JavaScriptCore/ChangeLog

   1 2006-07-19  Anders Carlsson  <acarlsson@apple.com>
   2
   3         Reviewed by Darin.
   4
   5         <rdar://problem/4620655> REGRESSION(10.4.7-10.5): preview button for a blogger.com post doesn't work
   6
   7         * kjs/nodes2string.cpp:
   8         (StringNode::streamTo):
   9         Return the escaped string.
  10
  11         (RegExpNode::streamTo):
  12         Use the correct syntax.
  13
  14         * kjs/function.cpp:
  15         (KJS::escapeStringForPrettyPrinting):
  16         * kjs/function.h:
  17         Add escape function which escapes a string for pretty-printing so it can be parsed again.
  18
  19         * wtf/unicode/icu/UnicodeIcu.h:
  20         (WTF::Unicode::isPrintableChar):
  21         New function.
  22
  23 === Safari-521.19 ===
  24
  25 2006-07-18  Maciej Stachowiak  <mjs@apple.com>
  26
  27         Reviewed by Adele Peterson.
  28
  29         <rdar://problem/4589530> REGRESSION: null character in JS string causes parse error (works in Tiger and in other browsers)
  30
  31         * kjs/lexer.cpp:
  32         (Lexer::shift):
  33         (Lexer::lex):
  34         (Lexer::record16):
  35         (Lexer::scanRegExp):
  36         * kjs/lexer.h:
  37
  38 2006-07-18  Tim Omernick  <timo@apple.com>
  39
  40         Reviewed by Tim Hatcher.
  41
  42         Removed a misleading comment; we recently added support for the NPNVPluginElementNPObject
  43         variable.
  44
  45         * bindings/npapi.h:
  46
  47 === Safari-521.18 ===
  48
  49 2006-07-18  Timothy Hatcher  <timothy@apple.com>
  50
  51         Made the following headers public:
  52
  53         * JavaScriptCore.h
  54         * JSBase.h
  55         * JSContextRef.h
  56         * JSObjectRef.h
  57         * JSStringRef.h
  58         * JSValueRef.h
  59
  60         * JavaScriptCore.xcodeproj/project.pbxproj:
  61
  62 2006-07-17  Geoffrey Garen  <ggaren@apple.com>
  63
  64         Reviewed by Maciej.
  65
  66         - Added automatic prototype creation for classes.
  67
  68         A class stores a weak reference to a prototype, which is cleared when
  69         the prototype is garbage collected, to avoid a reference cycle.
  70
  71         We now have an attributes field in JSClassDefinition, that currently is
  72         used only to override automatic prototype creation when you want to manage your
  73         own prototypes, but can be extended in the future for other nefarious purposes.
  74
  75         Similarly, we have JSObjectMake and JSObjectMakeWithPrototype, the latter
  76         allowing you to manage your own prototypes.
  77
  78         JSObjectMakeConstructor is more interesting now, able to make a constructor
  79         on your behalf if you just give it a class.
  80
  81         - Removed bogus old code from minidom.js.
  82
  83         - Tweaked the headerdocs.
  84
  85         - Added more GC testing, which caught some leaks, and tested more funny
  86         edge cases in lookup, which caught a lookup bug. Removed some testing
  87         we used to do with MyObject because it was redundant with the new, cool
  88         stuff.
  89
  90         While fixing the lookup bug I retracted this change:
  91
  92             "If a static setProperty callback returns 'false', to indicate that the
  93             property was not set, we no longer forward the set request up the class
  94             chain, because that's almost certainly not what the programmer expected."
  95
  96         Returning false when setting a static property is a little silly, but you can see
  97         it being useful when shadowing a base class's static properties, and, regardless
  98         of usefullness, this is the defined behavior of the setProperty callback.
  99
 100         - Plus a little ASCII art, for the kids.
 101
 102 2006-07-17  Timothy Hatcher  <timothy@apple.com>
 103
 104         Reviewed by Maciej.
 105
 106         <rdar://problem/4634874> WebScriptObject and WebUndefined are no longer defined by WebKit
 107
 108         Moves WebScriptObject and WebUndefined up to WebCore.
 109         This change does create an upwards-dependancy on WebScriptObject existing
 110         in the loaded process, but this code path in JavaScriptCore does not get used
 111         unless it is through WebKit/WebCore. Moving all of the binding code out of
 112         JavaScriptCore might make sense in the future.
 113
 114         * JavaScriptCore.exp:
 115         * JavaScriptCore.xcodeproj/project.pbxproj:
 116         * bindings/objc/WebScriptObject.h: Replaced.
 117         * bindings/objc/WebScriptObject.mm: Removed.
 118         * bindings/objc/WebScriptObjectPrivate.h: Removed.
 119         * bindings/objc/objc_class.h:
 120         * bindings/objc/objc_instance.h:
 121         * bindings/objc/objc_instance.mm:
 122         (ObjcInstance::~ObjcInstance):
 123         * bindings/objc/objc_runtime.h:
 124         * bindings/objc/objc_runtime.mm:
 125         (convertValueToObjcObject):
 126         * bindings/objc/objc_utility.mm:
 127         (KJS::Bindings::convertValueToObjcValue):
 128         (KJS::Bindings::convertObjcValueToValue):
 129         (KJS::Bindings::createObjcInstanceForValue):
 130
 131 2006-07-17  Darin Adler  <darin@apple.com>
 132
 133         * API/JSBase.h: Fix comment formatting where things used to be lined up but
 134         are now ragged. Got rid of spaces that attempted to line things up.
 135         * API/JSObjectRef.h: Ditto. Also add missing periods for a couple of comments.
 136
 137 2006-07-17  Geoffrey Garen  <ggaren@apple.com>
 138
 139         Reviewed by Maciej.
 140
 141         - Removed the exception parameter from the initialize callback and, by extension,
 142         JSObjectMake. We have never had a need for exceptions when iniitializing,
 143         so the parameter seemed likely to "get in the way."
 144
 145         Also, an exception in JavaScript is thrown in response to input --
 146         "invalid URL", "index not a number", etc., so it's the job of the
 147         constructor function, not the initialize method, to throw.
 148
 149         If initialize *really* wants to throw, it can communicate the throw to
 150         the constructor through the constructed object's private data (e.g., set
 151         it to NULL, signaling to the consntructor that initialization failed.)
 152
 153         - Added JSObjectMakeWithData, which enables a constructor to set private
 154         data on an object *before* it has been initialized. That way, the initialize
 155         methods can properly operate on the data.
 156
 157         * API/JSNode.c: Moved ref into the initialize method, for better encapsulation,
 158         now that it's possible.
 159         * API/JSNodeList.c: ditto
 160         * API/minidom.c:
 161         (main): Do more aggressive garbage collection to test ref/deref and
 162         initialize/finalize.
 163         * API/minidom.js: store childNodes in a temporary so it doesn't get re-created
 164         like a thousand times. This makes debugging ref/deref easier
 165
 166 2006-07-17  Geoffrey Garen  <ggaren@apple.com>
 167
 168         Reviewed by Maciej.
 169
 170         - Changed the initialize callback to run from least derived class (parent
 171         class) to most derived class. This enables C++ style initialization,
 172         and derived class overriding of member data.
 173
 174         - Added excpetion propopgation to JSObjectMake, to support initialize
 175         exceptions, and generally round out our policy of making function
 176         signatures as long as possible.
 177
 178         * API/JSCallbackObject.h: Use ExecState instead of ContextRef, cuz we're
 179         in C++ land now.
 180
 181 2006-07-17  Geoffrey Garen  <ggaren@apple.com>
 182
 183         Reviewed by Maciej.
 184
 185         - Changed JSObjectMakeConstructor to JSObjectMakeConstructorWithCallback,
 186         to match JSObjectMakeFunctionWithCallback.
 187
 188         - Added prototype parameter, so the generated constructor
 189         automatically works with hasInstance / instanceof
 190
 191         - Moved hasInstance implementation from InternalFunctionImp to JSObject
 192         so that subclasses can inherit it without inheriting function-related baggage.
 193         More refactoring here would be good, but this seems like a good short-term
 194         solution.
 195
 196         (KJS::JSCallbackFunction::implementsHasInstance): override and return false,
 197         because callback functions aren't constructors.
 198
 199 2006-07-17  Maciej Stachowiak  <mjs@apple.com>
 200
 201         Reviewed by Geoff.
 202
 203         - add a JSContextRef parameter to all JSValueRef, JSObjectRef, and JSContextRef operations;
 204         except JSObject{Get,Set}PrivateData which can be assumed to be simple pure accessors.
 205
 206         Also renamed the parameter "context" to "ctx" because it makes the code read better with this pervasive
 207         but usually uninteresting parameter.
 208
 209         * API/JSBase.cpp:
 210         (JSEvaluateScript):
 211         (JSCheckScriptSyntax):
 212         (JSGarbageCollect):
 213         * API/JSBase.h:
 214         * API/JSCallbackObject.cpp:
 215         (KJS::JSCallbackObject::JSCallbackObject):
 216         (KJS::JSCallbackObject::init):
 217         (KJS::JSCallbackObject::getOwnPropertySlot):
 218         (KJS::JSCallbackObject::put):
 219         (KJS::JSCallbackObject::deleteProperty):
 220         (KJS::JSCallbackObject::toNumber):
 221         (KJS::JSCallbackObject::toString):
 222         * API/JSContextRef.cpp:
 223         (JSGlobalContextCreate):
 224         (JSGlobalContextRetain):
 225         (JSGlobalContextRelease):
 226         (JSContextGetGlobalObject):
 227         * API/JSContextRef.h:
 228         * API/JSNode.c:
 229         (JSNodePrototype_appendChild):
 230         (JSNodePrototype_removeChild):
 231         (JSNodePrototype_replaceChild):
 232         (JSNode_getNodeType):
 233         (JSNode_getFirstChild):
 234         (JSNode_prototype):
 235         * API/JSNodeList.c:
 236         (JSNodeListPrototype_item):
 237         (JSNodeList_length):
 238         (JSNodeList_getProperty):
 239         (JSNodeList_prototype):
 240         * API/JSObjectRef.cpp:
 241         (JSObjectMake):
 242         (JSObjectMakeFunctionWithCallback):
 243         (JSObjectMakeConstructor):
 244         (JSObjectMakeFunction):
 245         (JSObjectGetPrototype):
 246         (JSObjectSetPrototype):
 247         (JSObjectHasProperty):
 248         (JSObjectGetProperty):
 249         (JSObjectSetProperty):
 250         (JSObjectGetPropertyAtIndex):
 251         (JSObjectSetPropertyAtIndex):
 252         (JSObjectDeleteProperty):
 253         (JSObjectIsFunction):
 254         (JSObjectCallAsFunction):
 255         (JSObjectIsConstructor):
 256         (JSObjectCallAsConstructor):
 257         (JSObjectCopyPropertyNames):
 258         * API/JSObjectRef.h:
 259         * API/JSStringRef.cpp:
 260         * API/JSValueRef.cpp:
 261         (JSValueGetType):
 262         (JSValueIsUndefined):
 263         (JSValueIsNull):
 264         (JSValueIsBoolean):
 265         (JSValueIsNumber):
 266         (JSValueIsString):
 267         (JSValueIsObject):
 268         (JSValueIsObjectOfClass):
 269         (JSValueIsEqual):
 270         (JSValueIsStrictEqual):
 271         (JSValueIsInstanceOfConstructor):
 272         (JSValueMakeUndefined):
 273         (JSValueMakeNull):
 274         (JSValueMakeBoolean):
 275         (JSValueMakeNumber):
 276         (JSValueMakeString):
 277         (JSValueToBoolean):
 278         (JSValueToNumber):
 279         (JSValueToStringCopy):
 280         (JSValueToObject):
 281         (JSValueProtect):
 282         (JSValueUnprotect):
 283         * API/JSValueRef.h:
 284         * API/minidom.c:
 285         (print):
 286         * API/testapi.c:
 287         (MyObject_getProperty):
 288         (MyObject_deleteProperty):
 289         (MyObject_callAsFunction):
 290         (MyObject_callAsConstructor):
 291         (MyObject_convertToType):
 292         (print_callAsFunction):
 293         (main):
 294
 295 2006-07-16  Geoffrey Garen  <ggaren@apple.com>
 296
 297         Approved by Maciej, RS by Beth.
 298
 299         JSObjectMakeFunction -> JSObjectMakeFunctionWithCallback
 300         JSObjectMakeFunctionWithBody -> JSObjectMakeFunction
 301
 302         because the latter is more common, and more fundamental, than the former.
 303
 304         * API/APICast.h:
 305         (toJS):
 306         * API/JSBase.h:
 307         * API/JSCallbackObject.cpp:
 308         (KJS::JSCallbackObject::getOwnPropertySlot):
 309         (KJS::JSCallbackObject::put):
 310         (KJS::JSCallbackObject::deleteProperty):
 311         (KJS::JSCallbackObject::getPropertyNames):
 312         (KJS::JSCallbackObject::staticValueGetter):
 313         (KJS::JSCallbackObject::staticFunctionGetter):
 314         * API/JSClassRef.cpp:
 315         (OpaqueJSClass::OpaqueJSClass):
 316         (OpaqueJSClass::~OpaqueJSClass):
 317         * API/JSClassRef.h:
 318         * API/JSObjectRef.cpp:
 319         (JSClassCreate):
 320         (JSObjectMakeFunctionWithCallback):
 321         (JSObjectMakeFunction):
 322         (OpaqueJSPropertyNameArray::OpaqueJSPropertyNameArray):
 323         (JSObjectCopyPropertyNames):
 324         * API/JSObjectRef.h:
 325         * API/minidom.c:
 326         (main):
 327         * API/testapi.c:
 328         (main):
 329         * ChangeLog:
 330         * JavaScriptCore.exp:
 331
 332 2006-07-16  Geoffrey Garen  <ggaren@apple.com>
 333
 334         Laughed at by Beth.
 335
 336         Replace __JS with OpaqueJS because the former, while used by CF, is
 337         a prefix that's triply-reserved by the compiler. (_* is reserved in global
 338         names, _[A-Z] is reserved in all names, and __ is reserved in all names
 339         in C++.)
 340
 341         Opaque is an alternative used by other Mac OS X framewokrs.
 342
 343         * API/APICast.h:
 344         (toJS):
 345         * API/JSBase.h:
 346         * API/JSCallbackObject.cpp:
 347         (KJS::JSCallbackObject::getOwnPropertySlot):
 348         (KJS::JSCallbackObject::put):
 349         (KJS::JSCallbackObject::deleteProperty):
 350         (KJS::JSCallbackObject::getPropertyNames):
 351         (KJS::JSCallbackObject::staticValueGetter):
 352         (KJS::JSCallbackObject::staticFunctionGetter):
 353         * API/JSClassRef.cpp:
 354         (OpaqueJSClass::OpaqueJSClass):
 355         (OpaqueJSClass::~OpaqueJSClass):
 356         * API/JSClassRef.h:
 357         * API/JSObjectRef.cpp:
 358         (JSClassCreate):
 359         (OpaqueJSPropertyNameArray::OpaqueJSPropertyNameArray):
 360         (JSObjectCopyPropertyNames):
 361
 362 2006-07-16  Darin Adler  <darin@apple.com>
 363
 364         - try to fix Windows build
 365
 366         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
 367         Added some recently added files, removed some recently removed.
 368
 369 2006-07-16  Geoffrey Garen  <ggaren@apple.com>
 370
 371         Reviewed by Maciej.
 372
 373         - Change getProperty* to return undefined, rather than NULL, for missing
 374         properties, since that's what the spec says. Also added exception out
 375         parameters to the *Index functions, because they can call through to the
 376         regular functions, which can throw for custom objects.
 377
 378         * API/JSObjectRef.cpp:
 379         (JSObjectGetProperty):
 380         (JSObjectGetPropertyAtIndex):
 381         (JSObjectSetPropertyAtIndex):
 382         * API/JSObjectRef.h:
 383         * API/testapi.c:
 384         (main):
 385
 386 2006-07-16  Geoffrey Garen  <ggaren@apple.com>
 387
 388         Reviewed by Maciej.
 389
 390         - Properly document and handle NULL callbacks for static properties. We
 391         throw an exception in any case other than a ReadOnly property with a NULL
 392         setProperty callback, because a NULL callback almost certainly indicates
 393         a programming error. Also throw an exception if hasProperty returns true
 394         for a property that getProperty can't get.
 395
 396         - If a static setProperty callback returns 'false', to indicate that the
 397         property was not set, we no longer forward the set request up the class
 398         chain, because that's almost certainly not what the programmer expected.
 399
 400         * API/JSCallbackObject.cpp:
 401         (KJS::JSCallbackObject::getOwnPropertySlot):
 402         (KJS::JSCallbackObject::put):
 403         (KJS::JSCallbackObject::staticValueGetter):
 404         (KJS::JSCallbackObject::staticFunctionGetter):
 405         (KJS::JSCallbackObject::callbackGetter):
 406         * API/JSObjectRef.h:
 407         * API/minidom.js:
 408         * API/testapi.c:
 409         (MyObject_hasProperty):
 410         * API/testapi.js:
 411
 412 2006-07-16  Geoffrey Garen  <ggaren@apple.com>
 413
 414         Reviewed by Maciej.
 415
 416         - Added names to functions.
 417
 418         - Removed GetPrivate/SetPrivate from callbackFunctions and callbackConstructors.
 419         The private data idiom is that a JS object stores its native implementation
 420         as private data. For functions and constructors, the native implementation is nothing
 421         more than the callback they already store, so supporting private data, too,
 422         confuses the idiom. If you *really* want, you can still create a custom
 423         function with private data.
 424
 425         * API/JSCallbackConstructor.cpp:
 426         * API/JSCallbackConstructor.h:
 427         * API/JSCallbackFunction.cpp:
 428         (KJS::JSCallbackFunction::JSCallbackFunction):
 429         * API/JSCallbackFunction.h:
 430         * API/JSCallbackObject.cpp:
 431         (KJS::JSCallbackObject::staticFunctionGetter):
 432         * API/JSObjectRef.cpp:
 433         (JSObjectMakeFunction):
 434         (JSObjectMakeFunctionWithBody):
 435         (JSObjectGetPrivate):
 436         (JSObjectSetPrivate):
 437         * API/JSObjectRef.h:
 438         * API/minidom.c:
 439         (main):
 440         * API/testapi.c:
 441         (main):
 442
 443 2006-07-15  Maciej Stachowiak  <mjs@apple.com>
 444
 445         Reviewed by Darin.
 446
 447         - switch property lists to be vector+set of Identifiers instead of list of References
 448
 449         This has the following benefits:
 450
 451         - no duplicates in property lists
 452         - simplifies API calls
 453         - probably more efficient, since linked list is gone
 454         - entirely removed Reference, ReferenceList and ProtectedReference types from the API
 455
 456         * kjs/PropertyNameArray.cpp: Added.
 457         (KJS::PropertyNameArray::add): Check set, if not already there, add to
 458         vector.
 459         * kjs/PropertyNameArray.h: Added.
 460         (KJS::PropertyNameArray::PropertyNameArray): Newly added type, combines
 461         a set and a vector to make a unique but ordered list of identifiers.
 462         (KJS::PropertyNameArray::begin): ditto
 463         (KJS::PropertyNameArray::end): ditto
 464         (KJS::PropertyNameArray::size): ditto
 465         (KJS::PropertyNameArray::operator[]): ditto
 466         * kjs/array_instance.h:
 467         * kjs/array_object.cpp:
 468         (ArrayInstance::getPropertyNames): renamed from getPropertyList, updated
 469         for PropertyNameArray
 470         (ArrayInstance::setLength): updated for PropertyNameArray
 471         (ArrayInstance::pushUndefinedObjectsToEnd): ditto
 472         * kjs/nodes.cpp:
 473         (ForInNode::execute): updated for PropertyNameArray
 474         * kjs/nodes.h:
 475         * kjs/object.cpp:
 476         (KJS::JSObject::getPropertyNames): renamed from getPropertyList, updated
 477         for PropertyNameArray
 478         * kjs/object.h:
 479         * kjs/property_map.cpp:
 480         (KJS::PropertyMap::getEnumerablePropertyNames): updated for PropertyNameArray
 481         (KJS::PropertyMap::getSparseArrayPropertyNames): ditto
 482         * kjs/property_map.h:
 483         * kjs/protected_reference.h: Removed.
 484         * kjs/reference.cpp: Removed.
 485         * kjs/reference.h: Removed.
 486         * kjs/reference_list.cpp: Removed.
 487         * kjs/reference_list.h: Removed.
 488         * kjs/scope_chain.cpp:
 489         (KJS::ScopeChain::print): Use PropertyNamesArray instead of ReferenceList.
 490         * kjs/string_object.cpp:
 491         (StringInstance::getPropertyNames): Updated for new approach.
 492         * kjs/string_object.h:
 493         * kjs/ustring.h:
 494         * API/APICast.h:
 495         (toJS): Added overload for PropertyNameAccumulatorRef / PropertyNameArray*
 496         (toRef): ditto
 497         * API/JSBase.h:
 498         * API/JSCallbackObject.cpp:
 499         (KJS::JSCallbackObject::getPropertyNames): Fixed for new API.
 500         * API/JSCallbackObject.h:
 501         * API/JSObjectRef.cpp:
 502         (__JSPropertyNameArray::__JSPropertyNameArray): Type used for a publicly vended
 503         JSPropertyNameArrayRef.
 504         (JSObjectCopyPropertyNames): New API call - renamed / refactored from
 505         JSObjectCreatePropertyList
 506         (JSPropertyNameArrayRetain): new retain call for JSPropertyNameArray.
 507         (JSPropertyNameArrayRelease): new release call for - " -.
 508         (JSPropertyNameArrayGetCount): Instead of having to use a stateful enumerator you
 509         can now get the count and items in any order.
 510         (JSPropertyNameArrayGetNameAtIndex): See above.
 511         (JSPropertyNameAccumulatorAddName): What you add properties to is now an opaque
 512         accumulator object.
 513         * API/JSObjectRef.h: Prototyped new functions, removed old ones
 514         * JavaScriptCore.exp: Updated exported symbols.
 515         * JavaScriptCore.xcodeproj/project.pbxproj: Added new files, removed old.
 516         * API/testapi.c:
 517         (MyObject_getPropertyNames): Renamed / fixed callback to fit new paradigm.
 518         (main): Updated for new API.
 519
 520 2006-07-15  Darin Adler  <darin@apple.com>
 521
 522         - oops, missed a few more arrays that had to be const
 523
 524         * API/JSNode.c:
 525         (JSNodePrototype_appendChild): Added const.
 526         (JSNodePrototype_removeChild): Ditto.
 527         (JSNodePrototype_replaceChild): Ditto.
 528         (JSNode_construct): Ditto.
 529         * API/JSNodeList.c:
 530         (JSNodeListPrototype_item): Ditto.
 531         * API/JSObjectRef.cpp:
 532         (JSObjectMakeFunctionWithBody): Ditto.
 533         (JSObjectCallAsFunction): Ditto.
 534         (JSObjectCallAsConstructor): Ditto.
 535         * API/minidom.c:
 536         (print): Ditto.
 537         * API/testapi.c:
 538         (MyObject_callAsFunction): Ditto.
 539         (MyObject_callAsConstructor): Ditto.
 540         (print_callAsFunction): Ditto.
 541         (myConstructor_callAsConstructor): Ditto.
 542
 543 2006-07-15  Darin Adler  <darin@apple.com>
 544
 545         Reviewed by Maciej.
 546
 547         * API/JSNode.h: Made an array parameter const.
 548         * API/JSObjectRef.h: Made array parameters const. Fixed a comment.
 549
 550 2006-07-15  Geoffrey Garen  <ggaren@apple.com>
 551
 552         Reviewed by Maciej.
 553
 554         - JSObjectMakeFunctionWithBody includes a function name and named parameters now.
 555
 556         * API/JSObjectRef.cpp:
 557         (JSObjectMakeFunctionWithBody):
 558         * API/JSObjectRef.h:
 559         * API/testapi.c:
 560         (assertEqualsAsUTF8String): More informative failure reporting.
 561         (main): Test more function cases.
 562
 563 2006-07-15  Geoffrey Garen  <ggaren@apple.com>
 564
 565         Reviewed by Maciej.
 566
 567         - Moved the arguments passed to JSClassCreate into a single structure,
 568         called JSClassDefinition. This will enable easier structure
 569         migration/versioning in the future, if necessary.
 570
 571         - Added support for class names.
 572
 573         - kJSClassDefinitionNull replaces kJSObjectCallbacksNone.
 574
 575         - JSClass is becoming a fairly complex struct, so I migrated all of its
 576         implementation other than reference counting to the sruct.
 577
 578         - Also moved JSClass* functions in the API to JSObjectRef.cpp, since they're
 579         declared in JSObjectRef.h
 580
 581         - Also added some more informative explanation to the class structure doc.
 582
 583 2006-07-15  Darin Adler  <darin@apple.com>
 584
 585         Reviewed by Geoff.
 586
 587         - fix http://bugzilla.opendarwin.org/show_bug.cgi?id=8395
 588           <rdar://problem/4613467>
 589           REGRESSION: RegEx seems broken for hex escaped non breaking space
 590
 591         Test: fast/js/regexp-extended-characters-more.html
 592
 593         * pcre/pcre_exec.c:
 594         (match): Got rid of utf16Length local variable to guarantee there's no
 595         extra stack usage in recursive calls. Fixed two places in the PCRE_UTF16
 596         code that were using the length variable, which is the UTF-8 length of
 597         a character in the pattern, to move in the UTF-16 subject string. Instead
 598         they hardcode lengths of 1 and 2 since the code already handles BMP
 599         characters and surrogate pairs separately. Also fixed some DPRINTF so
 600         I could compile with DEBUG on.
 601         (pcre_exec): Changed a place that was checking for multibyte characters
 602         in the subject string to use ISMIDCHAR. Instead it was using hardcoded
 603         logic that was right for UTF-8 but wrong for UTF-16.
 604
 605         * pcre/pcre_compile.c: (pcre_compile2): Fixed a DPRINTF so I could compile
 606         with DEBUG on.
 607
 608 2006-07-14  Geoffrey Garen  <ggaren@apple.com>
 609
 610         RS by Maciej.
 611
 612         Global replace in the API of argc/argv with argumentCount/arguments.
 613
 614 2006-07-14  Geoffrey Garen  <ggaren@apple.com>
 615
 616         Reviewed by Maciej.
 617
 618         - Finalized exception handling in the API.
 619
 620         setProperty can throw because it throws for built-in arrays. getProperty
 621         and deleteProperty can throw because setProperty can throw and we want
 622         to be consistent, and also because they seem like "actions." callAsFunction,
 623         callAsConstructor, and hasInstance can throw, because they caan throw for
 624         all built-ins.
 625
 626         toBoolean can't throw because it's defined that way in the spec.
 627
 628         - Documented that toBoolean and toObject can't be overridden by custom
 629         objects because they're defined that way in the spec.
 630
 631 === Safari-521.17 ===
 632
 633 2006-07-14  Geoffrey Garen  <ggaren@apple.com>
 634
 635         Reviewed by Maciej.
 636
 637         - Implemented ref-counting of JSContexts by splitting into two datatypes:
 638         JSGlobalContext, which you can create/retain/release, and JSContext, which
 639         you can't.
 640
 641         Internally, you retain a JSGlobalContext/ExecState by retaining its
 642         interpreter, which, in the case of a global ExecState, owns it.
 643
 644         - Also made ~Interpreter() protected to catch places where Interpreter
 645         is manually deleted. (Can't make it private because some crazy fool
 646         decided it would be a good idea to subclass Interpreter in other frameworks.
 647         I pity da fool.)
 648
 649         * API/APICast.h:
 650         (toJS): Added cast for new JSGlobalContext
 651         * API/JSStringRef.h: Changed vague "you must" language to more specific
 652         (but, ultimately, equally vague) "behavior is undefined if you don't"
 653         language.
 654         (KJS::Interpreter::Interpreter): Factored more common initialization into
 655         init()
 656         * kjs/interpreter.h:
 657         (KJS::Interpreter::ref): new
 658         (KJS::Interpreter::deref): new
 659         (KJS::Interpreter::refCount): new
 660         * kjs/testkjs.cpp:
 661         (doIt): Ref-count the interpreter.
 662
 663 2006-07-14  Maciej Stachowiak  <mjs@apple.com>
 664
 665         Reviewed by Geoff.
 666
 667         - removed bool return value from JSObjectSetProperty, since it is inefficient and
 668         also doesn't work quite right
 669         - added JSObjectGetPropertyAtIndex and JSObjectSetPropertyAtIndex
 670
 671         * API/JSObjectRef.cpp:
 672         (JSObjectSetProperty): Removed return value and canPut stuff.
 673         (JSObjectGetPropertyAtIndex): Added.
 674         (JSObjectSetPropertyAtIndex): Added.
 675         * API/JSObjectRef.h: Prototyped and documented new functions.
 676
 677 2006-07-14  Geoffrey Garen  <ggaren@apple.com>
 678
 679         Reviewed by Beth.
 680
 681         Moved JSCheckScriptSyntax, JSEvaluateScript, and JSGarbageCollect into
 682         JSBase.h/.cpp. They don't belong in the value-specific or context-specific
 683         files because they're not part of the value or context implementations.
 684
 685         * API/JSBase.h:
 686         * API/JSContextRef.cpp:
 687         (JSContextGetGlobalObject):
 688         * API/JSContextRef.h:
 689         * API/JSValueRef.cpp:
 690         (JSValueUnprotect):
 691         * API/JSValueRef.h:
 692         * JavaScriptCore.xcodeproj/project.pbxproj:
 693
 694 2006-07-13  Timothy Hatcher  <timothy@apple.com>
 695
 696         Reviewed by Maciej.
 697
 698         Moved JavaScriptCore to be a public framework.
 699
 700         * JavaScriptCore.xcodeproj/project.pbxproj:
 701
 702 2006-07-13  Mark Rowe  <opendarwin.org@bdash.net.nz>
 703
 704         Reviewed by Geoffrey.
 705
 706         http://bugzilla.opendarwin.org/show_bug.cgi?id=9742
 707         Bug 9742: REGRESSION: WebKit hangs when loading <http://www.vtbook.com>
 708
 709         * kjs/value.h:
 710         (KJS::JSValue::getUInt32): Only types tagged as numeric can be converted to UInt32.
 711
 712 2006-07-13  Geoffrey Garen  <ggaren@apple.com>
 713
 714         Pleasing to Maciej.
 715
 716         - Renamed JSEvaluate -> JSEvaluateScript, JSCheckSyntax -> JSCheckScriptSyntax
 717         - Added exception out parameters to JSValueTo* and JSValueIsEqual because
 718         they can throw
 719         - Removed JSObjectGetDescription because it's useless and vague, and
 720         JSValueToString/JSValueIsObjectOfClass do a better job, anyway
 721         - Clarified comments about "IsFunction/Constructor" to indicate that they
 722         are true of all functions/constructors, not just those created by JSObjectMake*
 723
 724 2006-07-12  Geoffrey Garen  <ggaren@apple.com>
 725
 726         RS by Beth.
 727
 728         Finished previously approved JSInternalString -> JSString conversion
 729         by renaming the files.
 730
 731         * API/JSCallbackObject.cpp:
 732         * API/JSInternalStringRef.cpp: Removed.
 733         * API/JSInternalStringRef.h: Removed.
 734         * API/JSStringRef.cpp: Added.
 735         * API/JSStringRef.h: Added.
 736         * API/JavaScriptCore.h:
 737         * JavaScriptCore.xcodeproj/project.pbxproj:
 738
 739 2006-07-12  Geoffrey Garen  <ggaren@apple.com>
 740
 741         Reviewed by Maciej.
 742
 743         - Removed context and exception parameters from JSObjectGetPropertyEnumerator,
 744         removing the spurious use of ExecState inside JavaScriptCore that made
 745         us think this was necessary in the first place.
 746
 747         (StringInstance::getPropertyList): Use getString instead of toString because
 748         we know we're dealing with a string -- we put it there in the first place.
 749         While we're at it, store the string's size instead of retrieving it each time
 750         through the loop, to avoid the unnecessary killing of puppies.
 751         * kjs/string_object.h:
 752
 753 2006-07-12  Maciej Stachowiak  <mjs@apple.com>
 754
 755         Reviewed by Geoff.
 756
 757         - add handling of hasInstance callback for API objects
 758
 759         * API/JSCallbackObject.cpp:
 760         (KJS::JSCallbackObject::implementsHasInstance): Check if callback is present.
 761         (KJS::JSCallbackObject::hasInstance): Invoke appropriate callback.
 762         * API/JSCallbackObject.h:
 763         * API/JSClassRef.cpp:
 764         * API/JSObjectRef.h:
 765         * API/testapi.c:
 766         (MyObject_hasInstance): Test case; should match what construct would do.
 767         * API/testapi.js:
 768
 769 2006-07-11  Geoffrey Garen  <ggaren@apple.com>
 770
 771         Reviewed by Maciej.
 772
 773         - Implemented a vast number of renames and comment clarifications
 774         suggested during API review.
 775
 776         JSInternalString -> JSString
 777         JS*Make -> JSValueMake*, JSObjectMake*
 778         JSTypeCode -> JSType
 779         JSValueIsInstanceOf -> JSValueIsInstanceOfConstructor (reads strangely well in client code)
 780         JSGC*Protect -> JSValue*Protect
 781         JS*Callback -> JSObject*Callback
 782         JSGetPropertyListCallback -> JSObjectAddPropertiesToListCallback
 783         JSPropertyEnumeratorGetNext -> JSPropertyEnumeratorGetNextName
 784         JSString* ->
 785             JSStringCreateWithUTF8CString, JSStringGetUTF8CString,
 786             JSStringGetMaximumUTF8CStringSize JSStringIsEqualToUTF8CString,
 787             JSStringCreateWithCFString, JSStringCopyCFString, JSStringCreateWithCharacters.
 788
 789         - Changed functions taking a JSValue out arg and returning a bool indicating
 790         whether it was set to simply return a JSValue or NULL.
 791
 792         - Removed JSStringGetCharacters because it's more documentation than code,
 793         and it's just a glorified memcpy built on existing API functionality.
 794
 795         - Moved standard library includes into the headers that actually require them.
 796
 797         - Standardized use of the phrase "Create Rule."
 798
 799         - Removed JSLock from make functions that don't allocate.
 800
 801         - Added exception handling to JSValueToBoolean, since we now allow
 802         callback objects to throw exceptions upon converting to boolean.
 803
 804         - Renamed JSGCCollect to JSGarbageCollect.
 805
 806 2006-07-10  Geoffrey Garen  <ggaren@apple.com>
 807
 808         Reviewed by Darin.
 809
 810         - Changed public header includes to the <JavaScriptCore/ style.
 811         - Changed instances of 'buffer' to 'string' since we decided on
 812         JSInternalString instead of JSStringBuffer.
 813
 814         * API/JSContextRef.h:
 815         * API/JSInternalStringRef.cpp:
 816         (JSStringMake):
 817         (JSInternalStringRetain):
 818         (JSInternalStringRelease):
 819         (JSValueCopyStringValue):
 820         (JSInternalStringGetLength):
 821         (JSInternalStringGetCharactersPtr):
 822         (JSInternalStringGetCharacters):
 823         (JSInternalStringGetMaxLengthUTF8):
 824         (JSInternalStringGetCharactersUTF8):
 825         (CFStringCreateWithJSInternalString):
 826         * API/JSInternalStringRef.h:
 827         * API/JSNode.c:
 828         (JSNodePrototype_appendChild):
 829         (JSNode_getNodeType):
 830         * API/JSObjectRef.cpp:
 831         (JSObjectCallAsConstructor):
 832         * API/JSValueRef.h:
 833         * API/JavaScriptCore.h:
 834         * API/minidom.c:
 835         (main):
 836         (print):
 837         * API/testapi.c:
 838         (MyObject_getPropertyList):
 839         (myConstructor_callAsConstructor):
 840         (main): I noticed that we were prematurely releasing some string buffers,
 841         so I moved their release calls to the end of main(). I got rid of 'Buf' in *Buf
 842         (sometimes changing to 'IString', when necessary to differentiate a variable)
 843         to match the buffer->string change.
 844
 845 === Safari-521.16 ===
 846
 847 2006-07-10  Darin Adler  <darin@apple.com>
 848
 849         * kjs/value.cpp: (KJS::JSValue::toInt32Inline): Added inline keyword one more place.
 850         Just in case.
 851
 852 2006-07-10  Darin Adler  <darin@apple.com>
 853
 854         - fix the release build
 855
 856         * kjs/value.h:
 857         * kjs/value.cpp:
 858         (KJS::JSValue::toInt32Inline): Move the code here to an inline.
 859         (KJS::JSValue::toInt32): Call the inline from both overloaded toInt32 functions.
 860
 861 2006-07-10  David Kilzer  <ddkilzer@kilzer.net>
 862
 863         Reviewed by Darin.
 864
 865         - fix http://bugzilla.opendarwin.org/show_bug.cgi?id=9179
 866           Implement select.options.add() method
 867
 868         * JavaScriptCore.exp: Added overloaded KJS::JSValue::toInt32() method.
 869         * JavaScriptCore.xcodeproj/project.pbxproj: Altered attributes metadata for
 870         kjs/value.h to make it available as a forwarded header.
 871         * kjs/lookup.h:
 872         (KJS::lookupPut): Extracted a lookupPut() method from the existing lookupPut() method.
 873         The new method returns a boolean value if no entry is found in the lookup table.
 874         * kjs/value.cpp:
 875         (KJS::JSValue::toInt32): Overloaded toInt32() method with boolean "Ok" argument.
 876         * kjs/value.h: Ditto.
 877
 878 2006-07-10  Geoffrey Garen  <ggaren@apple.com>
 879
 880         No review necessary. Removed bogus file I accidentally checked in before.
 881
 882         * API/JSInternalSringRef.h: Removed.
 883
 884 2006-07-10  Geoffrey Garen  <ggaren@apple.com>
 885
 886         Reviewed by Darin.
 887
 888         Added exception out parameter to API object callbacks, removed semi-bogus
 889         JSContext(.*)Exception functions.
 890
 891         To make these calls syntactically simple, I added an exceptionSlot()
 892         method to the ExecState class, which provides a JSValue** slot in which to
 893         store a JSValue* exception.
 894
 895         * API/APICast.h:
 896         (toRef):
 897         * API/JSCallbackConstructor.cpp:
 898         (KJS::JSCallbackConstructor::construct):
 899         * API/JSCallbackFunction.cpp:
 900         (KJS::JSCallbackFunction::callAsFunction):
 901         * API/JSCallbackObject.cpp:
 902         (KJS::JSCallbackObject::init):
 903         (KJS::JSCallbackObject::getOwnPropertySlot):
 904         (KJS::JSCallbackObject::put):
 905         (KJS::JSCallbackObject::deleteProperty):
 906         (KJS::JSCallbackObject::construct):
 907         (KJS::JSCallbackObject::callAsFunction):
 908         (KJS::JSCallbackObject::getPropertyList):
 909         (KJS::JSCallbackObject::toBoolean):
 910         (KJS::JSCallbackObject::toNumber):
 911         (KJS::JSCallbackObject::toString):
 912         (KJS::JSCallbackObject::staticValueGetter):
 913         (KJS::JSCallbackObject::callbackGetter):
 914         * API/JSContextRef.cpp:
 915         (JSCheckSyntax):
 916         * API/JSContextRef.h:
 917         * API/JSNode.c:
 918         (JSNodePrototype_appendChild):
 919         (JSNodePrototype_removeChild):
 920         (JSNodePrototype_replaceChild):
 921         (JSNode_getNodeType):
 922         (JSNode_getChildNodes):
 923         (JSNode_getFirstChild):
 924         (JSNode_construct):
 925         * API/JSNode.h:
 926         * API/JSNodeList.c:
 927         (JSNodeListPrototype_item):
 928         (JSNodeList_length):
 929         (JSNodeList_getProperty):
 930         * API/JSObjectRef.h:
 931         * API/minidom.c:
 932         (print):
 933         * API/testapi.c:
 934         (MyObject_initialize):
 935         (MyObject_hasProperty):
 936         (MyObject_getProperty):
 937         (MyObject_setProperty):
 938         (MyObject_deleteProperty):
 939         (MyObject_getPropertyList):
 940         (MyObject_callAsFunction):
 941         (MyObject_callAsConstructor):
 942         (MyObject_convertToType):
 943         (print_callAsFunction):
 944         (myConstructor_callAsConstructor):
 945         (main):
 946         * JavaScriptCore.exp:
 947         * kjs/ExecState.h:
 948         (KJS::ExecState::exceptionHandle):
 949
 950 2006-07-10  Geoffrey Garen  <ggaren@apple.com>
 951
 952         Reviewed by Darin.
 953
 954         Improved type safety by implementing opaque JSValue/JSObject typing through
 955         abuse of 'const', not void*. Also fixed an alarming number of bugs
 956         exposed by this new type safety.
 957
 958         I made one design change in JavaScriptCore, which is that the JSObject
 959         constructor should take a JSValue* as its prototype argument, not a JSObject*,
 960         since we allow the prototype to be any JSValue*, including jsNull(), for
 961         example.
 962
 963         * API/APICast.h:
 964         (toJS):
 965         * API/JSBase.h:
 966         * API/JSCallbackConstructor.cpp:
 967         (KJS::JSCallbackConstructor::construct):
 968         * API/JSCallbackFunction.cpp:
 969         (KJS::JSCallbackFunction::callAsFunction):
 970         * API/JSCallbackObject.cpp:
 971         (KJS::JSCallbackObject::JSCallbackObject):
 972         (KJS::JSCallbackObject::getOwnPropertySlot):
 973         (KJS::JSCallbackObject::put):
 974         (KJS::JSCallbackObject::construct):
 975         (KJS::JSCallbackObject::callAsFunction):
 976         (KJS::JSCallbackObject::staticFunctionGetter):
 977         * API/JSCallbackObject.h:
 978         * API/JSContextRef.cpp:
 979         (JSEvaluate):
 980         * API/JSNode.c:
 981         (JSNodePrototype_appendChild):
 982         (JSNodePrototype_removeChild):
 983         (JSNodePrototype_replaceChild):
 984         * API/JSObjectRef.cpp:
 985         (JSObjectMake):
 986         (JSFunctionMakeWithBody):
 987         (JSObjectGetProperty):
 988         (JSObjectCallAsFunction):
 989         (JSObjectCallAsConstructor):
 990         * API/JSObjectRef.h:
 991         * API/testapi.c:
 992         (main):
 993         * ChangeLog:
 994         * kjs/object.h:
 995         (KJS::JSObject::JSObject):
 996
 997 2006-07-10  Geoffrey Garen  <ggaren@apple.com>
 998
 999         Approved by Maciej, Darin.
1000
1001         Renamed JSStringBufferRef to JSInternalStringRef. "Internal string" means the
1002         JavaScript engine's internal string representation, which is the most
1003         low-level and efficient representation to use when interfacing with JavaScript.
1004
1005         * API/APICast.h:
1006         (toJS):
1007         (toRef):
1008         * API/JSBase.h:
1009         * API/JSCallbackObject.cpp:
1010         (KJS::JSCallbackObject::getOwnPropertySlot):
1011         (KJS::JSCallbackObject::put):
1012         (KJS::JSCallbackObject::deleteProperty):
1013         (KJS::JSCallbackObject::staticValueGetter):
1014         (KJS::JSCallbackObject::callbackGetter):
1015         * API/JSContextRef.cpp:
1016         (JSEvaluate):
1017         (JSCheckSyntax):
1018         * API/JSContextRef.h:
1019         * API/JSInternalStringRef.cpp: Added.
1020         (JSStringMake):
1021         (JSInternalStringCreate):
1022         (JSInternalStringCreateUTF8):
1023         (JSInternalStringRetain):
1024         (JSInternalStringRelease):
1025         (JSValueCopyStringValue):
1026         (JSInternalStringGetLength):
1027         (JSInternalStringGetCharactersPtr):
1028         (JSInternalStringGetCharacters):
1029         (JSInternalStringGetMaxLengthUTF8):
1030         (JSInternalStringGetCharactersUTF8):
1031         (JSInternalStringIsEqual):
1032         (JSInternalStringIsEqualUTF8):
1033         (JSInternalStringCreateCF):
1034         (CFStringCreateWithJSInternalString):
1035         * API/JSInternalStringRef.h: Added.
1036         * API/JSNode.c:
1037         (JSNodePrototype_appendChild):
1038         (JSNode_getNodeType):
1039         (JSNode_getChildNodes):
1040         (JSNode_getFirstChild):
1041         * API/JSNodeList.c:
1042         (JSNodeList_length):
1043         (JSNodeList_getProperty):
1044         * API/JSObjectRef.cpp:
1045         (JSFunctionMakeWithBody):
1046         (JSObjectGetDescription):
1047         (JSObjectHasProperty):
1048         (JSObjectGetProperty):
1049         (JSObjectSetProperty):
1050         (JSObjectDeleteProperty):
1051         (JSPropertyEnumeratorGetNext):
1052         (JSPropertyListAdd):
1053         * API/JSObjectRef.h:
1054         * API/JSStringBufferRef.cpp: Removed.
1055         * API/JSStringBufferRef.h: Removed.
1056         * API/JSValueRef.h:
1057         * API/JavaScriptCore.h:
1058         * API/minidom.c:
1059         (main):
1060         (print):
1061         * API/testapi.c:
1062         (assertEqualsAsUTF8String):
1063         (assertEqualsAsCharactersPtr):
1064         (assertEqualsAsCharacters):
1065         (MyObject_hasProperty):
1066         (MyObject_getProperty):
1067         (MyObject_setProperty):
1068         (MyObject_deleteProperty):
1069         (MyObject_getPropertyList):
1070         (print_callAsFunction):
1071         (myConstructor_callAsConstructor):
1072         (main):
1073         * JavaScriptCore.exp:
1074         * JavaScriptCore.xcodeproj/project.pbxproj:
1075
1076 2006-07-08  Tim Omernick  <timo@apple.com>
1077
1078         Reviewed by Maciej.
1079
1080         Added an OpenGL drawing model to the Netscape Plug-in API.
1081
1082         * bindings/npapi.h:
1083
1084 2006-07-08  Timothy Hatcher  <timothy@apple.com>
1085
1086         Reviewed by Maciej.
1087
1088         Moved KJS_GetCreatedJavaVMs to jni_utility.cpp.
1089         Switched KJS_GetCreatedJavaVMs over to use dlopen and dlsym
1090         now that NSAddImage, NSLookupSymbolInImage and NSAddressOfSymbol
1091         are deprecated in Leopard.
1092
1093         * JavaScriptCore.exp:
1094         * JavaScriptCore.xcodeproj/project.pbxproj:
1095         * bindings/jni/jni_utility.cpp:
1096         (KJS::Bindings::KJS_GetCreatedJavaVMs):
1097         * bindings/softlinking.c: Removed.
1098         * bindings/softlinking.h: Removed.
1099
1100 2006-07-08  Geoffrey Garen  <ggaren@apple.com>
1101
1102         Reviewed by Anders.
1103
1104         - Make JSObjectGetProperty return a JSValue or NULL, like JSEvaluate does.
1105
1106         * API/JSObjectRef.cpp:
1107         (JSObjectGetProperty):
1108         * API/JSObjectRef.h:
1109         * API/testapi.c:
1110         (main):
1111
1112 2006-07-08  Geoffrey Garen  <ggaren@apple.com>
1113
1114         Style change -- no review necessary.
1115
1116         Use 0 instead of NULL in API .cpp files, to match our style guidelines.
1117
1118         * API/JSContextRef.cpp:
1119         (JSEvaluate):
1120         * API/JSObjectRef.cpp:
1121         (JSFunctionMakeWithBody):
1122         (JSObjectCallAsFunction):
1123         (JSObjectCallAsConstructor):
1124         * API/JSValueRef.cpp:
1125         (JSValueToObject):
1126
1127 2006-07-08  Geoffrey Garen  <ggaren@apple.com>
1128
1129         Reviewed by TimO.
1130
1131         - Added ability to pass NULL for thisObject when calling JSObjectCallAsFunction,
1132         to match JSEvaluate.
1133
1134         * API/JSObjectRef.cpp:
1135         (JSObjectCallAsFunction):
1136         * API/JSObjectRef.h:
1137         * API/testapi.c:
1138         (main):
1139
1140 === Safari-521.15 ===
1141
1142 2006-07-07  Geoffrey Garen  <ggaren@apple.com>
1143
1144         Reviewed by Maciej.
1145
1146         - Standardized which functions take a JSContext as an argument. The rule is:
1147         if you might execute JavaScript, you take a JSContext, otherwise you don't.
1148
1149         The FIXME in JSObjectRef.h requires refactoring some parts of Interpreter,
1150         but not API changes, so I'm putting it off until later.
1151
1152         * API/JSCallbackObject.cpp:
1153         (KJS::JSCallbackObject::JSCallbackObject):
1154         (KJS::JSCallbackObject::init):
1155         * API/JSCallbackObject.h:
1156         * API/JSContextRef.cpp:
1157         (JSContextCreate):
1158         * API/JSContextRef.h:
1159         * API/JSObjectRef.cpp:
1160         (JSObjectMake):
1161         (JSPropertyEnumeratorGetNext):
1162         * API/JSObjectRef.h:
1163         * API/testapi.c:
1164         (MyObject_initialize):
1165         (main):
1166         * JavaScriptCore.exp:
1167         * kjs/array_object.cpp:
1168         (ArrayInstance::setLength):
1169         (ArrayInstance::pushUndefinedObjectsToEnd):
1170         * kjs/nodes.cpp:
1171         (ForInNode::execute):
1172         * kjs/reference.cpp:
1173         (KJS::Reference::getPropertyName):
1174         (KJS::Reference::getValue):
1175         * kjs/reference.h:
1176         * kjs/scope_chain.cpp:
1177         (KJS::ScopeChain::print):
1178
1179 2006-07-06  Geoffrey Garen  <ggaren@apple.com>
1180
1181         Reviewed by Maciej.
1182
1183         More API action.
1184
1185         - Headerdoc finished
1186
1187         Semantic Changes:
1188         - Added a JSContextRef argument to many functions, because you need a
1189         JSContextRef for doing virtually anything. I expect to add this argument
1190         to even more functions in a future patch.
1191
1192         - Removed the globalObjectPrototype argument to JSContextCreate because
1193         you can't create an object until you have a context, so it's impossible
1194         to pass a prototype object to JSContextCreate. That's OK because (1) there's
1195         no reason to give the global object a prototype and (2) if you really want
1196         to, you can just use a separate call to JSObjectSetPrototype.
1197
1198         - Removed the JSClassRef argument to JSClassCreate because it was unnecessary,
1199         and you need to be able to make the global object's class before you've
1200         created a JSContext.
1201
1202         - Added an optional exception parameter to JSFunctionMakeWithBody because anything
1203         less would be uncivilized.
1204
1205         - Made the return value parameter to JSObjectGetProperty optional to match
1206         all other return value parameters in the API.
1207
1208         - Made JSObjectSetPrivate/JSObjectGetPrivate work on JSCallbackFunctions
1209         and JSCallbackConstructors. You could use an abstract base class or strategic
1210         placement of m_privateData in the class structure to implement this, but
1211         the former seemed like overkill, and the latter seemed too dangerous.
1212
1213         - Fixed a bug where JSPropertyEnumeratorGetNext would skip the first property.
1214
1215         Cosmetic Changes:
1216         - Reversed the logic of the JSChar #ifdef to avoid confusing headerdoc
1217
1218         - Removed function names from @function declarations because headeroc
1219         can parse them automatically, and I wanted to rule out manual mismatch.
1220
1221         - Changed Error::create to take a const UString& instead of a UString*
1222         because it was looking at me funny.
1223
1224         - Renamed JSStringBufferCreateWithCFString to JSStringBufferCreateCF
1225         because the latter is more concise and it matches JSStringBufferCreateUTF8.
1226
1227         * API/JSCallbackObject.cpp:
1228         (KJS::JSCallbackObject::getOwnPropertySlot):
1229         (KJS::JSCallbackObject::put):
1230         (KJS::JSCallbackObject::deleteProperty):
1231         (KJS::JSCallbackObject::getPropertyList):
1232         (KJS::JSCallbackObject::toBoolean):
1233         (KJS::JSCallbackObject::toNumber):
1234         (KJS::JSCallbackObject::toString):
1235         * API/JSClassRef.cpp:
1236         (JSClassCreate):
1237         * API/JSContextRef.cpp:
1238         (JSContextCreate):
1239         (JSContextSetException):
1240         * API/JSContextRef.h:
1241         * API/JSNode.c:
1242         (JSNodePrototype_class):
1243         (JSNode_class):
1244         * API/JSNodeList.c:
1245         (JSNodeListPrototype_class):
1246         (JSNodeList_class):
1247         * API/JSObjectRef.cpp:
1248         (JSObjectGetProperty):
1249         (JSObjectGetPrivate):
1250         (JSObjectSetPrivate):
1251         (JSObjectCallAsFunction):
1252         (JSObjectCallAsConstructor):
1253         (JSPropertyEnumeratorGetNext):
1254         * API/JSObjectRef.h:
1255         * API/JSStringBufferRef.cpp:
1256         (JSStringBufferCreateCF):
1257         * API/JSStringBufferRef.h:
1258         * API/JSValueRef.cpp:
1259         (JSValueIsInstanceOf):
1260         * API/JSValueRef.h:
1261         * API/minidom.c:
1262         (main):
1263         * API/minidom.js:
1264         * API/testapi.c:
1265         (MyObject_hasProperty):
1266         (MyObject_setProperty):
1267         (MyObject_deleteProperty):
1268         (MyObject_getPropertyList):
1269         (MyObject_convertToType):
1270         (MyObject_class):
1271         (main):
1272         * JavaScriptCore.exp:
1273
1274 2006-07-07  Geoffrey Garen  <ggaren@apple.com>
1275
1276         Reviewed by John.
1277
1278         - Fixed a few crashes resulting from NULL parameters to JSClassCreate.
1279
1280         * API/JSClassRef.cpp:
1281         (JSClassCreate):
1282         (JSClassRelease):
1283         * API/testapi.c: Added test for NULL parameters.
1284         (main):
1285
1286 2006-07-07  Geoffrey Garen  <ggaren@apple.com>
1287
1288         Reviewed by John, mocked by Darin.
1289
1290         - Changed JSEvaluate to take a JSObjectRef instead of a JSValueRef as
1291         "this," since "this" must be an object.
1292
1293         * API/JSContextRef.cpp:
1294         (JSEvaluate):
1295         * API/JSContextRef.h:
1296
1297 2006-07-07  Geoffrey Garen  <ggaren@apple.com>
1298
1299         Reviewed by John.
1300
1301         - More headerdoc
1302
1303         * API/JSBase.h:
1304         * JavaScriptCore.xcodeproj/project.pbxproj:
1305
1306 2006-07-05  Geoffrey Garen  <ggaren@apple.com>
1307
1308         RS by Beth.
1309
1310         Renamed JSCharBufferRef, which was universally unpopular, to JSStringBufferRef,
1311         which, hopefully, will be less unpopular.
1312
1313         * API/APICast.h:
1314         (toJS):
1315         (toRef):
1316         * API/JSBase.h:
1317         * API/JSCallbackObject.cpp:
1318         (KJS::JSCallbackObject::getOwnPropertySlot):
1319         (KJS::JSCallbackObject::put):
1320         (KJS::JSCallbackObject::deleteProperty):
1321         (KJS::JSCallbackObject::staticValueGetter):
1322         (KJS::JSCallbackObject::callbackGetter):
1323         * API/JSCharBufferRef.cpp: Removed.
1324         * API/JSCharBufferRef.h: Removed.
1325         * API/JSContextRef.cpp:
1326         (JSEvaluate):
1327         (JSCheckSyntax):
1328         * API/JSContextRef.h:
1329         * API/JSNode.c:
1330         (JSNodePrototype_appendChild):
1331         (JSNode_getNodeType):
1332         (JSNode_getChildNodes):
1333         (JSNode_getFirstChild):
1334         * API/JSNodeList.c:
1335         (JSNodeList_length):
1336         (JSNodeList_getProperty):
1337         * API/JSObjectRef.cpp:
1338         (JSFunctionMakeWithBody):
1339         (JSObjectGetDescription):
1340         (JSObjectHasProperty):
1341         (JSObjectGetProperty):
1342         (JSObjectSetProperty):
1343         (JSObjectDeleteProperty):
1344         (JSPropertyEnumeratorGetNext):
1345         (JSPropertyListAdd):
1346         * API/JSObjectRef.h:
1347         * API/JSStringBufferRef.cpp: Added.
1348         (JSStringMake):
1349         (JSStringBufferCreate):
1350         (JSStringBufferCreateUTF8):
1351         (JSStringBufferRetain):
1352         (JSStringBufferRelease):
1353         (JSValueCopyStringValue):
1354         (JSStringBufferGetLength):
1355         (JSStringBufferGetCharactersPtr):
1356         (JSStringBufferGetCharacters):
1357         (JSStringBufferGetMaxLengthUTF8):
1358         (JSStringBufferGetCharactersUTF8):
1359         (JSStringBufferIsEqual):
1360         (JSStringBufferIsEqualUTF8):
1361         (JSStringBufferCreateWithCFString):
1362         (CFStringCreateWithJSStringBuffer):
1363         * API/JSStringBufferRef.h: Added.
1364         * API/JSValueRef.h:
1365         * API/JavaScriptCore.h:
1366         * API/minidom.c:
1367         (main):
1368         (print):
1369         * API/testapi.c:
1370         (assertEqualsAsUTF8String):
1371         (assertEqualsAsCharactersPtr):
1372         (assertEqualsAsCharacters):
1373         (MyObject_hasProperty):
1374         (MyObject_getProperty):
1375         (MyObject_setProperty):
1376         (MyObject_deleteProperty):
1377         (MyObject_getPropertyList):
1378         (print_callAsFunction):
1379         (myConstructor_callAsConstructor):
1380         (main):
1381         * JavaScriptCore.exp:
1382         * JavaScriptCore.xcodeproj/project.pbxproj:
1383
1384 2006-07-05  Geoffrey Garen  <ggaren@apple.com>
1385
1386         RS by Beth.
1387
1388         Moved some code around for more logical file separation.
1389
1390         * API/JSBase.h:
1391         * API/JSContextRef.h:
1392         * API/JSObjectRef.cpp:
1393         * API/JSValueRef.cpp:
1394         (JSValueToObject):
1395         * API/JSValueRef.h:
1396
1397 2006-07-03  Geoffrey Garen  <ggaren@apple.com>
1398
1399         Reviewed by Maciej.
1400
1401         Implemented JSFunctionMakeWithBody, which parses a script as a function body
1402         in the global scope, and returns the resulting anonymous function.
1403
1404         I also removed private data from JSCallbackFunction. It never worked,
1405         since JSCallbackFunction doesn't inherit from JSCallbackObject.
1406
1407         * API/JSCallbackConstructor.cpp: Removed.
1408         * API/JSCallbackConstructor.h: Removed.
1409         * API/JSCallbackFunction.cpp:
1410         (KJS::JSCallbackFunction::JSCallbackFunction):
1411         (KJS::JSCallbackFunction::implementsConstruct):
1412         (KJS::JSCallbackFunction::construct):
1413         (KJS::JSCallbackFunction::implementsCall):
1414         (KJS::JSCallbackFunction::callAsFunction):
1415         * API/JSCallbackFunction.h:
1416         * API/JSCallbackObject.cpp:
1417         (KJS::JSCallbackObject::staticFunctionGetter):
1418         * API/JSObjectRef.cpp:
1419         (JSFunctionMake):
1420         (JSFunctionMakeWithCallbacks):
1421         * API/JSObjectRef.h:
1422         * API/JSValueRef.h:
1423         * API/minidom.c:
1424         (main):
1425         * API/testapi.c:
1426         (main):
1427         * JavaScriptCore.exp: Programmatically added all symbols exported by
1428         API object files, and sorted results
1429         * JavaScriptCore.xcodeproj/project.pbxproj:
1430
1431 2006-07-03  Geoffrey Garen  <ggaren@apple.com>
1432
1433         Reviewed by Maciej.
1434
1435         - Return syntax error in JSCheckSyntax through a JSValueRef* exception
1436         argument
1437
1438         * API/JSBase.h:
1439         * API/JSContextRef.cpp:
1440         (JSCheckSyntax):
1441         * API/testapi.c:
1442         (main):
1443         * JavaScriptCore.exp:
1444         * kjs/interpreter.cpp:
1445         (KJS::Interpreter::checkSyntax):
1446         * kjs/interpreter.h:
1447
1448 2006-07-04  Darin Adler  <darin@apple.com>
1449
1450         - fixed build
1451
1452         * wtf/MathExtras.h: Oops. Added missing #endif.
1453
1454 2006-07-04  Bjoern Graf  <bjoern.graf@gmail.com>
1455
1456         Reviewed by Maciej.
1457         Tweaked a bit by Darin.
1458
1459         - http://bugzilla.opendarwin.org/show_bug.cgi?id=9678
1460           work around MSVCRT's fmod function returning NaN for fmod(x, infinity) instead of x
1461
1462         * wtf/MathExtras.h: Added include of <float.h>.
1463         (isinf): Fix to return false for NAN.
1464         (wtf_fmod): Added. An inline that works around the bug.
1465
1466         * kjs/nodes.cpp:
1467         * kjs/number_object.cpp:
1468         * kjs/operations.cpp:
1469         * kjs/value.cpp:
1470         Added includes of MathExtras.h to all files using fmod.
1471
1472         * JavaScriptCore.xcodeproj/project.pbxproj: Let Xcode 2.3 have its way with
1473         the project.
1474
1475 2006-07-01  Geoffrey Garen  <ggaren@apple.com>
1476
1477         Reviewed by Darin.
1478
1479         - Refined value conversions in the API:
1480             - failed toNumber returns NaN
1481             - failed toObject returns NULL
1482             - failed toString returns empty string
1483
1484         - Refined excpetion handling in the API:
1485             - failed value conversions do not throw exceptions
1486             - uncaught exceptions in JSEvaluate, JSObjectCallAsFunction, and
1487               JSObjectCallAsConstructor are returned through a JSValueRef* exception
1488               argument
1489             - removed JSContextHasException, because JSContextGetException does
1490               the same job
1491
1492         * API/JSBase.h:
1493         * API/JSCharBufferRef.cpp:
1494         (JSValueCopyStringValue):
1495         * API/JSContextRef.cpp:
1496         (JSEvaluate):
1497         * API/JSContextRef.h:
1498         * API/JSNodeList.c: Added test code demonstrating how you would use
1499         toNumber, and why you probably don't need toUInt32, etc.
1500         (JSNodeListPrototype_item):
1501         (JSNodeList_getProperty):
1502         * API/JSObjectRef.cpp:
1503         (JSValueToObject):
1504         (JSObjectCallAsFunction):
1505         (JSObjectCallAsConstructor):
1506         * API/JSObjectRef.h:
1507         * API/JSValueRef.cpp:
1508         (JSValueToNumber):
1509         * API/JSValueRef.h:
1510         * API/minidom.c:
1511         (main):
1512         * API/testapi.c:
1513         (main): Added tests for new rules, and call to JSGCProtect to fix Intel
1514         crash
1515         * JavaScriptCore.exp:
1516
1517 2006-07-03  Darin Adler  <darin@apple.com>
1518
1519         - Rolled out HashMap implementation of NPRuntime, at least temporarily.
1520
1521         Fixes hang in the bindings section of layout tests seen on the
1522         buildbot.
1523
1524         This code was using HashMap<const char*, PrivateIdentifier*>.
1525         But that hashes based on pointer identity, not string value.
1526         The default hash for any pointer type is to hash based on the pointer.
1527         And WTF doesn't currently have a string hash for char*.
1528         We'll need to fix that before re-landing this patch.
1529
1530         (Formatting was also incorrect -- extra spaces in parentheses.)
1531
1532         * bindings/npruntime.cpp: Rolled out last change.
1533
1534 2006-07-02  Justin Haygood  <jhaygood@spsu.edu>
1535
1536         Reviewed, tweaked, landed by ggaren.
1537
1538         - Port NPRuntime from CFDictionary to HashMap.
1539
1540         * bindings/npruntime.cpp:
1541         (getStringIdentifierDictionary):
1542         (getIntIdentifierDictionary):
1543         (_NPN_GetStringIdentifier):
1544         (_NPN_GetIntIdentifier):
1545         * bindings/npruntime.h:
1546
1547 2006-07-01  Geoffrey Garen  <ggaren@apple.com>
1548
1549         Reviewed by Adele.
1550
1551         - Fixed <rdar://problem/4611197> REGRESSION: Liveconnect with Java test
1552         fails at http://www-sor.inria.fr/~dedieu/notes/liveconnect/simple_example.html
1553
1554         * JavaScriptCore.exp: Export symbols used by liveconnect
1555
1556 2006-06-29  Geoffrey Garen  <ggaren@apple.com>
1557
1558         Reviewed by Maciej.
1559
1560         - Phase 2 in the JS API.
1561
1562         - Added support for specifying static tables of values -- this should
1563         obviate the need for using complicated callbacks for most lookups.
1564
1565         -  API objects are now created with classes (JSClassRef) -- in order to support
1566         static values, and in order to prevent API objects from storing their
1567         data inline, and thus falling into the oversized (read: slow and prone to
1568         giving Maciej the frowny face) heap.
1569
1570         - Added two specialized JSObject subclasses -- JSCallbackFunction and JSCallbackConstructor --
1571         to allow JSFunctionMake and JSConstructorMake to continue to work with
1572         the new class model. Another solution to this problem would be to create
1573         a custom class object for each function and constructor you make. This
1574         solution is more code but also more efficient.
1575
1576         - Substantially beefed up the minidom example to demonstrate and test a
1577         lot of these techniques. Its output is still pretty haphazard, though.
1578
1579         - Gave the <kjs/ preface to some includes -- I'm told this matters to
1580         building on some versions of Linux.
1581
1582         - Implemented JSValueIsInstanceOf and JSValueIsObjectOfClass
1583
1584         - Removed GetDescription callback. Something in the class datastructure
1585         should take care of this.
1586
1587         * API/JSBase.h:
1588         * API/JSCallbackConstructor.cpp: Added.
1589         (KJS::):
1590         (KJS::JSCallbackConstructor::JSCallbackConstructor):
1591         (KJS::JSCallbackConstructor::implementsConstruct):
1592         (KJS::JSCallbackConstructor::construct):
1593         (KJS::JSCallbackConstructor::setPrivate):
1594         (KJS::JSCallbackConstructor::getPrivate):
1595         * API/JSCallbackConstructor.h: Added.
1596         (KJS::JSCallbackConstructor::classInfo):
1597         * API/JSCallbackFunction.cpp: Added.
1598         (KJS::):
1599         (KJS::JSCallbackFunction::JSCallbackFunction):
1600         (KJS::JSCallbackFunction::implementsCall):
1601         (KJS::JSCallbackFunction::callAsFunction):
1602         (KJS::JSCallbackFunction::setPrivate):
1603         (KJS::JSCallbackFunction::getPrivate):
1604         * API/JSCallbackFunction.h: Added.
1605         (KJS::JSCallbackFunction::classInfo):
1606         * API/JSCallbackObject.cpp:
1607         (KJS::):
1608         (KJS::JSCallbackObject::JSCallbackObject):
1609         (KJS::JSCallbackObject::init):
1610         (KJS::JSCallbackObject::~JSCallbackObject):
1611         (KJS::JSCallbackObject::className):
1612         (KJS::JSCallbackObject::getOwnPropertySlot):
1613         (KJS::JSCallbackObject::put):
1614         (KJS::JSCallbackObject::deleteProperty):
1615         (KJS::JSCallbackObject::implementsConstruct):
1616         (KJS::JSCallbackObject::construct):
1617         (KJS::JSCallbackObject::implementsCall):
1618         (KJS::JSCallbackObject::callAsFunction):
1619         (KJS::JSCallbackObject::getPropertyList):
1620         (KJS::JSCallbackObject::toBoolean):
1621         (KJS::JSCallbackObject::toNumber):
1622         (KJS::JSCallbackObject::toString):
1623         (KJS::JSCallbackObject::inherits):
1624         (KJS::JSCallbackObject::staticValueGetter):
1625         (KJS::JSCallbackObject::staticFunctionGetter):
1626         (KJS::JSCallbackObject::callbackGetter):
1627         * API/JSCallbackObject.h:
1628         * API/JSCharBufferRef.cpp:
1629         * API/JSClassRef.cpp: Added.
1630         (JSClassCreate):
1631         (JSClassRetain):
1632         (JSClassRelease):
1633         * API/JSClassRef.h: Added.
1634         (StaticValueEntry::StaticValueEntry):
1635         (StaticFunctionEntry::StaticFunctionEntry):
1636         (__JSClass::__JSClass):
1637         * API/JSContextRef.cpp:
1638         (JSContextCreate):
1639         (JSEvaluate):
1640         * API/JSContextRef.h:
1641         * API/JSNode.c: Added.
1642         (JSNodePrototype_appendChild):
1643         (JSNodePrototype_removeChild):
1644         (JSNodePrototype_replaceChild):
1645         (JSNodePrototype_class):
1646         (JSNode_getNodeType):
1647         (JSNode_getChildNodes):
1648         (JSNode_getFirstChild):
1649         (JSNode_finalize):
1650         (JSNode_class):
1651         (JSNode_prototype):
1652         (JSNode_new):
1653         (JSNode_construct):
1654         * API/JSNode.h: Added.
1655         * API/JSNodeList.c: Added.
1656         (JSNodeListPrototype_item):
1657         (JSNodeListPrototype_class):
1658         (JSNodeList_length):
1659         (JSNodeList_getProperty):
1660         (JSNodeList_finalize):
1661         (JSNodeList_class):
1662         (JSNodeList_prototype):
1663         (JSNodeList_new):
1664         * API/JSNodeList.h: Added.
1665         * API/JSObjectRef.cpp:
1666         (JSObjectMake):
1667         (JSFunctionMake):
1668         (JSConstructorMake):
1669         (__JSPropertyEnumerator::__JSPropertyEnumerator):
1670         (JSObjectCreatePropertyEnumerator):
1671         (JSPropertyEnumeratorGetNext):
1672         (JSPropertyEnumeratorRetain):
1673         (JSPropertyEnumeratorRelease):
1674         * API/JSObjectRef.h:
1675         (__JSObjectCallbacks::):
1676         * API/JSValueRef.cpp:
1677         (JSValueIsObjectOfClass):
1678         (JSValueIsInstanceOf):
1679         * API/JSValueRef.h:
1680         * API/Node.c: Added.
1681         (Node_new):
1682         (Node_appendChild):
1683         (Node_removeChild):
1684         (Node_replaceChild):
1685         (Node_ref):
1686         (Node_deref):
1687         * API/Node.h: Added.
1688         * API/NodeList.c: Added.
1689         (NodeList_new):
1690         (NodeList_length):
1691         (NodeList_item):
1692         (NodeList_ref):
1693         (NodeList_deref):
1694         * API/NodeList.h: Added.
1695         * API/minidom.c:
1696         (main):
1697         (print):
1698         (createStringWithContentsOfFile):
1699         * API/minidom.js:
1700         * API/testapi.c:
1701         (assertEqualsAsCharacters):
1702         (MyObject_getProperty):
1703         (MyObject_class):
1704         (myConstructor_callAsConstructor):
1705         (main):
1706         * API/testapi.js:
1707         * JavaScriptCore.xcodeproj/project.pbxproj:
1708
1709 2006-06-26  Kevin Ollivier  <kevino@theolliviers.com>
1710
1711         Reviewed and tweaked by Darin.
1712
1713         - Compile fixes for wx port / gcc 4.0.2
1714
1715         * kjs/array_object.cpp:
1716         Added missing headers.
1717
1718         * kjs/ExecState.h:
1719         gcc needs class prototypes before defining those classes as friend classes
1720
1721 2006-06-30  Mike Emmel  <mike.emmel@gmail.com>
1722
1723         Reviewed by Darin.
1724
1725         Compilation fixes for Linux/Gdk.
1726
1727         * JavaScriptCore/kjs/interpreter.cpp: added include of signal.h
1728         * JavaScriptCore/kjs/ExecState.h: added missing class declaration
1729         * JavaScriptCore/kjs/ExecState.cpp: case wrong on include of context.h
1730         * JavaScriptCore/JavaScriptCoreSources.bkl: added Context.cpp and ExecState.cpp
1731
1732 === Safari-521.14 ===
1733
1734 2006-06-29  Maciej Stachowiak  <mjs@apple.com>
1735
1736         Reviewed by Geoff.
1737
1738         - add headerdoc comments to some of the new JS API headers
1739
1740         * API/JSBase.h:
1741         * API/JSValueRef.h:
1742
1743 2006-06-28  Timothy Hatcher  <timothy@apple.com>
1744
1745         Prefer the Stabs debugging symbols format until DWARF bugs are fixed.
1746
1747         * JavaScriptCore.xcodeproj/project.pbxproj:
1748
1749 2006-06-27  Timothy Hatcher  <timothy@apple.com>
1750
1751         Reviewed by Tim O.
1752
1753         <rdar://problem/4448350> Deprecated ObjC language API used in JavaScriptCore, WebCore, WebKit and WebBrowser
1754
1755         Switch to the new ObjC 2 API, ifdefed the old code around OBJC_API_VERSION so it still works on Tiger.
1756         Removed the use of the old stringWithCString, switched to the new Tiger version that accepts an encoding.
1757         Lots of code style cleanup.
1758
1759         * JavaScriptCore.xcodeproj/project.pbxproj:
1760         * bindings/objc/objc_class.h:
1761         * bindings/objc/objc_class.mm:
1762         (KJS::Bindings::ObjcClass::~ObjcClass):
1763         (KJS::Bindings::_createClassesByIsAIfNecessary):
1764         (KJS::Bindings::ObjcClass::classForIsA):
1765         (KJS::Bindings::ObjcClass::name):
1766         (KJS::Bindings::ObjcClass::methodsNamed):
1767         (KJS::Bindings::ObjcClass::fieldNamed):
1768         (KJS::Bindings::ObjcClass::fallbackObject):
1769         * bindings/objc/objc_header.h:
1770         * bindings/objc/objc_instance.h:
1771         * bindings/objc/objc_instance.mm:
1772         (ObjcInstance::ObjcInstance):
1773         (ObjcInstance::~ObjcInstance):
1774         (ObjcInstance::operator=):
1775         (ObjcInstance::begin):
1776         (ObjcInstance::end):
1777         (ObjcInstance::getClass):
1778         (ObjcInstance::invokeMethod):
1779         (ObjcInstance::invokeDefaultMethod):
1780         (ObjcInstance::setValueOfField):
1781         (ObjcInstance::supportsSetValueOfUndefinedField):
1782         (ObjcInstance::setValueOfUndefinedField):
1783         (ObjcInstance::getValueOfField):
1784         (ObjcInstance::getValueOfUndefinedField):
1785         (ObjcInstance::defaultValue):
1786         (ObjcInstance::stringValue):
1787         (ObjcInstance::numberValue):
1788         (ObjcInstance::booleanValue):
1789         (ObjcInstance::valueOf):
1790         * bindings/objc/objc_runtime.h:
1791         * bindings/objc/objc_runtime.mm:
1792         (ObjcMethod::ObjcMethod):
1793         (ObjcMethod::name):
1794         (ObjcMethod::getMethodSignature):
1795         (ObjcMethod::setJavaScriptName):
1796         (ObjcField::name):
1797         (ObjcField::type):
1798         (ObjcField::valueFromInstance):
1799         (convertValueToObjcObject):
1800         (ObjcField::setValueToInstance):
1801         (ObjcArray::operator=):
1802         (ObjcArray::setValueAt):
1803         (ObjcArray::valueAt):
1804         (ObjcFallbackObjectImp::ObjcFallbackObjectImp):
1805         (ObjcFallbackObjectImp::callAsFunction):
1806         (ObjcFallbackObjectImp::defaultValue):
1807
1808 2006-06-28  Anders Carlsson  <acarlsson@apple.com>
1809
1810         Reviewed by Geoff.
1811
1812         http://bugzilla.opendarwin.org/show_bug.cgi?id=8636
1813         REGRESSION: JavaScript access to Java applet causes hang (_webViewURL not implemented)
1814
1815         * bindings/jni/jni_objc.mm:
1816         (KJS::Bindings::dispatchJNICall):
1817         Just pass nil as the calling URL. This will cause the Java plugin to use the URL of the page
1818         containing the applet (which is what we used to do).
1819
1820 2006-06-27  Timothy Hatcher  <timothy@apple.com>
1821
1822         Reviewed by Darin.
1823
1824         <rdar://problem/4406785> Add an export file to TOT JavaScriptCore like the Safari-2-0-branch
1825
1826         * JavaScriptCore.exp: Added.
1827         * JavaScriptCore.xcodeproj/project.pbxproj:
1828
1829 2006-06-25  Geoffrey Garen  <ggaren@apple.com>
1830
1831         Reviewed by Adele.
1832
1833         - Added JSConstructorMake to match JSFunctionMake, along with test code.
1834
1835         [ I checked in the ChangeLog before without the actual files. ]
1836
1837         * API/JSObjectRef.cpp:
1838         (JSConstructorMake):
1839         * API/JSObjectRef.h:
1840         * API/testapi.c:
1841         (myConstructor_callAsConstructor):
1842         (main):
1843         * API/testapi.js:
1844         * ChangeLog:
1845         * JavaScriptCore.xcodeproj/project.pbxproj: Moved testapi.c to the testapi
1846         target -- this was an oversight in my earlier check-in.
1847
1848 2006-06-25  Timothy Hatcher  <timothy@apple.com>
1849
1850         Reviewed by Darin.
1851
1852         Bug 9574: Drosera should show inline scripts within the original HTML
1853         http://bugzilla.opendarwin.org/show_bug.cgi?id=9574
1854
1855         Pass the starting line number and error message to the debugger.
1856
1857         * kjs/debugger.cpp:
1858         (Debugger::sourceParsed):
1859         * kjs/debugger.h:
1860         * kjs/function.cpp:
1861         (KJS::GlobalFuncImp::callAsFunction):
1862         * kjs/function_object.cpp:
1863         (FunctionObjectImp::construct):
1864         * kjs/interpreter.cpp:
1865         (KJS::Interpreter::evaluate):
1866
1867 2006-06-24  Alexey Proskuryakov  <ap@nypop.com>
1868
1869         Rubber-stamped by Eric.
1870
1871         Add a -h (do not follow symlinks) option to ln in derived sources build script (without it,
1872         a symlink was created inside the source directory on second build).
1873
1874         * JavaScriptCore.xcodeproj/project.pbxproj:
1875
1876 2006-06-24  David Kilzer  <ddkilzer@kilzer.net>
1877
1878         Reviewed by Timothy.
1879
1880         * Info.plist: Fixed copyright to include 2003-2006.
1881
1882 2006-06-24  Alexey Proskuryakov  <ap@nypop.com>
1883
1884         Reviewed by Darin.
1885
1886         - http://bugzilla.opendarwin.org/show_bug.cgi?id=9418
1887         WebKit will not build when Space exists in path
1888
1889         * JavaScriptCore.xcodeproj/project.pbxproj: Enclose search paths in quotes; create symlinks to
1890         avoid passing paths with spaces to make.
1891
1892 2006-06-23  Timothy Hatcher  <timothy@apple.com>
1893
1894         Reviewed by Darin.
1895
1896         Adding more operator[] overloads for long and short types.
1897
1898         * wtf/Vector.h:
1899         (WTF::Vector::operator[]):
1900
1901 === JavaScriptCore-521.13 ===
1902
1903 2006-06-22  Alexey Proskuryakov  <ap@nypop.com>
1904
1905         Build fix.
1906
1907         - http://bugzilla.opendarwin.org/show_bug.cgi?id=9539
1908         Another case error preventing build
1909
1910         * API/JSObjectRef.cpp: Changed "identifier.h" to "Identifier.h"
1911
1912 2006-06-22  David Kilzer  <ddkilzer@kilzer.net>
1913
1914         Build fix.
1915
1916         http://bugzilla.opendarwin.org/show_bug.cgi?id=9539
1917         Another case error preventing build
1918
1919         * API/APICast.h: Changed "UString.h" to "ustring.h".
1920
1921 2006-06-21  Geoffrey Garen  <ggaren@apple.com>
1922
1923         Fixed release build, fixed accidental infinite recursion due to
1924         last minute global replace gone awry.
1925
1926         * API/APICast.h:
1927         (toRef):
1928         * API/testapi.c:
1929         (assertEqualsAsBoolean):
1930         (assertEqualsAsNumber):
1931         (assertEqualsAsUTF8String):
1932         (assertEqualsAsCharactersPtr):
1933         * JavaScriptCore.xcodeproj/project.pbxproj:
1934
1935 2006-06-21  Geoffrey Garen  <ggaren@apple.com>
1936
1937         Reviewed by Anders.
1938
1939         - First cut at C API to JavaScript. Includes a unit test, 'testapi.c',
1940         and the outline of a test app, 'minidom.c'.
1941
1942         Includes one change to JSC internals: Rename propList to getPropertyList and have it
1943         take its target property list by reference so that subclasses can
1944         add properties to the list before calling through to their superclasses.
1945
1946         Also, I just ran prepare-ChangeLog in about 10 seconds, and I would like
1947         to give a shout-out to that.
1948
1949         * API/APICast.h: Added.
1950         (toJS):
1951         (toRef):
1952         * API/JSBase.h: Added.
1953         * API/JSCallbackObject.cpp: Added.
1954         (KJS::):
1955         (KJS::JSCallbackObject::JSCallbackObject):
1956         (KJS::JSCallbackObject::~JSCallbackObject):
1957         (KJS::JSCallbackObject::className):
1958         (KJS::JSCallbackObject::getOwnPropertySlot):
1959         (KJS::JSCallbackObject::put):
1960         (KJS::JSCallbackObject::deleteProperty):
1961         (KJS::JSCallbackObject::implementsConstruct):
1962         (KJS::JSCallbackObject::construct):
1963         (KJS::JSCallbackObject::implementsCall):
1964         (KJS::JSCallbackObject::callAsFunction):
1965         (KJS::JSCallbackObject::getPropertyList):
1966         (KJS::JSCallbackObject::toBoolean):
1967         (KJS::JSCallbackObject::toNumber):
1968         (KJS::JSCallbackObject::toString):
1969         (KJS::JSCallbackObject::setPrivate):
1970         (KJS::JSCallbackObject::getPrivate):
1971         (KJS::JSCallbackObject::cachedValueGetter):
1972         (KJS::JSCallbackObject::callbackGetter):
1973         * API/JSCallbackObject.h: Added.
1974         (KJS::JSCallbackObject::classInfo):
1975         * API/JSCharBufferRef.cpp: Added.
1976         (JSStringMake):
1977         (JSCharBufferCreate):
1978         (JSCharBufferCreateUTF8):
1979         (JSCharBufferRetain):
1980         (JSCharBufferRelease):
1981         (JSValueCopyStringValue):
1982         (JSCharBufferGetLength):
1983         (JSCharBufferGetCharactersPtr):
1984         (JSCharBufferGetCharacters):
1985         (JSCharBufferGetMaxLengthUTF8):
1986         (JSCharBufferGetCharactersUTF8):
1987         (JSCharBufferIsEqual):
1988         (JSCharBufferIsEqualUTF8):
1989         (JSCharBufferCreateWithCFString):
1990         (CFStringCreateWithJSCharBuffer):
1991         * API/JSCharBufferRef.h: Added.
1992         * API/JSContextRef.cpp: Added.
1993         (JSContextCreate):
1994         (JSContextDestroy):
1995         (JSContextGetGlobalObject):
1996         (JSEvaluate):
1997         (JSCheckSyntax):
1998         (JSContextHasException):
1999         (JSContextGetException):
2000         (JSContextClearException):
2001         (JSContextSetException):
2002         * API/JSContextRef.h: Added.
2003         * API/JSObjectRef.cpp: Added.
2004         (JSValueToObject):
2005         (JSObjectMake):
2006         (JSFunctionMake):
2007         (JSObjectGetDescription):
2008         (JSObjectGetPrototype):
2009         (JSObjectSetPrototype):
2010         (JSObjectHasProperty):
2011         (JSObjectGetProperty):
2012         (JSObjectSetProperty):
2013         (JSObjectDeleteProperty):
2014         (JSObjectGetPrivate):
2015         (JSObjectSetPrivate):
2016         (JSObjectIsFunction):
2017         (JSObjectCallAsFunction):
2018         (JSObjectIsConstructor):
2019         (JSObjectCallAsConstructor):
2020         (__JSPropertyListEnumerator::__JSPropertyListEnumerator):
2021         (JSObjectCreatePropertyEnumerator):
2022         (JSPropertyEnumeratorGetNext):
2023         (JSPropertyEnumeratorRetain):
2024         (JSPropertyEnumeratorRelease):
2025         (JSPropertyListAdd):
2026         * API/JSObjectRef.h: Added.
2027         * API/JSValueRef.cpp: Added.
2028         (JSValueGetType):
2029         (JSValueIsUndefined):
2030         (JSValueIsNull):
2031         (JSValueIsBoolean):
2032         (JSValueIsNumber):
2033         (JSValueIsString):
2034         (JSValueIsObject):
2035         (JSValueIsEqual):
2036         (JSValueIsStrictEqual):
2037         (JSUndefinedMake):
2038         (JSNullMake):
2039         (JSBooleanMake):
2040         (JSNumberMake):
2041         (JSValueToBoolean):
2042         (JSValueToNumber):
2043         (JSGCProtect):
2044         (JSGCUnprotect):
2045         (JSGCCollect):
2046         * API/JSValueRef.h: Added.
2047         * API/JavaScriptCore.h: Added.
2048         * API/minidom.c: Added.
2049         (main):
2050         * API/minidom.html: Added.
2051         * API/minidom.js: Added.
2052         * API/testapi.c: Added.
2053         (assertEqualsAsBoolean):
2054         (assertEqualsAsNumber):
2055         (assertEqualsAsUTF8String):
2056         (assertEqualsAsCharactersPtr):
2057         (assertEqualsAsCharacters):
2058         (MyObject_initialize):
2059         (MyObject_copyDescription):
2060         (MyObject_hasProperty):
2061         (MyObject_getProperty):
2062         (MyObject_setProperty):
2063         (MyObject_deleteProperty):
2064         (MyObject_getPropertyList):
2065         (MyObject_callAsFunction):
2066         (MyObject_callAsConstructor):
2067         (MyObject_convertToType):
2068         (MyObject_finalize):
2069         (print_callAsFunction):
2070         (main):
2071         (createStringWithContentsOfFile):
2072         * API/testapi.js: Added.
2073         * ChangeLog:
2074         * JavaScriptCore.xcodeproj/project.pbxproj:
2075         * bindings/npruntime_impl.h:
2076         * kjs/array_instance.h:
2077         * kjs/array_object.cpp:
2078         (ArrayInstance::getPropertyList):
2079         * kjs/interpreter.cpp:
2080         (KJS::Interpreter::evaluate):
2081         * kjs/nodes.cpp:
2082         (ForInNode::execute):
2083         * kjs/object.cpp:
2084         (KJS::JSObject::put):
2085         (KJS::JSObject::canPut):
2086         (KJS::JSObject::deleteProperty):
2087         (KJS::JSObject::propertyIsEnumerable):
2088         (KJS::JSObject::getPropertyAttributes):
2089         (KJS::JSObject::getPropertyList):
2090         * kjs/object.h:
2091         * kjs/property_map.cpp:
2092         (KJS::PropertyMap::get):
2093         * kjs/property_map.h:
2094         * kjs/scope_chain.cpp:
2095         (KJS::ScopeChain::print):
2096         * kjs/string_object.cpp:
2097         (StringInstance::getPropertyList):
2098         * kjs/string_object.h:
2099         * kjs/ustring.h:
2100         (KJS::UString::Rep::ref):
2101
2102 2006-06-20  Timothy Hatcher  <timothy@apple.com>
2103
2104         Reviewed by Geoff.
2105
2106         Make sure we clear the exception before returning so
2107         that future calls will not fail because of an earlier
2108         exception state. Assert on entry that the WebScriptObject
2109         is working with an ExecState that dose not have an exception.
2110         Document that evaluateWebScript and callWebScriptMethod return
2111         WebUndefined when an exception is thrown.
2112
2113         * bindings/objc/WebScriptObject.h:
2114         * bindings/objc/WebScriptObject.mm:
2115         (-[WebScriptObject callWebScriptMethod:withArguments:]):
2116         (-[WebScriptObject evaluateWebScript:]):
2117         (-[WebScriptObject setValue:forKey:]):
2118         (-[WebScriptObject valueForKey:]):
2119         (-[WebScriptObject removeWebScriptKey:]):
2120         (-[WebScriptObject webScriptValueAtIndex:]):
2121         (-[WebScriptObject setWebScriptValueAtIndex:value:]):
2122
2123 2006-06-19  Anders Carlsson  <acarlsson@apple.com>
2124
2125         Reviewed by John.
2126
2127         * kjs/interpreter.cpp:
2128         (KJS::TimeoutChecker::pauseTimeoutCheck):
2129         (KJS::TimeoutChecker::resumeTimeoutCheck):
2130         Fix argument order in setitimer calls.
2131
2132 2006-06-18  Anders Carlsson  <acarlsson@apple.com>
2133
2134         Reviewed by Geoff.
2135
2136         * kjs/interpreter.cpp:
2137         (KJS::TimeoutChecker::pauseTimeoutCheck):
2138         Do nothing if the timeout check hasn't been started.
2139
2140         (KJS::TimeoutChecker::resumeTimeoutCheck):
2141         Do nothing if the timeout check hasn't been started.
2142         Use the right signal handler when unblocking.
2143
2144         (KJS::Interpreter::handleTimeout):
2145         pause/resume the timeout check around the call to
2146         shouldInterruptScript().
2147
2148 2006-06-16  Ben Goodger <beng@google.com>
2149
2150         Reviewed by Maciej
2151
2152         http://bugzilla.opendarwin.org/show_bug.cgi?id=9491
2153         Windows build breaks in interpreter.cpp
2154
2155         * kjs/interpreter.cpp
2156         (KJS::TimeoutChecker::pauseTimeoutCheck):
2157         (KJS::TimeoutChecker::resumeTimeoutCheck):
2158         Make sure to only assert equality with s_executingInterpreter when it
2159         is being used (i.e. when HAVE(SYS_TIME_H) == true)
2160
2161 2006-06-17  David Kilzer  <ddkilzer@kilzer.net>
2162
2163         Reviewed by darin.
2164
2165         http://bugzilla.opendarwin.org/show_bug.cgi?id=9477
2166         REGRESSION: fast/dom/replaceChild.html crashes on WebKit ToT in debug build
2167
2168         * kjs/function.cpp:
2169         (KJS::FunctionImp::callAsFunction): Refetch the debugger after executing the function
2170         in case the WebFrame it was running in has since been destroyed.
2171
2172 2006-06-17  David Kilzer  <ddkilzer@kilzer.net>
2173
2174         Reviewed by ggaren.
2175
2176         http://bugzilla.opendarwin.org/show_bug.cgi?id=9476
2177         REGRESSION: Reproducible crash after closing window after viewing
2178         css2.1/t0803-c5501-imrgn-t-00-b-ag.html
2179
2180         * kjs/debugger.cpp:
2181         (Debugger::detach): Call setDebugger(0) for all interpreters removed from
2182         the 'attached to a debugger' list.
2183
2184 2006-06-17  Anders Carlsson  <acarlsson@apple.com>
2185
2186         Reviewed by Maciej and Geoff.
2187
2188         http://bugzilla.opendarwin.org/show_bug.cgi?id=7080
2189         Provide some way to stop a JavaScript infinite loop
2190
2191         * kjs/completion.h:
2192         (KJS::):
2193         Add Interrupted completion type.
2194
2195         * kjs/function.cpp:
2196         (KJS::FunctionImp::callAsFunction):
2197         (KJS::GlobalFuncImp::callAsFunction):
2198         Only set the exception on the new ExecState if the current one has had one.
2199
2200         * kjs/interpreter.cpp:
2201         (KJS::TimeoutChecker::startTimeoutCheck):
2202         (KJS::TimeoutChecker::stopTimeoutCheck):
2203         (KJS::TimeoutChecker::alarmHandler):
2204         (KJS::TimeoutChecker::pauseTimeoutCheck):
2205         (KJS::TimeoutChecker::resumeTimeoutCheck):
2206         New TimeoutChecker class which handles setting Interpreter::m_timedOut flag after a given
2207         period of time. This currently only works on Unix platforms where setitimer and signals are used.
2208
2209         (KJS::Interpreter::Interpreter):
2210         Initialize new member variables.
2211
2212         (KJS::Interpreter::~Interpreter):
2213         Destroy the timeout checker.
2214
2215         (KJS::Interpreter::startTimeoutCheck):
2216         (KJS::Interpreter::stopTimeoutCheck):
2217         (KJS::Interpreter::pauseTimeoutCheck):
2218         (KJS::Interpreter::resumeTimeoutCheck):
2219         Call the timeout checker.
2220
2221         (KJS::Interpreter::handleTimeout):
2222         Called on timeout. Resets the m_timedOut flag and calls shouldInterruptScript.
2223
2224         * kjs/interpreter.h:
2225         (KJS::Interpreter::setTimeoutTime):
2226         New function for setting the timeout time.
2227
2228         (KJS::Interpreter::shouldInterruptScript):
2229         New function. The idea is that this should be overridden by subclasses in order to for example
2230         pop up a dialog asking the user if the script should be interrupted.
2231
2232         (KJS::Interpreter::checkTimeout):
2233         New function which checks the m_timedOut flag and calls handleTimeout if it's set.
2234
2235         * kjs/nodes.cpp:
2236         (DoWhileNode::execute):
2237         (WhileNode::execute):
2238         (ForNode::execute):
2239         Call Interpreter::checkTimeout after each iteration of the loop.
2240
2241 2006-06-15  Timothy Hatcher  <timothy@apple.com>
2242
2243         Reviewed by Geoff and Darin.
2244
2245         Prefer the DWARF debugging symbols format for use in Xcode 2.3.
2246
2247         * JavaScriptCore.xcodeproj/project.pbxproj:
2248
2249 2006-06-14  Geoffrey Garen  <ggaren@apple.com>
2250
2251         Reviewed by Beth.
2252
2253         - fixed http://bugzilla.opendarwin.org/show_bug.cgi?id=9438
2254         Someone broke ToT: cannot build
2255
2256         * JavaScriptCore.xcodeproj/project.pbxproj:
2257         * bindings/runtime_root.h: Changed "Interpreter.h" to "interpreter.h"
2258
2259 2006-06-12  Geoffrey Garen  <ggaren@apple.com>
2260
2261         build fix
2262
2263         * bindings/objc/WebScriptObject.mm:
2264         (+[WebScriptObject throwException:]): Restore assignment I accidentally
2265         deleted in previous commit
2266
2267 2006-06-12  Geoffrey Garen  <ggaren@apple.com>
2268
2269         Reviewed by TimO, Maciej.
2270
2271         - Merged InterpreterImp code into Interpreter, which implements
2272         all interpreter functionality now. This is part of my continuing quest
2273         to create an external notion of JS "execution context" that is unified and simple --
2274         something to replace the mix of Context, ContextImp, ExecState, Interpreter,
2275         InterpreterImp, and JSRun.
2276
2277         All tests pass. Leaks test has not regressed from its baseline ~207 leaks
2278         with ~3460 leaked nodes.
2279
2280         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
2281         * JavaScriptCore.xcodeproj/project.pbxproj:
2282         * bindings/NP_jsobject.cpp:
2283         * bindings/objc/WebScriptObject.mm:
2284         (+[WebScriptObject throwException:]):
2285         * bindings/runtime_root.cpp:
2286         * bindings/runtime_root.h:
2287         * kjs/Context.cpp:
2288         (KJS::Context::Context):
2289         * kjs/ExecState.cpp: Added.
2290         (KJS::ExecState::lexicalInterpreter):
2291         * kjs/ExecState.h: Added.
2292         (KJS::ExecState::dynamicInterpreter):
2293         * kjs/SavedBuiltins.h: Added.
2294         * kjs/bool_object.cpp:
2295         (BooleanPrototype::BooleanPrototype):
2296         * kjs/collector.cpp:
2297         (KJS::Collector::collect):
2298         (KJS::Collector::numInterpreters):
2299         * kjs/context.h:
2300         * kjs/debugger.cpp:
2301         (Debugger::attach):
2302         (Debugger::detach):
2303         * kjs/function.cpp:
2304         (KJS::FunctionImp::callAsFunction):
2305         (KJS::GlobalFuncImp::callAsFunction):
2306         * kjs/function_object.cpp:
2307         (FunctionObjectImp::construct):
2308         * kjs/internal.cpp:
2309         * kjs/internal.h:
2310         * kjs/interpreter.cpp:
2311         (KJS::interpreterMap):
2312         (KJS::Interpreter::Interpreter):
2313         (KJS::Interpreter::init):
2314         (KJS::Interpreter::~Interpreter):
2315         (KJS::Interpreter::globalObject):
2316         (KJS::Interpreter::initGlobalObject):
2317         (KJS::Interpreter::globalExec):
2318         (KJS::Interpreter::checkSyntax):
2319         (KJS::Interpreter::evaluate):
2320         (KJS::Interpreter::builtinObject):
2321         (KJS::Interpreter::builtinFunction):
2322         (KJS::Interpreter::builtinArray):
2323         (KJS::Interpreter::builtinBoolean):
2324         (KJS::Interpreter::builtinString):
2325         (KJS::Interpreter::builtinNumber):
2326         (KJS::Interpreter::builtinDate):
2327         (KJS::Interpreter::builtinRegExp):
2328         (KJS::Interpreter::builtinError):
2329         (KJS::Interpreter::builtinObjectPrototype):
2330         (KJS::Interpreter::builtinFunctionPrototype):
2331         (KJS::Interpreter::builtinArrayPrototype):
2332         (KJS::Interpreter::builtinBooleanPrototype):
2333         (KJS::Interpreter::builtinStringPrototype):
2334         (KJS::Interpreter::builtinNumberPrototype):
2335         (KJS::Interpreter::builtinDatePrototype):
2336         (KJS::Interpreter::builtinRegExpPrototype):
2337         (KJS::Interpreter::builtinErrorPrototype):
2338         (KJS::Interpreter::builtinEvalError):
2339         (KJS::Interpreter::builtinRangeError):
2340         (KJS::Interpreter::builtinReferenceError):
2341         (KJS::Interpreter::builtinSyntaxError):
2342         (KJS::Interpreter::builtinTypeError):
2343         (KJS::Interpreter::builtinURIError):
2344         (KJS::Interpreter::builtinEvalErrorPrototype):
2345         (KJS::Interpreter::builtinRangeErrorPrototype):
2346         (KJS::Interpreter::builtinReferenceErrorPrototype):
2347         (KJS::Interpreter::builtinSyntaxErrorPrototype):
2348         (KJS::Interpreter::builtinTypeErrorPrototype):
2349         (KJS::Interpreter::builtinURIErrorPrototype):
2350         (KJS::Interpreter::mark):
2351         (KJS::Interpreter::interpreterWithGlobalObject):
2352         (KJS::Interpreter::saveBuiltins):
2353         (KJS::Interpreter::restoreBuiltins):
2354         * kjs/interpreter.h:
2355         (KJS::Interpreter::setCompatMode):
2356         (KJS::Interpreter::compatMode):
2357         (KJS::Interpreter::firstInterpreter):
2358         (KJS::Interpreter::nextInterpreter):
2359         (KJS::Interpreter::prevInterpreter):
2360         (KJS::Interpreter::debugger):
2361         (KJS::Interpreter::setDebugger):
2362         (KJS::Interpreter::setContext):
2363         (KJS::Interpreter::context):
2364         * kjs/nodes.cpp:
2365         (StatementNode::hitStatement):
2366         (RegExpNode::evaluate):
2367         * kjs/protect.h:
2368
2369 2006-06-12  Geoffrey Garen  <ggaren@apple.com>
2370
2371         Reviewed by Maciej.
2372
2373         - Have *.lut.h files #include lookup.h to eliminate surprising header
2374         include order dependency.
2375
2376         * DerivedSources.make:
2377         * kjs/array_object.cpp:
2378         * kjs/date_object.cpp:
2379         * kjs/date_object.h:
2380         (KJS::DateProtoFunc::):
2381         * kjs/lexer.cpp:
2382         * kjs/math_object.cpp:
2383         * kjs/number_object.cpp:
2384         * kjs/regexp_object.cpp:
2385         * kjs/string_object.cpp:
2386
2387 2006-06-10  Geoffrey Garen  <ggaren@apple.com>
2388
2389         - http://bugzilla.opendarwin.org/show_bug.cgi?id=8515
2390         Linux porting compile bug
2391
2392         Fix by Mike Emmel, Reviewed by Darin.
2393
2394         * JavaScriptCoreSources.bkl:
2395         * jscore.bkl:
2396         * wtf/Platform.h:
2397
2398 2006-06-09  Geoffrey Garen  <ggaren@apple.com>
2399
2400         Build fix -- I think :).
2401
2402         * JavaScriptCore.xcodeproj/project.pbxproj:
2403         * kjs/context.h:
2404
2405 2006-06-09  Geoffrey Garen  <ggaren@apple.com>
2406
2407         Reviewed by Eric (yay!).
2408
2409         - Removed Context wrapper for ContextImp, renamed ContextImp to Context,
2410         split Context into its own file -- Context.cpp -- renamed _var to m_var,
2411         change ' *' to '* '.
2412
2413         * JavaScriptCore.xcodeproj/project.pbxproj:
2414         * kjs/Context.cpp: Added.
2415         (KJS::Context::Context):
2416         (KJS::Context::~Context):
2417         (KJS::Context::mark):
2418         * kjs/context.h:
2419         (KJS::Context::scopeChain):
2420         (KJS::Context::variableObject):
2421         (KJS::Context::setVariableObject):
2422         (KJS::Context::thisValue):
2423         (KJS::Context::callingContext):
2424         (KJS::Context::activationObject):
2425         (KJS::Context::currentBody):
2426         (KJS::Context::function):
2427         (KJS::Context::arguments):
2428         (KJS::Context::pushScope):
2429         (KJS::Context::seenLabels):
2430         * kjs/function.cpp:
2431         (KJS::FunctionImp::callAsFunction):
2432         (KJS::FunctionImp::processParameters):
2433         (KJS::FunctionImp::argumentsGetter):
2434         (KJS::GlobalFuncImp::callAsFunction):
2435         * kjs/internal.cpp:
2436         (KJS::InterpreterImp::evaluate):
2437         * kjs/internal.h:
2438         (KJS::InterpreterImp::setContext):
2439         (KJS::InterpreterImp::context):
2440         * kjs/interpreter.cpp:
2441         * kjs/interpreter.h:
2442         (KJS::ExecState::context):
2443         (KJS::ExecState::ExecState):
2444         * kjs/nodes.cpp:
2445         (currentSourceId):
2446         (currentSourceURL):
2447         (ThisNode::evaluate):
2448         (ResolveNode::evaluate):
2449         (FunctionCallResolveNode::evaluate):
2450         (PostfixResolveNode::evaluate):
2451         (DeleteResolveNode::evaluate):
2452         (TypeOfResolveNode::evaluate):
2453         (PrefixResolveNode::evaluate):
2454         (AssignResolveNode::evaluate):
2455         (VarDeclNode::evaluate):
2456         (VarDeclNode::processVarDecls):
2457         (DoWhileNode::execute):
2458         (WhileNode::execute):
2459         (ForNode::execute):
2460         (ForInNode::execute):
2461         (ContinueNode::execute):
2462         (BreakNode::execute):
2463         (ReturnNode::execute):
2464         (WithNode::execute):
2465         (SwitchNode::execute):
2466         (LabelNode::execute):
2467         (TryNode::execute):
2468         (FuncDeclNode::processFuncDecl):
2469         (FuncExprNode::evaluate):
2470
2471 2006-06-07  Geoffrey Garen  <ggaren@apple.com>
2472
2473         Removed API directory I prematurely/accidentally added.
2474
2475         * API: Removed.
2476
2477 2006-06-05  Mitz Pettel  <opendarwin.org@mitzpettel.com>
2478
2479         Reviewed and landed by Geoff.
2480
2481         - fix a regression in ecma_3/String/regress-104375.js
2482
2483         * kjs/string_object.cpp:
2484         (substituteBackreferences): If a 2-digit back reference is out of range,
2485         parse it as a 1-digit reference (followed by the other digit). This matches
2486         Firefox's behavior.
2487
2488 2006-06-05  Geoffrey Garen  <ggaren@apple.com>
2489
2490         Reviewed By Maciej.
2491         Darin already reviewed this change on the branch. See <rdar://problem/4317701>.
2492
2493         - Fixed <rdar://problem/4291345> PCRE overflow in Safari JavaScriptCore
2494
2495         No test case because there's no behavior change.
2496
2497         * pcre/pcre_compile.c:
2498         (read_repeat_counts): Check for integer overflow / out of bounds
2499
2500 2006-06-05  Geoffrey Garen  <ggaren@apple.com>
2501
2502         Reviewed by aliu.
2503
2504         - Changed CString length from int to size_t. We sould probably do this
2505         for UString, too. (Darin, if you're reading this: Maciej said so.)
2506
2507         * kjs/function.cpp:
2508         (KJS::encode):
2509         * kjs/ustring.cpp:
2510         (KJS::CString::CString):
2511         (KJS::operator==):
2512         * kjs/ustring.h:
2513         (KJS::CString::size):
2514
2515 2006-06-04  Geoffrey Garen  <ggaren@apple.com>
2516
2517         Reviewed by Maciej.
2518
2519         - http://bugzilla.opendarwin.org/show_bug.cgi?id=9304
2520         Minor cleanup in JavaScriptCore
2521
2522         * kjs/value.h: Removed redundant declarations
2523
2524 2006-06-04  Darin Adler  <darin@apple.com>
2525
2526         Reviewed by Anders.
2527
2528         - changed deleteAllValues so it can work on "const" collections
2529           Deleting the values affects the values, not the pointers in the
2530           collection, so it's legitimate to do it to a const collection,
2531           and a case of that actually came up in the XPath code.
2532
2533         * wtf/HashMap.h:
2534         (WTF::deleteAllPairSeconds): Use const iterators.
2535         (WTF::deleteAllValues): Take const HashMap reference as a parameter.
2536         * wtf/HashSet.h:
2537         (WTF::deleteAllValues): Take const HashSet reference as a parameter,
2538         and use const iterators.
2539         * wtf/Vector.h:
2540         (WTF::deleteAllValues): Take const Vector reference as a parameter.
2541
2542         - added more functions that are present in <math.h> on some platforms,
2543           but not on others; moved here from various files in WebCore
2544
2545         * wtf/MathExtras.h:
2546         (isinf): Added.
2547         (isnan): Added.
2548         (lround): Added.
2549         (lroundf): Tweaked.
2550         (round): Added.
2551         (roundf): Tweaked.
2552         (signbit): Added.
2553
2554 2006-06-02  Mitz Pettel  <opendarwin.org@mitzpettel.com>
2555
2556         Reviewed by ggaren.
2557
2558         - http://bugzilla.opendarwin.org/show_bug.cgi?id=9234
2559           Implement $&, $' and $` replacement codes in String.prototype.replace
2560
2561         Test: fast/js/string-replace-3.html
2562
2563         * kjs/string_object.cpp:
2564         (substituteBackreferences): Added support for $& (matched substring),
2565         $` (everything preceding matched substring), $' (everything following
2566         matched substring) and 2-digit back references, and cleaned up a little.
2567
2568 2006-06-02 Adele Peterson  <adele@apple.com>
2569
2570         Reviewed by Darin.
2571
2572         Set incremental linking to no. This seems to fix a build problem I was seeing
2573         where dftables couldn't find a dll.
2574
2575         * JavaScriptCore.vcproj/dftables/dftables.vcproj:
2576
2577 2006-05-26  Steve Falkenburg  <sfalken@apple.com>
2578
2579         Build fixes/tweaks
2580
2581         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
2582
2583 === JavaScriptCore-521.11 ===
2584
2585 2006-05-24  Geoffrey Garen  <ggaren@apple.com>
2586
2587         Reviewed by mjs.
2588
2589         - JSC half of fix for <rdar://problem/4557926> TOT REGRESSSION: Crash
2590         occurs when attempting to view image in slideshow mode at
2591         http://d.smugmug.com/gallery/581716 ( KJS::IfNode::execute
2592         (KJS::ExecState*) + 312)
2593
2594         On alternate threads, DOMObjects remain in the
2595         ScriptInterpreter's cache because they're not collected. So, they
2596         need an opportunity to mark their children.
2597
2598         I'm not particularly happy with this solution because it fails to
2599         resolve many outstanding issues with the DOM object cache. Since none
2600         of those issues is a crasher or a serious compatibility concern,
2601         and since the behavior of other browsers is not much to go on in this
2602         case, I've filed <rdar://problem/4561439> about that, and I'm moving on
2603         with my life.
2604
2605         * JavaScriptCore.xcodeproj/project.pbxproj:
2606         * kjs/collector.cpp:
2607         (KJS::Collector::collect):
2608         * kjs/internal.cpp:
2609         (KJS::InterpreterImp::mark):
2610         * kjs/internal.h:
2611         * kjs/interpreter.cpp:
2612         (KJS::Interpreter::mark):
2613         * kjs/interpreter.h:
2614
2615 === JavaScriptCore-521.10 ===
2616
2617 2006-05-22  Timothy Hatcher  <timothy@apple.com>
2618
2619         Reviewed by Eric, Kevin and Geoff.
2620
2621         Merge open source build fixes. <rdar://problem/4555500>
2622
2623         * kjs/collector.cpp: look at the rsp register in x86_64
2624         (KJS::Collector::markOtherThreadConservatively):
2625         * wtf/Platform.h: add x86_64 to the platform list
2626
2627 2006-05-19  Anders Carlsson  <acarlsson@apple.com>
2628
2629         Reviewed by Geoff.
2630
2631         http://bugzilla.opendarwin.org/show_bug.cgi?id=8993
2632         Support function declaration in case statements
2633
2634         * kjs/grammar.y: Get rid of StatementList and use SourceElements instead.
2635
2636         * kjs/nodes.cpp:
2637         (CaseClauseNode::evalStatements):
2638         (CaseClauseNode::processVarDecls):
2639         (CaseClauseNode::processFuncDecl):
2640         (ClauseListNode::processFuncDecl):
2641         (CaseBlockNode::processFuncDecl):
2642         (SwitchNode::processFuncDecl):
2643         * kjs/nodes.h:
2644         (KJS::CaseClauseNode::CaseClauseNode):
2645         (KJS::ClauseListNode::ClauseListNode):
2646         (KJS::ClauseListNode::getClause):
2647         (KJS::ClauseListNode::getNext):
2648         (KJS::ClauseListNode::releaseNext):
2649         (KJS::SwitchNode::SwitchNode):
2650         Add processFuncDecl for the relevant nodes.
2651
2652         * kjs/nodes2string.cpp:
2653         (CaseClauseNode::streamTo):
2654         next got renamed to source.
2655
2656 2006-05-17  George Staikos <staikos@kde.org>
2657
2658         Reviewed by Maciej, Alexey, and Eric.
2659
2660         * pcre/pcre_compile.c:
2661         * pcre/pcre_get.c:
2662         * pcre/pcre_exec.c:
2663         * wtf/UnusedParam.h:
2664         Use /**/ in .c files to compile with non-C99 and non-GCC compilers.
2665
2666         * kjs/testkjs.cpp:
2667         Change include to <wtf/HashTraits.h> from "HashTraits.h" to avoid -I
2668
2669         * wtf/unicode/qt4/UnicodeQt4.h:
2670         Use correct parentheses and correct mask for utf-32 support.
2671
2672 2006-05-17  Alexey Proskuryakov  <ap@nypop.com>
2673
2674         Reviewed by Darin.
2675
2676         - fix http://bugzilla.opendarwin.org/show_bug.cgi?id=8870
2677         Crash typing in Yahoo auto-complete widget.
2678
2679         Test: fast/js/regexp-stack-overflow.html
2680
2681         * pcre/pcre-config.h: Define NO_RECURSE.
2682
2683 2006-05-16  George Staikos <staikos@kde.org>
2684
2685         Reviewed by Maciej.
2686
2687         Fix some warnings and strict compilation errors.
2688
2689         * kjs/nodes.cpp:
2690         * kjs/value.cpp:
2691
2692 2006-05-15  Alexey Proskuryakov  <ap@nypop.com>
2693
2694         * make-generated-sources.sh: Changed to be executable and removed
2695         text in the file generated by "svn diff".
2696
2697 2006-05-15  Geoffrey Garen  <ggaren@apple.com>
2698
2699         Reviewed by Maciej.
2700
2701         - Fixed <rdar://problem/4534904> please do not treat "debugger" as
2702         a reserved word while parsing JavaScript (and other ECMA reserved
2703         words)
2704
2705         AKA
2706
2707         http://bugzilla.opendarwin.org/show_bug.cgi?id=6179
2708         We treat "char" as a reserved word in JavaScript and firefox/IE do
2709         not
2710
2711         (1) I unreserved most of the spec's "future reserved words" because
2712         they're not reserved in IE or FF. (Most, but not all, because IE
2713         somewhat randomly *does* reserve a few of them.)
2714         (2) I made 'debugger' a legitimate statement that acts like an empty
2715         statement because FF and IE support it.
2716
2717         * kjs/grammar.y:
2718         * kjs/keywords.table:
2719
2720 2006-05-15  Tim Omernick  <timo@apple.com>
2721
2722         Reviewed by John Sullivan.
2723
2724         Part of <rdar://problem/4466508> Add 64-bit support to the Netscape Plugin API
2725
2726         Added to the Netscape Plugin API the concept of "plugin drawing models".  The drawing model
2727         determines the kind of graphics context created by the browser for the plugin, as well as
2728         the Mac types of various Netscape Plugin API data structures.
2729
2730         There is a drawing model to represent the old QuickDraw-based API.  It is used by default
2731         if QuickDraw is available on the system, unless the plugin specifies another drawing model.
2732
2733         The big change is the addition of the CoreGraphics drawing model.  A plugin may request this
2734         drawing model to obtain access to a CGContextRef for drawing, instead of a QuickDraw CGrafPtr.
2735
2736         * bindings/npapi.h:
2737         Define NP_NO_QUICKDRAW when compiling 64-bit; there is no 64-bit QuickDraw.
2738         Added NPNVpluginDrawingModel, NPNVsupportsQuickDrawBool, and NPNVsupportsCoreGraphicsBool
2739         variables.
2740         Added NPDrawingModel enumeration.  Currently the only drawing models are QuickDraw and
2741         CoreGraphics.
2742         NPRegion's type now depends on the drawing model specified by the plugin.
2743         NP_Port is now only defined when QuickDraw is available.
2744         Added NP_CGContext, which is the type of the NPWindow's "window" member in CoreGraphics mode.
2745
2746 2006-05-13  Kevin M. Ollivier  <kevino@theolliviers.com>
2747
2748         Reviewed by Darin, landed by ap.
2749
2750         - http://bugzilla.opendarwin.org/show_bug.cgi?id=8528
2751           Bakefiles (and generated Makefiles) for wx and gdk ports
2752
2753         * make-generated-sources.sh:
2754         Added script to configure environment to run DerivedSources.make
2755
2756         * JavaScriptCoreSources.bkl:
2757         Added JavaScriptCore sources list for Bakefile.
2758
2759         * jscore.bkl:
2760         Bakefile used to generate JavaScriptCore project files
2761         (currently only used by wx and gdk ports)
2762
2763 2006-05-09  Steve Falkenburg  <sfalken@apple.com>
2764
2765         Fix Windows build.
2766         Minor fixes to WTF headers.
2767
2768         Reviewed by kevin.
2769
2770         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: Fix include dirs, paths to files.
2771         * JavaScriptCore.vcproj/testkjs/testkjs.vcproj: Fix include dirs.
2772         * wtf/Assertions.h: include Platform.h to get definition for COMPILER()
2773         * wtf/Vector.h: include FastMalloc.h for definition of fastMalloc, fastFree
2774
2775 2006-05-09  Maciej Stachowiak  <mjs@apple.com>
2776
2777         Rubber stamped by Anders.
2778
2779         - renamed kxmlcore to wtf
2780
2781         kxmlcore --> wtf
2782         KXMLCore --> WTF
2783         KXC --> WTF
2784
2785         * JavaScriptCore.xcodeproj/project.pbxproj:
2786         * bindings/c/c_instance.cpp:
2787         * bindings/objc/WebScriptObject.mm:
2788         * kjs/JSImmediate.h:
2789         * kjs/Parser.cpp:
2790         * kjs/Parser.h:
2791         * kjs/array_object.cpp:
2792         * kjs/collector.cpp:
2793         (KJS::Collector::registerThread):
2794         * kjs/collector.h:
2795         * kjs/config.h:
2796         * kjs/function.cpp:
2797         (KJS::isStrWhiteSpace):
2798         * kjs/function.h:
2799         * kjs/identifier.cpp:
2800         * kjs/internal.cpp:
2801         * kjs/internal.h:
2802         * kjs/lexer.cpp:
2803         (Lexer::shift):
2804         (Lexer::isWhiteSpace):
2805         (Lexer::isIdentStart):
2806         (Lexer::isIdentPart):
2807         * kjs/lookup.cpp:
2808         * kjs/nodes.cpp:
2809         * kjs/nodes.h:
2810         * kjs/number_object.cpp:
2811         * kjs/object.h:
2812         * kjs/property_map.cpp:
2813         * kjs/property_map.h:
2814         * kjs/string_object.cpp:
2815         (StringProtoFunc::callAsFunction):
2816         * kjs/testkjs.cpp:
2817         (testIsInteger):
2818         * kjs/ustring.cpp:
2819         * kjs/ustring.h:
2820         * kxmlcore: Removed.
2821         * kxmlcore/AlwaysInline.h: Removed.
2822         * kxmlcore/Assertions.cpp: Removed.
2823         * kxmlcore/Assertions.h: Removed.
2824         * kxmlcore/FastMalloc.cpp: Removed.
2825         * kxmlcore/FastMalloc.h: Removed.
2826         * kxmlcore/FastMallocInternal.h: Removed.
2827         * kxmlcore/Forward.h: Removed.
2828         * kxmlcore/HashCountedSet.h: Removed.
2829         * kxmlcore/HashFunctions.h: Removed.
2830         * kxmlcore/HashMap.h: Removed.
2831         * kxmlcore/HashSet.h: Removed.
2832         * kxmlcore/HashTable.cpp: Removed.
2833         * kxmlcore/HashTable.h: Removed.
2834         * kxmlcore/HashTraits.h: Removed.
2835         * kxmlcore/ListRefPtr.h: Removed.
2836         * kxmlcore/Noncopyable.h: Removed.
2837         * kxmlcore/OwnArrayPtr.h: Removed.
2838         * kxmlcore/OwnPtr.h: Removed.
2839         * kxmlcore/PassRefPtr.h: Removed.
2840         * kxmlcore/Platform.h: Removed.
2841         * kxmlcore/RefPtr.h: Removed.
2842         * kxmlcore/TCPageMap.h: Removed.
2843         * kxmlcore/TCSpinLock.h: Removed.
2844         * kxmlcore/TCSystemAlloc.cpp: Removed.
2845         * kxmlcore/TCSystemAlloc.h: Removed.
2846         * kxmlcore/UnusedParam.h: Removed.
2847         * kxmlcore/Vector.h: Removed.
2848         * kxmlcore/VectorTraits.h: Removed.
2849         * kxmlcore/unicode: Removed.
2850         * kxmlcore/unicode/Unicode.h: Removed.
2851         * kxmlcore/unicode/UnicodeCategory.h: Removed.
2852         * kxmlcore/unicode/icu: Removed.
2853         * kxmlcore/unicode/icu/UnicodeIcu.h: Removed.
2854         * kxmlcore/unicode/posix: Removed.
2855         * kxmlcore/unicode/qt3: Removed.
2856         * kxmlcore/unicode/qt4: Removed.
2857         * kxmlcore/unicode/qt4/UnicodeQt4.h: Removed.
2858         * pcre/pcre_get.c:
2859         * wtf: Added.
2860         * wtf/Assertions.cpp:
2861         * wtf/Assertions.h:
2862         * wtf/FastMalloc.cpp:
2863         (WTF::TCMalloc_ThreadCache::Scavenge):
2864         (WTF::do_malloc):
2865         (WTF::do_free):
2866         (WTF::TCMallocGuard::TCMallocGuard):
2867         (WTF::malloc):
2868         (WTF::free):
2869         (WTF::calloc):
2870         (WTF::cfree):
2871         (WTF::realloc):
2872         * wtf/FastMalloc.h:
2873         * wtf/FastMallocInternal.h:
2874         * wtf/Forward.h:
2875         * wtf/HashCountedSet.h:
2876         * wtf/HashFunctions.h:
2877         * wtf/HashMap.h:
2878         * wtf/HashSet.h:
2879         * wtf/HashTable.cpp:
2880         * wtf/HashTable.h:
2881         * wtf/HashTraits.h:
2882         * wtf/ListRefPtr.h:
2883         * wtf/Noncopyable.h:
2884         * wtf/OwnArrayPtr.h:
2885         * wtf/OwnPtr.h:
2886         * wtf/PassRefPtr.h:
2887         * wtf/RefPtr.h:
2888         * wtf/TCSystemAlloc.cpp:
2889         (TCMalloc_SystemAlloc):
2890         * wtf/Vector.h:
2891         * wtf/VectorTraits.h:
2892         * wtf/unicode/UnicodeCategory.h:
2893         * wtf/unicode/icu/UnicodeIcu.h:
2894
2895 2006-05-08  Timothy Hatcher  <timothy@apple.com>
2896
2897         Reviewed by Tim O.
2898
2899         * bindings/npapi.h: do not define #pragma options align=mac68k if we are 64-bit
2900
2901 2006-05-07  Darin Adler  <darin@apple.com>
2902
2903         Reviewed and landed by Maciej.
2904
2905         - fix http://bugzilla.opendarwin.org/show_bug.cgi?id=8765
2906         Random crashes on TOT since the form state change
2907
2908         I haven't figured out how to construct a test for this, but this does seem to fix the
2909         problem; Mitz mentioned that a double-destroy was occurring in these functions.
2910
2911         * kxmlcore/HashMap.h: (KXMLCore::HashMap::remove): Use RefCounter::deref instead of calling
2912         ~ValueType, because ~ValueType often results in a double-destroy, since the HashTable also
2913         destroys the element based on the storage type. The RefCounter template correctly does work
2914         only in cases where ValueType and ValueStorageType differ and this class is what's used
2915         elsewhere for the same purpose; I somehow missed this case when optimizing HashMap.
2916         * kxmlcore/HashSet.h: (KXMLCore::HashSet::remove): Ditto.
2917
2918 2006-05-05  Darin Adler  <darin@apple.com>
2919
2920         - http://bugzilla.opendarwin.org/show_bug.cgi?id=8722
2921           IE compatibility fix in date parsing
2922
2923         * kjs/date_object.cpp: (KJS::parseDate): Merged change that George Staikos provided
2924         from KDE 3.4.3 branch that allows day values of 0 and values that are > 1000.
2925
2926 2006-05-04  Anders Carlsson  <andersca@mac.com>
2927
2928         Reviewed by Maciej.
2929
2930         http://bugzilla.opendarwin.org/show_bug.cgi?id=8734
2931         Would like a Vector::append that takes another Vector
2932
2933         * kxmlcore/Vector.h:
2934         (KXMLCore::::append):
2935         New function that takes another array.
2936
2937 2006-05-02  Steve Falkenburg  <sfalken@apple.com>
2938
2939         Reviewed by eric.
2940
2941         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: set NDEBUG for release build
2942         * kxmlcore/FastMalloc.cpp: Prevent USE_SYSTEM_MALLOC from being defined twice
2943
2944 2006-05-02  Anders Carlsson  <andersca@mac.com>
2945
2946         Reviewed by Maciej.
2947
2948         * kxmlcore/HashMap.h:
2949         (KXMLCore::::operator):
2950         Return *this
2951
2952 2006-05-01  Tim Omernick  <timo@apple.com>
2953
2954         Reviewed by Tim Hatcher.
2955
2956         <rdar://problem/4476875> Support printing for embedded Netscape plugins
2957
2958         * bindings/npapi.h:
2959         Fixed struct alignment problem in our npapi.h.  Structs must be 68k-aligned on both pre-Mac OS X
2960         and Mac OS X systems, as this is what plugins expect.
2961
2962 2006-05-01  Timothy Hatcher  <timothy@apple.com>
2963
2964         Reviewed by Maciej.
2965
2966         <rdar://problem/4308243> 8F36 Regression: crash in malloc_consolidate if you use a .PAC file
2967
2968         The original fix missed the oversized cell case. Added a test for "currentThreadIsMainThread ||
2969         imp->m_destructorIsThreadSafe" where we collect oversized cells.
2970
2971         We don't have a way to test PAC files yet, so there's no test attached.
2972
2973         * kjs/collector.cpp:
2974         (KJS::Collector::collect): test the thread when we collect oversized cells
2975
2976 2006-05-01  Tim Omernick  <timo@apple.com>
2977
2978         Reviewed by Adele.
2979
2980         <rdar://problem/4526114> REGRESSION (two days ago): LOG() just prints @ for NSObject substitutions
2981
2982         * kxmlcore/Assertions.cpp:
2983         Changed sense of strstr("%@") check.  I already made the same fix to the WebBrowser assertions.
2984
2985 2006-04-28  Steve Falkenburg  <sfalken@apple.com>
2986
2987         Reviewed by kdecker
2988
2989         Actually apply the change that was reviewed insted of checking it in with an #if 0 (oops).
2990
2991         * kjs/testkjs.cpp:
2992         (main): Suppress C runtime alerts
2993
2994 2006-04-28  Steve Falkenburg  <sfalken@apple.com>
2995
2996         Reviewed by kdecker
2997
2998         Suppress error reporting dialog that blocks Javascript tests from completing.
2999
3000         Real error is due to an overflow in the date/time handling functions that needs
3001         to be addressed, but this will prevent the hang running the Javascript tests
3002         on the build bot (along with the related changes).
3003
3004         * kjs/testkjs.cpp:
3005         (main): Suppress C runtime alerts
3006
3007 2006-04-27  Geoffrey Garen  <ggaren@apple.com>
3008
3009         Reviewed by Maciej
3010
3011         - Minor fixups I discovered while working on the autogenerator.
3012
3013         * kjs/lookup.cpp:
3014         (findEntry): ASSERT that size is not 0, because otherwise we'll % by 0,
3015         compute a garbage address, and possibly crash.
3016         * kjs/lookup.h:
3017         (cacheGlobalObject): Don't enumerate cached objects -- ideally, they
3018         would be hidden entirely.
3019
3020 2006-04-21  Kevin M. Ollivier  <kevino@theolliviers.com>
3021
3022         Reviewed by Darin.
3023
3024         - http://bugzilla.opendarwin.org/show_bug.cgi?id=8507
3025           Compilation fixes for building on gcc 4.0.2, and without precomp headers
3026
3027         * kjs/operations.h:
3028         * kxmlcore/Assertions.cpp:
3029         * kxmlcore/FastMalloc.cpp:
3030         Added necessary headers to resolve compilation issues when not using
3031         precompiled headers.
3032
3033         * kjs/value.h: Declare the JSCell class before friend declaration
3034         to resolve compilation issues with gcc 4.0.2.
3035
3036         * kxmlcore/Platform.h: Set Unicode support to use ICU on platforms
3037         other than KDE (previously only defined for Win and Mac OS)
3038
3039 2006-04-18  Eric Seidel  <eseidel@apple.com>
3040
3041         Reviewed by ggaren.
3042
3043         Fix "new Function()" to correctly use lexical scoping.
3044         Add ScopeChain::print() function for debugging.
3045         <rdar://problem/4067864> REGRESSION (125-407): JavaScript failure on PeopleSoft REN Server
3046
3047         * kjs/function_object.cpp:
3048         (FunctionObjectImp::construct):
3049         * kjs/scope_chain.cpp:
3050         (KJS::ScopeChain::print):
3051         * kjs/scope_chain.h:
3052
3053 2006-04-14  James G. Speth  <speth@end.com>
3054
3055         Reviewed by Timothy.
3056
3057         Bug 8389: support for Cocoa bindings - binding an NSTreeController to the WebView's DOM
3058         http://bugzilla.opendarwin.org/show_bug.cgi?id=8389
3059
3060         Adds a category to WebScriptObject with array accessors for KVC/KVO.
3061
3062         If super valueForKey: fails it will call valueForUndefinedKey:, which is
3063         important because it causes the right behavior to happen with bindings using
3064         the "Raises for Not Applicable Keys" flag and the "Not Applicable Placeholder"
3065
3066         * bindings/objc/WebScriptObject.mm:
3067         (-[WebScriptObject valueForKey:]):
3068         (-[WebScriptObject count]):
3069         (-[WebScriptObject objectAtIndex:]):
3070         (-[WebUndefined description]): return "undefined"
3071
3072 2006-04-13  Geoffrey Garen  <ggaren@apple.com>
3073
3074         Reviewed by Darin.
3075
3076         * kjs/internal.cpp:
3077         (KJS::InterpreterImp::initGlobalObject): Add the built-in object
3078         prototype to the end of the global object's prototype chain instead of
3079         just blowing away its existing prototype. We need to do this because
3080         the window object has a meaningful prototype now.
3081
3082 2006-04-13  Maciej Stachowiak  <mjs@apple.com>
3083
3084         Reviewed by Geoff.
3085
3086         - fix testkjs to not show false-positive KJS::Node leaks in debug builds
3087
3088         * kjs/testkjs.cpp:
3089         (doIt):
3090         (kjsmain):
3091
3092 2006-04-11  Geoffrey Garen  <ggaren@apple.com>
3093
3094         Reviewed by Maciej.
3095
3096         Minor code cleanup -- passes all the JS tests.
3097
3098         * kjs/object_object.cpp:
3099         (ObjectObjectImp::construct):
3100         (ObjectObjectImp::callAsFunction):
3101
3102 2006-04-11  Darin Adler  <darin@apple.com>
3103
3104         - another attempt to fix Windows build -- Vector in Forward.h was not working
3105
3106         * kxmlcore/Forward.h: Remove Vector.
3107         * kxmlcore/Vector.h: Add back default arguments, remove include of
3108         Forward.h.
3109
3110 2006-04-11  Darin Adler  <darin@apple.com>
3111
3112         - try to fix Windows build -- HashForward.h was not working
3113
3114         * kxmlcore/HashForward.h: Removed.
3115
3116         * JavaScriptCore.xcodeproj/project.pbxproj: Remove HashForward.h.
3117         * kjs/collector.h: Remove use of HashForward.h.
3118         * kxmlcore/HashCountedSet.h: Remove include of HashForward.h, restore
3119         default arguments.
3120         * kxmlcore/HashMap.h: Ditto.
3121         * kxmlcore/HashSet.h: Ditto.
3122
3123 2006-04-11  David Harrison  <harrison@apple.com>
3124
3125         Reviewed by Darin.
3126
3127         - fixed clean build, broken by Darin's check-in
3128
3129         * kjs/date_object.cpp: Add needed include of lookup.h.
3130         * kjs/regexp_object.cpp: Move include of .lut.h file below other includes.
3131
3132 2006-04-10  Darin Adler  <darin@apple.com>
3133
3134         Rubber-stamped by John Sullivan.
3135
3136         - switched from a shell script to a makefile for generated files
3137         - removed lots of unneeded includes
3138         - added new Forward.h and HashForward.h headers that allow compiling with
3139           fewer unneeded templates
3140
3141         * DerivedSources.make: Added.
3142         * generate-derived-sources: Removed.
3143         * JavaScriptCore.xcodeproj/project.pbxproj: Added new files, changed to use
3144         DerivedSources.make.
3145
3146         * kxmlcore/Forward.h: Added.
3147         * kxmlcore/HashForward.h: Added.
3148
3149         * kxmlcore/HashCountedSet.h: Include HashForward for default args.
3150         * kxmlcore/HashMap.h: Ditto.
3151         * kxmlcore/HashSet.h: Ditto.
3152
3153         * kjs/object.h:
3154         * kjs/object.cpp:
3155         Moved KJS_MAX_STACK into the .cpp file.
3156
3157         * bindings/NP_jsobject.cpp:
3158         * bindings/c/c_instance.h:
3159         * bindings/jni/jni_class.h:
3160         * bindings/jni/jni_runtime.h:
3161         * bindings/jni/jni_utility.h:
3162         * bindings/objc/WebScriptObject.mm:
3163         * bindings/objc/WebScriptObjectPrivate.h:
3164         * bindings/objc/objc_class.h:
3165         * bindings/objc/objc_class.mm:
3166         * bindings/objc/objc_instance.h:
3167         * bindings/objc/objc_instance.mm:
3168         * bindings/objc/objc_runtime.mm:
3169         * bindings/objc/objc_utility.mm:
3170         * bindings/runtime.h:
3171         * bindings/runtime_array.cpp:
3172         * bindings/runtime_array.h:
3173         * bindings/runtime_method.cpp:
3174         * bindings/runtime_method.h:
3175         * bindings/runtime_object.cpp:
3176         * bindings/runtime_root.h:
3177         * kjs/JSImmediate.cpp:
3178         * kjs/Parser.h:
3179         * kjs/array_object.cpp:
3180         * kjs/array_object.h:
3181         * kjs/bool_object.cpp:
3182         * kjs/bool_object.h:
3183         * kjs/collector.h:
3184         * kjs/context.h:
3185         * kjs/debugger.cpp:
3186         * kjs/error_object.h:
3187         * kjs/function_object.h:
3188         * kjs/internal.h:
3189         * kjs/lexer.cpp:
3190         * kjs/math_object.cpp:
3191         * kjs/math_object.h:
3192         * kjs/nodes.cpp:
3193         * kjs/nodes.h:
3194         * kjs/number_object.cpp:
3195         * kjs/number_object.h:
3196         * kjs/object_object.cpp:
3197         * kjs/operations.cpp:
3198         * kjs/protected_reference.h:
3199         * kjs/reference.h:
3200         * kjs/reference_list.h:
3201         * kjs/regexp_object.h:
3202         * kjs/string_object.cpp:
3203         * kjs/string_object.h:
3204         * kjs/testkjs.cpp:
3205         * kjs/value.cpp:
3206         * kjs/value.h:
3207         * kxmlcore/HashTable.h:
3208         * kxmlcore/ListRefPtr.h:
3209         * kxmlcore/TCPageMap.h:
3210         * kxmlcore/Vector.h:
3211         Removed unneeded header includes.
3212
3213 2006-04-09  Geoffrey Garen  <ggaren@apple.com>
3214
3215         Reviewed by eric.
3216
3217         - Fixed http://bugzilla.opendarwin.org/show_bug.cgi?id=8284
3218         prevent unnecessary entries in the "nodes with extra refs" hash table
3219
3220         This patch switches manually RefPtr exchange with use of
3221         RefPtr::release to ensure that a node's ref count never tops 1
3222         (in the normal case).
3223
3224         * kjs/nodes.cpp:
3225         (BlockNode::BlockNode):
3226         (CaseBlockNode::CaseBlockNode):
3227         * kjs/nodes.h:
3228         (KJS::ArrayNode::ArrayNode):
3229         (KJS::ObjectLiteralNode::ObjectLiteralNode):
3230         (KJS::ArgumentsNode::ArgumentsNode):
3231         (KJS::VarStatementNode::VarStatementNode):
3232         (KJS::ForNode::ForNode):
3233         (KJS::CaseClauseNode::CaseClauseNode):
3234         (KJS::FuncExprNode::FuncExprNode):
3235         (KJS::FuncDeclNode::FuncDeclNode):
3236
3237 2006-04-08  Alexey Proskuryakov  <ap@nypop.com>
3238
3239         Reviewed by Darin.
3240
3241         One more attempt - use reinterpret_cast, rather than static_cast.
3242
3243 2006-04-08  Alexey Proskuryakov  <ap@nypop.com>
3244
3245         Reviewed by Darin.
3246
3247         An attempt to fix Win32 build - ICU uses wchar_t on Windows, so we need a type cast.
3248
3249         * kxmlcore/unicode/icu/UnicodeIcu.h:
3250         (KXMLCore::Unicode::toLower):
3251         (KXMLCore::Unicode::toUpper):
3252
3253 2006-04-08  Alexey Proskuryakov  <ap@nypop.com>
3254
3255         Reviewed by Darin.
3256
3257         - fix http://bugzilla.opendarwin.org/show_bug.cgi?id=8264
3258         toLowerCase and toUpperCase don't honor special mappings
3259
3260         Test: fast/js/string-capitalization.html
3261
3262         * JavaScriptCore.xcodeproj/project.pbxproj: Added KXMLCore::Unicode headers to the project.
3263         * icu/unicode/putil.h: Added (copied from WebCore).
3264         * icu/unicode/uiter.h: Ditto.
3265         * icu/unicode/ustring.h: Ditto.
3266         * kjs/string_object.cpp:
3267         (StringProtoFunc::callAsFunction): Use the new KXMLCore::Unicode::toUpper() and toLower().
3268         * kjs/ustring.cpp: Removed unused (and evil) UChar::toLower() and toUpper().
3269         * kjs/ustring.h: Ditto.
3270
3271         * kxmlcore/unicode/Unicode.h: Corrected capitalization of the word Unicode.
3272         * kxmlcore/unicode/UnicodeCategory.h: Renamed include guard macro to match file name.
3273
3274         * kxmlcore/unicode/icu/UnicodeIcu.h:
3275         (KXMLCore::Unicode::toLower): Work on strings, not individual characters. Use ICU root locale.
3276         (KXMLCore::Unicode::toUpper): Ditto.
3277         (KXMLCore::Unicode::isFormatChar): Use int32_t, which can hold a complete code point.
3278         (KXMLCore::Unicode::isSeparatorSpace): Ditto.
3279         (KXMLCore::Unicode::category): Ditto.
3280         * kxmlcore/unicode/qt4/UnicodeQt4.h:
3281         (KXMLCore::Unicode::toLower): Work on strings, not individual characters.
3282         (KXMLCore::Unicode::toUpper): Ditto.
3283         (KXMLCore::Unicode::isFormatChar): Use int32_t, which can hold a complete code point.
3284         (KXMLCore::Unicode::isSeparatorSpace): Ditto.
3285         (KXMLCore::Unicode::category): Ditto.
3286
3287         * tests/mozilla/ecma/String/15.5.4.12-1.js: Corrected expected results.
3288         * tests/mozilla/ecma/String/15.5.4.12-5.js: Corrected expected results.
3289
3290 2006-04-05  Darin Adler  <darin@apple.com>
3291
3292         - attempt to fix Windows build
3293
3294         * kxmlcore/HashMap.h: (KXMLCore::HashMap::remove): Use (*it). instead of it->.
3295         * kxmlcore/HashSet.h: (KXMLCore::HashSet::remove): Ditto.
3296
3297 2006-04-05  Darin Adler  <darin@apple.com>
3298
3299         - attempt to fix Windows build
3300
3301         * os-win32/stdint.h: Add int8_t, uint8_t, int64_t.
3302
3303 2006-04-05  Darin Adler  <darin@apple.com>
3304
3305         Reviewed by Maciej.
3306
3307         - fix memory leak introduced by the previous change
3308
3309         * kxmlcore/HashTable.h: Specialize NeedsRef so that it correctly returns true when
3310         the value in question is a pair where one of the pair needs a ref and the other
3311         of the pair does not.
3312
3313 2006-04-05  Darin Adler  <darin@apple.com>
3314
3315         Reviewed by Maciej.
3316
3317         - JavaScriptCore part of fix for http://bugzilla.opendarwin.org/show_bug.cgi?id=8049
3318           StringImpl hash traits deleted value creates an init routine for WebCore
3319           <rdar://problem/4442248> REGRESSION: WebCore has init routines (8049)
3320
3321         Change HashMap and HashSet implementation so they fold various types together.
3322         This allows us to implement maps and sets that use RefPtr<WebCore::StringImpl>
3323         and WebCore::String in terms of the underlying raw pointer type, and hence use
3324         -1 for the deleted value.
3325
3326         * kxmlcore/HashTraits.h: Added a new type to HashTraits, StorageTraits, which is a
3327         type to be used when storing a value that has the same layout as the type itself.
3328         This is used only for non-key cases. In the case of keys, the hash function must also
3329         be considered. Moved emptyValue out of GenericHashTraitsBase into GenericHashTraits.
3330         Added a new bool to HashTraits, needsRef, which indicates whether the type needs
3331         explicit reference counting. If the type itself has needsRef true, but the storage
3332         type has needsRef false, then the HashSet or HashMap has to handle the reference
3333         counting explicitly. Added hash trait specializations for all signed integer values
3334         that give -1 as the deleted value. Gave all integers StorageTraits of the canonical
3335         integer type of the same size so int and long will share code. Gave all pointers and
3336         RefPtrs StorageTraits of the appropriately sized integer type. Removed redundant
3337         TraitType and emptyValue definitions in the pointer specialization for HashTraits.
3338         Added PairBaseHashTraits, which doesn't try to set up needsDestruction and deletedValue.
3339         Useful for types where we don't want to force the existence of deletedValue, such as
3340         the type of a pair in a HashMap which is not the actual storage type. Removed an
3341         unneeded parameter from the DeletedValueAssigner template. Added HashKeyStorageTraits
3342         template, which determines what type can be used to store a given hash key type with
3343         a given hash function, and specialized it for pointers and RefPtr so that pointer
3344         hash tables share an underlying HashTable that uses IntHash.
3345
3346         * kxmlcore/HashTable.h: Added HashTableConstIteratorAdapter, HashTableIteratorAdapter,
3347         NeedsRef, RefCountManagerBase, RefCountManager, HashTableRefCountManagerBase, and
3348         HashTableRefCountManager. All are used by both HashSet and HashMap to handle hash
3349         tables where the type stored is not the same as the real value type.
3350
3351         * kxmlcore/HashFunctions.h: Added a new struct named IntTypes that finds an
3352         integer type given a sizeof value. Renamed pointerHash to intHash and made it
3353         use overloading and take integer parameters. Added an IntHash struct which is
3354         a hash function that works for integers. Changed PtrHash to call IntHash with
3355         an appropriately sized integer. Made IntHash the default hash function for
3356         many integer types. Made PtrHash the default hash function for RefPtr as well
3357         as for raw pointers.
3358
3359         * kxmlcore/HashSet.h: Changed implementation to use a separate "storage type"
3360         derived from the new traits. The HashTable will use the storage type and all
3361         necessary translation and ref/deref is done at the HashSet level. Also reorganized
3362         the file so that the HashSet is at the top and has no inline implementation inside
3363         it so it's easy to read the interface to HashSet.
3364
3365         * kxmlcore/HashMap.h: Changed implementation to use a separate "storage type"
3366         derived from the new traits. The HashTable will use the storage type and all
3367         necessary translation and ref/deref is done at the HashMap level. Also reorganized
3368         the file so that the HashMap is at the top and has no inline implementation inside
3369         it so it's easy to read the interface to HashMap.
3370
3371         * kxmlcore/HashMapPtrSpec.h: Removed. Superceded by optimizations in HashMap itself.
3372
3373         * JavaScriptCore.xcodeproj/project.pbxproj: Remove HashMapPtrSpec.h, resort files,
3374         and also remove some unnecessary build settings from the aggregate target that
3375         generates derived sources.
3376         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: Ditto.
3377
3378 2006-04-04  Timothy Hatcher  <timothy@apple.com>
3379
3380         Reviewed by Darin.
3381
3382         The Debug and Release frameworks are now built with install paths relative to the build products directory.
3383         This removes the need for other projects to build with -framework WebCore and -framework JavaScriptCore.
3384
3385         * JavaScriptCore.xcodeproj/project.pbxproj:
3386
3387 2006-04-04  Eric Seidel  <eseidel@apple.com>
3388
3389         Reviewed by ggaren.
3390
3391         Fix win32 build.
3392         Disable ASSERT redefinition warnings for now.
3393
3394         * JavaScriptCore.vcproj/testkjs/testkjs.vcproj:
3395         * kxmlcore/Assertions.h:
3396
3397 2006-04-04  Bjrn Graf  <bjoern.graf@gmail.com>
3398
3399         Reviewed by ggaren & darin.  Landed by eseidel.
3400
3401         Integrate CURL version of gettimeofday
3402         http://bugzilla.opendarwin.org/show_bug.cgi?id=7399
3403         Disable crash report dialogs for testkjs.exe in Release mode
3404         http://bugzilla.opendarwin.org/show_bug.cgi?id=8113
3405
3406         * kjs/testkjs.cpp:
3407         (StopWatch::start):
3408         (StopWatch::stop):
3409         (StopWatch::getElapsedMS):
3410         (main):
3411         (kjsmain):
3412
3413 2006-04-04  Eric Seidel  <eseidel@apple.com>
3414
3415         Reviewed by mjs.
3416
3417         * kjs/number_object.cpp:
3418         (NumberProtoFunc::callAsFunction): remove trunc() to fix win32.
3419
3420 2006-03-12  Maciej Stachowiak  <mjs@apple.com>
3421
3422         Reviewed by Darin.
3423
3424         - fixed "toPrecision sometimes messes up the last digit on intel Macs"
3425         http://bugzilla.opendarwin.org/show_bug.cgi?id=7748
3426
3427         * kjs/number_object.cpp:
3428         (intPow10): Compute integer powers of 10 using exponentiation by squaring.
3429         (NumberProtoFunc::callAsFunction): Use intPow10(n) in place of all pow(10.0, n),
3430         plus a bit of refactoring.
3431
3432 2006-04-03  Darin Adler  <darin@apple.com>
3433
3434         - tweak config.h and Platform.h to try to get buildbot working
3435           (making some small changes at the same time)
3436
3437         * kjs/config.h: Removed now-unneeded HAVE_ICU.
3438         * kxmlcore/Platform.h: Tweak how platform gets set up. Move all the
3439         USE stuff to the end.
3440
3441 2006-04-03  George Staikos   <staikos@opensource.apple.com>
3442
3443         Reviewed by Maciej.
3444
3445         Fix Win32 build breakage from previous commit, remove unused forward.
3446
3447 2006-04-03  George Staikos   <staikos@opensource.apple.com>
3448
3449         Reviewed by Maciej.
3450
3451         Implement a unicode abstraction layer to make JavaScriptCore much more
3452         easily ported to other platforms without having to take in libicu.  Also
3453         makes the unicode related code easier to understand.
3454
3455 2006-04-03  Timothy Hatcher  <timothy@apple.com>
3456
3457         Reviewed by Adele.
3458
3459         Fixes <rdar://problem/4498338> JavaScriptCore fails to compile for ppc64
3460         Other 64 bit build fixes.
3461
3462         * kjs/collector.cpp:
3463         (KJS::Collector::markOtherThreadConservatively): test for __DARWIN_UNIX03 and use __r1
3464         * kjs/dtoa.cpp:
3465         (Bigint::): cast PRIVATE_mem to unsigned to prevent warning
3466         * bindings/jni/jni_utility.cpp:
3467         (KJS::Bindings::getJavaVM): cast jniError to long to prevent format warning
3468         (KJS::Bindings::getJNIEnv): cast jniError to long to prevent format warning
3469         * bindings/runtime_root.cpp:
3470         (KJS::Bindings::addNativeReference): cast CFDictionaryGetValue to unsigned long to prevent warning
3471         (KJS::Bindings::removeNativeReference): cast CFDictionaryGetValue to unsigned long to prevent warning
3472
3473 2006-03-31  Darin Adler  <darin@apple.com>
3474
3475         Reviewed by Geoff.
3476
3477         - <rdar://problem/4395622> API: WebScriptObject.h incorrectly reports that -isSelectorExcludedFromWebScript returns NO by default
3478
3479         * bindings/objc/WebScriptObject.h: Fixed comment.
3480
3481 2006-03-31  Eric Seidel  <eseidel@apple.com>
3482
3483         Reviewed by mjs.
3484
3485         A bit more code cleanup.
3486
3487         * bindings/c/c_utility.cpp:
3488         (KJS::Bindings::convertValueToNPVariant):
3489         * bindings/objc/objc_runtime.mm:
3490         (convertValueToObjcObject):
3491         * bindings/objc/objc_utility.mm:
3492         (KJS::Bindings::convertValueToObjcValue):
3493         * kjs/function.cpp:
3494         (KJS::GlobalFuncImp::callAsFunction):
3495         * kjs/interpreter.cpp:
3496         (KJS::ExecState::lexicalInterpreter):
3497         * kjs/interpreter.h:
3498         * kjs/operations.cpp:
3499         (KJS::equal):
3500
3501 2006-03-30  Eric Seidel  <eseidel@apple.com>
3502
3503         Reviewed by anders.
3504
3505         Small code-style update.
3506
3507         * kjs/operations.cpp:
3508         (KJS::isNaN):
3509         (KJS::isInf):
3510         (KJS::isPosInf):
3511         (KJS::isNegInf):
3512         (KJS::equal):
3513         (KJS::strictEqual):
3514         (KJS::relation):
3515         (KJS::maxInt):
3516         (KJS::minInt):
3517         (KJS::add):
3518         (KJS::mult):
3519
3520 2006-03-31  Anders Carlsson  <andersca@mac.com>
3521
3522         Reviewed by Maciej.
3523
3524         Make sure the GetterSetterImp objects are marked as well.
3525
3526         * kjs/internal.cpp:
3527         (KJS::GetterSetterImp::mark):
3528         Call JSCell::mark().
3529
3530 2006-03-30  Eric Seidel  <eseidel@apple.com>
3531
3532         Reviewed by ggaren.
3533