JavaScriptCore/ChangeLog

   1 2006-09-11  Brady Eidson  <beidson@apple.com>
   2
   3         Build fix - I think Tim's last checkin wasn't tested on Tiger, possibly.  I simply
   4         commented out the undefined constants until he can have a chance to make the right call
   5
   6         * bindings/objc/objc_utility.mm:
   7         (KJS::Bindings::objcValueTypeForType): Commented out undefined symbols
   8
   9 2006-09-11  Timothy Hatcher  <timothy@apple.com>
  10
  11         Reviewed by Tim O. and Darin.
  12
  13         Add support for more method signatures affecting ObjC methods called from JavaScript:
  14         - Added unsigned types and long long.
  15         - Allow methods that use const, oneway, bycopy and byref type modifiers.
  16
  17         * bindings/objc/objc_instance.mm:
  18         (ObjcInstance::invokeMethod):
  19         * bindings/objc/objc_utility.h:
  20         (KJS::Bindings::):
  21         * bindings/objc/objc_utility.mm:
  22         (KJS::Bindings::convertValueToObjcValue):
  23         (KJS::Bindings::convertObjcValueToValue):
  24         (KJS::Bindings::objcValueTypeForType):
  25
  26 2006-09-05  Timothy Hatcher  <timothy@apple.com>
  27
  28         Reviewed by Tim O.
  29
  30         <rdar://problem/4715840> SEL is not char*
  31
  32         * bindings/objc/objc_class.mm:
  33         (KJS::Bindings::ObjcClass::methodsNamed): use sel_getName instead of a char* cast.
  34         * bindings/objc/objc_runtime.mm:
  35         (ObjcFallbackObjectImp::callAsFunction): ditto
  36
  37 2006-09-03  Alexey Proskuryakov  <ap@nypop.com>
  38
  39         Reviewed by Tim H.
  40
  41         http://bugzilla.opendarwin.org/show_bug.cgi?id=10693
  42         Convert JavaScript arrays to AppleScript lists
  43
  44         * JavaScriptCore.exp: Export ArrayInstance::info and ArrayInstance::getItem().
  45         * kjs/array_instance.h:
  46         * kjs/array_object.cpp:
  47         (ArrayInstance::getItem): Added a method to access array items from C++.
  48
  49 2006-09-02  Krzysztof Kowalczyk  <kkowalczyk@gmail.com>
  50
  51         Reviewed by Tim H.
  52
  53         Bug 10454: Unix bakefile fixes
  54         http://bugzilla.opendarwin.org/show_bug.cgi?id=10454
  55
  56         * JavaScriptCoreSources.bkl:
  57
  58 2006-09-01  Nikolas Zimmermann  <zimmermann@kde.org>
  59
  60         Reviewed by hyatt.  Landed by eseidel.
  61
  62         Fix build on Linux.
  63
  64         * pcre/CMakeLists.txt: Add wtf/ include.
  65
  66 2006-09-01  Nikolas Zimmermann  <zimmermann@kde.org>
  67
  68         Reviewed and landed by ap.
  69
  70         Fix build on Linux (C89 without gcc extensions enabled).
  71
  72         * pcre/pcre_internal.h: Use C style comments.
  73         * wtf/Assertions.h: Use C style comments.
  74         * wtf/Platform.h: Use C style comments.
  75
  76 2006-09-01  Steve Falkenburg  <sfalken@apple.com>
  77
  78         Fix build.
  79
  80         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
  81         * JavaScriptCore.vcproj/dftables/dftables.vcproj:
  82
  83 2006-08-31  Anders Carlsson  <acarlsson@apple.com>
  84
  85         Reviewed by Darin.
  86
  87         Add new portability functions to MathExtras.h and add StringExtras.h which is for
  88         string portability functions.
  89
  90         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
  91         * bindings/c/c_instance.cpp:
  92         * kjs/date_object.cpp:
  93         * wtf/MathExtras.h:
  94         (copysign):
  95         (isfinite):
  96         * wtf/StringExtras.h: Added.
  97         (snprintf):
  98         (strncasecmp):
  99
 100 2006-08-31  Anders Carlsson  <acarlsson@apple.com>
 101
 102         Reviewed by Tim H.
 103
 104         Fix Windows build.
 105
 106         * JavaScriptCore.vcproj/dftables/dftables.vcproj:
 107         * pcre/pcre_internal.h:
 108
 109 2006-08-31  Timothy Hatcher  <timothy@apple.com>
 110
 111         Reviewed by Geoff.
 112
 113         Band-aid fix for PCRE to compile for ppc64 and x86_64 now that
 114         we use -Wshorten-64-to-32. Adds an INT_CAST macro that ASSERTs
 115         the value <= INT_MAX.
 116
 117         I filed <rdar://problem/4712064> to track the need to verify
 118         PCRE's 64-bit compliance.
 119
 120         * pcre/pcre_compile.c:
 121         (complete_callout):
 122         (compile_branch):
 123         (compile_regex):
 124         (pcre_compile2):
 125         * pcre/pcre_exec.c:
 126         (match):
 127         (pcre_exec):
 128         * pcre/pcre_get.c:
 129         (pcre_get_substring_list):
 130         * pcre/pcre_internal.h:
 131         * pcre/pcre_tables.c:
 132         * pcre/pcre_try_flipped.c:
 133         (_pcre_try_flipped):
 134
 135 2006-08-30  Darin Adler  <darin@apple.com>
 136
 137         Reviewed by Tim Hatcher.
 138
 139         - add WTF::getPtr, a function template that makes it possible to write
 140           generic code that gets a raw pointer out of any of our pointer types
 141
 142         * JavaScriptCore.xcodeproj/project.pbxproj:
 143         * wtf/GetPtr.h: Added.
 144         * wtf/ListRefPtr.h: (WTF::getPtr): Added.
 145         * wtf/OwnArrayPtr.h: (WTF::getPtr): Added.
 146         * wtf/OwnPtr.h: (WTF::getPtr): Added.
 147         * wtf/PassRefPtr.h: (WTF::getPtr): Added.
 148         * wtf/RefPtr.h: (WTF::getPtr): Added.
 149
 150 2006-08-29  waylonis  <waylonis@google.com>
 151
 152         Reviewed, tweaked by ggaren.
 153
 154         - Added storage and accessor functions for ExecState as a fix for
 155           http://bugzilla.opendarwin.org/show_bug.cgi?id=10114
 156
 157         * kjs/ExecState.cpp:
 158         (KJS::ExecState::ExecState):
 159         * kjs/ExecState.h:
 160         * kjs/context.h:
 161         (KJS::Context::setExecState):
 162         (KJS::Context::execState):
 163
 164 2006-08-30  Nikolas Zimmermann  <zimmermann@kde.org>
 165
 166         Reviewed by Tim H.
 167
 168         Commit KDE related tweaks, to be able to
 169         differentiate between a Qt-only or a KDE build.
 170
 171         * CMakeLists.txt: Install wtf-unity library.
 172         * wtf/Platform.h: Add define for the KDE platform.
 173
 174 2006-08-28  Darin Adler  <darin@apple.com>
 175
 176         Reviewed by Geoff.
 177
 178         * kjs/list.h: Use explicit in constructor (as appropriate).
 179
 180 2006-08-24  Nikolas Zimmermann  <zimmermann@kde.org>
 181
 182         Reviewed, tweaked and landed by ap
 183
 184         http://bugzilla.opendarwin.org/show_bug.cgi?id=10467
 185         WebKit should have Qt platform support (Part II)
 186
 187         * CMakeLists.txt: Adjust to Anders' build fixes.
 188         * wtf/Platform.h: Fix define for the Qt platform (we don't use/need Cairo.)
 189
 190 2006-08-23  David Hyatt  <hyatt@apple.com>
 191
 192         Fix Platform.h to include #defines for graphics features.
 193
 194         Reviewed by darin
 195
 196         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
 197         * wtf/Platform.h:
 198
 199 2006-08-23  Anders Carlsson  <acarlsson@apple.com>
 200
 201         Reviewed by Darin.
 202
 203         Make the bindings compile without CoreFoundation.
 204
 205         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
 206         * bindings/c/c_instance.cpp:
 207         * bindings/c/c_utility.cpp:
 208         (KJS::Bindings::convertUTF8ToUTF16):
 209         * bindings/npapi.h:
 210         * bindings/runtime.cpp:
 211         (KJS::Bindings::Instance::createBindingForLanguageInstance):
 212         (KJS::Bindings::Instance::createLanguageInstanceForValue):
 213         * bindings/runtime_root.cpp:
 214         * bindings/runtime_root.h:
 215         * kjs/interpreter.cpp:
 216         (KJS::Interpreter::createLanguageInstanceForValue):
 217         * kjs/interpreter.h:
 218
 219 2006-08-22  Anders Carlsson  <acarlsson@apple.com>
 220
 221         Reviewed by Darin.
 222
 223         Move the npruntime code over to using HashMap and the runtime_root code over to using
 224         HashMap and HashCountedSet.
 225
 226         * bindings/NP_jsobject.cpp:
 227         * bindings/c/c_utility.cpp:
 228         (KJS::Bindings::identifierFromNPIdentifier):
 229         * bindings/c/c_utility.h:
 230         * bindings/jni/jni_jsobject.cpp:
 231         (JavaJSObject::invoke):
 232         * bindings/npruntime.cpp:
 233         (getStringIdentifierMap):
 234         (getIntIdentifierMap):
 235         (_NPN_GetStringIdentifier):
 236         (_NPN_GetIntIdentifier):
 237         * bindings/runtime_root.cpp:
 238         (getReferencesByRootMap):
 239         (getReferencesSet):
 240         (KJS::Bindings::findReferenceSet):
 241         (KJS::Bindings::rootForImp):
 242         (KJS::Bindings::rootForInterpreter):
 243         (KJS::Bindings::addNativeReference):
 244         (KJS::Bindings::removeNativeReference):
 245         (RootObject::removeAllNativeReferences):
 246         * bindings/runtime_root.h:
 247
 248 2006-08-22  Anders Carlsson  <acarlsson@apple.com>
 249
 250         Reviewed by Geoff.
 251
 252         Switch over the NPAPI and Java bindings to using HashMaps instead of dictionaries.
 253
 254         * JavaScriptCore.xcodeproj/project.pbxproj:
 255         * bindings/c/c_class.cpp:
 256         (KJS::Bindings::CClass::CClass):
 257         (KJS::Bindings::CClass::~CClass):
 258         (KJS::Bindings::CClass::classForIsA):
 259         (KJS::Bindings::CClass::methodsNamed):
 260         (KJS::Bindings::CClass::fieldNamed):
 261         * bindings/c/c_class.h:
 262         * bindings/jni/jni_class.cpp:
 263         (JavaClass::JavaClass):
 264         (JavaClass::~JavaClass):
 265         (JavaClass::methodsNamed):
 266         (JavaClass::fieldNamed):
 267         * bindings/jni/jni_class.h:
 268         * bindings/objc/objc_class.h:
 269         * bindings/objc/objc_class.mm:
 270         (KJS::Bindings::deleteMethod):
 271         (KJS::Bindings::deleteField):
 272         (KJS::Bindings::):
 273         (KJS::Bindings::ObjcClass::methodsNamed):
 274         (KJS::Bindings::ObjcClass::fieldNamed):
 275         * bindings/runtime.cpp:
 276         * bindings/runtime.h:
 277         * bindings/runtime_object.cpp:
 278         (RuntimeObjectImp::fieldGetter):
 279         (RuntimeObjectImp::getOwnPropertySlot):
 280         (RuntimeObjectImp::put):
 281         (RuntimeObjectImp::canPut):
 282
 283 2006-08-21  Vladimir Olexa  <vladimir.olexa@gmail.com>
 284
 285         Reviewed by Darin.
 286
 287         http://bugzilla.opendarwin.org/show_bug.cgi?id=6252
 288         JavaScript 1.6 Array.lastIndexOf
 289
 290         Test: fast/js/array-lastIndexOf.html
 291
 292         * kjs/array_object.cpp:
 293         (ArrayProtoFunc::callAsFunction): Added a LastIndexOf case.
 294         * kjs/array_object.h:
 295         (KJS::ArrayProtoFunc::): Added LastIndexOf to enum.
 296         * tests/mozilla/expected.html: Two more tests now pass.
 297
 298 2006-08-20  Nikolas Zimmermann  <zimmermann@kde.org>
 299
 300         Reviewed by Maciej. Landed by rwlbuis.
 301
 302         Fixes parts of: http://bugzilla.opendarwin.org/show_bug.cgi?id=10463
 303         WebKit should have Qt platform support
 304
 305         Removing obsolete QConstString/QString constructors in kjs code.
 306
 307         * kjs/identifier.h:
 308         * kjs/ustring.h:
 309
 310 2006-08-17  Nikolas Zimmermann  <zimmermann@kde.org>
 311
 312         Reviewed by Maciej. Landed by rwlbuis.
 313
 314         Fixes: http://bugzilla.opendarwin.org/show_bug.cgi?id=10463
 315         WTF Changes needed for Qt platform code.
 316
 317         * wtf/Platform.h:
 318         * wtf/unicode/UnicodeDecomposition.h: Added.
 319         (WTF::Unicode::):
 320         * wtf/unicode/UnicodeDirection.h: Added.
 321         (WTF::Unicode::):
 322         * wtf/unicode/qt4/UnicodeQt4.cpp: Added.
 323         (WTF::Unicode::direction):
 324         (WTF::Unicode::category):
 325         (WTF::Unicode::decomposition):
 326         * wtf/unicode/qt4/UnicodeQt4.h:
 327         (WTF::Unicode::toLower):
 328         (WTF::Unicode::toUpper):
 329         (WTF::Unicode::isPrintableChar):
 330         (WTF::Unicode::isSpace):
 331         (WTF::Unicode::isPunct):
 332         (WTF::Unicode::isDigit):
 333         (WTF::Unicode::mirroredChar):
 334         (WTF::Unicode::compare):
 335
 336 2006-08-17  Nikolas Zimmermann  <zimmermann@kde.org>
 337
 338         Reviewed by Eric. Landed by rwlbuis.
 339
 340         Fixes: http://bugzilla.opendarwin.org/show_bug.cgi?id=10464
 341         Offer a cmake build system for Qt platform.
 342
 343         * CMakeLists.txt: Added.
 344         * pcre/CMakeLists.txt: Added.
 345
 346 2006-08-17  Anders Carlsson  <acarlsson@apple.com>
 347
 348         Reviewed by Maciej.
 349
 350         * bindings/npapi.h:
 351         Fix ifdef.
 352
 353 2006-08-15  Steve Falkenburg  <sfalken@apple.com>
 354
 355         Reviewed by mjs.
 356
 357         Build fix.
 358
 359         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
 360         * wtf/Assertions.h:
 361
 362 2006-08-15  Mark Rowe  <opendarwin.org@bdash.net.nz>
 363
 364         Reviewed by Tim H.
 365
 366         Build fix:  DWARF and -gfull are incompatible with symbol separation.
 367
 368         * JavaScriptCore.xcodeproj/project.pbxproj:
 369
 370 2006-08-15  Mark Rowe  <opendarwin.org@bdash.net.nz>
 371
 372         Reviewed by Tim H.
 373
 374         http://bugzilla.opendarwin.org/show_bug.cgi?id=10394
 375         Bug 10394: WebKit Release and Production configurations should enable dead code stripping
 376
 377         * JavaScriptCore.xcodeproj/project.pbxproj:
 378
 379 2006-08-15  Mark Rowe  <opendarwin.org@bdash.net.nz>
 380
 381         Reviewed by Tim H.
 382
 383         http://bugzilla.opendarwin.org/show_bug.cgi?id=10384
 384         Bug 10384: Switch to DWARF for Release configuration
 385
 386         * JavaScriptCore.xcodeproj/project.pbxproj:
 387
 388 2006-08-13  Maks Orlovich  <maksim@kde.org>
 389
 390         Reviewed (and tweaked a little) by Maciej.
 391
 392         - shrank the size of JSObject by 8 bytes and made the corresponding reduction to the cell size, resulting
 393         in a 1.2% speed improvement on JS iBench (and probably overall memory savings).
 394
 395         This was done by removing _scope and _internalValue data members
 396         from JSObject and moving them only to the subclasses that actually
 397         make use of them.
 398
 399         * kjs/object.cpp:
 400         (KJS::JSObject::mark): No need to mark scope or internal value here.
 401         * kjs/object.h:
 402         (KJS::JSObject::JSObject): Don't initialize them.
 403         * kjs/JSWrapperObject.cpp: Added. New base class for object types that
 404         wrap primitive values (Number, String, Boolean, Date).
 405         (KJS::JSWrapperObject::mark):
 406         * kjs/JSWrapperObject.h: Added.
 407         (KJS::JSWrapperObject::JSWrapperObject):
 408         (KJS::JSWrapperObject::internalValue):
 409         (KJS::JSWrapperObject::setInternalValue):
 410         * kjs/array_object.cpp:
 411         (ArrayPrototype::ArrayPrototype): Don't set useless internal value.
 412         * kjs/bool_object.cpp:
 413         (BooleanInstance::BooleanInstance): Inherit from JSWrapperObject.
 414         (BooleanProtoFunc::callAsFunction): Fixed to account for fact that not all
 415         JSObjects have an internal value.
 416         (BooleanObjectImp::construct): ditto.
 417         * kjs/bool_object.h:
 418         * kjs/collector.cpp: Lowered cell size to 48.
 419         (KJS::Collector::allocate): meaningless whitespace change
 420         * kjs/date_object.cpp:
 421         (KJS::DateInstance::DateInstance): Inherit from JSWrapperObject.
 422         (KJS::DateProtoFunc::callAsFunction): adjusted for move of internalValue
 423         (KJS::DateObjectImp::construct): ditto
 424         * kjs/date_object.h:
 425         * kjs/error_object.cpp:
 426         (ErrorPrototype::ErrorPrototype): don't set internal value
 427         * kjs/function.cpp: move _scope and related handling here
 428         (KJS::FunctionImp::mark): mark scope
 429         * kjs/function.h:
 430         (KJS::FunctionImp::scope): moved here from JSObject
 431         (KJS::FunctionImp::setScope): ditto
 432         * kjs/number_object.cpp:
 433         (NumberInstance::NumberInstance): inherit from JSWrapperObject
 434         (NumberProtoFunc::callAsFunction): adjusted
 435         (NumberObjectImp::construct): adjusted
 436         * kjs/number_object.h: shring RegExp-related objects a little
 437         * kjs/regexp_object.cpp:
 438         (RegExpPrototype::RegExpPrototype): Adjust for size tweaks
 439         (RegExpObjectImp::RegExpObjectImp): ditto
 440         * kjs/regexp_object.h:
 441         * kjs/string_object.cpp:
 442         (StringInstance::StringInstance): inherit from JSWrapperObject
 443         (StringProtoFunc::callAsFunction): adjusted
 444         * kjs/string_object.h:
 445         * JavaScriptCore.exp: Exported new methods as needed.
 446         * JavaScriptCore.xcodeproj/project.pbxproj: Added new files to build.
 447
 448 2006-08-04  Brady Eidson  <beidson@apple.com>
 449
 450         Reviewed by Geoff's rubber stamp
 451
 452         Fix a build break on Intel hardware causes by adapting stricter
 453         compiler warnings (-Wshorten-64-to-32)
 454
 455         * API/testapi.c:
 456         (assertEqualsAsNumber): manually cast some doubles to floats
 457         (main): ditto
 458
 459 2006-08-04  Sam Weinig  <sam.weinig@gmail.com>
 460
 461         Reviewed by Darin.
 462
 463         - patch for http://bugzilla.opendarwin.org/show_bug.cgi?id=10192
 464           Make WebCore (and friends) compile with -Wshorten-64-to-32
 465
 466           * Adds -Wshorten-64-to-32 flag to Xcode project.
 467           * Adds explicit casts where OK.
 468
 469         * API/JSNodeList.c:
 470         (JSNodeList_item):
 471         (JSNodeList_getProperty):
 472         * JavaScriptCore.xcodeproj/project.pbxproj:
 473
 474 2006-08-04  Adam Roben  <aroben@apple.com>
 475
 476         Reviewed by Anders.
 477
 478         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: Convert
 479         spaces to tabs
 480
 481 2006-08-03  Sam Weinig  <sam.weinig@gmail.com>
 482
 483         Reviewed by Darin.
 484
 485         - patch for http://bugzilla.opendarwin.org/show_bug.cgi?id=10176
 486           Make WebCore compile with -Wundef
 487
 488           * Adds -Wundef flag to Xcode project
 489           * Converts #ifs to #ifdef and #ifndefs where needed.
 490           * Added #define YYMAXDEPTH 10000 in kjs/grammar.y
 491             to fix a warning from within Bison.
 492
 493         * JavaScriptCore.xcodeproj/project.pbxproj:
 494         * bindings/jni/jni_jsobject.cpp:
 495         (JavaJSObject::getSlot):
 496         (JavaJSObject::setSlot):
 497         * bindings/npapi.h:
 498         * bindings/objc/objc_class.mm:
 499         (KJS::Bindings::ObjcClass::methodsNamed):
 500         (KJS::Bindings::ObjcClass::fieldNamed):
 501         * bindings/objc/objc_instance.mm:
 502         (ObjcInstance::invokeMethod):
 503         * bindings/objc/objc_runtime.mm:
 504         (ObjcMethod::getMethodSignature):
 505         (ObjcField::name):
 506         (ObjcField::type):
 507         * kjs/grammar.y:
 508         * kjs/identifier.h:
 509
 510 2006-08-03  Anders Carlsson  <acarlsson@apple.com>
 511
 512         Reviewed by John Sullivan.
 513
 514        * wtf/HashSet.h:
 515         (WTF::::operator):
 516         Return *this in operator=
 517
 518 2006-08-03  Adam Roben  <aroben@apple.com>
 519
 520         Reviewed by Anders.
 521
 522         - Fixed Windows build
 523
 524         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
 525         * wtf/MathExtras.h: Implement inline versions of these functions
 526         (nextafter):
 527         (nextafterf):
 528
 529 2006-08-02  Adam Roben  <aroben@apple.com>
 530
 531         Reviewed by Darin.
 532
 533         - Fixed build
 534
 535         * kjs/date_object.cpp:
 536         (KJS::formatTime):
 537
 538 2006-07-29  Darin Adler  <darin@apple.com>
 539
 540         - Removed tabs from these source files that still had them.
 541           We don't use them; that way source files look fine in editors
 542           that have tabs set to 8 spaces or to 4 spaces.
 543         - Removed allow-tabs Subversion property from the files too.
 544
 545         * bindings/NP_jsobject.cpp:
 546         * bindings/c/c_utility.cpp:
 547         * bindings/jni/jni_runtime.cpp:
 548         * bindings/jni/jni_utility.cpp:
 549         * bindings/objc/objc_utility.mm:
 550         * bindings/runtime.cpp:
 551         * bindings/runtime_method.cpp:
 552         * bindings/testbindings.cpp:
 553         * bindings/testbindings.mm:
 554         * kjs/date_object.cpp:
 555         * kjs/function.cpp:
 556         * kjs/list.cpp:
 557         * kjs/nodes.cpp:
 558         * kjs/nodes.h:
 559         * kjs/string_object.cpp:
 560         * kjs/ustring.cpp:
 561
 562 2006-07-29  Darin Adler  <darin@apple.com>
 563
 564         * tests/mozilla/expected.html: Update test results now that regress-185165.js
 565         is succeeding. I suspect Anders fix for bug 4620655 is the reason.
 566
 567 2006-07-29  Sam Weinig  <sam.weinig@gmail.com>
 568
 569         Reviewed by Darin.
 570
 571         - patch for http://bugzilla.opendarwin.org/show_bug.cgi?id=10080
 572           Adopt pedantic changes from the Unity project to improve
 573           cross-compiler compatibility
 574
 575           Changes include:
 576           * Removing trailing semicolon from namespace braces.
 577           * Removing trailing comma from last enum declaration.
 578           * Updating to match style guidelines.
 579           * Adding missing newline to the end of the file.
 580           * Turning on gcc warning for missing newline at the end of a source file
 581             (GCC_WARN_ABOUT_MISSING_NEWLINE in Xcode, -Wnewline in gcc).
 582           * Alphabetical sorting of Xcode source list files.
 583           * Replace use of non-portable variable-size array with Vector.
 584           * Use C-style comments instead of C++ comments in files that might
 585             be included by either C or C++ files.
 586
 587         * API/JSCallbackConstructor.cpp:
 588         (KJS::JSCallbackConstructor::construct):
 589         * API/JSCallbackFunction.cpp:
 590         (KJS::JSCallbackFunction::callAsFunction):
 591         * API/JSCallbackObject.cpp:
 592         (KJS::JSCallbackObject::construct):
 593         (KJS::JSCallbackObject::callAsFunction):
 594         * JavaScriptCore.xcodeproj/project.pbxproj:
 595         * JavaScriptCorePrefix.h:
 596         * bindings/jni/jni_class.cpp:
 597         (JavaClass::fieldNamed):
 598         * bindings/jni/jni_class.h:
 599         * bindings/jni/jni_instance.cpp:
 600         (JavaInstance::JavaInstance):
 601         (JavaInstance::valueOf):
 602         * bindings/jni/jni_objc.mm:
 603         (KJS::Bindings::dispatchJNICall):
 604         * bindings/jni/jni_runtime.cpp:
 605         (JavaParameter::JavaParameter):
 606         (JavaArray::JavaArray):
 607         * bindings/jni/jni_runtime.h:
 608         * bindings/jni/jni_utility.h:
 609         * bindings/objc/objc_instance.h:
 610         * bindings/runtime_array.h:
 611         * kjs/collector.h:
 612         * kjs/config.h:
 613         * kjs/ustring.cpp:
 614         * wtf/Platform.h:
 615
 616 2006-07-29  Mike Emmel  <mike.emmel@gmail.com>
 617
 618         Reviewed by Darin.
 619
 620         - fixes for Linux build
 621
 622         * JavaScriptCoreSources.bkl: Added new files to build, kjs/PropertyNameArray.cpp
 623         and kjs/testkjs.cpp, and removed old files.
 624
 625 2006-07-24  Dan Waylonis  <waylonis@google.com>
 626
 627         Reviewed and tweaked a bit by Darin.
 628
 629         - fix http://bugzilla.opendarwin.org/show_bug.cgi?id=9902
 630           jsNull and NSNull not properly converted between JS and ObjC
 631
 632         * bindings/objc/objc_utility.mm: (KJS::Bindings::convertObjcValueToValue):
 633         Added case for converting NSNull to jsNull.
 634
 635 2006-07-24  Rob Buis  <buis@kde.org>
 636
 637         Reviewed by Darin.
 638
 639         http://bugzilla.opendarwin.org/show_bug.cgi?id=4258
 640         Date().toString() only includes GMT offset, not timezone string
 641
 642         Use the info in tm_zone to append timezone abbreviation
 643         to Date().toString().
 644
 645         * kjs/date_object.cpp:
 646         (KJS::formatTime):
 647
 648 2006-07-24  Rob Buis  <buis@kde.org>
 649
 650         Reviewed by Darin.
 651
 652         http://bugzilla.opendarwin.org/show_bug.cgi?id=5257
 653         setYear() does not match FireFox/IE behavior
 654
 655         Make sure the right values end up in tm_year.
 656
 657         * kjs/date_object.cpp:
 658         (KJS::formatTime):
 659
 660 2006-07-23  Mark Rowe  <opendarwin.org@bdash.net.nz>
 661
 662         Reviewed by Maciej.
 663
 664         Bug 9686: [Drosera] Need the ability to break into Drosera on Javascript exceptions
 665         http://bugzilla.opendarwin.org/show_bug.cgi?id=9686
 666
 667         JavaScriptCore portion of the fix.
 668
 669         * JavaScriptCore.exp: Update symbol for change in argument type.
 670         * kjs/debugger.cpp:
 671         (Debugger::detach): Clear map of recent exceptions.
 672         (Debugger::hasHandledException): Track the most recent exception
 673         thrown by an interpreter.
 674         (Debugger::exception): Change exception argument to a JSValue.
 675         * kjs/debugger.h:
 676         * kjs/nodes.cpp:
 677         (Node::debugExceptionIfNeeded): Notify the debugger of an exception
 678         if it hasn't seen it before.
 679         (ThrowNode::execute): Notify the debugger that an exception is being thrown.
 680         * kjs/nodes.h:
 681
 682     2006-07-23  Geoffrey Garen  <ggaren@apple.com>
 683
 684         Patch by Eric Albert, reviewed by Darin and me.
 685
 686         - Fixed <rdar://problem/4645931> JavaScriptCore stack-scanning code
 687         crashes (Collector::markStackObjectsConservatively)
 688
 689         * bindings/jni/jni_jsobject.cpp: On 64bit systems, jint is a long, not an
 690         int.
 691         (JavaJSObject::getSlot):
 692         (JavaJSObject::setSlot):
 693         * kjs/collector.cpp:
 694         (KJS::Collector::markCurrentThreadConservatively): Use a pointer instead of
 695         an int as 'dummy,' because on LP64 systems, an int is not pointer-aligned,
 696         and we want to scan the stack for pointers.
 697         * JavaScriptCore.xcodeproj/project.pbxproj: After a tense cease-fire, the
 698         XCode war has started up again!
 699
 700 === Safari-521.20 ===
 701
 702 2006-07-21  Geoffrey Garen  <ggaren@apple.com>
 703
 704         Reviewed by Darin.
 705
 706         <rdar://problem/4507265> REGRESSION: overlays don't work on HousingMaps.com (Google Maps-based site)
 707
 708         - Added support for strings that masquerade as undefined. Currently used
 709         by WebCore to implement undetectable style.filter.
 710
 711         The name is a little long, but it's only used in one line of code, so I
 712         thought clarity should win over brevity.
 713
 714         * JavaScriptCore.exp:
 715         * JavaScriptCore.xcodeproj/project.pbxproj:
 716         * kjs/object.h:
 717         * kjs/string_object.h:
 718         (KJS::StringInstanceThatMasqueradesAsUndefined::StringInstanceThatMasqueradesAsUndefined):
 719         (KJS::StringInstanceThatMasqueradesAsUndefined::masqueradeAsUndefined):
 720         (KJS::StringInstanceThatMasqueradesAsUndefined::toBoolean):
 721
 722 === Safari-521.19 ===
 723
 724 2006-07-20  Steve Falkenburg  <sfalken@apple.com>
 725
 726         Fix the build
 727
 728         * kjs/function.cpp:
 729         (KJS::escapeStringForPrettyPrinting):
 730
 731 2006-07-19  Anders Carlsson  <acarlsson@apple.com>
 732
 733         Reviewed by Darin.
 734
 735         <rdar://problem/4620655> REGRESSION(10.4.7-10.5): preview button for a blogger.com post doesn't work
 736
 737         * kjs/nodes2string.cpp:
 738         (StringNode::streamTo):
 739         Return the escaped string.
 740
 741         (RegExpNode::streamTo):
 742         Use the correct syntax.
 743
 744         * kjs/function.cpp:
 745         (KJS::escapeStringForPrettyPrinting):
 746         * kjs/function.h:
 747         Add escape function which escapes a string for pretty-printing so it can be parsed again.
 748
 749         * wtf/unicode/icu/UnicodeIcu.h:
 750         (WTF::Unicode::isPrintableChar):
 751         New function.
 752
 753 2006-07-18  Maciej Stachowiak  <mjs@apple.com>
 754
 755         Reviewed by Adele Peterson.
 756
 757         <rdar://problem/4589530> REGRESSION: null character in JS string causes parse error (works in Tiger and in other browsers)
 758
 759         * kjs/lexer.cpp:
 760         (Lexer::shift):
 761         (Lexer::lex):
 762         (Lexer::record16):
 763         (Lexer::scanRegExp):
 764         * kjs/lexer.h:
 765
 766 2006-07-18  Tim Omernick  <timo@apple.com>
 767
 768         Reviewed by Tim Hatcher.
 769
 770         Removed a misleading comment; we recently added support for the NPNVPluginElementNPObject
 771         variable.
 772
 773         * bindings/npapi.h:
 774
 775 === Safari-521.18 ===
 776
 777 2006-07-18  Timothy Hatcher  <timothy@apple.com>
 778
 779         Made the following headers public:
 780
 781         * JavaScriptCore.h
 782         * JSBase.h
 783         * JSContextRef.h
 784         * JSObjectRef.h
 785         * JSStringRef.h
 786         * JSValueRef.h
 787
 788         * JavaScriptCore.xcodeproj/project.pbxproj:
 789
 790 2006-07-17  Geoffrey Garen  <ggaren@apple.com>
 791
 792         Reviewed by Maciej.
 793
 794         - Added automatic prototype creation for classes.
 795
 796         A class stores a weak reference to a prototype, which is cleared when
 797         the prototype is garbage collected, to avoid a reference cycle.
 798
 799         We now have an attributes field in JSClassDefinition, that currently is
 800         used only to override automatic prototype creation when you want to manage your
 801         own prototypes, but can be extended in the future for other nefarious purposes.
 802
 803         Similarly, we have JSObjectMake and JSObjectMakeWithPrototype, the latter
 804         allowing you to manage your own prototypes.
 805
 806         JSObjectMakeConstructor is more interesting now, able to make a constructor
 807         on your behalf if you just give it a class.
 808
 809         - Removed bogus old code from minidom.js.
 810
 811         - Tweaked the headerdocs.
 812
 813         - Added more GC testing, which caught some leaks, and tested more funny
 814         edge cases in lookup, which caught a lookup bug. Removed some testing
 815         we used to do with MyObject because it was redundant with the new, cool
 816         stuff.
 817
 818         While fixing the lookup bug I retracted this change:
 819
 820             "If a static setProperty callback returns 'false', to indicate that the
 821             property was not set, we no longer forward the set request up the class
 822             chain, because that's almost certainly not what the programmer expected."
 823
 824         Returning false when setting a static property is a little silly, but you can see
 825         it being useful when shadowing a base class's static properties, and, regardless
 826         of usefullness, this is the defined behavior of the setProperty callback.
 827
 828         - Plus a little ASCII art, for the kids.
 829
 830 2006-07-17  Timothy Hatcher  <timothy@apple.com>
 831
 832         Reviewed by Maciej.
 833
 834         <rdar://problem/4634874> WebScriptObject and WebUndefined are no longer defined by WebKit
 835
 836         Moves WebScriptObject and WebUndefined up to WebCore.
 837         This change does create an upwards-dependancy on WebScriptObject existing
 838         in the loaded process, but this code path in JavaScriptCore does not get used
 839         unless it is through WebKit/WebCore. Moving all of the binding code out of
 840         JavaScriptCore might make sense in the future.
 841
 842         * JavaScriptCore.exp:
 843         * JavaScriptCore.xcodeproj/project.pbxproj:
 844         * bindings/objc/WebScriptObject.h: Replaced.
 845         * bindings/objc/WebScriptObject.mm: Removed.
 846         * bindings/objc/WebScriptObjectPrivate.h: Removed.
 847         * bindings/objc/objc_class.h:
 848         * bindings/objc/objc_instance.h:
 849         * bindings/objc/objc_instance.mm:
 850         (ObjcInstance::~ObjcInstance):
 851         * bindings/objc/objc_runtime.h:
 852         * bindings/objc/objc_runtime.mm:
 853         (convertValueToObjcObject):
 854         * bindings/objc/objc_utility.mm:
 855         (KJS::Bindings::convertValueToObjcValue):
 856         (KJS::Bindings::convertObjcValueToValue):
 857         (KJS::Bindings::createObjcInstanceForValue):
 858
 859 2006-07-17  Darin Adler  <darin@apple.com>
 860
 861         * API/JSBase.h: Fix comment formatting where things used to be lined up but
 862         are now ragged. Got rid of spaces that attempted to line things up.
 863         * API/JSObjectRef.h: Ditto. Also add missing periods for a couple of comments.
 864
 865 2006-07-17  Geoffrey Garen  <ggaren@apple.com>
 866
 867         Reviewed by Maciej.
 868
 869         - Removed the exception parameter from the initialize callback and, by extension,
 870         JSObjectMake. We have never had a need for exceptions when iniitializing,
 871         so the parameter seemed likely to "get in the way."
 872
 873         Also, an exception in JavaScript is thrown in response to input --
 874         "invalid URL", "index not a number", etc., so it's the job of the
 875         constructor function, not the initialize method, to throw.
 876
 877         If initialize *really* wants to throw, it can communicate the throw to
 878         the constructor through the constructed object's private data (e.g., set
 879         it to NULL, signaling to the consntructor that initialization failed.)
 880
 881         - Added JSObjectMakeWithData, which enables a constructor to set private
 882         data on an object *before* it has been initialized. That way, the initialize
 883         methods can properly operate on the data.
 884
 885         * API/JSNode.c: Moved ref into the initialize method, for better encapsulation,
 886         now that it's possible.
 887         * API/JSNodeList.c: ditto
 888         * API/minidom.c:
 889         (main): Do more aggressive garbage collection to test ref/deref and
 890         initialize/finalize.
 891         * API/minidom.js: store childNodes in a temporary so it doesn't get re-created
 892         like a thousand times. This makes debugging ref/deref easier
 893
 894 2006-07-17  Geoffrey Garen  <ggaren@apple.com>
 895
 896         Reviewed by Maciej.
 897
 898         - Changed the initialize callback to run from least derived class (parent
 899         class) to most derived class. This enables C++ style initialization,
 900         and derived class overriding of member data.
 901
 902         - Added excpetion propopgation to JSObjectMake, to support initialize
 903         exceptions, and generally round out our policy of making function
 904         signatures as long as possible.
 905
 906         * API/JSCallbackObject.h: Use ExecState instead of ContextRef, cuz we're
 907         in C++ land now.
 908
 909 2006-07-17  Geoffrey Garen  <ggaren@apple.com>
 910
 911         Reviewed by Maciej.
 912
 913         - Changed JSObjectMakeConstructor to JSObjectMakeConstructorWithCallback,
 914         to match JSObjectMakeFunctionWithCallback.
 915
 916         - Added prototype parameter, so the generated constructor
 917         automatically works with hasInstance / instanceof
 918
 919         - Moved hasInstance implementation from InternalFunctionImp to JSObject
 920         so that subclasses can inherit it without inheriting function-related baggage.
 921         More refactoring here would be good, but this seems like a good short-term
 922         solution.
 923
 924         (KJS::JSCallbackFunction::implementsHasInstance): override and return false,
 925         because callback functions aren't constructors.
 926
 927 2006-07-17  Maciej Stachowiak  <mjs@apple.com>
 928
 929         Reviewed by Geoff.
 930
 931         - add a JSContextRef parameter to all JSValueRef, JSObjectRef, and JSContextRef operations;
 932         except JSObject{Get,Set}PrivateData which can be assumed to be simple pure accessors.
 933
 934         Also renamed the parameter "context" to "ctx" because it makes the code read better with this pervasive
 935         but usually uninteresting parameter.
 936
 937         * API/JSBase.cpp:
 938         (JSEvaluateScript):
 939         (JSCheckScriptSyntax):
 940         (JSGarbageCollect):
 941         * API/JSBase.h:
 942         * API/JSCallbackObject.cpp:
 943         (KJS::JSCallbackObject::JSCallbackObject):
 944         (KJS::JSCallbackObject::init):
 945         (KJS::JSCallbackObject::getOwnPropertySlot):
 946         (KJS::JSCallbackObject::put):
 947         (KJS::JSCallbackObject::deleteProperty):
 948         (KJS::JSCallbackObject::toNumber):
 949         (KJS::JSCallbackObject::toString):
 950         * API/JSContextRef.cpp:
 951         (JSGlobalContextCreate):
 952         (JSGlobalContextRetain):
 953         (JSGlobalContextRelease):
 954         (JSContextGetGlobalObject):
 955         * API/JSContextRef.h:
 956         * API/JSNode.c:
 957         (JSNodePrototype_appendChild):
 958         (JSNodePrototype_removeChild):
 959         (JSNodePrototype_replaceChild):
 960         (JSNode_getNodeType):
 961         (JSNode_getFirstChild):
 962         (JSNode_prototype):
 963         * API/JSNodeList.c:
 964         (JSNodeListPrototype_item):
 965         (JSNodeList_length):
 966         (JSNodeList_getProperty):
 967         (JSNodeList_prototype):
 968         * API/JSObjectRef.cpp:
 969         (JSObjectMake):
 970         (JSObjectMakeFunctionWithCallback):
 971         (JSObjectMakeConstructor):
 972         (JSObjectMakeFunction):
 973         (JSObjectGetPrototype):
 974         (JSObjectSetPrototype):
 975         (JSObjectHasProperty):
 976         (JSObjectGetProperty):
 977         (JSObjectSetProperty):
 978         (JSObjectGetPropertyAtIndex):
 979         (JSObjectSetPropertyAtIndex):
 980         (JSObjectDeleteProperty):
 981         (JSObjectIsFunction):
 982         (JSObjectCallAsFunction):
 983         (JSObjectIsConstructor):
 984         (JSObjectCallAsConstructor):
 985         (JSObjectCopyPropertyNames):
 986         * API/JSObjectRef.h:
 987         * API/JSStringRef.cpp:
 988         * API/JSValueRef.cpp:
 989         (JSValueGetType):
 990         (JSValueIsUndefined):
 991         (JSValueIsNull):
 992         (JSValueIsBoolean):
 993         (JSValueIsNumber):
 994         (JSValueIsString):
 995         (JSValueIsObject):
 996         (JSValueIsObjectOfClass):
 997         (JSValueIsEqual):
 998         (JSValueIsStrictEqual):
 999         (JSValueIsInstanceOfConstructor):
1000         (JSValueMakeUndefined):
1001         (JSValueMakeNull):
1002         (JSValueMakeBoolean):
1003         (JSValueMakeNumber):
1004         (JSValueMakeString):
1005         (JSValueToBoolean):
1006         (JSValueToNumber):
1007         (JSValueToStringCopy):
1008         (JSValueToObject):
1009         (JSValueProtect):
1010         (JSValueUnprotect):
1011         * API/JSValueRef.h:
1012         * API/minidom.c:
1013         (print):
1014         * API/testapi.c:
1015         (MyObject_getProperty):
1016         (MyObject_deleteProperty):
1017         (MyObject_callAsFunction):
1018         (MyObject_callAsConstructor):
1019         (MyObject_convertToType):
1020         (print_callAsFunction):
1021         (main):
1022
1023 2006-07-16  Geoffrey Garen  <ggaren@apple.com>
1024
1025         Approved by Maciej, RS by Beth.
1026
1027         JSObjectMakeFunction -> JSObjectMakeFunctionWithCallback
1028         JSObjectMakeFunctionWithBody -> JSObjectMakeFunction
1029
1030         because the latter is more common, and more fundamental, than the former.
1031
1032         * API/APICast.h:
1033         (toJS):
1034         * API/JSBase.h:
1035         * API/JSCallbackObject.cpp:
1036         (KJS::JSCallbackObject::getOwnPropertySlot):
1037         (KJS::JSCallbackObject::put):
1038         (KJS::JSCallbackObject::deleteProperty):
1039         (KJS::JSCallbackObject::getPropertyNames):
1040         (KJS::JSCallbackObject::staticValueGetter):
1041         (KJS::JSCallbackObject::staticFunctionGetter):
1042         * API/JSClassRef.cpp:
1043         (OpaqueJSClass::OpaqueJSClass):
1044         (OpaqueJSClass::~OpaqueJSClass):
1045         * API/JSClassRef.h:
1046         * API/JSObjectRef.cpp:
1047         (JSClassCreate):
1048         (JSObjectMakeFunctionWithCallback):
1049         (JSObjectMakeFunction):
1050         (OpaqueJSPropertyNameArray::OpaqueJSPropertyNameArray):
1051         (JSObjectCopyPropertyNames):
1052         * API/JSObjectRef.h:
1053         * API/minidom.c:
1054         (main):
1055         * API/testapi.c:
1056         (main):
1057         * ChangeLog:
1058         * JavaScriptCore.exp:
1059
1060 2006-07-16  Geoffrey Garen  <ggaren@apple.com>
1061
1062         Laughed at by Beth.
1063
1064         Replace __JS with OpaqueJS because the former, while used by CF, is
1065         a prefix that's triply-reserved by the compiler. (_* is reserved in global
1066         names, _[A-Z] is reserved in all names, and __ is reserved in all names
1067         in C++.)
1068
1069         Opaque is an alternative used by other Mac OS X framewokrs.
1070
1071         * API/APICast.h:
1072         (toJS):
1073         * API/JSBase.h:
1074         * API/JSCallbackObject.cpp:
1075         (KJS::JSCallbackObject::getOwnPropertySlot):
1076         (KJS::JSCallbackObject::put):
1077         (KJS::JSCallbackObject::deleteProperty):
1078         (KJS::JSCallbackObject::getPropertyNames):
1079         (KJS::JSCallbackObject::staticValueGetter):
1080         (KJS::JSCallbackObject::staticFunctionGetter):
1081         * API/JSClassRef.cpp:
1082         (OpaqueJSClass::OpaqueJSClass):
1083         (OpaqueJSClass::~OpaqueJSClass):
1084         * API/JSClassRef.h:
1085         * API/JSObjectRef.cpp:
1086         (JSClassCreate):
1087         (OpaqueJSPropertyNameArray::OpaqueJSPropertyNameArray):
1088         (JSObjectCopyPropertyNames):
1089
1090 2006-07-16  Darin Adler  <darin@apple.com>
1091
1092         - try to fix Windows build
1093
1094         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
1095         Added some recently added files, removed some recently removed.
1096
1097 2006-07-16  Geoffrey Garen  <ggaren@apple.com>
1098
1099         Reviewed by Maciej.
1100
1101         - Change getProperty* to return undefined, rather than NULL, for missing
1102         properties, since that's what the spec says. Also added exception out
1103         parameters to the *Index functions, because they can call through to the
1104         regular functions, which can throw for custom objects.
1105
1106         * API/JSObjectRef.cpp:
1107         (JSObjectGetProperty):
1108         (JSObjectGetPropertyAtIndex):
1109         (JSObjectSetPropertyAtIndex):
1110         * API/JSObjectRef.h:
1111         * API/testapi.c:
1112         (main):
1113
1114 2006-07-16  Geoffrey Garen  <ggaren@apple.com>
1115
1116         Reviewed by Maciej.
1117
1118         - Properly document and handle NULL callbacks for static properties. We
1119         throw an exception in any case other than a ReadOnly property with a NULL
1120         setProperty callback, because a NULL callback almost certainly indicates
1121         a programming error. Also throw an exception if hasProperty returns true
1122         for a property that getProperty can't get.
1123
1124         - If a static setProperty callback returns 'false', to indicate that the
1125         property was not set, we no longer forward the set request up the class
1126         chain, because that's almost certainly not what the programmer expected.
1127
1128         * API/JSCallbackObject.cpp:
1129         (KJS::JSCallbackObject::getOwnPropertySlot):
1130         (KJS::JSCallbackObject::put):
1131         (KJS::JSCallbackObject::staticValueGetter):
1132         (KJS::JSCallbackObject::staticFunctionGetter):
1133         (KJS::JSCallbackObject::callbackGetter):
1134         * API/JSObjectRef.h:
1135         * API/minidom.js:
1136         * API/testapi.c:
1137         (MyObject_hasProperty):
1138         * API/testapi.js:
1139
1140 2006-07-16  Geoffrey Garen  <ggaren@apple.com>
1141
1142         Reviewed by Maciej.
1143
1144         - Added names to functions.
1145
1146         - Removed GetPrivate/SetPrivate from callbackFunctions and callbackConstructors.
1147         The private data idiom is that a JS object stores its native implementation
1148         as private data. For functions and constructors, the native implementation is nothing
1149         more than the callback they already store, so supporting private data, too,
1150         confuses the idiom. If you *really* want, you can still create a custom
1151         function with private data.
1152
1153         * API/JSCallbackConstructor.cpp:
1154         * API/JSCallbackConstructor.h:
1155         * API/JSCallbackFunction.cpp:
1156         (KJS::JSCallbackFunction::JSCallbackFunction):
1157         * API/JSCallbackFunction.h:
1158         * API/JSCallbackObject.cpp:
1159         (KJS::JSCallbackObject::staticFunctionGetter):
1160         * API/JSObjectRef.cpp:
1161         (JSObjectMakeFunction):
1162         (JSObjectMakeFunctionWithBody):
1163         (JSObjectGetPrivate):
1164         (JSObjectSetPrivate):
1165         * API/JSObjectRef.h:
1166         * API/minidom.c:
1167         (main):
1168         * API/testapi.c:
1169         (main):
1170
1171 2006-07-15  Maciej Stachowiak  <mjs@apple.com>
1172
1173         Reviewed by Darin.
1174
1175         - switch property lists to be vector+set of Identifiers instead of list of References
1176
1177         This has the following benefits:
1178
1179         - no duplicates in property lists
1180         - simplifies API calls
1181         - probably more efficient, since linked list is gone
1182         - entirely removed Reference, ReferenceList and ProtectedReference types from the API
1183
1184         * kjs/PropertyNameArray.cpp: Added.
1185         (KJS::PropertyNameArray::add): Check set, if not already there, add to
1186         vector.
1187         * kjs/PropertyNameArray.h: Added.
1188         (KJS::PropertyNameArray::PropertyNameArray): Newly added type, combines
1189         a set and a vector to make a unique but ordered list of identifiers.
1190         (KJS::PropertyNameArray::begin): ditto
1191         (KJS::PropertyNameArray::end): ditto
1192         (KJS::PropertyNameArray::size): ditto
1193         (KJS::PropertyNameArray::operator[]): ditto
1194         * kjs/array_instance.h:
1195         * kjs/array_object.cpp:
1196         (ArrayInstance::getPropertyNames): renamed from getPropertyList, updated
1197         for PropertyNameArray
1198         (ArrayInstance::setLength): updated for PropertyNameArray
1199         (ArrayInstance::pushUndefinedObjectsToEnd): ditto
1200         * kjs/nodes.cpp:
1201         (ForInNode::execute): updated for PropertyNameArray
1202         * kjs/nodes.h:
1203         * kjs/object.cpp:
1204         (KJS::JSObject::getPropertyNames): renamed from getPropertyList, updated
1205         for PropertyNameArray
1206         * kjs/object.h:
1207         * kjs/property_map.cpp:
1208         (KJS::PropertyMap::getEnumerablePropertyNames): updated for PropertyNameArray
1209         (KJS::PropertyMap::getSparseArrayPropertyNames): ditto
1210         * kjs/property_map.h:
1211         * kjs/protected_reference.h: Removed.
1212         * kjs/reference.cpp: Removed.
1213         * kjs/reference.h: Removed.
1214         * kjs/reference_list.cpp: Removed.
1215         * kjs/reference_list.h: Removed.
1216         * kjs/scope_chain.cpp:
1217         (KJS::ScopeChain::print): Use PropertyNamesArray instead of ReferenceList.
1218         * kjs/string_object.cpp:
1219         (StringInstance::getPropertyNames): Updated for new approach.
1220         * kjs/string_object.h:
1221         * kjs/ustring.h:
1222         * API/APICast.h:
1223         (toJS): Added overload for PropertyNameAccumulatorRef / PropertyNameArray*
1224         (toRef): ditto
1225         * API/JSBase.h:
1226         * API/JSCallbackObject.cpp:
1227         (KJS::JSCallbackObject::getPropertyNames): Fixed for new API.
1228         * API/JSCallbackObject.h:
1229         * API/JSObjectRef.cpp:
1230         (__JSPropertyNameArray::__JSPropertyNameArray): Type used for a publicly vended
1231         JSPropertyNameArrayRef.
1232         (JSObjectCopyPropertyNames): New API call - renamed / refactored from
1233         JSObjectCreatePropertyList
1234         (JSPropertyNameArrayRetain): new retain call for JSPropertyNameArray.
1235         (JSPropertyNameArrayRelease): new release call for - " -.
1236         (JSPropertyNameArrayGetCount): Instead of having to use a stateful enumerator you
1237         can now get the count and items in any order.
1238         (JSPropertyNameArrayGetNameAtIndex): See above.
1239         (JSPropertyNameAccumulatorAddName): What you add properties to is now an opaque
1240         accumulator object.
1241         * API/JSObjectRef.h: Prototyped new functions, removed old ones
1242         * JavaScriptCore.exp: Updated exported symbols.
1243         * JavaScriptCore.xcodeproj/project.pbxproj: Added new files, removed old.
1244         * API/testapi.c:
1245         (MyObject_getPropertyNames): Renamed / fixed callback to fit new paradigm.
1246         (main): Updated for new API.
1247
1248 2006-07-15  Darin Adler  <darin@apple.com>
1249
1250         - oops, missed a few more arrays that had to be const
1251
1252         * API/JSNode.c:
1253         (JSNodePrototype_appendChild): Added const.
1254         (JSNodePrototype_removeChild): Ditto.
1255         (JSNodePrototype_replaceChild): Ditto.
1256         (JSNode_construct): Ditto.
1257         * API/JSNodeList.c:
1258         (JSNodeListPrototype_item): Ditto.
1259         * API/JSObjectRef.cpp:
1260         (JSObjectMakeFunctionWithBody): Ditto.
1261         (JSObjectCallAsFunction): Ditto.
1262         (JSObjectCallAsConstructor): Ditto.
1263         * API/minidom.c:
1264         (print): Ditto.
1265         * API/testapi.c:
1266         (MyObject_callAsFunction): Ditto.
1267         (MyObject_callAsConstructor): Ditto.
1268         (print_callAsFunction): Ditto.
1269         (myConstructor_callAsConstructor): Ditto.
1270
1271 2006-07-15  Darin Adler  <darin@apple.com>
1272
1273         Reviewed by Maciej.
1274
1275         * API/JSNode.h: Made an array parameter const.
1276         * API/JSObjectRef.h: Made array parameters const. Fixed a comment.
1277
1278 2006-07-15  Geoffrey Garen  <ggaren@apple.com>
1279
1280         Reviewed by Maciej.
1281
1282         - JSObjectMakeFunctionWithBody includes a function name and named parameters now.
1283
1284         * API/JSObjectRef.cpp:
1285         (JSObjectMakeFunctionWithBody):
1286         * API/JSObjectRef.h:
1287         * API/testapi.c:
1288         (assertEqualsAsUTF8String): More informative failure reporting.
1289         (main): Test more function cases.
1290
1291 2006-07-15  Geoffrey Garen  <ggaren@apple.com>
1292
1293         Reviewed by Maciej.
1294
1295         - Moved the arguments passed to JSClassCreate into a single structure,
1296         called JSClassDefinition. This will enable easier structure
1297         migration/versioning in the future, if necessary.
1298
1299         - Added support for class names.
1300
1301         - kJSClassDefinitionNull replaces kJSObjectCallbacksNone.
1302
1303         - JSClass is becoming a fairly complex struct, so I migrated all of its
1304         implementation other than reference counting to the sruct.
1305
1306         - Also moved JSClass* functions in the API to JSObjectRef.cpp, since they're
1307         declared in JSObjectRef.h
1308
1309         - Also added some more informative explanation to the class structure doc.
1310
1311 2006-07-15  Darin Adler  <darin@apple.com>
1312
1313         Reviewed by Geoff.
1314
1315         - fix http://bugzilla.opendarwin.org/show_bug.cgi?id=8395
1316           <rdar://problem/4613467>
1317           REGRESSION: RegEx seems broken for hex escaped non breaking space
1318
1319         Test: fast/js/regexp-extended-characters-more.html
1320
1321         * pcre/pcre_exec.c:
1322         (match): Got rid of utf16Length local variable to guarantee there's no
1323         extra stack usage in recursive calls. Fixed two places in the PCRE_UTF16
1324         code that were using the length variable, which is the UTF-8 length of
1325         a character in the pattern, to move in the UTF-16 subject string. Instead
1326         they hardcode lengths of 1 and 2 since the code already handles BMP
1327         characters and surrogate pairs separately. Also fixed some DPRINTF so
1328         I could compile with DEBUG on.
1329         (pcre_exec): Changed a place that was checking for multibyte characters
1330         in the subject string to use ISMIDCHAR. Instead it was using hardcoded
1331         logic that was right for UTF-8 but wrong for UTF-16.
1332
1333         * pcre/pcre_compile.c: (pcre_compile2): Fixed a DPRINTF so I could compile
1334         with DEBUG on.
1335
1336 2006-07-14  Geoffrey Garen  <ggaren@apple.com>
1337
1338         RS by Maciej.
1339
1340         Global replace in the API of argc/argv with argumentCount/arguments.
1341
1342 2006-07-14  Geoffrey Garen  <ggaren@apple.com>
1343
1344         Reviewed by Maciej.
1345
1346         - Finalized exception handling in the API.
1347
1348         setProperty can throw because it throws for built-in arrays. getProperty
1349         and deleteProperty can throw because setProperty can throw and we want
1350         to be consistent, and also because they seem like "actions." callAsFunction,
1351         callAsConstructor, and hasInstance can throw, because they caan throw for
1352         all built-ins.
1353
1354         toBoolean can't throw because it's defined that way in the spec.
1355
1356         - Documented that toBoolean and toObject can't be overridden by custom
1357         objects because they're defined that way in the spec.
1358
1359 === Safari-521.17 ===
1360
1361 2006-07-14  Geoffrey Garen  <ggaren@apple.com>
1362
1363         Reviewed by Maciej.
1364
1365         - Implemented ref-counting of JSContexts by splitting into two datatypes:
1366         JSGlobalContext, which you can create/retain/release, and JSContext, which
1367         you can't.
1368
1369         Internally, you retain a JSGlobalContext/ExecState by retaining its
1370         interpreter, which, in the case of a global ExecState, owns it.
1371
1372         - Also made ~Interpreter() protected to catch places where Interpreter
1373         is manually deleted. (Can't make it private because some crazy fool
1374         decided it would be a good idea to subclass Interpreter in other frameworks.
1375         I pity da fool.)
1376
1377         * API/APICast.h:
1378         (toJS): Added cast for new JSGlobalContext
1379         * API/JSStringRef.h: Changed vague "you must" language to more specific
1380         (but, ultimately, equally vague) "behavior is undefined if you don't"
1381         language.
1382         (KJS::Interpreter::Interpreter): Factored more common initialization into
1383         init()
1384         * kjs/interpreter.h:
1385         (KJS::Interpreter::ref): new
1386         (KJS::Interpreter::deref): new
1387         (KJS::Interpreter::refCount): new
1388         * kjs/testkjs.cpp:
1389         (doIt): Ref-count the interpreter.
1390
1391 2006-07-14  Maciej Stachowiak  <mjs@apple.com>
1392
1393         Reviewed by Geoff.
1394
1395         - removed bool return value from JSObjectSetProperty, since it is inefficient and
1396         also doesn't work quite right
1397         - added JSObjectGetPropertyAtIndex and JSObjectSetPropertyAtIndex
1398
1399         * API/JSObjectRef.cpp:
1400         (JSObjectSetProperty): Removed return value and canPut stuff.
1401         (JSObjectGetPropertyAtIndex): Added.
1402         (JSObjectSetPropertyAtIndex): Added.
1403         * API/JSObjectRef.h: Prototyped and documented new functions.
1404
1405 2006-07-14  Geoffrey Garen  <ggaren@apple.com>
1406
1407         Reviewed by Beth.
1408
1409         Moved JSCheckScriptSyntax, JSEvaluateScript, and JSGarbageCollect into
1410         JSBase.h/.cpp. They don't belong in the value-specific or context-specific
1411         files because they're not part of the value or context implementations.
1412
1413         * API/JSBase.h:
1414         * API/JSContextRef.cpp:
1415         (JSContextGetGlobalObject):
1416         * API/JSContextRef.h:
1417         * API/JSValueRef.cpp:
1418         (JSValueUnprotect):
1419         * API/JSValueRef.h:
1420         * JavaScriptCore.xcodeproj/project.pbxproj:
1421
1422 2006-07-13  Timothy Hatcher  <timothy@apple.com>
1423
1424         Reviewed by Maciej.
1425
1426         Moved JavaScriptCore to be a public framework.
1427
1428         * JavaScriptCore.xcodeproj/project.pbxproj:
1429
1430 2006-07-13  Mark Rowe  <opendarwin.org@bdash.net.nz>
1431
1432         Reviewed by Geoffrey.
1433
1434         http://bugzilla.opendarwin.org/show_bug.cgi?id=9742
1435         Bug 9742: REGRESSION: WebKit hangs when loading <http://www.vtbook.com>
1436
1437         * kjs/value.h:
1438         (KJS::JSValue::getUInt32): Only types tagged as numeric can be converted to UInt32.
1439
1440 2006-07-13  Geoffrey Garen  <ggaren@apple.com>
1441
1442         Pleasing to Maciej.
1443
1444         - Renamed JSEvaluate -> JSEvaluateScript, JSCheckSyntax -> JSCheckScriptSyntax
1445         - Added exception out parameters to JSValueTo* and JSValueIsEqual because
1446         they can throw
1447         - Removed JSObjectGetDescription because it's useless and vague, and
1448         JSValueToString/JSValueIsObjectOfClass do a better job, anyway
1449         - Clarified comments about "IsFunction/Constructor" to indicate that they
1450         are true of all functions/constructors, not just those created by JSObjectMake*
1451
1452 2006-07-12  Geoffrey Garen  <ggaren@apple.com>
1453
1454         RS by Beth.
1455
1456         Finished previously approved JSInternalString -> JSString conversion
1457         by renaming the files.
1458
1459         * API/JSCallbackObject.cpp:
1460         * API/JSInternalStringRef.cpp: Removed.
1461         * API/JSInternalStringRef.h: Removed.
1462         * API/JSStringRef.cpp: Added.
1463         * API/JSStringRef.h: Added.
1464         * API/JavaScriptCore.h:
1465         * JavaScriptCore.xcodeproj/project.pbxproj:
1466
1467 2006-07-12  Geoffrey Garen  <ggaren@apple.com>
1468
1469         Reviewed by Maciej.
1470
1471         - Removed context and exception parameters from JSObjectGetPropertyEnumerator,
1472         removing the spurious use of ExecState inside JavaScriptCore that made
1473         us think this was necessary in the first place.
1474
1475         (StringInstance::getPropertyList): Use getString instead of toString because
1476         we know we're dealing with a string -- we put it there in the first place.
1477         While we're at it, store the string's size instead of retrieving it each time
1478         through the loop, to avoid the unnecessary killing of puppies.
1479         * kjs/string_object.h:
1480
1481 2006-07-12  Maciej Stachowiak  <mjs@apple.com>
1482
1483         Reviewed by Geoff.
1484
1485         - add handling of hasInstance callback for API objects
1486
1487         * API/JSCallbackObject.cpp:
1488         (KJS::JSCallbackObject::implementsHasInstance): Check if callback is present.
1489         (KJS::JSCallbackObject::hasInstance): Invoke appropriate callback.
1490         * API/JSCallbackObject.h:
1491         * API/JSClassRef.cpp:
1492         * API/JSObjectRef.h:
1493         * API/testapi.c:
1494         (MyObject_hasInstance): Test case; should match what construct would do.
1495         * API/testapi.js:
1496
1497 2006-07-11  Geoffrey Garen  <ggaren@apple.com>
1498
1499         Reviewed by Maciej.
1500
1501         - Implemented a vast number of renames and comment clarifications
1502         suggested during API review.
1503
1504         JSInternalString -> JSString
1505         JS*Make -> JSValueMake*, JSObjectMake*
1506         JSTypeCode -> JSType
1507         JSValueIsInstanceOf -> JSValueIsInstanceOfConstructor (reads strangely well in client code)
1508         JSGC*Protect -> JSValue*Protect
1509         JS*Callback -> JSObject*Callback
1510         JSGetPropertyListCallback -> JSObjectAddPropertiesToListCallback
1511         JSPropertyEnumeratorGetNext -> JSPropertyEnumeratorGetNextName
1512         JSString* ->
1513             JSStringCreateWithUTF8CString, JSStringGetUTF8CString,
1514             JSStringGetMaximumUTF8CStringSize JSStringIsEqualToUTF8CString,
1515             JSStringCreateWithCFString, JSStringCopyCFString, JSStringCreateWithCharacters.
1516
1517         - Changed functions taking a JSValue out arg and returning a bool indicating
1518         whether it was set to simply return a JSValue or NULL.
1519
1520         - Removed JSStringGetCharacters because it's more documentation than code,
1521         and it's just a glorified memcpy built on existing API functionality.
1522
1523         - Moved standard library includes into the headers that actually require them.
1524
1525         - Standardized use of the phrase "Create Rule."
1526
1527         - Removed JSLock from make functions that don't allocate.
1528
1529         - Added exception handling to JSValueToBoolean, since we now allow
1530         callback objects to throw exceptions upon converting to boolean.
1531
1532         - Renamed JSGCCollect to JSGarbageCollect.
1533
1534 2006-07-10  Geoffrey Garen  <ggaren@apple.com>
1535
1536         Reviewed by Darin.
1537
1538         - Changed public header includes to the <JavaScriptCore/ style.
1539         - Changed instances of 'buffer' to 'string' since we decided on
1540         JSInternalString instead of JSStringBuffer.
1541
1542         * API/JSContextRef.h:
1543         * API/JSInternalStringRef.cpp:
1544         (JSStringMake):
1545         (JSInternalStringRetain):
1546         (JSInternalStringRelease):
1547         (JSValueCopyStringValue):
1548         (JSInternalStringGetLength):
1549         (JSInternalStringGetCharactersPtr):
1550         (JSInternalStringGetCharacters):
1551         (JSInternalStringGetMaxLengthUTF8):
1552         (JSInternalStringGetCharactersUTF8):
1553         (CFStringCreateWithJSInternalString):
1554         * API/JSInternalStringRef.h:
1555         * API/JSNode.c:
1556         (JSNodePrototype_appendChild):
1557         (JSNode_getNodeType):
1558         * API/JSObjectRef.cpp:
1559         (JSObjectCallAsConstructor):
1560         * API/JSValueRef.h:
1561         * API/JavaScriptCore.h:
1562         * API/minidom.c:
1563         (main):
1564         (print):
1565         * API/testapi.c:
1566         (MyObject_getPropertyList):
1567         (myConstructor_callAsConstructor):
1568         (main): I noticed that we were prematurely releasing some string buffers,
1569         so I moved their release calls to the end of main(). I got rid of 'Buf' in *Buf
1570         (sometimes changing to 'IString', when necessary to differentiate a variable)
1571         to match the buffer->string change.
1572
1573 === Safari-521.16 ===
1574
1575 2006-07-10  Darin Adler  <darin@apple.com>
1576
1577         * kjs/value.cpp: (KJS::JSValue::toInt32Inline): Added inline keyword one more place.
1578         Just in case.
1579
1580 2006-07-10  Darin Adler  <darin@apple.com>
1581
1582         - fix the release build
1583
1584         * kjs/value.h:
1585         * kjs/value.cpp:
1586         (KJS::JSValue::toInt32Inline): Move the code here to an inline.
1587         (KJS::JSValue::toInt32): Call the inline from both overloaded toInt32 functions.
1588
1589 2006-07-10  David Kilzer  <ddkilzer@kilzer.net>
1590
1591         Reviewed by Darin.
1592
1593         - fix http://bugzilla.opendarwin.org/show_bug.cgi?id=9179
1594           Implement select.options.add() method
1595
1596         * JavaScriptCore.exp: Added overloaded KJS::JSValue::toInt32() method.
1597         * JavaScriptCore.xcodeproj/project.pbxproj: Altered attributes metadata for
1598         kjs/value.h to make it available as a forwarded header.
1599         * kjs/lookup.h:
1600         (KJS::lookupPut): Extracted a lookupPut() method from the existing lookupPut() method.
1601         The new method returns a boolean value if no entry is found in the lookup table.
1602         * kjs/value.cpp:
1603         (KJS::JSValue::toInt32): Overloaded toInt32() method with boolean "Ok" argument.
1604         * kjs/value.h: Ditto.
1605
1606 2006-07-10  Geoffrey Garen  <ggaren@apple.com>
1607
1608         No review necessary. Removed bogus file I accidentally checked in before.
1609
1610         * API/JSInternalSringRef.h: Removed.
1611
1612 2006-07-10  Geoffrey Garen  <ggaren@apple.com>
1613
1614         Reviewed by Darin.
1615
1616         Added exception out parameter to API object callbacks, removed semi-bogus
1617         JSContext(.*)Exception functions.
1618
1619         To make these calls syntactically simple, I added an exceptionSlot()
1620         method to the ExecState class, which provides a JSValue** slot in which to
1621         store a JSValue* exception.
1622
1623         * API/APICast.h:
1624         (toRef):
1625         * API/JSCallbackConstructor.cpp:
1626         (KJS::JSCallbackConstructor::construct):
1627         * API/JSCallbackFunction.cpp:
1628         (KJS::JSCallbackFunction::callAsFunction):
1629         * API/JSCallbackObject.cpp:
1630         (KJS::JSCallbackObject::init):
1631         (KJS::JSCallbackObject::getOwnPropertySlot):
1632         (KJS::JSCallbackObject::put):
1633         (KJS::JSCallbackObject::deleteProperty):
1634         (KJS::JSCallbackObject::construct):
1635         (KJS::JSCallbackObject::callAsFunction):
1636         (KJS::JSCallbackObject::getPropertyList):
1637         (KJS::JSCallbackObject::toBoolean):
1638         (KJS::JSCallbackObject::toNumber):
1639         (KJS::JSCallbackObject::toString):
1640         (KJS::JSCallbackObject::staticValueGetter):
1641         (KJS::JSCallbackObject::callbackGetter):
1642         * API/JSContextRef.cpp:
1643         (JSCheckSyntax):
1644         * API/JSContextRef.h:
1645         * API/JSNode.c:
1646         (JSNodePrototype_appendChild):
1647         (JSNodePrototype_removeChild):
1648         (JSNodePrototype_replaceChild):
1649         (JSNode_getNodeType):
1650         (JSNode_getChildNodes):
1651         (JSNode_getFirstChild):
1652         (JSNode_construct):
1653         * API/JSNode.h:
1654         * API/JSNodeList.c:
1655         (JSNodeListPrototype_item):
1656         (JSNodeList_length):
1657         (JSNodeList_getProperty):
1658         * API/JSObjectRef.h:
1659         * API/minidom.c:
1660         (print):
1661         * API/testapi.c:
1662         (MyObject_initialize):
1663         (MyObject_hasProperty):
1664         (MyObject_getProperty):
1665         (MyObject_setProperty):
1666         (MyObject_deleteProperty):
1667         (MyObject_getPropertyList):
1668         (MyObject_callAsFunction):
1669         (MyObject_callAsConstructor):
1670         (MyObject_convertToType):
1671         (print_callAsFunction):
1672         (myConstructor_callAsConstructor):
1673         (main):
1674         * JavaScriptCore.exp:
1675         * kjs/ExecState.h:
1676         (KJS::ExecState::exceptionHandle):
1677
1678 2006-07-10  Geoffrey Garen  <ggaren@apple.com>
1679
1680         Reviewed by Darin.
1681
1682         Improved type safety by implementing opaque JSValue/JSObject typing through
1683         abuse of 'const', not void*. Also fixed an alarming number of bugs
1684         exposed by this new type safety.
1685
1686         I made one design change in JavaScriptCore, which is that the JSObject
1687         constructor should take a JSValue* as its prototype argument, not a JSObject*,
1688         since we allow the prototype to be any JSValue*, including jsNull(), for
1689         example.
1690
1691         * API/APICast.h:
1692         (toJS):
1693         * API/JSBase.h:
1694         * API/JSCallbackConstructor.cpp:
1695         (KJS::JSCallbackConstructor::construct):
1696         * API/JSCallbackFunction.cpp:
1697         (KJS::JSCallbackFunction::callAsFunction):
1698         * API/JSCallbackObject.cpp:
1699         (KJS::JSCallbackObject::JSCallbackObject):
1700         (KJS::JSCallbackObject::getOwnPropertySlot):
1701         (KJS::JSCallbackObject::put):
1702         (KJS::JSCallbackObject::construct):
1703         (KJS::JSCallbackObject::callAsFunction):
1704         (KJS::JSCallbackObject::staticFunctionGetter):
1705         * API/JSCallbackObject.h:
1706         * API/JSContextRef.cpp:
1707         (JSEvaluate):
1708         * API/JSNode.c:
1709         (JSNodePrototype_appendChild):
1710         (JSNodePrototype_removeChild):
1711         (JSNodePrototype_replaceChild):
1712         * API/JSObjectRef.cpp:
1713         (JSObjectMake):
1714         (JSFunctionMakeWithBody):
1715         (JSObjectGetProperty):
1716         (JSObjectCallAsFunction):
1717         (JSObjectCallAsConstructor):
1718         * API/JSObjectRef.h:
1719         * API/testapi.c:
1720         (main):
1721         * ChangeLog:
1722         * kjs/object.h:
1723         (KJS::JSObject::JSObject):
1724
1725 2006-07-10  Geoffrey Garen  <ggaren@apple.com>
1726
1727         Approved by Maciej, Darin.
1728
1729         Renamed JSStringBufferRef to JSInternalStringRef. "Internal string" means the
1730         JavaScript engine's internal string representation, which is the most
1731         low-level and efficient representation to use when interfacing with JavaScript.
1732
1733         * API/APICast.h:
1734         (toJS):
1735         (toRef):
1736         * API/JSBase.h:
1737         * API/JSCallbackObject.cpp:
1738         (KJS::JSCallbackObject::getOwnPropertySlot):
1739         (KJS::JSCallbackObject::put):
1740         (KJS::JSCallbackObject::deleteProperty):
1741         (KJS::JSCallbackObject::staticValueGetter):
1742         (KJS::JSCallbackObject::callbackGetter):
1743         * API/JSContextRef.cpp:
1744         (JSEvaluate):
1745         (JSCheckSyntax):
1746         * API/JSContextRef.h:
1747         * API/JSInternalStringRef.cpp: Added.
1748         (JSStringMake):
1749         (JSInternalStringCreate):
1750         (JSInternalStringCreateUTF8):
1751         (JSInternalStringRetain):
1752         (JSInternalStringRelease):
1753         (JSValueCopyStringValue):
1754         (JSInternalStringGetLength):
1755         (JSInternalStringGetCharactersPtr):
1756         (JSInternalStringGetCharacters):
1757         (JSInternalStringGetMaxLengthUTF8):
1758         (JSInternalStringGetCharactersUTF8):
1759         (JSInternalStringIsEqual):
1760         (JSInternalStringIsEqualUTF8):
1761         (JSInternalStringCreateCF):
1762         (CFStringCreateWithJSInternalString):
1763         * API/JSInternalStringRef.h: Added.
1764         * API/JSNode.c:
1765         (JSNodePrototype_appendChild):
1766         (JSNode_getNodeType):
1767         (JSNode_getChildNodes):
1768         (JSNode_getFirstChild):
1769         * API/JSNodeList.c:
1770         (JSNodeList_length):
1771         (JSNodeList_getProperty):
1772         * API/JSObjectRef.cpp:
1773         (JSFunctionMakeWithBody):
1774         (JSObjectGetDescription):
1775         (JSObjectHasProperty):
1776         (JSObjectGetProperty):
1777         (JSObjectSetProperty):
1778         (JSObjectDeleteProperty):
1779         (JSPropertyEnumeratorGetNext):
1780         (JSPropertyListAdd):
1781         * API/JSObjectRef.h:
1782         * API/JSStringBufferRef.cpp: Removed.
1783         * API/JSStringBufferRef.h: Removed.
1784         * API/JSValueRef.h:
1785         * API/JavaScriptCore.h:
1786         * API/minidom.c:
1787         (main):
1788         (print):
1789         * API/testapi.c:
1790         (assertEqualsAsUTF8String):
1791         (assertEqualsAsCharactersPtr):
1792         (assertEqualsAsCharacters):
1793         (MyObject_hasProperty):
1794         (MyObject_getProperty):
1795         (MyObject_setProperty):
1796         (MyObject_deleteProperty):
1797         (MyObject_getPropertyList):
1798         (print_callAsFunction):
1799         (myConstructor_callAsConstructor):
1800         (main):
1801         * JavaScriptCore.exp:
1802         * JavaScriptCore.xcodeproj/project.pbxproj:
1803
1804 2006-07-08  Tim Omernick  <timo@apple.com>
1805
1806         Reviewed by Maciej.
1807
1808         Added an OpenGL drawing model to the Netscape Plug-in API.
1809
1810         * bindings/npapi.h:
1811
1812 2006-07-08  Timothy Hatcher  <timothy@apple.com>
1813
1814         Reviewed by Maciej.
1815
1816         Moved KJS_GetCreatedJavaVMs to jni_utility.cpp.
1817         Switched KJS_GetCreatedJavaVMs over to use dlopen and dlsym
1818         now that NSAddImage, NSLookupSymbolInImage and NSAddressOfSymbol
1819         are deprecated in Leopard.
1820
1821         * JavaScriptCore.exp:
1822         * JavaScriptCore.xcodeproj/project.pbxproj:
1823         * bindings/jni/jni_utility.cpp:
1824         (KJS::Bindings::KJS_GetCreatedJavaVMs):
1825         * bindings/softlinking.c: Removed.
1826         * bindings/softlinking.h: Removed.
1827
1828 2006-07-08  Geoffrey Garen  <ggaren@apple.com>
1829
1830         Reviewed by Anders.
1831
1832         - Make JSObjectGetProperty return a JSValue or NULL, like JSEvaluate does.
1833
1834         * API/JSObjectRef.cpp:
1835         (JSObjectGetProperty):
1836         * API/JSObjectRef.h:
1837         * API/testapi.c:
1838         (main):
1839
1840 2006-07-08  Geoffrey Garen  <ggaren@apple.com>
1841
1842         Style change -- no review necessary.
1843
1844         Use 0 instead of NULL in API .cpp files, to match our style guidelines.
1845
1846         * API/JSContextRef.cpp:
1847         (JSEvaluate):
1848         * API/JSObjectRef.cpp:
1849         (JSFunctionMakeWithBody):
1850         (JSObjectCallAsFunction):
1851         (JSObjectCallAsConstructor):
1852         * API/JSValueRef.cpp:
1853         (JSValueToObject):
1854
1855 2006-07-08  Geoffrey Garen  <ggaren@apple.com>
1856
1857         Reviewed by TimO.
1858
1859         - Added ability to pass NULL for thisObject when calling JSObjectCallAsFunction,
1860         to match JSEvaluate.
1861
1862         * API/JSObjectRef.cpp:
1863         (JSObjectCallAsFunction):
1864         * API/JSObjectRef.h:
1865         * API/testapi.c:
1866         (main):
1867
1868 === Safari-521.15 ===
1869
1870 2006-07-07  Geoffrey Garen  <ggaren@apple.com>
1871
1872         Reviewed by Maciej.
1873
1874         - Standardized which functions take a JSContext as an argument. The rule is:
1875         if you might execute JavaScript, you take a JSContext, otherwise you don't.
1876
1877         The FIXME in JSObjectRef.h requires refactoring some parts of Interpreter,
1878         but not API changes, so I'm putting it off until later.
1879
1880         * API/JSCallbackObject.cpp:
1881         (KJS::JSCallbackObject::JSCallbackObject):
1882         (KJS::JSCallbackObject::init):
1883         * API/JSCallbackObject.h:
1884         * API/JSContextRef.cpp:
1885         (JSContextCreate):
1886         * API/JSContextRef.h:
1887         * API/JSObjectRef.cpp:
1888         (JSObjectMake):
1889         (JSPropertyEnumeratorGetNext):
1890         * API/JSObjectRef.h:
1891         * API/testapi.c:
1892         (MyObject_initialize):
1893         (main):
1894         * JavaScriptCore.exp:
1895         * kjs/array_object.cpp:
1896         (ArrayInstance::setLength):
1897         (ArrayInstance::pushUndefinedObjectsToEnd):
1898         * kjs/nodes.cpp:
1899         (ForInNode::execute):
1900         * kjs/reference.cpp:
1901         (KJS::Reference::getPropertyName):
1902         (KJS::Reference::getValue):
1903         * kjs/reference.h:
1904         * kjs/scope_chain.cpp:
1905         (KJS::ScopeChain::print):
1906
1907 2006-07-06  Geoffrey Garen  <ggaren@apple.com>
1908
1909         Reviewed by Maciej.
1910
1911         More API action.
1912
1913         - Headerdoc finished
1914
1915         Semantic Changes:
1916         - Added a JSContextRef argument to many functions, because you need a
1917         JSContextRef for doing virtually anything. I expect to add this argument
1918         to even more functions in a future patch.
1919
1920         - Removed the globalObjectPrototype argument to JSContextCreate because
1921         you can't create an object until you have a context, so it's impossible
1922         to pass a prototype object to JSContextCreate. That's OK because (1) there's
1923         no reason to give the global object a prototype and (2) if you really want
1924         to, you can just use a separate call to JSObjectSetPrototype.
1925
1926         - Removed the JSClassRef argument to JSClassCreate because it was unnecessary,
1927         and you need to be able to make the global object's class before you've
1928         created a JSContext.
1929
1930         - Added an optional exception parameter to JSFunctionMakeWithBody because anything
1931         less would be uncivilized.
1932
1933         - Made the return value parameter to JSObjectGetProperty optional to match
1934         all other return value parameters in the API.
1935
1936         - Made JSObjectSetPrivate/JSObjectGetPrivate work on JSCallbackFunctions
1937         and JSCallbackConstructors. You could use an abstract base class or strategic
1938         placement of m_privateData in the class structure to implement this, but
1939         the former seemed like overkill, and the latter seemed too dangerous.
1940
1941         - Fixed a bug where JSPropertyEnumeratorGetNext would skip the first property.
1942
1943         Cosmetic Changes:
1944         - Reversed the logic of the JSChar #ifdef to avoid confusing headerdoc
1945
1946         - Removed function names from @function declarations because headeroc
1947         can parse them automatically, and I wanted to rule out manual mismatch.
1948
1949         - Changed Error::create to take a const UString& instead of a UString*
1950         because it was looking at me funny.
1951
1952         - Renamed JSStringBufferCreateWithCFString to JSStringBufferCreateCF
1953         because the latter is more concise and it matches JSStringBufferCreateUTF8.
1954
1955         * API/JSCallbackObject.cpp:
1956         (KJS::JSCallbackObject::getOwnPropertySlot):
1957         (KJS::JSCallbackObject::put):
1958         (KJS::JSCallbackObject::deleteProperty):
1959         (KJS::JSCallbackObject::getPropertyList):
1960         (KJS::JSCallbackObject::toBoolean):
1961         (KJS::JSCallbackObject::toNumber):
1962         (KJS::JSCallbackObject::toString):
1963         * API/JSClassRef.cpp:
1964         (JSClassCreate):
1965         * API/JSContextRef.cpp:
1966         (JSContextCreate):
1967         (JSContextSetException):
1968         * API/JSContextRef.h:
1969         * API/JSNode.c:
1970         (JSNodePrototype_class):
1971         (JSNode_class):
1972         * API/JSNodeList.c:
1973         (JSNodeListPrototype_class):
1974         (JSNodeList_class):
1975         * API/JSObjectRef.cpp:
1976         (JSObjectGetProperty):
1977         (JSObjectGetPrivate):
1978         (JSObjectSetPrivate):
1979         (JSObjectCallAsFunction):
1980         (JSObjectCallAsConstructor):
1981         (JSPropertyEnumeratorGetNext):
1982         * API/JSObjectRef.h:
1983         * API/JSStringBufferRef.cpp:
1984         (JSStringBufferCreateCF):
1985         * API/JSStringBufferRef.h:
1986         * API/JSValueRef.cpp:
1987         (JSValueIsInstanceOf):
1988         * API/JSValueRef.h:
1989         * API/minidom.c:
1990         (main):
1991         * API/minidom.js:
1992         * API/testapi.c:
1993         (MyObject_hasProperty):
1994         (MyObject_setProperty):
1995         (MyObject_deleteProperty):
1996         (MyObject_getPropertyList):
1997         (MyObject_convertToType):
1998         (MyObject_class):
1999         (main):
2000         * JavaScriptCore.exp:
2001
2002 2006-07-07  Geoffrey Garen  <ggaren@apple.com>
2003
2004         Reviewed by John.
2005
2006         - Fixed a few crashes resulting from NULL parameters to JSClassCreate.
2007
2008         * API/JSClassRef.cpp:
2009         (JSClassCreate):
2010         (JSClassRelease):
2011         * API/testapi.c: Added test for NULL parameters.
2012         (main):
2013
2014 2006-07-07  Geoffrey Garen  <ggaren@apple.com>
2015
2016         Reviewed by John, mocked by Darin.
2017
2018         - Changed JSEvaluate to take a JSObjectRef instead of a JSValueRef as
2019         "this," since "this" must be an object.
2020
2021         * API/JSContextRef.cpp:
2022         (JSEvaluate):
2023         * API/JSContextRef.h:
2024
2025 2006-07-07  Geoffrey Garen  <ggaren@apple.com>
2026
2027         Reviewed by John.
2028
2029         - More headerdoc
2030
2031         * API/JSBase.h:
2032         * JavaScriptCore.xcodeproj/project.pbxproj:
2033
2034 2006-07-05  Geoffrey Garen  <ggaren@apple.com>
2035
2036         RS by Beth.
2037
2038         Renamed JSCharBufferRef, which was universally unpopular, to JSStringBufferRef,
2039         which, hopefully, will be less unpopular.
2040
2041         * API/APICast.h:
2042         (toJS):
2043         (toRef):
2044         * API/JSBase.h:
2045         * API/JSCallbackObject.cpp:
2046         (KJS::JSCallbackObject::getOwnPropertySlot):
2047         (KJS::JSCallbackObject::put):
2048         (KJS::JSCallbackObject::deleteProperty):
2049         (KJS::JSCallbackObject::staticValueGetter):
2050         (KJS::JSCallbackObject::callbackGetter):
2051         * API/JSCharBufferRef.cpp: Removed.
2052         * API/JSCharBufferRef.h: Removed.
2053         * API/JSContextRef.cpp:
2054         (JSEvaluate):
2055         (JSCheckSyntax):
2056         * API/JSContextRef.h:
2057         * API/JSNode.c:
2058         (JSNodePrototype_appendChild):
2059         (JSNode_getNodeType):
2060         (JSNode_getChildNodes):
2061         (JSNode_getFirstChild):
2062         * API/JSNodeList.c:
2063         (JSNodeList_length):
2064         (JSNodeList_getProperty):
2065         * API/JSObjectRef.cpp:
2066         (JSFunctionMakeWithBody):
2067         (JSObjectGetDescription):
2068         (JSObjectHasProperty):
2069         (JSObjectGetProperty):
2070         (JSObjectSetProperty):
2071         (JSObjectDeleteProperty):
2072         (JSPropertyEnumeratorGetNext):
2073         (JSPropertyListAdd):
2074         * API/JSObjectRef.h:
2075         * API/JSStringBufferRef.cpp: Added.
2076         (JSStringMake):
2077         (JSStringBufferCreate):
2078         (JSStringBufferCreateUTF8):
2079         (JSStringBufferRetain):
2080         (JSStringBufferRelease):
2081         (JSValueCopyStringValue):
2082         (JSStringBufferGetLength):
2083         (JSStringBufferGetCharactersPtr):
2084         (JSStringBufferGetCharacters):
2085         (JSStringBufferGetMaxLengthUTF8):
2086         (JSStringBufferGetCharactersUTF8):
2087         (JSStringBufferIsEqual):
2088         (JSStringBufferIsEqualUTF8):
2089         (JSStringBufferCreateWithCFString):
2090         (CFStringCreateWithJSStringBuffer):
2091         * API/JSStringBufferRef.h: Added.
2092         * API/JSValueRef.h:
2093         * API/JavaScriptCore.h:
2094         * API/minidom.c:
2095         (main):
2096         (print):
2097         * API/testapi.c:
2098         (assertEqualsAsUTF8String):
2099         (assertEqualsAsCharactersPtr):
2100         (assertEqualsAsCharacters):
2101         (MyObject_hasProperty):
2102         (MyObject_getProperty):
2103         (MyObject_setProperty):
2104         (MyObject_deleteProperty):
2105         (MyObject_getPropertyList):
2106         (print_callAsFunction):
2107         (myConstructor_callAsConstructor):
2108         (main):
2109         * JavaScriptCore.exp:
2110         * JavaScriptCore.xcodeproj/project.pbxproj:
2111
2112 2006-07-05  Geoffrey Garen  <ggaren@apple.com>
2113
2114         RS by Beth.
2115
2116         Moved some code around for more logical file separation.
2117
2118         * API/JSBase.h:
2119         * API/JSContextRef.h:
2120         * API/JSObjectRef.cpp:
2121         * API/JSValueRef.cpp:
2122         (JSValueToObject):
2123         * API/JSValueRef.h:
2124
2125 2006-07-03  Geoffrey Garen  <ggaren@apple.com>
2126
2127         Reviewed by Maciej.
2128
2129         Implemented JSFunctionMakeWithBody, which parses a script as a function body
2130         in the global scope, and returns the resulting anonymous function.
2131
2132         I also removed private data from JSCallbackFunction. It never worked,
2133         since JSCallbackFunction doesn't inherit from JSCallbackObject.
2134
2135         * API/JSCallbackConstructor.cpp: Removed.
2136         * API/JSCallbackConstructor.h: Removed.
2137         * API/JSCallbackFunction.cpp:
2138         (KJS::JSCallbackFunction::JSCallbackFunction):
2139         (KJS::JSCallbackFunction::implementsConstruct):
2140         (KJS::JSCallbackFunction::construct):
2141         (KJS::JSCallbackFunction::implementsCall):
2142         (KJS::JSCallbackFunction::callAsFunction):
2143         * API/JSCallbackFunction.h:
2144         * API/JSCallbackObject.cpp:
2145         (KJS::JSCallbackObject::staticFunctionGetter):
2146         * API/JSObjectRef.cpp:
2147         (JSFunctionMake):
2148         (JSFunctionMakeWithCallbacks):
2149         * API/JSObjectRef.h:
2150         * API/JSValueRef.h:
2151         * API/minidom.c:
2152         (main):
2153         * API/testapi.c:
2154         (main):
2155         * JavaScriptCore.exp: Programmatically added all symbols exported by
2156         API object files, and sorted results
2157         * JavaScriptCore.xcodeproj/project.pbxproj:
2158
2159 2006-07-03  Geoffrey Garen  <ggaren@apple.com>
2160
2161         Reviewed by Maciej.
2162
2163         - Return syntax error in JSCheckSyntax through a JSValueRef* exception
2164         argument
2165
2166         * API/JSBase.h:
2167         * API/JSContextRef.cpp:
2168         (JSCheckSyntax):
2169         * API/testapi.c:
2170         (main):
2171         * JavaScriptCore.exp:
2172         * kjs/interpreter.cpp:
2173         (KJS::Interpreter::checkSyntax):
2174         * kjs/interpreter.h:
2175
2176 2006-07-04  Darin Adler  <darin@apple.com>
2177
2178         - fixed build
2179
2180         * wtf/MathExtras.h: Oops. Added missing #endif.
2181
2182 2006-07-04  Bjoern Graf  <bjoern.graf@gmail.com>
2183
2184         Reviewed by Maciej.
2185         Tweaked a bit by Darin.
2186
2187         - http://bugzilla.opendarwin.org/show_bug.cgi?id=9678
2188           work around MSVCRT's fmod function returning NaN for fmod(x, infinity) instead of x
2189
2190         * wtf/MathExtras.h: Added include of <float.h>.
2191         (isinf): Fix to return false for NAN.
2192         (wtf_fmod): Added. An inline that works around the bug.
2193
2194         * kjs/nodes.cpp:
2195         * kjs/number_object.cpp:
2196         * kjs/operations.cpp:
2197         * kjs/value.cpp:
2198         Added includes of MathExtras.h to all files using fmod.
2199
2200         * JavaScriptCore.xcodeproj/project.pbxproj: Let Xcode 2.3 have its way with
2201         the project.
2202
2203 2006-07-01  Geoffrey Garen  <ggaren@apple.com>
2204
2205         Reviewed by Darin.
2206
2207         - Refined value conversions in the API:
2208             - failed toNumber returns NaN
2209             - failed toObject returns NULL
2210             - failed toString returns empty string
2211
2212         - Refined excpetion handling in the API:
2213             - failed value conversions do not throw exceptions
2214             - uncaught exceptions in JSEvaluate, JSObjectCallAsFunction, and
2215               JSObjectCallAsConstructor are returned through a JSValueRef* exception
2216               argument
2217             - removed JSContextHasException, because JSContextGetException does
2218               the same job
2219
2220         * API/JSBase.h:
2221         * API/JSCharBufferRef.cpp:
2222         (JSValueCopyStringValue):
2223         * API/JSContextRef.cpp:
2224         (JSEvaluate):
2225         * API/JSContextRef.h:
2226         * API/JSNodeList.c: Added test code demonstrating how you would use
2227         toNumber, and why you probably don't need toUInt32, etc.
2228         (JSNodeListPrototype_item):
2229         (JSNodeList_getProperty):
2230         * API/JSObjectRef.cpp:
2231         (JSValueToObject):
2232         (JSObjectCallAsFunction):
2233         (JSObjectCallAsConstructor):
2234         * API/JSObjectRef.h:
2235         * API/JSValueRef.cpp:
2236         (JSValueToNumber):
2237         * API/JSValueRef.h:
2238         * API/minidom.c:
2239         (main):
2240         * API/testapi.c:
2241         (main): Added tests for new rules, and call to JSGCProtect to fix Intel
2242         crash
2243         * JavaScriptCore.exp:
2244
2245 2006-07-03  Darin Adler  <darin@apple.com>
2246
2247         - Rolled out HashMap implementation of NPRuntime, at least temporarily.
2248
2249         Fixes hang in the bindings section of layout tests seen on the
2250         buildbot.
2251
2252         This code was using HashMap<const char*, PrivateIdentifier*>.
2253         But that hashes based on pointer identity, not string value.
2254         The default hash for any pointer type is to hash based on the pointer.
2255         And WTF doesn't currently have a string hash for char*.
2256         We'll need to fix that before re-landing this patch.
2257
2258         (Formatting was also incorrect -- extra spaces in parentheses.)
2259
2260         * bindings/npruntime.cpp: Rolled out last change.
2261
2262 2006-07-02  Justin Haygood  <jhaygood@spsu.edu>
2263
2264         Reviewed, tweaked, landed by ggaren.
2265
2266         - Port NPRuntime from CFDictionary to HashMap.
2267
2268         * bindings/npruntime.cpp:
2269         (getStringIdentifierDictionary):
2270         (getIntIdentifierDictionary):
2271         (_NPN_GetStringIdentifier):
2272         (_NPN_GetIntIdentifier):
2273         * bindings/npruntime.h:
2274
2275 2006-07-01  Geoffrey Garen  <ggaren@apple.com>
2276
2277         Reviewed by Adele.
2278
2279         - Fixed <rdar://problem/4611197> REGRESSION: Liveconnect with Java test
2280         fails at http://www-sor.inria.fr/~dedieu/notes/liveconnect/simple_example.html
2281
2282         * JavaScriptCore.exp: Export symbols used by liveconnect
2283
2284 2006-06-29  Geoffrey Garen  <ggaren@apple.com>
2285
2286         Reviewed by Maciej.
2287
2288         - Phase 2 in the JS API.
2289
2290         - Added support for specifying static tables of values -- this should
2291         obviate the need for using complicated callbacks for most lookups.
2292
2293         -  API objects are now created with classes (JSClassRef) -- in order to support
2294         static values, and in order to prevent API objects from storing their
2295         data inline, and thus falling into the oversized (read: slow and prone to
2296         giving Maciej the frowny face) heap.
2297
2298         - Added two specialized JSObject subclasses -- JSCallbackFunction and JSCallbackConstructor --
2299         to allow JSFunctionMake and JSConstructorMake to continue to work with
2300         the new class model. Another solution to this problem would be to create
2301         a custom class object for each function and constructor you make. This
2302         solution is more code but also more efficient.
2303
2304         - Substantially beefed up the minidom example to demonstrate and test a
2305         lot of these techniques. Its output is still pretty haphazard, though.
2306
2307         - Gave the <kjs/ preface to some includes -- I'm told this matters to
2308         building on some versions of Linux.
2309
2310         - Implemented JSValueIsInstanceOf and JSValueIsObjectOfClass
2311
2312         - Removed GetDescription callback. Something in the class datastructure
2313         should take care of this.
2314
2315         * API/JSBase.h:
2316         * API/JSCallbackConstructor.cpp: Added.
2317         (KJS::):
2318         (KJS::JSCallbackConstructor::JSCallbackConstructor):
2319         (KJS::JSCallbackConstructor::implementsConstruct):
2320         (KJS::JSCallbackConstructor::construct):
2321         (KJS::JSCallbackConstructor::setPrivate):
2322         (KJS::JSCallbackConstructor::getPrivate):
2323         * API/JSCallbackConstructor.h: Added.
2324         (KJS::JSCallbackConstructor::classInfo):
2325         * API/JSCallbackFunction.cpp: Added.
2326         (KJS::):
2327         (KJS::JSCallbackFunction::JSCallbackFunction):
2328         (KJS::JSCallbackFunction::implementsCall):
2329         (KJS::JSCallbackFunction::callAsFunction):
2330         (KJS::JSCallbackFunction::setPrivate):
2331         (KJS::JSCallbackFunction::getPrivate):
2332         * API/JSCallbackFunction.h: Added.
2333         (KJS::JSCallbackFunction::classInfo):
2334         * API/JSCallbackObject.cpp:
2335         (KJS::):
2336         (KJS::JSCallbackObject::JSCallbackObject):
2337         (KJS::JSCallbackObject::init):
2338         (KJS::JSCallbackObject::~JSCallbackObject):
2339         (KJS::JSCallbackObject::className):
2340         (KJS::JSCallbackObject::getOwnPropertySlot):
2341         (KJS::JSCallbackObject::put):
2342         (KJS::JSCallbackObject::deleteProperty):
2343         (KJS::JSCallbackObject::implementsConstruct):
2344         (KJS::JSCallbackObject::construct):
2345         (KJS::JSCallbackObject::implementsCall):
2346         (KJS::JSCallbackObject::callAsFunction):
2347         (KJS::JSCallbackObject::getPropertyList):
2348         (KJS::JSCallbackObject::toBoolean):
2349         (KJS::JSCallbackObject::toNumber):
2350         (KJS::JSCallbackObject::toString):
2351         (KJS::JSCallbackObject::inherits):
2352         (KJS::JSCallbackObject::staticValueGetter):
2353         (KJS::JSCallbackObject::staticFunctionGetter):
2354         (KJS::JSCallbackObject::callbackGetter):
2355         * API/JSCallbackObject.h:
2356         * API/JSCharBufferRef.cpp:
2357         * API/JSClassRef.cpp: Added.
2358         (JSClassCreate):
2359         (JSClassRetain):
2360         (JSClassRelease):
2361         * API/JSClassRef.h: Added.
2362         (StaticValueEntry::StaticValueEntry):
2363         (StaticFunctionEntry::StaticFunctionEntry):
2364         (__JSClass::__JSClass):
2365         * API/JSContextRef.cpp:
2366         (JSContextCreate):
2367         (JSEvaluate):
2368         * API/JSContextRef.h:
2369         * API/JSNode.c: Added.
2370         (JSNodePrototype_appendChild):
2371         (JSNodePrototype_removeChild):
2372         (JSNodePrototype_replaceChild):
2373         (JSNodePrototype_class):
2374         (JSNode_getNodeType):
2375         (JSNode_getChildNodes):
2376         (JSNode_getFirstChild):
2377         (JSNode_finalize):
2378         (JSNode_class):
2379         (JSNode_prototype):
2380         (JSNode_new):
2381         (JSNode_construct):
2382         * API/JSNode.h: Added.
2383         * API/JSNodeList.c: Added.
2384         (JSNodeListPrototype_item):
2385         (JSNodeListPrototype_class):
2386         (JSNodeList_length):
2387         (JSNodeList_getProperty):
2388         (JSNodeList_finalize):
2389         (JSNodeList_class):
2390         (JSNodeList_prototype):
2391         (JSNodeList_new):
2392         * API/JSNodeList.h: Added.
2393         * API/JSObjectRef.cpp:
2394         (JSObjectMake):
2395         (JSFunctionMake):
2396         (JSConstructorMake):
2397         (__JSPropertyEnumerator::__JSPropertyEnumerator):
2398         (JSObjectCreatePropertyEnumerator):
2399         (JSPropertyEnumeratorGetNext):
2400         (JSPropertyEnumeratorRetain):
2401         (JSPropertyEnumeratorRelease):
2402         * API/JSObjectRef.h:
2403         (__JSObjectCallbacks::):
2404         * API/JSValueRef.cpp:
2405         (JSValueIsObjectOfClass):
2406         (JSValueIsInstanceOf):
2407         * API/JSValueRef.h:
2408         * API/Node.c: Added.
2409         (Node_new):
2410         (Node_appendChild):
2411         (Node_removeChild):
2412         (Node_replaceChild):
2413         (Node_ref):
2414         (Node_deref):
2415         * API/Node.h: Added.
2416         * API/NodeList.c: Added.
2417         (NodeList_new):
2418         (NodeList_length):
2419         (NodeList_item):
2420         (NodeList_ref):
2421         (NodeList_deref):
2422         * API/NodeList.h: Added.
2423         * API/minidom.c:
2424         (main):
2425         (print):
2426         (createStringWithContentsOfFile):
2427         * API/minidom.js:
2428         * API/testapi.c:
2429         (assertEqualsAsCharacters):
2430         (MyObject_getProperty):
2431         (MyObject_class):
2432         (myConstructor_callAsConstructor):
2433         (main):
2434         * API/testapi.js:
2435         * JavaScriptCore.xcodeproj/project.pbxproj:
2436
2437 2006-06-26  Kevin Ollivier  <kevino@theolliviers.com>
2438
2439         Reviewed and tweaked by Darin.
2440
2441         - Compile fixes for wx port / gcc 4.0.2
2442
2443         * kjs/array_object.cpp:
2444         Added missing headers.
2445
2446         * kjs/ExecState.h:
2447         gcc needs class prototypes before defining those classes as friend classes
2448
2449 2006-06-30  Mike Emmel  <mike.emmel@gmail.com>
2450
2451         Reviewed by Darin.
2452
2453         Compilation fixes for Linux/Gdk.
2454
2455         * JavaScriptCore/kjs/interpreter.cpp: added include of signal.h
2456         * JavaScriptCore/kjs/ExecState.h: added missing class declaration
2457         * JavaScriptCore/kjs/ExecState.cpp: case wrong on include of context.h
2458         * JavaScriptCore/JavaScriptCoreSources.bkl: added Context.cpp and ExecState.cpp
2459
2460 === Safari-521.14 ===
2461
2462 2006-06-29  Maciej Stachowiak  <mjs@apple.com>
2463
2464         Reviewed by Geoff.
2465
2466         - add headerdoc comments to some of the new JS API headers
2467
2468         * API/JSBase.h:
2469         * API/JSValueRef.h:
2470
2471 2006-06-28  Timothy Hatcher  <timothy@apple.com>
2472
2473         Prefer the Stabs debugging symbols format until DWARF bugs are fixed.
2474
2475         * JavaScriptCore.xcodeproj/project.pbxproj:
2476
2477 2006-06-27  Timothy Hatcher  <timothy@apple.com>
2478
2479         Reviewed by Tim O.
2480
2481         <rdar://problem/4448350> Deprecated ObjC language API used in JavaScriptCore, WebCore, WebKit and WebBrowser
2482
2483         Switch to the new ObjC 2 API, ifdefed the old code around OBJC_API_VERSION so it still works on Tiger.
2484         Removed the use of the old stringWithCString, switched to the new Tiger version that accepts an encoding.
2485         Lots of code style cleanup.
2486
2487         * JavaScriptCore.xcodeproj/project.pbxproj:
2488         * bindings/objc/objc_class.h:
2489         * bindings/objc/objc_class.mm:
2490         (KJS::Bindings::ObjcClass::~ObjcClass):
2491         (KJS::Bindings::_createClassesByIsAIfNecessary):
2492         (KJS::Bindings::ObjcClass::classForIsA):
2493         (KJS::Bindings::ObjcClass::name):
2494         (KJS::Bindings::ObjcClass::methodsNamed):
2495         (KJS::Bindings::ObjcClass::fieldNamed):
2496         (KJS::Bindings::ObjcClass::fallbackObject):
2497         * bindings/objc/objc_header.h:
2498         * bindings/objc/objc_instance.h:
2499         * bindings/objc/objc_instance.mm:
2500         (ObjcInstance::ObjcInstance):
2501         (ObjcInstance::~ObjcInstance):
2502         (ObjcInstance::operator=):
2503         (ObjcInstance::begin):
2504         (ObjcInstance::end):
2505         (ObjcInstance::getClass):
2506         (ObjcInstance::invokeMethod):
2507         (ObjcInstance::invokeDefaultMethod):
2508         (ObjcInstance::setValueOfField):
2509         (ObjcInstance::supportsSetValueOfUndefinedField):
2510         (ObjcInstance::setValueOfUndefinedField):
2511         (ObjcInstance::getValueOfField):
2512         (ObjcInstance::getValueOfUndefinedField):
2513         (ObjcInstance::defaultValue):
2514         (ObjcInstance::stringValue):
2515         (ObjcInstance::numberValue):
2516         (ObjcInstance::booleanValue):
2517         (ObjcInstance::valueOf):
2518         * bindings/objc/objc_runtime.h:
2519         * bindings/objc/objc_runtime.mm:
2520         (ObjcMethod::ObjcMethod):
2521         (ObjcMethod::name):
2522         (ObjcMethod::getMethodSignature):
2523         (ObjcMethod::setJavaScriptName):
2524         (ObjcField::name):
2525         (ObjcField::type):
2526         (ObjcField::valueFromInstance):
2527         (convertValueToObjcObject):
2528         (ObjcField::setValueToInstance):
2529         (ObjcArray::operator=):
2530         (ObjcArray::setValueAt):
2531         (ObjcArray::valueAt):
2532         (ObjcFallbackObjectImp::ObjcFallbackObjectImp):
2533         (ObjcFallbackObjectImp::callAsFunction):
2534         (ObjcFallbackObjectImp::defaultValue):
2535
2536 2006-06-28  Anders Carlsson  <acarlsson@apple.com>
2537
2538         Reviewed by Geoff.
2539
2540         http://bugzilla.opendarwin.org/show_bug.cgi?id=8636
2541         REGRESSION: JavaScript access to Java applet causes hang (_webViewURL not implemented)
2542
2543         * bindings/jni/jni_objc.mm:
2544         (KJS::Bindings::dispatchJNICall):
2545         Just pass nil as the calling URL. This will cause the Java plugin to use the URL of the page
2546         containing the applet (which is what we used to do).
2547
2548 2006-06-27  Timothy Hatcher  <timothy@apple.com>
2549
2550         Reviewed by Darin.
2551
2552         <rdar://problem/4406785> Add an export file to TOT JavaScriptCore like the Safari-2-0-branch
2553
2554         * JavaScriptCore.exp: Added.
2555         * JavaScriptCore.xcodeproj/project.pbxproj:
2556
2557 2006-06-25  Geoffrey Garen  <ggaren@apple.com>
2558
2559         Reviewed by Adele.
2560
2561         - Added JSConstructorMake to match JSFunctionMake, along with test code.
2562
2563         [ I checked in the ChangeLog before without the actual files. ]
2564
2565         * API/JSObjectRef.cpp:
2566         (JSConstructorMake):
2567         * API/JSObjectRef.h:
2568         * API/testapi.c:
2569         (myConstructor_callAsConstructor):
2570         (main):
2571         * API/testapi.js:
2572         * ChangeLog:
2573         * JavaScriptCore.xcodeproj/project.pbxproj: Moved testapi.c to the testapi
2574         target -- this was an oversight in my earlier check-in.
2575
2576 2006-06-25  Timothy Hatcher  <timothy@apple.com>
2577
2578         Reviewed by Darin.
2579
2580         Bug 9574: Drosera should show inline scripts within the original HTML
2581         http://bugzilla.opendarwin.org/show_bug.cgi?id=9574
2582
2583         Pass the starting line number and error message to the debugger.
2584
2585         * kjs/debugger.cpp:
2586         (Debugger::sourceParsed):
2587         * kjs/debugger.h:
2588         * kjs/function.cpp:
2589         (KJS::GlobalFuncImp::callAsFunction):
2590         * kjs/function_object.cpp:
2591         (FunctionObjectImp::construct):
2592         * kjs/interpreter.cpp:
2593         (KJS::Interpreter::evaluate):
2594
2595 2006-06-24  Alexey Proskuryakov  <ap@nypop.com>
2596
2597         Rubber-stamped by Eric.
2598
2599         Add a -h (do not follow symlinks) option to ln in derived sources build script (without it,
2600         a symlink was created inside the source directory on second build).
2601
2602         * JavaScriptCore.xcodeproj/project.pbxproj:
2603
2604 2006-06-24  David Kilzer  <ddkilzer@kilzer.net>
2605
2606         Reviewed by Timothy.
2607
2608         * Info.plist: Fixed copyright to include 2003-2006.
2609
2610 2006-06-24  Alexey Proskuryakov  <ap@nypop.com>
2611
2612         Reviewed by Darin.
2613
2614         - http://bugzilla.opendarwin.org/show_bug.cgi?id=9418
2615         WebKit will not build when Space exists in path
2616
2617         * JavaScriptCore.xcodeproj/project.pbxproj: Enclose search paths in quotes; create symlinks to
2618         avoid passing paths with spaces to make.
2619
2620 2006-06-23  Timothy Hatcher  <timothy@apple.com>
2621
2622         Reviewed by Darin.
2623
2624         Adding more operator[] overloads for long and short types.
2625
2626         * wtf/Vector.h:
2627         (WTF::Vector::operator[]):
2628
2629 === JavaScriptCore-521.13 ===
2630
2631 2006-06-22  Alexey Proskuryakov  <ap@nypop.com>
2632
2633         Build fix.
2634
2635         - http://bugzilla.opendarwin.org/show_bug.cgi?id=9539
2636         Another case error preventing build
2637
2638         * API/JSObjectRef.cpp: Changed "identifier.h" to "Identifier.h"
2639
2640 2006-06-22  David Kilzer  <ddkilzer@kilzer.net>
2641
2642         Build fix.
2643
2644         http://bugzilla.opendarwin.org/show_bug.cgi?id=9539
2645         Another case error preventing build
2646
2647         * API/APICast.h: Changed "UString.h" to "ustring.h".
2648
2649 2006-06-21  Geoffrey Garen  <ggaren@apple.com>
2650
2651         Fixed release build, fixed accidental infinite recursion due to
2652         last minute global replace gone awry.
2653
2654         * API/APICast.h:
2655         (toRef):
2656         * API/testapi.c:
2657         (assertEqualsAsBoolean):
2658         (assertEqualsAsNumber):
2659         (assertEqualsAsUTF8String):
2660         (assertEqualsAsCharactersPtr):
2661         * JavaScriptCore.xcodeproj/project.pbxproj:
2662
2663 2006-06-21  Geoffrey Garen  <ggaren@apple.com>
2664
2665         Reviewed by Anders.
2666
2667         - First cut at C API to JavaScript. Includes a unit test, 'testapi.c',
2668         and the outline of a test app, 'minidom.c'.
2669
2670         Includes one change to JSC internals: Rename propList to getPropertyList and have it
2671         take its target property list by reference so that subclasses can
2672         add properties to the list before calling through to their superclasses.
2673
2674         Also, I just ran prepare-ChangeLog in about 10 seconds, and I would like
2675         to give a shout-out to that.
2676
2677         * API/APICast.h: Added.
2678         (toJS):
2679         (toRef):
2680         * API/JSBase.h: Added.
2681         * API/JSCallbackObject.cpp: Added.
2682         (KJS::):
2683         (KJS::JSCallbackObject::JSCallbackObject):
2684         (KJS::JSCallbackObject::~JSCallbackObject):
2685         (KJS::JSCallbackObject::className):
2686         (KJS::JSCallbackObject::getOwnPropertySlot):
2687         (KJS::JSCallbackObject::put):
2688         (KJS::JSCallbackObject::deleteProperty):
2689         (KJS::JSCallbackObject::implementsConstruct):
2690         (KJS::JSCallbackObject::construct):
2691         (KJS::JSCallbackObject::implementsCall):
2692         (KJS::JSCallbackObject::callAsFunction):
2693         (KJS::JSCallbackObject::getPropertyList):
2694         (KJS::JSCallbackObject::toBoolean):
2695         (KJS::JSCallbackObject::toNumber):
2696         (KJS::JSCallbackObject::toString):
2697         (KJS::JSCallbackObject::setPrivate):
2698         (KJS::JSCallbackObject::getPrivate):
2699         (KJS::JSCallbackObject::cachedValueGetter):
2700         (KJS::JSCallbackObject::callbackGetter):
2701         * API/JSCallbackObject.h: Added.
2702         (KJS::JSCallbackObject::classInfo):
2703         * API/JSCharBufferRef.cpp: Added.
2704         (JSStringMake):
2705         (JSCharBufferCreate):
2706         (JSCharBufferCreateUTF8):
2707         (JSCharBufferRetain):
2708         (JSCharBufferRelease):
2709         (JSValueCopyStringValue):
2710         (JSCharBufferGetLength):
2711         (JSCharBufferGetCharactersPtr):
2712         (JSCharBufferGetCharacters):
2713         (JSCharBufferGetMaxLengthUTF8):
2714         (JSCharBufferGetCharactersUTF8):
2715         (JSCharBufferIsEqual):
2716         (JSCharBufferIsEqualUTF8):
2717         (JSCharBufferCreateWithCFString):
2718         (CFStringCreateWithJSCharBuffer):
2719         * API/JSCharBufferRef.h: Added.
2720         * API/JSContextRef.cpp: Added.
2721         (JSContextCreate):
2722         (JSContextDestroy):
2723         (JSContextGetGlobalObject):
2724         (JSEvaluate):
2725         (JSCheckSyntax):
2726         (JSContextHasException):
2727         (JSContextGetException):
2728         (JSContextClearException):
2729         (JSContextSetException):
2730         * API/JSContextRef.h: Added.
2731         * API/JSObjectRef.cpp: Added.
2732         (JSValueToObject):
2733         (JSObjectMake):
2734         (JSFunctionMake):
2735         (JSObjectGetDescription):
2736         (JSObjectGetPrototype):
2737         (JSObjectSetPrototype):
2738         (JSObjectHasProperty):
2739         (JSObjectGetProperty):
2740         (JSObjectSetProperty):
2741         (JSObjectDeleteProperty):
2742         (JSObjectGetPrivate):
2743         (JSObjectSetPrivate):
2744         (JSObjectIsFunction):
2745         (JSObjectCallAsFunction):
2746         (JSObjectIsConstructor):
2747         (JSObjectCallAsConstructor):
2748         (__JSPropertyListEnumerator::__JSPropertyListEnumerator):
2749         (JSObjectCreatePropertyEnumerator):
2750         (JSPropertyEnumeratorGetNext):
2751         (JSPropertyEnumeratorRetain):
2752         (JSPropertyEnumeratorRelease):
2753         (JSPropertyListAdd):
2754         * API/JSObjectRef.h: Added.
2755         * API/JSValueRef.cpp: Added.
2756         (JSValueGetType):
2757         (JSValueIsUndefined):
2758         (JSValueIsNull):
2759         (JSValueIsBoolean):
2760         (JSValueIsNumber):
2761         (JSValueIsString):
2762         (JSValueIsObject):
2763         (JSValueIsEqual):
2764         (JSValueIsStrictEqual):
2765         (JSUndefinedMake):
2766         (JSNullMake):
2767         (JSBooleanMake):
2768         (JSNumberMake):
2769         (JSValueToBoolean):
2770         (JSValueToNumber):
2771         (JSGCProtect):
2772         (JSGCUnprotect):
2773         (JSGCCollect):
2774         * API/JSValueRef.h: Added.
2775         * API/JavaScriptCore.h: Added.
2776         * API/minidom.c: Added.
2777         (main):
2778         * API/minidom.html: Added.
2779         * API/minidom.js: Added.
2780         * API/testapi.c: Added.
2781         (assertEqualsAsBoolean):
2782         (assertEqualsAsNumber):
2783         (assertEqualsAsUTF8String):
2784         (assertEqualsAsCharactersPtr):
2785         (assertEqualsAsCharacters):
2786         (MyObject_initialize):
2787         (MyObject_copyDescription):
2788         (MyObject_hasProperty):
2789         (MyObject_getProperty):
2790         (MyObject_setProperty):
2791         (MyObject_deleteProperty):
2792         (MyObject_getPropertyList):
2793         (MyObject_callAsFunction):
2794         (MyObject_callAsConstructor):
2795         (MyObject_convertToType):
2796         (MyObject_finalize):
2797         (print_callAsFunction):
2798         (main):
2799         (createStringWithContentsOfFile):
2800         * API/testapi.js: Added.
2801         * ChangeLog:
2802         * JavaScriptCore.xcodeproj/project.pbxproj:
2803         * bindings/npruntime_impl.h:
2804         * kjs/array_instance.h:
2805         * kjs/array_object.cpp:
2806         (ArrayInstance::getPropertyList):
2807         * kjs/interpreter.cpp:
2808         (KJS::Interpreter::evaluate):
2809         * kjs/nodes.cpp:
2810         (ForInNode::execute):
2811         * kjs/object.cpp:
2812         (KJS::JSObject::put):
2813         (KJS::JSObject::canPut):
2814         (KJS::JSObject::deleteProperty):
2815         (KJS::JSObject::propertyIsEnumerable):
2816         (KJS::JSObject::getPropertyAttributes):
2817         (KJS::JSObject::getPropertyList):
2818         * kjs/object.h:
2819         * kjs/property_map.cpp:
2820         (KJS::PropertyMap::get):
2821         * kjs/property_map.h:
2822         * kjs/scope_chain.cpp:
2823         (KJS::ScopeChain::print):
2824         * kjs/string_object.cpp:
2825         (StringInstance::getPropertyList):
2826         * kjs/string_object.h:
2827         * kjs/ustring.h:
2828         (KJS::UString::Rep::ref):
2829
2830 2006-06-20  Timothy Hatcher  <timothy@apple.com>
2831
2832         Reviewed by Geoff.
2833
2834         Make sure we clear the exception before returning so
2835         that future calls will not fail because of an earlier
2836         exception state. Assert on entry that the WebScriptObject
2837         is working with an ExecState that dose not have an exception.
2838         Document that evaluateWebScript and callWebScriptMethod return
2839         WebUndefined when an exception is thrown.
2840
2841         * bindings/objc/WebScriptObject.h:
2842         * bindings/objc/WebScriptObject.mm:
2843         (-[WebScriptObject callWebScriptMethod:withArguments:]):
2844         (-[WebScriptObject evaluateWebScript:]):
2845         (-[WebScriptObject setValue:forKey:]):
2846         (-[WebScriptObject valueForKey:]):
2847         (-[WebScriptObject removeWebScriptKey:]):
2848         (-[WebScriptObject webScriptValueAtIndex:]):
2849         (-[WebScriptObject setWebScriptValueAtIndex:value:]):
2850
2851 2006-06-19  Anders Carlsson  <acarlsson@apple.com>
2852
2853         Reviewed by John.
2854
2855         * kjs/interpreter.cpp:
2856         (KJS::TimeoutChecker::pauseTimeoutCheck):
2857         (KJS::TimeoutChecker::resumeTimeoutCheck):
2858         Fix argument order in setitimer calls.
2859
2860 2006-06-18  Anders Carlsson  <acarlsson@apple.com>
2861
2862         Reviewed by Geoff.
2863
2864         * kjs/interpreter.cpp:
2865         (KJS::TimeoutChecker::pauseTimeoutCheck):
2866         Do nothing if the timeout check hasn't been started.
2867
2868         (KJS::TimeoutChecker::resumeTimeoutCheck):
2869         Do nothing if the timeout check hasn't been started.
2870         Use the right signal handler when unblocking.
2871
2872         (KJS::Interpreter::handleTimeout):
2873         pause/resume the timeout check around the call to
2874         shouldInterruptScript().
2875
2876 2006-06-16  Ben Goodger <beng@google.com>
2877
2878         Reviewed by Maciej
2879
2880         http://bugzilla.opendarwin.org/show_bug.cgi?id=9491
2881         Windows build breaks in interpreter.cpp
2882
2883         * kjs/interpreter.cpp
2884         (KJS::TimeoutChecker::pauseTimeoutCheck):
2885         (KJS::TimeoutChecker::resumeTimeoutCheck):
2886         Make sure to only assert equality with s_executingInterpreter when it
2887         is being used (i.e. when HAVE(SYS_TIME_H) == true)
2888
2889 2006-06-17  David Kilzer  <ddkilzer@kilzer.net>
2890
2891         Reviewed by darin.
2892
2893         http://bugzilla.opendarwin.org/show_bug.cgi?id=9477
2894         REGRESSION: fast/dom/replaceChild.html crashes on WebKit ToT in debug build
2895
2896         * kjs/function.cpp:
2897         (KJS::FunctionImp::callAsFunction): Refetch the debugger after executing the function
2898         in case the WebFrame it was running in has since been destroyed.
2899
2900 2006-06-17  David Kilzer  <ddkilzer@kilzer.net>
2901
2902         Reviewed by ggaren.
2903
2904         http://bugzilla.opendarwin.org/show_bug.cgi?id=9476
2905         REGRESSION: Reproducible crash after closing window after viewing
2906         css2.1/t0803-c5501-imrgn-t-00-b-ag.html
2907
2908         * kjs/debugger.cpp:
2909         (Debugger::detach): Call setDebugger(0) for all interpreters removed from
2910         the 'attached to a debugger' list.
2911
2912 2006-06-17  Anders Carlsson  <acarlsson@apple.com>
2913
2914         Reviewed by Maciej and Geoff.
2915
2916         http://bugzilla.opendarwin.org/show_bug.cgi?id=7080
2917         Provide some way to stop a JavaScript infinite loop
2918
2919         * kjs/completion.h:
2920         (KJS::):
2921         Add Interrupted completion type.
2922
2923         * kjs/function.cpp:
2924         (KJS::FunctionImp::callAsFunction):
2925         (KJS::GlobalFuncImp::callAsFunction):
2926         Only set the exception on the new ExecState if the current one has had one.
2927
2928         * kjs/interpreter.cpp:
2929         (KJS::TimeoutChecker::startTimeoutCheck):
2930         (KJS::TimeoutChecker::stopTimeoutCheck):
2931         (KJS::TimeoutChecker::alarmHandler):
2932         (KJS::TimeoutChecker::pauseTimeoutCheck):
2933         (KJS::TimeoutChecker::resumeTimeoutCheck):
2934         New TimeoutChecker class which handles setting Interpreter::m_timedOut flag after a given
2935         period of time. This currently only works on Unix platforms where setitimer and signals are used.
2936
2937         (KJS::Interpreter::Interpreter):
2938         Initialize new member variables.
2939
2940         (KJS::Interpreter::~Interpreter):
2941         Destroy the timeout checker.
2942
2943         (KJS::Interpreter::startTimeoutCheck):
2944         (KJS::Interpreter::stopTimeoutCheck):
2945         (KJS::Interpreter::pauseTimeoutCheck):
2946         (KJS::Interpreter::resumeTimeoutCheck):
2947         Call the timeout checker.
2948
2949         (KJS::Interpreter::handleTimeout):
2950         Called on timeout. Resets the m_timedOut flag and calls shouldInterruptScript.
2951
2952         * kjs/interpreter.h:
2953         (KJS::Interpreter::setTimeoutTime):
2954         New function for setting the timeout time.
2955
2956         (KJS::Interpreter::shouldInterruptScript):
2957         New function. The idea is that this should be overridden by subclasses in order to for example
2958         pop up a dialog asking the user if the script should be interrupted.
2959
2960         (KJS::Interpreter::checkTimeout):
2961         New function which checks the m_timedOut flag and calls handleTimeout if it's set.
2962
2963         * kjs/nodes.cpp:
2964         (DoWhileNode::execute):
2965         (WhileNode::execute):
2966         (ForNode::execute):
2967         Call Interpreter::checkTimeout after each iteration of the loop.
2968
2969 2006-06-15  Timothy Hatcher  <timothy@apple.com>
2970
2971         Reviewed by Geoff and Darin.
2972
2973         Prefer the DWARF debugging symbols format for use in Xcode 2.3.
2974
2975         * JavaScriptCore.xcodeproj/project.pbxproj:
2976
2977 2006-06-14  Geoffrey Garen  <ggaren@apple.com>
2978
2979         Reviewed by Beth.
2980
2981         - fixed http://bugzilla.opendarwin.org/show_bug.cgi?id=9438
2982         Someone broke ToT: cannot build
2983
2984         * JavaScriptCore.xcodeproj/project.pbxproj:
2985         * bindings/runtime_root.h: Changed "Interpreter.h" to "interpreter.h"
2986
2987 2006-06-12  Geoffrey Garen  <ggaren@apple.com>
2988
2989         build fix
2990
2991         * bindings/objc/WebScriptObject.mm:
2992         (+[WebScriptObject throwException:]): Restore assignment I accidentally
2993         deleted in previous commit
2994
2995 2006-06-12  Geoffrey Garen  <ggaren@apple.com>
2996
2997         Reviewed by TimO, Maciej.
2998
2999         - Merged InterpreterImp code into Interpreter, which implements
3000         all interpreter functionality now. This is part of my continuing quest
3001         to create an external notion of JS "execution context" that is unified and simple --
3002         something to replace the mix of Context, ContextImp, ExecState, Interpreter,
3003         InterpreterImp, and JSRun.
3004
3005         All tests pass. Leaks test has not regressed from its baseline ~207 leaks
3006         with ~3460 leaked nodes.
3007
3008         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
3009         * JavaScriptCore.xcodeproj/project.pbxproj:
3010         * bindings/NP_jsobject.cpp:
3011         * bindings/objc/WebScriptObject.mm:
3012         (+[WebScriptObject throwException:]):
3013         * bindings/runtime_root.cpp:
3014         * bindings/runtime_root.h:
3015         * kjs/Context.cpp:
3016         (KJS::Context::Context):
3017         * kjs/ExecState.cpp: Added.
3018         (KJS::ExecState::lexicalInterpreter):
3019         * kjs/ExecState.h: Added.
3020         (KJS::ExecState::dynamicInterpreter):
3021         * kjs/SavedBuiltins.h: Added.
3022         * kjs/bool_object.cpp:
3023         (BooleanPrototype::BooleanPrototype):
3024         * kjs/collector.cpp:
3025         (KJS::Collector::collect):
3026         (KJS::Collector::numInterpreters):
3027         * kjs/context.h:
3028         * kjs/debugger.cpp:
3029         (Debugger::attach):
3030         (Debugger::detach):
3031         * kjs/function.cpp:
3032         (KJS::FunctionImp::callAsFunction):
3033         (KJS::GlobalFuncImp::callAsFunction):
3034         * kjs/function_object.cpp:
3035         (FunctionObjectImp::construct):
3036         * kjs/internal.cpp:
3037         * kjs/internal.h:
3038         * kjs/interpreter.cpp:
3039         (KJS::interpreterMap):
3040         (KJS::Interpreter::Interpreter):
3041         (KJS::Interpreter::init):
3042         (KJS::Interpreter::~Interpreter):
3043         (KJS::Interpreter::globalObject):
3044         (KJS::Interpreter::initGlobalObject):
3045         (KJS::Interpreter::globalExec):
3046         (KJS::Interpreter::checkSyntax):
3047         (KJS::Interpreter::evaluate):
3048         (KJS::Interpreter::builtinObject):
3049         (KJS::Interpreter::builtinFunction):
3050         (KJS::Interpreter::builtinArray):
3051         (KJS::Interpreter::builtinBoolean):
3052         (KJS::Interpreter::builtinString):
3053         (KJS::Interpreter::builtinNumber):
3054         (KJS::Interpreter::builtinDate):
3055         (KJS::Interpreter::builtinRegExp):
3056         (KJS::Interpreter::builtinError):
3057         (KJS::Interpreter::builtinObjectPrototype):
3058         (KJS::Interpreter::builtinFunctionPrototype):
3059         (KJS::Interpreter::builtinArrayPrototype):
3060         (KJS::Interpreter::builtinBooleanPrototype):
3061         (KJS::Interpreter::builtinStringPrototype):
3062         (KJS::Interpreter::builtinNumberPrototype):
3063         (KJS::Interpreter::builtinDatePrototype):
3064         (KJS::Interpreter::builtinRegExpPrototype):
3065         (KJS::Interpreter::builtinErrorPrototype):
3066         (KJS::Interpreter::builtinEvalError):
3067         (KJS::Interpreter::builtinRangeError):
3068         (KJS::Interpreter::builtinReferenceError):
3069         (KJS::Interpreter::builtinSyntaxError):
3070         (KJS::Interpreter::builtinTypeError):
3071         (KJS::Interpreter::builtinURIError):
3072         (KJS::Interpreter::builtinEvalErrorPrototype):
3073         (KJS::Interpreter::builtinRangeErrorPrototype):
3074         (KJS::Interpreter::builtinReferenceErrorPrototype):
3075         (KJS::Interpreter::builtinSyntaxErrorPrototype):
3076         (KJS::Interpreter::builtinTypeErrorPrototype):
3077         (KJS::Interpreter::builtinURIErrorPrototype):
3078         (KJS::Interpreter::mark):
3079         (KJS::Interpreter::interpreterWithGlobalObject):
3080         (KJS::Interpreter::saveBuiltins):
3081         (KJS::Interpreter::restoreBuiltins):
3082         * kjs/interpreter.h:
3083         (KJS::Interpreter::setCompatMode):
3084         (KJS::Interpreter::compatMode):
3085         (KJS::Interpreter::firstInterpreter):
3086         (KJS::Interpreter::nextInterpreter):
3087         (KJS::Interpreter::prevInterpreter):
3088         (KJS::Interpreter::debugger):
3089         (KJS::Interpreter::setDebugger):
3090         (KJS::Interpreter::setContext):
3091         (KJS::Interpreter::context):
3092         * kjs/nodes.cpp:
3093         (StatementNode::hitStatement):
3094         (RegExpNode::evaluate):
3095         * kjs/protect.h:
3096
3097 2006-06-12  Geoffrey Garen  <ggaren@apple.com>
3098
3099         Reviewed by Maciej.
3100
3101         - Have *.lut.h files #include lookup.h to eliminate surprising header
3102         include order dependency.
3103
3104         * DerivedSources.make:
3105         * kjs/array_object.cpp:
3106         * kjs/date_object.cpp:
3107         * kjs/date_object.h:
3108         (KJS::DateProtoFunc::):
3109         * kjs/lexer.cpp:
3110         * kjs/math_object.cpp:
3111         * kjs/number_object.cpp:
3112         * kjs/regexp_object.cpp:
3113         * kjs/string_object.cpp:
3114
3115 2006-06-10  Geoffrey Garen  <ggaren@apple.com>
3116
3117         - http://bugzilla.opendarwin.org/show_bug.cgi?id=8515
3118         Linux porting compile bug
3119
3120         Fix by Mike Emmel, Reviewed by Darin.
3121
3122         * JavaScriptCoreSources.bkl:
3123         * jscore.bkl:
3124         * wtf/Platform.h:
3125
3126 2006-06-09  Geoffrey Garen  <ggaren@apple.com>
3127
3128         Build fix -- I think :).
3129
3130         * JavaScriptCore.xcodeproj/project.pbxproj:
3131         * kjs/context.h:
3132
3133 2006-06-09  Geoffrey Garen  <ggaren@apple.com>
3134
3135         Reviewed by Eric (yay!).
3136
3137         - Removed Context wrapper for ContextImp, renamed ContextImp to Context,
3138         split Context into its own file -- Context.cpp -- renamed _var to m_var,
3139         change ' *' to '* '.
3140
3141         * JavaScriptCore.xcodeproj/project.pbxproj:
3142         * kjs/Context.cpp: Added.
3143         (KJS::Context::Context):
3144         (KJS::Context::~Context):
3145         (KJS::Context::mark):
3146         * kjs/context.h:
3147         (KJS::Context::scopeChain):
3148         (KJS::Context::variableObject):
3149         (KJS::Context::setVariableObject):
3150         (KJS::Context::thisValue):
3151         (KJS::Context::callingContext):
3152         (KJS::Context::activationObject):
3153         (KJS::Context::currentBody):
3154         (KJS::Context::function):
3155         (KJS::Context::arguments):
3156         (KJS::Context::pushScope):
3157         (KJS::Context::seenLabels):
3158         * kjs/function.cpp:
3159         (KJS::FunctionImp::callAsFunction):
3160         (KJS::FunctionImp::processParameters):
3161         (KJS::FunctionImp::argumentsGetter):
3162         (KJS::GlobalFuncImp::callAsFunction):
3163         * kjs/internal.cpp:
3164         (KJS::InterpreterImp::evaluate):
3165         * kjs/internal.h:
3166         (KJS::InterpreterImp::setContext):
3167         (KJS::InterpreterImp::context):
3168         * kjs/interpreter.cpp:
3169         * kjs/interpreter.h:
3170         (KJS::ExecState::context):
3171         (KJS::ExecState::ExecState):
3172         * kjs/nodes.cpp:
3173         (currentSourceId):
3174         (currentSourceURL):
3175         (ThisNode::evaluate):
3176         (ResolveNode::evaluate):
3177         (FunctionCallResolveNode::evaluate):
3178         (PostfixResolveNode::evaluate):
3179         (DeleteResolveNode::evaluate):
3180         (TypeOfResolveNode::evaluate):
3181         (PrefixResolveNode::evaluate):
3182         (AssignResolveNode::evaluate):
3183         (VarDeclNode::evaluate):
3184         (VarDeclNode::processVarDecls):
3185         (DoWhileNode::execute):
3186         (WhileNode::execute):
3187         (ForNode::execute):
3188         (ForInNode::execute):
3189         (ContinueNode::execute):
3190         (BreakNode::execute):
3191         (ReturnNode::execute):
3192         (WithNode::execute):
3193         (SwitchNode::execute):
3194         (LabelNode::execute):
3195         (TryNode::execute):
3196         (FuncDeclNode::processFuncDecl):
3197         (FuncExprNode::evaluate):
3198
3199 2006-06-07  Geoffrey Garen  <ggaren@apple.com>
3200
3201         Removed API directory I prematurely/accidentally added.
3202
3203         * API: Removed.
3204
3205 2006-06-05  Mitz Pettel  <opendarwin.org@mitzpettel.com>
3206
3207         Reviewed and landed by Geoff.
3208
3209         - fix a regression in ecma_3/String/regress-104375.js
3210
3211         * kjs/string_object.cpp:
3212         (substituteBackreferences): If a 2-digit back reference is out of range,
3213         parse it as a 1-digit reference (followed by the other digit). This matches
3214         Firefox's behavior.
3215
3216 2006-06-05  Geoffrey Garen  <ggaren@apple.com>
3217
3218         Reviewed By Maciej.
3219         Darin already reviewed this change on the branch. See <rdar://problem/4317701>.
3220
3221         - Fixed <rdar://problem/4291345> PCRE overflow in Safari JavaScriptCore
3222
3223         No test case because there's no behavior change.
3224
3225         * pcre/pcre_compile.c:
3226         (read_repeat_counts): Check for integer overflow / out of bounds
3227
3228 2006-06-05  Geoffrey Garen  <ggaren@apple.com>
3229
3230         Reviewed by aliu.
3231
3232         - Changed CString length from int to size_t. We sould probably do this
3233         for UString, too. (Darin, if you're reading this: Maciej said so.)
3234
3235         * kjs/function.cpp:
3236         (KJS::encode):
3237         * kjs/ustring.cpp:
3238         (KJS::CString::CString):
3239         (KJS::operator==):
3240         * kjs/ustring.h:
3241         (KJS::CString::size):
3242
3243 2006-06-04  Geoffrey Garen  <ggaren@apple.com>
3244
3245         Reviewed by Maciej.
3246
3247         - http://bugzilla.opendarwin.org/show_bug.cgi?id=9304
3248         Minor cleanup in JavaScriptCore
3249
3250         * kjs/value.h: Removed redundant declarations
3251
3252 2006-06-04  Darin Adler  <darin@apple.com>
3253
3254         Reviewed by Anders.
3255
3256         - changed deleteAllValues so it can work on "const" collections
3257           Deleting the values affects the values, not the pointers in the
3258           collection, so it's legitimate to do it to a const collection,
3259           and a case of that actually came up in the XPath code.
3260
3261         * wtf/HashMap.h:
3262         (WTF::deleteAllPairSeconds): Use const iterators.
3263         (WTF::deleteAllValues): Take const HashMap reference as a parameter.
3264         * wtf/HashSet.h:
3265         (WTF::deleteAllValues): Take const HashSet reference as a parameter,
3266         and use const iterators.
3267         * wtf/Vector.h:
3268         (WTF::deleteAllValues): Take const Vector reference as a parameter.
3269
3270         - added more functions that are present in <math.h> on some platforms,
3271           but not on others; moved here from various files in WebCore
3272
3273         * wtf/MathExtras.h:
3274         (isinf): Added.
3275         (isnan): Added.
3276         (lround): Added.
3277         (lroundf): Tweaked.
3278         (round): Added.
3279         (roundf): Tweaked.
3280         (signbit): Added.
3281
3282 2006-06-02  Mitz Pettel  <opendarwin.org@mitzpettel.com>
3283
3284         Reviewed by ggaren.
3285
3286         - http://bugzilla.opendarwin.org/show_bug.cgi?id=9234
3287           Implement $&, $' and $` replacement codes in String.prototype.replace
3288
3289         Test: fast/js/string-replace-3.html
3290
3291         * kjs/string_object.cpp:
3292         (substituteBackreferences): Added support for $& (matched substring),
3293         $` (everything preceding matched substring), $' (everything following
3294         matched substring) and 2-digit back references, and cleaned up a little.
3295
3296 2006-06-02 Adele Peterson  <adele@apple.com>
3297
3298         Reviewed by Darin.
3299
3300         Set incremental linking to no. This seems to fix a build problem I was seeing
3301         where dftables couldn't find a dll.
3302
3303         * JavaScriptCore.vcproj/dftables/dftables.vcproj:
3304
3305 2006-05-26  Steve Falkenburg  <sfalken@apple.com>
3306
3307         Build fixes/tweaks
3308
3309         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
3310
3311 === JavaScriptCore-521.11 ===
3312
3313 2006-05-24  Geoffrey Garen  <ggaren@apple.com>
3314
3315         Reviewed by mjs.
3316
3317         - JSC half of fix for <rdar://problem/4557926> TOT REGRESSSION: Crash
3318         occurs when attempting to view image in slideshow mode at
3319         http://d.smugmug.com/gallery/581716 ( KJS::IfNode::execute
3320         (KJS::ExecState*) + 312)
3321
3322         On alternate threads, DOMObjects remain in the
3323         ScriptInterpreter's cache because they're not collected. So, they
3324         need an opportunity to mark their children.
3325
3326         I'm not particularly happy with this solution because it fails to
3327         resolve many outstanding issues with the DOM object cache. Since none
3328         of those issues is a crasher or a serious compatibility concern,
3329         and since the behavior of other browsers is not much to go on in this
3330         case, I've filed <rdar://problem/4561439> about that, and I'm moving on
3331         with my life.
3332
3333         * JavaScriptCore.xcodeproj/project.pbxproj:
3334         * kjs/collector.cpp:
3335         (KJS::Collector::collect):
3336         * kjs/internal.cpp:
3337         (KJS::InterpreterImp::mark):
3338         * kjs/internal.h:
3339         * kjs/interpreter.cpp:
3340         (KJS::Interpreter::mark):
3341         * kjs/interpreter.h:
3342
3343 === JavaScriptCore-521.10 ===
3344
3345 2006-05-22  Timothy Hatcher  <timothy@apple.com>
3346
3347         Reviewed by Eric, Kevin and Geoff.
3348
3349         Merge open source build fixes. <rdar://problem/4555500>
3350
3351         * kjs/collector.cpp: look at the rsp register in x86_64
3352         (KJS::Collector::markOtherThreadConservatively):
3353         * wtf/Platform.h: add x86_64 to the platform list
3354
3355 2006-05-19  Anders Carlsson  <acarlsson@apple.com>
3356
3357         Reviewed by Geoff.
3358
3359         http://bugzilla.opendarwin.org/show_bug.cgi?id=8993
3360         Support function declaration in case statements
3361
3362         * kjs/grammar.y: Get rid of StatementList and use SourceElements instead.
3363
3364         * kjs/nodes.cpp:
3365         (CaseClauseNode::evalStatements):
3366         (CaseClauseNode::processVarDecls):
3367         (CaseClauseNode::processFuncDecl):
3368         (ClauseListNode::processFuncDecl):
3369         (CaseBlockNode::processFuncDecl):
3370         (SwitchNode::processFuncDecl):
3371         * kjs/nodes.h:
3372         (KJS::CaseClauseNode::CaseClauseNode):
3373         (KJS::ClauseListNode::ClauseListNode):
3374         (KJS::ClauseListNode::getClause):
3375         (KJS::ClauseListNode::getNext):
3376         (KJS::ClauseListNode::releaseNext):
3377         (KJS::SwitchNode::SwitchNode):
3378         Add processFuncDecl for the relevant nodes.
3379
3380         * kjs/nodes2string.cpp:
3381         (CaseClauseNode::streamTo):
3382         next got renamed to source.
3383
3384 2006-05-17  George Staikos <staikos@kde.org>
3385
3386         Reviewed by Maciej, Alexey, and Eric.
3387
3388         * pcre/pcre_compile.c:
3389         * pcre/pcre_get.c:
3390         * pcre/pcre_exec.c:
3391         * wtf/UnusedParam.h:
3392         Use /**/ in .c files to compile with non-C99 and non-GCC compilers.
3393
3394         * kjs/testkjs.cpp:
3395         Change include to <wtf/HashTraits.h> from "HashTraits.h" to avoid -I
3396
3397         * wtf/unicode/qt4/UnicodeQt4.h:
3398         Use correct parentheses and correct mask for utf-32 support.
3399
3400 2006-05-17  Alexey Proskuryakov  <ap@nypop.com>
3401
3402         Reviewed by Darin.
3403
3404         - fix http://bugzilla.opendarwin.org/show_bug.cgi?id=8870
3405         Crash typing in Yahoo auto-complete widget.
3406
3407         Test: fast/js/regexp-stack-overflow.html
3408
3409         * pcre/pcre-config.h: Define NO_RECURSE.
3410
3411 2006-05-16  George Staikos <staikos@kde.org>
3412
3413         Reviewed by Maciej.
3414
3415         Fix some warnings and strict compilation errors.
3416
3417         * kjs/nodes.cpp:
3418         * kjs/value.cpp:
3419
3420 2006-05-15  Alexey Proskuryakov  <ap@nypop.com>
3421
3422         * make-generated-sources.sh: Changed to be executable and removed
3423         text in the file generated by "svn diff".
3424
3425 2006-05-15  Geoffrey Garen  <ggaren@apple.com>
3426
3427         Reviewed by Maciej.
3428
3429         - Fixed <rdar://problem/4534904> please do not treat "debugger" as
3430         a reserved word while parsing JavaScript (and other ECMA reserved
3431         words)
3432
3433         AKA
3434
3435         http://bugzilla.opendarwin.org/show_bug.cgi?id=6179
3436         We treat "char" as a reserved word in JavaScript and firefox/IE do
3437         not
3438
3439         (1) I unreserved most of the spec's "future reserved words" because
3440         they're not reserved in IE or FF. (Most, but not all, because IE
3441         somewhat randomly *does* reserve a few of them.)
3442         (2) I made 'debugger' a legitimate statement that acts like an empty
3443         statement because FF and IE support it.
3444
3445         * kjs/grammar.y:
3446         * kjs/keywords.table:
3447
3448 2006-05-15  Tim Omernick  <timo@apple.com>
3449
3450         Reviewed by John Sullivan.
3451
3452         Part of <rdar://problem/4466508> Add 64-bit support to the Netscape Plugin API
3453
3454         Added to the Netscape Plugin API the concept of "plugin drawing models".  The drawing model
3455         determines the kind of graphics context created by the browser for the plugin, as well as
3456         the Mac types of various Netscape Plugin API data structures.
3457
3458         There is a drawing model to represent the old QuickDraw-based API.  It is used by default
3459         if QuickDraw is available on the system, unless the plugin specifies another drawing model.
3460
3461         The big change is the addition of the CoreGraphics drawing model.  A plugin may request this
3462         drawing model to obtain access to a CGContextRef for drawing, instead of a QuickDraw CGrafPtr.
3463
3464         * bindings/npapi.h:
3465         Define NP_NO_QUICKDRAW when compiling 64-bit; there is no 64-bit QuickDraw.
3466         Added NPNVpluginDrawingModel, NPNVsupportsQuickDrawBool, and NPNVsupportsCoreGraphicsBool
3467         variables.
3468         Added NPDrawingModel enumeration.  Currently the only drawing models are QuickDraw and
3469         CoreGraphics.
3470         NPRegion's type now depends on the drawing model specified by the plugin.
3471         NP_Port is now only defined when QuickDraw is available.
3472         Added NP_CGContext, which is the type of the NPWindow's "window" member in CoreGraphics mode.
3473
3474 2006-05-13  Kevin M. Ollivier  <kevino@theolliviers.com>
3475
3476         Reviewed by Darin, landed by ap.
3477
3478         - http://bugzilla.opendarwin.org/show_bug.cgi?id=8528
3479           Bakefiles (and generated Makefiles) for wx and gdk ports
3480
3481         * make-generated-sources.sh:
3482         Added script to configure environment to run DerivedSources.make
3483
3484         * JavaScriptCoreSources.bkl:
3485         Added JavaScriptCore sources list for Bakefile.
3486
3487         * jscore.bkl:
3488         Bakefile used to generate JavaScriptCore project files
3489         (currently only used by wx and gdk ports)
3490
3491 2006-05-09  Steve Falkenburg  <sfalken@apple.com>
3492
3493         Fix Windows build.
3494         Minor fixes to WTF headers.
3495
3496         Reviewed by kevin.
3497
3498         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: Fix include dirs, paths to files.
3499         * JavaScriptCore.vcproj/testkjs/testkjs.vcproj: Fix include dirs.
3500         * wtf/Assertions.h: include Platform.h to get definition for COMPILER()
3501         * wtf/Vector.h: include FastMalloc.h for definition of fastMalloc, fastFree
3502
3503 2006-05-09  Maciej Stachowiak  <mjs@apple.com>
3504
3505         Rubber stamped by Anders.
3506
3507         - renamed kxmlcore to wtf
3508
3509         kxmlcore --> wtf
3510         KXMLCore --> WTF
3511         KXC --> WTF
3512
3513         * JavaScriptCore.xcodeproj/project.pbxproj:
3514         * bindings/c/c_instance.cpp:
3515         * bindings/objc/WebScriptObject.mm:
3516         * kjs/JSImmediate.h:
3517         * kjs/Parser.cpp:
3518         * kjs/Parser.h:
3519         * kjs/array_object.cpp:
3520         * kjs/collector.cpp:
3521         (KJS::Collector::registerThread):
3522         * kjs/collector.h:
3523         * kjs/config.h:
3524         * kjs/function.cpp:
3525         (KJS::isStrWhiteSpace):
3526         * kjs/function.h:
3527         * kjs/identifier.cpp:
3528         * kjs/internal.cpp:
3529         * kjs/internal.h:
3530         * kjs/lexer.cpp:
3531         (Lexer::shift):
3532         (Lexer::isWhiteSpace):
3533         (Lexer::isIdentStart):
3534         (Lexer::isIdentPart):
3535         * kjs/lookup.cpp:
3536         * kjs/nodes.cpp:
3537         * kjs/nodes.h:
3538         * kjs/number_object.cpp:
3539         * kjs/object.h:
3540         * kjs/property_map.cpp:
3541         * kjs/property_map.h:
3542         * kjs/string_object.cpp:
3543         (StringProtoFunc::callAsFunction):
3544         * kjs/testkjs.cpp:
3545         (testIsInteger):
3546         * kjs/ustring.cpp:
3547         * kjs/ustring.h:
3548         * kxmlcore: Removed.
3549         * kxmlcore/AlwaysInline.h: Removed.
3550         * kxmlcore/Assertions.cpp: Removed.
3551         * kxmlcore/Assertions.h: Removed.
3552         * kxmlcore/FastMalloc.cpp: Removed.
3553         * kxmlcore/FastMalloc.h: Removed.
3554         * kxmlcore/FastMallocInternal.h: Removed.
3555         * kxmlcore/Forward.h: Removed.
3556         * kxmlcore/HashCountedSet.h: Removed.
3557         * kxmlcore/HashFunctions.h: Removed.
3558         * kxmlcore/HashMap.h: Removed.
3559         * kxmlcore/HashSet.h: Removed.
3560         * kxmlcore/HashTable.cpp: Removed.
3561         * kxmlcore/HashTable.h: Removed.
3562         * kxmlcore/HashTraits.h: Removed.
3563         * kxmlcore/ListRefPtr.h: Removed.
3564         * kxmlcore/Noncopyable.h: Removed.
3565         * kxmlcore/OwnArrayPtr.h: Removed.
3566         * kxmlcore/OwnPtr.h: Removed.
3567         * kxmlcore/PassRefPtr.h: Removed.
3568         * kxmlcore/Platform.h: Removed.
3569         * kxmlcore/RefPtr.h: Removed.
3570         * kxmlcore/TCPageMap.h: Removed.
3571         * kxmlcore/TCSpinLock.h: Removed.
3572         * kxmlcore/TCSystemAlloc.cpp: Removed.
3573         * kxmlcore/TCSystemAlloc.h: Removed.
3574         * kxmlcore/UnusedParam.h: Removed.
3575         * kxmlcore/Vector.h: Removed.
3576         * kxmlcore/VectorTraits.h: Removed.
3577         * kxmlcore/unicode: Removed.
3578         * kxmlcore/unicode/Unicode.h: Removed.
3579         * kxmlcore/unicode/UnicodeCategory.h: Removed.
3580         * kxmlcore/unicode/icu: Removed.
3581         * kxmlcore/unicode/icu/UnicodeIcu.h: Removed.
3582         * kxmlcore/unicode/posix: Removed.
3583         * kxmlcore/unicode/qt3: Removed.
3584         * kxmlcore/unicode/qt4: Removed.
3585         * kxmlcore/unicode/qt4/UnicodeQt4.h: Removed.
3586         * pcre/pcre_get.c:
3587         * wtf: Added.
3588         * wtf/Assertions.cpp:
3589         * wtf/Assertions.h:
3590         * wtf/FastMalloc.cpp:
3591         (WTF::TCMalloc_ThreadCache::Scavenge):
3592         (WTF::do_malloc):
3593         (WTF::do_free):
3594         (WTF::TCMallocGuard::TCMallocGuard):
3595         (WTF::malloc):
3596         (WTF::free):
3597         (WTF::calloc):
3598         (WTF::cfree):
3599         (WTF::realloc):
3600         * wtf/FastMalloc.h:
3601         * wtf/FastMallocInternal.h:
3602         * wtf/Forward.h:
3603         * wtf/HashCountedSet.h:
3604         * wtf/HashFunctions.h:
3605         * wtf/HashMap.h:
3606         * wtf/HashSet.h:
3607         * wtf/HashTable.cpp:
3608         * wtf/HashTable.h:
3609         * wtf/HashTraits.h:
3610         * wtf/ListRefPtr.h:
3611         * wtf/Noncopyable.h:
3612         * wtf/OwnArrayPtr.h:
3613         * wtf/OwnPtr.h:
3614         * wtf/PassRefPtr.h:
3615         * wtf/RefPtr.h:
3616         * wtf/TCSystemAlloc.cpp:
3617         (TCMalloc_SystemAlloc):
3618         * wtf/Vector.h:
3619         * wtf/VectorTraits.h:
3620         * wtf/unicode/UnicodeCategory.h:
3621         * wtf/unicode/icu/UnicodeIcu.h:
3622
3623 2006-05-08  Timothy Hatcher  <timothy@apple.com>
3624
3625         Reviewed by Tim O.
3626
3627         * bindings/npapi.h: do not define #pragma options align=mac68k if we are 64-bit
3628
3629 2006-05-07  Darin Adler  <darin@apple.com>
3630
3631         Reviewed and landed by Maciej.
3632
3633         - fix http://bugzilla.opendarwin.org/show_bug.cgi?id=8765
3634         Random crashes on TOT since the form state change
3635
3636         I haven't figured out how to construct a test for this, but this does seem to fix the
3637         problem; Mitz mentioned that a double-destroy was occurring in these functions.
3638
3639         * kxmlcore/HashMap.h: (KXMLCore::HashMap::remove): Use RefCounter::deref instead of calling
3640         ~ValueType, because ~ValueType often results in a double-destroy, since the HashTable also
3641         destroys the element based on the storage type. The RefCounter template correctly does work
3642         only in cases where ValueType and ValueStorageType differ and this class is what's used
3643         elsewhere for the same purpose; I somehow missed this case when optimizing HashMap.
3644         * kxmlcore/HashSet.h: (KXMLCore::HashSet::remove): Ditto.
3645
3646 2006-05-05  Darin Adler  <darin@apple.com>
3647
3648         - http://bugzilla.opendarwin.org/show_bug.cgi?id=8722
3649           IE compatibility fix in date parsing
3650
3651         * kjs/date_object.cpp: (KJS::parseDate): Merged change that George Staikos provided
3652         from KDE 3.4.3 branch that allows day values of 0 and values that are > 1000.
3653
3654 2006-05-04  Anders Carlsson  <andersca@mac.com>
3655
3656         Reviewed by Maciej.
3657
3658         http://bugzilla.opendarwin.org/show_bug.cgi?id=8734
3659         Would like a Vector::append that takes another Vector
3660
3661         * kxmlcore/Vector.h:
3662         (KXMLCore::::append):
3663         New function that takes another array.
3664
3665 2006-05-02  Steve Falkenburg  <sfalken@apple.com>
3666
3667         Reviewed by eric.
3668
3669         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: set NDEBUG for release build
3670         * kxmlcore/FastMalloc.cpp: Prevent USE_SYSTEM_MALLOC from being defined twice
3671
3672 2006-05-02  Anders Carlsson  <andersca@mac.com>
3673
3674         Reviewed by Maciej.
3675
3676         * kxmlcore/HashMap.h:
3677         (KXMLCore::::operator):
3678         Return *this
3679
3680 2006-05-01  Tim Omernick  <timo@apple.com>
3681
3682         Reviewed by Tim Hatcher.
3683
3684         <rdar://problem/4476875> Support printing for embedded Netscape plugins
3685
3686         * bindings/npapi.h:
3687         Fixed struct alignment problem in our npapi.h.  Structs must be 68k-aligned on both pre-Mac OS X
3688         and Mac OS X systems, as this is what plugins expect.
3689
3690 2006-05-01  Timothy Hatcher  <timothy@apple.com>
3691
3692         Reviewed by Maciej.
3693
3694         <rdar://problem/4308243> 8F36 Regression: crash in malloc_consolidate if you use a .PAC file
3695
3696         The original fix missed the oversized cell case. Added a test for "currentThreadIsMainThread ||
3697         imp->m_destructorIsThreadSafe" where we collect oversized cells.
3698
3699         We don't have a way to test PAC files yet, so there's no test attached.
3700
3701         * kjs/collector.cpp:
3702         (KJS::Collector::collect): test the thread when we collect oversized cells
3703
3704 2006-05-01  Tim Omernick  <timo@apple.com>
3705
3706         Reviewed by Adele.
3707
3708         <rdar://problem/4526114> REGRESSION (two days ago): LOG() just prints @ for NSObject substitutions
3709
3710         * kxmlcore/Assertions.cpp:
3711         Changed sense of strstr("%@") check.  I already made the same fix to the WebBrowser assertions.
3712
3713 2006-04-28  Steve Falkenburg  <sfalken@apple.com>
3714
3715         Reviewed by kdecker
3716
3717         Actually apply the change that was reviewed insted of checking it in with an #if 0 (oops).
3718
3719         * kjs/testkjs.cpp:
3720         (main): Suppress C runtime alerts
3721
3722 2006-04-28  Steve Falkenburg  <sfalken@apple.com>
3723
3724         Reviewed by kdecker
3725
3726         Suppress error reporting dialog that blocks Javascript tests from completing.
3727
3728         Real error is due to an overflow in the date/time handling functions that needs
3729         to be addressed, but this will prevent the hang running the Javascript tests
3730         on the build bot (along with the related changes).
3731
3732         * kjs/testkjs.cpp:
3733         (main): Suppress C runtime alerts
3734
3735 2006-04-27  Geoffrey Garen  <ggaren@apple.com>
3736
3737         Reviewed by Maciej
3738
3739         - Minor fixups I discovered while working on the autogenerator.
3740
3741         * kjs/lookup.cpp:
3742         (findEntry): ASSERT that size is not 0, because otherwise we'll % by 0,
3743         compute a garbage address, and possibly crash.
3744         * kjs/lookup.h:
3745         (cacheGlobalObject): Don't enumerate cached objects -- ideally, they
3746         would be hidden entirely.
3747
3748 2006-04-21  Kevin M. Ollivier  <kevino@theolliviers.com>
3749
3750         Reviewed by Darin.
3751
3752         - http://bugzilla.opendarwin.org/show_bug.cgi?id=8507
3753           Compilation fixes for building on gcc 4.0.2, and without precomp headers
3754
3755         * kjs/operations.h:
3756         * kxmlcore/Assertions.cpp:
3757         * kxmlcore/FastMalloc.cpp:
3758         Added necessary headers to resolve compilation issues when not using
3759         precompiled headers.
3760
3761         * kjs/value.h: Declare the JSCell class before friend declaration
3762         to resolve compilation issues with gcc 4.0.2.
3763
3764         * kxmlcore/Platform.h: Set Unicode support to use ICU on platforms
3765         other than KDE (previously only defined for Win and Mac OS)
3766
3767 2006-04-18  Eric Seidel  <eseidel@apple.com>
3768
3769         Reviewed by ggaren.
3770
3771         Fix "new Function()" to correctly use lexical scoping.
3772         Add ScopeChain::print() function for debugging.
3773         <rdar://problem/4067864> REGRESSION (125-407): JavaScript failure on PeopleSoft REN Server
3774
3775         * kjs/function_object.cpp:
3776         (FunctionObjectImp::construct):
3777         * kjs/scope_chain.cpp:
3778         (KJS::ScopeChain::print):
3779         * kjs/scope_chain.h:
3780
3781 2006-04-14  James G. Speth  <speth@end.com>
3782
3783         Reviewed by Timothy.
3784
3785         Bug 8389: support for Cocoa bindings - binding an NSTreeController to the WebView's DOM
3786         http://bugzilla.opendarwin.org/show_bug.cgi?id=8389
3787
3788         Adds a category to WebScriptObject with array accessors for KVC/KVO.
3789
3790         If super valueForKey: fails it will call valueForUndefinedKey:, which is
3791         important because it causes the right behavior to happen with bindings using
3792         the "Raises for Not Applicable Keys" flag and the "Not Applicable Placeholder"
3793
3794         * bindings/objc/WebScriptObject.mm:
3795         (-[WebScriptObject valueForKey:]):
3796         (-[WebScriptObject count]):
3797         (-[WebScriptObject objectAtIndex:]):
3798         (-[WebUndefined description]): return "undefined"
3799
3800 2006-04-13  Geoffrey Garen  <ggaren@apple.com>
3801
3802         Reviewed by Darin.
3803
3804         * kjs/internal.cpp:
3805         (KJS::InterpreterImp::initGlobalObject): Add the built-in object
3806         prototype to the end of the global object's prototype chain instead of
3807         just blowing away its existing prototype. We need to do this because
3808         the window object has a meaningful prototype now.
3809
3810 2006-04-13  Maciej Stachowiak  <mjs@apple.com>
3811
3812         Reviewed by Geoff.
3813
3814         - fix testkjs to not show false-positive KJS::Node leaks in debug builds
3815
3816         * kjs/testkjs.cpp:
3817         (doIt):
3818         (kjsmain):
3819
3820 2006-04-11  Geoffrey Garen  <ggaren@apple.com>
3821
3822         Reviewed by Maciej.
3823
3824         Minor code cleanup -- passes all the JS tests.
3825
3826         * kjs/object_object.cpp:
3827         (ObjectObjectImp::construct):
3828         (ObjectObjectImp::callAsFunction):
3829
3830 2006-04-11  Darin Adler  <darin@apple.com>
3831
3832         - another attempt to fix Windows build -- Vector in Forward.h was not working
3833
3834         * kxmlcore/Forward.h: Remove Vector.
3835         * kxmlcore/Vector.h: Add back default arguments, remove include of
3836         Forward.h.
3837
3838 2006-04-11  Darin Adler  <darin@apple.com>
3839
3840         - try to fix Windows build -- HashForward.h was not working
3841
3842         * kxmlcore/HashForward.h: Removed.
3843
3844         * JavaScriptCore.xcodeproj/project.pbxproj: Remove HashForward.h.
3845         * kjs/collector.h: Remove use of HashForward.h.
3846         * kxmlcore/HashCountedSet.h: Remove include of HashForward.h, restore
3847         default arguments.
3848         * kxmlcore/HashMap.h: Ditto.
3849         * kxmlcore/HashSet.h: Ditto.
3850
3851 2006-04-11  David Harrison  <harrison@apple.com>
3852
3853         Reviewed by Darin.
3854
3855         - fixed clean build, broken by Darin's check-in
3856
3857         * kjs/date_object.cpp: Add needed include of lookup.h.
3858         * kjs/regexp_object.cpp: Move include of .lut.h file below other includes.
3859
3860 2006-04-10  Darin Adler  <darin@apple.com>
3861
3862         Rubber-stamped by John Sullivan.
3863
3864         - switched from a shell script to a makefile for generated files
3865         - removed lots of unneeded includes
3866         - added new Forward.h and HashForward.h headers that allow compiling with
3867           fewer unneeded templates
3868
3869         * DerivedSources.make: Added.
3870         * generate-derived-sources: Removed.
3871         * JavaScriptCore.xcodeproj/project.pbxproj: Added new files, changed to use
3872         DerivedSources.make.
3873
3874         * kxmlcore/Forward.h: Added.
3875         * kxmlcore/HashForward.h: Added.
3876
3877         * kxmlcore/HashCountedSet.h: Include HashForward for default args.
3878         * kxmlcore/HashMap.h: Ditto.
3879         * kxmlcore/HashSet.h: Ditto.
3880
3881         * kjs/object.h:
3882         * kjs/object.cpp:
3883         Moved KJS_MAX_STACK into the .cpp file.
3884
3885         * bindings/NP_jsobject.cpp:
3886         * bindings/c/c_instance.h:
3887         * bindings/jni/jni_class.h:
3888         * bindings/jni/jni_runtime.h:
3889         * bindings/jni/jni_utility.h:
3890         * bindings/objc/WebScriptObject.mm:
3891         * bindings/objc/WebScriptObjectPrivate.h:
3892         * bindings/objc/objc_class.h:
3893         * bindings/objc/objc_class.mm:
3894         * bindings/objc/objc_instance.h:
3895         * bindings/objc/objc_instance.mm:
3896         * bindings/objc/objc_runtime.mm:
3897         * bindings/objc/objc_utility.mm:
3898         * bindings/runtime.h:
3899         * bindings/runtime_array.cpp:
3900         * bindings/runtime_array.h:
3901         * bindings/runtime_method.cpp:
3902         * bindings/runtime_method.h:
3903         * bindings/runtime_object.cpp:
3904         * bindings/runtime_root.h:
3905         * kjs/JSImmediate.cpp:
3906         * kjs/Parser.h:
3907         * kjs/array_object.cpp:
3908         * kjs/array_object.h:
3909         * kjs/bool_object.cpp:
3910         * kjs/bool_object.h:
3911         * kjs/collector.h:
3912         * kjs/context.h:
3913         * kjs/debugger.cpp:
3914         * kjs/error_object.h:
3915         * kjs/function_object.h:
3916         * kjs/internal.h:
3917         * kjs/lexer.cpp:
3918         * kjs/math_object.cpp:
3919         * kjs/math_object.h:
3920         * kjs/nodes.cpp:
3921         * kjs/nodes.h:
3922         * kjs/number_object.cpp:
3923         * kjs/number_object.h:
3924         * kjs/object_object.cpp:
3925         * kjs/operations.cpp:
3926         * kjs/protected_reference.h:
3927         * kjs/reference.h:
3928         * kjs/reference_list.h:
3929         * kjs/regexp_object.h:
3930         * kjs/string_object.cpp:
3931         * kjs/string_object.h:
3932         * kjs/testkjs.cpp:
3933         * kjs/value.cpp:
3934         * kjs/value.h:
3935         * kxmlcore/HashTable.h:
3936         * kxmlcore/ListRefPtr.h:
3937         * kxmlcore/TCPageMap.h:
3938         * kxmlcore/Vector.h:
3939         Removed unneeded header includes.
3940
3941 2006-04-09  Geoffrey Garen  <ggaren@apple.com>
3942
3943         Reviewed by eric.
3944
3945         - Fixed http://bugzilla.opendarwin.org/show_bug.cgi?id=8284
3946         prevent unnecessary entries in the "nodes with extra refs" hash table
3947
3948         This patch switches manually RefPtr exchange with use of
3949         RefPtr::release to ensure that a node's ref count never tops 1
3950         (in the normal case).
3951
3952         * kjs/nodes.cpp:
3953         (BlockNode::BlockNode):
3954         (CaseBlockNode::CaseBlockNode):
3955         * kjs/nodes.h:
3956         (KJS::ArrayNode::ArrayNode):
3957         (KJS::ObjectLiteralNode::ObjectLiteralNode):
3958         (KJS::ArgumentsNode::ArgumentsNode):
3959         (KJS::VarStatementNode::VarStatementNode):
3960         (KJS::ForNode::ForNode):
3961         (KJS::CaseClauseNode::CaseClauseNode):
3962         (KJS::FuncExprNode::FuncExprNode):
3963         (KJS::FuncDeclNode::FuncDeclNode):
3964
3965 2006-04-08  Alexey Proskuryakov  <ap@nypop.com>
3966
3967         Reviewed by Darin.
3968
3969         One more attempt - use reinterpret_cast, rather than static_cast.
3970
3971 2006-04-08  Alexey Proskuryakov  <ap@nypop.com>
3972
3973         Reviewed by Darin.
3974
3975         An attempt to fix Win32 build - ICU uses wchar_t on Windows, so we need a type cast.
3976
3977         * kxmlcore/unicode/icu/UnicodeIcu.h:
3978         (KXMLCore::Unicode::toLower):
3979         (KXMLCore::Unicode::toUpper):
3980
3981 2006-04-08  Alexey Proskuryakov  <ap@nypop.com>
3982
3983         Reviewed by Darin.
3984
3985         - fix http://bugzilla.opendarwin.org/show_bug.cgi?id=8264
3986         toLowerCase and toUpperCase don't honor special mappings
3987
3988         Test: fast/js/string-capitalization.html
3989
3990         * JavaScriptCore.xcodeproj/project.pbxproj: Added KXMLCore::Unicode headers to the project.
3991         * icu/unicode/putil.h: Added (copied from WebCore).
3992         * icu/unicode/uiter.h: Ditto.
3993         * icu/unicode/ustring.h: Ditto.
3994         * kjs/string_object.cpp:
3995         (StringProtoFunc::callAsFunction): Use the new KXMLCore::Unicode::toUpper() and toLower().
3996         * kjs/ustring.cpp: Removed unused (and evil) UChar::toLower() and toUpper().
3997         * kjs/ustring.h: Ditto.
3998
3999         * kxmlcore/unicode/Unicode.h: Corrected capitalization of the word Unicode.
4000         * kxmlcore/unicode/UnicodeCategory.h: Renamed include guard macro to match file name.
4001
4002         * kxmlcore/unicode/icu/UnicodeIcu.h:
4003         (KXMLCore::Unicode::toLower): Work on strings, not individual characters. Use ICU root locale.
4004         (KXMLCore::Unicode::toUpper): Ditto.
4005         (KXMLCore::Unicode::isFormatChar): Use int32_t, which can hold a complete code point.
4006         (KXMLCore::Unicode::isSeparatorSpace): Ditto.
4007         (KXMLCore::Unicode::category): Ditto.
4008         * kxmlcore/unicode/qt4/UnicodeQt4.h:
4009         (KXMLCore::Unicode::toLower): Work on strings, not individual characters.
4010         (KXMLCore::Unicode::toUpper): Ditto.
4011         (KXMLCore::Unicode::isFormatChar): Use int32_t, which can hold a complete code point.
4012         (KXMLCore::Unicode::isSeparatorSpace): Ditto.
4013         (KXMLCore::Unicode::category): Ditto.
4014
4015         * tests/mozilla/ecma/String/15.5.4.12-1.js: Corrected expected results.
4016         * tests/mozilla/ecma/String/15.5.4.12-5.js: Corrected expected results.
4017
4018 2006-04-05  Darin Adler  <darin@apple.com>
4019
4020         - attempt to fix Windows build
4021
4022         * kxmlcore/HashMap.h: (KXMLCore::HashMap::remove): Use (*it). instead of it->.
4023         * kxmlcore/HashSet.h: (KXMLCore::HashSet::remove): Ditto.
4024
4025 2006-04-05  Darin Adler  <darin@apple.com>
4026
4027         - attempt to fix Windows build
4028
4029         * os-win32/stdint.h: Add int8_t, uint8_t, int64_t.
4030
4031 2006-04-05  Darin Adler  <darin@apple.com>
4032
4033         Reviewed by Maciej.
4034
4035         - fix memory leak introduced by the previous change
4036
4037         * kxmlcore/HashTable.h: Specialize NeedsRef so that it correctly returns true when
4038         the value in question is a pair where one of the pair needs a ref and the other
4039         of the pair does not.
4040
4041 2006-04-05  Darin Adler  <darin@apple.com>
4042
4043         Reviewed by Maciej.
4044
4045         - JavaScriptCore part of fix for http://bugzilla.opendarwin.org/show_bug.cgi?id=8049
4046           StringImpl hash traits deleted value creates an init routine for WebCore
4047           <rdar://problem/4442248> REGRESSION: WebCore has init routines (8049)
4048
4049         Change HashMap and HashSet implementation so they fold various types together.
4050         This allows us to implement maps and sets that use RefPtr<WebCore::StringImpl>
4051         and WebCore::String in terms of the underlying raw pointer type, and hence use
4052         -1 for the deleted value.
4053
4054         * kxmlcore/HashTraits.h: Added a new type to HashTraits, StorageTraits, which is a
4055         type to be used when storing a value that has the same layout as the type itself.
4056         This is used only for non-key cases. In the case of keys, the hash function must also
4057         be considered. Moved emptyValue out of GenericHashTraitsBase into GenericHashTraits.
4058         Added a new bool to HashTraits, needsRef, which indicates whether the type needs
4059         explicit reference counting. If the type itself has needsRef true, but the storage
4060         type has needsRef false, then the HashSet or HashMap has to handle the reference
4061         counting explicitly. Added hash trait specializations for all signed integer values
4062         that give -1 as the deleted value. Gave all integers StorageTraits of the canonical
4063         integer type of the same size so int and long will share code. Gave all pointers and
4064         RefPtrs StorageTraits of the appropriately sized integer type. Removed redundant
4065         TraitType and emptyValue definitions in the pointer specialization for HashTraits.
4066         Added PairBaseHashTraits, which doesn't try to set up needsDestruction and deletedValue.
4067         Useful for types where we don't want to force the existence of deletedValue, such as
4068         the type of a pair in a HashMap which is not the actual storage type. Removed an
4069         unneeded parameter from the DeletedValueAssigner template. Added HashKeyStorageTraits
4070         template, which determines what type can be used to store a given hash key type with
4071         a given hash function, and specialized it for pointers and RefPtr so that pointer
4072         hash tables share an underlying HashTable that uses IntHash.
4073
4074         * kxmlcore/HashTable.h: Added HashTableConstIteratorAdapter, HashTableIteratorAdapter,
4075         NeedsRef, RefCountManagerBase, RefCountManager, HashTableRefCountManagerBase, and
4076         HashTableRefCountManager. All are used by both HashSet and HashMap to handle hash
4077         tables where the type stored is not the same as the real value type.
4078
4079         * kxmlcore/HashFunctions.h: Added a new struct named IntTypes that finds an
4080         integer type given a sizeof value. Renamed pointerHash to intHash and made it
4081         use overloading and take integer parameters. Added an IntHash struct which is
4082         a hash function that works for integers. Changed PtrHash to call IntHash with
4083         an appropriately sized integer. Made IntHash the default hash function for
4084         many integer types. Made PtrHash the default hash function for RefPtr as well
4085         as for raw pointers.
4086
4087         * kxmlcore/HashSet.h: Changed implementation to use a separate "storage type"
4088         derived from the new traits. The HashTable will use the storage type and all
4089         necessary translation and ref/deref is done at the HashSet level. Also reorganized
4090         the file so that the HashSet is at the top and has no inline implementation inside
4091         it so it's easy to read the interface to HashSet.
4092
4093         * kxmlcore/HashMap.h: Changed implementation to use a separate "storage type"
4094         derived from the new traits. The HashTable will use the storage type and all
4095         necessary translation and ref/deref is done at the HashMap level. Also reorganized
4096         the file so that the HashMap is at the top and has no inline implementation inside
4097         it so it's easy to read the interface to HashMap.
4098
4099         * kxmlcore/HashMapPtrSpec.h: Removed. Superceded by optimizations in HashMap itself.
4100
4101         * JavaScriptCore.xcodeproj/project.pbxproj: Remove HashMapPtrSpec.h, resort files,
4102         and also remove some unnecessary build settings from the aggregate target that
4103         generates derived sources.
4104         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: Ditto.
4105
4106 2006-04-04  Timothy Hatcher  <timothy@apple.com>
4107
4108         Reviewed by Darin.
4109
4110         The Debug and Release frameworks are now built with install paths relative to the build products directory.
4111         This removes the need for other projects to build with -framework WebCore and -framework JavaScriptCore.
4112
4113         * JavaScriptCore.xcodeproj/project.pbxproj:
4114
4115 2006-04-04  Eric Seidel  <eseidel@apple.com>
4116
4117         Reviewed by ggaren.
4118
4119         Fix win32 build.
4120         Disable ASSERT redefinition warnings for now.
4121
4122         * JavaScriptCore.vcproj/testkjs/testkjs.vcproj:
4123         * kxmlcore/Assertions.h:
4124
4125 2006-04-04  Bjrn Graf  <bjoern.graf@gmail.com>
4126
4127         Reviewed by ggaren & darin.  Landed by eseidel.
4128
4129         Integrate CURL version of gettimeofday
4130         http://bugzilla.opendarwin.org/show_bug.cgi?id=7399
4131         Disable crash report dialogs for testkjs.exe in Release mode
4132         http://bugzilla.opendarwin.org/show_bug.cgi?id=8113
4133
4134         * kjs/testkjs.cpp:
4135         (StopWatch::start):
4136         (StopWatch::stop):
4137         (StopWatch::getElapsedMS):
4138         (main):
4139         (kjsmain):
4140
4141 2006-04-04  Eric Seidel  <eseidel@apple.com>
4142
4143         Reviewed by mjs.
4144
4145         * kjs/number_object.cpp:
4146         (NumberProtoFunc::callAsFunction): remove trunc() to fix win32.
4147
4148 2006-03-12  Maciej Stachowiak  <mjs@apple.com>
4149
4150         Reviewed by Darin.
4151
4152         - fixed "toPrecision sometimes messes up the last digit on intel Macs"
4153         http://bugzilla.opendarwin.org/show_bug.cgi?id=7748
4154
4155         * kjs/number_object.cpp:
4156         (intPow10): Compute integer powers of 10 using exponentiation by squaring.
4157         (NumberProtoFunc::callAsFunction): Use intPow10(n) in place of all pow(10.0, n),
4158         plus a bit of refactoring.
4159
4160 2006-04-03  Darin Adler  <darin@apple.com>
4161
4162         - tweak config.h and Platform.h to try to get buildbot working
4163           (making some small changes at the same time)
4164
4165         * kjs/config.h: Removed now-unneeded HAVE_ICU.
4166         * kxmlcore/