JavaScriptCore/ChangeLog

   1 2006-09-28  Maciej Stachowiak  <mjs@apple.com>
   2
   3         Reviewed by Mitz.
   4
   5         - fixed REGRESSION (r16606): javascriptCore Crash on website load
   6
   7         Plus style fixes.
   8
   9             - fixed some possible off-by-one bugs
  10             - use indexing, not iterators, for Vectors
  11             - store Vector by pointer instead of by value to avoid blowing out FunctionImp size
  12
  13         * kjs/function.cpp:
  14         (KJS::FunctionImp::addParameter):
  15         (KJS::FunctionImp::parameterString):
  16         (KJS::FunctionImp::processParameters):
  17         (KJS::FunctionImp::lengthGetter):
  18         (KJS::FunctionImp::getParameterName):
  19         * kjs/function.h:
  20
  21 2006-09-27  Steve Falkenburg  <sfalken@apple.com>
  22
  23         Reviewed by Maciej.
  24
  25         More build tweaks
  26
  27         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
  28         * JavaScriptCore.vcproj/JavaScriptCore/dstroot-to-sdk.cmd: Removed.
  29
  30 2006-09-27  John Sullivan  <sullivan@apple.com>
  31
  32         * kjs/function.cpp:
  33         (KJS::FunctionImp::getParameterName):
  34         removed assertion that displeased gcc 4.0.1 (build 5420):
  35         ASSERT(static_cast<size_t>(index) == index);
  36
  37 2006-09-27  Kevin McCullough  <KMcCullough@apple.com>
  38
  39         Reviewed by GGaren.
  40
  41         Cleanup of previous fix which was to address Radar: 4752492
  42
  43         * kjs/function.cpp:
  44         (KJS::FunctionImp::addParameter):
  45         (KJS::FunctionImp::parameterString):
  46         (KJS::FunctionImp::processParameters):
  47         (KJS::FunctionImp::lengthGetter):
  48         (KJS::FunctionImp::getParameterName):
  49         * kjs/function.h:
  50
  51 2006-09-27  Kevin McCullough  <KMcCullough@apple.com>
  52
  53         Reviewed by Adele.
  54
  55         Fixes a GC stack overflow crash.
  56         The change is to move from a linked list implementation of Parameters to a Vector.
  57         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.
  58
  59         * kjs/function.cpp:
  60         (KJS::Parameter::Parameter):
  61         (KJS::FunctionImp::addParameter):
  62         (KJS::FunctionImp::parameterString):
  63         (KJS::FunctionImp::processParameters):
  64         (KJS::FunctionImp::lengthGetter):
  65         (KJS::FunctionImp::getParameterName):
  66         * kjs/function.h:
  67
  68 2006-09-27  Steve Falkenburg  <sfalken@apple.com>
  69
  70         Fix last path fix.
  71
  72         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
  73
  74 2006-09-27  Steve Falkenburg  <sfalken@apple.com>
  75
  76         Set path before build.
  77
  78         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
  79
  80 2006-09-27  Sean Gies  <seangies@apple.com>
  81
  82         Reviewed by Adam Roben.
  83
  84         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: Debug config should link to debug runtime.
  85         * JavaScriptCore.vcproj/dftables/dftables.vcproj: Debug config should link to debug runtime.
  86
  87 2006-09-27  Don Melton  <gramps@apple.com>
  88
  89         Reviewed by Adam Roben.
  90
  91         Changed line ending from DOS to UNIX format so it doesn't die running
  92         on my machine. ;)
  93
  94         * JavaScriptCore.vcproj/JavaScriptCore/build-generated-files.sh:
  95
  96 2006-09-23  Alexey Proskuryakov  <ap@nypop.com>
  97
  98         Reviewed by Maciej.
  99
 100         http://bugzilla.opendarwin.org/show_bug.cgi?id=10183
 101         REGRESSION: obfuscated JS decoding breaks because of soft hyphen removal
 102         (Fanfiction.net author pages not listing stories)
 103
 104         Rolled out the fix for bug 4139.
 105
 106         * kjs/lexer.cpp:
 107         (Lexer::setCode):
 108         (Lexer::shift):
 109         * tests/mozilla/ecma/Array/15.4.5.1-1.js:
 110         * tests/mozilla/expected.html:
 111
 112 2006-09-22  Steve Falkenburg  <sfalken@apple.com>
 113
 114         Build fix.
 115
 116         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
 117
 118 2006-09-22  Darin Adler  <darin@apple.com>
 119
 120         Reviewed by Alice.
 121
 122         * wtf/Vector.h: Add an append that takes a pointer and length.
 123         Generalize the existing Vector append to work on vectors with
 124         any value for inlineCapacity. Change the append algorithm so
 125         it doesn't check capacity each time through the loop.
 126
 127 2006-09-22  Steve Falkenburg  <sfalken@apple.com>
 128
 129         Fix release build.
 130
 131         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
 132
 133 2006-09-21  Geoffrey Garen  <ggaren@apple.com>
 134
 135         Reviewed by Maciej.
 136
 137         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
 138         Updated to include the right path.
 139         * wtf/FastMalloc.h: #include Platform.h, since we use Platform macros.
 140
 141 === Safari-521.27 ===
 142
 143 2006-09-20  Anders Carlsson  <acarlsson@apple.com>
 144
 145         Reviewed by Dave Hyatt.
 146
 147         * wtf/MathExtras.h:
 148         Get rid of lrint.
 149
 150 2006-09-20  Sean Gies  <seangies@apple.com>
 151
 152         Reviewed by Steve Falkenburg.
 153
 154         * wtf/Assertions.cpp: Debug messages should go into debugger console.
 155
 156 2006-09-20  David Hyatt  <hyatt@apple.com>
 157
 158         Add an implementation of lrint for Win32.
 159
 160         Reviewed by anders
 161
 162         * wtf/MathExtras.h:
 163         (lrint):
 164
 165 2006-09-15  Krzysztof Kowalczyk  <kkowalczyk@gmail.com>
 166
 167         Reviewed by Adam.
 168
 169         http://bugzilla.opendarwin.org/show_bug.cgi?id=10864
 170         Bug 10864: Linux\GDK build fixes
 171
 172         * JavaScriptCoreSources.bkl:
 173         * jscore.bkl:
 174
 175 2006-09-15  Adam Roben  <aroben@apple.com>
 176
 177         Windows build fix.
 178
 179         * JavaScriptCore.vcproj/JavaScriptCore/build-generated-files.sh:
 180
 181 2006-09-15  Anders Carlsson  <acarlsson@apple.com>
 182
 183         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
 184         Fix the release build.
 185
 186 2006-09-15  Anders Carlsson  <acarlsson@apple.com>
 187
 188         Reviewed by Steve.
 189
 190         Add JavaScriptCore API to the build.
 191         * API/JSBase.cpp:
 192         * API/JSCallbackConstructor.cpp:
 193         * API/JSCallbackFunction.cpp:
 194         * API/JSCallbackObject.cpp:
 195         * API/JSClassRef.cpp:
 196         * API/JSContextRef.cpp:
 197         * API/JSObjectRef.cpp:
 198         * API/JSStringRef.cpp:
 199         * API/JSValueRef.cpp:
 200         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
 201         * os-win32/stdbool.h: Added.
 202
 203 2006-09-12  Steve Falkenburg  <sfalken@apple.com>
 204
 205         Reviewed by Ada.
 206
 207         Build tweaks (doing JavaScriptCore now since it doesn't have
 208         dependencies).
 209
 210         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
 211         * JavaScriptCore.vcproj/JavaScriptCore/build-generated-files.sh:
 212         * JavaScriptCore.vcproj/JavaScriptCore/dstroot-to-sdk.cmd: Added.
 213         * JavaScriptCore.vcproj/dftables/dftables.vcproj:
 214         * JavaScriptCore.vcproj/testkjs/testkjs.vcproj:
 215
 216 2006-09-11  Brady Eidson  <beidson@apple.com>
 217
 218         Build fix - I think Tim's last checkin wasn't tested on Tiger, possibly.  I simply
 219         commented out the undefined constants until he can have a chance to make the right call
 220
 221         * bindings/objc/objc_utility.mm:
 222         (KJS::Bindings::objcValueTypeForType): Commented out undefined symbols
 223
 224 2006-09-11  Timothy Hatcher  <timothy@apple.com>
 225
 226         Reviewed by Tim O. and Darin.
 227
 228         Add support for more method signatures affecting ObjC methods called from JavaScript:
 229         - Added unsigned types and long long.
 230         - Allow methods that use const, oneway, bycopy and byref type modifiers.
 231
 232         * bindings/objc/objc_instance.mm:
 233         (ObjcInstance::invokeMethod):
 234         * bindings/objc/objc_utility.h:
 235         (KJS::Bindings::):
 236         * bindings/objc/objc_utility.mm:
 237         (KJS::Bindings::convertValueToObjcValue):
 238         (KJS::Bindings::convertObjcValueToValue):
 239         (KJS::Bindings::objcValueTypeForType):
 240
 241 2006-09-05  Timothy Hatcher  <timothy@apple.com>
 242
 243         Reviewed by Tim O.
 244
 245         <rdar://problem/4715840> SEL is not char*
 246
 247         * bindings/objc/objc_class.mm:
 248         (KJS::Bindings::ObjcClass::methodsNamed): use sel_getName instead of a char* cast.
 249         * bindings/objc/objc_runtime.mm:
 250         (ObjcFallbackObjectImp::callAsFunction): ditto
 251
 252 2006-09-03  Alexey Proskuryakov  <ap@nypop.com>
 253
 254         Reviewed by Tim H.
 255
 256         http://bugzilla.opendarwin.org/show_bug.cgi?id=10693
 257         Convert JavaScript arrays to AppleScript lists
 258
 259         * JavaScriptCore.exp: Export ArrayInstance::info and ArrayInstance::getItem().
 260         * kjs/array_instance.h:
 261         * kjs/array_object.cpp:
 262         (ArrayInstance::getItem): Added a method to access array items from C++.
 263
 264 2006-09-02  Krzysztof Kowalczyk  <kkowalczyk@gmail.com>
 265
 266         Reviewed by Tim H.
 267
 268         Bug 10454: Unix bakefile fixes
 269         http://bugzilla.opendarwin.org/show_bug.cgi?id=10454
 270
 271         * JavaScriptCoreSources.bkl:
 272
 273 2006-09-01  Nikolas Zimmermann  <zimmermann@kde.org>
 274
 275         Reviewed by hyatt.  Landed by eseidel.
 276
 277         Fix build on Linux.
 278
 279         * pcre/CMakeLists.txt: Add wtf/ include.
 280
 281 2006-09-01  Nikolas Zimmermann  <zimmermann@kde.org>
 282
 283         Reviewed and landed by ap.
 284
 285         Fix build on Linux (C89 without gcc extensions enabled).
 286
 287         * pcre/pcre_internal.h: Use C style comments.
 288         * wtf/Assertions.h: Use C style comments.
 289         * wtf/Platform.h: Use C style comments.
 290
 291 2006-09-01  Steve Falkenburg  <sfalken@apple.com>
 292
 293         Fix build.
 294
 295         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
 296         * JavaScriptCore.vcproj/dftables/dftables.vcproj:
 297
 298 2006-08-31  Anders Carlsson  <acarlsson@apple.com>
 299
 300         Reviewed by Darin.
 301
 302         Add new portability functions to MathExtras.h and add StringExtras.h which is for
 303         string portability functions.
 304
 305         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
 306         * bindings/c/c_instance.cpp:
 307         * kjs/date_object.cpp:
 308         * wtf/MathExtras.h:
 309         (copysign):
 310         (isfinite):
 311         * wtf/StringExtras.h: Added.
 312         (snprintf):
 313         (strncasecmp):
 314
 315 2006-08-31  Anders Carlsson  <acarlsson@apple.com>
 316
 317         Reviewed by Tim H.
 318
 319         Fix Windows build.
 320
 321         * JavaScriptCore.vcproj/dftables/dftables.vcproj:
 322         * pcre/pcre_internal.h:
 323
 324 2006-08-31  Timothy Hatcher  <timothy@apple.com>
 325
 326         Reviewed by Geoff.
 327
 328         Band-aid fix for PCRE to compile for ppc64 and x86_64 now that
 329         we use -Wshorten-64-to-32. Adds an INT_CAST macro that ASSERTs
 330         the value <= INT_MAX.
 331
 332         I filed <rdar://problem/4712064> to track the need to verify
 333         PCRE's 64-bit compliance.
 334
 335         * pcre/pcre_compile.c:
 336         (complete_callout):
 337         (compile_branch):
 338         (compile_regex):
 339         (pcre_compile2):
 340         * pcre/pcre_exec.c:
 341         (match):
 342         (pcre_exec):
 343         * pcre/pcre_get.c:
 344         (pcre_get_substring_list):
 345         * pcre/pcre_internal.h:
 346         * pcre/pcre_tables.c:
 347         * pcre/pcre_try_flipped.c:
 348         (_pcre_try_flipped):
 349
 350 2006-08-30  Darin Adler  <darin@apple.com>
 351
 352         Reviewed by Tim Hatcher.
 353
 354         - add WTF::getPtr, a function template that makes it possible to write
 355           generic code that gets a raw pointer out of any of our pointer types
 356
 357         * JavaScriptCore.xcodeproj/project.pbxproj:
 358         * wtf/GetPtr.h: Added.
 359         * wtf/ListRefPtr.h: (WTF::getPtr): Added.
 360         * wtf/OwnArrayPtr.h: (WTF::getPtr): Added.
 361         * wtf/OwnPtr.h: (WTF::getPtr): Added.
 362         * wtf/PassRefPtr.h: (WTF::getPtr): Added.
 363         * wtf/RefPtr.h: (WTF::getPtr): Added.
 364
 365 2006-08-29  waylonis  <waylonis@google.com>
 366
 367         Reviewed, tweaked by ggaren.
 368
 369         - Added storage and accessor functions for ExecState as a fix for
 370           http://bugzilla.opendarwin.org/show_bug.cgi?id=10114
 371
 372         * kjs/ExecState.cpp:
 373         (KJS::ExecState::ExecState):
 374         * kjs/ExecState.h:
 375         * kjs/context.h:
 376         (KJS::Context::setExecState):
 377         (KJS::Context::execState):
 378
 379 2006-08-30  Nikolas Zimmermann  <zimmermann@kde.org>
 380
 381         Reviewed by Tim H.
 382
 383         Commit KDE related tweaks, to be able to
 384         differentiate between a Qt-only or a KDE build.
 385
 386         * CMakeLists.txt: Install wtf-unity library.
 387         * wtf/Platform.h: Add define for the KDE platform.
 388
 389 2006-08-28  Darin Adler  <darin@apple.com>
 390
 391         Reviewed by Geoff.
 392
 393         * kjs/list.h: Use explicit in constructor (as appropriate).
 394
 395 2006-08-24  Nikolas Zimmermann  <zimmermann@kde.org>
 396
 397         Reviewed, tweaked and landed by ap
 398
 399         http://bugzilla.opendarwin.org/show_bug.cgi?id=10467
 400         WebKit should have Qt platform support (Part II)
 401
 402         * CMakeLists.txt: Adjust to Anders' build fixes.
 403         * wtf/Platform.h: Fix define for the Qt platform (we don't use/need Cairo.)
 404
 405 2006-08-23  David Hyatt  <hyatt@apple.com>
 406
 407         Fix Platform.h to include #defines for graphics features.
 408
 409         Reviewed by darin
 410
 411         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
 412         * wtf/Platform.h:
 413
 414 2006-08-23  Anders Carlsson  <acarlsson@apple.com>
 415
 416         Reviewed by Darin.
 417
 418         Make the bindings compile without CoreFoundation.
 419
 420         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
 421         * bindings/c/c_instance.cpp:
 422         * bindings/c/c_utility.cpp:
 423         (KJS::Bindings::convertUTF8ToUTF16):
 424         * bindings/npapi.h:
 425         * bindings/runtime.cpp:
 426         (KJS::Bindings::Instance::createBindingForLanguageInstance):
 427         (KJS::Bindings::Instance::createLanguageInstanceForValue):
 428         * bindings/runtime_root.cpp:
 429         * bindings/runtime_root.h:
 430         * kjs/interpreter.cpp:
 431         (KJS::Interpreter::createLanguageInstanceForValue):
 432         * kjs/interpreter.h:
 433
 434 2006-08-22  Anders Carlsson  <acarlsson@apple.com>
 435
 436         Reviewed by Darin.
 437
 438         Move the npruntime code over to using HashMap and the runtime_root code over to using
 439         HashMap and HashCountedSet.
 440
 441         * bindings/NP_jsobject.cpp:
 442         * bindings/c/c_utility.cpp:
 443         (KJS::Bindings::identifierFromNPIdentifier):
 444         * bindings/c/c_utility.h:
 445         * bindings/jni/jni_jsobject.cpp:
 446         (JavaJSObject::invoke):
 447         * bindings/npruntime.cpp:
 448         (getStringIdentifierMap):
 449         (getIntIdentifierMap):
 450         (_NPN_GetStringIdentifier):
 451         (_NPN_GetIntIdentifier):
 452         * bindings/runtime_root.cpp:
 453         (getReferencesByRootMap):
 454         (getReferencesSet):
 455         (KJS::Bindings::findReferenceSet):
 456         (KJS::Bindings::rootForImp):
 457         (KJS::Bindings::rootForInterpreter):
 458         (KJS::Bindings::addNativeReference):
 459         (KJS::Bindings::removeNativeReference):
 460         (RootObject::removeAllNativeReferences):
 461         * bindings/runtime_root.h:
 462
 463 2006-08-22  Anders Carlsson  <acarlsson@apple.com>
 464
 465         Reviewed by Geoff.
 466
 467         Switch over the NPAPI and Java bindings to using HashMaps instead of dictionaries.
 468
 469         * JavaScriptCore.xcodeproj/project.pbxproj:
 470         * bindings/c/c_class.cpp:
 471         (KJS::Bindings::CClass::CClass):
 472         (KJS::Bindings::CClass::~CClass):
 473         (KJS::Bindings::CClass::classForIsA):
 474         (KJS::Bindings::CClass::methodsNamed):
 475         (KJS::Bindings::CClass::fieldNamed):
 476         * bindings/c/c_class.h:
 477         * bindings/jni/jni_class.cpp:
 478         (JavaClass::JavaClass):
 479         (JavaClass::~JavaClass):
 480         (JavaClass::methodsNamed):
 481         (JavaClass::fieldNamed):
 482         * bindings/jni/jni_class.h:
 483         * bindings/objc/objc_class.h:
 484         * bindings/objc/objc_class.mm:
 485         (KJS::Bindings::deleteMethod):
 486         (KJS::Bindings::deleteField):
 487         (KJS::Bindings::):
 488         (KJS::Bindings::ObjcClass::methodsNamed):
 489         (KJS::Bindings::ObjcClass::fieldNamed):
 490         * bindings/runtime.cpp:
 491         * bindings/runtime.h:
 492         * bindings/runtime_object.cpp:
 493         (RuntimeObjectImp::fieldGetter):
 494         (RuntimeObjectImp::getOwnPropertySlot):
 495         (RuntimeObjectImp::put):
 496         (RuntimeObjectImp::canPut):
 497
 498 2006-08-21  Vladimir Olexa  <vladimir.olexa@gmail.com>
 499
 500         Reviewed by Darin.
 501
 502         http://bugzilla.opendarwin.org/show_bug.cgi?id=6252
 503         JavaScript 1.6 Array.lastIndexOf
 504
 505         Test: fast/js/array-lastIndexOf.html
 506
 507         * kjs/array_object.cpp:
 508         (ArrayProtoFunc::callAsFunction): Added a LastIndexOf case.
 509         * kjs/array_object.h:
 510         (KJS::ArrayProtoFunc::): Added LastIndexOf to enum.
 511         * tests/mozilla/expected.html: Two more tests now pass.
 512
 513 2006-08-20  Nikolas Zimmermann  <zimmermann@kde.org>
 514
 515         Reviewed by Maciej. Landed by rwlbuis.
 516
 517         Fixes parts of: http://bugzilla.opendarwin.org/show_bug.cgi?id=10463
 518         WebKit should have Qt platform support
 519
 520         Removing obsolete QConstString/QString constructors in kjs code.
 521
 522         * kjs/identifier.h:
 523         * kjs/ustring.h:
 524
 525 2006-08-17  Nikolas Zimmermann  <zimmermann@kde.org>
 526
 527         Reviewed by Maciej. Landed by rwlbuis.
 528
 529         Fixes: http://bugzilla.opendarwin.org/show_bug.cgi?id=10463
 530         WTF Changes needed for Qt platform code.
 531
 532         * wtf/Platform.h:
 533         * wtf/unicode/UnicodeDecomposition.h: Added.
 534         (WTF::Unicode::):
 535         * wtf/unicode/UnicodeDirection.h: Added.
 536         (WTF::Unicode::):
 537         * wtf/unicode/qt4/UnicodeQt4.cpp: Added.
 538         (WTF::Unicode::direction):
 539         (WTF::Unicode::category):
 540         (WTF::Unicode::decomposition):
 541         * wtf/unicode/qt4/UnicodeQt4.h:
 542         (WTF::Unicode::toLower):
 543         (WTF::Unicode::toUpper):
 544         (WTF::Unicode::isPrintableChar):
 545         (WTF::Unicode::isSpace):
 546         (WTF::Unicode::isPunct):
 547         (WTF::Unicode::isDigit):
 548         (WTF::Unicode::mirroredChar):
 549         (WTF::Unicode::compare):
 550
 551 2006-08-17  Nikolas Zimmermann  <zimmermann@kde.org>
 552
 553         Reviewed by Eric. Landed by rwlbuis.
 554
 555         Fixes: http://bugzilla.opendarwin.org/show_bug.cgi?id=10464
 556         Offer a cmake build system for Qt platform.
 557
 558         * CMakeLists.txt: Added.
 559         * pcre/CMakeLists.txt: Added.
 560
 561 2006-08-17  Anders Carlsson  <acarlsson@apple.com>
 562
 563         Reviewed by Maciej.
 564
 565         * bindings/npapi.h:
 566         Fix ifdef.
 567
 568 2006-08-15  Steve Falkenburg  <sfalken@apple.com>
 569
 570         Reviewed by mjs.
 571
 572         Build fix.
 573
 574         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
 575         * wtf/Assertions.h:
 576
 577 2006-08-15  Mark Rowe  <opendarwin.org@bdash.net.nz>
 578
 579         Reviewed by Tim H.
 580
 581         Build fix:  DWARF and -gfull are incompatible with symbol separation.
 582
 583         * JavaScriptCore.xcodeproj/project.pbxproj:
 584
 585 2006-08-15  Mark Rowe  <opendarwin.org@bdash.net.nz>
 586
 587         Reviewed by Tim H.
 588
 589         http://bugzilla.opendarwin.org/show_bug.cgi?id=10394
 590         Bug 10394: WebKit Release and Production configurations should enable dead code stripping
 591
 592         * JavaScriptCore.xcodeproj/project.pbxproj:
 593
 594 2006-08-15  Mark Rowe  <opendarwin.org@bdash.net.nz>
 595
 596         Reviewed by Tim H.
 597
 598         http://bugzilla.opendarwin.org/show_bug.cgi?id=10384
 599         Bug 10384: Switch to DWARF for Release configuration
 600
 601         * JavaScriptCore.xcodeproj/project.pbxproj:
 602
 603 2006-08-13  Maks Orlovich  <maksim@kde.org>
 604
 605         Reviewed (and tweaked a little) by Maciej.
 606
 607         - shrank the size of JSObject by 8 bytes and made the corresponding reduction to the cell size, resulting
 608         in a 1.2% speed improvement on JS iBench (and probably overall memory savings).
 609
 610         This was done by removing _scope and _internalValue data members
 611         from JSObject and moving them only to the subclasses that actually
 612         make use of them.
 613
 614         * kjs/object.cpp:
 615         (KJS::JSObject::mark): No need to mark scope or internal value here.
 616         * kjs/object.h:
 617         (KJS::JSObject::JSObject): Don't initialize them.
 618         * kjs/JSWrapperObject.cpp: Added. New base class for object types that
 619         wrap primitive values (Number, String, Boolean, Date).
 620         (KJS::JSWrapperObject::mark):
 621         * kjs/JSWrapperObject.h: Added.
 622         (KJS::JSWrapperObject::JSWrapperObject):
 623         (KJS::JSWrapperObject::internalValue):
 624         (KJS::JSWrapperObject::setInternalValue):
 625         * kjs/array_object.cpp:
 626         (ArrayPrototype::ArrayPrototype): Don't set useless internal value.
 627         * kjs/bool_object.cpp:
 628         (BooleanInstance::BooleanInstance): Inherit from JSWrapperObject.
 629         (BooleanProtoFunc::callAsFunction): Fixed to account for fact that not all
 630         JSObjects have an internal value.
 631         (BooleanObjectImp::construct): ditto.
 632         * kjs/bool_object.h:
 633         * kjs/collector.cpp: Lowered cell size to 48.
 634         (KJS::Collector::allocate): meaningless whitespace change
 635         * kjs/date_object.cpp:
 636         (KJS::DateInstance::DateInstance): Inherit from JSWrapperObject.
 637         (KJS::DateProtoFunc::callAsFunction): adjusted for move of internalValue
 638         (KJS::DateObjectImp::construct): ditto
 639         * kjs/date_object.h:
 640         * kjs/error_object.cpp:
 641         (ErrorPrototype::ErrorPrototype): don't set internal value
 642         * kjs/function.cpp: move _scope and related handling here
 643         (KJS::FunctionImp::mark): mark scope
 644         * kjs/function.h:
 645         (KJS::FunctionImp::scope): moved here from JSObject
 646         (KJS::FunctionImp::setScope): ditto
 647         * kjs/number_object.cpp:
 648         (NumberInstance::NumberInstance): inherit from JSWrapperObject
 649         (NumberProtoFunc::callAsFunction): adjusted
 650         (NumberObjectImp::construct): adjusted
 651         * kjs/number_object.h: shring RegExp-related objects a little
 652         * kjs/regexp_object.cpp:
 653         (RegExpPrototype::RegExpPrototype): Adjust for size tweaks
 654         (RegExpObjectImp::RegExpObjectImp): ditto
 655         * kjs/regexp_object.h:
 656         * kjs/string_object.cpp:
 657         (StringInstance::StringInstance): inherit from JSWrapperObject
 658         (StringProtoFunc::callAsFunction): adjusted
 659         * kjs/string_object.h:
 660         * JavaScriptCore.exp: Exported new methods as needed.
 661         * JavaScriptCore.xcodeproj/project.pbxproj: Added new files to build.
 662
 663 2006-08-04  Brady Eidson  <beidson@apple.com>
 664
 665         Reviewed by Geoff's rubber stamp
 666
 667         Fix a build break on Intel hardware causes by adapting stricter
 668         compiler warnings (-Wshorten-64-to-32)
 669
 670         * API/testapi.c:
 671         (assertEqualsAsNumber): manually cast some doubles to floats
 672         (main): ditto
 673
 674 2006-08-04  Sam Weinig  <sam.weinig@gmail.com>
 675
 676         Reviewed by Darin.
 677
 678         - patch for http://bugzilla.opendarwin.org/show_bug.cgi?id=10192
 679           Make WebCore (and friends) compile with -Wshorten-64-to-32
 680
 681           * Adds -Wshorten-64-to-32 flag to Xcode project.
 682           * Adds explicit casts where OK.
 683
 684         * API/JSNodeList.c:
 685         (JSNodeList_item):
 686         (JSNodeList_getProperty):
 687         * JavaScriptCore.xcodeproj/project.pbxproj:
 688
 689 2006-08-04  Adam Roben  <aroben@apple.com>
 690
 691         Reviewed by Anders.
 692
 693         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: Convert
 694         spaces to tabs
 695
 696 2006-08-03  Sam Weinig  <sam.weinig@gmail.com>
 697
 698         Reviewed by Darin.
 699
 700         - patch for http://bugzilla.opendarwin.org/show_bug.cgi?id=10176
 701           Make WebCore compile with -Wundef
 702
 703           * Adds -Wundef flag to Xcode project
 704           * Converts #ifs to #ifdef and #ifndefs where needed.
 705           * Added #define YYMAXDEPTH 10000 in kjs/grammar.y
 706             to fix a warning from within Bison.
 707
 708         * JavaScriptCore.xcodeproj/project.pbxproj:
 709         * bindings/jni/jni_jsobject.cpp:
 710         (JavaJSObject::getSlot):
 711         (JavaJSObject::setSlot):
 712         * bindings/npapi.h:
 713         * bindings/objc/objc_class.mm:
 714         (KJS::Bindings::ObjcClass::methodsNamed):
 715         (KJS::Bindings::ObjcClass::fieldNamed):
 716         * bindings/objc/objc_instance.mm:
 717         (ObjcInstance::invokeMethod):
 718         * bindings/objc/objc_runtime.mm:
 719         (ObjcMethod::getMethodSignature):
 720         (ObjcField::name):
 721         (ObjcField::type):
 722         * kjs/grammar.y:
 723         * kjs/identifier.h:
 724
 725 2006-08-03  Anders Carlsson  <acarlsson@apple.com>
 726
 727         Reviewed by John Sullivan.
 728
 729        * wtf/HashSet.h:
 730         (WTF::::operator):
 731         Return *this in operator=
 732
 733 2006-08-03  Adam Roben  <aroben@apple.com>
 734
 735         Reviewed by Anders.
 736
 737         - Fixed Windows build
 738
 739         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
 740         * wtf/MathExtras.h: Implement inline versions of these functions
 741         (nextafter):
 742         (nextafterf):
 743
 744 2006-08-02  Adam Roben  <aroben@apple.com>
 745
 746         Reviewed by Darin.
 747
 748         - Fixed build
 749
 750         * kjs/date_object.cpp:
 751         (KJS::formatTime):
 752
 753 2006-07-29  Darin Adler  <darin@apple.com>
 754
 755         - Removed tabs from these source files that still had them.
 756           We don't use them; that way source files look fine in editors
 757           that have tabs set to 8 spaces or to 4 spaces.
 758         - Removed allow-tabs Subversion property from the files too.
 759
 760         * bindings/NP_jsobject.cpp:
 761         * bindings/c/c_utility.cpp:
 762         * bindings/jni/jni_runtime.cpp:
 763         * bindings/jni/jni_utility.cpp:
 764         * bindings/objc/objc_utility.mm:
 765         * bindings/runtime.cpp:
 766         * bindings/runtime_method.cpp:
 767         * bindings/testbindings.cpp:
 768         * bindings/testbindings.mm:
 769         * kjs/date_object.cpp:
 770         * kjs/function.cpp:
 771         * kjs/list.cpp:
 772         * kjs/nodes.cpp:
 773         * kjs/nodes.h:
 774         * kjs/string_object.cpp:
 775         * kjs/ustring.cpp:
 776
 777 2006-07-29  Darin Adler  <darin@apple.com>
 778
 779         * tests/mozilla/expected.html: Update test results now that regress-185165.js
 780         is succeeding. I suspect Anders fix for bug 4620655 is the reason.
 781
 782 2006-07-29  Sam Weinig  <sam.weinig@gmail.com>
 783
 784         Reviewed by Darin.
 785
 786         - patch for http://bugzilla.opendarwin.org/show_bug.cgi?id=10080
 787           Adopt pedantic changes from the Unity project to improve
 788           cross-compiler compatibility
 789
 790           Changes include:
 791           * Removing trailing semicolon from namespace braces.
 792           * Removing trailing comma from last enum declaration.
 793           * Updating to match style guidelines.
 794           * Adding missing newline to the end of the file.
 795           * Turning on gcc warning for missing newline at the end of a source file
 796             (GCC_WARN_ABOUT_MISSING_NEWLINE in Xcode, -Wnewline in gcc).
 797           * Alphabetical sorting of Xcode source list files.
 798           * Replace use of non-portable variable-size array with Vector.
 799           * Use C-style comments instead of C++ comments in files that might
 800             be included by either C or C++ files.
 801
 802         * API/JSCallbackConstructor.cpp:
 803         (KJS::JSCallbackConstructor::construct):
 804         * API/JSCallbackFunction.cpp:
 805         (KJS::JSCallbackFunction::callAsFunction):
 806         * API/JSCallbackObject.cpp:
 807         (KJS::JSCallbackObject::construct):
 808         (KJS::JSCallbackObject::callAsFunction):
 809         * JavaScriptCore.xcodeproj/project.pbxproj:
 810         * JavaScriptCorePrefix.h:
 811         * bindings/jni/jni_class.cpp:
 812         (JavaClass::fieldNamed):
 813         * bindings/jni/jni_class.h:
 814         * bindings/jni/jni_instance.cpp:
 815         (JavaInstance::JavaInstance):
 816         (JavaInstance::valueOf):
 817         * bindings/jni/jni_objc.mm:
 818         (KJS::Bindings::dispatchJNICall):
 819         * bindings/jni/jni_runtime.cpp:
 820         (JavaParameter::JavaParameter):
 821         (JavaArray::JavaArray):
 822         * bindings/jni/jni_runtime.h:
 823         * bindings/jni/jni_utility.h:
 824         * bindings/objc/objc_instance.h:
 825         * bindings/runtime_array.h:
 826         * kjs/collector.h:
 827         * kjs/config.h:
 828         * kjs/ustring.cpp:
 829         * wtf/Platform.h:
 830
 831 2006-07-29  Mike Emmel  <mike.emmel@gmail.com>
 832
 833         Reviewed by Darin.
 834
 835         - fixes for Linux build
 836
 837         * JavaScriptCoreSources.bkl: Added new files to build, kjs/PropertyNameArray.cpp
 838         and kjs/testkjs.cpp, and removed old files.
 839
 840 2006-07-24  Dan Waylonis  <waylonis@google.com>
 841
 842         Reviewed and tweaked a bit by Darin.
 843
 844         - fix http://bugzilla.opendarwin.org/show_bug.cgi?id=9902
 845           jsNull and NSNull not properly converted between JS and ObjC
 846
 847         * bindings/objc/objc_utility.mm: (KJS::Bindings::convertObjcValueToValue):
 848         Added case for converting NSNull to jsNull.
 849
 850 2006-07-24  Rob Buis  <buis@kde.org>
 851
 852         Reviewed by Darin.
 853
 854         http://bugzilla.opendarwin.org/show_bug.cgi?id=4258
 855         Date().toString() only includes GMT offset, not timezone string
 856
 857         Use the info in tm_zone to append timezone abbreviation
 858         to Date().toString().
 859
 860         * kjs/date_object.cpp:
 861         (KJS::formatTime):
 862
 863 2006-07-24  Rob Buis  <buis@kde.org>
 864
 865         Reviewed by Darin.
 866
 867         http://bugzilla.opendarwin.org/show_bug.cgi?id=5257
 868         setYear() does not match FireFox/IE behavior
 869
 870         Make sure the right values end up in tm_year.
 871
 872         * kjs/date_object.cpp:
 873         (KJS::formatTime):
 874
 875 2006-07-23  Mark Rowe  <opendarwin.org@bdash.net.nz>
 876
 877         Reviewed by Maciej.
 878
 879         Bug 9686: [Drosera] Need the ability to break into Drosera on Javascript exceptions
 880         http://bugzilla.opendarwin.org/show_bug.cgi?id=9686
 881
 882         JavaScriptCore portion of the fix.
 883
 884         * JavaScriptCore.exp: Update symbol for change in argument type.
 885         * kjs/debugger.cpp:
 886         (Debugger::detach): Clear map of recent exceptions.
 887         (Debugger::hasHandledException): Track the most recent exception
 888         thrown by an interpreter.
 889         (Debugger::exception): Change exception argument to a JSValue.
 890         * kjs/debugger.h:
 891         * kjs/nodes.cpp:
 892         (Node::debugExceptionIfNeeded): Notify the debugger of an exception
 893         if it hasn't seen it before.
 894         (ThrowNode::execute): Notify the debugger that an exception is being thrown.
 895         * kjs/nodes.h:
 896
 897     2006-07-23  Geoffrey Garen  <ggaren@apple.com>
 898
 899         Patch by Eric Albert, reviewed by Darin and me.
 900
 901         - Fixed <rdar://problem/4645931> JavaScriptCore stack-scanning code
 902         crashes (Collector::markStackObjectsConservatively)
 903
 904         * bindings/jni/jni_jsobject.cpp: On 64bit systems, jint is a long, not an
 905         int.
 906         (JavaJSObject::getSlot):
 907         (JavaJSObject::setSlot):
 908         * kjs/collector.cpp:
 909         (KJS::Collector::markCurrentThreadConservatively): Use a pointer instead of
 910         an int as 'dummy,' because on LP64 systems, an int is not pointer-aligned,
 911         and we want to scan the stack for pointers.
 912         * JavaScriptCore.xcodeproj/project.pbxproj: After a tense cease-fire, the
 913         XCode war has started up again!
 914
 915 === Safari-521.20 ===
 916
 917 2006-07-21  Geoffrey Garen  <ggaren@apple.com>
 918
 919         Reviewed by Darin.
 920
 921         <rdar://problem/4507265> REGRESSION: overlays don't work on HousingMaps.com (Google Maps-based site)
 922
 923         - Added support for strings that masquerade as undefined. Currently used
 924         by WebCore to implement undetectable style.filter.
 925
 926         The name is a little long, but it's only used in one line of code, so I
 927         thought clarity should win over brevity.
 928
 929         * JavaScriptCore.exp:
 930         * JavaScriptCore.xcodeproj/project.pbxproj:
 931         * kjs/object.h:
 932         * kjs/string_object.h:
 933         (KJS::StringInstanceThatMasqueradesAsUndefined::StringInstanceThatMasqueradesAsUndefined):
 934         (KJS::StringInstanceThatMasqueradesAsUndefined::masqueradeAsUndefined):
 935         (KJS::StringInstanceThatMasqueradesAsUndefined::toBoolean):
 936
 937 === Safari-521.19 ===
 938
 939 2006-07-20  Steve Falkenburg  <sfalken@apple.com>
 940
 941         Fix the build
 942
 943         * kjs/function.cpp:
 944         (KJS::escapeStringForPrettyPrinting):
 945
 946 2006-07-19  Anders Carlsson  <acarlsson@apple.com>
 947
 948         Reviewed by Darin.
 949
 950         <rdar://problem/4620655> REGRESSION(10.4.7-10.5): preview button for a blogger.com post doesn't work
 951
 952         * kjs/nodes2string.cpp:
 953         (StringNode::streamTo):
 954         Return the escaped string.
 955
 956         (RegExpNode::streamTo):
 957         Use the correct syntax.
 958
 959         * kjs/function.cpp:
 960         (KJS::escapeStringForPrettyPrinting):
 961         * kjs/function.h:
 962         Add escape function which escapes a string for pretty-printing so it can be parsed again.
 963
 964         * wtf/unicode/icu/UnicodeIcu.h:
 965         (WTF::Unicode::isPrintableChar):
 966         New function.
 967
 968 2006-07-18  Maciej Stachowiak  <mjs@apple.com>
 969
 970         Reviewed by Adele Peterson.
 971
 972         <rdar://problem/4589530> REGRESSION: null character in JS string causes parse error (works in Tiger and in other browsers)
 973
 974         * kjs/lexer.cpp:
 975         (Lexer::shift):
 976         (Lexer::lex):
 977         (Lexer::record16):
 978         (Lexer::scanRegExp):
 979         * kjs/lexer.h:
 980
 981 2006-07-18  Tim Omernick  <timo@apple.com>
 982
 983         Reviewed by Tim Hatcher.
 984
 985         Removed a misleading comment; we recently added support for the NPNVPluginElementNPObject
 986         variable.
 987
 988         * bindings/npapi.h:
 989
 990 === Safari-521.18 ===
 991
 992 2006-07-18  Timothy Hatcher  <timothy@apple.com>
 993
 994         Made the following headers public:
 995
 996         * JavaScriptCore.h
 997         * JSBase.h
 998         * JSContextRef.h
 999         * JSObjectRef.h
1000         * JSStringRef.h
1001         * JSValueRef.h
1002
1003         * JavaScriptCore.xcodeproj/project.pbxproj:
1004
1005 2006-07-17  Geoffrey Garen  <ggaren@apple.com>
1006
1007         Reviewed by Maciej.
1008
1009         - Added automatic prototype creation for classes.
1010
1011         A class stores a weak reference to a prototype, which is cleared when
1012         the prototype is garbage collected, to avoid a reference cycle.
1013
1014         We now have an attributes field in JSClassDefinition, that currently is
1015         used only to override automatic prototype creation when you want to manage your
1016         own prototypes, but can be extended in the future for other nefarious purposes.
1017
1018         Similarly, we have JSObjectMake and JSObjectMakeWithPrototype, the latter
1019         allowing you to manage your own prototypes.
1020
1021         JSObjectMakeConstructor is more interesting now, able to make a constructor
1022         on your behalf if you just give it a class.
1023
1024         - Removed bogus old code from minidom.js.
1025
1026         - Tweaked the headerdocs.
1027
1028         - Added more GC testing, which caught some leaks, and tested more funny
1029         edge cases in lookup, which caught a lookup bug. Removed some testing
1030         we used to do with MyObject because it was redundant with the new, cool
1031         stuff.
1032
1033         While fixing the lookup bug I retracted this change:
1034
1035             "If a static setProperty callback returns 'false', to indicate that the
1036             property was not set, we no longer forward the set request up the class
1037             chain, because that's almost certainly not what the programmer expected."
1038
1039         Returning false when setting a static property is a little silly, but you can see
1040         it being useful when shadowing a base class's static properties, and, regardless
1041         of usefullness, this is the defined behavior of the setProperty callback.
1042
1043         - Plus a little ASCII art, for the kids.
1044
1045 2006-07-17  Timothy Hatcher  <timothy@apple.com>
1046
1047         Reviewed by Maciej.
1048
1049         <rdar://problem/4634874> WebScriptObject and WebUndefined are no longer defined by WebKit
1050
1051         Moves WebScriptObject and WebUndefined up to WebCore.
1052         This change does create an upwards-dependancy on WebScriptObject existing
1053         in the loaded process, but this code path in JavaScriptCore does not get used
1054         unless it is through WebKit/WebCore. Moving all of the binding code out of
1055         JavaScriptCore might make sense in the future.
1056
1057         * JavaScriptCore.exp:
1058         * JavaScriptCore.xcodeproj/project.pbxproj:
1059         * bindings/objc/WebScriptObject.h: Replaced.
1060         * bindings/objc/WebScriptObject.mm: Removed.
1061         * bindings/objc/WebScriptObjectPrivate.h: Removed.
1062         * bindings/objc/objc_class.h:
1063         * bindings/objc/objc_instance.h:
1064         * bindings/objc/objc_instance.mm:
1065         (ObjcInstance::~ObjcInstance):
1066         * bindings/objc/objc_runtime.h:
1067         * bindings/objc/objc_runtime.mm:
1068         (convertValueToObjcObject):
1069         * bindings/objc/objc_utility.mm:
1070         (KJS::Bindings::convertValueToObjcValue):
1071         (KJS::Bindings::convertObjcValueToValue):
1072         (KJS::Bindings::createObjcInstanceForValue):
1073
1074 2006-07-17  Darin Adler  <darin@apple.com>
1075
1076         * API/JSBase.h: Fix comment formatting where things used to be lined up but
1077         are now ragged. Got rid of spaces that attempted to line things up.
1078         * API/JSObjectRef.h: Ditto. Also add missing periods for a couple of comments.
1079
1080 2006-07-17  Geoffrey Garen  <ggaren@apple.com>
1081
1082         Reviewed by Maciej.
1083
1084         - Removed the exception parameter from the initialize callback and, by extension,
1085         JSObjectMake. We have never had a need for exceptions when iniitializing,
1086         so the parameter seemed likely to "get in the way."
1087
1088         Also, an exception in JavaScript is thrown in response to input --
1089         "invalid URL", "index not a number", etc., so it's the job of the
1090         constructor function, not the initialize method, to throw.
1091
1092         If initialize *really* wants to throw, it can communicate the throw to
1093         the constructor through the constructed object's private data (e.g., set
1094         it to NULL, signaling to the consntructor that initialization failed.)
1095
1096         - Added JSObjectMakeWithData, which enables a constructor to set private
1097         data on an object *before* it has been initialized. That way, the initialize
1098         methods can properly operate on the data.
1099
1100         * API/JSNode.c: Moved ref into the initialize method, for better encapsulation,
1101         now that it's possible.
1102         * API/JSNodeList.c: ditto
1103         * API/minidom.c:
1104         (main): Do more aggressive garbage collection to test ref/deref and
1105         initialize/finalize.
1106         * API/minidom.js: store childNodes in a temporary so it doesn't get re-created
1107         like a thousand times. This makes debugging ref/deref easier
1108
1109 2006-07-17  Geoffrey Garen  <ggaren@apple.com>
1110
1111         Reviewed by Maciej.
1112
1113         - Changed the initialize callback to run from least derived class (parent
1114         class) to most derived class. This enables C++ style initialization,
1115         and derived class overriding of member data.
1116
1117         - Added excpetion propopgation to JSObjectMake, to support initialize
1118         exceptions, and generally round out our policy of making function
1119         signatures as long as possible.
1120
1121         * API/JSCallbackObject.h: Use ExecState instead of ContextRef, cuz we're
1122         in C++ land now.
1123
1124 2006-07-17  Geoffrey Garen  <ggaren@apple.com>
1125
1126         Reviewed by Maciej.
1127
1128         - Changed JSObjectMakeConstructor to JSObjectMakeConstructorWithCallback,
1129         to match JSObjectMakeFunctionWithCallback.
1130
1131         - Added prototype parameter, so the generated constructor
1132         automatically works with hasInstance / instanceof
1133
1134         - Moved hasInstance implementation from InternalFunctionImp to JSObject
1135         so that subclasses can inherit it without inheriting function-related baggage.
1136         More refactoring here would be good, but this seems like a good short-term
1137         solution.
1138
1139         (KJS::JSCallbackFunction::implementsHasInstance): override and return false,
1140         because callback functions aren't constructors.
1141
1142 2006-07-17  Maciej Stachowiak  <mjs@apple.com>
1143
1144         Reviewed by Geoff.
1145
1146         - add a JSContextRef parameter to all JSValueRef, JSObjectRef, and JSContextRef operations;
1147         except JSObject{Get,Set}PrivateData which can be assumed to be simple pure accessors.
1148
1149         Also renamed the parameter "context" to "ctx" because it makes the code read better with this pervasive
1150         but usually uninteresting parameter.
1151
1152         * API/JSBase.cpp:
1153         (JSEvaluateScript):
1154         (JSCheckScriptSyntax):
1155         (JSGarbageCollect):
1156         * API/JSBase.h:
1157         * API/JSCallbackObject.cpp:
1158         (KJS::JSCallbackObject::JSCallbackObject):
1159         (KJS::JSCallbackObject::init):
1160         (KJS::JSCallbackObject::getOwnPropertySlot):
1161         (KJS::JSCallbackObject::put):
1162         (KJS::JSCallbackObject::deleteProperty):
1163         (KJS::JSCallbackObject::toNumber):
1164         (KJS::JSCallbackObject::toString):
1165         * API/JSContextRef.cpp:
1166         (JSGlobalContextCreate):
1167         (JSGlobalContextRetain):
1168         (JSGlobalContextRelease):
1169         (JSContextGetGlobalObject):
1170         * API/JSContextRef.h:
1171         * API/JSNode.c:
1172         (JSNodePrototype_appendChild):
1173         (JSNodePrototype_removeChild):
1174         (JSNodePrototype_replaceChild):
1175         (JSNode_getNodeType):
1176         (JSNode_getFirstChild):
1177         (JSNode_prototype):
1178         * API/JSNodeList.c:
1179         (JSNodeListPrototype_item):
1180         (JSNodeList_length):
1181         (JSNodeList_getProperty):
1182         (JSNodeList_prototype):
1183         * API/JSObjectRef.cpp:
1184         (JSObjectMake):
1185         (JSObjectMakeFunctionWithCallback):
1186         (JSObjectMakeConstructor):
1187         (JSObjectMakeFunction):
1188         (JSObjectGetPrototype):
1189         (JSObjectSetPrototype):
1190         (JSObjectHasProperty):
1191         (JSObjectGetProperty):
1192         (JSObjectSetProperty):
1193         (JSObjectGetPropertyAtIndex):
1194         (JSObjectSetPropertyAtIndex):
1195         (JSObjectDeleteProperty):
1196         (JSObjectIsFunction):
1197         (JSObjectCallAsFunction):
1198         (JSObjectIsConstructor):
1199         (JSObjectCallAsConstructor):
1200         (JSObjectCopyPropertyNames):
1201         * API/JSObjectRef.h:
1202         * API/JSStringRef.cpp:
1203         * API/JSValueRef.cpp:
1204         (JSValueGetType):
1205         (JSValueIsUndefined):
1206         (JSValueIsNull):
1207         (JSValueIsBoolean):
1208         (JSValueIsNumber):
1209         (JSValueIsString):
1210         (JSValueIsObject):
1211         (JSValueIsObjectOfClass):
1212         (JSValueIsEqual):
1213         (JSValueIsStrictEqual):
1214         (JSValueIsInstanceOfConstructor):
1215         (JSValueMakeUndefined):
1216         (JSValueMakeNull):
1217         (JSValueMakeBoolean):
1218         (JSValueMakeNumber):
1219         (JSValueMakeString):
1220         (JSValueToBoolean):
1221         (JSValueToNumber):
1222         (JSValueToStringCopy):
1223         (JSValueToObject):
1224         (JSValueProtect):
1225         (JSValueUnprotect):
1226         * API/JSValueRef.h:
1227         * API/minidom.c:
1228         (print):
1229         * API/testapi.c:
1230         (MyObject_getProperty):
1231         (MyObject_deleteProperty):
1232         (MyObject_callAsFunction):
1233         (MyObject_callAsConstructor):
1234         (MyObject_convertToType):
1235         (print_callAsFunction):
1236         (main):
1237
1238 2006-07-16  Geoffrey Garen  <ggaren@apple.com>
1239
1240         Approved by Maciej, RS by Beth.
1241
1242         JSObjectMakeFunction -> JSObjectMakeFunctionWithCallback
1243         JSObjectMakeFunctionWithBody -> JSObjectMakeFunction
1244
1245         because the latter is more common, and more fundamental, than the former.
1246
1247         * API/APICast.h:
1248         (toJS):
1249         * API/JSBase.h:
1250         * API/JSCallbackObject.cpp:
1251         (KJS::JSCallbackObject::getOwnPropertySlot):
1252         (KJS::JSCallbackObject::put):
1253         (KJS::JSCallbackObject::deleteProperty):
1254         (KJS::JSCallbackObject::getPropertyNames):
1255         (KJS::JSCallbackObject::staticValueGetter):
1256         (KJS::JSCallbackObject::staticFunctionGetter):
1257         * API/JSClassRef.cpp:
1258         (OpaqueJSClass::OpaqueJSClass):
1259         (OpaqueJSClass::~OpaqueJSClass):
1260         * API/JSClassRef.h:
1261         * API/JSObjectRef.cpp:
1262         (JSClassCreate):
1263         (JSObjectMakeFunctionWithCallback):
1264         (JSObjectMakeFunction):
1265         (OpaqueJSPropertyNameArray::OpaqueJSPropertyNameArray):
1266         (JSObjectCopyPropertyNames):
1267         * API/JSObjectRef.h:
1268         * API/minidom.c:
1269         (main):
1270         * API/testapi.c:
1271         (main):
1272         * ChangeLog:
1273         * JavaScriptCore.exp:
1274
1275 2006-07-16  Geoffrey Garen  <ggaren@apple.com>
1276
1277         Laughed at by Beth.
1278
1279         Replace __JS with OpaqueJS because the former, while used by CF, is
1280         a prefix that's triply-reserved by the compiler. (_* is reserved in global
1281         names, _[A-Z] is reserved in all names, and __ is reserved in all names
1282         in C++.)
1283
1284         Opaque is an alternative used by other Mac OS X framewokrs.
1285
1286         * API/APICast.h:
1287         (toJS):
1288         * API/JSBase.h:
1289         * API/JSCallbackObject.cpp:
1290         (KJS::JSCallbackObject::getOwnPropertySlot):
1291         (KJS::JSCallbackObject::put):
1292         (KJS::JSCallbackObject::deleteProperty):
1293         (KJS::JSCallbackObject::getPropertyNames):
1294         (KJS::JSCallbackObject::staticValueGetter):
1295         (KJS::JSCallbackObject::staticFunctionGetter):
1296         * API/JSClassRef.cpp:
1297         (OpaqueJSClass::OpaqueJSClass):
1298         (OpaqueJSClass::~OpaqueJSClass):
1299         * API/JSClassRef.h:
1300         * API/JSObjectRef.cpp:
1301         (JSClassCreate):
1302         (OpaqueJSPropertyNameArray::OpaqueJSPropertyNameArray):
1303         (JSObjectCopyPropertyNames):
1304
1305 2006-07-16  Darin Adler  <darin@apple.com>
1306
1307         - try to fix Windows build
1308
1309         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
1310         Added some recently added files, removed some recently removed.
1311
1312 2006-07-16  Geoffrey Garen  <ggaren@apple.com>
1313
1314         Reviewed by Maciej.
1315
1316         - Change getProperty* to return undefined, rather than NULL, for missing
1317         properties, since that's what the spec says. Also added exception out
1318         parameters to the *Index functions, because they can call through to the
1319         regular functions, which can throw for custom objects.
1320
1321         * API/JSObjectRef.cpp:
1322         (JSObjectGetProperty):
1323         (JSObjectGetPropertyAtIndex):
1324         (JSObjectSetPropertyAtIndex):
1325         * API/JSObjectRef.h:
1326         * API/testapi.c:
1327         (main):
1328
1329 2006-07-16  Geoffrey Garen  <ggaren@apple.com>
1330
1331         Reviewed by Maciej.
1332
1333         - Properly document and handle NULL callbacks for static properties. We
1334         throw an exception in any case other than a ReadOnly property with a NULL
1335         setProperty callback, because a NULL callback almost certainly indicates
1336         a programming error. Also throw an exception if hasProperty returns true
1337         for a property that getProperty can't get.
1338
1339         - If a static setProperty callback returns 'false', to indicate that the
1340         property was not set, we no longer forward the set request up the class
1341         chain, because that's almost certainly not what the programmer expected.
1342
1343         * API/JSCallbackObject.cpp:
1344         (KJS::JSCallbackObject::getOwnPropertySlot):
1345         (KJS::JSCallbackObject::put):
1346         (KJS::JSCallbackObject::staticValueGetter):
1347         (KJS::JSCallbackObject::staticFunctionGetter):
1348         (KJS::JSCallbackObject::callbackGetter):
1349         * API/JSObjectRef.h:
1350         * API/minidom.js:
1351         * API/testapi.c:
1352         (MyObject_hasProperty):
1353         * API/testapi.js:
1354
1355 2006-07-16  Geoffrey Garen  <ggaren@apple.com>
1356
1357         Reviewed by Maciej.
1358
1359         - Added names to functions.
1360
1361         - Removed GetPrivate/SetPrivate from callbackFunctions and callbackConstructors.
1362         The private data idiom is that a JS object stores its native implementation
1363         as private data. For functions and constructors, the native implementation is nothing
1364         more than the callback they already store, so supporting private data, too,
1365         confuses the idiom. If you *really* want, you can still create a custom
1366         function with private data.
1367
1368         * API/JSCallbackConstructor.cpp:
1369         * API/JSCallbackConstructor.h:
1370         * API/JSCallbackFunction.cpp:
1371         (KJS::JSCallbackFunction::JSCallbackFunction):
1372         * API/JSCallbackFunction.h:
1373         * API/JSCallbackObject.cpp:
1374         (KJS::JSCallbackObject::staticFunctionGetter):
1375         * API/JSObjectRef.cpp:
1376         (JSObjectMakeFunction):
1377         (JSObjectMakeFunctionWithBody):
1378         (JSObjectGetPrivate):
1379         (JSObjectSetPrivate):
1380         * API/JSObjectRef.h:
1381         * API/minidom.c:
1382         (main):
1383         * API/testapi.c:
1384         (main):
1385
1386 2006-07-15  Maciej Stachowiak  <mjs@apple.com>
1387
1388         Reviewed by Darin.
1389
1390         - switch property lists to be vector+set of Identifiers instead of list of References
1391
1392         This has the following benefits:
1393
1394         - no duplicates in property lists
1395         - simplifies API calls
1396         - probably more efficient, since linked list is gone
1397         - entirely removed Reference, ReferenceList and ProtectedReference types from the API
1398
1399         * kjs/PropertyNameArray.cpp: Added.
1400         (KJS::PropertyNameArray::add): Check set, if not already there, add to
1401         vector.
1402         * kjs/PropertyNameArray.h: Added.
1403         (KJS::PropertyNameArray::PropertyNameArray): Newly added type, combines
1404         a set and a vector to make a unique but ordered list of identifiers.
1405         (KJS::PropertyNameArray::begin): ditto
1406         (KJS::PropertyNameArray::end): ditto
1407         (KJS::PropertyNameArray::size): ditto
1408         (KJS::PropertyNameArray::operator[]): ditto
1409         * kjs/array_instance.h:
1410         * kjs/array_object.cpp:
1411         (ArrayInstance::getPropertyNames): renamed from getPropertyList, updated
1412         for PropertyNameArray
1413         (ArrayInstance::setLength): updated for PropertyNameArray
1414         (ArrayInstance::pushUndefinedObjectsToEnd): ditto
1415         * kjs/nodes.cpp:
1416         (ForInNode::execute): updated for PropertyNameArray
1417         * kjs/nodes.h:
1418         * kjs/object.cpp:
1419         (KJS::JSObject::getPropertyNames): renamed from getPropertyList, updated
1420         for PropertyNameArray
1421         * kjs/object.h:
1422         * kjs/property_map.cpp:
1423         (KJS::PropertyMap::getEnumerablePropertyNames): updated for PropertyNameArray
1424         (KJS::PropertyMap::getSparseArrayPropertyNames): ditto
1425         * kjs/property_map.h:
1426         * kjs/protected_reference.h: Removed.
1427         * kjs/reference.cpp: Removed.
1428         * kjs/reference.h: Removed.
1429         * kjs/reference_list.cpp: Removed.
1430         * kjs/reference_list.h: Removed.
1431         * kjs/scope_chain.cpp:
1432         (KJS::ScopeChain::print): Use PropertyNamesArray instead of ReferenceList.
1433         * kjs/string_object.cpp:
1434         (StringInstance::getPropertyNames): Updated for new approach.
1435         * kjs/string_object.h:
1436         * kjs/ustring.h:
1437         * API/APICast.h:
1438         (toJS): Added overload for PropertyNameAccumulatorRef / PropertyNameArray*
1439         (toRef): ditto
1440         * API/JSBase.h:
1441         * API/JSCallbackObject.cpp:
1442         (KJS::JSCallbackObject::getPropertyNames): Fixed for new API.
1443         * API/JSCallbackObject.h:
1444         * API/JSObjectRef.cpp:
1445         (__JSPropertyNameArray::__JSPropertyNameArray): Type used for a publicly vended
1446         JSPropertyNameArrayRef.
1447         (JSObjectCopyPropertyNames): New API call - renamed / refactored from
1448         JSObjectCreatePropertyList
1449         (JSPropertyNameArrayRetain): new retain call for JSPropertyNameArray.
1450         (JSPropertyNameArrayRelease): new release call for - " -.
1451         (JSPropertyNameArrayGetCount): Instead of having to use a stateful enumerator you
1452         can now get the count and items in any order.
1453         (JSPropertyNameArrayGetNameAtIndex): See above.
1454         (JSPropertyNameAccumulatorAddName): What you add properties to is now an opaque
1455         accumulator object.
1456         * API/JSObjectRef.h: Prototyped new functions, removed old ones
1457         * JavaScriptCore.exp: Updated exported symbols.
1458         * JavaScriptCore.xcodeproj/project.pbxproj: Added new files, removed old.
1459         * API/testapi.c:
1460         (MyObject_getPropertyNames): Renamed / fixed callback to fit new paradigm.
1461         (main): Updated for new API.
1462
1463 2006-07-15  Darin Adler  <darin@apple.com>
1464
1465         - oops, missed a few more arrays that had to be const
1466
1467         * API/JSNode.c:
1468         (JSNodePrototype_appendChild): Added const.
1469         (JSNodePrototype_removeChild): Ditto.
1470         (JSNodePrototype_replaceChild): Ditto.
1471         (JSNode_construct): Ditto.
1472         * API/JSNodeList.c:
1473         (JSNodeListPrototype_item): Ditto.
1474         * API/JSObjectRef.cpp:
1475         (JSObjectMakeFunctionWithBody): Ditto.
1476         (JSObjectCallAsFunction): Ditto.
1477         (JSObjectCallAsConstructor): Ditto.
1478         * API/minidom.c:
1479         (print): Ditto.
1480         * API/testapi.c:
1481         (MyObject_callAsFunction): Ditto.
1482         (MyObject_callAsConstructor): Ditto.
1483         (print_callAsFunction): Ditto.
1484         (myConstructor_callAsConstructor): Ditto.
1485
1486 2006-07-15  Darin Adler  <darin@apple.com>
1487
1488         Reviewed by Maciej.
1489
1490         * API/JSNode.h: Made an array parameter const.
1491         * API/JSObjectRef.h: Made array parameters const. Fixed a comment.
1492
1493 2006-07-15  Geoffrey Garen  <ggaren@apple.com>
1494
1495         Reviewed by Maciej.
1496
1497         - JSObjectMakeFunctionWithBody includes a function name and named parameters now.
1498
1499         * API/JSObjectRef.cpp:
1500         (JSObjectMakeFunctionWithBody):
1501         * API/JSObjectRef.h:
1502         * API/testapi.c:
1503         (assertEqualsAsUTF8String): More informative failure reporting.
1504         (main): Test more function cases.
1505
1506 2006-07-15  Geoffrey Garen  <ggaren@apple.com>
1507
1508         Reviewed by Maciej.
1509
1510         - Moved the arguments passed to JSClassCreate into a single structure,
1511         called JSClassDefinition. This will enable easier structure
1512         migration/versioning in the future, if necessary.
1513
1514         - Added support for class names.
1515
1516         - kJSClassDefinitionNull replaces kJSObjectCallbacksNone.
1517
1518         - JSClass is becoming a fairly complex struct, so I migrated all of its
1519         implementation other than reference counting to the sruct.
1520
1521         - Also moved JSClass* functions in the API to JSObjectRef.cpp, since they're
1522         declared in JSObjectRef.h
1523
1524         - Also added some more informative explanation to the class structure doc.
1525
1526 2006-07-15  Darin Adler  <darin@apple.com>
1527
1528         Reviewed by Geoff.
1529
1530         - fix http://bugzilla.opendarwin.org/show_bug.cgi?id=8395
1531           <rdar://problem/4613467>
1532           REGRESSION: RegEx seems broken for hex escaped non breaking space
1533
1534         Test: fast/js/regexp-extended-characters-more.html
1535
1536         * pcre/pcre_exec.c:
1537         (match): Got rid of utf16Length local variable to guarantee there's no
1538         extra stack usage in recursive calls. Fixed two places in the PCRE_UTF16
1539         code that were using the length variable, which is the UTF-8 length of
1540         a character in the pattern, to move in the UTF-16 subject string. Instead
1541         they hardcode lengths of 1 and 2 since the code already handles BMP
1542         characters and surrogate pairs separately. Also fixed some DPRINTF so
1543         I could compile with DEBUG on.
1544         (pcre_exec): Changed a place that was checking for multibyte characters
1545         in the subject string to use ISMIDCHAR. Instead it was using hardcoded
1546         logic that was right for UTF-8 but wrong for UTF-16.
1547
1548         * pcre/pcre_compile.c: (pcre_compile2): Fixed a DPRINTF so I could compile
1549         with DEBUG on.
1550
1551 2006-07-14  Geoffrey Garen  <ggaren@apple.com>
1552
1553         RS by Maciej.
1554
1555         Global replace in the API of argc/argv with argumentCount/arguments.
1556
1557 2006-07-14  Geoffrey Garen  <ggaren@apple.com>
1558
1559         Reviewed by Maciej.
1560
1561         - Finalized exception handling in the API.
1562
1563         setProperty can throw because it throws for built-in arrays. getProperty
1564         and deleteProperty can throw because setProperty can throw and we want
1565         to be consistent, and also because they seem like "actions." callAsFunction,
1566         callAsConstructor, and hasInstance can throw, because they caan throw for
1567         all built-ins.
1568
1569         toBoolean can't throw because it's defined that way in the spec.
1570
1571         - Documented that toBoolean and toObject can't be overridden by custom
1572         objects because they're defined that way in the spec.
1573
1574 === Safari-521.17 ===
1575
1576 2006-07-14  Geoffrey Garen  <ggaren@apple.com>
1577
1578         Reviewed by Maciej.
1579
1580         - Implemented ref-counting of JSContexts by splitting into two datatypes:
1581         JSGlobalContext, which you can create/retain/release, and JSContext, which
1582         you can't.
1583
1584         Internally, you retain a JSGlobalContext/ExecState by retaining its
1585         interpreter, which, in the case of a global ExecState, owns it.
1586
1587         - Also made ~Interpreter() protected to catch places where Interpreter
1588         is manually deleted. (Can't make it private because some crazy fool
1589         decided it would be a good idea to subclass Interpreter in other frameworks.
1590         I pity da fool.)
1591
1592         * API/APICast.h:
1593         (toJS): Added cast for new JSGlobalContext
1594         * API/JSStringRef.h: Changed vague "you must" language to more specific
1595         (but, ultimately, equally vague) "behavior is undefined if you don't"
1596         language.
1597         (KJS::Interpreter::Interpreter): Factored more common initialization into
1598         init()
1599         * kjs/interpreter.h:
1600         (KJS::Interpreter::ref): new
1601         (KJS::Interpreter::deref): new
1602         (KJS::Interpreter::refCount): new
1603         * kjs/testkjs.cpp:
1604         (doIt): Ref-count the interpreter.
1605
1606 2006-07-14  Maciej Stachowiak  <mjs@apple.com>
1607
1608         Reviewed by Geoff.
1609
1610         - removed bool return value from JSObjectSetProperty, since it is inefficient and
1611         also doesn't work quite right
1612         - added JSObjectGetPropertyAtIndex and JSObjectSetPropertyAtIndex
1613
1614         * API/JSObjectRef.cpp:
1615         (JSObjectSetProperty): Removed return value and canPut stuff.
1616         (JSObjectGetPropertyAtIndex): Added.
1617         (JSObjectSetPropertyAtIndex): Added.
1618         * API/JSObjectRef.h: Prototyped and documented new functions.
1619
1620 2006-07-14  Geoffrey Garen  <ggaren@apple.com>
1621
1622         Reviewed by Beth.
1623
1624         Moved JSCheckScriptSyntax, JSEvaluateScript, and JSGarbageCollect into
1625         JSBase.h/.cpp. They don't belong in the value-specific or context-specific
1626         files because they're not part of the value or context implementations.
1627
1628         * API/JSBase.h:
1629         * API/JSContextRef.cpp:
1630         (JSContextGetGlobalObject):
1631         * API/JSContextRef.h:
1632         * API/JSValueRef.cpp:
1633         (JSValueUnprotect):
1634         * API/JSValueRef.h:
1635         * JavaScriptCore.xcodeproj/project.pbxproj:
1636
1637 2006-07-13  Timothy Hatcher  <timothy@apple.com>
1638
1639         Reviewed by Maciej.
1640
1641         Moved JavaScriptCore to be a public framework.
1642
1643         * JavaScriptCore.xcodeproj/project.pbxproj:
1644
1645 2006-07-13  Mark Rowe  <opendarwin.org@bdash.net.nz>
1646
1647         Reviewed by Geoffrey.
1648
1649         http://bugzilla.opendarwin.org/show_bug.cgi?id=9742
1650         Bug 9742: REGRESSION: WebKit hangs when loading <http://www.vtbook.com>
1651
1652         * kjs/value.h:
1653         (KJS::JSValue::getUInt32): Only types tagged as numeric can be converted to UInt32.
1654
1655 2006-07-13  Geoffrey Garen  <ggaren@apple.com>
1656
1657         Pleasing to Maciej.
1658
1659         - Renamed JSEvaluate -> JSEvaluateScript, JSCheckSyntax -> JSCheckScriptSyntax
1660         - Added exception out parameters to JSValueTo* and JSValueIsEqual because
1661         they can throw
1662         - Removed JSObjectGetDescription because it's useless and vague, and
1663         JSValueToString/JSValueIsObjectOfClass do a better job, anyway
1664         - Clarified comments about "IsFunction/Constructor" to indicate that they
1665         are true of all functions/constructors, not just those created by JSObjectMake*
1666
1667 2006-07-12  Geoffrey Garen  <ggaren@apple.com>
1668
1669         RS by Beth.
1670
1671         Finished previously approved JSInternalString -> JSString conversion
1672         by renaming the files.
1673
1674         * API/JSCallbackObject.cpp:
1675         * API/JSInternalStringRef.cpp: Removed.
1676         * API/JSInternalStringRef.h: Removed.
1677         * API/JSStringRef.cpp: Added.
1678         * API/JSStringRef.h: Added.
1679         * API/JavaScriptCore.h:
1680         * JavaScriptCore.xcodeproj/project.pbxproj:
1681
1682 2006-07-12  Geoffrey Garen  <ggaren@apple.com>
1683
1684         Reviewed by Maciej.
1685
1686         - Removed context and exception parameters from JSObjectGetPropertyEnumerator,
1687         removing the spurious use of ExecState inside JavaScriptCore that made
1688         us think this was necessary in the first place.
1689
1690         (StringInstance::getPropertyList): Use getString instead of toString because
1691         we know we're dealing with a string -- we put it there in the first place.
1692         While we're at it, store the string's size instead of retrieving it each time
1693         through the loop, to avoid the unnecessary killing of puppies.
1694         * kjs/string_object.h:
1695
1696 2006-07-12  Maciej Stachowiak  <mjs@apple.com>
1697
1698         Reviewed by Geoff.
1699
1700         - add handling of hasInstance callback for API objects
1701
1702         * API/JSCallbackObject.cpp:
1703         (KJS::JSCallbackObject::implementsHasInstance): Check if callback is present.
1704         (KJS::JSCallbackObject::hasInstance): Invoke appropriate callback.
1705         * API/JSCallbackObject.h:
1706         * API/JSClassRef.cpp:
1707         * API/JSObjectRef.h:
1708         * API/testapi.c:
1709         (MyObject_hasInstance): Test case; should match what construct would do.
1710         * API/testapi.js:
1711
1712 2006-07-11  Geoffrey Garen  <ggaren@apple.com>
1713
1714         Reviewed by Maciej.
1715
1716         - Implemented a vast number of renames and comment clarifications
1717         suggested during API review.
1718
1719         JSInternalString -> JSString
1720         JS*Make -> JSValueMake*, JSObjectMake*
1721         JSTypeCode -> JSType
1722         JSValueIsInstanceOf -> JSValueIsInstanceOfConstructor (reads strangely well in client code)
1723         JSGC*Protect -> JSValue*Protect
1724         JS*Callback -> JSObject*Callback
1725         JSGetPropertyListCallback -> JSObjectAddPropertiesToListCallback
1726         JSPropertyEnumeratorGetNext -> JSPropertyEnumeratorGetNextName
1727         JSString* ->
1728             JSStringCreateWithUTF8CString, JSStringGetUTF8CString,
1729             JSStringGetMaximumUTF8CStringSize JSStringIsEqualToUTF8CString,
1730             JSStringCreateWithCFString, JSStringCopyCFString, JSStringCreateWithCharacters.
1731
1732         - Changed functions taking a JSValue out arg and returning a bool indicating
1733         whether it was set to simply return a JSValue or NULL.
1734
1735         - Removed JSStringGetCharacters because it's more documentation than code,
1736         and it's just a glorified memcpy built on existing API functionality.
1737
1738         - Moved standard library includes into the headers that actually require them.
1739
1740         - Standardized use of the phrase "Create Rule."
1741
1742         - Removed JSLock from make functions that don't allocate.
1743
1744         - Added exception handling to JSValueToBoolean, since we now allow
1745         callback objects to throw exceptions upon converting to boolean.
1746
1747         - Renamed JSGCCollect to JSGarbageCollect.
1748
1749 2006-07-10  Geoffrey Garen  <ggaren@apple.com>
1750
1751         Reviewed by Darin.
1752
1753         - Changed public header includes to the <JavaScriptCore/ style.
1754         - Changed instances of 'buffer' to 'string' since we decided on
1755         JSInternalString instead of JSStringBuffer.
1756
1757         * API/JSContextRef.h:
1758         * API/JSInternalStringRef.cpp:
1759         (JSStringMake):
1760         (JSInternalStringRetain):
1761         (JSInternalStringRelease):
1762         (JSValueCopyStringValue):
1763         (JSInternalStringGetLength):
1764         (JSInternalStringGetCharactersPtr):
1765         (JSInternalStringGetCharacters):
1766         (JSInternalStringGetMaxLengthUTF8):
1767         (JSInternalStringGetCharactersUTF8):
1768         (CFStringCreateWithJSInternalString):
1769         * API/JSInternalStringRef.h:
1770         * API/JSNode.c:
1771         (JSNodePrototype_appendChild):
1772         (JSNode_getNodeType):
1773         * API/JSObjectRef.cpp:
1774         (JSObjectCallAsConstructor):
1775         * API/JSValueRef.h:
1776         * API/JavaScriptCore.h:
1777         * API/minidom.c:
1778         (main):
1779         (print):
1780         * API/testapi.c:
1781         (MyObject_getPropertyList):
1782         (myConstructor_callAsConstructor):
1783         (main): I noticed that we were prematurely releasing some string buffers,
1784         so I moved their release calls to the end of main(). I got rid of 'Buf' in *Buf
1785         (sometimes changing to 'IString', when necessary to differentiate a variable)
1786         to match the buffer->string change.
1787
1788 === Safari-521.16 ===
1789
1790 2006-07-10  Darin Adler  <darin@apple.com>
1791
1792         * kjs/value.cpp: (KJS::JSValue::toInt32Inline): Added inline keyword one more place.
1793         Just in case.
1794
1795 2006-07-10  Darin Adler  <darin@apple.com>
1796
1797         - fix the release build
1798
1799         * kjs/value.h:
1800         * kjs/value.cpp:
1801         (KJS::JSValue::toInt32Inline): Move the code here to an inline.
1802         (KJS::JSValue::toInt32): Call the inline from both overloaded toInt32 functions.
1803
1804 2006-07-10  David Kilzer  <ddkilzer@kilzer.net>
1805
1806         Reviewed by Darin.
1807
1808         - fix http://bugzilla.opendarwin.org/show_bug.cgi?id=9179
1809           Implement select.options.add() method
1810
1811         * JavaScriptCore.exp: Added overloaded KJS::JSValue::toInt32() method.
1812         * JavaScriptCore.xcodeproj/project.pbxproj: Altered attributes metadata for
1813         kjs/value.h to make it available as a forwarded header.
1814         * kjs/lookup.h:
1815         (KJS::lookupPut): Extracted a lookupPut() method from the existing lookupPut() method.
1816         The new method returns a boolean value if no entry is found in the lookup table.
1817         * kjs/value.cpp:
1818         (KJS::JSValue::toInt32): Overloaded toInt32() method with boolean "Ok" argument.
1819         * kjs/value.h: Ditto.
1820
1821 2006-07-10  Geoffrey Garen  <ggaren@apple.com>
1822
1823         No review necessary. Removed bogus file I accidentally checked in before.
1824
1825         * API/JSInternalSringRef.h: Removed.
1826
1827 2006-07-10  Geoffrey Garen  <ggaren@apple.com>
1828
1829         Reviewed by Darin.
1830
1831         Added exception out parameter to API object callbacks, removed semi-bogus
1832         JSContext(.*)Exception functions.
1833
1834         To make these calls syntactically simple, I added an exceptionSlot()
1835         method to the ExecState class, which provides a JSValue** slot in which to
1836         store a JSValue* exception.
1837
1838         * API/APICast.h:
1839         (toRef):
1840         * API/JSCallbackConstructor.cpp:
1841         (KJS::JSCallbackConstructor::construct):
1842         * API/JSCallbackFunction.cpp:
1843         (KJS::JSCallbackFunction::callAsFunction):
1844         * API/JSCallbackObject.cpp:
1845         (KJS::JSCallbackObject::init):
1846         (KJS::JSCallbackObject::getOwnPropertySlot):
1847         (KJS::JSCallbackObject::put):
1848         (KJS::JSCallbackObject::deleteProperty):
1849         (KJS::JSCallbackObject::construct):
1850         (KJS::JSCallbackObject::callAsFunction):
1851         (KJS::JSCallbackObject::getPropertyList):
1852         (KJS::JSCallbackObject::toBoolean):
1853         (KJS::JSCallbackObject::toNumber):
1854         (KJS::JSCallbackObject::toString):
1855         (KJS::JSCallbackObject::staticValueGetter):
1856         (KJS::JSCallbackObject::callbackGetter):
1857         * API/JSContextRef.cpp:
1858         (JSCheckSyntax):
1859         * API/JSContextRef.h:
1860         * API/JSNode.c:
1861         (JSNodePrototype_appendChild):
1862         (JSNodePrototype_removeChild):
1863         (JSNodePrototype_replaceChild):
1864         (JSNode_getNodeType):
1865         (JSNode_getChildNodes):
1866         (JSNode_getFirstChild):
1867         (JSNode_construct):
1868         * API/JSNode.h:
1869         * API/JSNodeList.c:
1870         (JSNodeListPrototype_item):
1871         (JSNodeList_length):
1872         (JSNodeList_getProperty):
1873         * API/JSObjectRef.h:
1874         * API/minidom.c:
1875         (print):
1876         * API/testapi.c:
1877         (MyObject_initialize):
1878         (MyObject_hasProperty):
1879         (MyObject_getProperty):
1880         (MyObject_setProperty):
1881         (MyObject_deleteProperty):
1882         (MyObject_getPropertyList):
1883         (MyObject_callAsFunction):
1884         (MyObject_callAsConstructor):
1885         (MyObject_convertToType):
1886         (print_callAsFunction):
1887         (myConstructor_callAsConstructor):
1888         (main):
1889         * JavaScriptCore.exp:
1890         * kjs/ExecState.h:
1891         (KJS::ExecState::exceptionHandle):
1892
1893 2006-07-10  Geoffrey Garen  <ggaren@apple.com>
1894
1895         Reviewed by Darin.
1896
1897         Improved type safety by implementing opaque JSValue/JSObject typing through
1898         abuse of 'const', not void*. Also fixed an alarming number of bugs
1899         exposed by this new type safety.
1900
1901         I made one design change in JavaScriptCore, which is that the JSObject
1902         constructor should take a JSValue* as its prototype argument, not a JSObject*,
1903         since we allow the prototype to be any JSValue*, including jsNull(), for
1904         example.
1905
1906         * API/APICast.h:
1907         (toJS):
1908         * API/JSBase.h:
1909         * API/JSCallbackConstructor.cpp:
1910         (KJS::JSCallbackConstructor::construct):
1911         * API/JSCallbackFunction.cpp:
1912         (KJS::JSCallbackFunction::callAsFunction):
1913         * API/JSCallbackObject.cpp:
1914         (KJS::JSCallbackObject::JSCallbackObject):
1915         (KJS::JSCallbackObject::getOwnPropertySlot):
1916         (KJS::JSCallbackObject::put):
1917         (KJS::JSCallbackObject::construct):
1918         (KJS::JSCallbackObject::callAsFunction):
1919         (KJS::JSCallbackObject::staticFunctionGetter):
1920         * API/JSCallbackObject.h:
1921         * API/JSContextRef.cpp:
1922         (JSEvaluate):
1923         * API/JSNode.c:
1924         (JSNodePrototype_appendChild):
1925         (JSNodePrototype_removeChild):
1926         (JSNodePrototype_replaceChild):
1927         * API/JSObjectRef.cpp:
1928         (JSObjectMake):
1929         (JSFunctionMakeWithBody):
1930         (JSObjectGetProperty):
1931         (JSObjectCallAsFunction):
1932         (JSObjectCallAsConstructor):
1933         * API/JSObjectRef.h:
1934         * API/testapi.c:
1935         (main):
1936         * ChangeLog:
1937         * kjs/object.h:
1938         (KJS::JSObject::JSObject):
1939
1940 2006-07-10  Geoffrey Garen  <ggaren@apple.com>
1941
1942         Approved by Maciej, Darin.
1943
1944         Renamed JSStringBufferRef to JSInternalStringRef. "Internal string" means the
1945         JavaScript engine's internal string representation, which is the most
1946         low-level and efficient representation to use when interfacing with JavaScript.
1947
1948         * API/APICast.h:
1949         (toJS):
1950         (toRef):
1951         * API/JSBase.h:
1952         * API/JSCallbackObject.cpp:
1953         (KJS::JSCallbackObject::getOwnPropertySlot):
1954         (KJS::JSCallbackObject::put):
1955         (KJS::JSCallbackObject::deleteProperty):
1956         (KJS::JSCallbackObject::staticValueGetter):
1957         (KJS::JSCallbackObject::callbackGetter):
1958         * API/JSContextRef.cpp:
1959         (JSEvaluate):
1960         (JSCheckSyntax):
1961         * API/JSContextRef.h:
1962         * API/JSInternalStringRef.cpp: Added.
1963         (JSStringMake):
1964         (JSInternalStringCreate):
1965         (JSInternalStringCreateUTF8):
1966         (JSInternalStringRetain):
1967         (JSInternalStringRelease):
1968         (JSValueCopyStringValue):
1969         (JSInternalStringGetLength):
1970         (JSInternalStringGetCharactersPtr):
1971         (JSInternalStringGetCharacters):
1972         (JSInternalStringGetMaxLengthUTF8):
1973         (JSInternalStringGetCharactersUTF8):
1974         (JSInternalStringIsEqual):
1975         (JSInternalStringIsEqualUTF8):
1976         (JSInternalStringCreateCF):
1977         (CFStringCreateWithJSInternalString):
1978         * API/JSInternalStringRef.h: Added.
1979         * API/JSNode.c:
1980         (JSNodePrototype_appendChild):
1981         (JSNode_getNodeType):
1982         (JSNode_getChildNodes):
1983         (JSNode_getFirstChild):
1984         * API/JSNodeList.c:
1985         (JSNodeList_length):
1986         (JSNodeList_getProperty):
1987         * API/JSObjectRef.cpp:
1988         (JSFunctionMakeWithBody):
1989         (JSObjectGetDescription):
1990         (JSObjectHasProperty):
1991         (JSObjectGetProperty):
1992         (JSObjectSetProperty):
1993         (JSObjectDeleteProperty):
1994         (JSPropertyEnumeratorGetNext):
1995         (JSPropertyListAdd):
1996         * API/JSObjectRef.h:
1997         * API/JSStringBufferRef.cpp: Removed.
1998         * API/JSStringBufferRef.h: Removed.
1999         * API/JSValueRef.h:
2000         * API/JavaScriptCore.h:
2001         * API/minidom.c:
2002         (main):
2003         (print):
2004         * API/testapi.c:
2005         (assertEqualsAsUTF8String):
2006         (assertEqualsAsCharactersPtr):
2007         (assertEqualsAsCharacters):
2008         (MyObject_hasProperty):
2009         (MyObject_getProperty):
2010         (MyObject_setProperty):
2011         (MyObject_deleteProperty):
2012         (MyObject_getPropertyList):
2013         (print_callAsFunction):
2014         (myConstructor_callAsConstructor):
2015         (main):
2016         * JavaScriptCore.exp:
2017         * JavaScriptCore.xcodeproj/project.pbxproj:
2018
2019 2006-07-08  Tim Omernick  <timo@apple.com>
2020
2021         Reviewed by Maciej.
2022
2023         Added an OpenGL drawing model to the Netscape Plug-in API.
2024
2025         * bindings/npapi.h:
2026
2027 2006-07-08  Timothy Hatcher  <timothy@apple.com>
2028
2029         Reviewed by Maciej.
2030
2031         Moved KJS_GetCreatedJavaVMs to jni_utility.cpp.
2032         Switched KJS_GetCreatedJavaVMs over to use dlopen and dlsym
2033         now that NSAddImage, NSLookupSymbolInImage and NSAddressOfSymbol
2034         are deprecated in Leopard.
2035
2036         * JavaScriptCore.exp:
2037         * JavaScriptCore.xcodeproj/project.pbxproj:
2038         * bindings/jni/jni_utility.cpp:
2039         (KJS::Bindings::KJS_GetCreatedJavaVMs):
2040         * bindings/softlinking.c: Removed.
2041         * bindings/softlinking.h: Removed.
2042
2043 2006-07-08  Geoffrey Garen  <ggaren@apple.com>
2044
2045         Reviewed by Anders.
2046
2047         - Make JSObjectGetProperty return a JSValue or NULL, like JSEvaluate does.
2048
2049         * API/JSObjectRef.cpp:
2050         (JSObjectGetProperty):
2051         * API/JSObjectRef.h:
2052         * API/testapi.c:
2053         (main):
2054
2055 2006-07-08  Geoffrey Garen  <ggaren@apple.com>
2056
2057         Style change -- no review necessary.
2058
2059         Use 0 instead of NULL in API .cpp files, to match our style guidelines.
2060
2061         * API/JSContextRef.cpp:
2062         (JSEvaluate):
2063         * API/JSObjectRef.cpp:
2064         (JSFunctionMakeWithBody):
2065         (JSObjectCallAsFunction):
2066         (JSObjectCallAsConstructor):
2067         * API/JSValueRef.cpp:
2068         (JSValueToObject):
2069
2070 2006-07-08  Geoffrey Garen  <ggaren@apple.com>
2071
2072         Reviewed by TimO.
2073
2074         - Added ability to pass NULL for thisObject when calling JSObjectCallAsFunction,
2075         to match JSEvaluate.
2076
2077         * API/JSObjectRef.cpp:
2078         (JSObjectCallAsFunction):
2079         * API/JSObjectRef.h:
2080         * API/testapi.c:
2081         (main):
2082
2083 === Safari-521.15 ===
2084
2085 2006-07-07  Geoffrey Garen  <ggaren@apple.com>
2086
2087         Reviewed by Maciej.
2088
2089         - Standardized which functions take a JSContext as an argument. The rule is:
2090         if you might execute JavaScript, you take a JSContext, otherwise you don't.
2091
2092         The FIXME in JSObjectRef.h requires refactoring some parts of Interpreter,
2093         but not API changes, so I'm putting it off until later.
2094
2095         * API/JSCallbackObject.cpp:
2096         (KJS::JSCallbackObject::JSCallbackObject):
2097         (KJS::JSCallbackObject::init):
2098         * API/JSCallbackObject.h:
2099         * API/JSContextRef.cpp:
2100         (JSContextCreate):
2101         * API/JSContextRef.h:
2102         * API/JSObjectRef.cpp:
2103         (JSObjectMake):
2104         (JSPropertyEnumeratorGetNext):
2105         * API/JSObjectRef.h:
2106         * API/testapi.c:
2107         (MyObject_initialize):
2108         (main):
2109         * JavaScriptCore.exp:
2110         * kjs/array_object.cpp:
2111         (ArrayInstance::setLength):
2112         (ArrayInstance::pushUndefinedObjectsToEnd):
2113         * kjs/nodes.cpp:
2114         (ForInNode::execute):
2115         * kjs/reference.cpp:
2116         (KJS::Reference::getPropertyName):
2117         (KJS::Reference::getValue):
2118         * kjs/reference.h:
2119         * kjs/scope_chain.cpp:
2120         (KJS::ScopeChain::print):
2121
2122 2006-07-06  Geoffrey Garen  <ggaren@apple.com>
2123
2124         Reviewed by Maciej.
2125
2126         More API action.
2127
2128         - Headerdoc finished
2129
2130         Semantic Changes:
2131         - Added a JSContextRef argument to many functions, because you need a
2132         JSContextRef for doing virtually anything. I expect to add this argument
2133         to even more functions in a future patch.
2134
2135         - Removed the globalObjectPrototype argument to JSContextCreate because
2136         you can't create an object until you have a context, so it's impossible
2137         to pass a prototype object to JSContextCreate. That's OK because (1) there's
2138         no reason to give the global object a prototype and (2) if you really want
2139         to, you can just use a separate call to JSObjectSetPrototype.
2140
2141         - Removed the JSClassRef argument to JSClassCreate because it was unnecessary,
2142         and you need to be able to make the global object's class before you've
2143         created a JSContext.
2144
2145         - Added an optional exception parameter to JSFunctionMakeWithBody because anything
2146         less would be uncivilized.
2147
2148         - Made the return value parameter to JSObjectGetProperty optional to match
2149         all other return value parameters in the API.
2150
2151         - Made JSObjectSetPrivate/JSObjectGetPrivate work on JSCallbackFunctions
2152         and JSCallbackConstructors. You could use an abstract base class or strategic
2153         placement of m_privateData in the class structure to implement this, but
2154         the former seemed like overkill, and the latter seemed too dangerous.
2155
2156         - Fixed a bug where JSPropertyEnumeratorGetNext would skip the first property.
2157
2158         Cosmetic Changes:
2159         - Reversed the logic of the JSChar #ifdef to avoid confusing headerdoc
2160
2161         - Removed function names from @function declarations because headeroc
2162         can parse them automatically, and I wanted to rule out manual mismatch.
2163
2164         - Changed Error::create to take a const UString& instead of a UString*
2165         because it was looking at me funny.
2166
2167         - Renamed JSStringBufferCreateWithCFString to JSStringBufferCreateCF
2168         because the latter is more concise and it matches JSStringBufferCreateUTF8.
2169
2170         * API/JSCallbackObject.cpp:
2171         (KJS::JSCallbackObject::getOwnPropertySlot):
2172         (KJS::JSCallbackObject::put):
2173         (KJS::JSCallbackObject::deleteProperty):
2174         (KJS::JSCallbackObject::getPropertyList):
2175         (KJS::JSCallbackObject::toBoolean):
2176         (KJS::JSCallbackObject::toNumber):
2177         (KJS::JSCallbackObject::toString):
2178         * API/JSClassRef.cpp:
2179         (JSClassCreate):
2180         * API/JSContextRef.cpp:
2181         (JSContextCreate):
2182         (JSContextSetException):
2183         * API/JSContextRef.h:
2184         * API/JSNode.c:
2185         (JSNodePrototype_class):
2186         (JSNode_class):
2187         * API/JSNodeList.c:
2188         (JSNodeListPrototype_class):
2189         (JSNodeList_class):
2190         * API/JSObjectRef.cpp:
2191         (JSObjectGetProperty):
2192         (JSObjectGetPrivate):
2193         (JSObjectSetPrivate):
2194         (JSObjectCallAsFunction):
2195         (JSObjectCallAsConstructor):
2196         (JSPropertyEnumeratorGetNext):
2197         * API/JSObjectRef.h:
2198         * API/JSStringBufferRef.cpp:
2199         (JSStringBufferCreateCF):
2200         * API/JSStringBufferRef.h:
2201         * API/JSValueRef.cpp:
2202         (JSValueIsInstanceOf):
2203         * API/JSValueRef.h:
2204         * API/minidom.c:
2205         (main):
2206         * API/minidom.js:
2207         * API/testapi.c:
2208         (MyObject_hasProperty):
2209         (MyObject_setProperty):
2210         (MyObject_deleteProperty):
2211         (MyObject_getPropertyList):
2212         (MyObject_convertToType):
2213         (MyObject_class):
2214         (main):
2215         * JavaScriptCore.exp:
2216
2217 2006-07-07  Geoffrey Garen  <ggaren@apple.com>
2218
2219         Reviewed by John.
2220
2221         - Fixed a few crashes resulting from NULL parameters to JSClassCreate.
2222
2223         * API/JSClassRef.cpp:
2224         (JSClassCreate):
2225         (JSClassRelease):
2226         * API/testapi.c: Added test for NULL parameters.
2227         (main):
2228
2229 2006-07-07  Geoffrey Garen  <ggaren@apple.com>
2230
2231         Reviewed by John, mocked by Darin.
2232
2233         - Changed JSEvaluate to take a JSObjectRef instead of a JSValueRef as
2234         "this," since "this" must be an object.
2235
2236         * API/JSContextRef.cpp:
2237         (JSEvaluate):
2238         * API/JSContextRef.h:
2239
2240 2006-07-07  Geoffrey Garen  <ggaren@apple.com>
2241
2242         Reviewed by John.
2243
2244         - More headerdoc
2245
2246         * API/JSBase.h:
2247         * JavaScriptCore.xcodeproj/project.pbxproj:
2248
2249 2006-07-05  Geoffrey Garen  <ggaren@apple.com>
2250
2251         RS by Beth.
2252
2253         Renamed JSCharBufferRef, which was universally unpopular, to JSStringBufferRef,
2254         which, hopefully, will be less unpopular.
2255
2256         * API/APICast.h:
2257         (toJS):
2258         (toRef):
2259         * API/JSBase.h:
2260         * API/JSCallbackObject.cpp:
2261         (KJS::JSCallbackObject::getOwnPropertySlot):
2262         (KJS::JSCallbackObject::put):
2263         (KJS::JSCallbackObject::deleteProperty):
2264         (KJS::JSCallbackObject::staticValueGetter):
2265         (KJS::JSCallbackObject::callbackGetter):
2266         * API/JSCharBufferRef.cpp: Removed.
2267         * API/JSCharBufferRef.h: Removed.
2268         * API/JSContextRef.cpp:
2269         (JSEvaluate):
2270         (JSCheckSyntax):
2271         * API/JSContextRef.h:
2272         * API/JSNode.c:
2273         (JSNodePrototype_appendChild):
2274         (JSNode_getNodeType):
2275         (JSNode_getChildNodes):
2276         (JSNode_getFirstChild):
2277         * API/JSNodeList.c:
2278         (JSNodeList_length):
2279         (JSNodeList_getProperty):
2280         * API/JSObjectRef.cpp:
2281         (JSFunctionMakeWithBody):
2282         (JSObjectGetDescription):
2283         (JSObjectHasProperty):
2284         (JSObjectGetProperty):
2285         (JSObjectSetProperty):
2286         (JSObjectDeleteProperty):
2287         (JSPropertyEnumeratorGetNext):
2288         (JSPropertyListAdd):
2289         * API/JSObjectRef.h:
2290         * API/JSStringBufferRef.cpp: Added.
2291         (JSStringMake):
2292         (JSStringBufferCreate):
2293         (JSStringBufferCreateUTF8):
2294         (JSStringBufferRetain):
2295         (JSStringBufferRelease):
2296         (JSValueCopyStringValue):
2297         (JSStringBufferGetLength):
2298         (JSStringBufferGetCharactersPtr):
2299         (JSStringBufferGetCharacters):
2300         (JSStringBufferGetMaxLengthUTF8):
2301         (JSStringBufferGetCharactersUTF8):
2302         (JSStringBufferIsEqual):
2303         (JSStringBufferIsEqualUTF8):
2304         (JSStringBufferCreateWithCFString):
2305         (CFStringCreateWithJSStringBuffer):
2306         * API/JSStringBufferRef.h: Added.
2307         * API/JSValueRef.h:
2308         * API/JavaScriptCore.h:
2309         * API/minidom.c:
2310         (main):
2311         (print):
2312         * API/testapi.c:
2313         (assertEqualsAsUTF8String):
2314         (assertEqualsAsCharactersPtr):
2315         (assertEqualsAsCharacters):
2316         (MyObject_hasProperty):
2317         (MyObject_getProperty):
2318         (MyObject_setProperty):
2319         (MyObject_deleteProperty):
2320         (MyObject_getPropertyList):
2321         (print_callAsFunction):
2322         (myConstructor_callAsConstructor):
2323         (main):
2324         * JavaScriptCore.exp:
2325         * JavaScriptCore.xcodeproj/project.pbxproj:
2326
2327 2006-07-05  Geoffrey Garen  <ggaren@apple.com>
2328
2329         RS by Beth.
2330
2331         Moved some code around for more logical file separation.
2332
2333         * API/JSBase.h:
2334         * API/JSContextRef.h:
2335         * API/JSObjectRef.cpp:
2336         * API/JSValueRef.cpp:
2337         (JSValueToObject):
2338         * API/JSValueRef.h:
2339
2340 2006-07-03  Geoffrey Garen  <ggaren@apple.com>
2341
2342         Reviewed by Maciej.
2343
2344         Implemented JSFunctionMakeWithBody, which parses a script as a function body
2345         in the global scope, and returns the resulting anonymous function.
2346
2347         I also removed private data from JSCallbackFunction. It never worked,
2348         since JSCallbackFunction doesn't inherit from JSCallbackObject.
2349
2350         * API/JSCallbackConstructor.cpp: Removed.
2351         * API/JSCallbackConstructor.h: Removed.
2352         * API/JSCallbackFunction.cpp:
2353         (KJS::JSCallbackFunction::JSCallbackFunction):
2354         (KJS::JSCallbackFunction::implementsConstruct):
2355         (KJS::JSCallbackFunction::construct):
2356         (KJS::JSCallbackFunction::implementsCall):
2357         (KJS::JSCallbackFunction::callAsFunction):
2358         * API/JSCallbackFunction.h:
2359         * API/JSCallbackObject.cpp:
2360         (KJS::JSCallbackObject::staticFunctionGetter):
2361         * API/JSObjectRef.cpp:
2362         (JSFunctionMake):
2363         (JSFunctionMakeWithCallbacks):
2364         * API/JSObjectRef.h:
2365         * API/JSValueRef.h:
2366         * API/minidom.c:
2367         (main):
2368         * API/testapi.c:
2369         (main):
2370         * JavaScriptCore.exp: Programmatically added all symbols exported by
2371         API object files, and sorted results
2372         * JavaScriptCore.xcodeproj/project.pbxproj:
2373
2374 2006-07-03  Geoffrey Garen  <ggaren@apple.com>
2375
2376         Reviewed by Maciej.
2377
2378         - Return syntax error in JSCheckSyntax through a JSValueRef* exception
2379         argument
2380
2381         * API/JSBase.h:
2382         * API/JSContextRef.cpp:
2383         (JSCheckSyntax):
2384         * API/testapi.c:
2385         (main):
2386         * JavaScriptCore.exp:
2387         * kjs/interpreter.cpp:
2388         (KJS::Interpreter::checkSyntax):
2389         * kjs/interpreter.h:
2390
2391 2006-07-04  Darin Adler  <darin@apple.com>
2392
2393         - fixed build
2394
2395         * wtf/MathExtras.h: Oops. Added missing #endif.
2396
2397 2006-07-04  Bjoern Graf  <bjoern.graf@gmail.com>
2398
2399         Reviewed by Maciej.
2400         Tweaked a bit by Darin.
2401
2402         - http://bugzilla.opendarwin.org/show_bug.cgi?id=9678
2403           work around MSVCRT's fmod function returning NaN for fmod(x, infinity) instead of x
2404
2405         * wtf/MathExtras.h: Added include of <float.h>.
2406         (isinf): Fix to return false for NAN.
2407         (wtf_fmod): Added. An inline that works around the bug.
2408
2409         * kjs/nodes.cpp:
2410         * kjs/number_object.cpp:
2411         * kjs/operations.cpp:
2412         * kjs/value.cpp:
2413         Added includes of MathExtras.h to all files using fmod.
2414
2415         * JavaScriptCore.xcodeproj/project.pbxproj: Let Xcode 2.3 have its way with
2416         the project.
2417
2418 2006-07-01  Geoffrey Garen  <ggaren@apple.com>
2419
2420         Reviewed by Darin.
2421
2422         - Refined value conversions in the API:
2423             - failed toNumber returns NaN
2424             - failed toObject returns NULL
2425             - failed toString returns empty string
2426
2427         - Refined excpetion handling in the API:
2428             - failed value conversions do not throw exceptions
2429             - uncaught exceptions in JSEvaluate, JSObjectCallAsFunction, and
2430               JSObjectCallAsConstructor are returned through a JSValueRef* exception
2431               argument
2432             - removed JSContextHasException, because JSContextGetException does
2433               the same job
2434
2435         * API/JSBase.h:
2436         * API/JSCharBufferRef.cpp:
2437         (JSValueCopyStringValue):
2438         * API/JSContextRef.cpp:
2439         (JSEvaluate):
2440         * API/JSContextRef.h:
2441         * API/JSNodeList.c: Added test code demonstrating how you would use
2442         toNumber, and why you probably don't need toUInt32, etc.
2443         (JSNodeListPrototype_item):
2444         (JSNodeList_getProperty):
2445         * API/JSObjectRef.cpp:
2446         (JSValueToObject):
2447         (JSObjectCallAsFunction):
2448         (JSObjectCallAsConstructor):
2449         * API/JSObjectRef.h:
2450         * API/JSValueRef.cpp:
2451         (JSValueToNumber):
2452         * API/JSValueRef.h:
2453         * API/minidom.c:
2454         (main):
2455         * API/testapi.c:
2456         (main): Added tests for new rules, and call to JSGCProtect to fix Intel
2457         crash
2458         * JavaScriptCore.exp:
2459
2460 2006-07-03  Darin Adler  <darin@apple.com>
2461
2462         - Rolled out HashMap implementation of NPRuntime, at least temporarily.
2463
2464         Fixes hang in the bindings section of layout tests seen on the
2465         buildbot.
2466
2467         This code was using HashMap<const char*, PrivateIdentifier*>.
2468         But that hashes based on pointer identity, not string value.
2469         The default hash for any pointer type is to hash based on the pointer.
2470         And WTF doesn't currently have a string hash for char*.
2471         We'll need to fix that before re-landing this patch.
2472
2473         (Formatting was also incorrect -- extra spaces in parentheses.)
2474
2475         * bindings/npruntime.cpp: Rolled out last change.
2476
2477 2006-07-02  Justin Haygood  <jhaygood@spsu.edu>
2478
2479         Reviewed, tweaked, landed by ggaren.
2480
2481         - Port NPRuntime from CFDictionary to HashMap.
2482
2483         * bindings/npruntime.cpp:
2484         (getStringIdentifierDictionary):
2485         (getIntIdentifierDictionary):
2486         (_NPN_GetStringIdentifier):
2487         (_NPN_GetIntIdentifier):
2488         * bindings/npruntime.h:
2489
2490 2006-07-01  Geoffrey Garen  <ggaren@apple.com>
2491
2492         Reviewed by Adele.
2493
2494         - Fixed <rdar://problem/4611197> REGRESSION: Liveconnect with Java test
2495         fails at http://www-sor.inria.fr/~dedieu/notes/liveconnect/simple_example.html
2496
2497         * JavaScriptCore.exp: Export symbols used by liveconnect
2498
2499 2006-06-29  Geoffrey Garen  <ggaren@apple.com>
2500
2501         Reviewed by Maciej.
2502
2503         - Phase 2 in the JS API.
2504
2505         - Added support for specifying static tables of values -- this should
2506         obviate the need for using complicated callbacks for most lookups.
2507
2508         -  API objects are now created with classes (JSClassRef) -- in order to support
2509         static values, and in order to prevent API objects from storing their
2510         data inline, and thus falling into the oversized (read: slow and prone to
2511         giving Maciej the frowny face) heap.
2512
2513         - Added two specialized JSObject subclasses -- JSCallbackFunction and JSCallbackConstructor --
2514         to allow JSFunctionMake and JSConstructorMake to continue to work with
2515         the new class model. Another solution to this problem would be to create
2516         a custom class object for each function and constructor you make. This
2517         solution is more code but also more efficient.
2518
2519         - Substantially beefed up the minidom example to demonstrate and test a
2520         lot of these techniques. Its output is still pretty haphazard, though.
2521
2522         - Gave the <kjs/ preface to some includes -- I'm told this matters to
2523         building on some versions of Linux.
2524
2525         - Implemented JSValueIsInstanceOf and JSValueIsObjectOfClass
2526
2527         - Removed GetDescription callback. Something in the class datastructure
2528         should take care of this.
2529
2530         * API/JSBase.h:
2531         * API/JSCallbackConstructor.cpp: Added.
2532         (KJS::):
2533         (KJS::JSCallbackConstructor::JSCallbackConstructor):
2534         (KJS::JSCallbackConstructor::implementsConstruct):
2535         (KJS::JSCallbackConstructor::construct):
2536         (KJS::JSCallbackConstructor::setPrivate):
2537         (KJS::JSCallbackConstructor::getPrivate):
2538         * API/JSCallbackConstructor.h: Added.
2539         (KJS::JSCallbackConstructor::classInfo):
2540         * API/JSCallbackFunction.cpp: Added.
2541         (KJS::):
2542         (KJS::JSCallbackFunction::JSCallbackFunction):
2543         (KJS::JSCallbackFunction::implementsCall):
2544         (KJS::JSCallbackFunction::callAsFunction):
2545         (KJS::JSCallbackFunction::setPrivate):
2546         (KJS::JSCallbackFunction::getPrivate):
2547         * API/JSCallbackFunction.h: Added.
2548         (KJS::JSCallbackFunction::classInfo):
2549         * API/JSCallbackObject.cpp:
2550         (KJS::):
2551         (KJS::JSCallbackObject::JSCallbackObject):
2552         (KJS::JSCallbackObject::init):
2553         (KJS::JSCallbackObject::~JSCallbackObject):
2554         (KJS::JSCallbackObject::className):
2555         (KJS::JSCallbackObject::getOwnPropertySlot):
2556         (KJS::JSCallbackObject::put):
2557         (KJS::JSCallbackObject::deleteProperty):
2558         (KJS::JSCallbackObject::implementsConstruct):
2559         (KJS::JSCallbackObject::construct):
2560         (KJS::JSCallbackObject::implementsCall):
2561         (KJS::JSCallbackObject::callAsFunction):
2562         (KJS::JSCallbackObject::getPropertyList):
2563         (KJS::JSCallbackObject::toBoolean):
2564         (KJS::JSCallbackObject::toNumber):
2565         (KJS::JSCallbackObject::toString):
2566         (KJS::JSCallbackObject::inherits):
2567         (KJS::JSCallbackObject::staticValueGetter):
2568         (KJS::JSCallbackObject::staticFunctionGetter):
2569         (KJS::JSCallbackObject::callbackGetter):
2570         * API/JSCallbackObject.h:
2571         * API/JSCharBufferRef.cpp:
2572         * API/JSClassRef.cpp: Added.
2573         (JSClassCreate):
2574         (JSClassRetain):
2575         (JSClassRelease):
2576         * API/JSClassRef.h: Added.
2577         (StaticValueEntry::StaticValueEntry):
2578         (StaticFunctionEntry::StaticFunctionEntry):
2579         (__JSClass::__JSClass):
2580         * API/JSContextRef.cpp:
2581         (JSContextCreate):
2582         (JSEvaluate):
2583         * API/JSContextRef.h:
2584         * API/JSNode.c: Added.
2585         (JSNodePrototype_appendChild):
2586         (JSNodePrototype_removeChild):
2587         (JSNodePrototype_replaceChild):
2588         (JSNodePrototype_class):
2589         (JSNode_getNodeType):
2590         (JSNode_getChildNodes):
2591         (JSNode_getFirstChild):
2592         (JSNode_finalize):
2593         (JSNode_class):
2594         (JSNode_prototype):
2595         (JSNode_new):
2596         (JSNode_construct):
2597         * API/JSNode.h: Added.
2598         * API/JSNodeList.c: Added.
2599         (JSNodeListPrototype_item):
2600         (JSNodeListPrototype_class):
2601         (JSNodeList_length):
2602         (JSNodeList_getProperty):
2603         (JSNodeList_finalize):
2604         (JSNodeList_class):
2605         (JSNodeList_prototype):
2606         (JSNodeList_new):
2607         * API/JSNodeList.h: Added.
2608         * API/JSObjectRef.cpp:
2609         (JSObjectMake):
2610         (JSFunctionMake):
2611         (JSConstructorMake):
2612         (__JSPropertyEnumerator::__JSPropertyEnumerator):
2613         (JSObjectCreatePropertyEnumerator):
2614         (JSPropertyEnumeratorGetNext):
2615         (JSPropertyEnumeratorRetain):
2616         (JSPropertyEnumeratorRelease):
2617         * API/JSObjectRef.h:
2618         (__JSObjectCallbacks::):
2619         * API/JSValueRef.cpp:
2620         (JSValueIsObjectOfClass):
2621         (JSValueIsInstanceOf):
2622         * API/JSValueRef.h:
2623         * API/Node.c: Added.
2624         (Node_new):
2625         (Node_appendChild):
2626         (Node_removeChild):
2627         (Node_replaceChild):
2628         (Node_ref):
2629         (Node_deref):
2630         * API/Node.h: Added.
2631         * API/NodeList.c: Added.
2632         (NodeList_new):
2633         (NodeList_length):
2634         (NodeList_item):
2635         (NodeList_ref):
2636         (NodeList_deref):
2637         * API/NodeList.h: Added.
2638         * API/minidom.c:
2639         (main):
2640         (print):
2641         (createStringWithContentsOfFile):
2642         * API/minidom.js:
2643         * API/testapi.c:
2644         (assertEqualsAsCharacters):
2645         (MyObject_getProperty):
2646         (MyObject_class):
2647         (myConstructor_callAsConstructor):
2648         (main):
2649         * API/testapi.js:
2650         * JavaScriptCore.xcodeproj/project.pbxproj:
2651
2652 2006-06-26  Kevin Ollivier  <kevino@theolliviers.com>
2653
2654         Reviewed and tweaked by Darin.
2655
2656         - Compile fixes for wx port / gcc 4.0.2
2657
2658         * kjs/array_object.cpp:
2659         Added missing headers.
2660
2661         * kjs/ExecState.h:
2662         gcc needs class prototypes before defining those classes as friend classes
2663
2664 2006-06-30  Mike Emmel  <mike.emmel@gmail.com>
2665
2666         Reviewed by Darin.
2667
2668         Compilation fixes for Linux/Gdk.
2669
2670         * JavaScriptCore/kjs/interpreter.cpp: added include of signal.h
2671         * JavaScriptCore/kjs/ExecState.h: added missing class declaration
2672         * JavaScriptCore/kjs/ExecState.cpp: case wrong on include of context.h
2673         * JavaScriptCore/JavaScriptCoreSources.bkl: added Context.cpp and ExecState.cpp
2674
2675 === Safari-521.14 ===
2676
2677 2006-06-29  Maciej Stachowiak  <mjs@apple.com>
2678
2679         Reviewed by Geoff.
2680
2681         - add headerdoc comments to some of the new JS API headers
2682
2683         * API/JSBase.h:
2684         * API/JSValueRef.h:
2685
2686 2006-06-28  Timothy Hatcher  <timothy@apple.com>
2687
2688         Prefer the Stabs debugging symbols format until DWARF bugs are fixed.
2689
2690         * JavaScriptCore.xcodeproj/project.pbxproj:
2691
2692 2006-06-27  Timothy Hatcher  <timothy@apple.com>
2693
2694         Reviewed by Tim O.
2695
2696         <rdar://problem/4448350> Deprecated ObjC language API used in JavaScriptCore, WebCore, WebKit and WebBrowser
2697
2698         Switch to the new ObjC 2 API, ifdefed the old code around OBJC_API_VERSION so it still works on Tiger.
2699         Removed the use of the old stringWithCString, switched to the new Tiger version that accepts an encoding.
2700         Lots of code style cleanup.
2701
2702         * JavaScriptCore.xcodeproj/project.pbxproj:
2703         * bindings/objc/objc_class.h:
2704         * bindings/objc/objc_class.mm:
2705         (KJS::Bindings::ObjcClass::~ObjcClass):
2706         (KJS::Bindings::_createClassesByIsAIfNecessary):
2707         (KJS::Bindings::ObjcClass::classForIsA):
2708         (KJS::Bindings::ObjcClass::name):
2709         (KJS::Bindings::ObjcClass::methodsNamed):
2710         (KJS::Bindings::ObjcClass::fieldNamed):
2711         (KJS::Bindings::ObjcClass::fallbackObject):
2712         * bindings/objc/objc_header.h:
2713         * bindings/objc/objc_instance.h:
2714         * bindings/objc/objc_instance.mm:
2715         (ObjcInstance::ObjcInstance):
2716         (ObjcInstance::~ObjcInstance):
2717         (ObjcInstance::operator=):
2718         (ObjcInstance::begin):
2719         (ObjcInstance::end):
2720         (ObjcInstance::getClass):
2721         (ObjcInstance::invokeMethod):
2722         (ObjcInstance::invokeDefaultMethod):
2723         (ObjcInstance::setValueOfField):
2724         (ObjcInstance::supportsSetValueOfUndefinedField):
2725         (ObjcInstance::setValueOfUndefinedField):
2726         (ObjcInstance::getValueOfField):
2727         (ObjcInstance::getValueOfUndefinedField):
2728         (ObjcInstance::defaultValue):
2729         (ObjcInstance::stringValue):
2730         (ObjcInstance::numberValue):
2731         (ObjcInstance::booleanValue):
2732         (ObjcInstance::valueOf):
2733         * bindings/objc/objc_runtime.h:
2734         * bindings/objc/objc_runtime.mm:
2735         (ObjcMethod::ObjcMethod):
2736         (ObjcMethod::name):
2737         (ObjcMethod::getMethodSignature):
2738         (ObjcMethod::setJavaScriptName):
2739         (ObjcField::name):
2740         (ObjcField::type):
2741         (ObjcField::valueFromInstance):
2742         (convertValueToObjcObject):
2743         (ObjcField::setValueToInstance):
2744         (ObjcArray::operator=):
2745         (ObjcArray::setValueAt):
2746         (ObjcArray::valueAt):
2747         (ObjcFallbackObjectImp::ObjcFallbackObjectImp):
2748         (ObjcFallbackObjectImp::callAsFunction):
2749         (ObjcFallbackObjectImp::defaultValue):
2750
2751 2006-06-28  Anders Carlsson  <acarlsson@apple.com>
2752
2753         Reviewed by Geoff.
2754
2755         http://bugzilla.opendarwin.org/show_bug.cgi?id=8636
2756         REGRESSION: JavaScript access to Java applet causes hang (_webViewURL not implemented)
2757
2758         * bindings/jni/jni_objc.mm:
2759         (KJS::Bindings::dispatchJNICall):
2760         Just pass nil as the calling URL. This will cause the Java plugin to use the URL of the page
2761         containing the applet (which is what we used to do).
2762
2763 2006-06-27  Timothy Hatcher  <timothy@apple.com>
2764
2765         Reviewed by Darin.
2766
2767         <rdar://problem/4406785> Add an export file to TOT JavaScriptCore like the Safari-2-0-branch
2768
2769         * JavaScriptCore.exp: Added.
2770         * JavaScriptCore.xcodeproj/project.pbxproj:
2771
2772 2006-06-25  Geoffrey Garen  <ggaren@apple.com>
2773
2774         Reviewed by Adele.
2775
2776         - Added JSConstructorMake to match JSFunctionMake, along with test code.
2777
2778         [ I checked in the ChangeLog before without the actual files. ]
2779
2780         * API/JSObjectRef.cpp:
2781         (JSConstructorMake):
2782         * API/JSObjectRef.h:
2783         * API/testapi.c:
2784         (myConstructor_callAsConstructor):
2785         (main):
2786         * API/testapi.js:
2787         * ChangeLog:
2788         * JavaScriptCore.xcodeproj/project.pbxproj: Moved testapi.c to the testapi
2789         target -- this was an oversight in my earlier check-in.
2790
2791 2006-06-25  Timothy Hatcher  <timothy@apple.com>
2792
2793         Reviewed by Darin.
2794
2795         Bug 9574: Drosera should show inline scripts within the original HTML
2796         http://bugzilla.opendarwin.org/show_bug.cgi?id=9574
2797
2798         Pass the starting line number and error message to the debugger.
2799
2800         * kjs/debugger.cpp:
2801         (Debugger::sourceParsed):
2802         * kjs/debugger.h:
2803         * kjs/function.cpp:
2804         (KJS::GlobalFuncImp::callAsFunction):
2805         * kjs/function_object.cpp:
2806         (FunctionObjectImp::construct):
2807         * kjs/interpreter.cpp:
2808         (KJS::Interpreter::evaluate):
2809
2810 2006-06-24  Alexey Proskuryakov  <ap@nypop.com>
2811
2812         Rubber-stamped by Eric.
2813
2814         Add a -h (do not follow symlinks) option to ln in derived sources build script (without it,
2815         a symlink was created inside the source directory on second build).
2816
2817         * JavaScriptCore.xcodeproj/project.pbxproj:
2818
2819 2006-06-24  David Kilzer  <ddkilzer@kilzer.net>
2820
2821         Reviewed by Timothy.
2822
2823         * Info.plist: Fixed copyright to include 2003-2006.
2824
2825 2006-06-24  Alexey Proskuryakov  <ap@nypop.com>
2826
2827         Reviewed by Darin.
2828
2829         - http://bugzilla.opendarwin.org/show_bug.cgi?id=9418
2830         WebKit will not build when Space exists in path
2831
2832         * JavaScriptCore.xcodeproj/project.pbxproj: Enclose search paths in quotes; create symlinks to
2833         avoid passing paths with spaces to make.
2834
2835 2006-06-23  Timothy Hatcher  <timothy@apple.com>
2836
2837         Reviewed by Darin.
2838
2839         Adding more operator[] overloads for long and short types.
2840
2841         * wtf/Vector.h:
2842         (WTF::Vector::operator[]):
2843
2844 === JavaScriptCore-521.13 ===
2845
2846 2006-06-22  Alexey Proskuryakov  <ap@nypop.com>
2847
2848         Build fix.
2849
2850         - http://bugzilla.opendarwin.org/show_bug.cgi?id=9539
2851         Another case error preventing build
2852
2853         * API/JSObjectRef.cpp: Changed "identifier.h" to "Identifier.h"
2854
2855 2006-06-22  David Kilzer  <ddkilzer@kilzer.net>
2856
2857         Build fix.
2858
2859         http://bugzilla.opendarwin.org/show_bug.cgi?id=9539
2860         Another case error preventing build
2861
2862         * API/APICast.h: Changed "UString.h" to "ustring.h".
2863
2864 2006-06-21  Geoffrey Garen  <ggaren@apple.com>
2865
2866         Fixed release build, fixed accidental infinite recursion due to
2867         last minute global replace gone awry.
2868
2869         * API/APICast.h:
2870         (toRef):
2871         * API/testapi.c:
2872         (assertEqualsAsBoolean):
2873         (assertEqualsAsNumber):
2874         (assertEqualsAsUTF8String):
2875         (assertEqualsAsCharactersPtr):
2876         * JavaScriptCore.xcodeproj/project.pbxproj:
2877
2878 2006-06-21  Geoffrey Garen  <ggaren@apple.com>
2879
2880         Reviewed by Anders.
2881
2882         - First cut at C API to JavaScript. Includes a unit test, 'testapi.c',
2883         and the outline of a test app, 'minidom.c'.
2884
2885         Includes one change to JSC internals: Rename propList to getPropertyList and have it
2886         take its target property list by reference so that subclasses can
2887         add properties to the list before calling through to their superclasses.
2888
2889         Also, I just ran prepare-ChangeLog in about 10 seconds, and I would like
2890         to give a shout-out to that.
2891
2892         * API/APICast.h: Added.
2893         (toJS):
2894         (toRef):
2895         * API/JSBase.h: Added.
2896         * API/JSCallbackObject.cpp: Added.
2897         (KJS::):
2898         (KJS::JSCallbackObject::JSCallbackObject):
2899         (KJS::JSCallbackObject::~JSCallbackObject):
2900         (KJS::JSCallbackObject::className):
2901         (KJS::JSCallbackObject::getOwnPropertySlot):
2902         (KJS::JSCallbackObject::put):
2903         (KJS::JSCallbackObject::deleteProperty):
2904         (KJS::JSCallbackObject::implementsConstruct):
2905         (KJS::JSCallbackObject::construct):
2906         (KJS::JSCallbackObject::implementsCall):
2907         (KJS::JSCallbackObject::callAsFunction):
2908         (KJS::JSCallbackObject::getPropertyList):
2909         (KJS::JSCallbackObject::toBoolean):
2910         (KJS::JSCallbackObject::toNumber):
2911         (KJS::JSCallbackObject::toString):
2912         (KJS::JSCallbackObject::setPrivate):
2913         (KJS::JSCallbackObject::getPrivate):
2914         (KJS::JSCallbackObject::cachedValueGetter):
2915         (KJS::JSCallbackObject::callbackGetter):
2916         * API/JSCallbackObject.h: Added.
2917         (KJS::JSCallbackObject::classInfo):
2918         * API/JSCharBufferRef.cpp: Added.
2919         (JSStringMake):
2920         (JSCharBufferCreate):
2921         (JSCharBufferCreateUTF8):
2922         (JSCharBufferRetain):
2923         (JSCharBufferRelease):
2924         (JSValueCopyStringValue):
2925         (JSCharBufferGetLength):
2926         (JSCharBufferGetCharactersPtr):
2927         (JSCharBufferGetCharacters):
2928         (JSCharBufferGetMaxLengthUTF8):
2929         (JSCharBufferGetCharactersUTF8):
2930         (JSCharBufferIsEqual):
2931         (JSCharBufferIsEqualUTF8):
2932         (JSCharBufferCreateWithCFString):
2933         (CFStringCreateWithJSCharBuffer):
2934         * API/JSCharBufferRef.h: Added.
2935         * API/JSContextRef.cpp: Added.
2936         (JSContextCreate):
2937         (JSContextDestroy):
2938         (JSContextGetGlobalObject):
2939         (JSEvaluate):
2940         (JSCheckSyntax):
2941         (JSContextHasException):
2942         (JSContextGetException):
2943         (JSContextClearException):
2944         (JSContextSetException):
2945         * API/JSContextRef.h: Added.
2946         * API/JSObjectRef.cpp: Added.
2947         (JSValueToObject):
2948         (JSObjectMake):
2949         (JSFunctionMake):
2950         (JSObjectGetDescription):
2951         (JSObjectGetPrototype):
2952         (JSObjectSetPrototype):
2953         (JSObjectHasProperty):
2954         (JSObjectGetProperty):
2955         (JSObjectSetProperty):
2956         (JSObjectDeleteProperty):
2957         (JSObjectGetPrivate):
2958         (JSObjectSetPrivate):
2959         (JSObjectIsFunction):
2960         (JSObjectCallAsFunction):
2961         (JSObjectIsConstructor):
2962         (JSObjectCallAsConstructor):
2963         (__JSPropertyListEnumerator::__JSPropertyListEnumerator):
2964         (JSObjectCreatePropertyEnumerator):
2965         (JSPropertyEnumeratorGetNext):
2966         (JSPropertyEnumeratorRetain):
2967         (JSPropertyEnumeratorRelease):
2968         (JSPropertyListAdd):
2969         * API/JSObjectRef.h: Added.
2970         * API/JSValueRef.cpp: Added.
2971         (JSValueGetType):
2972         (JSValueIsUndefined):
2973         (JSValueIsNull):
2974         (JSValueIsBoolean):
2975         (JSValueIsNumber):
2976         (JSValueIsString):
2977         (JSValueIsObject):
2978         (JSValueIsEqual):
2979         (JSValueIsStrictEqual):
2980         (JSUndefinedMake):
2981         (JSNullMake):
2982         (JSBooleanMake):
2983         (JSNumberMake):
2984         (JSValueToBoolean):
2985         (JSValueToNumber):
2986         (JSGCProtect):
2987         (JSGCUnprotect):
2988         (JSGCCollect):
2989         * API/JSValueRef.h: Added.
2990         * API/JavaScriptCore.h: Added.
2991         * API/minidom.c: Added.
2992         (main):
2993         * API/minidom.html: Added.
2994         * API/minidom.js: Added.
2995         * API/testapi.c: Added.
2996         (assertEqualsAsBoolean):
2997         (assertEqualsAsNumber):
2998         (assertEqualsAsUTF8String):
2999         (assertEqualsAsCharactersPtr):
3000         (assertEqualsAsCharacters):
3001         (MyObject_initialize):
3002         (MyObject_copyDescription):
3003         (MyObject_hasProperty):
3004         (MyObject_getProperty):
3005         (MyObject_setProperty):
3006         (MyObject_deleteProperty):
3007         (MyObject_getPropertyList):
3008         (MyObject_callAsFunction):
3009         (MyObject_callAsConstructor):
3010         (MyObject_convertToType):
3011         (MyObject_finalize):
3012         (print_callAsFunction):
3013         (main):
3014         (createStringWithContentsOfFile):
3015         * API/testapi.js: Added.
3016         * ChangeLog:
3017         * JavaScriptCore.xcodeproj/project.pbxproj:
3018         * bindings/npruntime_impl.h:
3019         * kjs/array_instance.h:
3020         * kjs/array_object.cpp:
3021         (ArrayInstance::getPropertyList):
3022         * kjs/interpreter.cpp:
3023         (KJS::Interpreter::evaluate):
3024         * kjs/nodes.cpp:
3025         (ForInNode::execute):
3026         * kjs/object.cpp:
3027         (KJS::JSObject::put):
3028         (KJS::JSObject::canPut):
3029         (KJS::JSObject::deleteProperty):
3030         (KJS::JSObject::propertyIsEnumerable):
3031         (KJS::JSObject::getPropertyAttributes):
3032         (KJS::JSObject::getPropertyList):
3033         * kjs/object.h:
3034         * kjs/property_map.cpp:
3035         (KJS::PropertyMap::get):
3036         * kjs/property_map.h:
3037         * kjs/scope_chain.cpp:
3038         (KJS::ScopeChain::print):
3039         * kjs/string_object.cpp:
3040         (StringInstance::getPropertyList):
3041         * kjs/string_object.h:
3042         * kjs/ustring.h:
3043         (KJS::UString::Rep::ref):
3044
3045 2006-06-20  Timothy Hatcher  <timothy@apple.com>
3046
3047         Reviewed by Geoff.
3048
3049         Make sure we clear the exception before returning so
3050         that future calls will not fail because of an earlier
3051         exception state. Assert on entry that the WebScriptObject
3052         is working with an ExecState that dose not have an exception.
3053         Document that evaluateWebScript and callWebScriptMethod return
3054         WebUndefined when an exception is thrown.
3055
3056         * bindings/objc/WebScriptObject.h:
3057         * bindings/objc/WebScriptObject.mm:
3058         (-[WebScriptObject callWebScriptMethod:withArguments:]):
3059         (-[WebScriptObject evaluateWebScript:]):
3060         (-[WebScriptObject setValue:forKey:]):
3061         (-[WebScriptObject valueForKey:]):
3062         (-[WebScriptObject removeWebScriptKey:]):
3063         (-[WebScriptObject webScriptValueAtIndex:]):
3064         (-[WebScriptObject setWebScriptValueAtIndex:value:]):
3065
3066 2006-06-19  Anders Carlsson  <acarlsson@apple.com>
3067
3068         Reviewed by John.
3069
3070         * kjs/interpreter.cpp:
3071         (KJS::TimeoutChecker::pauseTimeoutCheck):
3072         (KJS::TimeoutChecker::resumeTimeoutCheck):
3073         Fix argument order in setitimer calls.
3074
3075 2006-06-18  Anders Carlsson  <acarlsson@apple.com>
3076
3077         Reviewed by Geoff.
3078
3079         * kjs/interpreter.cpp:
3080         (KJS::TimeoutChecker::pauseTimeoutCheck):
3081         Do nothing if the timeout check hasn't been started.
3082
3083         (KJS::TimeoutChecker::resumeTimeoutCheck):
3084         Do nothing if the timeout check hasn't been started.
3085         Use the right signal handler when unblocking.
3086
3087         (KJS::Interpreter::handleTimeout):
3088         pause/resume the timeout check around the call to
3089         shouldInterruptScript().
3090
3091 2006-06-16  Ben Goodger <beng@google.com>
3092
3093         Reviewed by Maciej
3094
3095         http://bugzilla.opendarwin.org/show_bug.cgi?id=9491
3096         Windows build breaks in interpreter.cpp
3097
3098         * kjs/interpreter.cpp
3099         (KJS::TimeoutChecker::pauseTimeoutCheck):
3100         (KJS::TimeoutChecker::resumeTimeoutCheck):
3101         Make sure to only assert equality with s_executingInterpreter when it
3102         is being used (i.e. when HAVE(SYS_TIME_H) == true)
3103
3104 2006-06-17  David Kilzer  <ddkilzer@kilzer.net>
3105
3106         Reviewed by darin.
3107
3108         http://bugzilla.opendarwin.org/show_bug.cgi?id=9477
3109         REGRESSION: fast/dom/replaceChild.html crashes on WebKit ToT in debug build
3110
3111         * kjs/function.cpp:
3112         (KJS::FunctionImp::callAsFunction): Refetch the debugger after executing the function
3113         in case the WebFrame it was running in has since been destroyed.
3114
3115 2006-06-17  David Kilzer  <ddkilzer@kilzer.net>
3116
3117         Reviewed by ggaren.
3118
3119         http://bugzilla.opendarwin.org/show_bug.cgi?id=9476
3120         REGRESSION: Reproducible crash after closing window after viewing
3121         css2.1/t0803-c5501-imrgn-t-00-b-ag.html
3122
3123         * kjs/debugger.cpp:
3124         (Debugger::detach): Call setDebugger(0) for all interpreters removed from
3125         the 'attached to a debugger' list.
3126
3127 2006-06-17  Anders Carlsson  <acarlsson@apple.com>
3128
3129         Reviewed by Maciej and Geoff.
3130
3131         http://bugzilla.opendarwin.org/show_bug.cgi?id=7080
3132         Provide some way to stop a JavaScript infinite loop
3133
3134         * kjs/completion.h:
3135         (KJS::):
3136         Add Interrupted completion type.
3137
3138         * kjs/function.cpp:
3139         (KJS::FunctionImp::callAsFunction):
3140         (KJS::GlobalFuncImp::callAsFunction):
3141         Only set the exception on the new ExecState if the current one has had one.
3142
3143         * kjs/interpreter.cpp:
3144         (KJS::TimeoutChecker::startTimeoutCheck):
3145         (KJS::TimeoutChecker::stopTimeoutCheck):
3146         (KJS::TimeoutChecker::alarmHandler):
3147         (KJS::TimeoutChecker::pauseTimeoutCheck):
3148         (KJS::TimeoutChecker::resumeTimeoutCheck):
3149         New TimeoutChecker class which handles setting Interpreter::m_timedOut flag after a given
3150         period of time. This currently only works on Unix platforms where setitimer and signals are used.
3151
3152         (KJS::Interpreter::Interpreter):
3153         Initialize new member variables.
3154
3155         (KJS::Interpreter::~Interpreter):
3156         Destroy the timeout checker.
3157
3158         (KJS::Interpreter::startTimeoutCheck):
3159         (KJS::Interpreter::stopTimeoutCheck):
3160         (KJS::Interpreter::pauseTimeoutCheck):
3161         (KJS::Interpreter::resumeTimeoutCheck):
3162         Call the timeout checker.
3163
3164         (KJS::Interpreter::handleTimeout):
3165         Called on timeout. Resets the m_timedOut flag and calls shouldInterruptScript.
3166
3167         * kjs/interpreter.h:
3168         (KJS::Interpreter::setTimeoutTime):
3169         New function for setting the timeout time.
3170
3171         (KJS::Interpreter::shouldInterruptScript):
3172         New function. The idea is that this should be overridden by subclasses in order to for example
3173         pop up a dialog asking the user if the script should be interrupted.
3174
3175         (KJS::Interpreter::checkTimeout):
3176         New function which checks the m_timedOut flag and calls handleTimeout if it's set.
3177
3178         * kjs/nodes.cpp:
3179         (DoWhileNode::execute):
3180         (WhileNode::execute):
3181         (ForNode::execute):
3182         Call Interpreter::checkTimeout after each iteration of the loop.
3183
3184 2006-06-15  Timothy Hatcher  <timothy@apple.com>
3185
3186         Reviewed by Geoff and Darin.
3187
3188         Prefer the DWARF debugging symbols format for use in Xcode 2.3.
3189
3190         * JavaScriptCore.xcodeproj/project.pbxproj:
3191
3192 2006-06-14  Geoffrey Garen  <ggaren@apple.com>
3193
3194         Reviewed by Beth.
3195
3196         - fixed http://bugzilla.opendarwin.org/show_bug.cgi?id=9438
3197         Someone broke ToT: cannot build
3198
3199         * JavaScriptCore.xcodeproj/project.pbxproj:
3200         * bindings/runtime_root.h: Changed "Interpreter.h" to "interpreter.h"
3201
3202 2006-06-12  Geoffrey Garen  <ggaren@apple.com>
3203
3204         build fix
3205
3206         * bindings/objc/WebScriptObject.mm:
3207         (+[WebScriptObject throwException:]): Restore assignment I accidentally
3208         deleted in previous commit
3209
3210 2006-06-12  Geoffrey Garen  <ggaren@apple.com>
3211
3212         Reviewed by TimO, Maciej.
3213
3214         - Merged InterpreterImp code into Interpreter, which implements
3215         all interpreter functionality now. This is part of my continuing quest
3216         to create an external notion of JS "execution context" that is unified and simple --
3217         something to replace the mix of Context, ContextImp, ExecState, Interpreter,
3218         InterpreterImp, and JSRun.
3219
3220         All tests pass. Leaks test has not regressed from its baseline ~207 leaks
3221         with ~3460 leaked nodes.
3222
3223         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
3224         * JavaScriptCore.xcodeproj/project.pbxproj:
3225         * bindings/NP_jsobject.cpp:
3226         * bindings/objc/WebScriptObject.mm:
3227         (+[WebScriptObject throwException:]):
3228         * bindings/runtime_root.cpp:
3229         * bindings/runtime_root.h:
3230         * kjs/Context.cpp:
3231         (KJS::Context::Context):
3232         * kjs/ExecState.cpp: Added.
3233         (KJS::ExecState::lexicalInterpreter):
3234         * kjs/ExecState.h: Added.
3235         (KJS::ExecState::dynamicInterpreter):
3236         * kjs/SavedBuiltins.h: Added.
3237         * kjs/bool_object.cpp:
3238         (BooleanPrototype::BooleanPrototype):
3239         * kjs/collector.cpp:
3240         (KJS::Collector::collect):
3241         (KJS::Collector::numInterpreters):
3242         * kjs/context.h:
3243         * kjs/debugger.cpp:
3244         (Debugger::attach):
3245         (Debugger::detach):
3246         * kjs/function.cpp:
3247         (KJS::FunctionImp::callAsFunction):
3248         (KJS::GlobalFuncImp::callAsFunction):
3249         * kjs/function_object.cpp:
3250         (FunctionObjectImp::construct):
3251         * kjs/internal.cpp:
3252         * kjs/internal.h:
3253         * kjs/interpreter.cpp:
3254         (KJS::interpreterMap):
3255         (KJS::Interpreter::Interpreter):
3256         (KJS::Interpreter::init):
3257         (KJS::Interpreter::~Interpreter):
3258         (KJS::Interpreter::globalObject):
3259         (KJS::Interpreter::initGlobalObject):
3260         (KJS::Interpreter::globalExec):
3261         (KJS::Interpreter::checkSyntax):
3262         (KJS::Interpreter::evaluate):
3263         (KJS::Interpreter::builtinObject):
3264         (KJS::Interpreter::builtinFunction):
3265         (KJS::Interpreter::builtinArray):
3266         (KJS::Interpreter::builtinBoolean):
3267         (KJS::Interpreter::builtinString):
3268         (KJS::Interpreter::builtinNumber):
3269         (KJS::Interpreter::builtinDate):
3270         (KJS::Interpreter::builtinRegExp):
3271         (KJS::Interpreter::builtinError):
3272         (KJS::Interpreter::builtinObjectPrototype):
3273         (KJS::Interpreter::builtinFunctionPrototype):
3274         (KJS::Interpreter::builtinArrayPrototype):
3275         (KJS::Interpreter::builtinBooleanPrototype):
3276         (KJS::Interpreter::builtinStringPrototype):
3277         (KJS::Interpreter::builtinNumberPrototype):
3278         (KJS::Interpreter::builtinDatePrototype):
3279         (KJS::Interpreter::builtinRegExpPrototype):
3280         (KJS::Interpreter::builtinErrorPrototype):
3281         (KJS::Interpreter::builtinEvalError):
3282         (KJS::Interpreter::builtinRangeError):
3283         (KJS::Interpreter::builtinReferenceError):
3284         (KJS::Interpreter::builtinSyntaxError):
3285         (KJS::Interpreter::builtinTypeError):
3286         (KJS::Interpreter::builtinURIError):
3287         (KJS::Interpreter::builtinEvalErrorPrototype):
3288         (KJS::Interpreter::builtinRangeErrorPrototype):
3289         (KJS::Interpreter::builtinReferenceErrorPrototype):
3290         (KJS::Interpreter::builtinSyntaxErrorPrototype):
3291         (KJS::Interpreter::builtinTypeErrorPrototype):
3292         (KJS::Interpreter::builtinURIErrorPrototype):
3293         (KJS::Interpreter::mark):
3294         (KJS::Interpreter::interpreterWithGlobalObject):
3295         (KJS::Interpreter::saveBuiltins):
3296         (KJS::Interpreter::restoreBuiltins):
3297         * kjs/interpreter.h:
3298         (KJS::Interpreter::setCompatMode):
3299         (KJS::Interpreter::compatMode):
3300         (KJS::Interpreter::firstInterpreter):
3301         (KJS::Interpreter::nextInterpreter):
3302         (KJS::Interpreter::prevInterpreter):
3303         (KJS::Interpreter::debugger):
3304         (KJS::Interpreter::setDebugger):
3305         (KJS::Interpreter::setContext):
3306         (KJS::Interpreter::context):
3307         * kjs/nodes.cpp:
3308         (StatementNode::hitStatement):
3309         (RegExpNode::evaluate):
3310         * kjs/protect.h:
3311
3312 2006-06-12  Geoffrey Garen  <ggaren@apple.com>
3313
3314         Reviewed by Maciej.
3315
3316         - Have *.lut.h files #include lookup.h to eliminate surprising header
3317         include order dependency.
3318
3319         * DerivedSources.make:
3320         * kjs/array_object.cpp:
3321         * kjs/date_object.cpp:
3322         * kjs/date_object.h:
3323         (KJS::DateProtoFunc::):
3324         * kjs/lexer.cpp:
3325         * kjs/math_object.cpp:
3326         * kjs/number_object.cpp:
3327         * kjs/regexp_object.cpp:
3328         * kjs/string_object.cpp:
3329
3330 2006-06-10  Geoffrey Garen  <ggaren@apple.com>
3331
3332         - http://bugzilla.opendarwin.org/show_bug.cgi?id=8515
3333         Linux porting compile bug
3334
3335         Fix by Mike Emmel, Reviewed by Darin.
3336
3337         * JavaScriptCoreSources.bkl:
3338         * jscore.bkl:
3339         * wtf/Platform.h:
3340
3341 2006-06-09  Geoffrey Garen  <ggaren@apple.com>
3342
3343         Build fix -- I think :).
3344
3345         * JavaScriptCore.xcodeproj/project.pbxproj:
3346         * kjs/context.h:
3347
3348 2006-06-09  Geoffrey Garen  <ggaren@apple.com>
3349
3350         Reviewed by Eric (yay!).
3351
3352         - Removed Context wrapper for ContextImp, renamed ContextImp to Context,
3353         split Context into its own file -- Context.cpp -- renamed _var to m_var,
3354         change ' *' to '* '.
3355
3356         * JavaScriptCore.xcodeproj/project.pbxproj:
3357         * kjs/Context.cpp: Added.
3358         (KJS::Context::Context):
3359         (KJS::Context::~Context):
3360         (KJS::Context::mark):
3361         * kjs/context.h:
3362         (KJS::Context::scopeChain):
3363         (KJS::Context::variableObject):
3364         (KJS::Context::setVariableObject):
3365         (KJS::Context::thisValue):
3366         (KJS::Context::callingContext):
3367         (KJS::Context::activationObject):
3368         (KJS::Context::currentBody):
3369         (KJS::Context::function):
3370         (KJS::Context::arguments):
3371         (KJS::Context::pushScope):
3372         (KJS::Context::seenLabels):
3373         * kjs/function.cpp:
3374         (KJS::FunctionImp::callAsFunction):
3375         (KJS::FunctionImp::processParameters):
3376         (KJS::FunctionImp::argumentsGetter):
3377         (KJS::GlobalFuncImp::callAsFunction):
3378         * kjs/internal.cpp:
3379         (KJS::InterpreterImp::evaluate):
3380         * kjs/internal.h:
3381         (KJS::InterpreterImp::setContext):
3382         (KJS::InterpreterImp::context):
3383         * kjs/interpreter.cpp:
3384         * kjs/interpreter.h:
3385         (KJS::ExecState::context):
3386         (KJS::ExecState::ExecState):
3387         * kjs/nodes.cpp:
3388         (currentSourceId):
3389         (currentSourceURL):
3390         (ThisNode::evaluate):
3391         (ResolveNode::evaluate):
3392         (FunctionCallResolveNode::evaluate):
3393         (PostfixResolveNode::evaluate):
3394         (DeleteResolveNode::evaluate):
3395         (TypeOfResolveNode::evaluate):
3396         (PrefixResolveNode::evaluate):
3397         (AssignResolveNode::evaluate):
3398         (VarDeclNode::evaluate):
3399         (VarDeclNode::processVarDecls):
3400         (DoWhileNode::execute):
3401         (WhileNode::execute):
3402         (ForNode::execute):
3403         (ForInNode::execute):
3404         (ContinueNode::execute):
3405         (BreakNode::execute):
3406         (ReturnNode::execute):
3407         (WithNode::execute):
3408         (SwitchNode::execute):
3409         (LabelNode::execute):
3410         (TryNode::execute):
3411         (FuncDeclNode::processFuncDecl):
3412         (FuncExprNode::evaluate):
3413
3414 2006-06-07  Geoffrey Garen  <ggaren@apple.com>
3415
3416         Removed API directory I prematurely/accidentally added.
3417
3418         * API: Removed.
3419
3420 2006-06-05  Mitz Pettel  <opendarwin.org@mitzpettel.com>
3421
3422         Reviewed and landed by Geoff.
3423
3424         - fix a regression in ecma_3/String/regress-104375.js
3425
3426         * kjs/string_object.cpp:
3427         (substituteBackreferences): If a 2-digit back reference is out of range,
3428         parse it as a 1-digit reference (followed by the other digit). This matches
3429         Firefox's behavior.
3430
3431 2006-06-05  Geoffrey Garen  <ggaren@apple.com>
3432
3433         Reviewed By Maciej.
3434         Darin already reviewed this change on the branch. See <rdar://problem/4317701>.
3435
3436         - Fixed <rdar://problem/4291345> PCRE overflow in Safari JavaScriptCore
3437
3438         No test case because there's no behavior change.
3439
3440         * pcre/pcre_compile.c:
3441         (read_repeat_counts): Check for integer overflow / out of bounds
3442
3443 2006-06-05  Geoffrey Garen  <ggaren@apple.com>
3444
3445         Reviewed by aliu.
3446
3447         - Changed CString length from int to size_t. We sould probably do this
3448         for UString, too. (Darin, if you're reading this: Maciej said so.)
3449
3450         * kjs/function.cpp:
3451         (KJS::encode):
3452         * kjs/ustring.cpp:
3453         (KJS::CString::CString):
3454         (KJS::operator==):
3455         * kjs/ustring.h:
3456         (KJS::CString::size):
3457
3458 2006-06-04  Geoffrey Garen  <ggaren@apple.com>
3459
3460         Reviewed by Maciej.
3461
3462         - http://bugzilla.opendarwin.org/show_bug.cgi?id=9304
3463         Minor cleanup in JavaScriptCore
3464
3465         * kjs/value.h: Removed redundant declarations
3466
3467 2006-06-04  Darin Adler  <darin@apple.com>
3468
3469         Reviewed by Anders.
3470
3471         - changed deleteAllValues so it can work on "const" collections
3472           Deleting the values affects the values, not the pointers in the
3473           collection, so it's legitimate to do it to a const collection,
3474           and a case of that actually came up in the XPath code.
3475
3476         * wtf/HashMap.h:
3477         (WTF::deleteAllPairSeconds): Use const iterators.
3478         (WTF::deleteAllValues): Take const HashMap reference as a parameter.
3479         * wtf/HashSet.h:
3480         (WTF::deleteAllValues): Take const HashSet reference as a parameter,
3481         and use const iterators.
3482         * wtf/Vector.h:
3483         (WTF::deleteAllValues): Take const Vector reference as a parameter.
3484
3485         - added more functions that are present in <math.h> on some platforms,
3486           but not on others; moved here from various files in WebCore
3487
3488         * wtf/MathExtras.h:
3489         (isinf): Added.
3490         (isnan): Added.
3491         (lround): Added.
3492         (lroundf): Tweaked.
3493         (round): Added.
3494         (roundf): Tweaked.
3495         (signbit): Added.
3496
3497 2006-06-02  Mitz Pettel  <opendarwin.org@mitzpettel.com>
3498
3499         Reviewed by ggaren.
3500
3501         - http://bugzilla.opendarwin.org/show_bug.cgi?id=9234
3502           Implement $&, $' and $` replacement codes in String.prototype.replace
3503
3504         Test: fast/js/string-replace-3.html
3505
3506         * kjs/string_object.cpp:
3507         (substituteBackreferences): Added support for $& (matched substring),
3508         $` (everything preceding matched substring), $' (everything following
3509         matched substring) and 2-digit back references, and cleaned up a little.
3510
3511 2006-06-02 Adele Peterson  <adele@apple.com>
3512
3513         Reviewed by Darin.
3514
3515         Set incremental linking to no. This seems to fix a build problem I was seeing
3516         where dftables couldn't find a dll.
3517
3518         * JavaScriptCore.vcproj/dftables/dftables.vcproj:
3519
3520 2006-05-26  Steve Falkenburg  <sfalken@apple.com>
3521
3522         Build fixes/tweaks
3523
3524         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
3525
3526 === JavaScriptCore-521.11 ===
3527
3528 2006-05-24  Geoffrey Garen  <ggaren@apple.com>
3529
3530         Reviewed by mjs.
3531
3532         - JSC half of fix for <rdar://problem/4557926> TOT REGRESSSION: Crash
3533         occurs when attempting to view image in slideshow mode at
3534         http://d.smugmug.com/gallery/581716 ( KJS::IfNode::execute
3535         (KJS::ExecState*) + 312)
3536
3537         On alternate threads, DOMObjects remain in the
3538         ScriptInterpreter's cache because they're not collected. So, they
3539         need an opportunity to mark their children.
3540
3541         I'm not particularly happy with this solution because it fails to
3542         resolve many outstanding issues with the DOM object cache. Since none
3543         of those issues is a crasher or a serious compatibility concern,
3544         and since the behavior of other browsers is not much to go on in this
3545         case, I've filed <rdar://problem/4561439> about that, and I'm moving on
3546         with my life.
3547
3548         * JavaScriptCore.xcodeproj/project.pbxproj:
3549         * kjs/collector.cpp:
3550         (KJS::Collector::collect):
3551         * kjs/internal.cpp:
3552         (KJS::InterpreterImp::mark):
3553         * kjs/internal.h:
3554         * kjs/interpreter.cpp:
3555         (KJS::Interpreter::mark):
3556         * kjs/interpreter.h:
3557
3558 === JavaScriptCore-521.10 ===
3559
3560 2006-05-22  Timothy Hatcher  <timothy@apple.com>
3561
3562         Reviewed by Eric, Kevin and Geoff.
3563
3564         Merge open source build fixes. <rdar://problem/4555500>
3565
3566         * kjs/collector.cpp: look at the rsp register in x86_64
3567         (KJS::Collector::markOtherThreadConservatively):
3568         * wtf/Platform.h: add x86_64 to the platform list
3569
3570 2006-05-19  Anders Carlsson  <acarlsson@apple.com>
3571
3572         Reviewed by Geoff.
3573
3574         http://bugzilla.opendarwin.org/show_bug.cgi?id=8993
3575         Support function declaration in case statements
3576
3577         * kjs/grammar.y: Get rid of StatementList and use SourceElements instead.
3578
3579         * kjs/nodes.cpp:
3580         (CaseClauseNode::evalStatements):
3581         (CaseClauseNode::processVarDecls):
3582         (CaseClauseNode::processFuncDecl):
3583         (ClauseListNode::processFuncDecl):
3584         (CaseBlockNode::processFuncDecl):
3585         (SwitchNode::processFuncDecl):
3586         * kjs/nodes.h:
3587         (KJS::CaseClauseNode::CaseClauseNode):
3588         (KJS::ClauseListNode::ClauseListNode):
3589         (KJS::ClauseListNode::getClause):
3590         (KJS::ClauseListNode::getNext):
3591         (KJS::ClauseListNode::releaseNext):
3592         (KJS::SwitchNode::SwitchNode):
3593         Add processFuncDecl for the relevant nodes.
3594
3595         * kjs/nodes2string.cpp:
3596         (CaseClauseNode::streamTo):
3597         next got renamed to source.
3598
3599 2006-05-17  George Staikos <staikos@kde.org>
3600
3601         Reviewed by Maciej, Alexey, and Eric.
3602
3603         * pcre/pcre_compile.c:
3604         * pcre/pcre_get.c:
3605         * pcre/pcre_exec.c:
3606         * wtf/UnusedParam.h:
3607         Use /**/ in .c files to compile with non-C99 and non-GCC compilers.
3608
3609         * kjs/testkjs.cpp:
3610         Change include to <wtf/HashTraits.h> from "HashTraits.h" to avoid -I
3611
3612         * wtf/unicode/qt4/UnicodeQt4.h:
3613         Use correct parentheses and correct mask for utf-32 support.
3614
3615 2006-05-17  Alexey Proskuryakov  <ap@nypop.com>
3616
3617         Reviewed by Darin.
3618
3619         - fix http://bugzilla.opendarwin.org/show_bug.cgi?id=8870
3620         Crash typing in Yahoo auto-complete widget.
3621
3622         Test: fast/js/regexp-stack-overflow.html
3623
3624         * pcre/pcre-config.h: Define NO_RECURSE.
3625
3626 2006-05-16  George Staikos <staikos@kde.org>
3627
3628         Reviewed by Maciej.
3629
3630         Fix some warnings and strict compilation errors.
3631
3632         * kjs/nodes.cpp:
3633         * kjs/value.cpp:
3634
3635 2006-05-15  Alexey Proskuryakov  <ap@nypop.com>
3636
3637         * make-generated-sources.sh: Changed to be executable and removed
3638         text in the file generated by "svn diff".
3639
3640 2006-05-15  Geoffrey Garen  <ggaren@apple.com>
3641
3642         Reviewed by Maciej.
3643
3644         - Fixed <rdar://problem/4534904> please do not treat "debugger" as
3645         a reserved word while parsing JavaScript (and other ECMA reserved
3646         words)
3647
3648         AKA
3649
3650         http://bugzilla.opendarwin.org/show_bug.cgi?id=6179
3651         We treat "char" as a reserved word in JavaScript and firefox/IE do
3652         not
3653
3654         (1) I unreserved most of the spec's "future reserved words" because
3655         they're not reserved in IE or FF. (Most, but not all, because IE
3656         somewhat randomly *does* reserve a few of them.)
3657         (2) I made 'debugger' a legitimate statement that acts like an empty
3658         statement because FF and IE support it.
3659
3660         * kjs/grammar.y:
3661         * kjs/keywords.table:
3662
3663 2006-05-15  Tim Omernick  <timo@apple.com>
3664
3665         Reviewed by John Sullivan.
3666
3667         Part of <rdar://problem/4466508> Add 64-bit support to the Netscape Plugin API
3668
3669         Added to the Netscape Plugin API the concept of "plugin drawing models".  The drawing model
3670         determines the kind of graphics context created by the browser for the plugin, as well as
3671         the Mac types of various Netscape Plugin API data structures.
3672
3673         There is a drawing model to represent the old QuickDraw-based API.  It is used by default
3674         if QuickDraw is available on the system, unless the plugin specifies another drawing model.
3675
3676         The big change is the addition of the CoreGraphics drawing model.  A plugin may request this
3677         drawing model to obtain access to a CGContextRef for drawing, instead of a QuickDraw CGrafPtr.
3678
3679         * bindings/npapi.h:
3680         Define NP_NO_QUICKDRAW when compiling 64-bit; there is no 64-bit QuickDraw.
3681         Added NPNVpluginDrawingModel, NPNVsupportsQuickDrawBool, and NPNVsupportsCoreGraphicsBool
3682         variables.
3683         Added NPDrawingModel enumeration.  Currently the only drawing models are QuickDraw and
3684         CoreGraphics.
3685         NPRegion's type now depends on the drawing model specified by the plugin.
3686         NP_Port is now only defined when QuickDraw is available.
3687         Added NP_CGContext, which is the type of the NPWindow's "window" member in CoreGraphics mode.
3688
3689 2006-05-13  Kevin M. Ollivier  <kevino@theolliviers.com>
3690
3691         Reviewed by Darin, landed by ap.
3692
3693         - http://bugzilla.opendarwin.org/show_bug.cgi?id=8528
3694           Bakefiles (and generated Makefiles) for wx and gdk ports
3695
3696         * make-generated-sources.sh:
3697         Added script to configure environment to run DerivedSources.make
3698
3699         * JavaScriptCoreSources.bkl:
3700         Added JavaScriptCore sources list for Bakefile.
3701
3702         * jscore.bkl:
3703         Bakefile used to generate JavaScriptCore project files
3704         (currently only used by wx and gdk ports)
3705
3706 2006-05-09  Steve Falkenburg  <sfalken@apple.com>
3707
3708         Fix Windows build.
3709         Minor fixes to WTF headers.
3710
3711         Reviewed by kevin.
3712
3713         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: Fix include dirs, paths to files.
3714         * JavaScriptCore.vcproj/testkjs/testkjs.vcproj: Fix include dirs.
3715         * wtf/Assertions.h: include Platform.h to get definition for COMPILER()
3716         * wtf/Vector.h: include FastMalloc.h for definition of fastMalloc, fastFree
3717
3718 2006-05-09  Maciej Stachowiak  <mjs@apple.com>
3719
3720         Rubber stamped by Anders.
3721
3722         - renamed kxmlcore to wtf
3723
3724         kxmlcore --> wtf
3725         KXMLCore --> WTF
3726         KXC --> WTF
3727
3728         * JavaScriptCore.xcodeproj/project.pbxproj:
3729         * bindings/c/c_instance.cpp:
3730         * bindings/objc/WebScriptObject.mm:
3731         * kjs/JSImmediate.h:
3732         * kjs/Parser.cpp:
3733         * kjs/Parser.h:
3734         * kjs/array_object.cpp:
3735         * kjs/collector.cpp:
3736         (KJS::Collector::registerThread):
3737         * kjs/collector.h:
3738         * kjs/config.h:
3739         * kjs/function.cpp:
3740         (KJS::isStrWhiteSpace):
3741         * kjs/function.h:
3742         * kjs/identifier.cpp:
3743         * kjs/internal.cpp:
3744         * kjs/internal.h:
3745         * kjs/lexer.cpp:
3746         (Lexer::shift):
3747         (Lexer::isWhiteSpace):
3748         (Lexer::isIdentStart):
3749         (Lexer::isIdentPart):
3750         * kjs/lookup.cpp:
3751         * kjs/nodes.cpp:
3752         * kjs/nodes.h:
3753         * kjs/number_object.cpp:
3754         * kjs/object.h:
3755         * kjs/property_map.cpp:
3756         * kjs/property_map.h:
3757         * kjs/string_object.cpp:
3758         (StringProtoFunc::callAsFunction):
3759         * kjs/testkjs.cpp:
3760         (testIsInteger):
3761         * kjs/ustring.cpp:
3762         * kjs/ustring.h:
3763         * kxmlcore: Removed.
3764         * kxmlcore/AlwaysInline.h: Removed.
3765         * kxmlcore/Assertions.cpp: Removed.
3766         * kxmlcore/Assertions.h: Removed.
3767         * kxmlcore/FastMalloc.cpp: Removed.
3768         * kxmlcore/FastMalloc.h: Removed.
3769         * kxmlcore/FastMallocInternal.h: Removed.
3770         * kxmlcore/Forward.h: Removed.
3771         * kxmlcore/HashCountedSet.h: Removed.
3772         * kxmlcore/HashFunctions.h: Removed.
3773         * kxmlcore/HashMap.h: Removed.
3774         * kxmlcore/HashSet.h: Removed.
3775         * kxmlcore/HashTable.cpp: Removed.
3776         * kxmlcore/HashTable.h: Removed.
3777         * kxmlcore/HashTraits.h: Removed.
3778         * kxmlcore/ListRefPtr.h: Removed.
3779         * kxmlcore/Noncopyable.h: Removed.
3780         * kxmlcore/OwnArrayPtr.h: Removed.
3781         * kxmlcore/OwnPtr.h: Removed.
3782         * kxmlcore/PassRefPtr.h: Removed.
3783         * kxmlcore/Platform.h: Removed.
3784         * kxmlcore/RefPtr.h: Removed.
3785         * kxmlcore/TCPageMap.h: Removed.
3786         * kxmlcore/TCSpinLock.h: Removed.
3787         * kxmlcore/TCSystemAlloc.cpp: Removed.
3788         * kxmlcore/TCSystemAlloc.h: Removed.
3789         * kxmlcore/UnusedParam.h: Removed.
3790         * kxmlcore/Vector.h: Removed.
3791         * kxmlcore/VectorTraits.h: Removed.
3792         * kxmlcore/unicode: Removed.
3793         * kxmlcore/unicode/Unicode.h: Removed.
3794         * kxmlcore/unicode/UnicodeCategory.h: Removed.
3795         * kxmlcore/unicode/icu: Removed.
3796         * kxmlcore/unicode/icu/UnicodeIcu.h: Removed.
3797         * kxmlcore/unicode/posix: Removed.
3798         * kxmlcore/unicode/qt3: Removed.
3799         * kxmlcore/unicode/qt4: Removed.
3800         * kxmlcore/unicode/qt4/UnicodeQt4.h: Removed.
3801         * pcre/pcre_get.c:
3802         * wtf: Added.
3803         * wtf/Assertions.cpp:
3804         * wtf/Assertions.h:
3805         * wtf/FastMalloc.cpp:
3806         (WTF::TCMalloc_ThreadCache::Scavenge):
3807         (WTF::do_malloc):
3808         (WTF::do_free):
3809         (WTF::TCMallocGuard::TCMallocGuard):
3810         (WTF::malloc):
3811         (WTF::free):
3812         (WTF::calloc):
3813         (WTF::cfree):
3814         (WTF::realloc):
3815         * wtf/FastMalloc.h:
3816         * wtf/FastMallocInternal.h:
3817         * wtf/Forward.h:
3818         * wtf/HashCountedSet.h:
3819         * wtf/HashFunctions.h:
3820         * wtf/HashMap.h:
3821         * wtf/HashSet.h:
3822         * wtf/HashTable.cpp:
3823         * wtf/HashTable.h:
3824         * wtf/HashTraits.h:
3825         * wtf/ListRefPtr.h:
3826         * wtf/Noncopyable.h:
3827         * wtf/OwnArrayPtr.h:
3828         * wtf/OwnPtr.h:
3829         * wtf/PassRefPtr.h:
3830         * wtf/RefPtr.h:
3831         * wtf/TCSystemAlloc.cpp:
3832         (TCMalloc_SystemAlloc):
3833         * wtf/Vector.h:
3834         * wtf/VectorTraits.h:
3835         * wtf/unicode/UnicodeCategory.h:
3836         * wtf/unicode/icu/UnicodeIcu.h:
3837
3838 2006-05-08  Timothy Hatcher  <timothy@apple.com>
3839
3840         Reviewed by Tim O.
3841
3842         * bindings/npapi.h: do not define #pragma options align=mac68k if we are 64-bit
3843
3844 2006-05-07  Darin Adler  <darin@apple.com>
3845
3846         Reviewed and landed by Maciej.
3847
3848         - fix http://bugzilla.opendarwin.org/show_bug.cgi?id=8765
3849         Random crashes on TOT since the form state change
3850
3851         I haven't figured out how to construct a test for this, but this does seem to fix the
3852         problem; Mitz mentioned that a double-destroy was occurring in these functions.
3853
3854         * kxmlcore/HashMap.h: (KXMLCore::HashMap::remove): Use RefCounter::deref instead of calling
3855         ~ValueType, because ~ValueType often results in a double-destroy, since the HashTable also
3856         destroys the element based on the storage type. The RefCounter template correctly does work
3857         only in cases where ValueType and ValueStorageType differ and this class is what's used
3858         elsewhere for the same purpose; I somehow missed this case when optimizing HashMap.
3859         * kxmlcore/HashSet.h: (KXMLCore::HashSet::remove): Ditto.
3860
3861 2006-05-05  Darin Adler  <darin@apple.com>
3862
3863         - http://bugzilla.opendarwin.org/show_bug.cgi?id=8722
3864           IE compatibility fix in date parsing
3865
3866         * kjs/date_object.cpp: (KJS::parseDate): Merged change that George Staikos provided
3867         from KDE 3.4.3 branch that allows day values of 0 and values that are > 1000.
3868
3869 2006-05-04  Anders Carlsson  <andersca@mac.com>
3870
3871         Reviewed by Maciej.
3872
3873         http://bugzilla.opendarwin.org/show_bug.cgi?id=8734
3874         Would like a Vector::append that takes another Vector
3875
3876         * kxmlcore/Vector.h:
3877         (KXMLCore::::append):
3878         New function that takes another array.
3879
3880 2006-05-02  Steve Falkenburg  <sfalken@apple.com>
3881
3882         Reviewed by eric.
3883
3884         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: set NDEBUG for release build
3885         * kxmlcore/FastMalloc.cpp: Prevent USE_SYSTEM_MALLOC from being defined twice
3886
3887 2006-05-02  Anders Carlsson  <andersca@mac.com>
3888
3889         Reviewed by Maciej.
3890
3891         * kxmlcore/HashMap.h:
3892         (KXMLCore::::operator):
3893         Return *this
3894
3895 2006-05-01  Tim Omernick  <timo@apple.com>
3896
3897         Reviewed by Tim Hatcher.
3898
3899         <rdar://problem/4476875> Support printing for embedded Netscape plugins
3900
3901         * bindings/npapi.h:
3902         Fixed struct alignment problem in our npapi.h.  Structs must be 68k-aligned on both pre-Mac OS X
3903         and Mac OS X systems, as this is what plugins expect.
3904
3905 2006-05-01  Timothy Hatcher  <timothy@apple.com>
3906
3907         Reviewed by Maciej.
3908
3909         <rdar://problem/4308243> 8F36 Regression: crash in malloc_consolidate if you use a .PAC file
3910
3911         The original fix missed the oversized cell case. Added a test for "currentThreadIsMainThread ||
3912         imp->m_destructorIsThreadSafe" where we collect oversized cells.
3913
3914         We don't have a way to test PAC files yet, so there's no test attached.
3915
3916         * kjs/collector.cpp:
3917         (KJS::Collector::collect): test the thread when we collect oversized cells
3918
3919 2006-05-01  Tim Omernick  <timo@apple.com>
3920
3921         Reviewed by Adele.
3922
3923         <rdar://problem/4526114> REGRESSION (two days ago): LOG() just prints @ for NSObject substitutions
3924
3925         * kxmlcore/Assertions.cpp:
3926         Changed sense of strstr("%@") check.  I already made the same fix to the WebBrowser assertions.
3927
3928 2006-04-28  Steve Falkenburg  <sfalken@apple.com>
3929
3930         Reviewed by kdecker
3931
3932         Actually apply the change that was reviewed insted of checking it in with an #if 0 (oops).
3933
3934         * kjs/testkjs.cpp:
3935         (main): Suppress C runtime alerts
3936
3937 2006-04-28  Steve Falkenburg  <sfalken@apple.com>
3938
3939         Reviewed by kdecker
3940
3941         Suppress error reporting dialog that blocks Javascript tests from completing.
3942
3943         Real error is due to an overflow in the date/time handling functions that needs
3944         to be addressed, but this will prevent the hang running the Javascript tests
3945         on the build bot (along with the related changes).
3946
3947         * kjs/testkjs.cpp:
3948         (main): Suppress C runtime alerts
3949
3950 2006-04-27  Geoffrey Garen  <ggaren@apple.com>
3951
3952         Reviewed by Maciej
3953
3954         - Minor fixups I discovered while working on the autogenerator.
3955
3956         * kjs/lookup.cpp:
3957         (findEntry): ASSERT that size is not 0, because otherwise we'll % by 0,
3958         compute a garbage address, and possibly crash.
3959         * kjs/lookup.h:
3960         (cacheGlobalObject): Don't enumerate cached objects -- ideally, they
3961         would be hidden entirely.
3962
3963 2006-04-21  Kevin M. Ollivier  <kevino@theolliviers.com>
3964
3965         Reviewed by Darin.
3966
3967         - http://bugzilla.opendarwin.org/show_bug.cgi?id=8507
3968           Compilation fixes for building on gcc 4.0.2, and without precomp headers
3969
3970         * kjs/operations.h:
3971         * kxmlcore/Assertions.cpp:
3972         * kxmlcore/FastMalloc.cpp:
3973         Added necessary headers to resolve compilation issues when not using
3974         precompiled headers.
3975
3976         * kjs/value.h: Declare the JSCell class before friend declaration
3977         to resolve compilation issues with gcc 4.0.2.
3978
3979         * kxmlcore/Platform.h: Set Unicode support to use ICU on platforms
3980         other than KDE (previously only defined for Win and Mac OS)
3981
3982 2006-04-18  Eric Seidel  <eseidel@apple.com>
3983
3984         Reviewed by ggaren.
3985
3986         Fix "new Function()" to correctly use lexical scoping.
3987         Add ScopeChain::print() function for debugging.
3988         <rdar://problem/4067864> REGRESSION (125-407): JavaScript failure on PeopleSoft REN Server
3989
3990         * kjs/function_object.cpp:
3991         (FunctionObjectImp::construct):
3992         * kjs/scope_chain.cpp:
3993         (KJS::ScopeChain::print):
3994         * kjs/scope_chain.h:
3995
3996 2006-04-14  James G. Speth  <speth@end.com>
3997
3998         Reviewed by Timothy.
3999
4000         Bug 8389: support for Cocoa bindings - binding an NSTreeController to the WebView's DOM
4001         http://bugzilla.opendarwin.org/show_bug.cgi?id=8389
4002
4003         Adds a category to WebScriptObject with array accessors for KVC/KVO.
4004
4005         If super valueForKey: fails it will call valueForUndefinedKey:, which is
4006         important because it causes the right behavior to happen with bindings using
4007         the "Raises for Not Applicable Keys" flag and the "Not Applicable Placeholder"
4008
4009         * bindings/objc/WebScriptObject.mm:
4010         (-[WebScriptObject valueForKey:]):
4011         (-[WebScriptObject count]):
4012         (-[WebScriptObject objectAtIndex:]):
4013         (-[WebUndefined description]): return "undefined"
4014
4015 2006-04-13  Geoffrey Garen  <ggaren@apple.com>
4016
4017         Reviewed by Darin.
4018
4019         * kjs/internal.cpp:
4020         (KJS::InterpreterImp::initGlobalObject): Add the built-in object
4021         prototype to the end of the global object's prototype chain instead of
4022         just blowing away its existing prototype. We need to do this because
4023         the window object has a meaningful prototype now.
4024
4025 2006-04-13  Maciej Stachowiak  <mjs@apple.com>
4026
4027         Reviewed by Geoff.
4028
4029         - fix testkjs to not show false-positive KJS::Node leaks in debug builds
4030
4031         * kjs/testkjs.cpp:
4032         (doIt):
4033         (kjsmain):
4034
4035 2006-04-11  Geoffrey Garen  <ggaren@apple.com>
4036
4037         Reviewed by Maciej.
4038
4039         Minor code cleanup -- passes all the JS tests.
4040
4041         * kjs/object_object.cpp:
4042         (ObjectObjectImp::construct):
4043         (ObjectObjectImp::callAsFunction):
4044
4045 2006-04-11  Darin Adler  <darin@apple.com>
4046
4047         - another attempt to fix Windows build -- Vector in Forward.h was not working
4048
4049         * kxmlcore/Forward.h: Remove Vector.
4050         * kxmlcore/Vector.h: Add back default arguments, remove include of
4051         Forward.h.
4052
4053 2006-04-11  Darin Adler  <darin@apple.com>
4054
4055         - try to fix Windows build -- HashForward.h was not working
4056
4057         * kxmlcore/HashForward.h: Removed.
4058
4059         * JavaScriptCore.xcodeproj/project.pbxproj: Remove HashForward.h.
4060         * kjs/collector.h: Remove use of HashForward.h.
4061         * kxmlcore/HashCountedSet.h: Remove include of HashForward.h, restore
4062         default arguments.
4063         * kxmlcore/HashMap.h: Ditto.
4064         * kxmlcore/HashSet.h: Ditto.
4065
4066 2006-04-11  David Harrison  <harrison@apple.com>
4067
4068         Reviewed by Darin.
4069
4070         - fixed clean build, broken by Darin's check-in
4071
4072         * kjs/date_object.cpp: Add needed include of lookup.h.
4073         * kjs/regexp_object.cpp: Move include of .lut.h file below other includes.
4074
4075 2006-04-10  Darin Adler  <darin@apple.com>
4076
4077         Rubber-stamped by John Sullivan.
4078
4079         - switched from a shell script to a makefile for generated files
4080         - removed lots of unneeded includes
4081         - added new Forward.h and HashForward.h headers that allow compiling with
4082           fewer unneeded templates
4083
4084         * DerivedSources.make: Added.
4085         * generate-derived-sources: Removed.
4086         * JavaScriptCore.xcodeproj/project.pbxproj: Added new files, changed to use
4087         DerivedSources.make.
4088
4089         * kxmlcore/Forward.h: Added.
4090         * kxmlcore/HashForward.h: Added.
4091
4092         * kxmlcore/HashCountedSet.h: Include HashForward for default args.
4093         * kxmlcore/HashMap.h: Ditto.
4094         * kxmlcore/HashSet.h: Ditto.
4095
4096         * kjs/object.h:
4097         * kjs/object.cpp:
4098         Moved KJS_MAX_STACK into the .cpp file.
4099
4100         * bindings/NP_jsobject.cpp:
4101         * bindings/c/c_instance.h:
4102         * bindings/jni/jni_class.h:
4103         * bindings/jni/jni_runtime.h:
4104         * bindings/jni/jni_utility.h:
4105         * bindings/objc/WebScriptObject.mm:
4106         * bindings/objc/WebScriptObjectPrivate.h:
4107         * bindings/objc/objc_class.h:
4108         * bindings/objc/objc_class.mm:
4109         * bindings/objc/objc_instance.h:
4110         * bindings/objc/objc_instance.mm:
4111         * bindings/objc/objc_runtime.mm:
4112         * bindings/objc/objc_utility.mm:
4113         * bindings/runtime.h:
4114         * bindings/runtime_array.cpp:
4115         * bindings/runtime_array.h:
4116         * bindings/runtime_method.cpp:
4117         * bindings/runtime_method.h:
4118         * bindings/runtime_object.cpp:
4119         * bindings/runtime_root.h:
4120         * kjs/JSImmediate.cpp:
4121         * kjs/Parser.h:
4122         * kjs/array_object.cpp:
4123         * kjs/array_object.h:
4124         * kjs/bool_object.cpp:
4125         * kjs/bool_object.h:
4126         * kjs/collector.h:
4127         * kjs/context.h:
4128         * kjs/debugger.cpp:
4129         * kjs/error_object.h:
4130         * kjs/function_object.h:
4131         * kjs/internal.h:
4132         * kjs/lexer.cpp:
4133         * kjs/math_object.cpp:
4134         * kjs/math_object.h:
4135         * kjs/nodes.cpp:
4136         * kjs/nodes.h:
4137         * kjs/number_object.cpp:
4138         * kjs/number_object.h:
4139         * kjs/object_object.cpp:
4140         * kjs/operations.cpp:
4141         * kjs/protected_reference.h:
4142         * kjs/reference.h:
4143         * kjs/reference_list.h:
4144         * kjs/regexp_object.h:
4145         * kjs/string_object.cpp:
4146         * kjs/string_object.h:
4147         * kjs/testkjs.cpp:
4148         * kjs/value.cpp:
4149         * kjs/value.h:
4150         * kxmlcore/HashTable.h:
4151         * kxmlcore/ListRefPtr.h:
4152         * kxmlcore/TCPageMap.h:
4153         * kxmlcore/Vector.h:
4154         Removed unneeded header includes.
4155
4156 2006-04-09  Geoffrey Garen  <ggaren@apple.com>
4157
4158         Reviewed by eric.
4159
4160         - Fixed http://bugzilla.opendarwin.org/show_bug.cgi?id=8284
4161         prevent unnecessary entries in the "nodes with extra refs" hash table
4162
4163         This patch switches manually RefPtr exchange with use of
4164         RefPtr::release to ensure that a node's ref count never tops 1
4165         (in the normal case).
4166
4167         * kjs/nodes.cpp:
4168         (BlockNode::BlockNode):
4169         (CaseBlockNode::CaseBlockNode):
4170         * kjs/nodes.h:
4171         (KJS::ArrayNode::ArrayNode):
4172         (KJS::ObjectLiteralNode::ObjectLiteralNode):
4173         (KJS::ArgumentsNode::ArgumentsNode):
4174         (KJS::VarStatementNode::VarStatementNode):
4175         (KJS::ForNode::ForNode):
4176         (KJS::CaseClauseNode::CaseClauseNode):
4177         (KJS::FuncExprNode::FuncExprNode):
4178         (KJS::FuncDeclNode::FuncDeclNode):
4179
4180 2006-04-08  Alexey Proskuryakov  <ap@nypop.com>
4181
4182         Reviewed by Darin.
4183
4184         One more attempt - use reinterpret_cast, rather than static_cast.
4185
4186 2006-04-08  Alexey Proskuryakov  <ap@nypop.com>
4187
4188         Reviewed by Darin.
4189
4190         An attempt to fix Win32 build - ICU uses wchar_t on Windows, so we need a type cast.
4191
4192         * kxmlcore/unicode/icu/UnicodeIcu.h:
4193         (KXMLCore::Unicode::toLower):
4194         (KXMLCore::Unicode::toUpper):