JavaScriptCore/ChangeLog

   1     2006-07-23  Geoffrey Garen  <ggaren@apple.com>
   2
   3         Patch by Eric Albert, reviewed by Darin and me.
   4
   5         - Fixed <rdar://problem/4645931> JavaScriptCore stack-scanning code
   6         crashes (Collector::markStackObjectsConservatively)
   7
   8         * bindings/jni/jni_jsobject.cpp: On 64bit systems, jint is a long, not an
   9         int.
  10         (JavaJSObject::getSlot):
  11         (JavaJSObject::setSlot):
  12         * kjs/collector.cpp:
  13         (KJS::Collector::markCurrentThreadConservatively): Use a pointer instead of
  14         an int as 'dummy,' because on LP64 systems, an int is not pointer-aligned,
  15         and we want to scan the stack for pointers.
  16         * JavaScriptCore.xcodeproj/project.pbxproj: After a tense cease-fire, the
  17         XCode war has started up again!
  18
  19 === Safari-521.20 ===
  20
  21 2006-07-21  Geoffrey Garen  <ggaren@apple.com>
  22
  23         Reviewed by Darin.
  24
  25         <rdar://problem/4507265> REGRESSION: overlays don't work on HousingMaps.com (Google Maps-based site)
  26
  27         - Added support for strings that masquerade as undefined. Currently used
  28         by WebCore to implement undetectable style.filter.
  29
  30         The name is a little long, but it's only used in one line of code, so I
  31         thought clarity should win over brevity.
  32
  33         * JavaScriptCore.exp:
  34         * JavaScriptCore.xcodeproj/project.pbxproj:
  35         * kjs/object.h:
  36         * kjs/string_object.h:
  37         (KJS::StringInstanceThatMasqueradesAsUndefined::StringInstanceThatMasqueradesAsUndefined):
  38         (KJS::StringInstanceThatMasqueradesAsUndefined::masqueradeAsUndefined):
  39         (KJS::StringInstanceThatMasqueradesAsUndefined::toBoolean):
  40
  41 === Safari-521.19 ===
  42
  43 2006-07-20  Steve Falkenburg  <sfalken@apple.com>
  44
  45         Fix the build
  46
  47         * kjs/function.cpp:
  48         (KJS::escapeStringForPrettyPrinting):
  49
  50 2006-07-19  Anders Carlsson  <acarlsson@apple.com>
  51
  52         Reviewed by Darin.
  53
  54         <rdar://problem/4620655> REGRESSION(10.4.7-10.5): preview button for a blogger.com post doesn't work
  55
  56         * kjs/nodes2string.cpp:
  57         (StringNode::streamTo):
  58         Return the escaped string.
  59
  60         (RegExpNode::streamTo):
  61         Use the correct syntax.
  62
  63         * kjs/function.cpp:
  64         (KJS::escapeStringForPrettyPrinting):
  65         * kjs/function.h:
  66         Add escape function which escapes a string for pretty-printing so it can be parsed again.
  67
  68         * wtf/unicode/icu/UnicodeIcu.h:
  69         (WTF::Unicode::isPrintableChar):
  70         New function.
  71
  72 2006-07-18  Maciej Stachowiak  <mjs@apple.com>
  73
  74         Reviewed by Adele Peterson.
  75
  76         <rdar://problem/4589530> REGRESSION: null character in JS string causes parse error (works in Tiger and in other browsers)
  77
  78         * kjs/lexer.cpp:
  79         (Lexer::shift):
  80         (Lexer::lex):
  81         (Lexer::record16):
  82         (Lexer::scanRegExp):
  83         * kjs/lexer.h:
  84
  85 2006-07-18  Tim Omernick  <timo@apple.com>
  86
  87         Reviewed by Tim Hatcher.
  88
  89         Removed a misleading comment; we recently added support for the NPNVPluginElementNPObject
  90         variable.
  91
  92         * bindings/npapi.h:
  93
  94 === Safari-521.18 ===
  95
  96 2006-07-18  Timothy Hatcher  <timothy@apple.com>
  97
  98         Made the following headers public:
  99
 100         * JavaScriptCore.h
 101         * JSBase.h
 102         * JSContextRef.h
 103         * JSObjectRef.h
 104         * JSStringRef.h
 105         * JSValueRef.h
 106
 107         * JavaScriptCore.xcodeproj/project.pbxproj:
 108
 109 2006-07-17  Geoffrey Garen  <ggaren@apple.com>
 110
 111         Reviewed by Maciej.
 112
 113         - Added automatic prototype creation for classes.
 114
 115         A class stores a weak reference to a prototype, which is cleared when
 116         the prototype is garbage collected, to avoid a reference cycle.
 117
 118         We now have an attributes field in JSClassDefinition, that currently is
 119         used only to override automatic prototype creation when you want to manage your
 120         own prototypes, but can be extended in the future for other nefarious purposes.
 121
 122         Similarly, we have JSObjectMake and JSObjectMakeWithPrototype, the latter
 123         allowing you to manage your own prototypes.
 124
 125         JSObjectMakeConstructor is more interesting now, able to make a constructor
 126         on your behalf if you just give it a class.
 127
 128         - Removed bogus old code from minidom.js.
 129
 130         - Tweaked the headerdocs.
 131
 132         - Added more GC testing, which caught some leaks, and tested more funny
 133         edge cases in lookup, which caught a lookup bug. Removed some testing
 134         we used to do with MyObject because it was redundant with the new, cool
 135         stuff.
 136
 137         While fixing the lookup bug I retracted this change:
 138
 139             "If a static setProperty callback returns 'false', to indicate that the
 140             property was not set, we no longer forward the set request up the class
 141             chain, because that's almost certainly not what the programmer expected."
 142
 143         Returning false when setting a static property is a little silly, but you can see
 144         it being useful when shadowing a base class's static properties, and, regardless
 145         of usefullness, this is the defined behavior of the setProperty callback.
 146
 147         - Plus a little ASCII art, for the kids.
 148
 149 2006-07-17  Timothy Hatcher  <timothy@apple.com>
 150
 151         Reviewed by Maciej.
 152
 153         <rdar://problem/4634874> WebScriptObject and WebUndefined are no longer defined by WebKit
 154
 155         Moves WebScriptObject and WebUndefined up to WebCore.
 156         This change does create an upwards-dependancy on WebScriptObject existing
 157         in the loaded process, but this code path in JavaScriptCore does not get used
 158         unless it is through WebKit/WebCore. Moving all of the binding code out of
 159         JavaScriptCore might make sense in the future.
 160
 161         * JavaScriptCore.exp:
 162         * JavaScriptCore.xcodeproj/project.pbxproj:
 163         * bindings/objc/WebScriptObject.h: Replaced.
 164         * bindings/objc/WebScriptObject.mm: Removed.
 165         * bindings/objc/WebScriptObjectPrivate.h: Removed.
 166         * bindings/objc/objc_class.h:
 167         * bindings/objc/objc_instance.h:
 168         * bindings/objc/objc_instance.mm:
 169         (ObjcInstance::~ObjcInstance):
 170         * bindings/objc/objc_runtime.h:
 171         * bindings/objc/objc_runtime.mm:
 172         (convertValueToObjcObject):
 173         * bindings/objc/objc_utility.mm:
 174         (KJS::Bindings::convertValueToObjcValue):
 175         (KJS::Bindings::convertObjcValueToValue):
 176         (KJS::Bindings::createObjcInstanceForValue):
 177
 178 2006-07-17  Darin Adler  <darin@apple.com>
 179
 180         * API/JSBase.h: Fix comment formatting where things used to be lined up but
 181         are now ragged. Got rid of spaces that attempted to line things up.
 182         * API/JSObjectRef.h: Ditto. Also add missing periods for a couple of comments.
 183
 184 2006-07-17  Geoffrey Garen  <ggaren@apple.com>
 185
 186         Reviewed by Maciej.
 187
 188         - Removed the exception parameter from the initialize callback and, by extension,
 189         JSObjectMake. We have never had a need for exceptions when iniitializing,
 190         so the parameter seemed likely to "get in the way."
 191
 192         Also, an exception in JavaScript is thrown in response to input --
 193         "invalid URL", "index not a number", etc., so it's the job of the
 194         constructor function, not the initialize method, to throw.
 195
 196         If initialize *really* wants to throw, it can communicate the throw to
 197         the constructor through the constructed object's private data (e.g., set
 198         it to NULL, signaling to the consntructor that initialization failed.)
 199
 200         - Added JSObjectMakeWithData, which enables a constructor to set private
 201         data on an object *before* it has been initialized. That way, the initialize
 202         methods can properly operate on the data.
 203
 204         * API/JSNode.c: Moved ref into the initialize method, for better encapsulation,
 205         now that it's possible.
 206         * API/JSNodeList.c: ditto
 207         * API/minidom.c:
 208         (main): Do more aggressive garbage collection to test ref/deref and
 209         initialize/finalize.
 210         * API/minidom.js: store childNodes in a temporary so it doesn't get re-created
 211         like a thousand times. This makes debugging ref/deref easier
 212
 213 2006-07-17  Geoffrey Garen  <ggaren@apple.com>
 214
 215         Reviewed by Maciej.
 216
 217         - Changed the initialize callback to run from least derived class (parent
 218         class) to most derived class. This enables C++ style initialization,
 219         and derived class overriding of member data.
 220
 221         - Added excpetion propopgation to JSObjectMake, to support initialize
 222         exceptions, and generally round out our policy of making function
 223         signatures as long as possible.
 224
 225         * API/JSCallbackObject.h: Use ExecState instead of ContextRef, cuz we're
 226         in C++ land now.
 227
 228 2006-07-17  Geoffrey Garen  <ggaren@apple.com>
 229
 230         Reviewed by Maciej.
 231
 232         - Changed JSObjectMakeConstructor to JSObjectMakeConstructorWithCallback,
 233         to match JSObjectMakeFunctionWithCallback.
 234
 235         - Added prototype parameter, so the generated constructor
 236         automatically works with hasInstance / instanceof
 237
 238         - Moved hasInstance implementation from InternalFunctionImp to JSObject
 239         so that subclasses can inherit it without inheriting function-related baggage.
 240         More refactoring here would be good, but this seems like a good short-term
 241         solution.
 242
 243         (KJS::JSCallbackFunction::implementsHasInstance): override and return false,
 244         because callback functions aren't constructors.
 245
 246 2006-07-17  Maciej Stachowiak  <mjs@apple.com>
 247
 248         Reviewed by Geoff.
 249
 250         - add a JSContextRef parameter to all JSValueRef, JSObjectRef, and JSContextRef operations;
 251         except JSObject{Get,Set}PrivateData which can be assumed to be simple pure accessors.
 252
 253         Also renamed the parameter "context" to "ctx" because it makes the code read better with this pervasive
 254         but usually uninteresting parameter.
 255
 256         * API/JSBase.cpp:
 257         (JSEvaluateScript):
 258         (JSCheckScriptSyntax):
 259         (JSGarbageCollect):
 260         * API/JSBase.h:
 261         * API/JSCallbackObject.cpp:
 262         (KJS::JSCallbackObject::JSCallbackObject):
 263         (KJS::JSCallbackObject::init):
 264         (KJS::JSCallbackObject::getOwnPropertySlot):
 265         (KJS::JSCallbackObject::put):
 266         (KJS::JSCallbackObject::deleteProperty):
 267         (KJS::JSCallbackObject::toNumber):
 268         (KJS::JSCallbackObject::toString):
 269         * API/JSContextRef.cpp:
 270         (JSGlobalContextCreate):
 271         (JSGlobalContextRetain):
 272         (JSGlobalContextRelease):
 273         (JSContextGetGlobalObject):
 274         * API/JSContextRef.h:
 275         * API/JSNode.c:
 276         (JSNodePrototype_appendChild):
 277         (JSNodePrototype_removeChild):
 278         (JSNodePrototype_replaceChild):
 279         (JSNode_getNodeType):
 280         (JSNode_getFirstChild):
 281         (JSNode_prototype):
 282         * API/JSNodeList.c:
 283         (JSNodeListPrototype_item):
 284         (JSNodeList_length):
 285         (JSNodeList_getProperty):
 286         (JSNodeList_prototype):
 287         * API/JSObjectRef.cpp:
 288         (JSObjectMake):
 289         (JSObjectMakeFunctionWithCallback):
 290         (JSObjectMakeConstructor):
 291         (JSObjectMakeFunction):
 292         (JSObjectGetPrototype):
 293         (JSObjectSetPrototype):
 294         (JSObjectHasProperty):
 295         (JSObjectGetProperty):
 296         (JSObjectSetProperty):
 297         (JSObjectGetPropertyAtIndex):
 298         (JSObjectSetPropertyAtIndex):
 299         (JSObjectDeleteProperty):
 300         (JSObjectIsFunction):
 301         (JSObjectCallAsFunction):
 302         (JSObjectIsConstructor):
 303         (JSObjectCallAsConstructor):
 304         (JSObjectCopyPropertyNames):
 305         * API/JSObjectRef.h:
 306         * API/JSStringRef.cpp:
 307         * API/JSValueRef.cpp:
 308         (JSValueGetType):
 309         (JSValueIsUndefined):
 310         (JSValueIsNull):
 311         (JSValueIsBoolean):
 312         (JSValueIsNumber):
 313         (JSValueIsString):
 314         (JSValueIsObject):
 315         (JSValueIsObjectOfClass):
 316         (JSValueIsEqual):
 317         (JSValueIsStrictEqual):
 318         (JSValueIsInstanceOfConstructor):
 319         (JSValueMakeUndefined):
 320         (JSValueMakeNull):
 321         (JSValueMakeBoolean):
 322         (JSValueMakeNumber):
 323         (JSValueMakeString):
 324         (JSValueToBoolean):
 325         (JSValueToNumber):
 326         (JSValueToStringCopy):
 327         (JSValueToObject):
 328         (JSValueProtect):
 329         (JSValueUnprotect):
 330         * API/JSValueRef.h:
 331         * API/minidom.c:
 332         (print):
 333         * API/testapi.c:
 334         (MyObject_getProperty):
 335         (MyObject_deleteProperty):
 336         (MyObject_callAsFunction):
 337         (MyObject_callAsConstructor):
 338         (MyObject_convertToType):
 339         (print_callAsFunction):
 340         (main):
 341
 342 2006-07-16  Geoffrey Garen  <ggaren@apple.com>
 343
 344         Approved by Maciej, RS by Beth.
 345
 346         JSObjectMakeFunction -> JSObjectMakeFunctionWithCallback
 347         JSObjectMakeFunctionWithBody -> JSObjectMakeFunction
 348
 349         because the latter is more common, and more fundamental, than the former.
 350
 351         * API/APICast.h:
 352         (toJS):
 353         * API/JSBase.h:
 354         * API/JSCallbackObject.cpp:
 355         (KJS::JSCallbackObject::getOwnPropertySlot):
 356         (KJS::JSCallbackObject::put):
 357         (KJS::JSCallbackObject::deleteProperty):
 358         (KJS::JSCallbackObject::getPropertyNames):
 359         (KJS::JSCallbackObject::staticValueGetter):
 360         (KJS::JSCallbackObject::staticFunctionGetter):
 361         * API/JSClassRef.cpp:
 362         (OpaqueJSClass::OpaqueJSClass):
 363         (OpaqueJSClass::~OpaqueJSClass):
 364         * API/JSClassRef.h:
 365         * API/JSObjectRef.cpp:
 366         (JSClassCreate):
 367         (JSObjectMakeFunctionWithCallback):
 368         (JSObjectMakeFunction):
 369         (OpaqueJSPropertyNameArray::OpaqueJSPropertyNameArray):
 370         (JSObjectCopyPropertyNames):
 371         * API/JSObjectRef.h:
 372         * API/minidom.c:
 373         (main):
 374         * API/testapi.c:
 375         (main):
 376         * ChangeLog:
 377         * JavaScriptCore.exp:
 378
 379 2006-07-16  Geoffrey Garen  <ggaren@apple.com>
 380
 381         Laughed at by Beth.
 382
 383         Replace __JS with OpaqueJS because the former, while used by CF, is
 384         a prefix that's triply-reserved by the compiler. (_* is reserved in global
 385         names, _[A-Z] is reserved in all names, and __ is reserved in all names
 386         in C++.)
 387
 388         Opaque is an alternative used by other Mac OS X framewokrs.
 389
 390         * API/APICast.h:
 391         (toJS):
 392         * API/JSBase.h:
 393         * API/JSCallbackObject.cpp:
 394         (KJS::JSCallbackObject::getOwnPropertySlot):
 395         (KJS::JSCallbackObject::put):
 396         (KJS::JSCallbackObject::deleteProperty):
 397         (KJS::JSCallbackObject::getPropertyNames):
 398         (KJS::JSCallbackObject::staticValueGetter):
 399         (KJS::JSCallbackObject::staticFunctionGetter):
 400         * API/JSClassRef.cpp:
 401         (OpaqueJSClass::OpaqueJSClass):
 402         (OpaqueJSClass::~OpaqueJSClass):
 403         * API/JSClassRef.h:
 404         * API/JSObjectRef.cpp:
 405         (JSClassCreate):
 406         (OpaqueJSPropertyNameArray::OpaqueJSPropertyNameArray):
 407         (JSObjectCopyPropertyNames):
 408
 409 2006-07-16  Darin Adler  <darin@apple.com>
 410
 411         - try to fix Windows build
 412
 413         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
 414         Added some recently added files, removed some recently removed.
 415
 416 2006-07-16  Geoffrey Garen  <ggaren@apple.com>
 417
 418         Reviewed by Maciej.
 419
 420         - Change getProperty* to return undefined, rather than NULL, for missing
 421         properties, since that's what the spec says. Also added exception out
 422         parameters to the *Index functions, because they can call through to the
 423         regular functions, which can throw for custom objects.
 424
 425         * API/JSObjectRef.cpp:
 426         (JSObjectGetProperty):
 427         (JSObjectGetPropertyAtIndex):
 428         (JSObjectSetPropertyAtIndex):
 429         * API/JSObjectRef.h:
 430         * API/testapi.c:
 431         (main):
 432
 433 2006-07-16  Geoffrey Garen  <ggaren@apple.com>
 434
 435         Reviewed by Maciej.
 436
 437         - Properly document and handle NULL callbacks for static properties. We
 438         throw an exception in any case other than a ReadOnly property with a NULL
 439         setProperty callback, because a NULL callback almost certainly indicates
 440         a programming error. Also throw an exception if hasProperty returns true
 441         for a property that getProperty can't get.
 442
 443         - If a static setProperty callback returns 'false', to indicate that the
 444         property was not set, we no longer forward the set request up the class
 445         chain, because that's almost certainly not what the programmer expected.
 446
 447         * API/JSCallbackObject.cpp:
 448         (KJS::JSCallbackObject::getOwnPropertySlot):
 449         (KJS::JSCallbackObject::put):
 450         (KJS::JSCallbackObject::staticValueGetter):
 451         (KJS::JSCallbackObject::staticFunctionGetter):
 452         (KJS::JSCallbackObject::callbackGetter):
 453         * API/JSObjectRef.h:
 454         * API/minidom.js:
 455         * API/testapi.c:
 456         (MyObject_hasProperty):
 457         * API/testapi.js:
 458
 459 2006-07-16  Geoffrey Garen  <ggaren@apple.com>
 460
 461         Reviewed by Maciej.
 462
 463         - Added names to functions.
 464
 465         - Removed GetPrivate/SetPrivate from callbackFunctions and callbackConstructors.
 466         The private data idiom is that a JS object stores its native implementation
 467         as private data. For functions and constructors, the native implementation is nothing
 468         more than the callback they already store, so supporting private data, too,
 469         confuses the idiom. If you *really* want, you can still create a custom
 470         function with private data.
 471
 472         * API/JSCallbackConstructor.cpp:
 473         * API/JSCallbackConstructor.h:
 474         * API/JSCallbackFunction.cpp:
 475         (KJS::JSCallbackFunction::JSCallbackFunction):
 476         * API/JSCallbackFunction.h:
 477         * API/JSCallbackObject.cpp:
 478         (KJS::JSCallbackObject::staticFunctionGetter):
 479         * API/JSObjectRef.cpp:
 480         (JSObjectMakeFunction):
 481         (JSObjectMakeFunctionWithBody):
 482         (JSObjectGetPrivate):
 483         (JSObjectSetPrivate):
 484         * API/JSObjectRef.h:
 485         * API/minidom.c:
 486         (main):
 487         * API/testapi.c:
 488         (main):
 489
 490 2006-07-15  Maciej Stachowiak  <mjs@apple.com>
 491
 492         Reviewed by Darin.
 493
 494         - switch property lists to be vector+set of Identifiers instead of list of References
 495
 496         This has the following benefits:
 497
 498         - no duplicates in property lists
 499         - simplifies API calls
 500         - probably more efficient, since linked list is gone
 501         - entirely removed Reference, ReferenceList and ProtectedReference types from the API
 502
 503         * kjs/PropertyNameArray.cpp: Added.
 504         (KJS::PropertyNameArray::add): Check set, if not already there, add to
 505         vector.
 506         * kjs/PropertyNameArray.h: Added.
 507         (KJS::PropertyNameArray::PropertyNameArray): Newly added type, combines
 508         a set and a vector to make a unique but ordered list of identifiers.
 509         (KJS::PropertyNameArray::begin): ditto
 510         (KJS::PropertyNameArray::end): ditto
 511         (KJS::PropertyNameArray::size): ditto
 512         (KJS::PropertyNameArray::operator[]): ditto
 513         * kjs/array_instance.h:
 514         * kjs/array_object.cpp:
 515         (ArrayInstance::getPropertyNames): renamed from getPropertyList, updated
 516         for PropertyNameArray
 517         (ArrayInstance::setLength): updated for PropertyNameArray
 518         (ArrayInstance::pushUndefinedObjectsToEnd): ditto
 519         * kjs/nodes.cpp:
 520         (ForInNode::execute): updated for PropertyNameArray
 521         * kjs/nodes.h:
 522         * kjs/object.cpp:
 523         (KJS::JSObject::getPropertyNames): renamed from getPropertyList, updated
 524         for PropertyNameArray
 525         * kjs/object.h:
 526         * kjs/property_map.cpp:
 527         (KJS::PropertyMap::getEnumerablePropertyNames): updated for PropertyNameArray
 528         (KJS::PropertyMap::getSparseArrayPropertyNames): ditto
 529         * kjs/property_map.h:
 530         * kjs/protected_reference.h: Removed.
 531         * kjs/reference.cpp: Removed.
 532         * kjs/reference.h: Removed.
 533         * kjs/reference_list.cpp: Removed.
 534         * kjs/reference_list.h: Removed.
 535         * kjs/scope_chain.cpp:
 536         (KJS::ScopeChain::print): Use PropertyNamesArray instead of ReferenceList.
 537         * kjs/string_object.cpp:
 538         (StringInstance::getPropertyNames): Updated for new approach.
 539         * kjs/string_object.h:
 540         * kjs/ustring.h:
 541         * API/APICast.h:
 542         (toJS): Added overload for PropertyNameAccumulatorRef / PropertyNameArray*
 543         (toRef): ditto
 544         * API/JSBase.h:
 545         * API/JSCallbackObject.cpp:
 546         (KJS::JSCallbackObject::getPropertyNames): Fixed for new API.
 547         * API/JSCallbackObject.h:
 548         * API/JSObjectRef.cpp:
 549         (__JSPropertyNameArray::__JSPropertyNameArray): Type used for a publicly vended
 550         JSPropertyNameArrayRef.
 551         (JSObjectCopyPropertyNames): New API call - renamed / refactored from
 552         JSObjectCreatePropertyList
 553         (JSPropertyNameArrayRetain): new retain call for JSPropertyNameArray.
 554         (JSPropertyNameArrayRelease): new release call for - " -.
 555         (JSPropertyNameArrayGetCount): Instead of having to use a stateful enumerator you
 556         can now get the count and items in any order.
 557         (JSPropertyNameArrayGetNameAtIndex): See above.
 558         (JSPropertyNameAccumulatorAddName): What you add properties to is now an opaque
 559         accumulator object.
 560         * API/JSObjectRef.h: Prototyped new functions, removed old ones
 561         * JavaScriptCore.exp: Updated exported symbols.
 562         * JavaScriptCore.xcodeproj/project.pbxproj: Added new files, removed old.
 563         * API/testapi.c:
 564         (MyObject_getPropertyNames): Renamed / fixed callback to fit new paradigm.
 565         (main): Updated for new API.
 566
 567 2006-07-15  Darin Adler  <darin@apple.com>
 568
 569         - oops, missed a few more arrays that had to be const
 570
 571         * API/JSNode.c:
 572         (JSNodePrototype_appendChild): Added const.
 573         (JSNodePrototype_removeChild): Ditto.
 574         (JSNodePrototype_replaceChild): Ditto.
 575         (JSNode_construct): Ditto.
 576         * API/JSNodeList.c:
 577         (JSNodeListPrototype_item): Ditto.
 578         * API/JSObjectRef.cpp:
 579         (JSObjectMakeFunctionWithBody): Ditto.
 580         (JSObjectCallAsFunction): Ditto.
 581         (JSObjectCallAsConstructor): Ditto.
 582         * API/minidom.c:
 583         (print): Ditto.
 584         * API/testapi.c:
 585         (MyObject_callAsFunction): Ditto.
 586         (MyObject_callAsConstructor): Ditto.
 587         (print_callAsFunction): Ditto.
 588         (myConstructor_callAsConstructor): Ditto.
 589
 590 2006-07-15  Darin Adler  <darin@apple.com>
 591
 592         Reviewed by Maciej.
 593
 594         * API/JSNode.h: Made an array parameter const.
 595         * API/JSObjectRef.h: Made array parameters const. Fixed a comment.
 596
 597 2006-07-15  Geoffrey Garen  <ggaren@apple.com>
 598
 599         Reviewed by Maciej.
 600
 601         - JSObjectMakeFunctionWithBody includes a function name and named parameters now.
 602
 603         * API/JSObjectRef.cpp:
 604         (JSObjectMakeFunctionWithBody):
 605         * API/JSObjectRef.h:
 606         * API/testapi.c:
 607         (assertEqualsAsUTF8String): More informative failure reporting.
 608         (main): Test more function cases.
 609
 610 2006-07-15  Geoffrey Garen  <ggaren@apple.com>
 611
 612         Reviewed by Maciej.
 613
 614         - Moved the arguments passed to JSClassCreate into a single structure,
 615         called JSClassDefinition. This will enable easier structure
 616         migration/versioning in the future, if necessary.
 617
 618         - Added support for class names.
 619
 620         - kJSClassDefinitionNull replaces kJSObjectCallbacksNone.
 621
 622         - JSClass is becoming a fairly complex struct, so I migrated all of its
 623         implementation other than reference counting to the sruct.
 624
 625         - Also moved JSClass* functions in the API to JSObjectRef.cpp, since they're
 626         declared in JSObjectRef.h
 627
 628         - Also added some more informative explanation to the class structure doc.
 629
 630 2006-07-15  Darin Adler  <darin@apple.com>
 631
 632         Reviewed by Geoff.
 633
 634         - fix http://bugzilla.opendarwin.org/show_bug.cgi?id=8395
 635           <rdar://problem/4613467>
 636           REGRESSION: RegEx seems broken for hex escaped non breaking space
 637
 638         Test: fast/js/regexp-extended-characters-more.html
 639
 640         * pcre/pcre_exec.c:
 641         (match): Got rid of utf16Length local variable to guarantee there's no
 642         extra stack usage in recursive calls. Fixed two places in the PCRE_UTF16
 643         code that were using the length variable, which is the UTF-8 length of
 644         a character in the pattern, to move in the UTF-16 subject string. Instead
 645         they hardcode lengths of 1 and 2 since the code already handles BMP
 646         characters and surrogate pairs separately. Also fixed some DPRINTF so
 647         I could compile with DEBUG on.
 648         (pcre_exec): Changed a place that was checking for multibyte characters
 649         in the subject string to use ISMIDCHAR. Instead it was using hardcoded
 650         logic that was right for UTF-8 but wrong for UTF-16.
 651
 652         * pcre/pcre_compile.c: (pcre_compile2): Fixed a DPRINTF so I could compile
 653         with DEBUG on.
 654
 655 2006-07-14  Geoffrey Garen  <ggaren@apple.com>
 656
 657         RS by Maciej.
 658
 659         Global replace in the API of argc/argv with argumentCount/arguments.
 660
 661 2006-07-14  Geoffrey Garen  <ggaren@apple.com>
 662
 663         Reviewed by Maciej.
 664
 665         - Finalized exception handling in the API.
 666
 667         setProperty can throw because it throws for built-in arrays. getProperty
 668         and deleteProperty can throw because setProperty can throw and we want
 669         to be consistent, and also because they seem like "actions." callAsFunction,
 670         callAsConstructor, and hasInstance can throw, because they caan throw for
 671         all built-ins.
 672
 673         toBoolean can't throw because it's defined that way in the spec.
 674
 675         - Documented that toBoolean and toObject can't be overridden by custom
 676         objects because they're defined that way in the spec.
 677
 678 === Safari-521.17 ===
 679
 680 2006-07-14  Geoffrey Garen  <ggaren@apple.com>
 681
 682         Reviewed by Maciej.
 683
 684         - Implemented ref-counting of JSContexts by splitting into two datatypes:
 685         JSGlobalContext, which you can create/retain/release, and JSContext, which
 686         you can't.
 687
 688         Internally, you retain a JSGlobalContext/ExecState by retaining its
 689         interpreter, which, in the case of a global ExecState, owns it.
 690
 691         - Also made ~Interpreter() protected to catch places where Interpreter
 692         is manually deleted. (Can't make it private because some crazy fool
 693         decided it would be a good idea to subclass Interpreter in other frameworks.
 694         I pity da fool.)
 695
 696         * API/APICast.h:
 697         (toJS): Added cast for new JSGlobalContext
 698         * API/JSStringRef.h: Changed vague "you must" language to more specific
 699         (but, ultimately, equally vague) "behavior is undefined if you don't"
 700         language.
 701         (KJS::Interpreter::Interpreter): Factored more common initialization into
 702         init()
 703         * kjs/interpreter.h:
 704         (KJS::Interpreter::ref): new
 705         (KJS::Interpreter::deref): new
 706         (KJS::Interpreter::refCount): new
 707         * kjs/testkjs.cpp:
 708         (doIt): Ref-count the interpreter.
 709
 710 2006-07-14  Maciej Stachowiak  <mjs@apple.com>
 711
 712         Reviewed by Geoff.
 713
 714         - removed bool return value from JSObjectSetProperty, since it is inefficient and
 715         also doesn't work quite right
 716         - added JSObjectGetPropertyAtIndex and JSObjectSetPropertyAtIndex
 717
 718         * API/JSObjectRef.cpp:
 719         (JSObjectSetProperty): Removed return value and canPut stuff.
 720         (JSObjectGetPropertyAtIndex): Added.
 721         (JSObjectSetPropertyAtIndex): Added.
 722         * API/JSObjectRef.h: Prototyped and documented new functions.
 723
 724 2006-07-14  Geoffrey Garen  <ggaren@apple.com>
 725
 726         Reviewed by Beth.
 727
 728         Moved JSCheckScriptSyntax, JSEvaluateScript, and JSGarbageCollect into
 729         JSBase.h/.cpp. They don't belong in the value-specific or context-specific
 730         files because they're not part of the value or context implementations.
 731
 732         * API/JSBase.h:
 733         * API/JSContextRef.cpp:
 734         (JSContextGetGlobalObject):
 735         * API/JSContextRef.h:
 736         * API/JSValueRef.cpp:
 737         (JSValueUnprotect):
 738         * API/JSValueRef.h:
 739         * JavaScriptCore.xcodeproj/project.pbxproj:
 740
 741 2006-07-13  Timothy Hatcher  <timothy@apple.com>
 742
 743         Reviewed by Maciej.
 744
 745         Moved JavaScriptCore to be a public framework.
 746
 747         * JavaScriptCore.xcodeproj/project.pbxproj:
 748
 749 2006-07-13  Mark Rowe  <opendarwin.org@bdash.net.nz>
 750
 751         Reviewed by Geoffrey.
 752
 753         http://bugzilla.opendarwin.org/show_bug.cgi?id=9742
 754         Bug 9742: REGRESSION: WebKit hangs when loading <http://www.vtbook.com>
 755
 756         * kjs/value.h:
 757         (KJS::JSValue::getUInt32): Only types tagged as numeric can be converted to UInt32.
 758
 759 2006-07-13  Geoffrey Garen  <ggaren@apple.com>
 760
 761         Pleasing to Maciej.
 762
 763         - Renamed JSEvaluate -> JSEvaluateScript, JSCheckSyntax -> JSCheckScriptSyntax
 764         - Added exception out parameters to JSValueTo* and JSValueIsEqual because
 765         they can throw
 766         - Removed JSObjectGetDescription because it's useless and vague, and
 767         JSValueToString/JSValueIsObjectOfClass do a better job, anyway
 768         - Clarified comments about "IsFunction/Constructor" to indicate that they
 769         are true of all functions/constructors, not just those created by JSObjectMake*
 770
 771 2006-07-12  Geoffrey Garen  <ggaren@apple.com>
 772
 773         RS by Beth.
 774
 775         Finished previously approved JSInternalString -> JSString conversion
 776         by renaming the files.
 777
 778         * API/JSCallbackObject.cpp:
 779         * API/JSInternalStringRef.cpp: Removed.
 780         * API/JSInternalStringRef.h: Removed.
 781         * API/JSStringRef.cpp: Added.
 782         * API/JSStringRef.h: Added.
 783         * API/JavaScriptCore.h:
 784         * JavaScriptCore.xcodeproj/project.pbxproj:
 785
 786 2006-07-12  Geoffrey Garen  <ggaren@apple.com>
 787
 788         Reviewed by Maciej.
 789
 790         - Removed context and exception parameters from JSObjectGetPropertyEnumerator,
 791         removing the spurious use of ExecState inside JavaScriptCore that made
 792         us think this was necessary in the first place.
 793
 794         (StringInstance::getPropertyList): Use getString instead of toString because
 795         we know we're dealing with a string -- we put it there in the first place.
 796         While we're at it, store the string's size instead of retrieving it each time
 797         through the loop, to avoid the unnecessary killing of puppies.
 798         * kjs/string_object.h:
 799
 800 2006-07-12  Maciej Stachowiak  <mjs@apple.com>
 801
 802         Reviewed by Geoff.
 803
 804         - add handling of hasInstance callback for API objects
 805
 806         * API/JSCallbackObject.cpp:
 807         (KJS::JSCallbackObject::implementsHasInstance): Check if callback is present.
 808         (KJS::JSCallbackObject::hasInstance): Invoke appropriate callback.
 809         * API/JSCallbackObject.h:
 810         * API/JSClassRef.cpp:
 811         * API/JSObjectRef.h:
 812         * API/testapi.c:
 813         (MyObject_hasInstance): Test case; should match what construct would do.
 814         * API/testapi.js:
 815
 816 2006-07-11  Geoffrey Garen  <ggaren@apple.com>
 817
 818         Reviewed by Maciej.
 819
 820         - Implemented a vast number of renames and comment clarifications
 821         suggested during API review.
 822
 823         JSInternalString -> JSString
 824         JS*Make -> JSValueMake*, JSObjectMake*
 825         JSTypeCode -> JSType
 826         JSValueIsInstanceOf -> JSValueIsInstanceOfConstructor (reads strangely well in client code)
 827         JSGC*Protect -> JSValue*Protect
 828         JS*Callback -> JSObject*Callback
 829         JSGetPropertyListCallback -> JSObjectAddPropertiesToListCallback
 830         JSPropertyEnumeratorGetNext -> JSPropertyEnumeratorGetNextName
 831         JSString* ->
 832             JSStringCreateWithUTF8CString, JSStringGetUTF8CString,
 833             JSStringGetMaximumUTF8CStringSize JSStringIsEqualToUTF8CString,
 834             JSStringCreateWithCFString, JSStringCopyCFString, JSStringCreateWithCharacters.
 835
 836         - Changed functions taking a JSValue out arg and returning a bool indicating
 837         whether it was set to simply return a JSValue or NULL.
 838
 839         - Removed JSStringGetCharacters because it's more documentation than code,
 840         and it's just a glorified memcpy built on existing API functionality.
 841
 842         - Moved standard library includes into the headers that actually require them.
 843
 844         - Standardized use of the phrase "Create Rule."
 845
 846         - Removed JSLock from make functions that don't allocate.
 847
 848         - Added exception handling to JSValueToBoolean, since we now allow
 849         callback objects to throw exceptions upon converting to boolean.
 850
 851         - Renamed JSGCCollect to JSGarbageCollect.
 852
 853 2006-07-10  Geoffrey Garen  <ggaren@apple.com>
 854
 855         Reviewed by Darin.
 856
 857         - Changed public header includes to the <JavaScriptCore/ style.
 858         - Changed instances of 'buffer' to 'string' since we decided on
 859         JSInternalString instead of JSStringBuffer.
 860
 861         * API/JSContextRef.h:
 862         * API/JSInternalStringRef.cpp:
 863         (JSStringMake):
 864         (JSInternalStringRetain):
 865         (JSInternalStringRelease):
 866         (JSValueCopyStringValue):
 867         (JSInternalStringGetLength):
 868         (JSInternalStringGetCharactersPtr):
 869         (JSInternalStringGetCharacters):
 870         (JSInternalStringGetMaxLengthUTF8):
 871         (JSInternalStringGetCharactersUTF8):
 872         (CFStringCreateWithJSInternalString):
 873         * API/JSInternalStringRef.h:
 874         * API/JSNode.c:
 875         (JSNodePrototype_appendChild):
 876         (JSNode_getNodeType):
 877         * API/JSObjectRef.cpp:
 878         (JSObjectCallAsConstructor):
 879         * API/JSValueRef.h:
 880         * API/JavaScriptCore.h:
 881         * API/minidom.c:
 882         (main):
 883         (print):
 884         * API/testapi.c:
 885         (MyObject_getPropertyList):
 886         (myConstructor_callAsConstructor):
 887         (main): I noticed that we were prematurely releasing some string buffers,
 888         so I moved their release calls to the end of main(). I got rid of 'Buf' in *Buf
 889         (sometimes changing to 'IString', when necessary to differentiate a variable)
 890         to match the buffer->string change.
 891
 892 === Safari-521.16 ===
 893
 894 2006-07-10  Darin Adler  <darin@apple.com>
 895
 896         * kjs/value.cpp: (KJS::JSValue::toInt32Inline): Added inline keyword one more place.
 897         Just in case.
 898
 899 2006-07-10  Darin Adler  <darin@apple.com>
 900
 901         - fix the release build
 902
 903         * kjs/value.h:
 904         * kjs/value.cpp:
 905         (KJS::JSValue::toInt32Inline): Move the code here to an inline.
 906         (KJS::JSValue::toInt32): Call the inline from both overloaded toInt32 functions.
 907
 908 2006-07-10  David Kilzer  <ddkilzer@kilzer.net>
 909
 910         Reviewed by Darin.
 911
 912         - fix http://bugzilla.opendarwin.org/show_bug.cgi?id=9179
 913           Implement select.options.add() method
 914
 915         * JavaScriptCore.exp: Added overloaded KJS::JSValue::toInt32() method.
 916         * JavaScriptCore.xcodeproj/project.pbxproj: Altered attributes metadata for
 917         kjs/value.h to make it available as a forwarded header.
 918         * kjs/lookup.h:
 919         (KJS::lookupPut): Extracted a lookupPut() method from the existing lookupPut() method.
 920         The new method returns a boolean value if no entry is found in the lookup table.
 921         * kjs/value.cpp:
 922         (KJS::JSValue::toInt32): Overloaded toInt32() method with boolean "Ok" argument.
 923         * kjs/value.h: Ditto.
 924
 925 2006-07-10  Geoffrey Garen  <ggaren@apple.com>
 926
 927         No review necessary. Removed bogus file I accidentally checked in before.
 928
 929         * API/JSInternalSringRef.h: Removed.
 930
 931 2006-07-10  Geoffrey Garen  <ggaren@apple.com>
 932
 933         Reviewed by Darin.
 934
 935         Added exception out parameter to API object callbacks, removed semi-bogus
 936         JSContext(.*)Exception functions.
 937
 938         To make these calls syntactically simple, I added an exceptionSlot()
 939         method to the ExecState class, which provides a JSValue** slot in which to
 940         store a JSValue* exception.
 941
 942         * API/APICast.h:
 943         (toRef):
 944         * API/JSCallbackConstructor.cpp:
 945         (KJS::JSCallbackConstructor::construct):
 946         * API/JSCallbackFunction.cpp:
 947         (KJS::JSCallbackFunction::callAsFunction):
 948         * API/JSCallbackObject.cpp:
 949         (KJS::JSCallbackObject::init):
 950         (KJS::JSCallbackObject::getOwnPropertySlot):
 951         (KJS::JSCallbackObject::put):
 952         (KJS::JSCallbackObject::deleteProperty):
 953         (KJS::JSCallbackObject::construct):
 954         (KJS::JSCallbackObject::callAsFunction):
 955         (KJS::JSCallbackObject::getPropertyList):
 956         (KJS::JSCallbackObject::toBoolean):
 957         (KJS::JSCallbackObject::toNumber):
 958         (KJS::JSCallbackObject::toString):
 959         (KJS::JSCallbackObject::staticValueGetter):
 960         (KJS::JSCallbackObject::callbackGetter):
 961         * API/JSContextRef.cpp:
 962         (JSCheckSyntax):
 963         * API/JSContextRef.h:
 964         * API/JSNode.c:
 965         (JSNodePrototype_appendChild):
 966         (JSNodePrototype_removeChild):
 967         (JSNodePrototype_replaceChild):
 968         (JSNode_getNodeType):
 969         (JSNode_getChildNodes):
 970         (JSNode_getFirstChild):
 971         (JSNode_construct):
 972         * API/JSNode.h:
 973         * API/JSNodeList.c:
 974         (JSNodeListPrototype_item):
 975         (JSNodeList_length):
 976         (JSNodeList_getProperty):
 977         * API/JSObjectRef.h:
 978         * API/minidom.c:
 979         (print):
 980         * API/testapi.c:
 981         (MyObject_initialize):
 982         (MyObject_hasProperty):
 983         (MyObject_getProperty):
 984         (MyObject_setProperty):
 985         (MyObject_deleteProperty):
 986         (MyObject_getPropertyList):
 987         (MyObject_callAsFunction):
 988         (MyObject_callAsConstructor):
 989         (MyObject_convertToType):
 990         (print_callAsFunction):
 991         (myConstructor_callAsConstructor):
 992         (main):
 993         * JavaScriptCore.exp:
 994         * kjs/ExecState.h:
 995         (KJS::ExecState::exceptionHandle):
 996
 997 2006-07-10  Geoffrey Garen  <ggaren@apple.com>
 998
 999         Reviewed by Darin.
1000
1001         Improved type safety by implementing opaque JSValue/JSObject typing through
1002         abuse of 'const', not void*. Also fixed an alarming number of bugs
1003         exposed by this new type safety.
1004
1005         I made one design change in JavaScriptCore, which is that the JSObject
1006         constructor should take a JSValue* as its prototype argument, not a JSObject*,
1007         since we allow the prototype to be any JSValue*, including jsNull(), for
1008         example.
1009
1010         * API/APICast.h:
1011         (toJS):
1012         * API/JSBase.h:
1013         * API/JSCallbackConstructor.cpp:
1014         (KJS::JSCallbackConstructor::construct):
1015         * API/JSCallbackFunction.cpp:
1016         (KJS::JSCallbackFunction::callAsFunction):
1017         * API/JSCallbackObject.cpp:
1018         (KJS::JSCallbackObject::JSCallbackObject):
1019         (KJS::JSCallbackObject::getOwnPropertySlot):
1020         (KJS::JSCallbackObject::put):
1021         (KJS::JSCallbackObject::construct):
1022         (KJS::JSCallbackObject::callAsFunction):
1023         (KJS::JSCallbackObject::staticFunctionGetter):
1024         * API/JSCallbackObject.h:
1025         * API/JSContextRef.cpp:
1026         (JSEvaluate):
1027         * API/JSNode.c:
1028         (JSNodePrototype_appendChild):
1029         (JSNodePrototype_removeChild):
1030         (JSNodePrototype_replaceChild):
1031         * API/JSObjectRef.cpp:
1032         (JSObjectMake):
1033         (JSFunctionMakeWithBody):
1034         (JSObjectGetProperty):
1035         (JSObjectCallAsFunction):
1036         (JSObjectCallAsConstructor):
1037         * API/JSObjectRef.h:
1038         * API/testapi.c:
1039         (main):
1040         * ChangeLog:
1041         * kjs/object.h:
1042         (KJS::JSObject::JSObject):
1043
1044 2006-07-10  Geoffrey Garen  <ggaren@apple.com>
1045
1046         Approved by Maciej, Darin.
1047
1048         Renamed JSStringBufferRef to JSInternalStringRef. "Internal string" means the
1049         JavaScript engine's internal string representation, which is the most
1050         low-level and efficient representation to use when interfacing with JavaScript.
1051
1052         * API/APICast.h:
1053         (toJS):
1054         (toRef):
1055         * API/JSBase.h:
1056         * API/JSCallbackObject.cpp:
1057         (KJS::JSCallbackObject::getOwnPropertySlot):
1058         (KJS::JSCallbackObject::put):
1059         (KJS::JSCallbackObject::deleteProperty):
1060         (KJS::JSCallbackObject::staticValueGetter):
1061         (KJS::JSCallbackObject::callbackGetter):
1062         * API/JSContextRef.cpp:
1063         (JSEvaluate):
1064         (JSCheckSyntax):
1065         * API/JSContextRef.h:
1066         * API/JSInternalStringRef.cpp: Added.
1067         (JSStringMake):
1068         (JSInternalStringCreate):
1069         (JSInternalStringCreateUTF8):
1070         (JSInternalStringRetain):
1071         (JSInternalStringRelease):
1072         (JSValueCopyStringValue):
1073         (JSInternalStringGetLength):
1074         (JSInternalStringGetCharactersPtr):
1075         (JSInternalStringGetCharacters):
1076         (JSInternalStringGetMaxLengthUTF8):
1077         (JSInternalStringGetCharactersUTF8):
1078         (JSInternalStringIsEqual):
1079         (JSInternalStringIsEqualUTF8):
1080         (JSInternalStringCreateCF):
1081         (CFStringCreateWithJSInternalString):
1082         * API/JSInternalStringRef.h: Added.
1083         * API/JSNode.c:
1084         (JSNodePrototype_appendChild):
1085         (JSNode_getNodeType):
1086         (JSNode_getChildNodes):
1087         (JSNode_getFirstChild):
1088         * API/JSNodeList.c:
1089         (JSNodeList_length):
1090         (JSNodeList_getProperty):
1091         * API/JSObjectRef.cpp:
1092         (JSFunctionMakeWithBody):
1093         (JSObjectGetDescription):
1094         (JSObjectHasProperty):
1095         (JSObjectGetProperty):
1096         (JSObjectSetProperty):
1097         (JSObjectDeleteProperty):
1098         (JSPropertyEnumeratorGetNext):
1099         (JSPropertyListAdd):
1100         * API/JSObjectRef.h:
1101         * API/JSStringBufferRef.cpp: Removed.
1102         * API/JSStringBufferRef.h: Removed.
1103         * API/JSValueRef.h:
1104         * API/JavaScriptCore.h:
1105         * API/minidom.c:
1106         (main):
1107         (print):
1108         * API/testapi.c:
1109         (assertEqualsAsUTF8String):
1110         (assertEqualsAsCharactersPtr):
1111         (assertEqualsAsCharacters):
1112         (MyObject_hasProperty):
1113         (MyObject_getProperty):
1114         (MyObject_setProperty):
1115         (MyObject_deleteProperty):
1116         (MyObject_getPropertyList):
1117         (print_callAsFunction):
1118         (myConstructor_callAsConstructor):
1119         (main):
1120         * JavaScriptCore.exp:
1121         * JavaScriptCore.xcodeproj/project.pbxproj:
1122
1123 2006-07-08  Tim Omernick  <timo@apple.com>
1124
1125         Reviewed by Maciej.
1126
1127         Added an OpenGL drawing model to the Netscape Plug-in API.
1128
1129         * bindings/npapi.h:
1130
1131 2006-07-08  Timothy Hatcher  <timothy@apple.com>
1132
1133         Reviewed by Maciej.
1134
1135         Moved KJS_GetCreatedJavaVMs to jni_utility.cpp.
1136         Switched KJS_GetCreatedJavaVMs over to use dlopen and dlsym
1137         now that NSAddImage, NSLookupSymbolInImage and NSAddressOfSymbol
1138         are deprecated in Leopard.
1139
1140         * JavaScriptCore.exp:
1141         * JavaScriptCore.xcodeproj/project.pbxproj:
1142         * bindings/jni/jni_utility.cpp:
1143         (KJS::Bindings::KJS_GetCreatedJavaVMs):
1144         * bindings/softlinking.c: Removed.
1145         * bindings/softlinking.h: Removed.
1146
1147 2006-07-08  Geoffrey Garen  <ggaren@apple.com>
1148
1149         Reviewed by Anders.
1150
1151         - Make JSObjectGetProperty return a JSValue or NULL, like JSEvaluate does.
1152
1153         * API/JSObjectRef.cpp:
1154         (JSObjectGetProperty):
1155         * API/JSObjectRef.h:
1156         * API/testapi.c:
1157         (main):
1158
1159 2006-07-08  Geoffrey Garen  <ggaren@apple.com>
1160
1161         Style change -- no review necessary.
1162
1163         Use 0 instead of NULL in API .cpp files, to match our style guidelines.
1164
1165         * API/JSContextRef.cpp:
1166         (JSEvaluate):
1167         * API/JSObjectRef.cpp:
1168         (JSFunctionMakeWithBody):
1169         (JSObjectCallAsFunction):
1170         (JSObjectCallAsConstructor):
1171         * API/JSValueRef.cpp:
1172         (JSValueToObject):
1173
1174 2006-07-08  Geoffrey Garen  <ggaren@apple.com>
1175
1176         Reviewed by TimO.
1177
1178         - Added ability to pass NULL for thisObject when calling JSObjectCallAsFunction,
1179         to match JSEvaluate.
1180
1181         * API/JSObjectRef.cpp:
1182         (JSObjectCallAsFunction):
1183         * API/JSObjectRef.h:
1184         * API/testapi.c:
1185         (main):
1186
1187 === Safari-521.15 ===
1188
1189 2006-07-07  Geoffrey Garen  <ggaren@apple.com>
1190
1191         Reviewed by Maciej.
1192
1193         - Standardized which functions take a JSContext as an argument. The rule is:
1194         if you might execute JavaScript, you take a JSContext, otherwise you don't.
1195
1196         The FIXME in JSObjectRef.h requires refactoring some parts of Interpreter,
1197         but not API changes, so I'm putting it off until later.
1198
1199         * API/JSCallbackObject.cpp:
1200         (KJS::JSCallbackObject::JSCallbackObject):
1201         (KJS::JSCallbackObject::init):
1202         * API/JSCallbackObject.h:
1203         * API/JSContextRef.cpp:
1204         (JSContextCreate):
1205         * API/JSContextRef.h:
1206         * API/JSObjectRef.cpp:
1207         (JSObjectMake):
1208         (JSPropertyEnumeratorGetNext):
1209         * API/JSObjectRef.h:
1210         * API/testapi.c:
1211         (MyObject_initialize):
1212         (main):
1213         * JavaScriptCore.exp:
1214         * kjs/array_object.cpp:
1215         (ArrayInstance::setLength):
1216         (ArrayInstance::pushUndefinedObjectsToEnd):
1217         * kjs/nodes.cpp:
1218         (ForInNode::execute):
1219         * kjs/reference.cpp:
1220         (KJS::Reference::getPropertyName):
1221         (KJS::Reference::getValue):
1222         * kjs/reference.h:
1223         * kjs/scope_chain.cpp:
1224         (KJS::ScopeChain::print):
1225
1226 2006-07-06  Geoffrey Garen  <ggaren@apple.com>
1227
1228         Reviewed by Maciej.
1229
1230         More API action.
1231
1232         - Headerdoc finished
1233
1234         Semantic Changes:
1235         - Added a JSContextRef argument to many functions, because you need a
1236         JSContextRef for doing virtually anything. I expect to add this argument
1237         to even more functions in a future patch.
1238
1239         - Removed the globalObjectPrototype argument to JSContextCreate because
1240         you can't create an object until you have a context, so it's impossible
1241         to pass a prototype object to JSContextCreate. That's OK because (1) there's
1242         no reason to give the global object a prototype and (2) if you really want
1243         to, you can just use a separate call to JSObjectSetPrototype.
1244
1245         - Removed the JSClassRef argument to JSClassCreate because it was unnecessary,
1246         and you need to be able to make the global object's class before you've
1247         created a JSContext.
1248
1249         - Added an optional exception parameter to JSFunctionMakeWithBody because anything
1250         less would be uncivilized.
1251
1252         - Made the return value parameter to JSObjectGetProperty optional to match
1253         all other return value parameters in the API.
1254
1255         - Made JSObjectSetPrivate/JSObjectGetPrivate work on JSCallbackFunctions
1256         and JSCallbackConstructors. You could use an abstract base class or strategic
1257         placement of m_privateData in the class structure to implement this, but
1258         the former seemed like overkill, and the latter seemed too dangerous.
1259
1260         - Fixed a bug where JSPropertyEnumeratorGetNext would skip the first property.
1261
1262         Cosmetic Changes:
1263         - Reversed the logic of the JSChar #ifdef to avoid confusing headerdoc
1264
1265         - Removed function names from @function declarations because headeroc
1266         can parse them automatically, and I wanted to rule out manual mismatch.
1267
1268         - Changed Error::create to take a const UString& instead of a UString*
1269         because it was looking at me funny.
1270
1271         - Renamed JSStringBufferCreateWithCFString to JSStringBufferCreateCF
1272         because the latter is more concise and it matches JSStringBufferCreateUTF8.
1273
1274         * API/JSCallbackObject.cpp:
1275         (KJS::JSCallbackObject::getOwnPropertySlot):
1276         (KJS::JSCallbackObject::put):
1277         (KJS::JSCallbackObject::deleteProperty):
1278         (KJS::JSCallbackObject::getPropertyList):
1279         (KJS::JSCallbackObject::toBoolean):
1280         (KJS::JSCallbackObject::toNumber):
1281         (KJS::JSCallbackObject::toString):
1282         * API/JSClassRef.cpp:
1283         (JSClassCreate):
1284         * API/JSContextRef.cpp:
1285         (JSContextCreate):
1286         (JSContextSetException):
1287         * API/JSContextRef.h:
1288         * API/JSNode.c:
1289         (JSNodePrototype_class):
1290         (JSNode_class):
1291         * API/JSNodeList.c:
1292         (JSNodeListPrototype_class):
1293         (JSNodeList_class):
1294         * API/JSObjectRef.cpp:
1295         (JSObjectGetProperty):
1296         (JSObjectGetPrivate):
1297         (JSObjectSetPrivate):
1298         (JSObjectCallAsFunction):
1299         (JSObjectCallAsConstructor):
1300         (JSPropertyEnumeratorGetNext):
1301         * API/JSObjectRef.h:
1302         * API/JSStringBufferRef.cpp:
1303         (JSStringBufferCreateCF):
1304         * API/JSStringBufferRef.h:
1305         * API/JSValueRef.cpp:
1306         (JSValueIsInstanceOf):
1307         * API/JSValueRef.h:
1308         * API/minidom.c:
1309         (main):
1310         * API/minidom.js:
1311         * API/testapi.c:
1312         (MyObject_hasProperty):
1313         (MyObject_setProperty):
1314         (MyObject_deleteProperty):
1315         (MyObject_getPropertyList):
1316         (MyObject_convertToType):
1317         (MyObject_class):
1318         (main):
1319         * JavaScriptCore.exp:
1320
1321 2006-07-07  Geoffrey Garen  <ggaren@apple.com>
1322
1323         Reviewed by John.
1324
1325         - Fixed a few crashes resulting from NULL parameters to JSClassCreate.
1326
1327         * API/JSClassRef.cpp:
1328         (JSClassCreate):
1329         (JSClassRelease):
1330         * API/testapi.c: Added test for NULL parameters.
1331         (main):
1332
1333 2006-07-07  Geoffrey Garen  <ggaren@apple.com>
1334
1335         Reviewed by John, mocked by Darin.
1336
1337         - Changed JSEvaluate to take a JSObjectRef instead of a JSValueRef as
1338         "this," since "this" must be an object.
1339
1340         * API/JSContextRef.cpp:
1341         (JSEvaluate):
1342         * API/JSContextRef.h:
1343
1344 2006-07-07  Geoffrey Garen  <ggaren@apple.com>
1345
1346         Reviewed by John.
1347
1348         - More headerdoc
1349
1350         * API/JSBase.h:
1351         * JavaScriptCore.xcodeproj/project.pbxproj:
1352
1353 2006-07-05  Geoffrey Garen  <ggaren@apple.com>
1354
1355         RS by Beth.
1356
1357         Renamed JSCharBufferRef, which was universally unpopular, to JSStringBufferRef,
1358         which, hopefully, will be less unpopular.
1359
1360         * API/APICast.h:
1361         (toJS):
1362         (toRef):
1363         * API/JSBase.h:
1364         * API/JSCallbackObject.cpp:
1365         (KJS::JSCallbackObject::getOwnPropertySlot):
1366         (KJS::JSCallbackObject::put):
1367         (KJS::JSCallbackObject::deleteProperty):
1368         (KJS::JSCallbackObject::staticValueGetter):
1369         (KJS::JSCallbackObject::callbackGetter):
1370         * API/JSCharBufferRef.cpp: Removed.
1371         * API/JSCharBufferRef.h: Removed.
1372         * API/JSContextRef.cpp:
1373         (JSEvaluate):
1374         (JSCheckSyntax):
1375         * API/JSContextRef.h:
1376         * API/JSNode.c:
1377         (JSNodePrototype_appendChild):
1378         (JSNode_getNodeType):
1379         (JSNode_getChildNodes):
1380         (JSNode_getFirstChild):
1381         * API/JSNodeList.c:
1382         (JSNodeList_length):
1383         (JSNodeList_getProperty):
1384         * API/JSObjectRef.cpp:
1385         (JSFunctionMakeWithBody):
1386         (JSObjectGetDescription):
1387         (JSObjectHasProperty):
1388         (JSObjectGetProperty):
1389         (JSObjectSetProperty):
1390         (JSObjectDeleteProperty):
1391         (JSPropertyEnumeratorGetNext):
1392         (JSPropertyListAdd):
1393         * API/JSObjectRef.h:
1394         * API/JSStringBufferRef.cpp: Added.
1395         (JSStringMake):
1396         (JSStringBufferCreate):
1397         (JSStringBufferCreateUTF8):
1398         (JSStringBufferRetain):
1399         (JSStringBufferRelease):
1400         (JSValueCopyStringValue):
1401         (JSStringBufferGetLength):
1402         (JSStringBufferGetCharactersPtr):
1403         (JSStringBufferGetCharacters):
1404         (JSStringBufferGetMaxLengthUTF8):
1405         (JSStringBufferGetCharactersUTF8):
1406         (JSStringBufferIsEqual):
1407         (JSStringBufferIsEqualUTF8):
1408         (JSStringBufferCreateWithCFString):
1409         (CFStringCreateWithJSStringBuffer):
1410         * API/JSStringBufferRef.h: Added.
1411         * API/JSValueRef.h:
1412         * API/JavaScriptCore.h:
1413         * API/minidom.c:
1414         (main):
1415         (print):
1416         * API/testapi.c:
1417         (assertEqualsAsUTF8String):
1418         (assertEqualsAsCharactersPtr):
1419         (assertEqualsAsCharacters):
1420         (MyObject_hasProperty):
1421         (MyObject_getProperty):
1422         (MyObject_setProperty):
1423         (MyObject_deleteProperty):
1424         (MyObject_getPropertyList):
1425         (print_callAsFunction):
1426         (myConstructor_callAsConstructor):
1427         (main):
1428         * JavaScriptCore.exp:
1429         * JavaScriptCore.xcodeproj/project.pbxproj:
1430
1431 2006-07-05  Geoffrey Garen  <ggaren@apple.com>
1432
1433         RS by Beth.
1434
1435         Moved some code around for more logical file separation.
1436
1437         * API/JSBase.h:
1438         * API/JSContextRef.h:
1439         * API/JSObjectRef.cpp:
1440         * API/JSValueRef.cpp:
1441         (JSValueToObject):
1442         * API/JSValueRef.h:
1443
1444 2006-07-03  Geoffrey Garen  <ggaren@apple.com>
1445
1446         Reviewed by Maciej.
1447
1448         Implemented JSFunctionMakeWithBody, which parses a script as a function body
1449         in the global scope, and returns the resulting anonymous function.
1450
1451         I also removed private data from JSCallbackFunction. It never worked,
1452         since JSCallbackFunction doesn't inherit from JSCallbackObject.
1453
1454         * API/JSCallbackConstructor.cpp: Removed.
1455         * API/JSCallbackConstructor.h: Removed.
1456         * API/JSCallbackFunction.cpp:
1457         (KJS::JSCallbackFunction::JSCallbackFunction):
1458         (KJS::JSCallbackFunction::implementsConstruct):
1459         (KJS::JSCallbackFunction::construct):
1460         (KJS::JSCallbackFunction::implementsCall):
1461         (KJS::JSCallbackFunction::callAsFunction):
1462         * API/JSCallbackFunction.h:
1463         * API/JSCallbackObject.cpp:
1464         (KJS::JSCallbackObject::staticFunctionGetter):
1465         * API/JSObjectRef.cpp:
1466         (JSFunctionMake):
1467         (JSFunctionMakeWithCallbacks):
1468         * API/JSObjectRef.h:
1469         * API/JSValueRef.h:
1470         * API/minidom.c:
1471         (main):
1472         * API/testapi.c:
1473         (main):
1474         * JavaScriptCore.exp: Programmatically added all symbols exported by
1475         API object files, and sorted results
1476         * JavaScriptCore.xcodeproj/project.pbxproj:
1477
1478 2006-07-03  Geoffrey Garen  <ggaren@apple.com>
1479
1480         Reviewed by Maciej.
1481
1482         - Return syntax error in JSCheckSyntax through a JSValueRef* exception
1483         argument
1484
1485         * API/JSBase.h:
1486         * API/JSContextRef.cpp:
1487         (JSCheckSyntax):
1488         * API/testapi.c:
1489         (main):
1490         * JavaScriptCore.exp:
1491         * kjs/interpreter.cpp:
1492         (KJS::Interpreter::checkSyntax):
1493         * kjs/interpreter.h:
1494
1495 2006-07-04  Darin Adler  <darin@apple.com>
1496
1497         - fixed build
1498
1499         * wtf/MathExtras.h: Oops. Added missing #endif.
1500
1501 2006-07-04  Bjoern Graf  <bjoern.graf@gmail.com>
1502
1503         Reviewed by Maciej.
1504         Tweaked a bit by Darin.
1505
1506         - http://bugzilla.opendarwin.org/show_bug.cgi?id=9678
1507           work around MSVCRT's fmod function returning NaN for fmod(x, infinity) instead of x
1508
1509         * wtf/MathExtras.h: Added include of <float.h>.
1510         (isinf): Fix to return false for NAN.
1511         (wtf_fmod): Added. An inline that works around the bug.
1512
1513         * kjs/nodes.cpp:
1514         * kjs/number_object.cpp:
1515         * kjs/operations.cpp:
1516         * kjs/value.cpp:
1517         Added includes of MathExtras.h to all files using fmod.
1518
1519         * JavaScriptCore.xcodeproj/project.pbxproj: Let Xcode 2.3 have its way with
1520         the project.
1521
1522 2006-07-01  Geoffrey Garen  <ggaren@apple.com>
1523
1524         Reviewed by Darin.
1525
1526         - Refined value conversions in the API:
1527             - failed toNumber returns NaN
1528             - failed toObject returns NULL
1529             - failed toString returns empty string
1530
1531         - Refined excpetion handling in the API:
1532             - failed value conversions do not throw exceptions
1533             - uncaught exceptions in JSEvaluate, JSObjectCallAsFunction, and
1534               JSObjectCallAsConstructor are returned through a JSValueRef* exception
1535               argument
1536             - removed JSContextHasException, because JSContextGetException does
1537               the same job
1538
1539         * API/JSBase.h:
1540         * API/JSCharBufferRef.cpp:
1541         (JSValueCopyStringValue):
1542         * API/JSContextRef.cpp:
1543         (JSEvaluate):
1544         * API/JSContextRef.h:
1545         * API/JSNodeList.c: Added test code demonstrating how you would use
1546         toNumber, and why you probably don't need toUInt32, etc.
1547         (JSNodeListPrototype_item):
1548         (JSNodeList_getProperty):
1549         * API/JSObjectRef.cpp:
1550         (JSValueToObject):
1551         (JSObjectCallAsFunction):
1552         (JSObjectCallAsConstructor):
1553         * API/JSObjectRef.h:
1554         * API/JSValueRef.cpp:
1555         (JSValueToNumber):
1556         * API/JSValueRef.h:
1557         * API/minidom.c:
1558         (main):
1559         * API/testapi.c:
1560         (main): Added tests for new rules, and call to JSGCProtect to fix Intel
1561         crash
1562         * JavaScriptCore.exp:
1563
1564 2006-07-03  Darin Adler  <darin@apple.com>
1565
1566         - Rolled out HashMap implementation of NPRuntime, at least temporarily.
1567
1568         Fixes hang in the bindings section of layout tests seen on the
1569         buildbot.
1570
1571         This code was using HashMap<const char*, PrivateIdentifier*>.
1572         But that hashes based on pointer identity, not string value.
1573         The default hash for any pointer type is to hash based on the pointer.
1574         And WTF doesn't currently have a string hash for char*.
1575         We'll need to fix that before re-landing this patch.
1576
1577         (Formatting was also incorrect -- extra spaces in parentheses.)
1578
1579         * bindings/npruntime.cpp: Rolled out last change.
1580
1581 2006-07-02  Justin Haygood  <jhaygood@spsu.edu>
1582
1583         Reviewed, tweaked, landed by ggaren.
1584
1585         - Port NPRuntime from CFDictionary to HashMap.
1586
1587         * bindings/npruntime.cpp:
1588         (getStringIdentifierDictionary):
1589         (getIntIdentifierDictionary):
1590         (_NPN_GetStringIdentifier):
1591         (_NPN_GetIntIdentifier):
1592         * bindings/npruntime.h:
1593
1594 2006-07-01  Geoffrey Garen  <ggaren@apple.com>
1595
1596         Reviewed by Adele.
1597
1598         - Fixed <rdar://problem/4611197> REGRESSION: Liveconnect with Java test
1599         fails at http://www-sor.inria.fr/~dedieu/notes/liveconnect/simple_example.html
1600
1601         * JavaScriptCore.exp: Export symbols used by liveconnect
1602
1603 2006-06-29  Geoffrey Garen  <ggaren@apple.com>
1604
1605         Reviewed by Maciej.
1606
1607         - Phase 2 in the JS API.
1608
1609         - Added support for specifying static tables of values -- this should
1610         obviate the need for using complicated callbacks for most lookups.
1611
1612         -  API objects are now created with classes (JSClassRef) -- in order to support
1613         static values, and in order to prevent API objects from storing their
1614         data inline, and thus falling into the oversized (read: slow and prone to
1615         giving Maciej the frowny face) heap.
1616
1617         - Added two specialized JSObject subclasses -- JSCallbackFunction and JSCallbackConstructor --
1618         to allow JSFunctionMake and JSConstructorMake to continue to work with
1619         the new class model. Another solution to this problem would be to create
1620         a custom class object for each function and constructor you make. This
1621         solution is more code but also more efficient.
1622
1623         - Substantially beefed up the minidom example to demonstrate and test a
1624         lot of these techniques. Its output is still pretty haphazard, though.
1625
1626         - Gave the <kjs/ preface to some includes -- I'm told this matters to
1627         building on some versions of Linux.
1628
1629         - Implemented JSValueIsInstanceOf and JSValueIsObjectOfClass
1630
1631         - Removed GetDescription callback. Something in the class datastructure
1632         should take care of this.
1633
1634         * API/JSBase.h:
1635         * API/JSCallbackConstructor.cpp: Added.
1636         (KJS::):
1637         (KJS::JSCallbackConstructor::JSCallbackConstructor):
1638         (KJS::JSCallbackConstructor::implementsConstruct):
1639         (KJS::JSCallbackConstructor::construct):
1640         (KJS::JSCallbackConstructor::setPrivate):
1641         (KJS::JSCallbackConstructor::getPrivate):
1642         * API/JSCallbackConstructor.h: Added.
1643         (KJS::JSCallbackConstructor::classInfo):
1644         * API/JSCallbackFunction.cpp: Added.
1645         (KJS::):
1646         (KJS::JSCallbackFunction::JSCallbackFunction):
1647         (KJS::JSCallbackFunction::implementsCall):
1648         (KJS::JSCallbackFunction::callAsFunction):
1649         (KJS::JSCallbackFunction::setPrivate):
1650         (KJS::JSCallbackFunction::getPrivate):
1651         * API/JSCallbackFunction.h: Added.
1652         (KJS::JSCallbackFunction::classInfo):
1653         * API/JSCallbackObject.cpp:
1654         (KJS::):
1655         (KJS::JSCallbackObject::JSCallbackObject):
1656         (KJS::JSCallbackObject::init):
1657         (KJS::JSCallbackObject::~JSCallbackObject):
1658         (KJS::JSCallbackObject::className):
1659         (KJS::JSCallbackObject::getOwnPropertySlot):
1660         (KJS::JSCallbackObject::put):
1661         (KJS::JSCallbackObject::deleteProperty):
1662         (KJS::JSCallbackObject::implementsConstruct):
1663         (KJS::JSCallbackObject::construct):
1664         (KJS::JSCallbackObject::implementsCall):
1665         (KJS::JSCallbackObject::callAsFunction):
1666         (KJS::JSCallbackObject::getPropertyList):
1667         (KJS::JSCallbackObject::toBoolean):
1668         (KJS::JSCallbackObject::toNumber):
1669         (KJS::JSCallbackObject::toString):
1670         (KJS::JSCallbackObject::inherits):
1671         (KJS::JSCallbackObject::staticValueGetter):
1672         (KJS::JSCallbackObject::staticFunctionGetter):
1673         (KJS::JSCallbackObject::callbackGetter):
1674         * API/JSCallbackObject.h:
1675         * API/JSCharBufferRef.cpp:
1676         * API/JSClassRef.cpp: Added.
1677         (JSClassCreate):
1678         (JSClassRetain):
1679         (JSClassRelease):
1680         * API/JSClassRef.h: Added.
1681         (StaticValueEntry::StaticValueEntry):
1682         (StaticFunctionEntry::StaticFunctionEntry):
1683         (__JSClass::__JSClass):
1684         * API/JSContextRef.cpp:
1685         (JSContextCreate):
1686         (JSEvaluate):
1687         * API/JSContextRef.h:
1688         * API/JSNode.c: Added.
1689         (JSNodePrototype_appendChild):
1690         (JSNodePrototype_removeChild):
1691         (JSNodePrototype_replaceChild):
1692         (JSNodePrototype_class):
1693         (JSNode_getNodeType):
1694         (JSNode_getChildNodes):
1695         (JSNode_getFirstChild):
1696         (JSNode_finalize):
1697         (JSNode_class):
1698         (JSNode_prototype):
1699         (JSNode_new):
1700         (JSNode_construct):
1701         * API/JSNode.h: Added.
1702         * API/JSNodeList.c: Added.
1703         (JSNodeListPrototype_item):
1704         (JSNodeListPrototype_class):
1705         (JSNodeList_length):
1706         (JSNodeList_getProperty):
1707         (JSNodeList_finalize):
1708         (JSNodeList_class):
1709         (JSNodeList_prototype):
1710         (JSNodeList_new):
1711         * API/JSNodeList.h: Added.
1712         * API/JSObjectRef.cpp:
1713         (JSObjectMake):
1714         (JSFunctionMake):
1715         (JSConstructorMake):
1716         (__JSPropertyEnumerator::__JSPropertyEnumerator):
1717         (JSObjectCreatePropertyEnumerator):
1718         (JSPropertyEnumeratorGetNext):
1719         (JSPropertyEnumeratorRetain):
1720         (JSPropertyEnumeratorRelease):
1721         * API/JSObjectRef.h:
1722         (__JSObjectCallbacks::):
1723         * API/JSValueRef.cpp:
1724         (JSValueIsObjectOfClass):
1725         (JSValueIsInstanceOf):
1726         * API/JSValueRef.h:
1727         * API/Node.c: Added.
1728         (Node_new):
1729         (Node_appendChild):
1730         (Node_removeChild):
1731         (Node_replaceChild):
1732         (Node_ref):
1733         (Node_deref):
1734         * API/Node.h: Added.
1735         * API/NodeList.c: Added.
1736         (NodeList_new):
1737         (NodeList_length):
1738         (NodeList_item):
1739         (NodeList_ref):
1740         (NodeList_deref):
1741         * API/NodeList.h: Added.
1742         * API/minidom.c:
1743         (main):
1744         (print):
1745         (createStringWithContentsOfFile):
1746         * API/minidom.js:
1747         * API/testapi.c:
1748         (assertEqualsAsCharacters):
1749         (MyObject_getProperty):
1750         (MyObject_class):
1751         (myConstructor_callAsConstructor):
1752         (main):
1753         * API/testapi.js:
1754         * JavaScriptCore.xcodeproj/project.pbxproj:
1755
1756 2006-06-26  Kevin Ollivier  <kevino@theolliviers.com>
1757
1758         Reviewed and tweaked by Darin.
1759
1760         - Compile fixes for wx port / gcc 4.0.2
1761
1762         * kjs/array_object.cpp:
1763         Added missing headers.
1764
1765         * kjs/ExecState.h:
1766         gcc needs class prototypes before defining those classes as friend classes
1767
1768 2006-06-30  Mike Emmel  <mike.emmel@gmail.com>
1769
1770         Reviewed by Darin.
1771
1772         Compilation fixes for Linux/Gdk.
1773
1774         * JavaScriptCore/kjs/interpreter.cpp: added include of signal.h
1775         * JavaScriptCore/kjs/ExecState.h: added missing class declaration
1776         * JavaScriptCore/kjs/ExecState.cpp: case wrong on include of context.h
1777         * JavaScriptCore/JavaScriptCoreSources.bkl: added Context.cpp and ExecState.cpp
1778
1779 === Safari-521.14 ===
1780
1781 2006-06-29  Maciej Stachowiak  <mjs@apple.com>
1782
1783         Reviewed by Geoff.
1784
1785         - add headerdoc comments to some of the new JS API headers
1786
1787         * API/JSBase.h:
1788         * API/JSValueRef.h:
1789
1790 2006-06-28  Timothy Hatcher  <timothy@apple.com>
1791
1792         Prefer the Stabs debugging symbols format until DWARF bugs are fixed.
1793
1794         * JavaScriptCore.xcodeproj/project.pbxproj:
1795
1796 2006-06-27  Timothy Hatcher  <timothy@apple.com>
1797
1798         Reviewed by Tim O.
1799
1800         <rdar://problem/4448350> Deprecated ObjC language API used in JavaScriptCore, WebCore, WebKit and WebBrowser
1801
1802         Switch to the new ObjC 2 API, ifdefed the old code around OBJC_API_VERSION so it still works on Tiger.
1803         Removed the use of the old stringWithCString, switched to the new Tiger version that accepts an encoding.
1804         Lots of code style cleanup.
1805
1806         * JavaScriptCore.xcodeproj/project.pbxproj:
1807         * bindings/objc/objc_class.h:
1808         * bindings/objc/objc_class.mm:
1809         (KJS::Bindings::ObjcClass::~ObjcClass):
1810         (KJS::Bindings::_createClassesByIsAIfNecessary):
1811         (KJS::Bindings::ObjcClass::classForIsA):
1812         (KJS::Bindings::ObjcClass::name):
1813         (KJS::Bindings::ObjcClass::methodsNamed):
1814         (KJS::Bindings::ObjcClass::fieldNamed):
1815         (KJS::Bindings::ObjcClass::fallbackObject):
1816         * bindings/objc/objc_header.h:
1817         * bindings/objc/objc_instance.h:
1818         * bindings/objc/objc_instance.mm:
1819         (ObjcInstance::ObjcInstance):
1820         (ObjcInstance::~ObjcInstance):
1821         (ObjcInstance::operator=):
1822         (ObjcInstance::begin):
1823         (ObjcInstance::end):
1824         (ObjcInstance::getClass):
1825         (ObjcInstance::invokeMethod):
1826         (ObjcInstance::invokeDefaultMethod):
1827         (ObjcInstance::setValueOfField):
1828         (ObjcInstance::supportsSetValueOfUndefinedField):
1829         (ObjcInstance::setValueOfUndefinedField):
1830         (ObjcInstance::getValueOfField):
1831         (ObjcInstance::getValueOfUndefinedField):
1832         (ObjcInstance::defaultValue):
1833         (ObjcInstance::stringValue):
1834         (ObjcInstance::numberValue):
1835         (ObjcInstance::booleanValue):
1836         (ObjcInstance::valueOf):
1837         * bindings/objc/objc_runtime.h:
1838         * bindings/objc/objc_runtime.mm:
1839         (ObjcMethod::ObjcMethod):
1840         (ObjcMethod::name):
1841         (ObjcMethod::getMethodSignature):
1842         (ObjcMethod::setJavaScriptName):
1843         (ObjcField::name):
1844         (ObjcField::type):
1845         (ObjcField::valueFromInstance):
1846         (convertValueToObjcObject):
1847         (ObjcField::setValueToInstance):
1848         (ObjcArray::operator=):
1849         (ObjcArray::setValueAt):
1850         (ObjcArray::valueAt):
1851         (ObjcFallbackObjectImp::ObjcFallbackObjectImp):
1852         (ObjcFallbackObjectImp::callAsFunction):
1853         (ObjcFallbackObjectImp::defaultValue):
1854
1855 2006-06-28  Anders Carlsson  <acarlsson@apple.com>
1856
1857         Reviewed by Geoff.
1858
1859         http://bugzilla.opendarwin.org/show_bug.cgi?id=8636
1860         REGRESSION: JavaScript access to Java applet causes hang (_webViewURL not implemented)
1861
1862         * bindings/jni/jni_objc.mm:
1863         (KJS::Bindings::dispatchJNICall):
1864         Just pass nil as the calling URL. This will cause the Java plugin to use the URL of the page
1865         containing the applet (which is what we used to do).
1866
1867 2006-06-27  Timothy Hatcher  <timothy@apple.com>
1868
1869         Reviewed by Darin.
1870
1871         <rdar://problem/4406785> Add an export file to TOT JavaScriptCore like the Safari-2-0-branch
1872
1873         * JavaScriptCore.exp: Added.
1874         * JavaScriptCore.xcodeproj/project.pbxproj:
1875
1876 2006-06-25  Geoffrey Garen  <ggaren@apple.com>
1877
1878         Reviewed by Adele.
1879
1880         - Added JSConstructorMake to match JSFunctionMake, along with test code.
1881
1882         [ I checked in the ChangeLog before without the actual files. ]
1883
1884         * API/JSObjectRef.cpp:
1885         (JSConstructorMake):
1886         * API/JSObjectRef.h:
1887         * API/testapi.c:
1888         (myConstructor_callAsConstructor):
1889         (main):
1890         * API/testapi.js:
1891         * ChangeLog:
1892         * JavaScriptCore.xcodeproj/project.pbxproj: Moved testapi.c to the testapi
1893         target -- this was an oversight in my earlier check-in.
1894
1895 2006-06-25  Timothy Hatcher  <timothy@apple.com>
1896
1897         Reviewed by Darin.
1898
1899         Bug 9574: Drosera should show inline scripts within the original HTML
1900         http://bugzilla.opendarwin.org/show_bug.cgi?id=9574
1901
1902         Pass the starting line number and error message to the debugger.
1903
1904         * kjs/debugger.cpp:
1905         (Debugger::sourceParsed):
1906         * kjs/debugger.h:
1907         * kjs/function.cpp:
1908         (KJS::GlobalFuncImp::callAsFunction):
1909         * kjs/function_object.cpp:
1910         (FunctionObjectImp::construct):
1911         * kjs/interpreter.cpp:
1912         (KJS::Interpreter::evaluate):
1913
1914 2006-06-24  Alexey Proskuryakov  <ap@nypop.com>
1915
1916         Rubber-stamped by Eric.
1917
1918         Add a -h (do not follow symlinks) option to ln in derived sources build script (without it,
1919         a symlink was created inside the source directory on second build).
1920
1921         * JavaScriptCore.xcodeproj/project.pbxproj:
1922
1923 2006-06-24  David Kilzer  <ddkilzer@kilzer.net>
1924
1925         Reviewed by Timothy.
1926
1927         * Info.plist: Fixed copyright to include 2003-2006.
1928
1929 2006-06-24  Alexey Proskuryakov  <ap@nypop.com>
1930
1931         Reviewed by Darin.
1932
1933         - http://bugzilla.opendarwin.org/show_bug.cgi?id=9418
1934         WebKit will not build when Space exists in path
1935
1936         * JavaScriptCore.xcodeproj/project.pbxproj: Enclose search paths in quotes; create symlinks to
1937         avoid passing paths with spaces to make.
1938
1939 2006-06-23  Timothy Hatcher  <timothy@apple.com>
1940
1941         Reviewed by Darin.
1942
1943         Adding more operator[] overloads for long and short types.
1944
1945         * wtf/Vector.h:
1946         (WTF::Vector::operator[]):
1947
1948 === JavaScriptCore-521.13 ===
1949
1950 2006-06-22  Alexey Proskuryakov  <ap@nypop.com>
1951
1952         Build fix.
1953
1954         - http://bugzilla.opendarwin.org/show_bug.cgi?id=9539
1955         Another case error preventing build
1956
1957         * API/JSObjectRef.cpp: Changed "identifier.h" to "Identifier.h"
1958
1959 2006-06-22  David Kilzer  <ddkilzer@kilzer.net>
1960
1961         Build fix.
1962
1963         http://bugzilla.opendarwin.org/show_bug.cgi?id=9539
1964         Another case error preventing build
1965
1966         * API/APICast.h: Changed "UString.h" to "ustring.h".
1967
1968 2006-06-21  Geoffrey Garen  <ggaren@apple.com>
1969
1970         Fixed release build, fixed accidental infinite recursion due to
1971         last minute global replace gone awry.
1972
1973         * API/APICast.h:
1974         (toRef):
1975         * API/testapi.c:
1976         (assertEqualsAsBoolean):
1977         (assertEqualsAsNumber):
1978         (assertEqualsAsUTF8String):
1979         (assertEqualsAsCharactersPtr):
1980         * JavaScriptCore.xcodeproj/project.pbxproj:
1981
1982 2006-06-21  Geoffrey Garen  <ggaren@apple.com>
1983
1984         Reviewed by Anders.
1985
1986         - First cut at C API to JavaScript. Includes a unit test, 'testapi.c',
1987         and the outline of a test app, 'minidom.c'.
1988
1989         Includes one change to JSC internals: Rename propList to getPropertyList and have it
1990         take its target property list by reference so that subclasses can
1991         add properties to the list before calling through to their superclasses.
1992
1993         Also, I just ran prepare-ChangeLog in about 10 seconds, and I would like
1994         to give a shout-out to that.
1995
1996         * API/APICast.h: Added.
1997         (toJS):
1998         (toRef):
1999         * API/JSBase.h: Added.
2000         * API/JSCallbackObject.cpp: Added.
2001         (KJS::):
2002         (KJS::JSCallbackObject::JSCallbackObject):
2003         (KJS::JSCallbackObject::~JSCallbackObject):
2004         (KJS::JSCallbackObject::className):
2005         (KJS::JSCallbackObject::getOwnPropertySlot):
2006         (KJS::JSCallbackObject::put):
2007         (KJS::JSCallbackObject::deleteProperty):
2008         (KJS::JSCallbackObject::implementsConstruct):
2009         (KJS::JSCallbackObject::construct):
2010         (KJS::JSCallbackObject::implementsCall):
2011         (KJS::JSCallbackObject::callAsFunction):
2012         (KJS::JSCallbackObject::getPropertyList):
2013         (KJS::JSCallbackObject::toBoolean):
2014         (KJS::JSCallbackObject::toNumber):
2015         (KJS::JSCallbackObject::toString):
2016         (KJS::JSCallbackObject::setPrivate):
2017         (KJS::JSCallbackObject::getPrivate):
2018         (KJS::JSCallbackObject::cachedValueGetter):
2019         (KJS::JSCallbackObject::callbackGetter):
2020         * API/JSCallbackObject.h: Added.
2021         (KJS::JSCallbackObject::classInfo):
2022         * API/JSCharBufferRef.cpp: Added.
2023         (JSStringMake):
2024         (JSCharBufferCreate):
2025         (JSCharBufferCreateUTF8):
2026         (JSCharBufferRetain):
2027         (JSCharBufferRelease):
2028         (JSValueCopyStringValue):
2029         (JSCharBufferGetLength):
2030         (JSCharBufferGetCharactersPtr):
2031         (JSCharBufferGetCharacters):
2032         (JSCharBufferGetMaxLengthUTF8):
2033         (JSCharBufferGetCharactersUTF8):
2034         (JSCharBufferIsEqual):
2035         (JSCharBufferIsEqualUTF8):
2036         (JSCharBufferCreateWithCFString):
2037         (CFStringCreateWithJSCharBuffer):
2038         * API/JSCharBufferRef.h: Added.
2039         * API/JSContextRef.cpp: Added.
2040         (JSContextCreate):
2041         (JSContextDestroy):
2042         (JSContextGetGlobalObject):
2043         (JSEvaluate):
2044         (JSCheckSyntax):
2045         (JSContextHasException):
2046         (JSContextGetException):
2047         (JSContextClearException):
2048         (JSContextSetException):
2049         * API/JSContextRef.h: Added.
2050         * API/JSObjectRef.cpp: Added.
2051         (JSValueToObject):
2052         (JSObjectMake):
2053         (JSFunctionMake):
2054         (JSObjectGetDescription):
2055         (JSObjectGetPrototype):
2056         (JSObjectSetPrototype):
2057         (JSObjectHasProperty):
2058         (JSObjectGetProperty):
2059         (JSObjectSetProperty):
2060         (JSObjectDeleteProperty):
2061         (JSObjectGetPrivate):
2062         (JSObjectSetPrivate):
2063         (JSObjectIsFunction):
2064         (JSObjectCallAsFunction):
2065         (JSObjectIsConstructor):
2066         (JSObjectCallAsConstructor):
2067         (__JSPropertyListEnumerator::__JSPropertyListEnumerator):
2068         (JSObjectCreatePropertyEnumerator):
2069         (JSPropertyEnumeratorGetNext):
2070         (JSPropertyEnumeratorRetain):
2071         (JSPropertyEnumeratorRelease):
2072         (JSPropertyListAdd):
2073         * API/JSObjectRef.h: Added.
2074         * API/JSValueRef.cpp: Added.
2075         (JSValueGetType):
2076         (JSValueIsUndefined):
2077         (JSValueIsNull):
2078         (JSValueIsBoolean):
2079         (JSValueIsNumber):
2080         (JSValueIsString):
2081         (JSValueIsObject):
2082         (JSValueIsEqual):
2083         (JSValueIsStrictEqual):
2084         (JSUndefinedMake):
2085         (JSNullMake):
2086         (JSBooleanMake):
2087         (JSNumberMake):
2088         (JSValueToBoolean):
2089         (JSValueToNumber):
2090         (JSGCProtect):
2091         (JSGCUnprotect):
2092         (JSGCCollect):
2093         * API/JSValueRef.h: Added.
2094         * API/JavaScriptCore.h: Added.
2095         * API/minidom.c: Added.
2096         (main):
2097         * API/minidom.html: Added.
2098         * API/minidom.js: Added.
2099         * API/testapi.c: Added.
2100         (assertEqualsAsBoolean):
2101         (assertEqualsAsNumber):
2102         (assertEqualsAsUTF8String):
2103         (assertEqualsAsCharactersPtr):
2104         (assertEqualsAsCharacters):
2105         (MyObject_initialize):
2106         (MyObject_copyDescription):
2107         (MyObject_hasProperty):
2108         (MyObject_getProperty):
2109         (MyObject_setProperty):
2110         (MyObject_deleteProperty):
2111         (MyObject_getPropertyList):
2112         (MyObject_callAsFunction):
2113         (MyObject_callAsConstructor):
2114         (MyObject_convertToType):
2115         (MyObject_finalize):
2116         (print_callAsFunction):
2117         (main):
2118         (createStringWithContentsOfFile):
2119         * API/testapi.js: Added.
2120         * ChangeLog:
2121         * JavaScriptCore.xcodeproj/project.pbxproj:
2122         * bindings/npruntime_impl.h:
2123         * kjs/array_instance.h:
2124         * kjs/array_object.cpp:
2125         (ArrayInstance::getPropertyList):
2126         * kjs/interpreter.cpp:
2127         (KJS::Interpreter::evaluate):
2128         * kjs/nodes.cpp:
2129         (ForInNode::execute):
2130         * kjs/object.cpp:
2131         (KJS::JSObject::put):
2132         (KJS::JSObject::canPut):
2133         (KJS::JSObject::deleteProperty):
2134         (KJS::JSObject::propertyIsEnumerable):
2135         (KJS::JSObject::getPropertyAttributes):
2136         (KJS::JSObject::getPropertyList):
2137         * kjs/object.h:
2138         * kjs/property_map.cpp:
2139         (KJS::PropertyMap::get):
2140         * kjs/property_map.h:
2141         * kjs/scope_chain.cpp:
2142         (KJS::ScopeChain::print):
2143         * kjs/string_object.cpp:
2144         (StringInstance::getPropertyList):
2145         * kjs/string_object.h:
2146         * kjs/ustring.h:
2147         (KJS::UString::Rep::ref):
2148
2149 2006-06-20  Timothy Hatcher  <timothy@apple.com>
2150
2151         Reviewed by Geoff.
2152
2153         Make sure we clear the exception before returning so
2154         that future calls will not fail because of an earlier
2155         exception state. Assert on entry that the WebScriptObject
2156         is working with an ExecState that dose not have an exception.
2157         Document that evaluateWebScript and callWebScriptMethod return
2158         WebUndefined when an exception is thrown.
2159
2160         * bindings/objc/WebScriptObject.h:
2161         * bindings/objc/WebScriptObject.mm:
2162         (-[WebScriptObject callWebScriptMethod:withArguments:]):
2163         (-[WebScriptObject evaluateWebScript:]):
2164         (-[WebScriptObject setValue:forKey:]):
2165         (-[WebScriptObject valueForKey:]):
2166         (-[WebScriptObject removeWebScriptKey:]):
2167         (-[WebScriptObject webScriptValueAtIndex:]):
2168         (-[WebScriptObject setWebScriptValueAtIndex:value:]):
2169
2170 2006-06-19  Anders Carlsson  <acarlsson@apple.com>
2171
2172         Reviewed by John.
2173
2174         * kjs/interpreter.cpp:
2175         (KJS::TimeoutChecker::pauseTimeoutCheck):
2176         (KJS::TimeoutChecker::resumeTimeoutCheck):
2177         Fix argument order in setitimer calls.
2178
2179 2006-06-18  Anders Carlsson  <acarlsson@apple.com>
2180
2181         Reviewed by Geoff.
2182
2183         * kjs/interpreter.cpp:
2184         (KJS::TimeoutChecker::pauseTimeoutCheck):
2185         Do nothing if the timeout check hasn't been started.
2186
2187         (KJS::TimeoutChecker::resumeTimeoutCheck):
2188         Do nothing if the timeout check hasn't been started.
2189         Use the right signal handler when unblocking.
2190
2191         (KJS::Interpreter::handleTimeout):
2192         pause/resume the timeout check around the call to
2193         shouldInterruptScript().
2194
2195 2006-06-16  Ben Goodger <beng@google.com>
2196
2197         Reviewed by Maciej
2198
2199         http://bugzilla.opendarwin.org/show_bug.cgi?id=9491
2200         Windows build breaks in interpreter.cpp
2201
2202         * kjs/interpreter.cpp
2203         (KJS::TimeoutChecker::pauseTimeoutCheck):
2204         (KJS::TimeoutChecker::resumeTimeoutCheck):
2205         Make sure to only assert equality with s_executingInterpreter when it
2206         is being used (i.e. when HAVE(SYS_TIME_H) == true)
2207
2208 2006-06-17  David Kilzer  <ddkilzer@kilzer.net>
2209
2210         Reviewed by darin.
2211
2212         http://bugzilla.opendarwin.org/show_bug.cgi?id=9477
2213         REGRESSION: fast/dom/replaceChild.html crashes on WebKit ToT in debug build
2214
2215         * kjs/function.cpp:
2216         (KJS::FunctionImp::callAsFunction): Refetch the debugger after executing the function
2217         in case the WebFrame it was running in has since been destroyed.
2218
2219 2006-06-17  David Kilzer  <ddkilzer@kilzer.net>
2220
2221         Reviewed by ggaren.
2222
2223         http://bugzilla.opendarwin.org/show_bug.cgi?id=9476
2224         REGRESSION: Reproducible crash after closing window after viewing
2225         css2.1/t0803-c5501-imrgn-t-00-b-ag.html
2226
2227         * kjs/debugger.cpp:
2228         (Debugger::detach): Call setDebugger(0) for all interpreters removed from
2229         the 'attached to a debugger' list.
2230
2231 2006-06-17  Anders Carlsson  <acarlsson@apple.com>
2232
2233         Reviewed by Maciej and Geoff.
2234
2235         http://bugzilla.opendarwin.org/show_bug.cgi?id=7080
2236         Provide some way to stop a JavaScript infinite loop
2237
2238         * kjs/completion.h:
2239         (KJS::):
2240         Add Interrupted completion type.
2241
2242         * kjs/function.cpp:
2243         (KJS::FunctionImp::callAsFunction):
2244         (KJS::GlobalFuncImp::callAsFunction):
2245         Only set the exception on the new ExecState if the current one has had one.
2246
2247         * kjs/interpreter.cpp:
2248         (KJS::TimeoutChecker::startTimeoutCheck):
2249         (KJS::TimeoutChecker::stopTimeoutCheck):
2250         (KJS::TimeoutChecker::alarmHandler):
2251         (KJS::TimeoutChecker::pauseTimeoutCheck):
2252         (KJS::TimeoutChecker::resumeTimeoutCheck):
2253         New TimeoutChecker class which handles setting Interpreter::m_timedOut flag after a given
2254         period of time. This currently only works on Unix platforms where setitimer and signals are used.
2255
2256         (KJS::Interpreter::Interpreter):
2257         Initialize new member variables.
2258
2259         (KJS::Interpreter::~Interpreter):
2260         Destroy the timeout checker.
2261
2262         (KJS::Interpreter::startTimeoutCheck):
2263         (KJS::Interpreter::stopTimeoutCheck):
2264         (KJS::Interpreter::pauseTimeoutCheck):
2265         (KJS::Interpreter::resumeTimeoutCheck):
2266         Call the timeout checker.
2267
2268         (KJS::Interpreter::handleTimeout):
2269         Called on timeout. Resets the m_timedOut flag and calls shouldInterruptScript.
2270
2271         * kjs/interpreter.h:
2272         (KJS::Interpreter::setTimeoutTime):
2273         New function for setting the timeout time.
2274
2275         (KJS::Interpreter::shouldInterruptScript):
2276         New function. The idea is that this should be overridden by subclasses in order to for example
2277         pop up a dialog asking the user if the script should be interrupted.
2278
2279         (KJS::Interpreter::checkTimeout):
2280         New function which checks the m_timedOut flag and calls handleTimeout if it's set.
2281
2282         * kjs/nodes.cpp:
2283         (DoWhileNode::execute):
2284         (WhileNode::execute):
2285         (ForNode::execute):
2286         Call Interpreter::checkTimeout after each iteration of the loop.
2287
2288 2006-06-15  Timothy Hatcher  <timothy@apple.com>
2289
2290         Reviewed by Geoff and Darin.
2291
2292         Prefer the DWARF debugging symbols format for use in Xcode 2.3.
2293
2294         * JavaScriptCore.xcodeproj/project.pbxproj:
2295
2296 2006-06-14  Geoffrey Garen  <ggaren@apple.com>
2297
2298         Reviewed by Beth.
2299
2300         - fixed http://bugzilla.opendarwin.org/show_bug.cgi?id=9438
2301         Someone broke ToT: cannot build
2302
2303         * JavaScriptCore.xcodeproj/project.pbxproj:
2304         * bindings/runtime_root.h: Changed "Interpreter.h" to "interpreter.h"
2305
2306 2006-06-12  Geoffrey Garen  <ggaren@apple.com>
2307
2308         build fix
2309
2310         * bindings/objc/WebScriptObject.mm:
2311         (+[WebScriptObject throwException:]): Restore assignment I accidentally
2312         deleted in previous commit
2313
2314 2006-06-12  Geoffrey Garen  <ggaren@apple.com>
2315
2316         Reviewed by TimO, Maciej.
2317
2318         - Merged InterpreterImp code into Interpreter, which implements
2319         all interpreter functionality now. This is part of my continuing quest
2320         to create an external notion of JS "execution context" that is unified and simple --
2321         something to replace the mix of Context, ContextImp, ExecState, Interpreter,
2322         InterpreterImp, and JSRun.
2323
2324         All tests pass. Leaks test has not regressed from its baseline ~207 leaks
2325         with ~3460 leaked nodes.
2326
2327         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
2328         * JavaScriptCore.xcodeproj/project.pbxproj:
2329         * bindings/NP_jsobject.cpp:
2330         * bindings/objc/WebScriptObject.mm:
2331         (+[WebScriptObject throwException:]):
2332         * bindings/runtime_root.cpp:
2333         * bindings/runtime_root.h:
2334         * kjs/Context.cpp:
2335         (KJS::Context::Context):
2336         * kjs/ExecState.cpp: Added.
2337         (KJS::ExecState::lexicalInterpreter):
2338         * kjs/ExecState.h: Added.
2339         (KJS::ExecState::dynamicInterpreter):
2340         * kjs/SavedBuiltins.h: Added.
2341         * kjs/bool_object.cpp:
2342         (BooleanPrototype::BooleanPrototype):
2343         * kjs/collector.cpp:
2344         (KJS::Collector::collect):
2345         (KJS::Collector::numInterpreters):
2346         * kjs/context.h:
2347         * kjs/debugger.cpp:
2348         (Debugger::attach):
2349         (Debugger::detach):
2350         * kjs/function.cpp:
2351         (KJS::FunctionImp::callAsFunction):
2352         (KJS::GlobalFuncImp::callAsFunction):
2353         * kjs/function_object.cpp:
2354         (FunctionObjectImp::construct):
2355         * kjs/internal.cpp:
2356         * kjs/internal.h:
2357         * kjs/interpreter.cpp:
2358         (KJS::interpreterMap):
2359         (KJS::Interpreter::Interpreter):
2360         (KJS::Interpreter::init):
2361         (KJS::Interpreter::~Interpreter):
2362         (KJS::Interpreter::globalObject):
2363         (KJS::Interpreter::initGlobalObject):
2364         (KJS::Interpreter::globalExec):
2365         (KJS::Interpreter::checkSyntax):
2366         (KJS::Interpreter::evaluate):
2367         (KJS::Interpreter::builtinObject):
2368         (KJS::Interpreter::builtinFunction):
2369         (KJS::Interpreter::builtinArray):
2370         (KJS::Interpreter::builtinBoolean):
2371         (KJS::Interpreter::builtinString):
2372         (KJS::Interpreter::builtinNumber):
2373         (KJS::Interpreter::builtinDate):
2374         (KJS::Interpreter::builtinRegExp):
2375         (KJS::Interpreter::builtinError):
2376         (KJS::Interpreter::builtinObjectPrototype):
2377         (KJS::Interpreter::builtinFunctionPrototype):
2378         (KJS::Interpreter::builtinArrayPrototype):
2379         (KJS::Interpreter::builtinBooleanPrototype):
2380         (KJS::Interpreter::builtinStringPrototype):
2381         (KJS::Interpreter::builtinNumberPrototype):
2382         (KJS::Interpreter::builtinDatePrototype):
2383         (KJS::Interpreter::builtinRegExpPrototype):
2384         (KJS::Interpreter::builtinErrorPrototype):
2385         (KJS::Interpreter::builtinEvalError):
2386         (KJS::Interpreter::builtinRangeError):
2387         (KJS::Interpreter::builtinReferenceError):
2388         (KJS::Interpreter::builtinSyntaxError):
2389         (KJS::Interpreter::builtinTypeError):
2390         (KJS::Interpreter::builtinURIError):
2391         (KJS::Interpreter::builtinEvalErrorPrototype):
2392         (KJS::Interpreter::builtinRangeErrorPrototype):
2393         (KJS::Interpreter::builtinReferenceErrorPrototype):
2394         (KJS::Interpreter::builtinSyntaxErrorPrototype):
2395         (KJS::Interpreter::builtinTypeErrorPrototype):
2396         (KJS::Interpreter::builtinURIErrorPrototype):
2397         (KJS::Interpreter::mark):
2398         (KJS::Interpreter::interpreterWithGlobalObject):
2399         (KJS::Interpreter::saveBuiltins):
2400         (KJS::Interpreter::restoreBuiltins):
2401         * kjs/interpreter.h:
2402         (KJS::Interpreter::setCompatMode):
2403         (KJS::Interpreter::compatMode):
2404         (KJS::Interpreter::firstInterpreter):
2405         (KJS::Interpreter::nextInterpreter):
2406         (KJS::Interpreter::prevInterpreter):
2407         (KJS::Interpreter::debugger):
2408         (KJS::Interpreter::setDebugger):
2409         (KJS::Interpreter::setContext):
2410         (KJS::Interpreter::context):
2411         * kjs/nodes.cpp:
2412         (StatementNode::hitStatement):
2413         (RegExpNode::evaluate):
2414         * kjs/protect.h:
2415
2416 2006-06-12  Geoffrey Garen  <ggaren@apple.com>
2417
2418         Reviewed by Maciej.
2419
2420         - Have *.lut.h files #include lookup.h to eliminate surprising header
2421         include order dependency.
2422
2423         * DerivedSources.make:
2424         * kjs/array_object.cpp:
2425         * kjs/date_object.cpp:
2426         * kjs/date_object.h:
2427         (KJS::DateProtoFunc::):
2428         * kjs/lexer.cpp:
2429         * kjs/math_object.cpp:
2430         * kjs/number_object.cpp:
2431         * kjs/regexp_object.cpp:
2432         * kjs/string_object.cpp:
2433
2434 2006-06-10  Geoffrey Garen  <ggaren@apple.com>
2435
2436         - http://bugzilla.opendarwin.org/show_bug.cgi?id=8515
2437         Linux porting compile bug
2438
2439         Fix by Mike Emmel, Reviewed by Darin.
2440
2441         * JavaScriptCoreSources.bkl:
2442         * jscore.bkl:
2443         * wtf/Platform.h:
2444
2445 2006-06-09  Geoffrey Garen  <ggaren@apple.com>
2446
2447         Build fix -- I think :).
2448
2449         * JavaScriptCore.xcodeproj/project.pbxproj:
2450         * kjs/context.h:
2451
2452 2006-06-09  Geoffrey Garen  <ggaren@apple.com>
2453
2454         Reviewed by Eric (yay!).
2455
2456         - Removed Context wrapper for ContextImp, renamed ContextImp to Context,
2457         split Context into its own file -- Context.cpp -- renamed _var to m_var,
2458         change ' *' to '* '.
2459
2460         * JavaScriptCore.xcodeproj/project.pbxproj:
2461         * kjs/Context.cpp: Added.
2462         (KJS::Context::Context):
2463         (KJS::Context::~Context):
2464         (KJS::Context::mark):
2465         * kjs/context.h:
2466         (KJS::Context::scopeChain):
2467         (KJS::Context::variableObject):
2468         (KJS::Context::setVariableObject):
2469         (KJS::Context::thisValue):
2470         (KJS::Context::callingContext):
2471         (KJS::Context::activationObject):
2472         (KJS::Context::currentBody):
2473         (KJS::Context::function):
2474         (KJS::Context::arguments):
2475         (KJS::Context::pushScope):
2476         (KJS::Context::seenLabels):
2477         * kjs/function.cpp:
2478         (KJS::FunctionImp::callAsFunction):
2479         (KJS::FunctionImp::processParameters):
2480         (KJS::FunctionImp::argumentsGetter):
2481         (KJS::GlobalFuncImp::callAsFunction):
2482         * kjs/internal.cpp:
2483         (KJS::InterpreterImp::evaluate):
2484         * kjs/internal.h:
2485         (KJS::InterpreterImp::setContext):
2486         (KJS::InterpreterImp::context):
2487         * kjs/interpreter.cpp:
2488         * kjs/interpreter.h:
2489         (KJS::ExecState::context):
2490         (KJS::ExecState::ExecState):
2491         * kjs/nodes.cpp:
2492         (currentSourceId):
2493         (currentSourceURL):
2494         (ThisNode::evaluate):
2495         (ResolveNode::evaluate):
2496         (FunctionCallResolveNode::evaluate):
2497         (PostfixResolveNode::evaluate):
2498         (DeleteResolveNode::evaluate):
2499         (TypeOfResolveNode::evaluate):
2500         (PrefixResolveNode::evaluate):
2501         (AssignResolveNode::evaluate):
2502         (VarDeclNode::evaluate):
2503         (VarDeclNode::processVarDecls):
2504         (DoWhileNode::execute):
2505         (WhileNode::execute):
2506         (ForNode::execute):
2507         (ForInNode::execute):
2508         (ContinueNode::execute):
2509         (BreakNode::execute):
2510         (ReturnNode::execute):
2511         (WithNode::execute):
2512         (SwitchNode::execute):
2513         (LabelNode::execute):
2514         (TryNode::execute):
2515         (FuncDeclNode::processFuncDecl):
2516         (FuncExprNode::evaluate):
2517
2518 2006-06-07  Geoffrey Garen  <ggaren@apple.com>
2519
2520         Removed API directory I prematurely/accidentally added.
2521
2522         * API: Removed.
2523
2524 2006-06-05  Mitz Pettel  <opendarwin.org@mitzpettel.com>
2525
2526         Reviewed and landed by Geoff.
2527
2528         - fix a regression in ecma_3/String/regress-104375.js
2529
2530         * kjs/string_object.cpp:
2531         (substituteBackreferences): If a 2-digit back reference is out of range,
2532         parse it as a 1-digit reference (followed by the other digit). This matches
2533         Firefox's behavior.
2534
2535 2006-06-05  Geoffrey Garen  <ggaren@apple.com>
2536
2537         Reviewed By Maciej.
2538         Darin already reviewed this change on the branch. See <rdar://problem/4317701>.
2539
2540         - Fixed <rdar://problem/4291345> PCRE overflow in Safari JavaScriptCore
2541
2542         No test case because there's no behavior change.
2543
2544         * pcre/pcre_compile.c:
2545         (read_repeat_counts): Check for integer overflow / out of bounds
2546
2547 2006-06-05  Geoffrey Garen  <ggaren@apple.com>
2548
2549         Reviewed by aliu.
2550
2551         - Changed CString length from int to size_t. We sould probably do this
2552         for UString, too. (Darin, if you're reading this: Maciej said so.)
2553
2554         * kjs/function.cpp:
2555         (KJS::encode):
2556         * kjs/ustring.cpp:
2557         (KJS::CString::CString):
2558         (KJS::operator==):
2559         * kjs/ustring.h:
2560         (KJS::CString::size):
2561
2562 2006-06-04  Geoffrey Garen  <ggaren@apple.com>
2563
2564         Reviewed by Maciej.
2565
2566         - http://bugzilla.opendarwin.org/show_bug.cgi?id=9304
2567         Minor cleanup in JavaScriptCore
2568
2569         * kjs/value.h: Removed redundant declarations
2570
2571 2006-06-04  Darin Adler  <darin@apple.com>
2572
2573         Reviewed by Anders.
2574
2575         - changed deleteAllValues so it can work on "const" collections
2576           Deleting the values affects the values, not the pointers in the
2577           collection, so it's legitimate to do it to a const collection,
2578           and a case of that actually came up in the XPath code.
2579
2580         * wtf/HashMap.h:
2581         (WTF::deleteAllPairSeconds): Use const iterators.
2582         (WTF::deleteAllValues): Take const HashMap reference as a parameter.
2583         * wtf/HashSet.h:
2584         (WTF::deleteAllValues): Take const HashSet reference as a parameter,
2585         and use const iterators.
2586         * wtf/Vector.h:
2587         (WTF::deleteAllValues): Take const Vector reference as a parameter.
2588
2589         - added more functions that are present in <math.h> on some platforms,
2590           but not on others; moved here from various files in WebCore
2591
2592         * wtf/MathExtras.h:
2593         (isinf): Added.
2594         (isnan): Added.
2595         (lround): Added.
2596         (lroundf): Tweaked.
2597         (round): Added.
2598         (roundf): Tweaked.
2599         (signbit): Added.
2600
2601 2006-06-02  Mitz Pettel  <opendarwin.org@mitzpettel.com>
2602
2603         Reviewed by ggaren.
2604
2605         - http://bugzilla.opendarwin.org/show_bug.cgi?id=9234
2606           Implement $&, $' and $` replacement codes in String.prototype.replace
2607
2608         Test: fast/js/string-replace-3.html
2609
2610         * kjs/string_object.cpp:
2611         (substituteBackreferences): Added support for $& (matched substring),
2612         $` (everything preceding matched substring), $' (everything following
2613         matched substring) and 2-digit back references, and cleaned up a little.
2614
2615 2006-06-02 Adele Peterson  <adele@apple.com>
2616
2617         Reviewed by Darin.
2618
2619         Set incremental linking to no. This seems to fix a build problem I was seeing
2620         where dftables couldn't find a dll.
2621
2622         * JavaScriptCore.vcproj/dftables/dftables.vcproj:
2623
2624 2006-05-26  Steve Falkenburg  <sfalken@apple.com>
2625
2626         Build fixes/tweaks
2627
2628         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
2629
2630 === JavaScriptCore-521.11 ===
2631
2632 2006-05-24  Geoffrey Garen  <ggaren@apple.com>
2633
2634         Reviewed by mjs.
2635
2636         - JSC half of fix for <rdar://problem/4557926> TOT REGRESSSION: Crash
2637         occurs when attempting to view image in slideshow mode at
2638         http://d.smugmug.com/gallery/581716 ( KJS::IfNode::execute
2639         (KJS::ExecState*) + 312)
2640
2641         On alternate threads, DOMObjects remain in the
2642         ScriptInterpreter's cache because they're not collected. So, they
2643         need an opportunity to mark their children.
2644
2645         I'm not particularly happy with this solution because it fails to
2646         resolve many outstanding issues with the DOM object cache. Since none
2647         of those issues is a crasher or a serious compatibility concern,
2648         and since the behavior of other browsers is not much to go on in this
2649         case, I've filed <rdar://problem/4561439> about that, and I'm moving on
2650         with my life.
2651
2652         * JavaScriptCore.xcodeproj/project.pbxproj:
2653         * kjs/collector.cpp:
2654         (KJS::Collector::collect):
2655         * kjs/internal.cpp:
2656         (KJS::InterpreterImp::mark):
2657         * kjs/internal.h:
2658         * kjs/interpreter.cpp:
2659         (KJS::Interpreter::mark):
2660         * kjs/interpreter.h:
2661
2662 === JavaScriptCore-521.10 ===
2663
2664 2006-05-22  Timothy Hatcher  <timothy@apple.com>
2665
2666         Reviewed by Eric, Kevin and Geoff.
2667
2668         Merge open source build fixes. <rdar://problem/4555500>
2669
2670         * kjs/collector.cpp: look at the rsp register in x86_64
2671         (KJS::Collector::markOtherThreadConservatively):
2672         * wtf/Platform.h: add x86_64 to the platform list
2673
2674 2006-05-19  Anders Carlsson  <acarlsson@apple.com>
2675
2676         Reviewed by Geoff.
2677
2678         http://bugzilla.opendarwin.org/show_bug.cgi?id=8993
2679         Support function declaration in case statements
2680
2681         * kjs/grammar.y: Get rid of StatementList and use SourceElements instead.
2682
2683         * kjs/nodes.cpp:
2684         (CaseClauseNode::evalStatements):
2685         (CaseClauseNode::processVarDecls):
2686         (CaseClauseNode::processFuncDecl):
2687         (ClauseListNode::processFuncDecl):
2688         (CaseBlockNode::processFuncDecl):
2689         (SwitchNode::processFuncDecl):
2690         * kjs/nodes.h:
2691         (KJS::CaseClauseNode::CaseClauseNode):
2692         (KJS::ClauseListNode::ClauseListNode):
2693         (KJS::ClauseListNode::getClause):
2694         (KJS::ClauseListNode::getNext):
2695         (KJS::ClauseListNode::releaseNext):
2696         (KJS::SwitchNode::SwitchNode):
2697         Add processFuncDecl for the relevant nodes.
2698
2699         * kjs/nodes2string.cpp:
2700         (CaseClauseNode::streamTo):
2701         next got renamed to source.
2702
2703 2006-05-17  George Staikos <staikos@kde.org>
2704
2705         Reviewed by Maciej, Alexey, and Eric.
2706
2707         * pcre/pcre_compile.c:
2708         * pcre/pcre_get.c:
2709         * pcre/pcre_exec.c:
2710         * wtf/UnusedParam.h:
2711         Use /**/ in .c files to compile with non-C99 and non-GCC compilers.
2712
2713         * kjs/testkjs.cpp:
2714         Change include to <wtf/HashTraits.h> from "HashTraits.h" to avoid -I
2715
2716         * wtf/unicode/qt4/UnicodeQt4.h:
2717         Use correct parentheses and correct mask for utf-32 support.
2718
2719 2006-05-17  Alexey Proskuryakov  <ap@nypop.com>
2720
2721         Reviewed by Darin.
2722
2723         - fix http://bugzilla.opendarwin.org/show_bug.cgi?id=8870
2724         Crash typing in Yahoo auto-complete widget.
2725
2726         Test: fast/js/regexp-stack-overflow.html
2727
2728         * pcre/pcre-config.h: Define NO_RECURSE.
2729
2730 2006-05-16  George Staikos <staikos@kde.org>
2731
2732         Reviewed by Maciej.
2733
2734         Fix some warnings and strict compilation errors.
2735
2736         * kjs/nodes.cpp:
2737         * kjs/value.cpp:
2738
2739 2006-05-15  Alexey Proskuryakov  <ap@nypop.com>
2740
2741         * make-generated-sources.sh: Changed to be executable and removed
2742         text in the file generated by "svn diff".
2743
2744 2006-05-15  Geoffrey Garen  <ggaren@apple.com>
2745
2746         Reviewed by Maciej.
2747
2748         - Fixed <rdar://problem/4534904> please do not treat "debugger" as
2749         a reserved word while parsing JavaScript (and other ECMA reserved
2750         words)
2751
2752         AKA
2753
2754         http://bugzilla.opendarwin.org/show_bug.cgi?id=6179
2755         We treat "char" as a reserved word in JavaScript and firefox/IE do
2756         not
2757
2758         (1) I unreserved most of the spec's "future reserved words" because
2759         they're not reserved in IE or FF. (Most, but not all, because IE
2760         somewhat randomly *does* reserve a few of them.)
2761         (2) I made 'debugger' a legitimate statement that acts like an empty
2762         statement because FF and IE support it.
2763
2764         * kjs/grammar.y:
2765         * kjs/keywords.table:
2766
2767 2006-05-15  Tim Omernick  <timo@apple.com>
2768
2769         Reviewed by John Sullivan.
2770
2771         Part of <rdar://problem/4466508> Add 64-bit support to the Netscape Plugin API
2772
2773         Added to the Netscape Plugin API the concept of "plugin drawing models".  The drawing model
2774         determines the kind of graphics context created by the browser for the plugin, as well as
2775         the Mac types of various Netscape Plugin API data structures.
2776
2777         There is a drawing model to represent the old QuickDraw-based API.  It is used by default
2778         if QuickDraw is available on the system, unless the plugin specifies another drawing model.
2779
2780         The big change is the addition of the CoreGraphics drawing model.  A plugin may request this
2781         drawing model to obtain access to a CGContextRef for drawing, instead of a QuickDraw CGrafPtr.
2782
2783         * bindings/npapi.h:
2784         Define NP_NO_QUICKDRAW when compiling 64-bit; there is no 64-bit QuickDraw.
2785         Added NPNVpluginDrawingModel, NPNVsupportsQuickDrawBool, and NPNVsupportsCoreGraphicsBool
2786         variables.
2787         Added NPDrawingModel enumeration.  Currently the only drawing models are QuickDraw and
2788         CoreGraphics.
2789         NPRegion's type now depends on the drawing model specified by the plugin.
2790         NP_Port is now only defined when QuickDraw is available.
2791         Added NP_CGContext, which is the type of the NPWindow's "window" member in CoreGraphics mode.
2792
2793 2006-05-13  Kevin M. Ollivier  <kevino@theolliviers.com>
2794
2795         Reviewed by Darin, landed by ap.
2796
2797         - http://bugzilla.opendarwin.org/show_bug.cgi?id=8528
2798           Bakefiles (and generated Makefiles) for wx and gdk ports
2799
2800         * make-generated-sources.sh:
2801         Added script to configure environment to run DerivedSources.make
2802
2803         * JavaScriptCoreSources.bkl:
2804         Added JavaScriptCore sources list for Bakefile.
2805
2806         * jscore.bkl:
2807         Bakefile used to generate JavaScriptCore project files
2808         (currently only used by wx and gdk ports)
2809
2810 2006-05-09  Steve Falkenburg  <sfalken@apple.com>
2811
2812         Fix Windows build.
2813         Minor fixes to WTF headers.
2814
2815         Reviewed by kevin.
2816
2817         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: Fix include dirs, paths to files.
2818         * JavaScriptCore.vcproj/testkjs/testkjs.vcproj: Fix include dirs.
2819         * wtf/Assertions.h: include Platform.h to get definition for COMPILER()
2820         * wtf/Vector.h: include FastMalloc.h for definition of fastMalloc, fastFree
2821
2822 2006-05-09  Maciej Stachowiak  <mjs@apple.com>
2823
2824         Rubber stamped by Anders.
2825
2826         - renamed kxmlcore to wtf
2827
2828         kxmlcore --> wtf
2829         KXMLCore --> WTF
2830         KXC --> WTF
2831
2832         * JavaScriptCore.xcodeproj/project.pbxproj:
2833         * bindings/c/c_instance.cpp:
2834         * bindings/objc/WebScriptObject.mm:
2835         * kjs/JSImmediate.h:
2836         * kjs/Parser.cpp:
2837         * kjs/Parser.h:
2838         * kjs/array_object.cpp:
2839         * kjs/collector.cpp:
2840         (KJS::Collector::registerThread):
2841         * kjs/collector.h:
2842         * kjs/config.h:
2843         * kjs/function.cpp:
2844         (KJS::isStrWhiteSpace):
2845         * kjs/function.h:
2846         * kjs/identifier.cpp:
2847         * kjs/internal.cpp:
2848         * kjs/internal.h:
2849         * kjs/lexer.cpp:
2850         (Lexer::shift):
2851         (Lexer::isWhiteSpace):
2852         (Lexer::isIdentStart):
2853         (Lexer::isIdentPart):
2854         * kjs/lookup.cpp:
2855         * kjs/nodes.cpp:
2856         * kjs/nodes.h:
2857         * kjs/number_object.cpp:
2858         * kjs/object.h:
2859         * kjs/property_map.cpp:
2860         * kjs/property_map.h:
2861         * kjs/string_object.cpp:
2862         (StringProtoFunc::callAsFunction):
2863         * kjs/testkjs.cpp:
2864         (testIsInteger):
2865         * kjs/ustring.cpp:
2866         * kjs/ustring.h:
2867         * kxmlcore: Removed.
2868         * kxmlcore/AlwaysInline.h: Removed.
2869         * kxmlcore/Assertions.cpp: Removed.
2870         * kxmlcore/Assertions.h: Removed.
2871         * kxmlcore/FastMalloc.cpp: Removed.
2872         * kxmlcore/FastMalloc.h: Removed.
2873         * kxmlcore/FastMallocInternal.h: Removed.
2874         * kxmlcore/Forward.h: Removed.
2875         * kxmlcore/HashCountedSet.h: Removed.
2876         * kxmlcore/HashFunctions.h: Removed.
2877         * kxmlcore/HashMap.h: Removed.
2878         * kxmlcore/HashSet.h: Removed.
2879         * kxmlcore/HashTable.cpp: Removed.
2880         * kxmlcore/HashTable.h: Removed.
2881         * kxmlcore/HashTraits.h: Removed.
2882         * kxmlcore/ListRefPtr.h: Removed.
2883         * kxmlcore/Noncopyable.h: Removed.
2884         * kxmlcore/OwnArrayPtr.h: Removed.
2885         * kxmlcore/OwnPtr.h: Removed.
2886         * kxmlcore/PassRefPtr.h: Removed.
2887         * kxmlcore/Platform.h: Removed.
2888         * kxmlcore/RefPtr.h: Removed.
2889         * kxmlcore/TCPageMap.h: Removed.
2890         * kxmlcore/TCSpinLock.h: Removed.
2891         * kxmlcore/TCSystemAlloc.cpp: Removed.
2892         * kxmlcore/TCSystemAlloc.h: Removed.
2893         * kxmlcore/UnusedParam.h: Removed.
2894         * kxmlcore/Vector.h: Removed.
2895         * kxmlcore/VectorTraits.h: Removed.
2896         * kxmlcore/unicode: Removed.
2897         * kxmlcore/unicode/Unicode.h: Removed.
2898         * kxmlcore/unicode/UnicodeCategory.h: Removed.
2899         * kxmlcore/unicode/icu: Removed.
2900         * kxmlcore/unicode/icu/UnicodeIcu.h: Removed.
2901         * kxmlcore/unicode/posix: Removed.
2902         * kxmlcore/unicode/qt3: Removed.
2903         * kxmlcore/unicode/qt4: Removed.
2904         * kxmlcore/unicode/qt4/UnicodeQt4.h: Removed.
2905         * pcre/pcre_get.c:
2906         * wtf: Added.
2907         * wtf/Assertions.cpp:
2908         * wtf/Assertions.h:
2909         * wtf/FastMalloc.cpp:
2910         (WTF::TCMalloc_ThreadCache::Scavenge):
2911         (WTF::do_malloc):
2912         (WTF::do_free):
2913         (WTF::TCMallocGuard::TCMallocGuard):
2914         (WTF::malloc):
2915         (WTF::free):
2916         (WTF::calloc):
2917         (WTF::cfree):
2918         (WTF::realloc):
2919         * wtf/FastMalloc.h:
2920         * wtf/FastMallocInternal.h:
2921         * wtf/Forward.h:
2922         * wtf/HashCountedSet.h:
2923         * wtf/HashFunctions.h:
2924         * wtf/HashMap.h:
2925         * wtf/HashSet.h:
2926         * wtf/HashTable.cpp:
2927         * wtf/HashTable.h:
2928         * wtf/HashTraits.h:
2929         * wtf/ListRefPtr.h:
2930         * wtf/Noncopyable.h:
2931         * wtf/OwnArrayPtr.h:
2932         * wtf/OwnPtr.h:
2933         * wtf/PassRefPtr.h:
2934         * wtf/RefPtr.h:
2935         * wtf/TCSystemAlloc.cpp:
2936         (TCMalloc_SystemAlloc):
2937         * wtf/Vector.h:
2938         * wtf/VectorTraits.h:
2939         * wtf/unicode/UnicodeCategory.h:
2940         * wtf/unicode/icu/UnicodeIcu.h:
2941
2942 2006-05-08  Timothy Hatcher  <timothy@apple.com>
2943
2944         Reviewed by Tim O.
2945
2946         * bindings/npapi.h: do not define #pragma options align=mac68k if we are 64-bit
2947
2948 2006-05-07  Darin Adler  <darin@apple.com>
2949
2950         Reviewed and landed by Maciej.
2951
2952         - fix http://bugzilla.opendarwin.org/show_bug.cgi?id=8765
2953         Random crashes on TOT since the form state change
2954
2955         I haven't figured out how to construct a test for this, but this does seem to fix the
2956         problem; Mitz mentioned that a double-destroy was occurring in these functions.
2957
2958         * kxmlcore/HashMap.h: (KXMLCore::HashMap::remove): Use RefCounter::deref instead of calling
2959         ~ValueType, because ~ValueType often results in a double-destroy, since the HashTable also
2960         destroys the element based on the storage type. The RefCounter template correctly does work
2961         only in cases where ValueType and ValueStorageType differ and this class is what's used
2962         elsewhere for the same purpose; I somehow missed this case when optimizing HashMap.
2963         * kxmlcore/HashSet.h: (KXMLCore::HashSet::remove): Ditto.
2964
2965 2006-05-05  Darin Adler  <darin@apple.com>
2966
2967         - http://bugzilla.opendarwin.org/show_bug.cgi?id=8722
2968           IE compatibility fix in date parsing
2969
2970         * kjs/date_object.cpp: (KJS::parseDate): Merged change that George Staikos provided
2971         from KDE 3.4.3 branch that allows day values of 0 and values that are > 1000.
2972
2973 2006-05-04  Anders Carlsson  <andersca@mac.com>
2974
2975         Reviewed by Maciej.
2976
2977         http://bugzilla.opendarwin.org/show_bug.cgi?id=8734
2978         Would like a Vector::append that takes another Vector
2979
2980         * kxmlcore/Vector.h:
2981         (KXMLCore::::append):
2982         New function that takes another array.
2983
2984 2006-05-02  Steve Falkenburg  <sfalken@apple.com>
2985
2986         Reviewed by eric.
2987
2988         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: set NDEBUG for release build
2989         * kxmlcore/FastMalloc.cpp: Prevent USE_SYSTEM_MALLOC from being defined twice
2990
2991 2006-05-02  Anders Carlsson  <andersca@mac.com>
2992
2993         Reviewed by Maciej.
2994
2995         * kxmlcore/HashMap.h:
2996         (KXMLCore::::operator):
2997         Return *this
2998
2999 2006-05-01  Tim Omernick  <timo@apple.com>
3000
3001         Reviewed by Tim Hatcher.
3002
3003         <rdar://problem/4476875> Support printing for embedded Netscape plugins
3004
3005         * bindings/npapi.h:
3006         Fixed struct alignment problem in our npapi.h.  Structs must be 68k-aligned on both pre-Mac OS X
3007         and Mac OS X systems, as this is what plugins expect.
3008
3009 2006-05-01  Timothy Hatcher  <timothy@apple.com>
3010
3011         Reviewed by Maciej.
3012
3013         <rdar://problem/4308243> 8F36 Regression: crash in malloc_consolidate if you use a .PAC file
3014
3015         The original fix missed the oversized cell case. Added a test for "currentThreadIsMainThread ||
3016         imp->m_destructorIsThreadSafe" where we collect oversized cells.
3017
3018         We don't have a way to test PAC files yet, so there's no test attached.
3019
3020         * kjs/collector.cpp:
3021         (KJS::Collector::collect): test the thread when we collect oversized cells
3022
3023 2006-05-01  Tim Omernick  <timo@apple.com>
3024
3025         Reviewed by Adele.
3026
3027         <rdar://problem/4526114> REGRESSION (two days ago): LOG() just prints @ for NSObject substitutions
3028
3029         * kxmlcore/Assertions.cpp:
3030         Changed sense of strstr("%@") check.  I already made the same fix to the WebBrowser assertions.
3031
3032 2006-04-28  Steve Falkenburg  <sfalken@apple.com>
3033
3034         Reviewed by kdecker
3035
3036         Actually apply the change that was reviewed insted of checking it in with an #if 0 (oops).
3037
3038         * kjs/testkjs.cpp:
3039         (main): Suppress C runtime alerts
3040
3041 2006-04-28  Steve Falkenburg  <sfalken@apple.com>
3042
3043         Reviewed by kdecker
3044
3045         Suppress error reporting dialog that blocks Javascript tests from completing.
3046
3047         Real error is due to an overflow in the date/time handling functions that needs
3048         to be addressed, but this will prevent the hang running the Javascript tests
3049         on the build bot (along with the related changes).
3050
3051         * kjs/testkjs.cpp:
3052         (main): Suppress C runtime alerts
3053
3054 2006-04-27  Geoffrey Garen  <ggaren@apple.com>
3055
3056         Reviewed by Maciej
3057
3058         - Minor fixups I discovered while working on the autogenerator.
3059
3060         * kjs/lookup.cpp:
3061         (findEntry): ASSERT that size is not 0, because otherwise we'll % by 0,
3062         compute a garbage address, and possibly crash.
3063         * kjs/lookup.h:
3064         (cacheGlobalObject): Don't enumerate cached objects -- ideally, they
3065         would be hidden entirely.
3066
3067 2006-04-21  Kevin M. Ollivier  <kevino@theolliviers.com>
3068
3069         Reviewed by Darin.
3070
3071         - http://bugzilla.opendarwin.org/show_bug.cgi?id=8507
3072           Compilation fixes for building on gcc 4.0.2, and without precomp headers
3073
3074         * kjs/operations.h:
3075         * kxmlcore/Assertions.cpp:
3076         * kxmlcore/FastMalloc.cpp:
3077         Added necessary headers to resolve compilation issues when not using
3078         precompiled headers.
3079
3080         * kjs/value.h: Declare the JSCell class before friend declaration
3081         to resolve compilation issues with gcc 4.0.2.
3082
3083         * kxmlcore/Platform.h: Set Unicode support to use ICU on platforms
3084         other than KDE (previously only defined for Win and Mac OS)
3085
3086 2006-04-18  Eric Seidel  <eseidel@apple.com>
3087
3088         Reviewed by ggaren.
3089
3090         Fix "new Function()" to correctly use lexical scoping.
3091         Add ScopeChain::print() function for debugging.
3092         <rdar://problem/4067864> REGRESSION (125-407): JavaScript failure on PeopleSoft REN Server
3093
3094         * kjs/function_object.cpp:
3095         (FunctionObjectImp::construct):
3096         * kjs/scope_chain.cpp:
3097         (KJS::ScopeChain::print):
3098         * kjs/scope_chain.h:
3099
3100 2006-04-14  James G. Speth  <speth@end.com>
3101
3102         Reviewed by Timothy.
3103
3104         Bug 8389: support for Cocoa bindings - binding an NSTreeController to the WebView's DOM
3105         http://bugzilla.opendarwin.org/show_bug.cgi?id=8389
3106
3107         Adds a category to WebScriptObject with array accessors for KVC/KVO.
3108
3109         If super valueForKey: fails it will call valueForUndefinedKey:, which is
3110         important because it causes the right behavior to happen with bindings using
3111         the "Raises for Not Applicable Keys" flag and the "Not Applicable Placeholder"
3112
3113         * bindings/objc/WebScriptObject.mm:
3114         (-[WebScriptObject valueForKey:]):
3115         (-[WebScriptObject count]):
3116         (-[WebScriptObject objectAtIndex:]):
3117         (-[WebUndefined description]): return "undefined"
3118
3119 2006-04-13  Geoffrey Garen  <ggaren@apple.com>
3120
3121         Reviewed by Darin.
3122
3123         * kjs/internal.cpp:
3124         (KJS::InterpreterImp::initGlobalObject): Add the built-in object
3125         prototype to the end of the global object's prototype chain instead of
3126         just blowing away its existing prototype. We need to do this because
3127         the window object has a meaningful prototype now.
3128
3129 2006-04-13  Maciej Stachowiak  <mjs@apple.com>
3130
3131         Reviewed by Geoff.
3132
3133         - fix testkjs to not show false-positive KJS::Node leaks in debug builds
3134
3135         * kjs/testkjs.cpp:
3136         (doIt):
3137         (kjsmain):
3138
3139 2006-04-11  Geoffrey Garen  <ggaren@apple.com>
3140
3141         Reviewed by Maciej.
3142
3143         Minor code cleanup -- passes all the JS tests.
3144
3145         * kjs/object_object.cpp:
3146         (ObjectObjectImp::construct):
3147         (ObjectObjectImp::callAsFunction):
3148
3149 2006-04-11  Darin Adler  <darin@apple.com>
3150
3151         - another attempt to fix Windows build -- Vector in Forward.h was not working
3152
3153         * kxmlcore/Forward.h: Remove Vector.
3154         * kxmlcore/Vector.h: Add back default arguments, remove include of
3155         Forward.h.
3156
3157 2006-04-11  Darin Adler  <darin@apple.com>
3158
3159         - try to fix Windows build -- HashForward.h was not working
3160
3161         * kxmlcore/HashForward.h: Removed.
3162
3163         * JavaScriptCore.xcodeproj/project.pbxproj: Remove HashForward.h.
3164         * kjs/collector.h: Remove use of HashForward.h.
3165         * kxmlcore/HashCountedSet.h: Remove include of HashForward.h, restore
3166         default arguments.
3167         * kxmlcore/HashMap.h: Ditto.
3168         * kxmlcore/HashSet.h: Ditto.
3169
3170 2006-04-11  David Harrison  <harrison@apple.com>
3171
3172         Reviewed by Darin.
3173
3174         - fixed clean build, broken by Darin's check-in
3175
3176         * kjs/date_object.cpp: Add needed include of lookup.h.
3177         * kjs/regexp_object.cpp: Move include of .lut.h file below other includes.
3178
3179 2006-04-10  Darin Adler  <darin@apple.com>
3180
3181         Rubber-stamped by John Sullivan.
3182
3183         - switched from a shell script to a makefile for generated files
3184         - removed lots of unneeded includes
3185         - added new Forward.h and HashForward.h headers that allow compiling with
3186           fewer unneeded templates
3187
3188         * DerivedSources.make: Added.
3189         * generate-derived-sources: Removed.
3190         * JavaScriptCore.xcodeproj/project.pbxproj: Added new files, changed to use
3191         DerivedSources.make.
3192
3193         * kxmlcore/Forward.h: Added.
3194         * kxmlcore/HashForward.h: Added.
3195
3196         * kxmlcore/HashCountedSet.h: Include HashForward for default args.
3197         * kxmlcore/HashMap.h: Ditto.
3198         * kxmlcore/HashSet.h: Ditto.
3199
3200         * kjs/object.h:
3201         * kjs/object.cpp:
3202         Moved KJS_MAX_STACK into the .cpp file.
3203
3204         * bindings/NP_jsobject.cpp:
3205         * bindings/c/c_instance.h:
3206         * bindings/jni/jni_class.h:
3207         * bindings/jni/jni_runtime.h:
3208         * bindings/jni/jni_utility.h:
3209         * bindings/objc/WebScriptObject.mm:
3210         * bindings/objc/WebScriptObjectPrivate.h:
3211         * bindings/objc/objc_class.h:
3212         * bindings/objc/objc_class.mm:
3213         * bindings/objc/objc_instance.h:
3214         * bindings/objc/objc_instance.mm:
3215         * bindings/objc/objc_runtime.mm:
3216         * bindings/objc/objc_utility.mm:
3217         * bindings/runtime.h:
3218         * bindings/runtime_array.cpp:
3219         * bindings/runtime_array.h:
3220         * bindings/runtime_method.cpp:
3221         * bindings/runtime_method.h:
3222         * bindings/runtime_object.cpp:
3223         * bindings/runtime_root.h:
3224         * kjs/JSImmediate.cpp:
3225         * kjs/Parser.h:
3226         * kjs/array_object.cpp:
3227         * kjs/array_object.h:
3228         * kjs/bool_object.cpp:
3229         * kjs/bool_object.h:
3230         * kjs/collector.h:
3231         * kjs/context.h:
3232         * kjs/debugger.cpp:
3233         * kjs/error_object.h:
3234         * kjs/function_object.h:
3235         * kjs/internal.h:
3236         * kjs/lexer.cpp:
3237         * kjs/math_object.cpp:
3238         * kjs/math_object.h:
3239         * kjs/nodes.cpp:
3240         * kjs/nodes.h:
3241         * kjs/number_object.cpp:
3242         * kjs/number_object.h:
3243         * kjs/object_object.cpp:
3244         * kjs/operations.cpp:
3245         * kjs/protected_reference.h:
3246         * kjs/reference.h:
3247         * kjs/reference_list.h:
3248         * kjs/regexp_object.h:
3249         * kjs/string_object.cpp:
3250         * kjs/string_object.h:
3251         * kjs/testkjs.cpp:
3252         * kjs/value.cpp:
3253         * kjs/value.h:
3254         * kxmlcore/HashTable.h:
3255         * kxmlcore/ListRefPtr.h:
3256         * kxmlcore/TCPageMap.h:
3257         * kxmlcore/Vector.h:
3258         Removed unneeded header includes.
3259
3260 2006-04-09  Geoffrey Garen  <ggaren@apple.com>
3261
3262         Reviewed by eric.
3263
3264         - Fixed http://bugzilla.opendarwin.org/show_bug.cgi?id=8284
3265         prevent unnecessary entries in the "nodes with extra refs" hash table
3266
3267         This patch switches manually RefPtr exchange with use of
3268         RefPtr::release to ensure that a node's ref count never tops 1
3269         (in the normal case).
3270
3271         * kjs/nodes.cpp:
3272         (BlockNode::BlockNode):
3273         (CaseBlockNode::CaseBlockNode):
3274         * kjs/nodes.h:
3275         (KJS::ArrayNode::ArrayNode):
3276         (KJS::ObjectLiteralNode::ObjectLiteralNode):
3277         (KJS::ArgumentsNode::ArgumentsNode):
3278         (KJS::VarStatementNode::VarStatementNode):
3279         (KJS::ForNode::ForNode):
3280         (KJS::CaseClauseNode::CaseClauseNode):
3281         (KJS::FuncExprNode::FuncExprNode):
3282         (KJS::FuncDeclNode::FuncDeclNode):
3283
3284 2006-04-08  Alexey Proskuryakov  <ap@nypop.com>
3285
3286         Reviewed by Darin.
3287
3288         One more attempt - use reinterpret_cast, rather than static_cast.
3289
3290 2006-04-08  Alexey Proskuryakov  <ap@nypop.com>
3291
3292         Reviewed by Darin.
3293
3294         An attempt to fix Win32 build - ICU uses wchar_t on Windows, so we need a type cast.
3295
3296         * kxmlcore/unicode/icu/UnicodeIcu.h:
3297         (KXMLCore::Unicode::toLower):
3298         (KXMLCore::Unicode::toUpper):
3299
3300 2006-04-08  Alexey Proskuryakov  <ap@nypop.com>
3301
3302         Reviewed by Darin.
3303
3304         - fix http://bugzilla.opendarwin.org/show_bug.cgi?id=8264
3305         toLowerCase and toUpperCase don't honor special mappings
3306
3307         Test: fast/js/string-capitalization.html
3308
3309         * JavaScriptCore.xcodeproj/project.pbxproj: Added KXMLCore::Unicode headers to the project.
3310         * icu/unicode/putil.h: Added (copied from WebCore).
3311         * icu/unicode/uiter.h: Ditto.
3312         * icu/unicode/ustring.h: Ditto.
3313         * kjs/string_object.cpp:
3314         (StringProtoFunc::callAsFunction): Use the new KXMLCore::Unicode::toUpper() and toLower().
3315         * kjs/ustring.cpp: Removed unused (and evil) UChar::toLower() and toUpper().
3316         * kjs/ustring.h: Ditto.
3317
3318         * kxmlcore/unicode/Unicode.h: Corrected capitalization of the word Unicode.
3319         * kxmlcore/unicode/UnicodeCategory.h: Renamed include guard macro to match file name.
3320
3321         * kxmlcore/unicode/icu/UnicodeIcu.h:
3322         (KXMLCore::Unicode::toLower): Work on strings, not individual characters. Use ICU root locale.
3323         (KXMLCore::Unicode::toUpper): Ditto.
3324         (KXMLCore::Unicode::isFormatChar): Use int32_t, which can hold a complete code point.
3325         (KXMLCore::Unicode::isSeparatorSpace): Ditto.
3326         (KXMLCore::Unicode::category): Ditto.
3327         * kxmlcore/unicode/qt4/UnicodeQt4.h:
3328         (KXMLCore::Unicode::toLower): Work on strings, not individual characters.
3329         (KXMLCore::Unicode::toUpper): Ditto.
3330         (KXMLCore::Unicode::isFormatChar): Use int32_t, which can hold a complete code point.
3331         (KXMLCore::Unicode::isSeparatorSpace): Ditto.
3332         (KXMLCore::Unicode::category): Ditto.
3333
3334         * tests/mozilla/ecma/String/15.5.4.12-1.js: Corrected expected results.
3335         * tests/mozilla/ecma/String/15.5.4.12-5.js: Corrected expected results.
3336
3337 2006-04-05  Darin Adler  <darin@apple.com>
3338
3339         - attempt to fix Windows build
3340
3341         * kxmlcore/HashMap.h: (KXMLCore::HashMap::remove): Use (*it). instead of it->.
3342         * kxmlcore/HashSet.h: (KXMLCore::HashSet::remove): Ditto.
3343
3344 2006-04-05  Darin Adler  <darin@apple.com>
3345
3346         - attempt to fix Windows build
3347
3348         * os-win32/stdint.h: Add int8_t, uint8_t, int64_t.
3349
3350 2006-04-05  Darin Adler  <darin@apple.com>
3351
3352         Reviewed by Maciej.
3353
3354         - fix memory leak introduced by the previous change
3355
3356         * kxmlcore/HashTable.h: Specialize NeedsRef so that it correctly returns true when
3357         the value in question is a pair where one of the pair needs a ref and the other
3358         of the pair does not.
3359
3360 2006-04-05  Darin Adler  <darin@apple.com>
3361
3362         Reviewed by Maciej.
3363
3364         - JavaScriptCore part of fix for http://bugzilla.opendarwin.org/show_bug.cgi?id=8049
3365           StringImpl hash traits deleted value creates an init routine for WebCore
3366           <rdar://problem/4442248> REGRESSION: WebCore has init routines (8049)
3367
3368         Change HashMap and HashSet implementation so they fold various types together.
3369         This allows us to implement maps and sets that use RefPtr<WebCore::StringImpl>
3370         and WebCore::String in terms of the underlying raw pointer type, and hence use
3371         -1 for the deleted value.
3372
3373         * kxmlcore/HashTraits.h: Added a new type to HashTraits, StorageTraits, which is a
3374         type to be used when storing a value that has the same layout as the type itself.
3375         This is used only for non-key cases. In the case of keys, the hash function must also
3376         be considered. Moved emptyValue out of GenericHashTraitsBase into GenericHashTraits.
3377         Added a new bool to HashTraits, needsRef, which indicates whether the type needs
3378         explicit reference counting. If the type itself has needsRef true, but the storage
3379         type has needsRef false, then the HashSet or HashMap has to handle the reference
3380         counting explicitly. Added hash trait specializations for all signed integer values
3381         that give -1 as the deleted value. Gave all integers StorageTraits of the canonical
3382         integer type of the same size so int and long will share code. Gave all pointers and
3383         RefPtrs StorageTraits of the appropriately sized integer type. Removed redundant
3384         TraitType and emptyValue definitions in the pointer specialization for HashTraits.
3385         Added PairBaseHashTraits, which doesn't try to set up needsDestruction and deletedValue.
3386         Useful for types where we don't want to force the existence of deletedValue, such as
3387         the type of a pair in a HashMap which is not the actual storage type. Removed an
3388         unneeded parameter from the DeletedValueAssigner template. Added HashKeyStorageTraits
3389         template, which determines what type can be used to store a given hash key type with
3390         a given hash function, and specialized it for pointers and RefPtr so that pointer
3391         hash tables share an underlying HashTable that uses IntHash.
3392
3393         * kxmlcore/HashTable.h: Added HashTableConstIteratorAdapter, HashTableIteratorAdapter,
3394         NeedsRef, RefCountManagerBase, RefCountManager, HashTableRefCountManagerBase, and
3395         HashTableRefCountManager. All are used by both HashSet and HashMap to handle hash
3396         tables where the type stored is not the same as the real value type.
3397
3398         * kxmlcore/HashFunctions.h: Added a new struct named IntTypes that finds an
3399         integer type given a sizeof value. Renamed pointerHash to intHash and made it
3400         use overloading and take integer parameters. Added an IntHash struct which is
3401         a hash function that works for integers. Changed PtrHash to call IntHash with
3402         an appropriately sized integer. Made IntHash the default hash function for
3403         many integer types. Made PtrHash the default hash function for RefPtr as well
3404         as for raw pointers.
3405
3406         * kxmlcore/HashSet.h: Changed implementation to use a separate "storage type"
3407         derived from the new traits. The HashTable will use the storage type and all
3408         necessary translation and ref/deref is done at the HashSet level. Also reorganized
3409         the file so that the HashSet is at the top and has no inline implementation inside
3410         it so it's easy to read the interface to HashSet.
3411
3412         * kxmlcore/HashMap.h: Changed implementation to use a separate "storage type"
3413         derived from the new traits. The HashTable will use the storage type and all
3414         necessary translation and ref/deref is done at the HashMap level. Also reorganized
3415         the file so that the HashMap is at the top and has no inline implementation inside
3416         it so it's easy to read the interface to HashMap.
3417
3418         * kxmlcore/HashMapPtrSpec.h: Removed. Superceded by optimizations in HashMap itself.
3419
3420         * JavaScriptCore.xcodeproj/project.pbxproj: Remove HashMapPtrSpec.h, resort files,
3421         and also remove some unnecessary build settings from the aggregate target that
3422         generates derived sources.
3423         * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: Ditto.
3424
3425 2006-04-04  Timothy Hatcher  <timothy@apple.com>
3426
3427         Reviewed by Darin.
3428
3429         The Debug and Release frameworks are now built with install paths relative to the build products directory.
3430         This removes the need for other projects to build with -framework WebCore and -framework JavaScriptCore.
3431
3432         * JavaScriptCore.xcodeproj/project.pbxproj:
3433
3434 2006-04-04  Eric Seidel  <eseidel@apple.com>
3435
3436         Reviewed by ggaren.
3437
3438         Fix win32 build.
3439         Disable ASSERT redefinition warnings for now.
3440
3441         * JavaScriptCore.vcproj/testkjs/testkjs.vcproj:
3442         * kxmlcore/Assertions.h:
3443
3444 2006-04-04  Bjrn Graf  <bjoern.graf@gmail.com>
3445
3446         Reviewed by ggaren & darin.  Landed by eseidel.
3447
3448         Integrate CURL version of gettimeofday
3449         http://bugzilla.opendarwin.org/show_bug.cgi?id=7399
3450         Disable crash report dialogs for testkjs.exe in Release mode
3451         http://bugzilla.opendarwin.org/show_bug.cgi?id=8113
3452
3453         * kjs/testkjs.cpp:
3454         (StopWatch::start):
3455         (StopWatch::stop):
3456         (StopWatch::getElapsedMS):
3457         (main):
3458         (kjsmain):
3459
3460 2006-04-04  Eric Seidel  <eseidel@apple.com>
3461
3462         Reviewed by mjs.
3463
3464         * kjs/number_object.cpp:
3465         (NumberProtoFunc::callAsFunction): remove trunc() to fix win32.
3466
3467 2006-03-12  Maciej Stachowiak  <mjs@apple.com>
3468
3469         Reviewed by Darin.
3470
3471         - fixed "toPrecision sometimes messes up the last digit on intel Macs"
3472         http://bugzilla.opendarwin.org/show_bug.cgi?id=7748
3473
3474         * kjs/number_object.cpp:
3475         (intPow10): Compute integer powers of 10 using exponentiation by squaring.
3476         (NumberProtoFunc::callAsFunction): Use intPow10(n) in place of all pow(10.0, n),
3477         plus a bit of refactoring.
3478
3479 2006-04-03  Darin Adler  <darin@apple.com>
3480
3481         - tweak config.h and Platform.h to try to get buildbot working
3482           (making some small changes at the same time)
3483
3484         * kjs/config.h: Removed now-unneeded HAVE_ICU.
3485         * kxmlcore/Platform.h: Tweak how platform gets set up. Move all the
3486         USE stuff to the end.
3487
3488 2006-04-03  George Staikos   <staikos@opensource.apple.com>
3489
3490         Reviewed by Maciej.
3491
3492         Fix Win32 build breakage from previous commit, remove unused forward.
3493
3494 2006-04-03  George Staikos   <staikos@opensource.apple.com>
3495
3496         Reviewed by Maciej.
3497
3498         Implement a unicode abstraction layer to make JavaScriptCore much more
3499         easily ported to other platforms without having to take in libicu.  Also
3500         makes the unicode related code easier to understand.
3501
3502 2006-04-03  Timothy Hatcher  <timothy@apple.com>
3503
3504         Reviewed by Adele.
3505
3506         Fixes <rdar://problem/4498338> JavaScriptCore fails to compile for ppc64
3507         Other 64 bit build fixes.
3508
3509         * kjs/collector.cpp:
3510         (KJS::Collector::markOtherThreadConservatively): test for __DARWIN_UNIX03 and use __r1
3511         * kjs/dtoa.cpp:
3512         (Bigint::): cast PRIVATE_mem to unsigned to prevent warning
3513         * bindings/jni/jni_utility.cpp:
3514         (KJS::Bindings::getJavaVM): cast jniError to long to prevent format warning
3515         (KJS::Bindings::getJNIEnv): cast jniError to long to prevent format warning
3516         * bindings/runtime_root.cpp:
3517         (KJS::Bindings::addNativeReference): cast CFDictionaryGetValue to unsigned long to prevent warning
3518         (KJS::Bindings::removeNativeReference): cast CFDictionaryGetValue to unsigned long to prevent warning
3519
3520 2006-03-31  Darin Adler  <darin@apple.com>
3521
3522         Reviewed by Geoff.
3523
3524         - <rdar://problem/4395622> API: WebScriptObject.h incorrectly reports that -isSelectorExcludedFromWebScript returns NO by default
3525