construct):
 767         (KJS::DateObjectImp::callAsFunction):
 768         (KJS::DateObjectFuncImp::callAsFunction):
 769         (KJS::parseDate):
 770         * kjs/date_object.h:
 771
 772 2006-10-09  Krzysztof Kowalczyk  <kkowalczyk@gmail.com>
 773
 774         Reviewed by Geoff.
 775
 776         Improve gdk build compiler flags (show warning, no rtti and exceptions).
 777
 778         * jscore.bkl:
 779
 780 2006-10-06  Kevin McCullough  <KMcCullough@apple.com>
 781
 782         Reviewed by Brady.
 783
 784         DST and TimeZones were wrong in some cases, specifically on some of the dates where DST changes.
 785
 786         * kjs/DateMath.cpp:
 787         (KJS::equivalentYearForDST):
 788         (KJS::getUTCOffset):
 789         (KJS::getDSTOffsetSimple):
 790         (KJS::getDSTOffset):
 791         (KJS::dateToMseconds):
 792         (KJS::msToTM):
 793         * kjs/DateMath.h:
 794         * kjs/date_object.cpp:
 795         (KJS::gmtoffset):
 796
 797 2006-10-05  Darin Adler  <darin@apple.com>
 798
 799         Reviewed by Kevin McCullough.
 800
 801         * wtf/Assertions.cpp: Fix build when _DEBUG is not defined.
 802
 803 2006-10-04  Kevin McCullough  <KMcCullough@apple.com>
 804
 805         Reviewed by Adam.
 806
 807         - Removed an unnecessary assert that was stopping many pages.  tm_gmtoff was not set for UTC time in mozilla but is always set for us.
 808
 809         * kjs/DateMath.cpp:
 810         (KJS::getUTCOffset):
 811         (KJS::msToTM):
 812         * kjs/date_object.cpp:
 813         (KJS::gmtoffset):
 814         (KJS::formatTime):
 815
 816 2006-10-04  Geoffrey Garen  <ggaren@apple.com>
 817
 818         Patch by Darin and me, reviewed by Maciej.
 819
 820         Fixed <rdar://problem/4518397> REGRESSION(?): Oft-seen but unrepro crash
 821               in JavaScript garbage collection (KJS::Collector::collect())
 822                <rdar://problem/4752492> Crash in KJS::collect
 823
 824         The issue here was allocating one garbage-collected object in the midst
 825         of allocating a second garbage-collected object. In such a case, the
 826         zeroIfFree word lies.
 827
 828         * kjs/collector.cpp:
 829         (KJS::Collector::allocate):
 830         (KJS::Collector::collect):
 831
 832 2006-10-04  Kevin McCullough  <KMcCullough@apple.com>
 833
 834         Reviewed by Adam.
 835
 836         - Layout test fix
 837
 838         * kjs/DateMath.cpp:
 839         (KJS::dateToDayInYear): accept and correctly handle negative months
 840
 841 2006-10-05  Kevin McCullough  <KMcCullough@apple.com>
 842
 843        build fix
 844
 845         * kjs/DateMath.cpp:
 846         (KJS::dateToDayInYear):
 847
 848 2006-10-05  Mark Rowe  <bdash@webkit.org>
 849
 850         Reviewed by maculloch.
 851
 852         Gdk build fix.
 853
 854         * JavaScriptCoreSources.bkl: Add DateMath.cpp to file list.
 855
 856 2006-10-05  Kevin McCullough  <KMcCullough@apple.com>
 857
 858         Reviewed by aroben
 859
 860         - build fix
 861
 862         * JavaScriptCore.xcodeproj/project.pbxproj:
 863
 864 2006-10-04  Nikolas Zimmermann  <zimmermann@kde.org>
 865
 866         Reviewed by Mitz.
 867
 868         Fix Qt/Linux build by adding DateMath.cpp to compilation.
 869
 870         * CMakeLists.txt: Also replace tabs with spaces.
 871
 872 2006-10-04  Kevin McCullough  <KMcCullough@apple.com>
 873
 874         Reviewed by DethBakin.
 875
 876         - Apparently the build bot uses an older version of XCode which warns about conversions and the newest version does not.  I hope this fixes the build but I cann't be sure on my system.
 877
 878         * kjs/DateMath.cpp:
 879         (KJS::msToYear):
 880         (KJS::dayInYear):
 881         (KJS::dateToDayInYear):
 882
 883 2006-10-05  Darin Adler  <darin@apple.com>
 884
 885         Reviewed by Adam.
 886
 887         * wtf/Assertions.cpp: Changed assertion formatting to omit the "======"
 888         lines so you can see more assertions in less space. Also improved format
 889         of file/line information so it works with more development environments.
 890
 891 2006-10-04  Kevin McCullough  <KMcCullough@apple.com>
 892
 893         Reviewed by Tim H.
 894
 895         - The build machine is more sensitive about automatic conversions.  These fixes exp
 896 licitly cast or change the input and return types of functions to avoid conversions.
 897
 898         * JavaScriptCore.xcodeproj/project.pbxproj:
 899         * kjs/DateMath.cpp:
 900         (KJS::):
 901         (KJS::msToDays):
 902         (KJS::msToYear):
 903         (KJS::dayInYear):
 904         (KJS::monthToDayInYear):
 905         (KJS::dateToDayInYear):
 906         (KJS::getDSTOffsetSimple):
 907         (KJS::getDSTOffset):
 908         (KJS::dateToMseconds):
 909         (KJS::msToTM):
 910
 911 2006-10-04  Kevin McCullough  <KMcCullough@apple.com>
 912
 913         Reviewed by GGaren
 914
 915         - This is a big makeover for our Date implemenetation.  This solves many platform specific issues, specifically dates before 1970, and simplifies some ugly code.  The purpose of this was to get us to pass many of the JavaScriptCore tests on windows.
 916
 917         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
 918         * JavaScriptCore.xcodeproj/project.pbxproj:
 919         * kjs/DateMath.cpp: Added.
 920         (KJS::):
 921         (KJS::daysInYear):
 922         (KJS::daysFrom1970ToYear):
 923         (KJS::msFrom1970ToYear):
 924         (KJS::msToDays):
 925         (KJS::msToYear):
 926         (KJS::isLeapYear):
 927         (KJS::isInLeapYear):
 928         (KJS::dayInYear):
 929         (KJS::msToMilliseconds):
 930         (KJS::msToWeekDay):
 931         (KJS::msToSeconds):
 932         (KJS::msToMinutes):
 933         (KJS::msToHours):
 934         (KJS::msToMonth):
 935         (KJS::msToDayInMonth):
 936         (KJS::monthToDayInYear):
 937         (KJS::timeToMseconds):
 938         (KJS::dateToDayInYear):
 939         (KJS::equivalentYearForDST):
 940         (KJS::getUTCOffset):
 941         (KJS::getDSTOffsetSimple):
 942         (KJS::getDSTOffset):
 943         (KJS::localTimeToUTC):
 944         (KJS::UTCToLocalTime):
 945         (KJS::dateToMseconds):
 946         (KJS::msToTM):
 947         (KJS::isDST):
 948         * kjs/DateMath.h: Added.
 949         (KJS::):
 950         * kjs/date_object.cpp:
 951         (KJS::gmtoffset):
 952         (KJS::formatTime):
 953         (KJS::DateInstance::getTime):
 954         (KJS::DateInstance::getUTCTime):
 955         (KJS::DateProtoFunc::callAsFunction):
 956         (KJS::DateObjectImp::construct):
 957         (KJS::DateObjectFuncImp::callAsFunction):
 958         (KJS::parseDate):
 959         * kjs/testkjs.cpp:
 960         * os-win32/stdint.h:
 961
 962 2006-10-02  Nikolas Zimmermann  <zimmermann@kde.org>
 963
 964         Reviewed/landed by Adam.
 965
 966         Build testkjs on Qt/Linux.
 967
 968         * CMakeLists.txt:
 969
 970 2006-10-02  Nikolas Zimmermann  <zimmermann@kde.org>
 971
 972         Reviewed by eseidel.  Landed by eseidel.
 973
 974         Fix win32 build, which has no inttypes.h
 975
 976         * wtf/Assertions.h:
 977
 978 2006-10-02  Nikolas Zimmermann <zimmermann@kde.org>
 979
 980         Reviewed by eseidel & mjs.  Landed by eseidel.
 981
 982         Fix Qt/Linux build with older gcc 3.3.4.
 983         http://bugs.webkit.org/show_bug.cgi?id=11116
 984
 985         * kjs/lookup.h: Move cacheGlobalObject into KJS namespace.
 986         (KJS::cacheGlobalObject): Also remove GCC_ROOT_NS_HACK.
 987         * wtf/Assertions.h: Include inttypes.h for uintptr_t.
 988
 989 2006-09-28  Steve Falkenburg  <sfalken@apple.com>
 990
 991         Reviewed by Maciej.
 992
 993         Use $(ConfigSuffix) set via vsprops files to add _debug
 994         to end of debug filenames.
 995
 996         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
 997         * JavaScriptCore.vcproj/debug.vsprops: Added.
 998         * JavaScriptCore.vcproj/dftables/dftables.vcproj:
 999         * JavaScriptCore.vcproj/release.vsprops: Added.
1000         * JavaScriptCore.vcproj/testkjs/testkjs.vcproj:
1001
1002 2006-09-28  Darin Adler  <darin@apple.com>
1003
1004         Reviewed by Alice.
1005
1006         - support for change that should fix <rdar://problem/4733044>
1007           REGRESSION: XML iBench shows 10% perf. regression (copying
1008           strings while decoding)
1009
1010         * wtf/Vector.h: Changed VectorBuffer so that the general case
1011         contains an instance of the 0 case, since deriving from it
1012         was violating the Liskov Substitution Principle.
1013         (WTF::VectorBuffer::releaseBuffer): Added. Releases the buffer so it can
1014         be adopted by another data structure that uses the FastMalloc.h allocator.
1015         Returns 0 if the internal buffer was being used.
1016         (WTF::Vector::releaseBuffer): Added. Releases the buffer as above or creates
1017         a new one in the case where the internal buffer was being used.
1018
1019 2006-09-28  Maciej Stachowiak  <mjs@apple.com>
1020
1021         Reviewed by Geoff.
1022
1023         - change garbage collection to happen at increments proportional to number of live objects, not always
1024         every 1000 allocations
1025
1026         * kjs/collector.cpp:
1027         (KJS::Collector::allocate):
1028
1029 2006-09-28  Maciej Stachowiak  <mjs@apple.com>
1030
1031         Reviewed by Mitz.
1032
1033         - fixed REGRESSION (r16606): javascriptCore Crash on website load
1034
1035         Plus style fixes.
1036
1037             - fixed some possible off-by-one bugs
1038             - use indexing, not iterators, for Vectors
1039             - store Vector by pointer instead of by value to avoid blowing out FunctionImp size
1040
1041         * kjs/function.cpp:
1042         (KJS::FunctionImp::addParameter):
1043         (KJS::FunctionImp::parameterString):
1044         (KJS::FunctionImp::processParameters):
1045         (KJS::FunctionImp::lengthGetter):
1046         (KJS::FunctionImp::getParameterName):
1047         * kjs/function.h:
1048
1049 2006-09-27  Steve Falkenburg  <sfalken@apple.com>
1050
1051         Reviewed by Maciej.
1052
1053         More build tweaks
1054
1055         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
1056         * JavaScriptCore.vcproj/JavaScriptCore/dstroot-to-sdk.cmd: Removed.
1057
1058 2006-09-27  John Sullivan  <sullivan@apple.com>
1059
1060         * kjs/function.cpp:
1061         (KJS::FunctionImp::getParameterName):
1062         removed assertion that displeased gcc 4.0.1 (build 5420):
1063         ASSERT(static_cast<size_t>(index) == index);
1064
1065 2006-09-27  Kevin McCullough  <KMcCullough@apple.com>
1066
1067         Reviewed by GGaren.
1068
1069         Cleanup of previous fix which was to address Radar: 4752492
1070
1071         * kjs/function.cpp:
1072         (KJS::FunctionImp::addParameter):
1073         (KJS::FunctionImp::parameterString):
1074         (KJS::FunctionImp::processParameters):
1075         (KJS::FunctionImp::lengthGetter):
1076         (KJS::FunctionImp::getParameterName):
1077         * kjs/function.h:
1078
1079 2006-09-27  Kevin McCullough  <KMcCullough@apple.com>
1080
1081         Reviewed by Adele.
1082
1083         Fixes a GC stack overflow crash.
1084         The change is to move from a linked list implementation of Parameters to a Vector.
1085         The problem with the linked list is that each one creates it's own stack frame when being destroyed and in extreme cases this caused the stack to overflow.
1086
1087         * kjs/function.cpp:
1088         (KJS::Parameter::Parameter):
1089         (KJS::FunctionImp::addParameter):
1090         (KJS::FunctionImp::parameterString):
1091         (KJS::FunctionImp::processParameters):
1092         (KJS::FunctionImp::lengthGetter):
1093         (KJS::FunctionImp::getParameterName):
1094         * kjs/function.h:
1095
1096 2006-09-27  Steve Falkenburg  <sfalken@apple.com>
1097
1098         Fix last path fix.
1099
1100         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
1101
1102 2006-09-27  Steve Falkenburg  <sfalken@apple.com>
1103
1104         Set path before build.
1105
1106         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
1107
1108 2006-09-27  Sean Gies  <seangies@apple.com>
1109
1110         Reviewed by Adam Roben.
1111
1112         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: Debug config should link to debug runtime.
1113         * JavaScriptCore.vcproj/dftables/dftables.vcproj: Debug config should link to debug runtime.
1114
1115 2006-09-27  Don Melton  <gramps@apple.com>
1116
1117         Reviewed by Adam Roben.
1118
1119         Changed line ending from DOS to UNIX format so it doesn't die running
1120         on my machine. ;)
1121
1122         * JavaScriptCore.vcproj/JavaScriptCore/build-generated-files.sh:
1123
1124 2006-09-23  Alexey Proskuryakov  <ap@nypop.com>
1125
1126         Reviewed by Maciej.
1127
1128         http://bugs.webkit.org/show_bug.cgi?id=10183
1129         REGRESSION: obfuscated JS decoding breaks because of soft hyphen removal
1130         (Fanfiction.net author pages not listing stories)
1131
1132         Rolled out the fix for bug 4139.
1133
1134         * kjs/lexer.cpp:
1135         (Lexer::setCode):
1136         (Lexer::shift):
1137         * tests/mozilla/ecma/Array/15.4.5.1-1.js:
1138         * tests/mozilla/expected.html:
1139
1140 2006-09-22  Steve Falkenburg  <sfalken@apple.com>
1141
1142         Build fix.
1143
1144         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
1145
1146 2006-09-22  Darin Adler  <darin@apple.com>
1147
1148         Reviewed by Alice.
1149
1150         * wtf/Vector.h: Add an append that takes a pointer and length.
1151         Generalize the existing Vector append to work on vectors with
1152         any value for inlineCapacity. Change the append algorithm so
1153         it doesn't check capacity each time through the loop.
1154
1155 2006-09-22  Steve Falkenburg  <sfalken@apple.com>
1156
1157         Fix release build.
1158
1159         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
1160
1161 2006-09-21  Geoffrey Garen  <ggaren@apple.com>
1162
1163         Reviewed by Maciej.
1164
1165         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
1166         Updated to include the right path.
1167         * wtf/FastMalloc.h: #include Platform.h, since we use Platform macros.
1168
1169 === Safari-521.27 ===
1170
1171 2006-09-20  Anders Carlsson  <acarlsson@apple.com>
1172
1173         Reviewed by Dave Hyatt.
1174
1175         * wtf/MathExtras.h:
1176         Get rid of lrint.
1177
1178 2006-09-20  Sean Gies  <seangies@apple.com>
1179
1180         Reviewed by Steve Falkenburg.
1181
1182         * wtf/Assertions.cpp: Debug messages should go into debugger console.
1183
1184 2006-09-20  David Hyatt  <hyatt@apple.com>
1185
1186         Add an implementation of lrint for Win32.
1187
1188         Reviewed by anders
1189
1190         * wtf/MathExtras.h:
1191         (lrint):
1192
1193 2006-09-15  Krzysztof Kowalczyk  <kkowalczyk@gmail.com>
1194
1195         Reviewed by Adam.
1196
1197         http://bugs.webkit.org/show_bug.cgi?id=10864
1198         Bug 10864: Linux\GDK build fixes
1199
1200         * JavaScriptCoreSources.bkl:
1201         * jscore.bkl:
1202
1203 2006-09-15  Adam Roben  <aroben@apple.com>
1204
1205         Windows build fix.
1206
1207         * JavaScriptCore.vcproj/JavaScriptCore/build-generated-files.sh:
1208
1209 2006-09-15  Anders Carlsson  <acarlsson@apple.com>
1210
1211         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
1212         Fix the release build.
1213
1214 2006-09-15  Anders Carlsson  <acarlsson@apple.com>
1215
1216         Reviewed by Steve.
1217
1218         Add JavaScriptCore API to the build.
1219         * API/JSBase.cpp:
1220         * API/JSCallbackConstructor.cpp:
1221         * API/JSCallbackFunction.cpp:
1222         * API/JSCallbackObject.cpp:
1223         * API/JSClassRef.cpp:
1224         * API/JSContextRef.cpp:
1225         * API/JSObjectRef.cpp:
1226         * API/JSStringRef.cpp:
1227         * API/JSValueRef.cpp:
1228         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
1229         * os-win32/stdbool.h: Added.
1230
1231 2006-09-12  Steve Falkenburg  <sfalken@apple.com>
1232
1233         Reviewed by Ada.
1234
1235         Build tweaks (doing JavaScriptCore now since it doesn't have
1236         dependencies).
1237
1238         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
1239         * JavaScriptCore.vcproj/JavaScriptCore/build-generated-files.sh:
1240         * JavaScriptCore.vcproj/JavaScriptCore/dstroot-to-sdk.cmd: Added.
1241         * JavaScriptCore.vcproj/dftables/dftables.vcproj:
1242         * JavaScriptCore.vcproj/testkjs/testkjs.vcproj:
1243
1244 2006-09-11  Brady Eidson  <beidson@apple.com>
1245
1246         Build fix - I think Tim's last checkin wasn't tested on Tiger, possibly.  I simply
1247         commented out the undefined constants until he can have a chance to make the right call
1248
1249         * bindings/objc/objc_utility.mm:
1250         (KJS::Bindings::objcValueTypeForType): Commented out undefined symbols
1251
1252 2006-09-11  Timothy Hatcher  <timothy@apple.com>
1253
1254         Reviewed by Tim O. and Darin.
1255
1256         Add support for more method signatures affecting ObjC methods called from JavaScript:
1257         - Added unsigned types and long long.
1258         - Allow methods that use const, oneway, bycopy and byref type modifiers.
1259
1260         * bindings/objc/objc_instance.mm:
1261         (ObjcInstance::invokeMethod):
1262         * bindings/objc/objc_utility.h:
1263         (KJS::Bindings::):
1264         * bindings/objc/objc_utility.mm:
1265         (KJS::Bindings::convertValueToObjcValue):
1266         (KJS::Bindings::convertObjcValueToValue):
1267         (KJS::Bindings::objcValueTypeForType):
1268
1269 2006-09-05  Timothy Hatcher  <timothy@apple.com>
1270
1271         Reviewed by Tim O.
1272
1273         <rdar://problem/4715840> SEL is not char*
1274
1275         * bindings/objc/objc_class.mm:
1276         (KJS::Bindings::ObjcClass::methodsNamed): use sel_getName instead of a char* cast.
1277         * bindings/objc/objc_runtime.mm:
1278         (ObjcFallbackObjectImp::callAsFunction): ditto
1279
1280 2006-09-03  Alexey Proskuryakov  <ap@nypop.com>
1281
1282         Reviewed by Tim H.
1283
1284         http://bugs.webkit.org/show_bug.cgi?id=10693
1285         Convert JavaScript arrays to AppleScript lists
1286
1287         * JavaScriptCore.exp: Export ArrayInstance::info and ArrayInstance::getItem().
1288         * kjs/array_instance.h:
1289         * kjs/array_object.cpp:
1290         (ArrayInstance::getItem): Added a method to access array items from C++.
1291
1292 2006-09-02  Krzysztof Kowalczyk  <kkowalczyk@gmail.com>
1293
1294         Reviewed by Tim H.
1295
1296         Bug 10454: Unix bakefile fixes
1297         http://bugs.webkit.org/show_bug.cgi?id=10454
1298
1299         * JavaScriptCoreSources.bkl:
1300
1301 2006-09-01  Nikolas Zimmermann  <zimmermann@kde.org>
1302
1303         Reviewed by hyatt.  Landed by eseidel.
1304
1305         Fix build on Linux.
1306
1307         * pcre/CMakeLists.txt: Add wtf/ include.
1308
1309 2006-09-01  Nikolas Zimmermann  <zimmermann@kde.org>
1310
1311         Reviewed and landed by ap.
1312
1313         Fix build on Linux (C89 without gcc extensions enabled).
1314
1315         * pcre/pcre_internal.h: Use C style comments.
1316         * wtf/Assertions.h: Use C style comments.
1317         * wtf/Platform.h: Use C style comments.
1318
1319 2006-09-01  Steve Falkenburg  <sfalken@apple.com>
1320
1321         Fix build.
1322
1323         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
1324         * JavaScriptCore.vcproj/dftables/dftables.vcproj:
1325
1326 2006-08-31  Anders Carlsson  <acarlsson@apple.com>
1327
1328         Reviewed by Darin.
1329
1330         Add new portability functions to MathExtras.h and add StringExtras.h which is for
1331         string portability functions.
1332
1333         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
1334         * bindings/c/c_instance.cpp:
1335         * kjs/date_object.cpp:
1336         * wtf/MathExtras.h:
1337         (copysign):
1338         (isfinite):
1339         * wtf/StringExtras.h: Added.
1340         (snprintf):
1341         (strncasecmp):
1342
1343 2006-08-31  Anders Carlsson  <acarlsson@apple.com>
1344
1345         Reviewed by Tim H.
1346
1347         Fix Windows build.
1348
1349         * JavaScriptCore.vcproj/dftables/dftables.vcproj:
1350         * pcre/pcre_internal.h:
1351
1352 2006-08-31  Timothy Hatcher  <timothy@apple.com>
1353
1354         Reviewed by Geoff.
1355
1356         Band-aid fix for PCRE to compile for ppc64 and x86_64 now that
1357         we use -Wshorten-64-to-32. Adds an INT_CAST macro that ASSERTs
1358         the value <= INT_MAX.
1359
1360         I filed <rdar://problem/4712064> to track the need to verify
1361         PCRE's 64-bit compliance.
1362
1363         * pcre/pcre_compile.c:
1364         (complete_callout):
1365         (compile_branch):
1366         (compile_regex):
1367         (pcre_compile2):
1368         * pcre/pcre_exec.c:
1369         (match):
1370         (pcre_exec):
1371         * pcre/pcre_get.c:
1372         (pcre_get_substring_list):
1373         * pcre/pcre_internal.h:
1374         * pcre/pcre_tables.c:
1375         * pcre/pcre_try_flipped.c:
1376         (_pcre_try_flipped):
1377
1378 2006-08-30  Darin Adler  <darin@apple.com>
1379
1380         Reviewed by Tim Hatcher.
1381
1382         - add WTF::getPtr, a function template that makes it possible to write
1383           generic code that gets a raw pointer out of any of our pointer types
1384
1385         * JavaScriptCore.xcodeproj/project.pbxproj:
1386         * wtf/GetPtr.h: Added.
1387         * wtf/ListRefPtr.h: (WTF::getPtr): Added.
1388         * wtf/OwnArrayPtr.h: (WTF::getPtr): Added.
1389         * wtf/OwnPtr.h: (WTF::getPtr): Added.
1390         * wtf/PassRefPtr.h: (WTF::getPtr): Added.
1391         * wtf/RefPtr.h: (WTF::getPtr): Added.
1392
1393 2006-08-29  waylonis  <waylonis@google.com>
1394
1395         Reviewed, tweaked by ggaren.
1396
1397         - Added storage and accessor functions for ExecState as a fix for
1398           http://bugs.webkit.org/show_bug.cgi?id=10114
1399
1400         * kjs/ExecState.cpp:
1401         (KJS::ExecState::ExecState):
1402         * kjs/ExecState.h:
1403         * kjs/context.h:
1404         (KJS::Context::setExecState):
1405         (KJS::Context::execState):
1406
1407 2006-08-30  Nikolas Zimmermann  <zimmermann@kde.org>
1408
1409         Reviewed by Tim H.
1410
1411         Commit KDE related tweaks, to be able to
1412         differentiate between a Qt-only or a KDE build.
1413
1414         * CMakeLists.txt: Install wtf-unity library.
1415         * wtf/Platform.h: Add define for the KDE platform.
1416
1417 2006-08-28  Darin Adler  <darin@apple.com>
1418
1419         Reviewed by Geoff.
1420
1421         * kjs/list.h: Use explicit in constructor (as appropriate).
1422
1423 2006-08-24  Nikolas Zimmermann  <zimmermann@kde.org>
1424
1425         Reviewed, tweaked and landed by ap
1426
1427         http://bugs.webkit.org/show_bug.cgi?id=10467
1428         WebKit should have Qt platform support (Part II)
1429
1430         * CMakeLists.txt: Adjust to Anders' build fixes.
1431         * wtf/Platform.h: Fix define for the Qt platform (we don't use/need Cairo.)
1432
1433 2006-08-23  David Hyatt  <hyatt@apple.com>
1434
1435         Fix Platform.h to include #defines for graphics features.
1436
1437         Reviewed by darin
1438
1439         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
1440         * wtf/Platform.h:
1441
1442 2006-08-23  Anders Carlsson  <acarlsson@apple.com>
1443
1444         Reviewed by Darin.
1445
1446         Make the bindings compile without CoreFoundation.
1447
1448         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
1449         * bindings/c/c_instance.cpp:
1450         * bindings/c/c_utility.cpp:
1451         (KJS::Bindings::convertUTF8ToUTF16):
1452         * bindings/npapi.h:
1453         * bindings/runtime.cpp:
1454         (KJS::Bindings::Instance::createBindingForLanguageInstance):
1455         (KJS::Bindings::Instance::createLanguageInstanceForValue):
1456         * bindings/runtime_root.cpp:
1457         * bindings/runtime_root.h:
1458         * kjs/interpreter.cpp:
1459         (KJS::Interpreter::createLanguageInstanceForValue):
1460         * kjs/interpreter.h:
1461
1462 2006-08-22  Anders Carlsson  <acarlsson@apple.com>
1463
1464         Reviewed by Darin.
1465
1466         Move the npruntime code over to using HashMap and the runtime_root code over to using
1467         HashMap and HashCountedSet.
1468
1469         * bindings/NP_jsobject.cpp:
1470         * bindings/c/c_utility.cpp:
1471         (KJS::Bindings::identifierFromNPIdentifier):
1472         * bindings/c/c_utility.h:
1473         * bindings/jni/jni_jsobject.cpp:
1474         (JavaJSObject::invoke):
1475         * bindings/npruntime.cpp:
1476         (getStringIdentifierMap):
1477         (getIntIdentifierMap):
1478         (_NPN_GetStringIdentifier):
1479         (_NPN_GetIntIdentifier):
1480         * bindings/runtime_root.cpp:
1481         (getReferencesByRootMap):
1482         (getReferencesSet):
1483         (KJS::Bindings::findReferenceSet):
1484         (KJS::Bindings::rootForImp):
1485         (KJS::Bindings::rootForInterpreter):
1486         (KJS::Bindings::addNativeReference):
1487         (KJS::Bindings::removeNativeReference):
1488         (RootObject::removeAllNativeReferences):
1489         * bindings/runtime_root.h:
1490
1491 2006-08-22  Anders Carlsson  <acarlsson@apple.com>
1492
1493         Reviewed by Geoff.
1494
1495         Switch over the NPAPI and Java bindings to using HashMaps instead of dictionaries.
1496
1497         * JavaScriptCore.xcodeproj/project.pbxproj:
1498         * bindings/c/c_class.cpp:
1499         (KJS::Bindings::CClass::CClass):
1500         (KJS::Bindings::CClass::~CClass):
1501         (KJS::Bindings::CClass::classForIsA):
1502         (KJS::Bindings::CClass::methodsNamed):
1503         (KJS::Bindings::CClass::fieldNamed):
1504         * bindings/c/c_class.h:
1505         * bindings/jni/jni_class.cpp:
1506         (JavaClass::JavaClass):
1507         (JavaClass::~JavaClass):
1508         (JavaClass::methodsNamed):
1509         (JavaClass::fieldNamed):
1510         * bindings/jni/jni_class.h:
1511         * bindings/objc/objc_class.h:
1512         * bindings/objc/objc_class.mm:
1513         (KJS::Bindings::deleteMethod):
1514         (KJS::Bindings::deleteField):
1515         (KJS::Bindings::):
1516         (KJS::Bindings::ObjcClass::methodsNamed):
1517         (KJS::Bindings::ObjcClass::fieldNamed):
1518         * bindings/runtime.cpp:
1519         * bindings/runtime.h:
1520         * bindings/runtime_object.cpp:
1521         (RuntimeObjectImp::fieldGetter):
1522         (RuntimeObjectImp::getOwnPropertySlot):
1523         (RuntimeObjectImp::put):
1524         (RuntimeObjectImp::canPut):
1525
1526 2006-08-21  Vladimir Olexa  <vladimir.olexa@gmail.com>
1527
1528         Reviewed by Darin.
1529
1530         http://bugs.webkit.org/show_bug.cgi?id=6252
1531         JavaScript 1.6 Array.lastIndexOf
1532
1533         Test: fast/js/array-lastIndexOf.html
1534
1535         * kjs/array_object.cpp:
1536         (ArrayProtoFunc::callAsFunction): Added a LastIndexOf case.
1537         * kjs/array_object.h:
1538         (KJS::ArrayProtoFunc::): Added LastIndexOf to enum.
1539         * tests/mozilla/expected.html: Two more tests now pass.
1540
1541 2006-08-20  Nikolas Zimmermann  <zimmermann@kde.org>
1542
1543         Reviewed by Maciej. Landed by rwlbuis.
1544
1545         Fixes parts of: http://bugs.webkit.org/show_bug.cgi?id=10463
1546         WebKit should have Qt platform support
1547
1548         Removing obsolete QConstString/QString constructors in kjs code.
1549
1550         * kjs/identifier.h:
1551         * kjs/ustring.h:
1552
1553 2006-08-17  Nikolas Zimmermann  <zimmermann@kde.org>
1554
1555         Reviewed by Maciej. Landed by rwlbuis.
1556
1557         Fixes: http://bugs.webkit.org/show_bug.cgi?id=10463
1558         WTF Changes needed for Qt platform code.
1559
1560         * wtf/Platform.h:
1561         * wtf/unicode/UnicodeDecomposition.h: Added.
1562         (WTF::Unicode::):
1563         * wtf/unicode/UnicodeDirection.h: Added.
1564         (WTF::Unicode::):
1565         * wtf/unicode/qt4/UnicodeQt4.cpp: Added.
1566         (WTF::Unicode::direction):
1567         (WTF::Unicode::category):
1568         (WTF::Unicode::decomposition):
1569         * wtf/unicode/qt4/UnicodeQt4.h:
1570         (WTF::Unicode::toLower):
1571         (WTF::Unicode::toUpper):
1572         (WTF::Unicode::isPrintableChar):
1573         (WTF::Unicode::isSpace):
1574         (WTF::Unicode::isPunct):
1575         (WTF::Unicode::isDigit):
1576         (WTF::Unicode::mirroredChar):
1577         (WTF::Unicode::compare):
1578
1579 2006-08-17  Nikolas Zimmermann  <zimmermann@kde.org>
1580
1581         Reviewed by Eric. Landed by rwlbuis.
1582
1583         Fixes: http://bugs.webkit.org/show_bug.cgi?id=10464
1584         Offer a cmake build system for Qt platform.
1585
1586         * CMakeLists.txt: Added.
1587         * pcre/CMakeLists.txt: Added.
1588
1589 2006-08-17  Anders Carlsson  <acarlsson@apple.com>
1590
1591         Reviewed by Maciej.
1592
1593         * bindings/npapi.h:
1594         Fix ifdef.
1595
1596 2006-08-15  Steve Falkenburg  <sfalken@apple.com>
1597
1598         Reviewed by mjs.
1599
1600         Build fix.
1601
1602         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
1603         * wtf/Assertions.h:
1604
1605 2006-08-15  Mark Rowe  <opendarwin.org@bdash.net.nz>
1606
1607         Reviewed by Tim H.
1608
1609         Build fix:  DWARF and -gfull are incompatible with symbol separation.
1610
1611         * JavaScriptCore.xcodeproj/project.pbxproj:
1612
1613 2006-08-15  Mark Rowe  <opendarwin.org@bdash.net.nz>
1614
1615         Reviewed by Tim H.
1616
1617         http://bugs.webkit.org/show_bug.cgi?id=10394
1618         Bug 10394: WebKit Release and Production configurations should enable dead code stripping
1619
1620         * JavaScriptCore.xcodeproj/project.pbxproj:
1621
1622 2006-08-15  Mark Rowe  <opendarwin.org@bdash.net.nz>
1623
1624         Reviewed by Tim H.
1625
1626         http://bugs.webkit.org/show_bug.cgi?id=10384
1627         Bug 10384: Switch to DWARF for Release configuration
1628
1629         * JavaScriptCore.xcodeproj/project.pbxproj:
1630
1631 2006-08-13  Maks Orlovich  <maksim@kde.org>
1632
1633         Reviewed (and tweaked a little) by Maciej.
1634
1635         - shrank the size of JSObject by 8 bytes and made the corresponding reduction to the cell size, resulting
1636         in a 1.2% speed improvement on JS iBench (and probably overall memory savings).
1637
1638         This was done by removing _scope and _internalValue data members
1639         from JSObject and moving them only to the subclasses that actually
1640         make use of them.
1641
1642         * kjs/object.cpp:
1643         (KJS::JSObject::mark): No need to mark scope or internal value here.
1644         * kjs/object.h:
1645         (KJS::JSObject::JSObject): Don't initialize them.
1646         * kjs/JSWrapperObject.cpp: Added. New base class for object types that
1647         wrap primitive values (Number, String, Boolean, Date).
1648         (KJS::JSWrapperObject::mark):
1649         * kjs/JSWrapperObject.h: Added.
1650         (KJS::JSWrapperObject::JSWrapperObject):
1651         (KJS::JSWrapperObject::internalValue):
1652         (KJS::JSWrapperObject::setInternalValue):
1653         * kjs/array_object.cpp:
1654         (ArrayPrototype::ArrayPrototype): Don't set useless internal value.
1655         * kjs/bool_object.cpp:
1656         (BooleanInstance::BooleanInstance): Inherit from JSWrapperObject.
1657         (BooleanProtoFunc::callAsFunction): Fixed to account for fact that not all
1658         JSObjects have an internal value.
1659         (BooleanObjectImp::construct): ditto.
1660         * kjs/bool_object.h:
1661         * kjs/collector.cpp: Lowered cell size to 48.
1662         (KJS::Collector::allocate): meaningless whitespace change
1663         * kjs/date_object.cpp:
1664         (KJS::DateInstance::DateInstance): Inherit from JSWrapperObject.
1665         (KJS::DateProtoFunc::callAsFunction): adjusted for move of internalValue
1666         (KJS::DateObjectImp::construct): ditto
1667         * kjs/date_object.h:
1668         * kjs/error_object.cpp:
1669         (ErrorPrototype::ErrorPrototype): don't set internal value
1670         * kjs/function.cpp: move _scope and related handling here
1671         (KJS::FunctionImp::mark): mark scope
1672         * kjs/function.h:
1673         (KJS::FunctionImp::scope): moved here from JSObject
1674         (KJS::FunctionImp::setScope): ditto
1675         * kjs/number_object.cpp:
1676         (NumberInstance::NumberInstance): inherit from JSWrapperObject
1677         (NumberProtoFunc::callAsFunction): adjusted
1678         (NumberObjectImp::construct): adjusted
1679         * kjs/number_object.h: shring RegExp-related objects a little
1680         * kjs/regexp_object.cpp:
1681         (RegExpPrototype::RegExpPrototype): Adjust for size tweaks
1682         (RegExpObjectImp::RegExpObjectImp): ditto
1683         * kjs/regexp_object.h:
1684         * kjs/string_object.cpp:
1685         (StringInstance::StringInstance): inherit from JSWrapperObject
1686         (StringProtoFunc::callAsFunction): adjusted
1687         * kjs/string_object.h:
1688         * JavaScriptCore.exp: Exported new methods as needed.
1689         * JavaScriptCore.xcodeproj/project.pbxproj: Added new files to build.
1690
1691 2006-08-04  Brady Eidson  <beidson@apple.com>
1692
1693         Reviewed by Geoff's rubber stamp
1694
1695         Fix a build break on Intel hardware causes by adapting stricter
1696         compiler warnings (-Wshorten-64-to-32)
1697
1698         * API/testapi.c:
1699         (assertEqualsAsNumber): manually cast some doubles to floats
1700         (main): ditto
1701
1702 2006-08-04  Sam Weinig  <sam.weinig@gmail.com>
1703
1704         Reviewed by Darin.
1705
1706         - patch for http://bugs.webkit.org/show_bug.cgi?id=10192
1707           Make WebCore (and friends) compile with -Wshorten-64-to-32
1708
1709           * Adds -Wshorten-64-to-32 flag to Xcode project.
1710           * Adds explicit casts where OK.
1711
1712         * API/JSNodeList.c:
1713         (JSNodeList_item):
1714         (JSNodeList_getProperty):
1715         * JavaScriptCore.xcodeproj/project.pbxproj:
1716
1717 2006-08-04  Adam Roben  <aroben@apple.com>
1718
1719         Reviewed by Anders.
1720
1721         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: Convert
1722         spaces to tabs
1723
1724 2006-08-03  Sam Weinig  <sam.weinig@gmail.com>
1725
1726         Reviewed by Darin.
1727
1728         - patch for http://bugs.webkit.org/show_bug.cgi?id=10176
1729           Make WebCore compile with -Wundef
1730
1731           * Adds -Wundef flag to Xcode project
1732           * Converts #ifs to #ifdef and #ifndefs where needed.
1733           * Added #define YYMAXDEPTH 10000 in kjs/grammar.y
1734             to fix a warning from within Bison.
1735
1736         * JavaScriptCore.xcodeproj/project.pbxproj:
1737         * bindings/jni/jni_jsobject.cpp:
1738         (JavaJSObject::getSlot):
1739         (JavaJSObject::setSlot):
1740         * bindings/npapi.h:
1741         * bindings/objc/objc_class.mm:
1742         (KJS::Bindings::ObjcClass::methodsNamed):
1743         (KJS::Bindings::ObjcClass::fieldNamed):
1744         * bindings/objc/objc_instance.mm:
1745         (ObjcInstance::invokeMethod):
1746         * bindings/objc/objc_runtime.mm:
1747         (ObjcMethod::getMethodSignature):
1748         (ObjcField::name):
1749         (ObjcField::type):
1750         * kjs/grammar.y:
1751         * kjs/identifier.h:
1752
1753 2006-08-03  Anders Carlsson  <acarlsson@apple.com>
1754
1755         Reviewed by John Sullivan.
1756
1757        * wtf/HashSet.h:
1758         (WTF::::operator):
1759         Return *this in operator=
1760
1761 2006-08-03  Adam Roben  <aroben@apple.com>
1762
1763         Reviewed by Anders.
1764
1765         - Fixed Windows build
1766
1767         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
1768         * wtf/MathExtras.h: Implement inline versions of these functions
1769         (nextafter):
1770         (nextafterf):
1771
1772 2006-08-02  Adam Roben  <aroben@apple.com>
1773
1774         Reviewed by Darin.
1775
1776         - Fixed build
1777
1778         * kjs/date_object.cpp:
1779         (KJS::formatTime):
1780
1781 2006-07-29  Darin Adler  <darin@apple.com>
1782
1783         - Removed tabs from these source files that still had them.
1784           We don't use them; that way source files look fine in editors
1785           that have tabs set to 8 spaces or to 4 spaces.
1786         - Removed allow-tabs Subversion property from the files too.
1787
1788         * bindings/NP_jsobject.cpp:
1789         * bindings/c/c_utility.cpp:
1790         * bindings/jni/jni_runtime.cpp:
1791         * bindings/jni/jni_utility.cpp:
1792         * bindings/objc/objc_utility.mm:
1793         * bindings/runtime.cpp:
1794         * bindings/runtime_method.cpp:
1795         * bindings/testbindings.cpp:
1796         * bindings/testbindings.mm:
1797         * kjs/date_object.cpp:
1798         * kjs/function.cpp:
1799         * kjs/list.cpp:
1800         * kjs/nodes.cpp:
1801         * kjs/nodes.h:
1802         * kjs/string_object.cpp:
1803         * kjs/ustring.cpp:
1804
1805 2006-07-29  Darin Adler  <darin@apple.com>
1806
1807         * tests/mozilla/expected.html: Update test results now that regress-185165.js
1808         is succeeding. I suspect Anders fix for bug 4620655 is the reason.
1809
1810 2006-07-29  Sam Weinig  <sam.weinig@gmail.com>
1811
1812         Reviewed by Darin.
1813
1814         - patch for http://bugs.webkit.org/show_bug.cgi?id=10080
1815           Adopt pedantic changes from the Unity project to improve
1816           cross-compiler compatibility
1817
1818           Changes include:
1819           * Removing trailing semicolon from namespace braces.
1820           * Removing trailing comma from last enum declaration.
1821           * Updating to match style guidelines.
1822           * Adding missing newline to the end of the file.
1823           * Turning on gcc warning for missing newline at the end of a source file
1824             (GCC_WARN_ABOUT_MISSING_NEWLINE in Xcode, -Wnewline in gcc).
1825           * Alphabetical sorting of Xcode source list files.
1826           * Replace use of non-portable variable-size array with Vector.
1827           * Use C-style comments instead of C++ comments in files that might
1828             be included by either C or C++ files.
1829
1830         * API/JSCallbackConstructor.cpp:
1831         (KJS::JSCallbackConstructor::construct):
1832         * API/JSCallbackFunction.cpp:
1833         (KJS::JSCallbackFunction::callAsFunction):
1834         * API/JSCallbackObject.cpp:
1835         (KJS::JSCallbackObject::construct):
1836         (KJS::JSCallbackObject::callAsFunction):
1837         * JavaScriptCore.xcodeproj/project.pbxproj:
1838         * JavaScriptCorePrefix.h:
1839         * bindings/jni/jni_class.cpp:
1840         (JavaClass::fieldNamed):
1841         * bindings/jni/jni_class.h:
1842         * bindings/jni/jni_instance.cpp:
1843         (JavaInstance::JavaInstance):
1844         (JavaInstance::valueOf):
1845         * bindings/jni/jni_objc.mm:
1846         (KJS::Bindings::dispatchJNICall):
1847         * bindings/jni/jni_runtime.cpp:
1848         (JavaParameter::JavaParameter):
1849         (JavaArray::JavaArray):
1850         * bindings/jni/jni_runtime.h:
1851         * bindings/jni/jni_utility.h:
1852         * bindings/objc/objc_instance.h:
1853         * bindings/runtime_array.h:
1854         * kjs/collector.h:
1855         * kjs/config.h:
1856         * kjs/ustring.cpp:
1857         * wtf/Platform.h:
1858
1859 2006-07-29  Mike Emmel  <mike.emmel@gmail.com>
1860
1861         Reviewed by Darin.
1862
1863         - fixes for Linux build
1864
1865         * JavaScriptCoreSources.bkl: Added new files to build, kjs/PropertyNameArray.cpp
1866         and kjs/testkjs.cpp, and removed old files.
1867
1868 2006-07-24  Dan Waylonis  <waylonis@google.com>
1869
1870         Reviewed and tweaked a bit by Darin.
1871
1872         - fix http://bugs.webkit.org/show_bug.cgi?id=9902
1873           jsNull and NSNull not properly converted between JS and ObjC
1874
1875         * bindings/objc/objc_utility.mm: (KJS::Bindings::convertObjcValueToValue):
1876         Added case for converting NSNull to jsNull.
1877
1878 2006-07-24  Rob Buis  <buis@kde.org>
1879
1880         Reviewed by Darin.
1881
1882         http://bugs.webkit.org/show_bug.cgi?id=4258
1883         Date().toString() only includes GMT offset, not timezone string
1884
1885         Use the info in tm_zone to append timezone abbreviation
1886         to Date().toString().
1887
1888         * kjs/date_object.cpp:
1889         (KJS::formatTime):
1890
1891 2006-07-24  Rob Buis  <buis@kde.org>
1892
1893         Reviewed by Darin.
1894
1895         http://bugs.webkit.org/show_bug.cgi?id=5257
1896         setYear() does not match FireFox/IE behavior
1897
1898         Make sure the right values end up in tm_year.
1899
1900         * kjs/date_object.cpp:
1901         (KJS::formatTime):
1902
1903 2006-07-23  Mark Rowe  <opendarwin.org@bdash.net.nz>
1904
1905         Reviewed by Maciej.
1906
1907         Bug 9686: [Drosera] Need the ability to break into Drosera on Javascript exceptions
1908         http://bugs.webkit.org/show_bug.cgi?id=9686
1909
1910         JavaScriptCore portion of the fix.
1911
1912         * JavaScriptCore.exp: Update symbol for change in argument type.
1913         * kjs/debugger.cpp:
1914         (Debugger::detach): Clear map of recent exceptions.
1915         (Debugger::hasHandledException): Track the most recent exception
1916         thrown by an interpreter.
1917         (Debugger::exception): Change exception argument to a JSValue.
1918         * kjs/debugger.h:
1919         * kjs/nodes.cpp:
1920         (Node::debugExceptionIfNeeded): Notify the debugger of an exception
1921         if it hasn't seen it before.
1922         (ThrowNode::execute): Notify the debugger that an exception is being thrown.
1923         * kjs/nodes.h:
1924
1925     2006-07-23  Geoffrey Garen  <ggaren@apple.com>
1926
1927         Patch by Eric Albert, reviewed by Darin and me.
1928
1929         - Fixed <rdar://problem/4645931> JavaScriptCore stack-scanning code
1930         crashes (Collector::markStackObjectsConservatively)
1931
1932         * bindings/jni/jni_jsobject.cpp: On 64bit systems, jint is a long, not an
1933         int.
1934         (JavaJSObject::getSlot):
1935         (JavaJSObject::setSlot):
1936         * kjs/collector.cpp:
1937         (KJS::Collector::markCurrentThreadConservatively): Use a pointer instead of
1938         an int as 'dummy,' because on LP64 systems, an int is not pointer-aligned,
1939         and we want to scan the stack for pointers.
1940         * JavaScriptCore.xcodeproj/project.pbxproj: After a tense cease-fire, the
1941         XCode war has started up again!
1942
1943 === Safari-521.20 ===
1944
1945 2006-07-21  Geoffrey Garen  <ggaren@apple.com>
1946
1947         Reviewed by Darin.
1948
1949         <rdar://problem/4507265> REGRESSION: overlays don't work on HousingMaps.com (Google Maps-based site)
1950
1951         - Added support for strings that masquerade as undefined. Currently used
1952         by WebCore to implement undetectable style.filter.
1953
1954         The name is a little long, but it's only used in one line of code, so I
1955         thought clarity should win over brevity.
1956
1957         * JavaScriptCore.exp:
1958         * JavaScriptCore.xcodeproj/project.pbxproj:
1959         * kjs/object.h:
1960         * kjs/string_object.h:
1961         (KJS::StringInstanceThatMasqueradesAsUndefined::StringInstanceThatMasqueradesAsUndefined):
1962         (KJS::StringInstanceThatMasqueradesAsUndefined::masqueradeAsUndefined):
1963         (KJS::StringInstanceThatMasqueradesAsUndefined::toBoolean):
1964
1965 === Safari-521.19 ===
1966
1967 2006-07-20  Steve Falkenburg  <sfalken@apple.com>
1968
1969         Fix the build
1970
1971         * kjs/function.cpp:
1972         (KJS::escapeStringForPrettyPrinting):
1973
1974 2006-07-19  Anders Carlsson  <acarlsson@apple.com>
1975
1976         Reviewed by Darin.
1977
1978         <rdar://problem/4620655> REGRESSION(10.4.7-10.5): preview button for a blogger.com post doesn't work
1979
1980         * kjs/nodes2string.cpp:
1981         (StringNode::streamTo):
1982         Return the escaped string.
1983
1984         (RegExpNode::streamTo):
1985         Use the correct syntax.
1986
1987         * kjs/function.cpp:
1988         (KJS::escapeStringForPrettyPrinting):
1989         * kjs/function.h:
1990         Add escape function which escapes a string for pretty-printing so it can be parsed again.
1991
1992         * wtf/unicode/icu/UnicodeIcu.h:
1993         (WTF::Unicode::isPrintableChar):
1994         New function.
1995
1996 2006-07-18  Maciej Stachowiak  <mjs@apple.com>
1997
1998         Reviewed by Adele Peterson.
1999
2000         <rdar://problem/4589530> REGRESSION: null character in JS string causes parse error (works in Tiger and in other browsers)
2001
2002         * kjs/lexer.cpp:
2003         (Lexer::shift):
2004         (Lexer::lex):
2005         (Lexer::record16):
2006         (Lexer::scanRegExp):
2007         * kjs/lexer.h:
2008
2009 2006-07-18  Tim Omernick  <timo@apple.com>
2010
2011         Reviewed by Tim Hatcher.
2012
2013         Removed a misleading comment; we recently added support for the NPNVPluginElementNPObject
2014         variable.
2015
2016         * bindings/npapi.h:
2017
2018 === Safari-521.18 ===
2019
2020 2006-07-18  Timothy Hatcher  <timothy@apple.com>
2021
2022         Made the following headers public:
2023
2024         * JavaScriptCore.h
2025         * JSBase.h
2026         * JSContextRef.h
2027         * JSObjectRef.h
2028         * JSStringRef.h
2029         * JSValueRef.h
2030
2031         * JavaScriptCore.xcodeproj/project.pbxproj:
2032
2033 2006-07-17  Geoffrey Garen  <ggaren@apple.com>
2034
2035         Reviewed by Maciej.
2036
2037         - Added automatic prototype creation for classes.
2038
2039         A class stores a weak reference to a prototype, which is cleared when
2040         the prototype is garbage collected, to avoid a reference cycle.
2041
2042         We now have an attributes field in JSClassDefinition, that currently is
2043         used only to override automatic prototype creation when you want to manage your
2044         own prototypes, but can be extended in the future for other nefarious purposes.
2045
2046         Similarly, we have JSObjectMake and JSObjectMakeWithPrototype, the latter
2047         allowing you to manage your own prototypes.
2048
2049         JSObjectMakeConstructor is more interesting now, able to make a constructor
2050         on your behalf if you just give it a class.
2051
2052         - Removed bogus old code from minidom.js.
2053
2054         - Tweaked the headerdocs.
2055
2056         - Added more GC testing, which caught some leaks, and tested more funny
2057         edge cases in lookup, which caught a lookup bug. Removed some testing
2058         we used to do with MyObject because it was redundant with the new, cool
2059         stuff.
2060
2061         While fixing the lookup bug I retracted this change:
2062
2063             "If a static setProperty callback returns 'false', to indicate that the
2064             property was not set, we no longer forward the set request up the class
2065             chain, because that's almost certainly not what the programmer expected."
2066
2067         Returning false when setting a static property is a little silly, but you can see
2068         it being useful when shadowing a base class's static properties, and, regardless
2069         of usefullness, this is the defined behavior of the setProperty callback.
2070
2071         - Plus a little ASCII art, for the kids.
2072
2073 2006-07-17  Timothy Hatcher  <timothy@apple.com>
2074
2075         Reviewed by Maciej.
2076
2077         <rdar://problem/4634874> WebScriptObject and WebUndefined are no longer defined by WebKit
2078
2079         Moves WebScriptObject and WebUndefined up to WebCore.
2080         This change does create an upwards-dependancy on WebScriptObject existing
2081         in the loaded process, but this code path in JavaScriptCore does not get used
2082         unless it is through WebKit/WebCore. Moving all of the binding code out of
2083         JavaScriptCore might make sense in the future.
2084
2085         * JavaScriptCore.exp:
2086         * JavaScriptCore.xcodeproj/project.pbxproj:
2087         * bindings/objc/WebScriptObject.h: Replaced.
2088         * bindings/objc/WebScriptObject.mm: Removed.
2089         * bindings/objc/WebScriptObjectPrivate.h: Removed.
2090         * bindings/objc/objc_class.h:
2091         * bindings/objc/objc_instance.h:
2092         * bindings/objc/objc_instance.mm:
2093         (ObjcInstance::~ObjcInstance):
2094         * bindings/objc/objc_runtime.h:
2095         * bindings/objc/objc_runtime.mm:
2096         (convertValueToObjcObject):
2097         * bindings/objc/objc_utility.mm:
2098         (KJS::Bindings::convertValueToObjcValue):
2099         (KJS::Bindings::convertObjcValueToValue):
2100         (KJS::Bindings::createObjcInstanceForValue):
2101
2102 2006-07-17  Darin Adler  <darin@apple.com>
2103
2104         * API/JSBase.h: Fix comment formatting where things used to be lined up but
2105         are now ragged. Got rid of spaces that attempted to line things up.
2106         * API/JSObjectRef.h: Ditto. Also add missing periods for a couple of comments.
2107
2108 2006-07-17  Geoffrey Garen  <ggaren@apple.com>
2109
2110         Reviewed by Maciej.
2111
2112         - Removed the exception parameter from the initialize callback and, by extension,
2113         JSObjectMake. We have never had a need for exceptions when iniitializing,
2114         so the parameter seemed likely to "get in the way."
2115
2116         Also, an exception in JavaScript is thrown in response to input --
2117         "invalid URL", "index not a number", etc., so it's the job of the
2118         constructor function, not the initialize method, to throw.
2119
2120         If initialize *really* wants to throw, it can communicate the throw to
2121         the constructor through the constructed object's private data (e.g., set
2122         it to NULL, signaling to the consntructor that initialization failed.)
2123
2124         - Added JSObjectMakeWithData, which enables a constructor to set private
2125         data on an object *before* it has been initialized. That way, the initialize
2126         methods can properly operate on the data.
2127
2128         * API/JSNode.c: Moved ref into the initialize method, for better encapsulation,
2129         now that it's possible.
2130         * API/JSNodeList.c: ditto
2131         * API/minidom.c:
2132         (main): Do more aggressive garbage collection to test ref/deref and
2133         initialize/finalize.
2134         * API/minidom.js: store childNodes in a temporary so it doesn't get re-created
2135         like a thousand times. This makes debugging ref/deref easier
2136
2137 2006-07-17  Geoffrey Garen  <ggaren@apple.com>
2138
2139         Reviewed by Maciej.
2140
2141         - Changed the initialize callback to run from least derived class (parent
2142         class) to most derived class. This enables C++ style initialization,
2143         and derived class overriding of member data.
2144
2145         - Added excpetion propopgation to JSObjectMake, to support initialize
2146         exceptions, and generally round out our policy of making function
2147         signatures as long as possible.
2148
2149         * API/JSCallbackObject.h: Use ExecState instead of ContextRef, cuz we're
2150         in C++ land now.
2151
2152 2006-07-17  Geoffrey Garen  <ggaren@apple.com>
2153
2154         Reviewed by Maciej.
2155
2156         - Changed JSObjectMakeConstructor to JSObjectMakeConstructorWithCallback,
2157         to match JSObjectMakeFunctionWithCallback.
2158
2159         - Added prototype parameter, so the generated constructor
2160         automatically works with hasInstance / instanceof
2161
2162         - Moved hasInstance implementation from InternalFunctionImp to JSObject
2163         so that subclasses can inherit it without inheriting function-related baggage.
2164         More refactoring here would be good, but this seems like a good short-term
2165         solution.
2166
2167         (KJS::JSCallbackFunction::implementsHasInstance): override and return false,
2168         because callback functions aren't constructors.
2169
2170 2006-07-17  Maciej Stachowiak  <mjs@apple.com>
2171
2172         Reviewed by Geoff.
2173
2174         - add a JSContextRef parameter to all JSValueRef, JSObjectRef, and JSContextRef operations;
2175         except JSObject{Get,Set}PrivateData which can be assumed to be simple pure accessors.
2176
2177         Also renamed the parameter "context" to "ctx" because it makes the code read better with this pervasive
2178         but usually uninteresting parameter.
2179
2180         * API/JSBase.cpp:
2181         (JSEvaluateScript):
2182         (JSCheckScriptSyntax):
2183         (JSGarbageCollect):
2184         * API/JSBase.h:
2185         * API/JSCallbackObject.cpp:
2186         (KJS::JSCallbackObject::JSCallbackObject):
2187         (KJS::JSCallbackObject::init):
2188         (KJS::JSCallbackObject::getOwnPropertySlot):
2189         (KJS::JSCallbackObject::put):
2190         (KJS::JSCallbackObject::deleteProperty):
2191         (KJS::JSCallbackObject::toNumber):
2192         (KJS::JSCallbackObject::toString):
2193         * API/JSContextRef.cpp:
2194         (JSGlobalContextCreate):
2195         (JSGlobalContextRetain):
2196         (JSGlobalContextRelease):
2197         (JSContextGetGlobalObject):
2198         * API/JSContextRef.h:
2199         * API/JSNode.c:
2200         (JSNodePrototype_appendChild):
2201         (JSNodePrototype_removeChild):
2202         (JSNodePrototype_replaceChild):
2203         (JSNode_getNodeType):
2204         (JSNode_getFirstChild):
2205         (JSNode_prototype):
2206         * API/JSNodeList.c:
2207         (JSNodeListPrototype_item):
2208         (JSNodeList_length):
2209         (JSNodeList_getProperty):
2210         (JSNodeList_prototype):
2211         * API/JSObjectRef.cpp:
2212         (JSObjectMake):
2213         (JSObjectMakeFunctionWithCallback):
2214         (JSObjectMakeConstructor):
2215         (JSObjectMakeFunction):
2216         (JSObjectGetPrototype):
2217         (JSObjectSetPrototype):
2218         (JSObjectHasProperty):
2219         (JSObjectGetProperty):
2220         (JSObjectSetProperty):
2221         (JSObjectGetPropertyAtIndex):
2222         (JSObjectSetPropertyAtIndex):
2223         (JSObjectDeleteProperty):
2224         (JSObjectIsFunction):
2225         (JSObjectCallAsFunction):
2226         (JSObjectIsConstructor):
2227         (JSObjectCallAsConstructor):
2228         (JSObjectCopyPropertyNames):
2229         * API/JSObjectRef.h:
2230         * API/JSStringRef.cpp:
2231         * API/JSValueRef.cpp:
2232         (JSValueGetType):
2233         (JSValueIsUndefined):
2234         (JSValueIsNull):
2235         (JSValueIsBoolean):
2236         (JSValueIsNumber):
2237         (JSValueIsString):
2238         (JSValueIsObject):
2239         (JSValueIsObjectOfClass):
2240         (JSValueIsEqual):
2241         (JSValueIsStrictEqual):
2242         (JSValueIsInstanceOfConstructor):
2243         (JSValueMakeUndefined):
2244         (JSValueMakeNull):
2245         (JSValueMakeBoolean):
2246         (JSValueMakeNumber):
2247         (JSValueMakeString):
2248         (JSValueToBoolean):
2249         (JSValueToNumber):
2250         (JSValueToStringCopy):
2251         (JSValueToObject):
2252         (JSValueProtect):
2253         (JSValueUnprotect):
2254         * API/JSValueRef.h:
2255         * API/minidom.c:
2256         (print):
2257         * API/testapi.c:
2258         (MyObject_getProperty):
2259         (MyObject_deleteProperty):
2260         (MyObject_callAsFunction):
2261         (MyObject_callAsConstructor):
2262         (MyObject_convertToType):
2263         (print_callAsFunction):
2264         (main):
2265
2266 2006-07-16  Geoffrey Garen  <ggaren@apple.com>
2267
2268         Approved by Maciej, RS by Beth.
2269
2270         JSObjectMakeFunction -> JSObjectMakeFunctionWithCallback
2271         JSObjectMakeFunctionWithBody -> JSObjectMakeFunction
2272
2273         because the latter is more common, and more fundamental, than the former.
2274
2275         * API/APICast.h:
2276         (toJS):
2277         * API/JSBase.h:
2278         * API/JSCallbackObject.cpp:
2279         (KJS::JSCallbackObject::getOwnPropertySlot):
2280         (KJS::JSCallbackObject::put):
2281         (KJS::JSCallbackObject::deleteProperty):
2282         (KJS::JSCallbackObject::getPropertyNames):
2283         (KJS::JSCallbackObject::staticValueGetter):
2284         (KJS::JSCallbackObject::staticFunctionGetter):
2285         * API/JSClassRef.cpp:
2286         (OpaqueJSClass::OpaqueJSClass):
2287         (OpaqueJSClass::~OpaqueJSClass):
2288         * API/JSClassRef.h:
2289         * API/JSObjectRef.cpp:
2290         (JSClassCreate):
2291         (JSObjectMakeFunctionWithCallback):
2292         (JSObjectMakeFunction):
2293         (OpaqueJSPropertyNameArray::OpaqueJSPropertyNameArray):
2294         (JSObjectCopyPropertyNames):
2295         * API/JSObjectRef.h:
2296         * API/minidom.c:
2297         (main):
2298         * API/testapi.c:
2299         (main):
2300         * ChangeLog:
2301         * JavaScriptCore.exp:
2302
2303 2006-07-16  Geoffrey Garen  <ggaren@apple.com>
2304
2305         Laughed at by Beth.
2306
2307         Replace __JS with OpaqueJS because the former, while used by CF, is
2308         a prefix that's triply-reserved by the compiler. (_* is reserved in global
2309         names, _[A-Z] is reserved in all names, and __ is reserved in all names
2310         in C++.)
2311
2312         Opaque is an alternative used by other Mac OS X framewokrs.
2313
2314         * API/APICast.h:
2315         (toJS):
2316         * API/JSBase.h:
2317         * API/JSCallbackObject.cpp:
2318         (KJS::JSCallbackObject::getOwnPropertySlot):
2319         (KJS::JSCallbackObject::put):
2320         (KJS::JSCallbackObject::deleteProperty):
2321         (KJS::JSCallbackObject::getPropertyNames):
2322         (KJS::JSCallbackObject::staticValueGetter):
2323         (KJS::JSCallbackObject::staticFunctionGetter):
2324         * API/JSClassRef.cpp:
2325         (OpaqueJSClass::OpaqueJSClass):
2326         (OpaqueJSClass::~OpaqueJSClass):
2327         * API/JSClassRef.h:
2328         * API/JSObjectRef.cpp:
2329         (JSClassCreate):
2330         (OpaqueJSPropertyNameArray::OpaqueJSPropertyNameArray):
2331         (JSObjectCopyPropertyNames):
2332
2333 2006-07-16  Darin Adler  <darin@apple.com>
2334
2335         - try to fix Windows build
2336
2337         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
2338         Added some recently added files, removed some recently removed.
2339
2340 2006-07-16  Geoffrey Garen  <ggaren@apple.com>
2341
2342         Reviewed by Maciej.
2343
2344         - Change getProperty* to return undefined, rather than NULL, for missing
2345         properties, since that's what the spec says. Also added exception out
2346         parameters to the *Index functions, because they can call through to the
2347         regular functions, which can throw for custom objects.
2348
2349         * API/JSObjectRef.cpp:
2350         (JSObjectGetProperty):
2351         (JSObjectGetPropertyAtIndex):
2352         (JSObjectSetPropertyAtIndex):
2353         * API/JSObjectRef.h:
2354         * API/testapi.c:
2355         (main):
2356
2357 2006-07-16  Geoffrey Garen  <ggaren@apple.com>
2358
2359         Reviewed by Maciej.
2360
2361         - Properly document and handle NULL callbacks for static properties. We
2362         throw an exception in any case other than a ReadOnly property with a NULL
2363         setProperty callback, because a NULL callback almost certainly indicates
2364         a programming error. Also throw an exception if hasProperty returns true
2365         for a property that getProperty can't get.
2366
2367         - If a static setProperty callback returns 'false', to indicate that the
2368         property was not set, we no longer forward the set request up the class
2369         chain, because that's almost certainly not what the programmer expected.
2370
2371         * API/JSCallbackObject.cpp:
2372         (KJS::JSCallbackObject::getOwnPropertySlot):
2373         (KJS::JSCallbackObject::put):
2374         (KJS::JSCallbackObject::staticValueGetter):
2375         (KJS::JSCallbackObject::staticFunctionGetter):
2376         (KJS::JSCallbackObject::callbackGetter):
2377         * API/JSObjectRef.h:
2378         * API/minidom.js:
2379         * API/testapi.c:
2380         (MyObject_hasProperty):
2381         * API/testapi.js:
2382
2383 2006-07-16  Geoffrey Garen  <ggaren@apple.com>
2384
2385         Reviewed by Maciej.
2386
2387         - Added names to functions.
2388
2389         - Removed GetPrivate/SetPrivate from callbackFunctions and callbackConstructors.
2390         The private data idiom is that a JS object stores its native implementation
2391         as private data. For functions and constructors, the native implementation is nothing
2392         more than the callback they already store, so supporting private data, too,
2393         confuses the idiom. If you *really* want, you can still create a custom
2394         function with private data.
2395
2396         * API/JSCallbackConstructor.cpp:
2397         * API/JSCallbackConstructor.h:
2398         * API/JSCallbackFunction.cpp:
2399         (KJS::JSCallbackFunction::JSCallbackFunction):
2400         * API/JSCallbackFunction.h:
2401         * API/JSCallbackObject.cpp:
2402         (KJS::JSCallbackObject::staticFunctionGetter):
2403         * API/JSObjectRef.cpp:
2404         (JSObjectMakeFunction):
2405         (JSObjectMakeFunctionWithBody):
2406         (JSObjectGetPrivate):
2407         (JSObjectSetPrivate):
2408         * API/JSObjectRef.h:
2409         * API/minidom.c:
2410         (main):
2411         * API/testapi.c:
2412         (main):
2413
2414 2006-07-15  Maciej Stachowiak  <mjs@apple.com>
2415
2416         Reviewed by Darin.
2417
2418         - switch property lists to be vector+set of Identifiers instead of list of References
2419
2420         This has the following benefits:
2421
2422         - no duplicates in property lists
2423         - simplifies API calls
2424         - probably more efficient, since linked list is gone
2425         - entirely removed Reference, ReferenceList and ProtectedReference types from the API
2426
2427         * kjs/PropertyNameArray.cpp: Added.
2428         (KJS::PropertyNameArray::add): Check set, if not already there, add to
2429         vector.
2430         * kjs/PropertyNameArray.h: Added.
2431         (KJS::PropertyNameArray::PropertyNameArray): Newly added type, combines
2432         a set and a vector to make a unique but ordered list of identifiers.
2433         (KJS::PropertyNameArray::begin): ditto
2434         (KJS::PropertyNameArray::end): ditto
2435         (KJS::PropertyNameArray::size): ditto
2436         (KJS::PropertyNameArray::operator[]): ditto
2437         * kjs/array_instance.h:
2438         * kjs/array_object.cpp:
2439         (ArrayInstance::getPropertyNames): renamed from getPropertyList, updated
2440         for PropertyNameArray
2441         (ArrayInstance::setLength): updated for PropertyNameArray
2442         (ArrayInstance::pushUndefinedObjectsToEnd): ditto
2443         * kjs/nodes.cpp:
2444         (ForInNode::execute): updated for PropertyNameArray
2445         * kjs/nodes.h:
2446         * kjs/object.cpp:
2447         (KJS::JSObject::getPropertyNames): renamed from getPropertyList, updated
2448         for PropertyNameArray
2449         * kjs/object.h:
2450         * kjs/property_map.cpp:
2451         (KJS::PropertyMap::getEnumerablePropertyNames): updated for PropertyNameArray
2452         (KJS::PropertyMap::getSparseArrayPropertyNames): ditto
2453         * kjs/property_map.h:
2454         * kjs/protected_reference.h: Removed.
2455         * kjs/reference.cpp: Removed.
2456         * kjs/reference.h: Removed.
2457         * kjs/reference_list.cpp: Removed.
2458         * kjs/reference_list.h: Removed.
2459         * kjs/scope_chain.cpp:
2460         (KJS::ScopeChain::print): Use PropertyNamesArray instead of ReferenceList.
2461         * kjs/string_object.cpp:
2462         (StringInstance::getPropertyNames): Updated for new approach.
2463         * kjs/string_object.h:
2464         * kjs/ustring.h:
2465         * API/APICast.h:
2466         (toJS): Added overload for PropertyNameAccumulatorRef / PropertyNameArray*
2467         (toRef): ditto
2468         * API/JSBase.h:
2469         * API/JSCallbackObject.cpp:
2470         (KJS::JSCallbackObject::getPropertyNames): Fixed for new API.
2471         * API/JSCallbackObject.h:
2472         * API/JSObjectRef.cpp:
2473         (__JSPropertyNameArray::__JSPropertyNameArray): Type used for a publicly vended
2474         JSPropertyNameArrayRef.
2475         (JSObjectCopyPropertyNames): New API call - renamed / refactored from
2476         JSObjectCreatePropertyList
2477         (JSPropertyNameArrayRetain): new retain call for JSPropertyNameArray.
2478         (JSPropertyNameArrayRelease): new release call for - " -.
2479         (JSPropertyNameArrayGetCount): Instead of having to use a stateful enumerator you
2480         can now get the count and items in any order.
2481         (JSPropertyNameArrayGetNameAtIndex): See above.
2482         (JSPropertyNameAccumulatorAddName): What you add properties to is now an opaque
2483         accumulator object.
2484         * API/JSObjectRef.h: Prototyped new functions, removed old ones
2485         * JavaScriptCore.exp: Updated exported symbols.
2486         * JavaScriptCore.xcodeproj/project.pbxproj: Added new files, removed old.
2487         * API/testapi.c:
2488         (MyObject_getPropertyNames): Renamed / fixed callback to fit new paradigm.
2489         (main): Updated for new API.
2490
2491 2006-07-15  Darin Adler  <darin@apple.com>
2492
2493         - oops, missed a few more arrays that had to be const
2494
2495         * API/JSNode.c:
2496         (JSNodePrototype_appendChild): Added const.
2497         (JSNodePrototype_removeChild): Ditto.
2498         (JSNodePrototype_replaceChild): Ditto.
2499         (JSNode_construct): Ditto.
2500         * API/JSNodeList.c:
2501         (JSNodeListPrototype_item): Ditto.
2502         * API/JSObjectRef.cpp:
2503         (JSObjectMakeFunctionWithBody): Ditto.
2504         (JSObjectCallAsFunction): Ditto.
2505         (JSObjectCallAsConstructor): Ditto.
2506         * API/minidom.c:
2507         (print): Ditto.
2508         * API/testapi.c:
2509         (MyObject_callAsFunction): Ditto.
2510         (MyObject_callAsConstructor): Ditto.
2511         (print_callAsFunction): Ditto.
2512         (myConstructor_callAsConstructor): Ditto.
2513
2514 2006-07-15  Darin Adler  <darin@apple.com>
2515
2516         Reviewed by Maciej.
2517
2518         * API/JSNode.h: Made an array parameter const.
2519         * API/JSObjectRef.h: Made array parameters const. Fixed a comment.
2520
2521 2006-07-15  Geoffrey Garen  <ggaren@apple.com>
2522
2523         Reviewed by Maciej.
2524
2525         - JSObjectMakeFunctionWithBody includes a function name and named parameters now.
2526
2527         * API/JSObjectRef.cpp:
2528         (JSObjectMakeFunctionWithBody):
2529         * API/JSObjectRef.h:
2530         * API/testapi.c:
2531         (assertEqualsAsUTF8String): More informative failure reporting.
2532         (main): Test more function cases.
2533
2534 2006-07-15  Geoffrey Garen  <ggaren@apple.com>
2535
2536         Reviewed by Maciej.
2537
2538         - Moved the arguments passed to JSClassCreate into a single structure,
2539         called JSClassDefinition. This will enable easier structure
2540         migration/versioning in the future, if necessary.
2541
2542         - Added support for class names.
2543
2544         - kJSClassDefinitionNull replaces kJSObjectCallbacksNone.
2545
2546         - JSClass is becoming a fairly complex struct, so I migrated all of its
2547         implementation other than reference counting to the sruct.
2548
2549         - Also moved JSClass* functions in the API to JSObjectRef.cpp, since they're
2550         declared in JSObjectRef.h
2551
2552         - Also added some more informative explanation to the class structure doc.
2553
2554 2006-07-15  Darin Adler  <darin@apple.com>
2555
2556         Reviewed by Geoff.
2557
2558         - fix http://bugs.webkit.org/show_bug.cgi?id=8395
2559           <rdar://problem/4613467>
2560           REGRESSION: RegEx seems broken for hex escaped non breaking space
2561
2562         Test: fast/js/regexp-extended-characters-more.html
2563
2564         * pcre/pcre_exec.c:
2565         (match): Got rid of utf16Length local variable to guarantee there's no
2566         extra stack usage in recursive calls. Fixed two places in the PCRE_UTF16
2567         code that were using the length variable, which is the UTF-8 length of
2568         a character in the pattern, to move in the UTF-16 subject string. Instead
2569         they hardcode lengths of 1 and 2 since the code already handles BMP
2570         characters and surrogate pairs separately. Also fixed some DPRINTF so
2571         I could compile with DEBUG on.
2572         (pcre_exec): Changed a place that was checking for multibyte characters
2573         in the subject string to use ISMIDCHAR. Instead it was using hardcoded
2574         logic that was right for UTF-8 but wrong for UTF-16.
2575
2576         * pcre/pcre_compile.c: (pcre_compile2): Fixed a DPRINTF so I could compile
2577         with DEBUG on.
2578
2579 2006-07-14  Geoffrey Garen  <ggaren@apple.com>
2580
2581         RS by Maciej.
2582
2583         Global replace in the API of argc/argv with argumentCount/arguments.
2584
2585 2006-07-14  Geoffrey Garen  <ggaren@apple.com>
2586
2587         Reviewed by Maciej.
2588
2589         - Finalized exception handling in the API.
2590
2591         setProperty can throw because it throws for built-in arrays. getProperty
2592         and deleteProperty can throw because setProperty can throw and we want
2593         to be consistent, and also because they seem like "actions." callAsFunction,
2594         callAsConstructor, and hasInstance can throw, because they caan throw for
2595         all built-ins.
2596
2597         toBoolean can't throw because it's defined that way in the spec.
2598
2599         - Documented that toBoolean and toObject can't be overridden by custom
2600         objects because they're defined that way in the spec.
2601
2602 === Safari-521.17 ===
2603
2604 2006-07-14  Geoffrey Garen  <ggaren@apple.com>
2605
2606         Reviewed by Maciej.
2607
2608         - Implemented ref-counting of JSContexts by splitting into two datatypes:
2609         JSGlobalContext, which you can create/retain/release, and JSContext, which
2610         you can't.
2611
2612         Internally, you retain a JSGlobalContext/ExecState by retaining its
2613         interpreter, which, in the case of a global ExecState, owns it.
2614
2615         - Also made ~Interpreter() protected to catch places where Interpreter
2616         is manually deleted. (Can't make it private because some crazy fool
2617         decided it would be a good idea to subclass Interpreter in other frameworks.
2618         I pity da fool.)
2619
2620         * API/APICast.h:
2621         (toJS): Added cast for new JSGlobalContext
2622         * API/JSStringRef.h: Changed vague "you must" language to more specific
2623         (but, ultimately, equally vague) "behavior is undefined if you don't"
2624         language.
2625         (KJS::Interpreter::Interpreter): Factored more common initialization into
2626         init()
2627         * kjs/interpreter.h:
2628         (KJS::Interpreter::ref): new
2629         (KJS::Interpreter::deref): new
2630         (KJS::Interpreter::refCount): new
2631         * kjs/testkjs.cpp:
2632         (doIt): Ref-count the interpreter.
2633
2634 2006-07-14  Maciej Stachowiak  <mjs@apple.com>
2635
2636         Reviewed by Geoff.
2637
2638         - removed bool return value from JSObjectSetProperty, since it is inefficient and
2639         also doesn't work quite right
2640         - added JSObjectGetPropertyAtIndex and JSObjectSetPropertyAtIndex
2641
2642         * API/JSObjectRef.cpp:
2643         (JSObjectSetProperty): Removed return value and canPut stuff.
2644         (JSObjectGetPropertyAtIndex): Added.
2645         (JSObjectSetPropertyAtIndex): Added.
2646         * API/JSObjectRef.h: Prototyped and documented new functions.
2647
2648 2006-07-14  Geoffrey Garen  <ggaren@apple.com>
2649
2650         Reviewed by Beth.
2651
2652         Moved JSCheckScriptSyntax, JSEvaluateScript, and JSGarbageCollect into
2653         JSBase.h/.cpp. They don't belong in the value-specific or context-specific
2654         files because they're not part of the value or context implementations.
2655
2656         * API/JSBase.h:
2657         * API/JSContextRef.cpp:
2658         (JSContextGetGlobalObject):
2659         * API/JSContextRef.h:
2660         * API/JSValueRef.cpp:
2661         (JSValueUnprotect):
2662         * API/JSValueRef.h:
2663         * JavaScriptCore.xcodeproj/project.pbxproj:
2664
2665 2006-07-13  Timothy Hatcher  <timothy@apple.com>
2666
2667         Reviewed by Maciej.
2668
2669         Moved JavaScriptCore to be a public framework.
2670
2671         * JavaScriptCore.xcodeproj/project.pbxproj:
2672
2673 2006-07-13  Mark Rowe  <opendarwin.org@bdash.net.nz>
2674
2675         Reviewed by Geoffrey.
2676
2677         http://bugs.webkit.org/show_bug.cgi?id=9742
2678         Bug 9742: REGRESSION: WebKit hangs when loading <http://www.vtbook.com>
2679
2680         * kjs/value.h:
2681         (KJS::JSValue::getUInt32): Only types tagged as numeric can be converted to UInt32.
2682
2683 2006-07-13  Geoffrey Garen  <ggaren@apple.com>
2684
2685         Pleasing to Maciej.
2686
2687         - Renamed JSEvaluate -> JSEvaluateScript, JSCheckSyntax -> JSCheckScriptSyntax
2688         - Added exception out parameters to JSValueTo* and JSValueIsEqual because
2689         they can throw
2690         - Removed JSObjectGetDescription because it's useless and vague, and
2691         JSValueToString/JSValueIsObjectOfClass do a better job, anyway
2692         - Clarified comments about "IsFunction/Constructor" to indicate that they
2693         are true of all functions/constructors, not just those created by JSObjectMake*
2694
2695 2006-07-12  Geoffrey Garen  <ggaren@apple.com>
2696
2697         RS by Beth.
2698
2699         Finished previously approved JSInternalString -> JSString conversion
2700         by renaming the files.
2701
2702         * API/JSCallbackObject.cpp:
2703         * API/JSInternalStringRef.cpp: Removed.
2704         * API/JSInternalStringRef.h: Removed.
2705         * API/JSStringRef.cpp: Added.
2706         * API/JSStringRef.h: Added.
2707         * API/JavaScriptCore.h:
2708         * JavaScriptCore.xcodeproj/project.pbxproj:
2709
2710 2006-07-12  Geoffrey Garen  <ggaren@apple.com>
2711
2712         Reviewed by Maciej.
2713
2714         - Removed context and exception parameters from JSObjectGetPropertyEnumerator,
2715         removing the spurious use of ExecState inside JavaScriptCore that made
2716         us think this was necessary in the first place.
2717
2718         (StringInstance::getPropertyList): Use getString instead of toString because
2719         we know we're dealing with a string -- we put it there in the first place.
2720         While we're at it, store the string's size instead of retrieving it each time
2721         through the loop, to avoid the unnecessary killing of puppies.
2722         * kjs/string_object.h:
2723
2724 2006-07-12  Maciej Stachowiak  <mjs@apple.com>
2725
2726         Reviewed by Geoff.
2727
2728         - add handling of hasInstance callback for API objects
2729
2730         * API/JSCallbackObject.cpp:
2731         (KJS::JSCallbackObject::implementsHasInstance): Check if callback is present.
2732         (KJS::JSCallbackObject::hasInstance): Invoke appropriate callback.
2733         * API/JSCallbackObject.h:
2734         * API/JSClassRef.cpp:
2735         * API/JSObjectRef.h:
2736         * API/testapi.c:
2737         (MyObject_hasInstance): Test case; should match what construct would do.
2738         * API/testapi.js:
2739
2740 2006-07-11  Geoffrey Garen  <ggaren@apple.com>
2741
2742         Reviewed by Maciej.
2743
2744         - Implemented a vast number of renames and comment clarifications
2745         suggested during API review.
2746
2747         JSInternalString -> JSString
2748         JS*Make -> JSValueMake*, JSObjectMake*
2749         JSTypeCode -> JSType
2750         JSValueIsInstanceOf -> JSValueIsInstanceOfConstructor (reads strangely well in client code)
2751         JSGC*Protect -> JSValue*Protect
2752         JS*Callback -> JSObject*Callback
2753         JSGetPropertyListCallback -> JSObjectAddPropertiesToListCallback
2754         JSPropertyEnumeratorGetNext -> JSPropertyEnumeratorGetNextName
2755         JSString* ->
2756             JSStringCreateWithUTF8CString, JSStringGetUTF8CString,
2757             JSStringGetMaximumUTF8CStringSize JSStringIsEqualToUTF8CString,
2758             JSStringCreateWithCFString, JSStringCopyCFString, JSStringCreateWithCharacters.
2759
2760         - Changed functions taking a JSValue out arg and returning a bool indicating
2761         whether it was set to simply return a JSValue or NULL.
2762
2763         - Removed JSStringGetCharacters because it's more documentation than code,
2764         and it's just a glorified memcpy built on existing API functionality.
2765
2766         - Moved standard library includes into the headers that actually require them.
2767
2768         - Standardized use of the phrase "Create Rule."
2769
2770         - Removed JSLock from make functions that don't allocate.
2771
2772         - Added exception handling to JSValueToBoolean, since we now allow
2773         callback objects to throw exceptions upon converting to boolean.
2774
2775         - Renamed JSGCCollect to JSGarbageCollect.
2776
2777 2006-07-10  Geoffrey Garen  <ggaren@apple.com>
2778
2779         Reviewed by Darin.
2780
2781         - Changed public header includes to the <JavaScriptCore/ style.
2782         - Changed instances of 'buffer' to 'string' since we decided on
2783         JSInternalString instead of JSStringBuffer.
2784
2785         * API/JSContextRef.h:
2786         * API/JSInternalStringRef.cpp:
2787         (JSStringMake):
2788         (JSInternalStringRetain):
2789         (JSInternalStringRelease):
2790         (JSValueCopyStringValue):
2791         (JSInternalStringGetLength):
2792         (JSInternalStringGetCharactersPtr):
2793         (JSInternalStringGetCharacters):
2794         (JSInternalStringGetMaxLengthUTF8):
2795         (JSInternalStringGetCharactersUTF8):
2796         (CFStringCreateWithJSInternalString):
2797         * API/JSInternalStringRef.h:
2798         * API/JSNode.c:
2799         (JSNodePrototype_appendChild):
2800         (JSNode_getNodeType):
2801         * API/JSObjectRef.cpp:
2802         (JSObjectCallAsConstructor):
2803         * API/JSValueRef.h:
2804         * API/JavaScriptCore.h:
2805         * API/minidom.c:
2806         (main):
2807         (print):
2808         * API/testapi.c:
2809         (MyObject_getPropertyList):
2810         (myConstructor_callAsConstructor):
2811         (main): I noticed that we were prematurely releasing some string buffers,
2812         so I moved their release calls to the end of main(). I got rid of 'Buf' in *Buf
2813         (sometimes changing to 'IString', when necessary to differentiate a variable)
2814         to match the buffer->string change.
2815
2816 === Safari-521.16 ===
2817
2818 2006-07-10  Darin Adler  <darin@apple.com>
2819
2820         * kjs/value.cpp: (KJS::JSValue::toInt32Inline): Added inline keyword one more place.
2821         Just in case.
2822
2823 2006-07-10  Darin Adler  <darin@apple.com>
2824
2825         - fix the release build
2826
2827         * kjs/value.h:
2828         * kjs/value.cpp:
2829         (KJS::JSValue::toInt32Inline): Move the code here to an inline.
2830         (KJS::JSValue::toInt32): Call the inline from both overloaded toInt32 functions.
2831
2832 2006-07-10  David Kilzer  <ddkilzer@kilzer.net>
2833
2834         Reviewed by Darin.
2835
2836         - fix http://bugs.webkit.org/show_bug.cgi?id=9179
2837           Implement select.options.add() method
2838
2839         * JavaScriptCore.exp: Added overloaded KJS::JSValue::toInt32() method.
2840         * JavaScriptCore.xcodeproj/project.pbxproj: Altered attributes metadata for
2841         kjs/value.h to make it available as a forwarded header.
2842         * kjs/lookup.h:
2843         (KJS::lookupPut): Extracted a lookupPut() method from the existing lookupPut() method.
2844         The new method returns a boolean value if no entry is found in the lookup table.
2845         * kjs/value.cpp:
2846         (KJS::JSValue::toInt32): Overloaded toInt32() method with boolean "Ok" argument.
2847         * kjs/value.h: Ditto.
2848
2849 2006-07-10  Geoffrey Garen  <ggaren@apple.com>
2850
2851         No review necessary. Removed bogus file I accidentally checked in before.
2852
2853         * API/JSInternalSringRef.h: Removed.
2854
2855 2006-07-10  Geoffrey Garen  <ggaren@apple.com>
2856
2857         Reviewed by Darin.
2858
2859         Added exception out parameter to API object callbacks, removed semi-bogus
2860         JSContext(.*)Exception functions.
2861
2862         To make these calls syntactically simple, I added an exceptionSlot()
2863         method to the ExecState class, which provides a JSValue** slot in which to
2864         store a JSValue* exception.
2865
2866         * API/APICast.h:
2867         (toRef):
2868         * API/JSCallbackConstructor.cpp:
2869         (KJS::JSCallbackConstructor::construct):
2870         * API/JSCallbackFunction.cpp:
2871         (KJS::JSCallbackFunction::callAsFunction):
2872         * API/JSCallbackObject.cpp:
2873         (KJS::JSCallbackObject::init):
2874         (KJS::JSCallbackObject::getOwnPropertySlot):
2875         (KJS::JSCallbackObject::put):
2876         (KJS::JSCallbackObject::deleteProperty):
2877         (KJS::JSCallbackObject::construct):
2878         (KJS::JSCallbackObject::callAsFunction):
2879         (KJS::JSCallbackObject::getPropertyList):
2880         (KJS::JSCallbackObject::toBoolean):
2881         (KJS::JSCallbackObject::toNumber):
2882         (KJS::JSCallbackObject::toString):
2883         (KJS::JSCallbackObject::staticValueGetter):
2884         (KJS::JSCallbackObject::callbackGetter):
2885         * API/JSContextRef.cpp:
2886         (JSCheckSyntax):
2887         * API/JSContextRef.h:
2888         * API/JSNode.c:
2889         (JSNodePrototype_appendChild):
2890         (JSNodePrototype_removeChild):
2891         (JSNodePrototype_replaceChild):
2892         (JSNode_getNodeType):
2893         (JSNode_getChildNodes):
2894         (JSNode_getFirstChild):
2895         (JSNode_construct):
2896         * API/JSNode.h:
2897         * API/JSNodeList.c:
2898         (JSNodeListPrototype_item):
2899         (JSNodeList_length):
2900         (JSNodeList_getProperty):
2901         * API/JSObjectRef.h:
2902         * API/minidom.c:
2903         (print):
2904         * API/testapi.c:
2905         (MyObject_initialize):
2906         (MyObject_hasProperty):
2907         (MyObject_getProperty):
2908         (MyObject_setProperty):
2909         (MyObject_deleteProperty):
2910         (MyObject_getPropertyList):
2911         (MyObject_callAsFunction):
2912         (MyObject_callAsConstructor):
2913         (MyObject_convertToType):
2914         (print_callAsFunction):
2915         (myConstructor_callAsConstructor):
2916         (main):
2917         * JavaScriptCore.exp:
2918         * kjs/ExecState.h:
2919         (KJS::ExecState::exceptionHandle):
2920
2921 2006-07-10  Geoffrey Garen  <ggaren@apple.com>
2922
2923         Reviewed by Darin.
2924
2925         Improved type safety by implementing opaque JSValue/JSObject typing through
2926         abuse of 'const', not void*. Also fixed an alarming number of bugs
2927         exposed by this new type safety.
2928
2929         I made one design change in JavaScriptCore, which is that the JSObject
2930         constructor should take a JSValue* as its prototype argument, not a JSObject*,
2931         since we allow the prototype to be any JSValue*, including jsNull(), for
2932         example.
2933
2934         * API/APICast.h:
2935         (toJS):
2936         * API/JSBase.h:
2937         * API/JSCallbackConstructor.cpp:
2938         (KJS::JSCallbackConstructor::construct):
2939         * API/JSCallbackFunction.cpp:
2940         (KJS::JSCallbackFunction::callAsFunction):
2941         * API/JSCallbackObject.cpp:
2942         (KJS::JSCallbackObject::JSCallbackObject):
2943         (KJS::JSCallbackObject::getOwnPropertySlot):
2944         (KJS::JSCallbackObject::put):
2945         (KJS::JSCallbackObject::construct):
2946         (KJS::JSCallbackObject::callAsFunction):
2947         (KJS::JSCallbackObject::staticFunctionGetter):
2948         * API/JSCallbackObject.h:
2949         * API/JSContextRef.cpp:
2950         (JSEvaluate):
2951         * API/JSNode.c:
2952         (JSNodePrototype_appendChild):
2953         (JSNodePrototype_removeChild):
2954         (JSNodePrototype_replaceChild):
2955         * API/JSObjectRef.cpp:
2956         (JSObjectMake):
2957         (JSFunctionMakeWithBody):
2958         (JSObjectGetProperty):
2959         (JSObjectCallAsFunction):
2960         (JSObjectCallAsConstructor):
2961         * API/JSObjectRef.h:
2962         * API/testapi.c:
2963         (main):
2964         * ChangeLog:
2965         * kjs/object.h:
2966         (KJS::JSObject::JSObject):
2967
2968 2006-07-10  Geoffrey Garen  <ggaren@apple.com>
2969
2970         Approved by Maciej, Darin.
2971
2972         Renamed JSStringBufferRef to JSInternalStringRef. "Internal string" means the
2973         JavaScript engine's internal string representation, which is the most
2974         low-level and efficient representation to use when interfacing with JavaScript.
2975
2976         * API/APICast.h:
2977         (toJS):
2978         (toRef):
2979         * API/JSBase.h:
2980         * API/JSCallbackObject.cpp:
2981         (KJS::JSCallbackObject::getOwnPropertySlot):
2982         (KJS::JSCallbackObject::put):
2983         (KJS::JSCallbackObject::deleteProperty):
2984         (KJS::JSCallbackObject::staticValueGetter):
2985         (KJS::JSCallbackObject::callbackGetter):
2986         * API/JSContextRef.cpp:
2987         (JSEvaluate):
2988         (JSCheckSyntax):
2989         * API/JSContextRef.h:
2990         * API/JSInternalStringRef.cpp: Added.
2991         (JSStringMake):
2992         (JSInternalStringCreate):
2993         (JSInternalStringCreateUTF8):
2994         (JSInternalStringRetain):
2995         (JSInternalStringRelease):
2996         (JSValueCopyStringValue):
2997         (JSInternalStringGetLength):
2998         (JSInternalStringGetCharactersPtr):
2999         (JSInternalStringGetCharacters):
3000         (JSInternalStringGetMaxLengthUTF8):
3001         (JSInternalStringGetCharactersUTF8):
3002         (JSInternalStringIsEqual):
3003         (JSInternalStringIsEqualUTF8):
3004         (JSInternalStringCreateCF):
3005         (CFStringCreateWithJSInternalString):
3006         * API/JSInternalStringRef.h: Added.
3007         * API/JSNode.c:
3008         (JSNodePrototype_appendChild):
3009         (JSNode_getNodeType):
3010         (JSNode_getChildNodes):
3011         (JSNode_getFirstChild):
3012         * API/JSNodeList.c:
3013         (JSNodeList_length):
3014         (JSNodeList_getProperty):
3015         * API/JSObjectRef.cpp:
3016         (JSFunctionMakeWithBody):
3017         (JSObjectGetDescription):
3018         (JSObjectHasProperty):
3019         (JSObjectGetProperty):
3020         (JSObjectSetProperty):
3021         (JSObjectDeleteProperty):
3022         (JSPropertyEnumeratorGetNext):
3023         (JSPropertyListAdd):
3024         * API/JSObjectRef.h:
3025         * API/JSStringBufferRef.cpp: Removed.
3026         * API/JSStringBufferRef.h: Removed.
3027         * API/JSValueRef.h:
3028         * API/JavaScriptCore.h:
3029         * API/minidom.c:
3030         (main):
3031         (print):
3032         * API/testapi.c:
3033         (assertEqualsAsUTF8String):
3034         (assertEqualsAsCharactersPtr):
3035         (assertEqualsAsCharacters):
3036         (MyObject_hasProperty):
3037         (MyObject_getProperty):
3038         (MyObject_setProperty):
3039         (MyObject_deleteProperty):
3040         (MyObject_getPropertyList):
3041         (print_callAsFunction):
3042         (myConstructor_callAsConstructor):
3043         (main):
3044         * JavaScriptCore.exp:
3045         * JavaScriptCore.xcodeproj/project.pbxproj:
3046
3047 2006-07-08  Tim Omernick  <timo@apple.com>
3048
3049         Reviewed by Maciej.
3050
3051         Added an OpenGL drawing model to the Netscape Plug-in API.
3052
3053         * bindings/npapi.h:
3054
3055 2006-07-08  Timothy Hatcher  <timothy@apple.com>
3056
3057         Reviewed by Maciej.
3058
3059         Moved KJS_GetCreatedJavaVMs to jni_utility.cpp.
3060         Switched KJS_GetCreatedJavaVMs over to use dlopen and dlsym
3061         now that NSAddImage, NSLookupSymbolInImage and NSAddressOfSymbol
3062         are deprecated in Leopard.
3063
3064         * JavaScriptCore.exp:
3065         * JavaScriptCore.xcodeproj/project.pbxproj:
3066         * bindings/jni/jni_utility.cpp:
3067         (KJS::Bindings::KJS_GetCreatedJavaVMs):
3068         * bindings/softlinking.c: Removed.
3069         * bindings/softlinking.h: Removed.
3070
3071 2006-07-08  Geoffrey Garen  <ggaren@apple.com>
3072
3073         Reviewed by Anders.
3074
3075         - Make JSObjectGetProperty return a JSValue or NULL, like JSEvaluate does.
3076
3077         * API/JSObjectRef.cpp:
3078         (JSObjectGetProperty):
3079         * API/JSObjectRef.h:
3080         * API/testapi.c:
3081         (main):
3082
3083 2006-07-08  Geoffrey Garen  <ggaren@apple.com>
3084
3085         Style change -- no review necessary.
3086
3087         Use 0 instead of NULL in API .cpp files, to match our style guidelines.
3088
3089         * API/JSContextRef.cpp:
3090         (JSEvaluate):
3091         * API/JSObjectRef.cpp:
3092         (JSFunctionMakeWithBody):
3093         (JSObjectCallAsFunction):
3094         (JSObjectCallAsConstructor):
3095         * API/JSValueRef.cpp:
3096         (JSValueToObject):
3097
3098 2006-07-08  Geoffrey Garen  <ggaren@apple.com>
3099
3100         Reviewed by TimO.
3101
3102         - Added ability to pass NULL for thisObject when calling JSObjectCallAsFunction,
3103         to match JSEvaluate.
3104
3105         * API/JSObjectRef.cpp:
3106         (JSObjectCallAsFunction):
3107         * API/JSObjectRef.h:
3108         * API/testapi.c:
3109         (main):
3110
3111 === Safari-521.15 ===
3112
3113 2006-07-07  Geoffrey Garen  <ggaren@apple.com>
3114
3115         Reviewed by Maciej.
3116
3117         - Standardized which functions take a JSContext as an argument. The rule is:
3118         if you might execute JavaScript, you take a JSContext, otherwise you don't.
3119
3120         The FIXME in JSObjectRef.h requires refactoring some parts of Interpreter,
3121         but not API changes, so I'm putting it off until later.
3122
3123         * API/JSCallbackObject.cpp:
3124         (KJS::JSCallbackObject::JSCallbackObject):
3125         (KJS::JSCallbackObject::init):
3126         * API/JSCallbackObject.h:
3127         * API/JSContextRef.cpp:
3128         (JSContextCreate):
3129         * API/JSContextRef.h:
3130         * API/JSObjectRef.cpp:
3131         (JSObjectMake):
3132         (JSPropertyEnumeratorGetNext):
3133         * API/JSObjectRef.h:
3134         * API/testapi.c:
3135         (MyObject_initialize):
3136         (main):
3137         * JavaScriptCore.exp:
3138         * kjs/array_object.cpp:
3139         (ArrayInstance::setLength):
3140         (ArrayInstance::pushUndefinedObjectsToEnd):
3141         * kjs/nodes.cpp:
3142         (ForInNode::execute):
3143         * kjs/reference.cpp:
3144         (KJS::Reference::getPropertyName):
3145         (KJS::Reference::getValue):
3146         * kjs/reference.h:
3147         * kjs/scope_chain.cpp:
3148         (KJS::ScopeChain::print):
3149
3150 2006-07-06  Geoffrey Garen  <ggaren@apple.com>
3151
3152         Reviewed by Maciej.
3153
3154         More API action.
3155
3156         - Headerdoc finished
3157
3158         Semantic Changes:
3159         - Added a JSContextRef argument to many functions, because you need a
3160         JSContextRef for doing virtually anything. I expect to add this argument
3161         to even more functions in a future patch.
3162
3163         - Removed the globalObjectPrototype argument to JSContextCreate because
3164         you can't create an object until you have a context, so it's impossible
3165         to pass a prototype object to JSContextCreate. That's OK because (1) there's
3166         no reason to give the global object a prototype and (2) if you really want
3167         to, you can just use a separate call to JSObjectSetPrototype.
3168
3169         - Removed the JSClassRef argument to JSClassCreate because it was unnecessary,
3170         and you need to be able to make the global object's class before you've
3171         created a JSContext.
3172
3173         - Added an optional exception parameter to JSFunctionMakeWithBody because anything
3174         less would be uncivilized.
3175
3176         - Made the return value parameter to JSObjectGetProperty optional to match
3177         all other return value parameters in the API.
3178
3179         - Made JSObjectSetPrivate/JSObjectGetPrivate work on JSCallbackFunctions
3180         and JSCallbackConstructors. You could use an abstract base class or strategic
3181         placement of m_privateData in the class structure to implement this, but
3182         the former seemed like overkill, and the latter seemed too dangerous.
3183
3184         - Fixed a bug where JSPropertyEnumeratorGetNext would skip the first property.
3185
3186         Cosmetic Changes:
3187         - Reversed the logic of the JSChar #ifdef to avoid confusing headerdoc
3188
3189         - Removed function names from @function declarations because headeroc
3190         can parse them automatically, and I wanted to rule out manual mismatch.
3191
3192         - Changed Error::create to take a const UString& instead of a UString*
3193         because it was looking at me funny.
3194
3195         - Renamed JSStringBufferCreateWithCFString to JSStringBufferCreateCF
3196         because the latter is more concise and it matches JSStringBufferCreateUTF8.
3197
3198         * API/JSCallbackObject.cpp:
3199         (KJS::JSCallbackObject::getOwnPropertySlot):
3200         (KJS::JSCallbackObject::put):
3201         (KJS::JSCallbackObject::deleteProperty):
3202         (KJS::JSCallbackObject::getPropertyList):
3203         (KJS::JSCallbackObject::toBoolean):
3204         (KJS::JSCallbackObject::toNumber):
3205         (KJS::JSCallbackObject::toString):
3206         * API/JSClassRef.cpp:
3207         (JSClassCreate):
3208         * API/JSContextRef.cpp:
3209         (JSContextCreate):
3210         (JSContextSetException):
3211         * API/JSContextRef.h:
3212         * API/JSNode.c:
3213         (JSNodePrototype_class):
3214         (JSNode_class):
3215         * API/JSNodeList.c:
3216         (JSNodeListPrototype_class):
3217         (JSNodeList_class):
3218         * API/JSObjectRef.cpp:
3219         (JSObjectGetProperty):
3220         (JSObjectGetPrivate):
3221         (JSObjectSetPrivate):
3222         (JSObjectCallAsFunction):
3223         (JSObjectCallAsConstructor):
3224         (JSPropertyEnumeratorGetNext):
3225         * API/JSObjectRef.h:
3226         * API/JSStringBufferRef.cpp:
3227         (JSStringBufferCreateCF):
3228         * API/JSStringBufferRef.h:
3229         * API/JSValueRef.cpp:
3230         (JSValueIsInstanceOf):
3231         * API/JSValueRef.h:
3232         * API/minidom.c:
3233         (main):
3234         * API/minidom.js:
3235         * API/testapi.c:
3236         (MyObject_hasProperty):
3237         (MyObject_setProperty):
3238         (MyObject_deleteProperty):
3239         (MyObject_getPropertyList):
3240         (MyObject_convertToType):
3241         (MyObject_class):
3242         (main):
3243         * JavaScriptCore.exp:
3244
3245 2006-07-07  Geoffrey Garen  <ggaren@apple.com>
3246
3247         Reviewed by John.
3248
3249         - Fixed a few crashes resulting from NULL parameters to JSClassCreate.
3250
3251         * API/JSClassRef.cpp:
3252         (JSClassCreate):
3253         (JSClassRelease):
3254         * API/testapi.c: Added test for NULL parameters.
3255         (main):
3256
3257 2006-07-07  Geoffrey Garen  <ggaren@apple.com>
3258
3259         Reviewed by John, mocked by Darin.
3260
3261         - Changed JSEvaluate to take a JSObjectRef instead of a JSValueRef as
3262         "this," since "this" must be an object.
3263
3264         * API/JSContextRef.cpp:
3265         (JSEvaluate):
3266         * API/JSContextRef.h:
3267
3268 2006-07-07  Geoffrey Garen  <ggaren@apple.com>
3269
3270         Reviewed by John.
3271
3272         - More headerdoc
3273
3274         * API/JSBase.h:
3275         * JavaScriptCore.xcodeproj/project.pbxproj:
3276
3277 2006-07-05  Geoffrey Garen  <ggaren@apple.com>
3278
3279         RS by Beth.
3280
3281         Renamed JSCharBufferRef, which was universally unpopular, to JSStringBufferRef,
3282         which, hopefully, will be less unpopular.
3283
3284         * API/APICast.h:
3285         (toJS):
3286         (toRef):
3287         * API/JSBase.h:
3288         * API/JSCallbackObject.cpp:
3289         (KJS::JSCallbackObject::getOwnPropertySlot):
3290         (KJS::JSCallbackObject::put):
3291         (KJS::JSCallbackObject::deleteProperty):
3292         (KJS::JSCallbackObject::staticValueGetter):
3293         (KJS::JSCallbackObject::callbackGetter):
3294         * API/JSCharBufferRef.cpp: Removed.
3295         * API/JSCharBufferRef.h: Removed.
3296         * API/JSContextRef.cpp:
3297         (JSEvaluate):
3298         (JSCheckSyntax):
3299         * API/JSContextRef.h:
3300         * API/JSNode.c:
3301         (JSNodePrototype_appendChild):
3302         (JSNode_getNodeType):
3303         (JSNode_getChildNodes):
3304         (JSNode_getFirstChild):
3305         * API/JSNodeList.c:
3306         (JSNodeList_length):
3307         (JSNodeList_getProperty):
3308         * API/JSObjectRef.cpp:
3309         (JSFunctionMakeWithBody):
3310         (JSObjectGetDescription):
3311         (JSObjectHasProperty):
3312         (JSObjectGetProperty):
3313         (JSObjectSetProperty):
3314         (JSObjectDeleteProperty):
3315         (JSPropertyEnumeratorGetNext):
3316         (JSPropertyListAdd):
3317         * API/JSObjectRef.h:
3318         * API/JSStringBufferRef.cpp: Added.
3319         (JSStringMake):
3320         (JSStringBufferCreate):
3321         (JSStringBufferCreateUTF8):
3322         (JSStringBufferRetain):
3323         (JSStringBufferRelease):
3324         (JSValueCopyStringValue):
3325         (JSStringBufferGetLength):
3326         (JSStringBufferGetCharactersPtr):
3327         (JSStringBufferGetCharacters):
3328         (JSStringBufferGetMaxLengthUTF8):
3329         (JSStringBufferGetCharactersUTF8):
3330         (JSStringBufferIsEqual):
3331         (JSStringBufferIsEqualUTF8):
3332         (JSStringBufferCreateWithCFString):
3333         (CFStringCreateWithJSStringBuffer):
3334         * API/JSStringBufferRef.h: Added.
3335         * API/JSValueRef.h:
3336         * API/JavaScriptCore.h:
3337         * API/minidom.c:
3338         (main):
3339         (print):
3340         * API/testapi.c:
3341         (assertEqualsAsUTF8String):
3342         (assertEqualsAsCharactersPtr):
3343         (assertEqualsAsCharacters):
3344         (MyObject_hasProperty):
3345         (MyObject_getProperty):
3346         (MyObject_setProperty):
3347         (MyObject_deleteProperty):
3348         (MyObject_getPropertyList):
3349         (print_callAsFunction):
3350         (myConstructor_callAsConstructor):
3351         (main):
3352         * JavaScriptCore.exp:
3353         * JavaScriptCore.xcodeproj/project.pbxproj:
3354
3355 2006-07-05  Geoffrey Garen  <ggaren@apple.com>
3356
3357         RS by Beth.
3358
3359         Moved some code around for more logical file separation.
3360
3361         * API/JSBase.h:
3362         * API/JSContextRef.h:
3363         * API/JSObjectRef.cpp:
3364         * API/JSValueRef.cpp:
3365         (JSValueToObject):
3366         * API/JSValueRef.h:
3367
3368 2006-07-03  Geoffrey Garen  <ggaren@apple.com>
3369
3370         Reviewed by Maciej.
3371
3372         Implemented JSFunctionMakeWithBody, which parses a script as a function body
3373         in the global scope, and returns the resulting anonymous function.
3374
3375         I also removed private data from JSCallbackFunction. It never worked,
3376         since JSCallbackFunction doesn't inherit from JSCallbackObject.
3377
3378         * API/JSCallbackConstructor.cpp: Removed.
3379         * API/JSCallbackConstructor.h: Removed.
3380         * API/JSCallbackFunction.cpp:
3381         (KJS::JSCallbackFunction::JSCallbackFunction):
3382         (KJS::JSCallbackFunction::implementsConstruct):
3383         (KJS::JSCallbackFunction::construct):
3384         (KJS::JSCallbackFunction::implementsCall):
3385         (KJS::JSCallbackFunction::callAsFunction):
3386         * API/JSCallbackFunction.h:
3387         * API/JSCallbackObject.cpp:
3388         (KJS::JSCallbackObject::staticFunctionGetter):
3389         * API/JSObjectRef.cpp:
3390         (JSFunctionMake):
3391         (JSFunctionMakeWithCallbacks):
3392         * API/JSObjectRef.h:
3393         * API/JSValueRef.h:
3394         * API/minidom.c:
3395         (main):
3396         * API/testapi.c:
3397         (main):
3398         * JavaScriptCore.exp: Programmatically added all symbols exported by
3399         API object files, and sorted results
3400         * JavaScriptCore.xcodeproj/project.pbxproj:
3401
3402 2006-07-03  Geoffrey Garen  <ggaren@apple.com>
3403
3404         Reviewed by Maciej.
3405
3406         - Return syntax error in JSCheckSyntax through a JSValueRef* exception
3407         argument
3408
3409         * API/JSBase.h:
3410         * API/JSContextRef.cpp:
3411         (JSCheckSyntax):
3412         * API/testapi.c:
3413         (main):
3414         * JavaScriptCore.exp:
3415         * kjs/interpreter.cpp:
3416         (KJS::Interpreter::checkSyntax):
3417         * kjs/interpreter.h:
3418
3419 2006-07-04  Darin Adler  <darin@apple.com>
3420
3421         - fixed build
3422
3423         * wtf/MathExtras.h: Oops. Added missing #endif.
3424
3425 2006-07-04  Bjoern Graf  <bjoern.graf@gmail.com>
3426
3427         Reviewed by Maciej.
3428         Tweaked a bit by Darin.
3429
3430         - http://bugs.webkit.org/show_bug.cgi?id=9678
3431           work around MSVCRT's fmod function returning NaN for fmod(x, infinity) instead of x
3432
3433         * wtf/MathExtras.h: Added include of <float.h>.
3434         (isinf): Fix to return false for NAN.
3435         (wtf_fmod): Added. An inline that works around the bug.
3436
3437         * kjs/nodes.cpp:
3438         * kjs/number_object.cpp:
3439         * kjs/operations.cpp:
3440         * kjs/value.cpp:
3441         Added includes of MathExtras.h to all files using fmod.
3442
3443         * JavaScriptCore.xcodeproj/project.pbxproj: Let Xcode 2.3 have its way with
3444         the project.
3445
3446 2006-07-01  Geoffrey Garen  <ggaren@apple.com>
3447
3448         Reviewed by Darin.
3449
3450         - Refined value conversions in the API:
3451             - failed toNumber returns NaN
3452             - failed toObject returns NULL
3453             - failed toString returns empty string
3454
3455         - Refined excpetion handling in the API:
3456             - failed value conversions do not throw exceptions
3457             - uncaught exceptions in JSEvaluate, JSObjectCallAsFunction, and
3458               JSObjectCallAsConstructor are returned through a JSValueRef* exception
3459               argument
3460             - removed JSContextHasException, because JSContextGetException does
3461               the same job
3462
3463         * API/JSBase.h:
3464         * API/JSCharBufferRef.cpp:
3465         (JSValueCopyStringValue):
3466         * API/JSContextRef.cpp:
3467         (JSEvaluate):
3468         * API/JSContextRef.h:
3469         * API/JSNodeList.c: Added test code demonstrating how you would use
3470         toNumber, and why you probably don't need toUInt32, etc.
3471         (JSNodeListPrototype_item):
3472         (JSNodeList_getProperty):
3473         * API/JSObjectRef.cpp:
3474         (JSValueToObject):
3475         (JSObjectCallAsFunction):
3476         (JSObjectCallAsConstructor):
3477         * API/JSObjectRef.h:
3478         * API/JSValueRef.cpp:
3479         (JSValueToNumber):
3480         * API/JSValueRef.h:
3481         * API/minidom.c:
3482         (main):
3483         * API/testapi.c:
3484         (main): Added tests for new rules, and call to JSGCProtect to fix Intel
3485         crash
3486         * JavaScriptCore.exp:
3487
3488 2006-07-03  Darin Adler  <darin@apple.com>
3489
3490         - Rolled out HashMap implementation of NPRuntime, at least temporarily.
3491
3492         Fixes hang in the bindings section of layout tests seen on the
3493         buildbot.
3494
3495         This code was using HashMap<const char*, PrivateIdentifier*>.
3496         But that hashes based on pointer identity, not string value.
3497         The default hash for any pointer type is to hash based on the pointer.
3498         And WTF doesn't currently have a string hash for char*.
3499         We'll need to fix that before re-landing this patch.
3500
3501         (Formatting was also incorrect -- extra spaces in parentheses.)
3502
3503         * bindings/npruntime.cpp: Rolled out last change.
3504
3505 2006-07-02  Justin Haygood  <jhaygood@spsu.edu>
3506
3507         Reviewed, tweaked, landed by ggaren.
3508
3509         - Port NPRuntime from CFDictionary to HashMap.
3510
3511         * bindings/npruntime.cpp:
3512         (getStringIdentifierDictionary):
3513         (getIntIdentifierDictionary):
3514         (_NPN_GetStringIdentifier):
3515         (_NPN_GetIntIdentifier):
3516         * bindings/npruntime.h:
3517
3518 2006-07-01  Geoffrey Garen  <ggaren@apple.com>
3519
3520         Reviewed by Adele.
3521
3522         - Fixed <rdar://problem/4611197> REGRESSION: Liveconnect with Java test
3523         fails at http://www-sor.inria.fr/~dedieu/notes/liveconnect/simple_example.html
3524
3525         * JavaScriptCore.exp: Export symbols used by liveconnect
3526
3527 2006-06-29  Geoffrey Garen  <ggaren@apple.com>
3528
3529         Reviewed by Maciej.
3530
3531         - Phase 2 in the JS API.
3532
3533         - Added support for specifying static tables of values -- this should
3534         obviate the need for using complicated callbacks for most lookups.
3535
3536         -  API objects are now created with classes (JSClassRef) -- in order to support
3537         static values, and in order to prevent API objects from storing their
3538         data inline, and thus falling into the oversized (read: slow and prone to
3539         giving Maciej the frowny face) heap.
3540
3541         - Added two specialized JSObject subclasses -- JSCallbackFunction and JSCallbackConstructor --
3542         to allow JSFunctionMake and JSConstructorMake to continue to work with
3543         the new class model. Another solution to this problem would be to create
3544         a custom class object for each function and constructor you make. This
3545         solution is more code but also more efficient.
3546
3547         - Substantially beefed up the minidom example to demonstrate and test a
3548         lot of these techniques. Its output is still pretty haphazard, though.
3549
3550         - Gave the <kjs/ preface to some includes -- I'm told this matters to
3551         building on some versions of Linux.
3552
3553         - Implemented JSValueIsInstanceOf and JSValueIsObjectOfClass
3554
3555         - Removed GetDescription callback. Something in the class datastructure
3556         should take care of this.
3557
3558         * API/JSBase.h:
3559         * API/JSCallbackConstructor.cpp: Added.
3560         (KJS::):
3561         (KJS::JSCallbackConstructor::JSCallbackConstructor):
3562         (KJS::JSCallbackConstructor::implementsConstruct):
3563         (KJS::JSCallbackConstructor::construct):
3564         (KJS::JSCallbackConstructor::setPrivate):
3565         (KJS::JSCallbackConstructor::getPrivate):
3566         * API/JSCallbackConstructor.h: Added.
3567         (KJS::JSCallbackConstructor::classInfo):
3568         * API/JSCallbackFunction.cpp: Added.
3569         (KJS::):
3570         (KJS::JSCallbackFunction::JSCallbackFunction):
3571         (KJS::JSCallbackFunction::implementsCall):
3572         (KJS::JSCallbackFunction::callAsFunction):
3573         (KJS::JSCallbackFunction::setPrivate):
3574         (KJS::JSCallbackFunction::getPrivate):
3575         * API/JSCallbackFunction.h: Added.
3576         (KJS::JSCallbackFunction::classInfo):
3577         * API/JSCallbackObject.cpp:
3578         (KJS::):
3579         (KJS::JSCallbackObject::JSCallbackObject):
3580         (KJS::JSCallbackObject::init):
3581         (KJS::JSCallbackObject::~JSCallbackObject):
3582         (KJS::JSCallbackObject::className):
3583         (KJS::JSCallbackObject::getOwnPropertySlot):
3584         (KJS::JSCallbackObject::put):
3585         (KJS::JSCallbackObject::deleteProperty):
3586         (KJS::JSCallbackObject::implementsConstruct):
3587         (KJS::JSCallbackObject::construct):
3588         (KJS::JSCallbackObject::implementsCall):
3589         (KJS::JSCallbackObject::callAsFunction):
3590         (KJS::JSCallbackObject::getPropertyList):
3591         (KJS::JSCallbackObject::toBoolean):