Roll over ChangeLogs for JavaScriptCore, LayoutTests, WebCore and WebKit.
authorbdash <bdash@268f45cc-cd09-0410-ab3c-d52691b4dbfc>
Sun, 14 Oct 2007 11:02:41 +0000 (11:02 +0000)
committerbdash <bdash@268f45cc-cd09-0410-ab3c-d52691b4dbfc>
Sun, 14 Oct 2007 11:02:41 +0000 (11:02 +0000)
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26574 268f45cc-cd09-0410-ab3c-d52691b4dbfc

JavaScriptCore/ChangeLog
JavaScriptCore/ChangeLog-2007-10-14 [new file with mode: 0644]
LayoutTests/ChangeLog
LayoutTests/ChangeLog-2007-10-14 [new file with mode: 0644]
WebCore/ChangeLog
WebCore/ChangeLog-2007-10-14 [new file with mode: 0644]
WebKit/ChangeLog
WebKit/ChangeLog-2007-10-14 [new file with mode: 0644]

index c44a39c..c039b41 100644 (file)
@@ -1,26221 +1 @@
-=== Start merge of feature-branch 2007-10-12 ===
-
-2007-10-11  Andrew Wellington  <proton@wiretapped.net>
-
-        Reviewed by Eric Seidel.
-        
-        Fix for http://bugs.webkit.org/show_bug.cgi?id=15076
-        "deg2rad has multiple definitions"
-        
-        Define deg2rad, rad2deg, deg2grad, grad2deg, rad2grad, grad2rad
-        These are used through WebKit.
-        
-        Change based on original patch by Rob Buis.
-
-        * wtf/MathExtras.h:
-        (deg2rad):
-        (rad2deg):
-        (deg2grad):
-        (grad2deg):
-        (rad2grad):
-        (grad2rad):
-
-2007-10-10  Maciej Stachowiak  <mjs@apple.com>
-
-        Reviewed by Eric.
-        
-        - fix assertion failures on quit.
-
-        * kjs/array_object.cpp:
-        (ArrayProtoFunc::callAsFunction): Dynamically alocate function-scope static
-        UStrings to avoid the static destructor getting called later.
-        * kjs/lookup.h: Dynamically alocate function-scope static
-        Identifiers to avoid the static destructor getting called later.
-
-2007-10-07  Ed Schouten  <ed@fxq.nl>
-
-        Reviewed and landed by Alexey Proskuryakov.
-
-        Add PLATFORM(FREEBSD), so we can fix the build on FreeBSD-like
-        systems by including <pthread_np.h>. Also fix some (disabled)
-        regcomp()/regexec() code; it seems some variable names have
-        changed.
-
-        * kjs/config.h:
-        * kjs/regexp.cpp:
-        (KJS::RegExp::RegExp):
-        * wtf/Platform.h:
-
-2007-10-02  Alexey Proskuryakov  <ap@webkit.org>
-
-        Reviewed by Darin.
-
-        http://bugs.webkit.org/show_bug.cgi?id=10370
-        RegExp fails to match non-ASCII characters against [\S\s]
-
-        Test: fast/js/regexp-negative-special-characters.html
-
-        * pcre/pcre_compile.c:
-        (compile_branch): Adjust opcode and bitmap as necessary to include (or exclude)
-        character codes >255. Fix suggested by Philip Hazel.
-
-        * pcre/pcre_exec.c:
-        (match): Merged fix for PCRE bug 580 (\S\S vs. \S{2}).
-
-        * tests/mozilla/expected.html: One test was fixed.
-        * pcre/MERGING: Added information about this fix.
-
-2007-10-02  Maciej Stachowiak  <mjs@apple.com>
-
-        Reviewed by Oliver.
-        
-        - skip extra hash lookup and avoid converting char* to UString for 19% speedup on CK JS array test 
-        http://bugs.webkit.org/show_bug.cgi?id=15350
-
-        * kjs/array_object.cpp:
-        (ArrayProtoFunc::callAsFunction): Implement the two mentioned optimizations.
-
-2007-10-02  Maciej Stachowiak  <mjs@apple.com>
-
-        Reviewed by Mark.
-        
-        - Efficiently handle regexp property identifiers for 19% speedup on Celtic Kane regexp test
-        http://bugs.webkit.org/show_bug.cgi?id=15337
-
-        * kjs/CommonIdentifiers.h:
-        * kjs/regexp_object.cpp:
-        (RegExpProtoFunc::callAsFunction):
-        (RegExpObjectImp::arrayOfMatches):
-        (RegExpObjectImp::construct):
-
-2007-10-02  Maciej Stachowiak  <mjs@apple.com>
-
-        Reviewed by Mark.
-        
-        - Cache global prorotypes more efficiently for 10% speedup on CK AJAX benchmark
-        http://bugs.webkit.org/show_bug.cgi?id=15335
-
-        * kjs/lookup.h:
-
-2007-10-01  Oliver Hunt  <oliver@apple.com>
-
-        Reviewed by Mark.
-
-        Enable Experimental SVG features by default when building from Xcode
-
-        * Configurations/JavaScriptCore.xcconfig:
-
-2007-09-29  Rob Buis  <buis@kde.org>
-
-        Reviewed by Adam.
-
-        http://bugs.webkit.org/show_bug.cgi?id=13472
-        Misparsing date in javascript leads to year value of -1
-        http://bugs.webkit.org/show_bug.cgi?id=14176
-        Some date values not handled consistently with IE/Firefox
-
-        Allow an optional comma between month and year, and year and time.
-
-        * kjs/date_object.cpp:
-        (KJS::parseDate):
-
-2007-07-11  Nikolas Zimmermann  <zimmermann@kde.org>
-
-        Reviewed by Mark.
-
-        Forwardport the hash table fix from CodeGeneratorJS.pm to create_hash_table.
-        Reran run-jsc-tests, couldn't find any regressions. Suggested by Darin.
-
-        * kjs/create_hash_table:
-
-2007-06-25  Antti Koivisto  <antti@apple.com>
-
-        Reviewed by Maciej.
-        
-        Use intHash to hash floats and doubles too.
-
-        * ChangeLog:
-        * wtf/HashFunctions.h:
-        (WTF::FloatHash::hash):
-        (WTF::FloatHash::equal):
-        (WTF::):
-        * wtf/HashTraits.h:
-        (WTF::FloatHashTraits::emptyValue):
-        (WTF::FloatHashTraits::deletedValue):
-        (WTF::):
-
-=== End merge of feature-branch 2007-10-12 ===
-
-2007-10-11  Mark Rowe  <mrowe@apple.com>
-
-        Reviewed by Tim Hatcher.
-
-        Fix for <rdar://problem/5488678>.  Disable debugging symbols in production builds for 10.4
-        PowerPC to prevent a huge STABS section from being generated.
-
-        * Configurations/Base.xcconfig:
-
-2007-10-08  George Staikos  <staikos@kde.org>
-
-        Reviewed by Adam Roben.
-
-        Fix Qt build on Win32.
-
-        * kjs/testkjs.cpp:
-        (main):
-
-2007-10-10  Simon Hausmann  <hausmann@kde.org>
-
-        Reviewed by Lars.
-
-        Fix compilation using gcc 4.3. Header files have been reorganized and as a result some extra
-        includes are needed for INT_MAX, std::auto_ptr and the like.
-
-        * kjs/collector.cpp:
-        * kjs/collector.h:
-        * kjs/lexer.cpp:
-        * kjs/scope_chain.cpp:
-        * kjs/ustring.cpp:
-        * wtf/Vector.h:
-
-2007-10-09  Lars Knoll  <lars@trolltech.com>
-
-        Reviewed by Simon.
-
-        fix the invokation of slots with return types. Add a JSLock around the conversion from QVariant to JSValue.
-
-        * bindings/qt/qt_instance.cpp:
-        (KJS::Bindings::QtInstance::invokeMethod):
-        * bindings/qt/qt_runtime.cpp:
-        (KJS::Bindings::convertValueToQVariant):
-        (KJS::Bindings::convertQVariantToValue):
-
-2007-10-05  Geoffrey Garen  <ggaren@apple.com>
-
-        Reviewed by Sam Weinig.
-        
-        Added JSObject::removeDirect, to support the fix for 
-        <rdar://problem/5522487> REGRESSION: With JavaScript disabled, any 
-        page load causes a crash in PropertyMap::put
-
-        * kjs/object.cpp:
-        (KJS::JSObject::removeDirect):
-        * kjs/object.h:
-
-2007-10-04  Mark Rowe  <mrowe@apple.com>
-
-        Reviewed by Oliver.
-
-        Switch to default level of debugging symbols to resolve <rdar://problem/5488678>.
-        The "full" level appears to offer no observable benefits even though the documentation
-        suggests it be used for dead code stripping.  This should also decrease link times.
-
-        * Configurations/Base.xcconfig:
-
-2007-10-03  Lars Knoll  <lars@trolltech.com>
-
-        Reviewed by Rob.
-
-        Fix a stupid bug in Unicode::toUpper/toLower.
-        Fixes all three test failures in the JavaScriptCore test
-        suite.
-
-        * wtf/unicode/qt4/UnicodeQt4.h:
-        (WTF::Unicode::toLower):
-        (WTF::Unicode::toUpper):
-
-2007-10-02  Darin Adler  <darin@apple.com>
-
-        Reviewed by Adam.
-
-        - add support for GDI objects to OwnPtr; I plan to use this
-          to fix some GDI handle leaks
-
-        * kjs/grammar.y: Change parser to avoid macros that conflict
-        with macros defined in Windows system headers: THIS, DELETE,
-        VOID, IN, and CONST. This is needed because OwnPtr.h will now
-        include <windows.h>.
-        * kjs/keywords.table: Ditto.
-
-        * wtf/OwnPtr.h: For PLATFORM(WIN), add support so that OwnPtr can be
-        a GDI handle, and it will call DeleteObject. Also change to use the
-        RemovePointer technique used by RetainPtr, so you can say OwnPtr<HBITMAP>
-        rather than having to pass in the type pointed to by HBITMAP.
-
-        * wtf/OwnPtrWin.cpp: Added.
-        (WebCore::deleteOwnedPtr): Put this in a separate file so that we
-        don't have to include <windows.h> in OwnPtr.h.
-
-        * JavaScriptCore.vcproj/WTF/WTF.vcproj: Added OwnPtrWin.cpp.
-
-2007-09-29  Holger Hans Peter Freyther  <zecke@selfish.org>
-
-        Reviewed by Mark.
-
-        -Fix http://bugs.webkit.org/show_bug.cgi?id=13226.
-        Remove Bakefiles from svn.
-
-        * JavaScriptCoreSources.bkl: Removed.
-        * jscore.bkl: Removed.
-
-2007-09-27  Kevin Decker  <kdecker@apple.com>
-
-        Rubber stamped by John Sullivan.
-
-        <rdar://problem/5493093>
-        
-        * JavaScriptCore.order: Added.
-        * JavaScriptCore.xcodeproj/project.pbxproj: We're changing from using an order file built by
-         another team to using one we actually check into our project repository. Linker settings for
-         Symbol Ordering Flags have been updated accordingly.
-
-2007-09-26  Adam Roben  <aroben@apple.com>
-
-        Make testkjs delay-load WebKit.dll so WebKitInitializer can work its magic
-
-        Rubberstamped by Anders.
-
-        * JavaScriptCore.vcproj/testkjs/testkjs.vcproj:
-
-2007-09-25  Adam Roben  <aroben@apple.com>
-
-        Make testkjs delay-load its dependencies
-
-        This lets WebKitInitializer re-route the dependencies to be loaded out
-        of the Safari installation directory.
-
-        Rubberstamped by Sam.
-
-        * JavaScriptCore.vcproj/testkjs/testkjs.vcproj:
-
-2007-09-25  David Kilzer  <ddkilzer@webkit.org>
-
-        Reviewed by Adam.
-
-        - Fix http://bugs.webkit.org/show_bug.cgi?id=14885
-          LGPL'ed files contain incorrect FSF address
-
-        * COPYING.LIB:
-        * bindings/testbindings.cpp:
-        * kjs/AllInOneFile.cpp:
-        * kjs/DateMath.cpp:
-        * kjs/PropertyNameArray.cpp:
-        * kjs/PropertyNameArray.h:
-        * kjs/config.h:
-
-2007-09-25  Sam Weinig  <sam@webkit.org>
-
-        Fix location for build products for Debug_Internal.
-
-        Reviewed by Adam Roben.
-
-        * JavaScriptCore.vcproj/testkjs/testkjs.vcproj:
-
-2007-09-25  Adam Roben  <aroben@apple.com>
-
-        Make testkjs use WebKitInitializer
-
-        Reviewed by Sam.
-
-        * JavaScriptCore.vcproj/JavaScriptCore.sln: Add WebKitInitializer and
-        make testkjs depend on it.
-        * JavaScriptCore.vcproj/testkjs/testkjs.vcproj: Link against
-        WebKitInitializer.lib.
-        * kjs/testkjs.cpp:
-        (main): Call initializeWebKit.
-
-2007-09-24  Kevin McCullough  <kmccullough@apple.com>
-
-        Reviewed by Sam.
-
-        - Continued to update project files to not use Edit and Continue for Debug Information since it doesn't work and breaks some functionality.
-
-        * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
-        * JavaScriptCore.vcproj/WTF/WTF.vcproj:
-
-2007-09-21  Kevin McCullough  <kmccullough@apple.com>
-
-        Reviewed by Sam.
-
-        - Updated project files to not use Edit and Continue for Debug Information since it doesn't work and breaks some functionality.
-
-        * JavaScriptCore.vcproj/dftables/dftables.vcproj:
-        * JavaScriptCore.vcproj/testkjs/testkjs.vcproj:
-
-2007-09-20  Holger Hans Peter Freyther  <zecke@selfish.org>
-
-        Rubber stamped by Adam.
-
-        Renamed files from *Gdk to *Gtk (see #14732) using the
-        work of Juan A. Suarez Romero as a base.
-
-        GDK -> GTK 
-
-        * JavaScriptCore.pri:
-        * kjs/testkjs.pro:
-        * pcre/dftables.pro:
-        * wtf/Platform.h: PLATFORM(GDK) to PLATFORM(GTK)
-
-2007-09-21  Mark Rowe  <mrowe@apple.com>
-
-        Reviewed by Antti Koivisto.
-
-        http://bugs.webkit.org/show_bug.cgi?id=15250
-        <rdar://problem/5496942> REGRESSION: Reproducible crash in Safari when evaluating script in Drosera console (15250)
-
-        * kjs/function.cpp:
-        (KJS::GlobalFuncImp::callAsFunction): Null-check thisObj before passing it to interpreterForGlobalObject.
-
-2007-09-19  Holger Hans Peter Freyther  <zecke@selfish.org>
-
-        Rubber stamped by Adam.
-
-        Make the guard/#if use the same name (ENABLE_FTPDIR) as the #define. This follows
-        the ENABLE_ICONDATABASE example from a couple of lines above.
-
-        * wtf/Platform.h:
-
-2007-09-19  Mark Rowe  <mrowe@apple.com>
-
-        Reviewed by Maciej.
-
-        <rdar://problem/5487107> NULL dereference crash in FastMallocZone::enumerate when running leaks against Safari
-
-        Storing remote pointers to their local equivalents in mapped memory was leading to the local pointer being
-        interpreted as a remote pointer.  This caused a crash when using the result of mapping this invalid remote pointer.
-        The fix is to follow the pattern used elsewhere in FastMallocZone by always doing the mapping after reading and
-        never storing the mapped pointer.
-
-        * wtf/FastMalloc.cpp:
-        (WTF::FastMallocZone::enumerate):
-
-2007-09-15  Darin Adler  <darin@apple.com>
-
-        - fix Mac build
-
-        * JavaScriptCore.exp: Export WTFLogVerbose.
-
-2007-09-14  Kevin McCullough  <kmccullough@apple.com>
-
-        Reviewed by Sam.
-
-        - Copy JSRetainPtr to include folder.
-
-        * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
-
-2007-09-13  Geoffrey Garen  <ggaren@apple.com>
-
-        Try to fix GDK build.
-
-        * wtf/MathExtras.h:
-        (wtf_random_init):
-
-2007-09-12  Geoff Garen  <ggaren@apple.com>
-
-        Reviewed by Sam Weinig.
-        
-        Fixed <rdar://problem/5429064> 141885 Safari JavaScript: Math.random() slightly less randomly distributed than on Safari / Mac
-
-        Math.random was skewed slightly upward because it assumed that RAND_MAX was outside the range of
-        values that rand() might return. This problem was particularly pronounced on Windows because
-        the range of values returned by rand() on Windows is 2^16 smaller than the range of values
-        return by rand() on Mac.
-        
-        Fixed by accounting for RAND_MAX return values. Also, switched Windows over to rand_s, which has
-        a range that's equal to rand()'s range on Mac.
-        
-        * kjs/config.h:
-
-        * kjs/math_object.cpp:
-        (MathFuncImp::callAsFunction): Use the new new thing.
-
-        * wtf/MathExtras.h: Platform abstraction for random numbers, to cover over differences on Windows.
-        (wtf_random_init):
-        (wtf_random):
-
-2007-09-13  Antti Koivisto  <antti@apple.com>
-
-        Reviewed by Maciej.
-        
-        Small addition to previous path to cover
-        http://bugs.webkit.org/show_bug.cgi?id=11399
-        window.eval runs in the global scope of the calling window
-        
-        Switch variable scope as well.
-
-        * kjs/function.cpp:
-        (KJS::GlobalFuncImp::callAsFunction):
-
-2007-09-12  Antti Koivisto  <antti@apple.com>
-
-        Reviewed by Geoff, Maciej.
-        
-        Fix <rdar://problem/5445058>
-        REGRESSION: Unable to upload picture to eBay auction due to domain security check
-        
-        eBay uses window.eval() between windows. In Firefox window.eval() switches execution
-        and security context to the target window, something WebKit did not do. With WebKit
-        security tightening in r24781, this broke picture uploads.
-        
-        Fix by making WebKit switch context in window.eval().
-        
-        * kjs/Context.cpp:
-        (KJS::Context::Context):
-        (KJS::Context::~Context):
-        * kjs/context.h:
-        Save and restore interpreter context independently from calling context.
-        
-        * kjs/function.cpp:
-        (KJS::GlobalFuncImp::callAsFunction):
-        If eval is called for global object different than current one, switch execution context
-        to that object and push it to scope.
-
-2007-09-12  Sam Weinig  <sam@webkit.org>
-
-        Reviewed by Geoffrey Garen.
-
-        <rdar://problem/5478717> JSStringCreateWithCFString leaks when passed a zero length CFStringRef
-
-        * API/JSStringRefCF.cpp:
-        (JSStringCreateWithCFString): Special case the zero length string and remove the
-        UTF16 optimized path since it will always leak due to the fact that we won't be 
-        able to free the backing store that the CFStringRef provides.
-
-2007-09-10  Timothy Hatcher  <timothy@apple.com>
-
-        Reviewed by Darin Adler.
-
-        <rdar://problem/5456224> CrashTracer: [USER] 2 crashes in Toast Titanium at com.apple.CoreServices.CarbonCore: CSMemDisposePtr + 37
-
-        Removed the implementation of these malloc zone functions. We do not have the ability to 
-        check if a pointer is valid or not, so we can't correctly implement them. The system free 
-        does not fail if you pass in a bad pointer.
-
-        * wtf/FastMalloc.cpp:
-        (WTF::FastMallocZone::size):
-        (WTF::FastMallocZone::zoneMalloc):
-        (WTF::FastMallocZone::zoneCalloc):
-        (WTF::FastMallocZone::zoneFree):
-        (WTF::FastMallocZone::zoneRealloc):
-
-2007-09-07  Darin Adler  <darin@apple.com>
-
-        Reviewed by Steve Falkenburg.
-
-        - fix crash seen on Windows release builds
-
-        * wtf/FastMalloc.cpp: Change pthread_getspecific optimization to be done only
-        on the DARWIN platform. Also correct a couple reinterpret_cast that should be
-        static_cast instead.
-
-2007-09-06  Kevin McCullough  <kmccullough@apple.com>
-
-        Reviewed by Maciej.
-
-        - Moved JSRetainPtr to the API.
-
-        * API/JSRetainPtr.h: Copied from kjs/JSRetainPtr.h.
-        (JSRetain):
-        (JSRelease):
-        (JSRetainPtr::JSRetainPtr):
-        (JSRetainPtr::~JSRetainPtr):
-        (JSRetainPtr::get):
-        (JSRetainPtr::releaseRef):
-        (JSRetainPtr::operator->):
-        (JSRetainPtr::operator!):
-        (JSRetainPtr::operator UnspecifiedBoolType):
-        (::operator):
-        (::adopt):
-        (::swap):
-        (swap):
-        (operator==):
-        (operator!=):
-        * JavaScriptCore.xcodeproj/project.pbxproj:
-        * kjs/JSRetainPtr.h: Removed.
-
-2007-09-05  Maciej Stachowiak  <mjs@apple.com>
-
-        Reviewed by Darin.
-        
-        - Remove single-threaded optimization for FastMalloc. 
-        
-        It does not appear to help anywhere but Mac OS X on PPC, due to
-        pthread_getspecific being slow there. On Intel, removing the
-        optimization results in a ~1.5% PLT speedup, a ~1-5% JS iBench
-        speedup, and a ~1.5% HTML iBench speedup. On PPC this change is a
-        speedup on some benchmarks, a slight hit on others.
-
-        * JavaScriptCore.xcodeproj/project.pbxproj:
-        * kjs/collector.cpp:
-        (KJS::Collector::registerThread):
-        * wtf/FastMalloc.cpp:
-        (WTF::TCMalloc_ThreadCache::GetCache):
-        (WTF::TCMalloc_ThreadCache::GetCacheIfPresent):
-        (WTF::TCMalloc_ThreadCache::CreateCacheIfNecessary):
-        (WTF::do_malloc):
-        * wtf/FastMallocInternal.h: Removed.
-
-2007-09-05  Kevin McCullough  <kmccullough@apple.com>
-
-        Reviewed by Adam, Sam, Darin.
-
-        - Created a JSRetainPtr specifically for JSStringRefs so they can be automatically refed and derefed.
-
-        * JavaScriptCore.xcodeproj/project.pbxproj:
-        * kjs/JSRetainPtr.h: Copied from wtf/RetainPtr.h.
-        (KJS::JSRetain):
-        (KJS::JSRelease):
-        (KJS::):
-        (KJS::JSRetainPtr::JSRetainPtr):
-        (KJS::JSRetainPtr::~JSRetainPtr):
-        (KJS::JSRetainPtr::get):
-        (KJS::JSRetainPtr::releaseRef):
-        (KJS::JSRetainPtr::operator->):
-        (KJS::JSRetainPtr::operator UnspecifiedBoolType):
-        (KJS::::operator):
-        (KJS::::adopt):
-        (KJS::::swap):
-        (KJS::swap):
-        (KJS::operator==):
-        (KJS::operator!=):
-
-2007-09-05  Mark Rowe  <mrowe@apple.com>
-
-        Unreviewed Qt build fix.
-
-        * wtf/unicode/qt4/UnicodeQt4.h: Fix the constness of the src argument to toUpper to prevent build failures.
-
-2007-09-04  Maciej Stachowiak  <mjs@apple.com>
-
-        Back out accidentally committed change.
-
-        * JavaScriptCore.xcodeproj/project.pbxproj:
-        * kjs/collector.cpp:
-        (KJS::Collector::registerThread):
-        * wtf/FastMalloc.cpp:
-        (WTF::fastMallocSetIsMultiThreaded):
-        (WTF::TCMalloc_ThreadCache::GetCache):
-        (WTF::TCMalloc_ThreadCache::GetCacheIfPresent):
-        (WTF::TCMalloc_ThreadCache::CreateCacheIfNecessary):
-        (WTF::do_malloc):
-        * wtf/FastMallocInternal.h: Added.
-
-2007-09-04  Maciej Stachowiak  <mjs@apple.com>
-
-        Reviewed by Darin.
-        
-        - Added Vector::appendRange(), which appends to a vector based on a given start and end iterator
-        - Added keys() and values() functions to HashMap iterators, which give keys-only and values-only iterators
-        
-        Together, these allow easy copying of a set, or the keys or values of a map, into a Vector. Examples:
-        
-        HashMap<int, int> map;
-        HashSet<int> set;
-        Vector<int> vec;
-        // ...
-        vec.appendRange(set.begin(), set.end());
-        vec.appendRange(map.begin().keys(), map.end().keys());
-        vec.appendRange(map.begin().values(), map.end().values());
-
-        This also allows for a slightly nicer syntax when iterating a map. Instead of saying 
-        (*it)->first, you can say *it.values(). Similarly for keys. Example:
-        
-        HashMap<int, int>::const_iterator end = map.end();
-        for (HashMap<int, int>::const_iterator it = map.begin(); it != end; ++it)
-        printf(" [%d => %d]", *it.keys(), *it.values());
-
-        * JavaScriptCore.xcodeproj/project.pbxproj:
-        * wtf/HashIterators.h: Added.
-        (WTF::):
-        (WTF::HashTableConstKeysIterator::HashTableConstKeysIterator):
-        (WTF::HashTableConstKeysIterator::get):
-        (WTF::HashTableConstKeysIterator::operator*):
-        (WTF::HashTableConstKeysIterator::operator->):
-        (WTF::HashTableConstKeysIterator::operator++):
-        (WTF::HashTableConstValuesIterator::HashTableConstValuesIterator):
-        (WTF::HashTableConstValuesIterator::get):
-        (WTF::HashTableConstValuesIterator::operator*):
-        (WTF::HashTableConstValuesIterator::operator->):
-        (WTF::HashTableConstValuesIterator::operator++):
-        (WTF::HashTableKeysIterator::HashTableKeysIterator):
-        (WTF::HashTableKeysIterator::get):
-        (WTF::HashTableKeysIterator::operator*):
-        (WTF::HashTableKeysIterator::operator->):
-        (WTF::HashTableKeysIterator::operator++):
-        (WTF::HashTableKeysIterator::operator HashTableConstKeysIterator<HashTableType, KeyType, MappedType>):
-        (WTF::HashTableValuesIterator::HashTableValuesIterator):
-        (WTF::HashTableValuesIterator::get):
-        (WTF::HashTableValuesIterator::operator*):
-        (WTF::HashTableValuesIterator::operator->):
-        (WTF::HashTableValuesIterator::operator++):
-        (WTF::HashTableValuesIterator::operator HashTableConstValuesIterator<HashTableType, KeyType, MappedType>):
-        (WTF::operator==):
-        (WTF::operator!=):
-        * wtf/HashTable.h:
-        * wtf/Vector.h:
-        (WTF::::appendRange):
-
-2007-09-04  Maciej Stachowiak  <mjs@apple.com>
-
-        Reviewed by Darin.
-        
-        - Remove single-threaded optimization for FastMalloc. 
-        
-        It does not appear to help anywhere but Mac OS X on PPC, due to
-        pthread_getspecific being slow there. On Intel, removing the
-        optimization results in a 1% PLT speedup, a 2% JS iBench speedup,
-        and no measurable effect on HTML iBench (maybe a slight speedup).
-
-        * JavaScriptCore.xcodeproj/project.pbxproj:
-        * kjs/collector.cpp:
-        (KJS::Collector::registerThread):
-        * wtf/FastMalloc.cpp:
-        (WTF::TCMalloc_ThreadCache::GetCache):
-        (WTF::TCMalloc_ThreadCache::GetCacheIfPresent):
-        (WTF::TCMalloc_ThreadCache::CreateCacheIfNecessary):
-        (WTF::do_malloc):
-        * wtf/FastMallocInternal.h: Removed.
-
-2007-09-03  Mark Rowe  <mrowe@apple.com>
-
-        Reviewed by Tim Hatcher.
-
-        <rdar://problem/5452164> Production build with in symbols directory has no debug info
-
-        Enable debug symbol generation on all build configurations.  Production builds are stripped
-        of symbols by Xcode during deployment post-processing.
-
-        * Configurations/Base.xcconfig:
-        * JavaScriptCore.xcodeproj/project.pbxproj:
-
-2007-08-30  Riku Voipio  <riku.voipio@iki.fi>
-
-        Reviewed by Dave Kilzer.
-
-        Better ARM defines.
-
-        * kjs/ustring.h: Update comments to reflect the change and update test
-        to fit changes to Platform.h.
-        * wtf/Platform.h: Forced packing is only needed on oldabi ARM.
-        Set middle-endian floats only for little-endian oldabi ARM.
-        Set big-endian define for big-endian ARM.
-
-2007-08-29  Ryan Leavengood <leavengood@gmail.com>
-
-        Reviewed by Maciej.
-
-        http://bugs.webkit.org/show_bug.cgi?id=15043
-        - posix_memalign takes a void** as its first parameter. My port makes use of this function call.
-
-        * kjs/collector.cpp:
-        (KJS::allocateBlock):
-
-2007-08-26  Darin Adler  <darin@apple.com>
-
-        - quick follow on to that last check-in
-
-        * API/JSCallbackObject.cpp: (KJS::JSCallbackObject::JSCallbackObject):
-        Need to initialize m_class to 0.
-
-2007-08-26  Mark Rowe  <mrowe@apple.com>
-
-        Reviewed by Darin Adler.
-
-        <rdar://problem/4949002> JSGlobalContextCreate can cause crashes because it passes a NULL JSContextRef to the globalObjectClass's initialize callback
-
-        JSCallbackObject now tracks whether it was constructed with a null ExecState.  This will happen when the object is being used as the global object,
-        as the Interpreter needs to be created after the global object.  In this situation the initialization is deferred until after the Interpreter's
-        ExecState is available to be passed down to the initialize callbacks.
-
-        * API/JSCallbackObject.cpp:
-        (KJS::JSCallbackObject::init): Track whether we successfully initialized.
-        (KJS::JSCallbackObject::initializeIfNeeded): Attempt to initialize with the new ExecState.
-        * API/JSCallbackObject.h:
-        * API/JSContextRef.cpp:
-        (JSGlobalContextCreate): Initialize the JSCallbackObject with the Interpreter's ExecState.
-        * API/testapi.c:
-        (testInitializeOfGlobalObjectClassHasNonNullContext):
-        (main): Verify that the context passed to the initialize callback is non-null.
-
-2007-08-26  Mark Rowe  <mrowe@apple.com>
-
-        Reviewed by Darin Adler.
-
-        <rdar://problem/5438496> JSGlobalContextCreate crashes when passed a custom class
-
-        * API/JSContextRef.cpp:
-        (JSGlobalContextCreate): Specify jsNull() as the prototype and let Interpreter's constructor fix it up to point at builtinObjectPrototype().
-        * API/testapi.c:
-        (main): Use an instance of a custom class as the global object to ensure the code path is exercised in the test.
-
-2007-08-26  Mike Hommey  <glandium@debian.org>
-
-        Reviewed by Mark Rowe and David Kilzer.
-
-        Fix build failure on arm.
-
-        * wtf/Platform.h: Also test if __arm__ is defined.
-
-2007-08-25  Peter Kasting <zerodpx@gmail.org>
-
-        Reviewed by Maciej Stachowiak.
-
-        Part 3 of http://bugs.webkit.org/show_bug.cgi?id=14967
-        Bug 14967: Reduce wtf::Vector::operator[]() overloads
-
-        * wtf/Vector.h:
-        (WTF::Vector::operator[]): Only provide versions of operator[] that takes a size_t argument.
-
-2007-08-25  Peter Kasting <zerodpx@gmail.org>
-
-        Reviewed by Sam Weinig.
-
-        Part 2 of http://bugs.webkit.org/show_bug.cgi?id=14967.
-        Eliminate all remaining implicit conversions of wtf::Vector<T> to T*.  Where code was
-        previously checking that the Vector's data pointer was non-NULL, check !Vector::isEmpty()
-        instead.
-
-        * wtf/Vector.h:
-        (WTF::Vector::data):
-
-2007-08-16  Kevin McCullough  <kmccullough@apple.com>
-
-        Reviewed by Geoff and Adam.
-
-        - Changing stack depth to 500 (from 100 on mac and win) to help out some apps specifically gmail.  <rdar://problem/3590522> JavaScript call stack limit of 99 is too small for some applications; needs to be closer to 500 (4045)
-
-        * kjs/object.cpp:
-
-2007-08-15  Peter Kasting  <zerodpx@gmail.org>
-
-        Reviewed by Darin.
-        
-        http://bugs.webkit.org/show_bug.cgi?id=14967 part 1 - Eliminate most implicit
-        conversions of wtf::Vector<T> to T* by explicitly calling .data()
-
-        * API/JSCallbackConstructor.cpp:
-        (KJS::JSCallbackConstructor::construct):
-        * API/JSCallbackFunction.cpp:
-        (KJS::JSCallbackFunction::callAsFunction):
-        * API/JSCallbackObject.cpp:
-        (KJS::JSCallbackObject::construct):
-        (KJS::JSCallbackObject::callAsFunction):
-        * bindings/c/c_instance.cpp:
-        (KJS::Bindings::CInstance::invokeMethod):
-        (KJS::Bindings::CInstance::invokeDefaultMethod):
-        * kjs/number_object.cpp:
-        (integer_part_noexp):
-        (char_sequence):
-        * kjs/ustring.cpp:
-        (KJS::UString::UTF8String):
-
-2007-08-14  Darin Adler  <darin@apple.com>
-
-        Reviewed by Sam.
-
-        - fix <rdar://problem/5410570> Global initializer introduced by use of std::numeric_limits in r24919
-
-        * kjs/ustring.cpp:
-        (KJS::overflowIndicator): Turned into a function.
-        (KJS::maxUChars): Ditto.
-        (KJS::allocChars): Use the functions.
-        (KJS::reallocChars): Ditto.
-        (KJS::UString::expandedSize): Ditto.
-
-2007-08-12  Darin Adler  <darin@apple.com>
-
-        Reviewed by Maciej.
-
-        - fix http://bugs.webkit.org/show_bug.cgi?id=14931
-          <rdar://problem/5403816> JavaScript regular expression non-participating capturing parentheses
-          fail in 3 different ways
-
-        Test: fast/js/regexp-non-capturing-groups.html
-
-        * kjs/string_object.cpp:
-        (KJS::replace): Add missing code to handle undefined backreferences; before we'd get the empty string
-        instead of a JavaScript "undefined" value.
-        (KJS::StringProtoFunc::callAsFunction): Implemented backreference support for split.
-        * pcre/pcre_exec.c: (match): Made backreferences to undefined groups match the empty string instead
-        of always failing. Only in JAVASCRIPT mode.
-
-        * tests/mozilla/expected.html: Add a new expected test success, since this fixed one test result.
-
-2007-08-10  Timothy Hatcher  <timothy@apple.com>
-
-        Reviewed by Adam.
-
-        <rdar://problem/5394449> Stop using some Carbon UI APIs for 64 bit
-
-        Disable the NPAPI for 64-bit on Mac OS X.
-
-        * Configurations/JavaScriptCore.xcconfig: Use the 64-bit export file.
-        * JavaScriptCore.xcodeproj/project.pbxproj: Create a 64-bit export file
-        that filters out the NPN fnctions.
-        * bindings/NP_jsobject.cpp: #ifdef out this for 64-bit on Mac OS X
-        * bindings/NP_jsobject.h: Ditto.
-        * bindings/c/c_class.cpp: Ditto.
-        * bindings/c/c_class.h: Ditto.
-        * bindings/c/c_instance.cpp: Ditto.
-        * bindings/c/c_instance.h: Ditto.
-        * bindings/c/c_runtime.cpp: Ditto.
-        * bindings/c/c_runtime.h: Ditto.
-        * bindings/c/c_utility.cpp: Ditto.
-        * bindings/c/c_utility.h: Ditto.
-        * bindings/npapi.h: Ditto.
-        * bindings/npruntime.cpp: Ditto.
-        * bindings/npruntime.h: Ditto.
-        * bindings/npruntime_impl.h: Ditto.
-        * bindings/npruntime_priv.h: Ditto.
-        * bindings/runtime.cpp:
-        (KJS::Bindings::Instance::createBindingForLanguageInstance):
-         don't creat an NPObject on Mac OS X in 64-bit.
-
-2007-08-09  Mark Rowe  <mrowe@apple.com>
-
-        Reviewed by Antti.
-
-        <rdar://problem/5400709> Versioning in debug and release builds should include minor and tiny version before +
-
-        * Configurations/Version.xcconfig:
-        * JavaScriptCore.xcodeproj/project.pbxproj: Add a shell script phase to make to dependency between
-        Version.xcconfig and Info.plist explicit to Xcode.
-
-2007-08-08  George Staikos  <staikos@kde.org>
-
-        Make it compile with Qt again.
-
-        * wtf/unicode/qt4/UnicodeQt4.h:
-        (WTF::Unicode::toUpper):
-
-2007-08-07  Sam Weinig  <sam@webkit.org>
-
-        Reviewed by Oliver.
-
-        Fix for http://bugs.webkit.org/show_bug.cgi?id=14897
-        Decompilation of double negation fails and produces invalid or incorrect code
-
-        Test: fast/js/function-decompilation-operators.html
-
-        * kjs/nodes2string.cpp:
-        (UnaryPlusNode::streamTo): Put space after unary operator.  Matches Firefox.
-        (NegateNode::streamTo): Diito.
-        (MultNode::streamTo): Put spaces around binary operator.  Matches Firefox.
-        (AddNode::streamTo): Ditto.
-
-2007-08-07  Darin Adler  <darin@apple.com>
-
-        Reviewed by Adele.
-
-        - fix <rdar://problem/5383104> REGRESSION: XHR.responseText is null instead of empty string
-          in http/tests/xmlhttprequest/zero-length-response.html
-
-        The new code to handle out of memory conditions was turning a "" into a null string.
-
-        * kjs/ustring.h: Removed UCharReference, which has long been obsolete and unused.
-        Removed copyForWriting, which was only used for the upper/lowercasing code and for
-        UCharReference.
-        * kjs/ustring.cpp:
-        (KJS::allocChars): Removed special case that made this fail (return 0) when passed 0.
-        Instead assert that we're not passed 0. Also added an overflow check for two reasons:
-        1) for sizes that aren't checked this prevents us from allocating a buffer that's too
-        small, and 2) for sizes where we overflowed in the expandedSize function and returned
-        overflowIndicator, it guarantees we fail.
-        (KJS::reallocChars): Ditto.
-        (KJS::UString::expandedSize): Return a large number, overflowIndicator, rather than 0
-        for cases where we overflow.
-        (KJS::UString::spliceSubstringsWithSeparators): Added a special case for empty string so
-        we don't call allocChars with a length of 0.
-        (KJS::UString::operator=): Added special characters for both 0 and empty string so we
-        match the behavior of the constructor. This avoids calling allocChars with a length of 0
-        and making a null string rather than an empty string in that case, and also matches the
-        pattern used in the rest of the functions.
-        (KJS::UString::operator[]): Made the return value const so code that tries to use the
-        operator to modify the string will fail.
-
-        * kjs/string_object.cpp: (KJS::StringProtoFunc::callAsFunction): Rewrote uppercasing and
-        lowercasing functions so they don't need copyForWriting any more -- it wasn't really doing
-        any good for optimization purposes. Instead use a Vector and releaseBuffer.
-
-        * wtf/unicode/icu/UnicodeIcu.h: Eliminate one of the versions of toLower/toUpper -- we now
-        only need the version where both a source and destination buffer is passed in, not the one
-        that works in place.
-        * wtf/unicode/qt4/UnicodeQt4.h: Ditto.
-
-2007-08-06  Sam Weinig  <sam@webkit.org>
-
-        Reviewed by Oliver.
-
-        Fix for http://bugs.webkit.org/show_bug.cgi?id=14891
-        Decompilation of try block immediately following "else" fails
-
-        Test: fast/js/toString-try-else.html
-
-        * kjs/nodes2string.cpp:
-        (TryNode::streamTo): Add newline before "try".
-
-2007-08-07  Mark Rowe  <mrowe@apple.com>
-
-        Reviewed by Maciej.
-
-        <rdar://problem/5388774> REGRESSION: Hang occurs after clicking "Attach a file " link in a new .Mac message
-
-        Attempting to acquire the JSLock inside CollectorHeap::forceLock can lead to a deadlock if the thread currently
-        holding the lock is waiting on the thread that is forking.  It is not considered safe to use system frameworks
-        after a fork without first execing[*] so it is not particularly important to ensure that the collector and
-        fastMalloc allocators are unlocked in the child process.  If the child process wishes to use JavaScriptCore it
-        should exec after forking like it would to use any other system framework.
-        [*]: <http://lists.apple.com/archives/Cocoa-dev/2005/Jan/msg00676.html>
-
-        * kjs/CollectorHeapIntrospector.cpp: Remove forceLock and forceUnlock implementations.
-        * kjs/CollectorHeapIntrospector.h: Stub out forceLock and forceUnlock methods.
-        * wtf/FastMalloc.cpp: Ditto.
-
-2007-08-06  Darin Adler  <darin@apple.com>
-
-        Rubber stamped by Geoff.
-
-        * kjs/ustring.h: Added an assertion which would have helped us find the
-        previous bug more easily.
-
-2007-08-06  Darin Adler  <darin@apple.com>
-
-        Reviewed by Anders.
-
-        - fix <rdar://problem/5387589> 9A514: Quartz Composer crash on launch in KJS::jsString
-
-        * API/JSBase.cpp:
-        (JSEvaluateScript): Turn NULL for sourceURL into UString::null(), just as JSObjectMakeFunction already does.
-        (JSCheckScriptSyntax): Ditto.
-
-2007-08-06  Matt Lilek  <pewtermoose@gmail.com>
-
-        Not reviewed, build fix.
-
-        * kjs/string_object.cpp:
-        (KJS::StringProtoFunc::callAsFunction):
-
-2007-08-04  Darin Adler  <darin@apple.com>
-
-        Reviewed by Maciej.
-
-        - fix <rdar://problem/5371862> crash in Dashcode due to Quartz Composer JavaScript garbage collector reentrancy
-
-        * API/JSBase.cpp: (JSGarbageCollect): Don't call collector() if isBusy() returns true.
-
-        * kjs/collector.h: Added isBusy(), removed the unused return value from collect()
-        * kjs/collector.cpp: Added an "operation in progress" flag to the allocator.
-        (KJS::Collector::allocate): Call abort() if an operation is already in progress. Set the new flag instead
-        of using the debug-only GCLock.
-        (KJS::Collector::collect): Ditto.
-        (KJS::Collector::isBusy): Added.
-
-2007-08-04  Maciej Stachowiak  <mjs@apple.com>
-
-        Reviewed by Darin and Adam.
-        
-        <rdar://problem/5368990> REGRESSION: newsgator.com sign-on 6x slower than Safari 3 beta due to GC changes (14808)
-
-        * kjs/string_object.cpp:
-        (KJS::replace): if the string didn't change (very common in some cases) reuse the original string value.
-        (KJS::StringProtoFunc::callAsFunction): Pass in the StringImp* when replacing, not just the UString.
-        * kjs/string_object.h:
-        (KJS::StringInstance::internalValue): covariant override to return StringImp for convenience
-
-2007-08-04  Mark Rowe  <mrowe@apple.com>
-
-        Reviewed by Oliver Hunt.
-
-        <rdar://problem/5385145> r24843 introduces a crash on calling fork() (14878)
-        http://bugs.webkit.org/show_bug.cgi?id=14878
-
-        Provide no-op functions for all members of the malloc_zone_t and malloc_introspection_t structures that we
-        register to avoid crashes in system code that assumes they will be non-null.
-
-        * kjs/CollectorHeapIntrospector.cpp:
-        (KJS::CollectorHeapIntrospector::CollectorHeapIntrospector):
-        (KJS::CollectorHeapIntrospector::forceLock): Grab the lock.
-        (KJS::CollectorHeapIntrospector::forceUnlock): Release the lock.
-        * kjs/CollectorHeapIntrospector.h:
-        (KJS::CollectorHeapIntrospector::goodSize):
-        (KJS::CollectorHeapIntrospector::check):
-        (KJS::CollectorHeapIntrospector::print):
-        (KJS::CollectorHeapIntrospector::log):
-        (KJS::CollectorHeapIntrospector::statistics):
-        (KJS::CollectorHeapIntrospector::size):
-        (KJS::CollectorHeapIntrospector::zoneMalloc):
-        (KJS::CollectorHeapIntrospector::zoneCalloc):
-        (KJS::CollectorHeapIntrospector::zoneFree):
-        * wtf/FastMalloc.cpp:
-        (WTF::FastMallocZone::goodSize):
-        (WTF::FastMallocZone::check):
-        (WTF::FastMallocZone::print):
-        (WTF::FastMallocZone::log):
-        (WTF::FastMallocZone::forceLock): Grab the TCMalloc locks.
-        (WTF::FastMallocZone::forceUnlock): Release the TCMalloc locks.
-        (WTF::FastMallocZone::FastMallocZone):
-
-2007-08-04  Mark Rowe  <mrowe@apple.com>
-
-        Rubber-stamped by Anders.
-
-        * pcre/pcre_compile.c: Remove non-ASCII character from a comment.
-
-2007-08-02  Mark Rowe  <mrowe@apple.com>
-
-        Reviewed by Geoff Garen.
-
-        <rdar://problem/4212199> 'leaks' reports false leaks in WebKit (because the WTF allocator uses mmap?)
-
-        Implement malloc zone introspection routines to allow leaks, heap, and friends to request information
-        about specific memory regions that were allocated by FastMalloc or the JavaScriptCore collector.
-
-        This requires tool-side support before the regions will be displayed.  The addition of that support is
-        tracked by <rdar://problems/5353057&5353060>.
-
-        * JavaScriptCore.exp: Export the two variables that are used by leaks to introspect the allocators.
-        * JavaScriptCore.xcodeproj/project.pbxproj:
-        * kjs/AllInOneFile.cpp:
-        * kjs/CollectorHeapIntrospector.cpp: Added.
-        (KJS::):
-        (KJS::CollectorHeapIntrospector::init): 
-        (KJS::CollectorHeapIntrospector::CollectorHeapIntrospector): Create and register our zone with the system.
-        (KJS::CollectorHeapIntrospector::enumerate): Iterate over the CollectorBlocks that are in use and report them to the caller as being used.
-        * kjs/CollectorHeapIntrospector.h: Added.
-        (KJS::CollectorHeapIntrospector::size): Return zero to indicate the specified pointer does not belong to this zone.
-        * kjs/collector.cpp:
-        (KJS::Collector::registerThread): Register the CollectorHeapIntrospector with the system when the first thread is registered with the collector.
-        * wtf/FastMalloc.cpp:
-        (WTF::TCMalloc_PageHeap::GetDescriptorEnsureSafe):
-        (WTF::TCMalloc_ThreadCache_FreeList::enumerateFreeObjects): Enumerate the objects on the free list.
-        (WTF::TCMalloc_ThreadCache::enumerateFreeObjects): Ditto.
-        (WTF::TCMalloc_Central_FreeList::enumerateFreeObjects): Ditto.
-        (WTF::TCMalloc_ThreadCache::InitModule): Register the FastMallocZone with the system when initializing TCMalloc.
-        (WTF::FreeObjectFinder::FreeObjectFinder):
-        (WTF::FreeObjectFinder::visit): Add an object to the free list.
-        (WTF::FreeObjectFinder::isFreeObject):
-        (WTF::FreeObjectFinder::freeObjectCount):
-        (WTF::FreeObjectFinder::findFreeObjects): Find the free objects within a thread cache or free list.
-        (WTF::PageMapFreeObjectFinder::PageMapFreeObjectFinder): Find the free objects within a TC_PageMap.
-        (WTF::PageMapFreeObjectFinder::visit): Called once per allocated span.  Record whether the span or any subobjects are free.
-        (WTF::PageMapMemoryUsageRecorder::PageMapMemoryUsageRecorder):
-        (WTF::PageMapMemoryUsageRecorder::visit): Called once per allocated span.  Report the range of memory as being allocated, and the span or
-        its subobjects as being used if they do not appear on the free list.
-        (WTF::FastMallocZone::enumerate): Map the key remote TCMalloc data structures into our address space.  We then locate all free memory ranges
-        before reporting the other ranges as being in use.
-        (WTF::FastMallocZone::size): Determine whether the given pointer originates from within our allocation zone.  If so,
-        we return its allocation size.
-        (WTF::FastMallocZone::zoneMalloc):
-        (WTF::FastMallocZone::zoneCalloc):
-        (WTF::FastMallocZone::zoneFree):
-        (WTF::FastMallocZone::zoneRealloc):
-        (WTF::):
-        (WTF::FastMallocZone::FastMallocZone): Create and register our zone with the system.
-        (WTF::FastMallocZone::init):
-        * wtf/MallocZoneSupport.h: Added.
-        (WTF::RemoteMemoryReader::RemoteMemoryReader): A helper class to ease the process of mapping memory in a different process into
-        our local address space
-        (WTF::RemoteMemoryReader::operator()):
-        * wtf/TCPageMap.h:
-        (TCMalloc_PageMap2::visit): Walk over the heap and visit each allocated span.
-        (TCMalloc_PageMap3::visit): Ditto.
-
-2007-08-02  Mark Rowe  <mrowe@apple.com>
-
-        Build fix.
-
-        * kjs/ustring.cpp:
-        (KJS::UString::expandedSize):  Use std::numeric_limits<size_t>::max() rather than the non-portable SIZE_T_MAX.
-
-2007-08-02  Mark Rowe  <mrowe@apple.com>
-
-        Reviewed by Maciej.
-
-        <rdar://problem/5352887> "Out of memory" error during repeated JS string concatenation leaks hundreds of MBs of RAM
-
-        A call to fastRealloc was failing which lead to UString::expandCapacity leaking the buffer it was trying to reallocate.
-        It also resulted in the underlying UString::rep having both a null baseString and buf field, which meant that attempting
-        to access the contents of the string after the failed memory reallocation would crash.
-
-        A third issue is that expandedSize size was calculating the new length in a way that led to an integer overflow occurring.
-        Attempting to allocate a string more than 190,000,000 characters long would fail a the integer overflow would lead to a
-        memory allocation of around 3.6GB being attempted rather than the expected 390MB.  Sizes that would lead to an overflow
-        are now  returned as zero and callers are updated to treat this as though the memory allocation has failed.
-
-        * kjs/array_object.cpp:
-        (ArrayProtoFunc::callAsFunction): Check whether the append failed and raise an "Out of memory" exception if it did.
-        * kjs/ustring.cpp:
-        (KJS::allocChars): Wrapper around fastMalloc that takes a length in characters.  It will return 0 when asked to allocate a zero-length buffer.
-        (KJS::reallocChars): Wrapper around fastRealloc that takes a length in characters.  It will return 0 when asked to allocate a zero-length buffer.
-        (KJS::UString::expandedSize): Split the size calculation in two and guard against overflow during each step.
-        (KJS::UString::expandCapacity): Don't leak r->buf if reallocation fails.  Instead free the memory and use the null representation.
-        (KJS::UString::expandPreCapacity): If fastMalloc fails then use the null representation rather than crashing in memcpy.
-        (KJS::UString::UString): If calls to expandCapacity, expandPreCapacity or fastMalloc fail then use the null representation rather than crashing in memcpy.
-        (KJS::UString::append): Ditto.
-        (KJS::UString::operator=): Ditto.
-        * kjs/ustring.h: Change return type of expandedSize from int to size_t.
-
-2007-08-01  Darin Adler  <darin@apple.com>
-
-        Reviewed by Kevin McCullough.
-
-        - fix <rdar://problem/5375186> pointers to pieces of class definition passed to JSClassCreate should all be const
-
-        * API/JSObjectRef.h: Added const.
-
-        * API/JSClassRef.cpp:
-        (OpaqueJSClass::OpaqueJSClass): Added const.
-        (OpaqueJSClass::create): Added const.
-        * API/JSObjectRef.cpp:
-        (JSClassCreate): Added const.
-
-2007-08-01  Steve Falkenburg  <sfalken@apple.com>
-
-        Build mod: Fix sln to match configs in vcproj.
-        
-        Reviewed by Adam.
-
-        * JavaScriptCore.vcproj/JavaScriptCore.make:
-        * JavaScriptCore.vcproj/JavaScriptCore.sln:
-
-2007-07-30  Simon Hausmann  <hausmann@kde.org>
-
-        Done with and reviewed by Lars.
-
-        Removed the __BUILDING_QT ifdef in JSStringRef.h and changed UChar for the Qt build to use wchar_t on Windows.
-
-        * API/JSStringRef.h:
-        * wtf/unicode/qt4/UnicodeQt4.h:
-
-2007-07-27  Simon Hausmann  <hausmann@kde.org>
-
-        Done with and reviewed by Lars and Zack.
-
-        Always define JSChar to be unsigned short for the Qt builds, to ensure compatibility with UChar.
-
-        * API/JSStringRef.h:
-
-2007-07-27  Simon Hausmann  <hausmann@kde.org>
-
-        Done with and reviewed by Lars and Zack.
-
-        Fix compilation with Qt on Windows with MingW: Implemented currentThreadStackBase() for this platform.
-
-        * kjs/collector.cpp:
-        (KJS::currentThreadStackBase):
-
-2007-07-27  Simon Hausmann  <hausmann@kde.org>
-
-        Done with and reviewed by Lars and Zack.
-
-        Fix compilation with Qt on Windows with MingW: The MingW headers do not provide a prototype for a reentrant version of localtime. But since we don't use multiple threads for the Qt build we can use the plain localtime() function.
-
-        * kjs/DateMath.cpp:
-        (KJS::getDSTOffsetSimple):
-
-2007-07-27  Simon Hausmann  <hausmann@kde.org>
-
-        Done with and reviewed by Lars and Zack.
-
-        Use $(MOVE) instead of mv to eliminated the shell dependency and replaced the long shell line to call bison and modify the css grammar file with a few lines of portable perl code.
-
-        * JavaScriptCore.pri:
-
-2007-07-27  Simon Hausmann  <hausmann@kde.org>
-
-        Done with and reviewed by Lars and Zack.
-
-        Implemented currentTime() in the interpreter by using QDateTime, so that we don't need timeGetTime() on Windows and therefore also don't need to link against Winmm.dll.
-
-        * kjs/interpreter.cpp:
-        (KJS::getCurrentTime):
-        * kjs/testkjs.cpp:
-        (StopWatch::start):
-        (StopWatch::stop):
-
-2007-07-27  Simon Hausmann  <hausmann@kde.org>
-
-        Done with and reviewed by Lars and Zack.
-
-        Replace the use of snprintf with QByteArray to compile under msvc 2005 express.
-
-        * bindings/qt/qt_instance.cpp:
-        (KJS::Bindings::QtInstance::stringValue):
-
-2007-07-27  Simon Hausmann  <hausmann@kde.org>
-
-        Done with and reviewed by Lars and Zack.
-
-        Don't use pthread.h unless thread support is enabled.
-
-        * kjs/collector.cpp:
-        (KJS::Collector::registerAsMainThread):
-        (KJS::onMainThread):
-
-2007-07-27  Simon Hausmann  <hausmann@kde.org>
-
-        Done with and reviewed by Lars and Zack.
-
-        Removed TCSystemMalloc from the Qt build, it's not necessary it seems.
-
-        * JavaScriptCore.pri:
-
-2007-07-27  Simon Hausmann  <hausmann@kde.org>
-
-        Done with and reviewed by Lars and Zack.
-
-        Added os-win32 to the include search path for the Qt windows build in order to provide the fake stdint.h header file.
-
-        * JavaScriptCore.pri:
-
-2007-07-25  Maciej Stachowiak  <mjs@apple.com>
-
-        Reviewed by Mark.
-
-        - follow-up to previous change
-        
-        * kjs/ustring.cpp:
-        (KJS::UString::operator=): Make sure to reset the length when
-        replacing the buffer contents for a single-owned string.
-
-2007-07-25  Maciej Stachowiak  <mjs@apple.com>
-
-        Reviewed by Darin.
-        
-        - JavaScriptCore part of fix for <rdar://problem/5300291> Optimize GC to reclaim big, temporary objects (like XMLHttpRequest.responseXML) quickly
-        
-        Also, as a side effect of optimizations included in this patch:
-        - 7% speedup on JavaScript iBench
-        - 4% speedup on "Celtic Kane" JS benchmark
-        
-        The basic idea is explained in a big comment in collector.cpp. When unusually 
-        large objecs are allocated, we push the next GC closer on the assumption that
-        most objects are short-lived.
-        
-        I also did the following two optimizations in the course of tuning
-        this not to be a performance regression:
-
-        1) Change UString::Rep to hold a self-pointer as the baseString in
-        the unshared case, instead of a null pointer; this removes a
-        number of null checks in hot code because many places already
-        wanted to use the rep itself or the baseString as appropriate.
-        
-        2) Avoid creating duplicate StringImpls when creating a
-        StringInstance (the object wrapper for a JS string) or calling
-        their methods. Since a temporary wrapper object is made every time
-        a string method is called, this resulted in two useless extra
-        StringImpls being allocated for no reason whenever a String method
-        was invoked on a string value. Now we bypass those.
-        
-        * kjs/collector.cpp:
-        (KJS::):
-        (KJS::Collector::recordExtraCost): Basics of the extra cost mechanism.
-        (KJS::Collector::allocate): ditto
-        (KJS::Collector::collect): ditto
-        * kjs/collector.h:
-        (KJS::Collector::reportExtraMemoryCost): ditto
-        * kjs/array_object.cpp:
-        (ArrayInstance::ArrayInstance): record extra cost
-        * kjs/internal.cpp:
-        (KJS::StringImp::toObject): don't create a whole new StringImpl just
-        to be the internal value of a StringInstance! StringImpls are immutable
-        so there's no point tot his.
-        * kjs/internal.h:
-        (KJS::StringImp::StringImp): report extra cost
-        * kjs/string_object.cpp:
-        (KJS::StringInstance::StringInstance): new version that takes a StringImp
-        (KJS::StringProtoFunc::callAsFunction): don't create a whole new StringImpl
-        just to convert self to string! we already have one in the internal value
-        * kjs/string_object.h: report extra cost        
-        * kjs/ustring.cpp: All changes to handle baseString being self instead of null in the 
-        unshared case.
-        (KJS::):
-        (KJS::UString::Rep::create):
-        (KJS::UString::Rep::destroy):
-        (KJS::UString::usedCapacity):
-        (KJS::UString::usedPreCapacity):
-        (KJS::UString::expandCapacity):
-        (KJS::UString::expandPreCapacity):
-        (KJS::UString::UString):
-        (KJS::UString::append):
-        (KJS::UString::operator=):
-        (KJS::UString::copyForWriting):
-        * kjs/ustring.h:
-        (KJS::UString::Rep::baseIsSelf): new method, now that baseString is
-        self instead of null in the unshared case we can't just null check.
-        (KJS::UString::Rep::data): adjusted as mentioned above
-        (KJS::UString::cost): new method to compute the cost for a UString, for
-        use by StringImpl.
-
-        * kjs/value.cpp:
-        (KJS::jsString): style fixups.
-        (KJS::jsOwnedString): new method, use this for strings allocated from UStrings
-        held by the parse tree. Tracking their cost as part of string cost is pointless,
-        because garbage collecting them will not actually free the relevant string buffer.
-        * kjs/value.h: prototyped jsOwnedString.
-        * kjs/nodes.cpp:
-        (StringNode::evaluate): use jsOwnedString as appropriate
-        (RegExpNode::evaluate): ditto
-        (PropertyNameNode::evaluate): ditto
-        (ForInNode::execute): ditto
-        
-        * JavaScriptCore.exp: Exported some new symbols.
-
-2007-07-23  Anders Carlsson  <andersca@apple.com>
-
-        Reviewed by Geoff.
-
-        <rdar://problem/5121461> REGRESSION: Unable to load JigZone puzzle
-        
-        * bindings/jni/jni_jsobject.cpp:
-        (JavaJSObject::createNative):
-        
-        Call RootObject::gcProtect on the global object, thereby putting it in the
-        "protect count" set which is used for checking if a native handle is valid.
-        
-2007-07-23  Darin Adler  <darin@apple.com>
-
-        * pcre/pcre_compile.c: Roll back a tiny accidental change in the unused !JAVASCRIPT
-        side of an #ifdef. This has no effect when using PCRE in JAVASCRIPT mode as we do,
-        but seems worth rolling back.
-
-2007-07-23  Maciej Stachowiak  <mjs@apple.com>
-
-        Reviewed by Oliver.
-        
-        - fix remaining problems with Window shadowing
-
-        * kjs/nodes.cpp:
-        (VarDeclNode::evaluate): Tweak the special case a little.
-
-2007-07-23  Maciej Stachowiak  <mjs@apple.com>
-
-        Reviewed by Oliver.
-        
-        - fix Window shadowing regressions caused by the previous commit.
-
-        * kjs/nodes.cpp:
-        (VarDeclNode::evaluate): Handle the case of global scope specially.
-
-2007-07-22  Maciej Stachowiak  <mjs@apple.com>
-
-        Reviewed by Darin.
-
-        -fixed <rdar://problem/5353293> REGRESSION (r24287): 1% i-Bench JS slowdown from JavaScript compatibility fix (14719)
-        http://bugs.webkit.org/show_bug.cgi?id=14719
-        
-        My fix for this actually resulted in JS iBench being 1% faster than before the regression
-        and the Celtic Kane benchmark being 5% faster than before the regression.
-        
-        * kjs/nodes.cpp:
-        (VarDeclNode::handleSlowCase): factored out the slow code path to be out of line.
-        (VarDeclNode::evaluate): I did a couple of things:
-        (1) Don't check if the variable is already declared by looking for the property in
-        the variable object, that code path was dead code.
-        (2) Special-case the common case where the top of the scope and the variable object
-        are the same; in that case the variable must always be in the variable object.
-        (3) Don't return a jsString() of the variable name, nothing uses the return value
-        from this node types evaluate method.
-        * kjs/nodes.h:
-
-2007-07-22  Darin Adler  <darin@apple.com>
-
-        Reviewed by Kevin Decker.
-
-        - fix <rdar://problem/5126394> REGRESSION: Crash after clicking back button in test application (13250)
-          http://bugs.webkit.org/show_bug.cgi?id=13250
-
-        * bindings/objc/objc_utility.mm: (KJS::Bindings::convertObjcValueToValue):
-        If the object returns 0 for _imp, convert that to "undefined", since callers
-        can't cope with a JSValue of 0.
-
-2007-07-19  Geoffrey Garen  <ggaren@apple.com>
-
-        Reviewed by Darin Adler.
-        
-        Fixed http://bugs.webkit.org/show_bug.cgi?id=10880 | <rdar://problem/5335694>
-        REGRESSION: JavaScript menu doesn't appear on pricepoint.com (14595)
-        
-        Though the ECMA spec says auto-semicolon insertion should not occur
-        without a newline or '}', Firefox treats do-while specially, and the
-        library used by pricepoint.com requires that special treatment.
-        
-        * JavaScriptCore.xcodeproj/project.pbxproj:
-        * kjs/grammar.y:
-
-2007-07-19  Darin Adler  <darin@apple.com>
-
-        Reviewed by Geoff.
-
-        - fix <rdar://problem/5345440> PCRE computes wrong length for expressions with quantifiers
-          on named recursion or subexpressions
-
-        It's challenging to implement proper preflighting for compiling these advanced features.
-        But we don't want them in the JavaScript engine anyway.
-
-        Turned off the following features of PCRE (some of these are simply parsed and not implemented):
-
-            \C \E \G \L \N \P \Q \U \X \Z
-            \e \l \p \u \z
-            [::] [..] [==]
-            (?#) (?<=) (?<!) (?>)
-            (?C) (?P) (?R)
-            (?0) (and 1-9)
-            (?imsxUX)
-
-        Added the following:
-
-            \u \v
-
-        Because of \v, the js1_2/regexp/special_characters.js test now passes.
-
-        To be conservative, I left some features that JavaScript doesn't want, such as
-        \012 and \x{2013}, in place. We can revisit these later; they're not directly-enough
-        related to avoiding the incorrect preflighting.
-
-        I also didn't try to remove unused opcodes and remove code from the execution engine.
-        That could save code size and speed things up a bit, but it would require more changes.
-
-        * kjs/regexp.h:
-        * kjs/regexp.cpp: (KJS::RegExp::RegExp): Remove the sanitizePattern workaround for
-        lack of \u support, since the PCRE code now has \u support.
-
-        * pcre/pcre-config.h: Set JAVASCRIPT to 1.
-        * pcre/pcre_internal.h: Added ESC_v.
-
-        * pcre/pcre_compile.c: Added a different escape table for when JAVASCRIPT is set that
-        omits all the escapes we don't want interpreted and includes '\v'.
-        (check_escape): Put !JAVASCRIPT around the code for '\l', '\L', '\N', '\u', and '\U',
-        and added code to handle '\u2013' inside JAVASCRIPT.
-        (compile_branch): Put !JAVASCRIPT if around all the code implementing the features we
-        don't want.
-        (pcre_compile2): Ditto.
-
-        * tests/mozilla/expected.html: Updated since js1_2/regexp/special_characters.js now
-        passes.
-
-2007-07-18  Darin Adler  <darin@apple.com>
-
-        Reviewed by Oliver Hunt.
-
-        - fix <rdar://problem/5345432> PCRE computes length wrong for expressions such as "[**]"
-
-        Test: fast/js/regexp-charclass-crash.html
-
-        * pcre/pcre_compile.c: (pcre_compile2): Fix the preflight code that calls
-        check_posix_syntax to match the actual regular expression compilation code;
-        before it was missing the check of the first character.
-
-2007-07-19  Holger Hans Peter Freyther  <zecke@selfish.org>
-
-        Reviewed by Mark.
-
-        Define __BUILDING_GDK when building for Gdk to fix building testkjs on OSX.
-
-        * JavaScriptCore.pri:
-
-2007-07-18  Simon Hausmann  <hausmann@kde.org>
-
-        * Fix the Qt build, call dftables from the right directory.
-
-        Reviewed by Adam Treat.
-
-        * pcre/pcre.pri:
-
-2007-07-18  Simon Hausmann  <hausmann@kde.org>
-
-        Reviewed by Zack.
-
-        Don't call gcc directly when building the dftables tool but use a separate .pro file for the Qt build.
-
-        * pcre/dftables.pro: Added.
-        * pcre/pcre.pri:
-
-2007-07-17  Cameron Zwarich  <cwzwarich@uwaterloo.ca>
-
-        Reviewed by Darin, Maciej, and Adam.
-
-        Fixes <http://bugs.webkit.org/show_bug.cgi?id=9697>,
-              the failure of ecma/GlobalObject/15.1.2.2-2.js,
-              the failure of ecma/LexicalConventions/7.7.3-1.js,
-              and most of the failures of tests in ecma/TypeConversion/9.3.1-3.js.
-
-        Bug 9697: parseInt results may be inaccurate for numbers greater than 2^53
-
-        This patch also fixes similar issues in the lexer and UString::toDouble().
-
-        * kjs/function.cpp:
-        (KJS::parseIntOverflow):
-        (KJS::parseInt):
-        * kjs/function.h:
-        * kjs/lexer.cpp:
-        (KJS::Lexer::lex):
-        * kjs/ustring.cpp:
-        (KJS::UString::toDouble):
-        * tests/mozilla/expected.html:
-
-2007-07-16  Sam Weinig  <sam@webkit.org>
-
-        Reviewed by Oliver.
-
-        Turn off -Wshorten-64-to-32 warning for 64-bit builds.
-
-        * Configurations/Base.xcconfig:
-
-2007-07-14  Brady Eidson  <beidson@apple.com>
-
-        Reviewed by Sam Weinig
-
-        Initial check-in for <rdar://problem/3154486> - Supporting FTP directory listings in the browser
-
-        * wtf/Platform.h: Add ENABLE_FTPDIR feature to handle building on platforms that don't have the
-          proper network-layer support
-
-2007-07-14  Cameron Zwarich  <cwzwarich@uwaterloo.ca>
-
-        Reviewed by Darin.
-
-        Fixes http://bugs.webkit.org/show_bug.cgi?id=13517,
-              http://bugs.webkit.org/show_bug.cgi?id=14237, and
-              the failure of test js1_5/Scope/regress-185485.js
-
-        Bug 13517: DOM Exception 8 in finance.aol.com sub-page
-        Bug 14237: Javascript "var" statement interprets initialization in the topmost function scope
-
-        * kjs/nodes.cpp:
-        (VarDeclNode::evaluate):
-        * tests/mozilla/expected.html:
-
-2007-07-12  Alexey Proskuryakov  <ap@webkit.org>
-
-        Reviewed by Mitz.
-
-        http://bugs.webkit.org/show_bug.cgi?id=14596
-        Fix JSC compilation with KJS_VERBOSE.
-
-        * kjs/function.cpp:
-        (KJS::FunctionImp::passInParameters):
-
-2007-07-11  George Staikos  <staikos@kde.org>
-
-        Make it compile.
-
-        * ForwardingHeaders: Added.
-        * ForwardingHeaders/JavaScriptCore: Added.
-        * ForwardingHeaders/JavaScriptCore/APICast.h: Added.
-        * ForwardingHeaders/JavaScriptCore/JSBase.h: Added.
-        * ForwardingHeaders/JavaScriptCore/JSContextRef.h: Added.
-        * ForwardingHeaders/JavaScriptCore/JSLock.h: Added.
-        * ForwardingHeaders/JavaScriptCore/JSObjectRef.h: Added.
-        * ForwardingHeaders/JavaScriptCore/JSStringRef.h: Added.
-        * ForwardingHeaders/JavaScriptCore/JSStringRefCF.h: Added.
-        * ForwardingHeaders/JavaScriptCore/JSValueRef.h: Added.
-        * ForwardingHeaders/JavaScriptCore/JavaScriptCore.h: Added.
-
-2007-07-11  Holger Hans Peter Freyther  <zecke@selfish.org>
-
-        Reviewed by Darin.
-
-        As of http://bugs.webkit.org/show_bug.cgi?id=14527 move the
-        WebCore/ForwardingHeader/JavaScriptCore to JavaScriptCore
-
-        * ForwardingHeaders: Added.
-        * ForwardingHeaders/JavaScriptCore: Copied from WebCore/ForwardingHeaders/JavaScriptCore.
-
-2007-07-11  Nikolas Zimmermann  <zimmermann@kde.org>
-
-        Reviewed by Mark.
-
-        Forwardport the hash table fix from CodeGeneratorJS.pm to create_hash_table.
-        Reran run-jsc-tests, couldn't find any regressions. Suggested by Darin.
-
-        * kjs/create_hash_table:
-
-2007-07-09  Maciej Stachowiak  <mjs@apple.com>
-
-        Reviewed by Oliver.
-        
-        - JavaScriptCore part of fix for: <rdar://problem/5295734> Repro crash closing tab/window @ maps.google.com in WTF::HashSet<KJS::RuntimeObjectImp*, WTF::PtrHash<KJS::RuntimeObjectImp*>, WTF::HashTraits<KJS::RuntimeObjectImp*> >::add + 11
-        
-        * JavaScriptCore.exp: Added needed export.
-
-2007-07-06  Maciej Stachowiak  <mjs@apple.com>
-
-        Reviewed by Antti.
-
-        - <rdar://problem/5311093> JavaScriptCore fails to build with strict-aliasing warnings
-        
-        * Configurations/Base.xcconfig: Re-enable -Wstrict-aliasing
-        * bindings/jni/jni_utility.cpp:
-        (KJS::Bindings::getJNIEnv): Type-pun via a union instead of a pointer cast.
-        * wtf/HashMap.h:
-        (WTF::): Instead of doing type-punned assignments via pointer cast, do one of three things:
-        (1) assign directly w/o cast if storage type matches real type; (2) assign using cast
-        via union if type does not need reffing; (3) copy with memcpy and ref/deref manually if type
-        needs reffing. This is ok peref-wise because memcpy of a constant length gets optomized.
-        HashTraits are now expected to make ref()/deref() take the storage type, not the true type.
-        * wtf/HashSet.h:
-        (WTF::): Same basic idea.
-        * wtf/HashTable.h:
-        (WTF::): Added Assigner template for use by HashMap/HashSet. Change RefCounter to call ref()
-        and deref() via storage type, avoiding the need to
-        type-pun.
-        (WTF::RefCounter::ref): ditto
-        (WTF::RefCounter::deref): ditto
-        * wtf/HashTraits.h:
-        (WTF::): Change ref() and deref() for RefPtr HashTraits to take the storage type; cast
-        via union to pointer type.
-        * wtf/FastMalloc.cpp:
-        (WTF::TCMalloc_PageHeap::init): Changed from constructor to init function so this can go in a union.
-        (WTF::): redefine pageheap macro in terms of getPageHeap().
-        (WTF::getPageHeap): new inline function, helper for pageheap macro. This hides the cast in a union.
-        (WTF::TCMalloc_ThreadCache::InitModule): Call init() instead of using placement new to initialize page
-        heap.
-        * wtf/TCPageMap.h:
-        (TCMalloc_PageMap1::init): Changed from constructor to init function.
-        (TCMalloc_PageMap2::init): ditto
-        (TCMalloc_PageMap3::init): ditto
-
-
-2007-07-06  George Staikos  <staikos@kde.org>
-
-        Reviewed by Maciej.
-
-        Switch USE(ICONDATABASE) to ENABLE(ICONDATABASE)
-
-        * wtf/Platform.h:
-
-2007-07-03  Sam Weinig  <sam@webkit.org>
-
-        Reviewed by Darin.
-
-        Eleventh round of fixes for implicit 64-32 bit conversion errors.
-        <rdar://problem/5292262>
-
-        - Fixes a real bug where where we were setting long long and unsigned long long
-          values to a long field.
-
-        * bindings/objc/objc_utility.mm:
-        (KJS::Bindings::convertValueToObjcValue):
-
-2007-07-03  Sam Weinig  <sam@webkit.org>
-
-        Reviewed by Brady Eidson.
-
-        Tenth round of fixes for implicit 64-32 bit conversion errors.
-        <rdar://problem/5292262>
-
-        - Add explicit casts.
-
-        * kjs/dtoa.cpp:
-        (Bigint::):
-
-2007-07-02  Sam Weinig  <sam@webkit.org>
-
-        Reviewed by Kevin McCullough.
-
-        Fourth round of fixes for implicit 64-32 bit conversion errors.
-        <rdar://problem/5292262>
-
-        Add custom piDouble and piFloat constants to use instead of M_PI.
-
-        * kjs/math_object.cpp:
-        (MathObjectImp::getValueProperty):
-        * wtf/MathExtras.h:
-        (wtf_atan2):
-
-2007-06-29  Sam Weinig  <sam@webkit.org>
-
-        Reviewed by Darin.
-
-        Second pass at fixing implicit 64-32 bit conversion errors.
-        <rdar://problem/5292262>
-
-        - Add a toFloat() method to JSValue for float conversion.
-
-        * JavaScriptCore.exp:
-        * kjs/value.cpp:
-        (KJS::JSValue::toFloat):
-        * kjs/value.h:
-
-2007-06-27  Kevin McCullough  <kmccullough@apple.com>
-
-        Reviewed by Darin.
-
-        - <rdar://problem/5271937> REGRESSION: Apparent WebKit JavaScript memory smasher when submitting comment to iWeb site (crashes in kjs_pcre_compile2)
-        - Correctly evaluate the return value of _pcre_ucp_findchar.
-
-        * pcre/pcre_compile.c:
-        (compile_branch):
-        * pcre/pcre_exec.c:
-        (match):
-
-2007-06-27  Sam Weinig  <sam@webkit.org>
-
-        Reviewed by Darin.
-
-        First pass at fixing implicit 64-32 bit conversion errors.
-        <rdar://problem/5292262>
-
-        - Add 'f' suffix where necessary.
-
-        * kjs/testkjs.cpp:
-        (StopWatch::getElapsedMS):
-
-2007-06-26  Geoffrey Garen  <ggaren@apple.com>
-
-        Reviewed by Maciej Stachowiak.
-        
-        Fixed <rdar://problem/5296627> JSGarbageCollect headerdoc suggests that 
-        using JavaScriptCore requires leaking memory
-
-        * API/JSBase.h: Changed documentation to explain that you can pass NULL
-        to JSGarbageCollect.
-
-2007-06-26  Adam Treat  <adam@staikos.net>
-
-        Reviewed by Adam Roben.
-
-        Make the SQLite icon database optional.
-
-        * wtf/Platform.h:
-
-2007-06-15  George Staikos  <staikos@kde.org>
-
-        More missing files for Qt.
-
-        * JavaScriptCore.pri:
-        * kjs/testkjs.pro:
-
-2007-06-15  George Staikos  <staikos@kde.org>
-
-        Another Qt build fix.
-
-        * JavaScriptCore.pri:
-        * kjs/testkjs.pro:
-
-2007-06-15  George Staikos  <staikos@kde.org>
-
-        Fixing Qt build.
-
-        * JavaScriptCore.pri:
-
-2007-06-20  Mark Rowe  <mrowe@apple.com>
-
-        Reviewed by Mitz.
-
-        Fix http://bugs.webkit.org/show_bug.cgi?id=14244
-        Bug 14244: Data corruption when using a replace() callback function with data containing "$"
-
-        * kjs/string_object.cpp:
-        (KJS::replace):  When 'replacement' is a function, do not replace $n placeholders in its return value.
-        This matches the behaviour described in ECMA 262 3rd Ed section 15.5.4.1, and as implemented in Firefox.
-
-2007-06-14  Anders Carlsson  <andersca@apple.com>
-
-        Fix Windows build.
-        
-        * bindings/runtime_object.cpp:
-        (RuntimeObjectImp::canPut):
-
-2007-06-14  Anders Carlsson  <andersca@apple.com>
-
-        Reviewed by Darin.
-
-        <rdar://problem/5103077> 
-        Crash at _NPN_ReleaseObject when quitting page at http://eshop.macsales.com/shop/ModBook
-        
-        <rdar://problem/5183692>
-        http://bugs.webkit.org/show_bug.cgi?id=13547
-        REGRESSION: Crash in _NPN_ReleaseObject when closing Safari on nba.com (13547)
-        
-        <rdar://problem/5261499>
-        CrashTracer: [USER] 75 crashes in Safari at com.apple.JavaScriptCore: KJS::Bindings::CInstance::~CInstance + 40
-        
-        Have the root object track all live instances of RuntimeObjectImp. When invalidating 
-        the root object, also invalidate all live runtime objects by zeroing out their instance ivar.
-        This prevents instances from outliving their plug-ins which lead to crashes.
-        
-        * bindings/c/c_utility.cpp:
-        (KJS::Bindings::convertValueToNPVariant):
-        * bindings/jni/jni_jsobject.cpp:
-        (JavaJSObject::convertValueToJObject):
-        * bindings/jni/jni_utility.cpp:
-        (KJS::Bindings::convertValueToJValue):
-        * bindings/objc/objc_runtime.mm:
-        (ObjcFallbackObjectImp::callAsFunction):
-        * bindings/runtime_array.cpp:
-        (RuntimeArray::RuntimeArray):
-        * bindings/runtime_array.h:
-        (KJS::RuntimeArray::getConcreteArray):
-        * bindings/runtime_method.cpp:
-        (RuntimeMethod::callAsFunction):
-        * bindings/runtime_method.h:
-        * bindings/runtime_object.cpp:
-        (RuntimeObjectImp::RuntimeObjectImp):
-        (RuntimeObjectImp::~RuntimeObjectImp):
-        (RuntimeObjectImp::invalidate):
-        (RuntimeObjectImp::fallbackObjectGetter):
-        (RuntimeObjectImp::fieldGetter):
-        (RuntimeObjectImp::methodGetter):
-        (RuntimeObjectImp::getOwnPropertySlot):
-        (RuntimeObjectImp::put):
-        (RuntimeObjectImp::canPut):
-        (RuntimeObjectImp::defaultValue):
-        (RuntimeObjectImp::implementsCall):
-        (RuntimeObjectImp::callAsFunction):
-        (RuntimeObjectImp::getPropertyNames):
-        (RuntimeObjectImp::throwInvalidAccessError):
-        * bindings/runtime_object.h:
-        * bindings/runtime_root.cpp:
-        (KJS::Bindings::RootObject::invalidate):
-        (KJS::Bindings::RootObject::addRuntimeObject):
-        (KJS::Bindings::RootObject::removeRuntimeObject):
-        * bindings/runtime_root.h:
-
-2007-06-14  Anders Carlsson  <andersca@apple.com>
-
-        Reviewed by Mitz.
-
-        <rdar://problem/5244948>
-        Safari keeps on complaining about slow script playing NBC TV video (14133)
-
-        http://bugs.webkit.org/show_bug.cgi?id=14133
-        Runaway JavaScript timer fires when spinning around in Google Maps street view
-
-        Make sure to start and stop the timeout checker around calls to JS.
-        
-        * bindings/NP_jsobject.cpp:
-        (_NPN_InvokeDefault):
-        (_NPN_Invoke):
-        (_NPN_Evaluate):
-        * bindings/jni/jni_jsobject.cpp:
-        (JavaJSObject::call):
-        (JavaJSObject::eval):
-
-2007-06-13  Darin Adler  <darin@apple.com>
-
-        Reviewed by Mark Rowe.
-
-        - fix http://bugs.webkit.org/show_bug.cgi?id=14132
-          array sort with > 10000 elements sets elements > 10000 undefined
-
-        Test: fast/js/sort-large-array.html
-
-        * kjs/array_instance.h: Replaced pushUndefinedObjectsToEnd with
-        compactForSorting, and removed ExecState parameters.
-
-        * kjs/array_object.cpp:
-        (ArrayInstance::sort): Changed to call compactForSorting.
-        (ArrayInstance::compactForSorting): Do the get and delete of the
-        properties directly on the property map instead of using public
-        calls from JSObject. The public calls would just read the undefined
-        values from the compacted sort results array!
-
-2007-06-13  George Staikos  <staikos@kde.org>
-
-        Reviewed by Lars.
-
-        Fix Mac OS X build after last checkin.
-
-        * wtf/FastMalloc.h:
-
-2007-06-14  Lars Knoll <lars@trolltech.com>
-
-        Reviewed by Maciej.
-
-        Disable FastMalloc for the Qt build and make sure we
-        don't reimplement the global new/delete operators
-        when using the system malloc.
-
-        * wtf/FastMalloc.cpp:
-        * wtf/FastMalloc.h:
-        * wtf/Platform.h:
-
-2007-06-13  Anders Carlsson  <andersca@apple.com>
-
-        Reviewed by Geoff.
-
-        Make sure that bindings instances get correct root objects.
-        
-        * JavaScriptCore.exp:
-        * bindings/NP_jsobject.cpp:
-        (listFromVariantArgs):
-        (_NPN_InvokeDefault):
-        (_NPN_Invoke):
-        (_NPN_SetProperty):
-        * bindings/c/c_instance.cpp:
-        (KJS::Bindings::CInstance::invokeMethod):
-        (KJS::Bindings::CInstance::invokeDefaultMethod):
-        * bindings/c/c_runtime.cpp:
-        (KJS::Bindings::CField::valueFromInstance):
-        * bindings/c/c_utility.cpp:
-        (KJS::Bindings::convertNPVariantToValue):
-        * bindings/c/c_utility.h:
-        * bindings/objc/objc_instance.mm:
-        (ObjcInstance::invokeMethod):
-        (ObjcInstance::invokeDefaultMethod):
-        (ObjcInstance::getValueOfUndefinedField):
-        * bindings/objc/objc_runtime.mm:
-        (ObjcField::valueFromInstance):
-        (ObjcArray::valueAt):
-        * bindings/objc/objc_utility.h:
-        * bindings/objc/objc_utility.mm:
-        (KJS::Bindings::convertObjcValueToValue):
-        * bindings/runtime.h:
-
-2007-06-13  Simon Hausmann  <hausmann@kde.org>
-
-        Reviewed by Lars.
-
-        * kjs/testkjs.pro: WebKitQt is now called QtWebKit.
-
-2007-06-12  Anders Carlsson  <andersca@apple.com>
-
-        Another build fix.
-        
-        * bindings/qt/qt_instance.cpp:
-        (KJS::Bindings::QtInstance::invokeMethod):
-
-2007-06-12  Anders Carlsson  <andersca@apple.com>
-
-        Reviewed by Geoff.
-
-        Move the notion of field type to the JNI runtime since that's the only 
-        one that was actually using it.
-        
-        * bindings/c/c_runtime.h:
-        (KJS::Bindings::CField::CField):
-        * bindings/jni/jni_runtime.h:
-        * bindings/objc/objc_runtime.h:
-        * bindings/objc/objc_runtime.mm:
-        * bindings/qt/qt_runtime.h:
-        * bindings/runtime.h:
-        * bindings/runtime_method.cpp:
-
-2007-06-12  Anders Carlsson  <andersca@apple.com>
-
-        Build fix.
-        
-        * bindings/qt/qt_class.cpp:
-        (KJS::Bindings::QtClass::methodsNamed):
-        * bindings/qt/qt_instance.cpp:
-        (KJS::Bindings::QtInstance::invokeMethod):
-
-2007-06-12  Anders Carlsson  <andersca@apple.com>
-
-        Reviewed by Oliver.
-        
-        Get rid of the MethodList class and use a good ol' Vector instead.
-
-        * bindings/c/c_class.cpp:
-        (KJS::Bindings::CClass::methodsNamed):
-        * bindings/c/c_instance.cpp:
-        (KJS::Bindings::CInstance::invokeMethod):
-        * bindings/jni/jni_class.cpp:
-        (JavaClass::JavaClass):
-        (JavaClass::~JavaClass):
-        * bindings/jni/jni_instance.cpp:
-        (JavaInstance::invokeMethod):
-        * bindings/objc/objc_class.mm:
-        (KJS::Bindings::ObjcClass::methodsNamed):
-        * bindings/objc/objc_instance.mm:
-        (ObjcInstance::invokeMethod):
-        * bindings/objc/objc_runtime.mm:
-        (ObjcFallbackObjectImp::callAsFunction):
-        * bindings/runtime.cpp:
-        * bindings/runtime.h:
-        * bindings/runtime_method.cpp:
-        (RuntimeMethod::lengthGetter):
-        (RuntimeMethod::callAsFunction):
-        * bindings/runtime_object.cpp:
-        (RuntimeObjectImp::getOwnPropertySlot):
-
-2007-06-12  Anders Carlsson  <andersca@apple.com>
-
-        Reviewed by Geoff.
-
-        Make RuntimeMethod's method list a pointer so that the object size doesn't
-        grow beyond 32 bytes when we later will replace MethodList with a Vector.
-        
-        * bindings/runtime_method.cpp:
-        (RuntimeMethod::RuntimeMethod):
-        (RuntimeMethod::lengthGetter):
-        (RuntimeMethod::callAsFunction):
-        * bindings/runtime_method.h:
-
-2007-06-12  Anders Carlsson  <andersca@apple.com>
-
-        Reviewed by Geoff.
-
-        Get rid of the Parameter class.
-        
-        * bindings/jni/jni_instance.cpp:
-        (JavaInstance::invokeMethod):
-        * bindings/jni/jni_runtime.cpp:
-        (JavaMethod::signature):
-        * bindings/jni/jni_runtime.h:
-        (KJS::Bindings::JavaParameter::JavaParameter):
-        (KJS::Bindings::JavaParameter::~JavaParameter):
-        (KJS::Bindings::JavaParameter::type):
-        (KJS::Bindings::JavaMethod::parameterAt):
-        (KJS::Bindings::JavaMethod::numParameters):
-        * bindings/runtime.h:
-
-2007-06-12  Anders Carlsson  <andersca@apple.com>
-
-        Build fix.
-        
-        * bindings/qt/qt_class.h:
-
-2007-06-12  Mark Rowe  <mrowe@apple.com>
-
-        Build fix.
-
-        * bindings/objc/objc_runtime.h:
-
-2007-06-12  Anders Carlsson  <andersca@apple.com>
-
-        Reviewed by Geoff.
-
-        Get rid of Constructor and its only subclass JavaConstructor.
-        
-        * bindings/c/c_class.h:
-        * bindings/jni/jni_class.cpp:
-        (JavaClass::JavaClass):
-        (JavaClass::~JavaClass):
-        * bindings/jni/jni_class.h:
-        * bindings/jni/jni_runtime.cpp:
-        * bindings/jni/jni_runtime.h:
-        * bindings/objc/objc_class.h:
-        * bindings/runtime.h:
-
-2007-06-12  Anders Carlsson  <andersca@apple.com>
-
-        Reviewed by Geoff.
-
-        Use RetainPtr throughout the bindings code.
-        
-        * bindings/objc/objc_class.h:
-        * bindings/objc/objc_class.mm:
-        (KJS::Bindings::ObjcClass::ObjcClass):
-        (KJS::Bindings::ObjcClass::methodsNamed):
-        (KJS::Bindings::ObjcClass::fieldNamed):
-        * bindings/objc/objc_instance.h:
-        (KJS::Bindings::ObjcInstance::getObject):
-        * bindings/objc/objc_instance.mm:
-        (ObjcInstance::ObjcInstance):
-        (ObjcInstance::~ObjcInstance):
-        (ObjcInstance::implementsCall):
-        (ObjcInstance::invokeMethod):
-        (ObjcInstance::invokeDefaultMethod):
-        (ObjcInstance::defaultValue):
-        * bindings/objc/objc_runtime.h:
-        (KJS::Bindings::ObjcMethod::setJavaScriptName):
-        (KJS::Bindings::ObjcMethod::javaScriptName):
-        (KJS::Bindings::ObjcArray::getObjcArray):
-        * bindings/objc/objc_runtime.mm:
-        (ObjcField::name):
-        (ObjcArray::ObjcArray):
-        (ObjcArray::setValueAt):
-        (ObjcArray::valueAt):
-        (ObjcArray::getLength):
-        * wtf/RetainPtr.h:
-
-2007-06-12  Anders Carlsson  <andersca@apple.com>
-
-        Reviewed by Maciej.
-
-        Have JSCell inherit from Noncopyable.
-        
-        * bindings/objc/objc_runtime.h:
-        * bindings/runtime_object.h:
-        * kjs/value.h:
-
-2007-06-12  Anders Carlsson  <andersca@apple.com>
-
-        Reviewed by Darin and Maciej.
-
-        More cleanup. Use our Noncopyable WTF class, add a root object member
-        to the Array class.
-        
-        * bindings/c/c_class.h:
-        * bindings/jni/jni_class.h:
-        * bindings/jni/jni_instance.h:
-        * bindings/jni/jni_runtime.cpp:
-        (JavaArray::JavaArray):
-        * bindings/jni/jni_runtime.h:
-        * bindings/objc/objc_class.h:
-        * bindings/objc/objc_runtime.h:
-        * bindings/objc/objc_runtime.mm:
-        (ObjcArray::ObjcArray):
-        * bindings/objc/objc_utility.mm:
-        (KJS::Bindings::convertObjcValueToValue):
-        * bindings/runtime.cpp:
-        (KJS::Bindings::Array::Array):
-        (KJS::Bindings::Array::~Array):
-        * bindings/runtime.h:
-        * bindings/runtime_object.h:
-        * bindings/runtime_root.h:
-
-2007-06-08  Zack Rusin  <zrusin@trolltech.com>
-
-        Fix the Qt build
-
-        * bindings/qt/qt_instance.cpp:
-        (KJS::Bindings::QtInstance::QtInstance):
-        * bindings/qt/qt_instance.h:
-
-2007-06-07  Anders Carlsson  <andersca@apple.com>
-
-        Reviewed by Geoff.
-
-        Get rid of Instance::setRootObject and pass the root object to the instance constructor instead.
-        
-        * bindings/c/c_instance.cpp:
-        (KJS::Bindings::CInstance::CInstance):
-        * bindings/c/c_instance.h:
-        * bindings/jni/jni_instance.cpp:
-        (JavaInstance::JavaInstance):
-        * bindings/jni/jni_instance.h:
-        * bindings/jni/jni_jsobject.cpp:
-        (JavaJSObject::convertJObjectToValue):
-        * bindings/objc/objc_instance.h:
-        * bindings/objc/objc_instance.mm:
-        (ObjcInstance::ObjcInstance):
-        * bindings/runtime.cpp:
-        (KJS::Bindings::Instance::Instance):
-        (KJS::Bindings::Instance::createBindingForLanguageInstance):
-        * bindings/runtime.h:
-
-2007-06-07  Anders Carlsson  <andersca@apple.com>
-
-        Reviewed by Adam.
-
-        Don't use a JavaInstance to store the field when all we want to do is to keep the field
-        from being garbage collected. Instead, use a JObjectWrapper.
-        
-        * bindings/jni/jni_instance.h:
-        * bindings/jni/jni_runtime.cpp:
-        (JavaField::JavaField):
-        (JavaField::dispatchValueFromInstance):
-        (JavaField::dispatchSetValueToInstance):
-        * bindings/jni/jni_runtime.h:
-        (KJS::Bindings::JavaField::JavaField):
-        (KJS::Bindings::JavaField::operator=):
-
-2007-05-30  Alp Toker  <alp.toker@collabora.co.uk>
-
-        Reviewed by Brady.
-
-        Enable logging in the Gdk port.
-        http://bugs.webkit.org/show_bug.cgi?id=13936
-
-        * wtf/Assertions.cpp:
-        * wtf/Assertions.h: Add WTFLogVerbose which also logs
-        the file, line number and function.
-
-2007-05-30  Mark Rowe  <mrowe@apple.com>
-
-        Mac build fix.  Update #include.
-
-        * API/JSCallbackFunction.h:
-
-2007-05-30  Luciano Montanaro  <mikelima@cirulla.net>
-
-        Reviewed by Maciej.
-
-        - cross-port Harri Porten's commits 636099 and 636108 from KJS: 
-        "publish a class anyway public already" and "class is being used from
-        outside for quite some time" in preparation for further syncronizations
-
-        * kjs/context.h:
-        * kjs/date_object.cpp:
-        * kjs/date_object.h:
-        * kjs/function.h:
-        (KJS::):
-        (KJS::InternalFunctionImp::classInfo):
-        (KJS::InternalFunctionImp::functionName):
-        * kjs/function_object.h:
-        * kjs/internal.h:
-        * kjs/lookup.h:
-        (KJS::getStaticPropertySlot):
-        (KJS::getStaticFunctionSlot):
-        (KJS::getStaticValueSlot):
-        * kjs/object_object.h:
-
-2007-05-29  Sam Weinig  <sam@webkit.org>
-
-        Reviewed by Adam Roben.
-
-        Cleanup function and fix to match comparison API.
-
-        * kjs/string_object.cpp:
-        (KJS::substituteBackreferences):
-        (KJS::localeCompare):
-
-2007-05-28  Geoffrey Garen  <ggaren@apple.com>
-
-        Slight clarification to an exception message.
-
-        * API/JSCallbackObject.cpp:
-        (KJS::JSCallbackObject::put):
-
-2007-05-27  Holger Freyther  <zecke@selfish.org>
-
-        Reviewed by Mark Rowe.
-
-        * wtf/Platform.h: Move Gdk up to allow building WebKit/Gdk on Darwin
-
-2007-05-27  Darin Adler  <darin@apple.com>
-
-        - fix a couple ifdefs that said WIN instead of WIN_OS
-
-        * kjs/collector.cpp:
-        (KJS::allocateBlock): WIN -> WIN_OS
-        (KJS::freeBlock): Ditto.
-
-2007-05-26  Sam Weinig  <sam@webkit.org>
-
-        Reviewed by Darin.
-
-        Patch for http://bugs.webkit.org/show_bug.cgi?id=13854
-        Port of commit 667785 from kjs
-
-        - special case calling String.localeCompare() with no parameters to return 0.
-
-        * kjs/string_object.cpp:
-        (KJS::StringProtoFunc::callAsFunction):
-
-2007-05-25  Kimmo Kinnunen  <kimmok@iki.fi>
-
-        Reviewed by Darin.
-
-        - Fix for http://bugs.webkit.org/show_bug.cgi?id=13456
-        REGRESSION: setTimeout "arguments" object gets shadowed by a local variable
-
-        - Add a explicit check for arguments. Previously check was done with getDirect,
-        but since the arguments is created on-demand in ActivationImp, it doesn't
-        show up in the test. 'arguments' should always be in the VarDeclNode's
-        evaluation scope.
-
-        * kjs/nodes.cpp:
-        (VarDeclNode::evaluate): Additional check if the var decl identifier is 'arguments'
-
-2007-05-25  George Staikos  <staikos@kde.org>
-
-        Reviewed by Maciej.
-
-        - Use COMPILER(GCC), not PLATFORM(GCC) - as Platform.h defines
-
-        * wtf/FastMalloc.h:
-
-2007-05-25  Kimmo Kinnunen  <kimmok@iki.fi>
-
-        Reviewed by Darin.
-
-        - http://bugs.webkit.org/show_bug.cgi?id=13623 (Decompilation of function
-          doesn't compile with "++(x,y)")
-        - Create the error node based on the actual node, not the node inside
-          parenthesis
-        - Fix applies to postfix, prefix and typeof operators
-        - Produces run-time ReferenceError like other non-lvalue assignments etc.
-
-        * kjs/grammar.y: Create {Prefix,Postfix}ErrorNode based on the actual node,
-        not the based on the node returned by "nodeInsideAllParens()". Same for
-        TypeOfValueNode.
-
-2007-05-25  Simon Hausmann  <hausmann@kde.org>
-
-        Reviewed by Zack.
-
-        Fix crash in Qt JavaScript bindings when the arguments used on the Qt side are not
-        registered with QMetaType.
-
-        * bindings/qt/qt_instance.cpp:
-        (KJS::Bindings::QtInstance::invokeMethod):
-        * bindings/qt/qt_runtime.cpp:
-
-2007-05-24  Luciano Montanaro  <mikelima@cirulla.net>
-
-        Reviewed by Darin
-
-        Patch for http://bugs.webkit.org/show_bug.cgi?id=13855
-        Port patch 666176 to JavaScriptCore
-
-        - Renamed JSValue::downcast() to JSValue::asCell() which makes the
-        function meaning cleaner. It's modeled after Harri Porten change in 
-        KDE trunk.
-
-        * kjs/collector.cpp:
-        (KJS::Collector::protect):
-        (KJS::Collector::unprotect):
-        (KJS::Collector::collectOnMainThreadOnly):
-        * kjs/object.h:
-        (KJS::JSValue::isObject):
-        * kjs/string_object.cpp:
-        (KJS::StringProtoFunc::callAsFunction):
-        * kjs/value.h:
-        (KJS::JSValue::asCell):
-        (KJS::JSValue::isNumber):
-        (KJS::JSValue::isString):
-        (KJS::JSValue::isObject):
-        (KJS::JSValue::getNumber):
-        (KJS::JSValue::getString):
-        (KJS::JSValue::getObject):
-        (KJS::JSValue::getUInt32):
-        (KJS::JSValue::mark):
-        (KJS::JSValue::marked):
-        (KJS::JSValue::type):
-        (KJS::JSValue::toPrimitive):
-        (KJS::JSValue::toBoolean):
-        (KJS::JSValue::toNumber):
-        (KJS::JSValue::toString):
-        (KJS::JSValue::toObject):
-
-2007-05-18  Holger Hans Peter Freyther  <zecke@selfish.org>
-
-        Reviewed by Mark Rowe.
-
-        * kjs/testkjs.pro: Make the Gdk port link to icu
-
-2007-05-15  Geoffrey Garen  <ggaren@apple.com>
-
-        Reviewed by Adele Peterson.
-        
-        It helps if you swap the right variable.
-
-        * wtf/HashSet.h:
-        (WTF::::operator):
-
-2007-05-15  Lars Knoll <lars@trolltech.com>
-
-        Reviewed by Zack
-
-        Extend the QObject JavaScript bindings to work for slots with
-        arguments.
-
-        * bindings/qt/qt_instance.cpp:
-        (KJS::Bindings::QtInstance::invokeMethod):
-
-2007-05-14  Kimmo Kinnunen  <kimmok@iki.fi>
-
-        Reviewed by Darin.
-
-        - Fixes http://bugs.webkit.org/show_bug.cgi?id=13622 (Decompiler
-          omits trailing comma in array literal)
-
-         * kjs/nodes2string.cpp:
-         (ArrayNode::streamTo): print extra ',' in case there was elision
-         commas (check opt member var) and array elements present
-         in the array expression
-
-2007-05-14  Geoffrey Garen  <ggaren@apple.com>
-
-        Reviewed by Oliver Hunt.
-        
-        Added HashMap::swap and HashSet::swap. WebCore now uses HashSet::swap.
-        I figured while I was in the neighborhood I might as well add HashMap::swap,
-        too.
-
-        * wtf/HashMap.h:
-        (WTF::::operator):
-        (WTF::::swap):
-        * wtf/HashSet.h:
-        (WTF::::operator):
-        (WTF::::swap):
-
-2007-05-11  Kimmo Kinnunen  <kimmok@iki.fi>
-
-        Reviewed by Darin.
-
-        - Fix for bug http://bugs.webkit.org/show_bug.cgi?id=13620
-          Bogus decompilation of "for (var j = 1 in [])"
-        - ForInNode toString()'ed to syntax error if there was var decl
-          and initializer
-        - ForNode toStringed()'ed lost 'var ' if it was present
-
-        * kjs/nodes2string.cpp:
-        (VarDeclListNode::streamTo): Print "var " here
-        (VarStatementNode::streamTo): Don't print "var " here
-        (ForNode::streamTo): Remove TODO comment, VarDeclListNode will
-        stream the "var "
-        (ForInNode::streamTo): ForIn initializer is printed by VarDeclNode
-
-2007-05-11  Kimmo Kinnunen  <kimmok@iki.fi>
-
-        Reviewed by Darin.
-
-        - Fixes http://bugs.webkit.org/show_bug.cgi?id=10878
-          (Incorrect decompilation for "4..x")
-        - Group numbers in dotted expressions in toString() output, so we
-          avoid the 4.x constructs  when the original input is 4..x.
-          4..x means the same as 4. .x or (4).x or Number(4).x
-
-        * kjs/nodes2string.cpp:
-        (KJS::SourceStream::):
-        Add boolean flag to indicate that if next item is a number, it should be grouped.
-        Add new formatting enum which turns on the boolean flag.
-        (KJS::SourceStream::SourceStream): Added. Initialize the flag.
-        (SourceStream::operator<<): Added. New overloaded operator with double value as parameter.
-        (NumberNode::streamTo): Use the double operator
-        (ArrayNode::streamTo):
-        (DotAccessorNode::streamTo):
-        (FunctionCallDotNode::streamTo):
-        (FunctionCallParenDotNode::streamTo):
-        (PostfixDotNode::streamTo):
-        (DeleteDotNode::streamTo):
-        (PrefixDotNode::streamTo):
-        (AssignDotNode::streamTo): Use the new formatting enum to turn on the grouping flag.
-
-2007-05-10  Lars Knoll <lars@trolltech.com>
-
-        Reviewed by Zack
-
-        Fix our last three test failures in the JavaScript
-        tests.
-
-        * wtf/unicode/qt4/UnicodeQt4.h:
-        (WTF::Unicode::toLower):
-        (WTF::Unicode::toUpper):
-
-2007-05-08  Geoffrey Garen  <ggaren@apple.com>
-
-        Reviewed by Darin Adler.
-        
-        Fixed #includes of JSStringRefCF.h and use of CF datatypes. I think I 
-        misunderstood this issue before.
-
-        * API/JavaScriptCore.h: #include JSStringRefCF.h. Platforms that don't
-        want this behavior can just #include individual headers, instead of the
-        umbrella framework header. But we definitely want Mac OS X clients to
-        get the #include of JSStringRefCF.h "for free."
-        * API/minidom.c: Don't #include JSStringRefCF.h. (Don't need to #include
-        JavaScriptCore.h, either.)
-        * API/testapi.c: Don't #include JSStringRefCF.h. Do use CF datatypes
-        regardless of whether __APPLE__ is defined. Platforms that don't support
-        CF just shouldn't compile this file.
-        (main):
-
-2007-05-09  Eric Seidel  <eric@webkit.org>
-
-        Reviewed by mjs.
-        
-        http://bugs.webkit.org/show_bug.cgi?id=6985
-        Cyclic __proto__ values cause WebKit to hang
-
-        * kjs/object.cpp:
-        (KJS::JSObject::put): do a cycle check before setting __proto__
-
-2007-05-08  Kimmo Kinnunen  <kimmok@iki.fi>
-
-        Reviewed by darin.  Landed by eseidel.
-
-        - http://bugs.webkit.org/show_bug.cgi?id=10880 (Do..while loop gains 
-        a semicolon each time it is toStringed)
-        Grammar in Ecma-66262, 12.6: "do Statement while ( Expression );"
-        EmptyStatement was created after every do..while(expr) which
-        had semicolon at the end.
-
-        * kjs/grammar.y: Require semicolon at the end of do..while
-
-2007-05-08  Geoffrey Garen  <ggaren@apple.com>
-
-        Build fix -- this time for sure.
-        
-        APICast.h, being private, ends up in a different folder than JSValueRef.h,
-        so we can't include one from the other using "". Instead, just forward
-        declare the relevant data types.
-
-        * API/APICast.h:
-
-2007-05-08  Geoffrey Garen  <ggaren@apple.com>
-
-        Build fix: export APICast.h for WebCore and WebKit.
-
-        * JavaScriptCore.xcodeproj/project.pbxproj:
-
-2007-05-04  Darin Adler  <darin@apple.com>
-
-        Reviewed by Adele.
-
-        - fix http://bugs.webkit.org/show_bug.cgi?id=12821
-          <rdar://problem/5007921> Number.toExponential doesn't work for negative numbers
-
-        * kjs/number_object.cpp: (NumberProtoFunc::callAsFunction):
-        Added a call to fabs before calling log10.
-
-2007-05-03  Holger Freyther <freyther@kde.org>
-
-        Reviewed by Zack, landed by Simon.
-        This is bugzilla bug 13499.
-
-        * JavaScriptCore.pri: Place Qt into the qt-port scope
-        * bindings/testbindings.pro: Place Qt into the qt-port scope
-        * kjs/testkjs.pro: Place Qt into the qt-port scope
-        * pcre/pcre.pri: Place Qt into the qt-port scope
-
-2007-05-02  David Harrison  <harrison@apple.com>
-
-        Reviewed by Antti.
-
-        <rdar://problem/5174862> Crash resulting from DeprecatedString::insert()
-
-        Added insertion support for more than one value.
-        
-        * wtf/Vector.h:
-        (WTF::::insert):
-        Added support for inserting multiple values.
-        
-        (WTF::::prepend):
-        New. Insert at the start of vectors. Convenient for vectors used as strings.
-
-2007-05-01  Jungshik Shin  <jungshik.shin@gmail.com>
-
-        Reviewed by Alexey.
-
-        - get rid of non-ASCII lteral characters : suppress compiler warnings
-        http://bugs.webkit.org/show_bug.cgi?id=13551
-         
-        * kjs/testkjs.cpp:
-        * pcre/pcre_compile.c:
-
-2007-04-28  Jungshik Shin  <jungshik.shin@gmail.com>
-
-        Reviewed by Sam Weinig.
-
-        - Replace copyright sign in Latin-1 (0xA9) with '(C)'
-        http://bugs.webkit.org/show_bug.cgi?id=13531
-
-        * bindings/npruntime.h:
-    
-2007-04-28  Darin Adler  <darin@apple.com>
-
-        Reviewed by Maciej.
-
-        - fix <rdar://problem/5154144> Hamachi test fails: assertion failure in ListHashSet
-
-        Test: fast/forms/add-remove-form-elements-stress-test.html
-
-        * wtf/ListHashSet.h:
-        (WTF::ListHashSetNodeAllocator::ListHashSetNodeAllocator): Initialize
-        m_isDoneWithInitialFreeList to false.
-        (WTF::ListHashSetNodeAllocator::allocate): Added assertions based on a debug-only
-        m_isAllocated flag that make sure we don't allocate a block that's already allocated.
-        These assertions helped pinpoint the bug. Set m_isDoneWithInitialFreeList when we
-        allocate the last block of the initial free list. Once we're done with the initial
-        free list, turn off the rule that says that the next node in the pool after the last
-        node in the free list is also free. This rule works because any free nodes are added
-        to the head of the free list, so a node that hasn't been allocated even once is always
-        at the tail of the free list and all the nodes after it also haven't been allocated
-        even once. But it doesn't work any longer once the entire pool has been used at least
-        once, because there's nothing special about the last node on the free list any more.
-        (WTF::ListHashSetNodeAllocator::deallocate): Set the node's m_isAllocated to false.
-        (WTF::ListHashSetNodeAllocator::pastPool): Added. Used above.
-        (WTF::ListHashSetNodeAllocator::inPool): Changed to use the pastPool function.
-        (WTF::ListHashSetNode::ListHashSetNode): Initialize m_isAllocated to true.
-        (WTF::ListHashSetNode::operator new): Removed variable name for unused size
-        parameter.
-        (WTF::ListHashSetNode::destroy): Changed to call the destructor rather than
-        delete -- this gets rid of the need to define an operator delete.
-
-2007-04-27  Christopher Brichford  <chrisb@adobe.com>
-
-        Reviewed by Timothy Hatcher.
-
-        Fix for: Bug 13211: Move JavaScriptCore mac project files for apollo port
-        http://bugs.webkit.org/show_bug.cgi?id=13211
-
-        * JavaScriptCore.apolloproj/mac/JavaScriptCore.Debug.xcconfig: Added.
-        * JavaScriptCore.apolloproj/mac/JavaScriptCore.Release.xcconfig: Added.
-        * JavaScriptCore.apolloproj/mac/JavaScriptCore.xcconfig: Added.
-        * JavaScriptCore.apolloproj/mac/JavaScriptCore.xcodeproj/project.pbxproj: Added.
-        * JavaScriptCore.apolloproj/mac/JavaScriptCore/JavaScriptCore.Debug.xcconfig: Removed.
-        * JavaScriptCore.apolloproj/mac/JavaScriptCore/JavaScriptCore.Release.xcconfig: Removed.
-        * JavaScriptCore.apolloproj/mac/JavaScriptCore/JavaScriptCore.xcconfig: Removed.
-        * JavaScriptCore.apolloproj/mac/JavaScriptCore/JavaScriptCore.xcodeproj/project.pbxproj: Removed.
-
-2007-04-27  Holger Freyther <freyther@kde.org>
-
-        Reviewed by Maciej.
-
-        Remove unmaintained CMake build system.
-
-        * CMakeLists.txt: Removed.
-        * pcre/CMakeLists.txt: Removed.
-
-2007-04-27  Mark Rowe  <mrowe@apple.com>
-
-        Reviewed by Oliver.
-
-        * JavaScriptCore.xcodeproj/project.pbxproj: Improve dependencies in Xcode project
-        by marking dftables as a dependency of Generate Derived Sources rather than of
-        JavaScriptCore itself.
-
-2007-04-26  Geoffrey Garen  <ggaren@apple.com>
-
-        Build fix -- added #includes that we used to get implicitly through
-        JSStringRef.h.
-
-        * API/JSNode.c:
-        * API/JSNodeList.c:
-        * API/minidom.c:
-        * API/testapi.c:
-
-2007-04-26  Geoffrey Garen  <ggaren@apple.com>
-
-        Reviewed by Maciej Stachowiak, Adam Roben.
-        
-        Fixed 
-        <rdar://problem/4885130> Remove #include of JSStringRefCF.h from JSStringRef.h
-        <rdar://problem/4885123> JavaScriptCore is not cross-platform -- JSStringRef.h references CF datatypes
-
-        * API/JSStringRef.h: Removed #include -- no clients need it anymore.
-
-2007-04-25  David Kilzer  <ddkilzer@apple.com>
-
-        Reviewed by Maciej.
-
-        Add assertions for debug builds.
-
-        * kjs/JSLock.cpp:
-        (KJS::JSLock::lock): Assert the return value of pthread_mutex_lock() in debug builds.
-        (KJS::JSLock::unlock): Assert the return value of pthread_mutex_unlock() in debug builds.
-
-2007-04-25  Maciej Stachowiak  <mjs@apple.com>
-
-        Reviewed by Anders.
-        
-        - fix build problems
-
-        * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: Disable warning that
-        gives often downright incorrect results based on guessing what will happen in 64-bit.
-
-2007-04-25  Darin Adler  <darin@apple.com>
-
-        Reviewed by Geoff.
-
-        - tweak the allocator for a small speedup -- Shark showed this was a win, but I can't
-          measure an improvement right now, but it's also clear these changes do no harm
-
-        * wtf/FastMalloc.cpp:
-        (WTF::LgFloor): Use ALWAYS_INLINE here; in testing I did a while back this was necessary
-        to get this single-instruction function to be inlined.
-        (WTF::SizeClass): Use ALWAYS_INLINE here too for the same reason. Also change the special
-        case for a size of 0 to work without a branch for a bit of extra speed.
-        (WTF::ByteSizeForClass): Use ALWAYS_INLINE here too for the same reason.
-
-2007-04-24  Maciej Stachowiak  <mjs@apple.com>
-
-        Reviewed by Oliver.
-        
-        - use custom calling convention for everything in nodes.cpp on intel gcc for 1.5% speed boost
-
-        Nearly all functions in nodes.cpp were marked up to use the
-        regparm(3) calling convention under GCC for x86, since this is
-        faster and they are all guaranteed to be called only internally to
-        kjs.
-        
-        The only exception is destructors, since delete doesn't know how to use a custom calling convention.
-        
-        * kjs/nodes.cpp:
-        (dotExprDoesNotAllowCallsString):
-        * kjs/nodes.h:
-        (KJS::Node::):
-        (KJS::StatementNode::):
-        (KJS::NullNode::):
-        (KJS::BooleanNode::):
-        (KJS::NumberNode::):
-        (KJS::StringNode::):
-        (KJS::RegExpNode::):
-        (KJS::ThisNode::):
-        (KJS::ResolveNode::):
-        (KJS::GroupNode::):
-        (KJS::ElementNode::):
-        (KJS::ArrayNode::):
-        (KJS::PropertyNameNode::):
-        (KJS::PropertyNode::):
-        (KJS::PropertyListNode::):
-        (KJS::ObjectLiteralNode::):
-        (KJS::BracketAccessorNode::):
-        (KJS::DotAccessorNode::):
-        (KJS::ArgumentListNode::):
-        (KJS::ArgumentsNode::):
-        (KJS::NewExprNode::):
-        (KJS::FunctionCallValueNode::):
-        (KJS::FunctionCallResolveNode::):
-        (KJS::FunctionCallBracketNode::):
-        (KJS::FunctionCallParenBracketNode::):
-        (KJS::FunctionCallDotNode::):
-        (KJS::FunctionCallParenDotNode::):
-        (KJS::PostfixResolveNode::):
-        (KJS::PostfixBracketNode::):
-        (KJS::PostfixDotNode::):
-        (KJS::PostfixErrorNode::):
-        (KJS::DeleteResolveNode::):
-        (KJS::DeleteBracketNode::):
-        (KJS::DeleteDotNode::):
-        (KJS::DeleteValueNode::):
-        (KJS::VoidNode::):
-        (KJS::TypeOfResolveNode::):
-        (KJS::TypeOfValueNode::):
-        (KJS::PrefixResolveNode::):
-        (KJS::PrefixBracketNode::):
-        (KJS::PrefixDotNode::):
-        (KJS::PrefixErrorNode::):
-        (KJS::UnaryPlusNode::):
-        (KJS::NegateNode::):
-        (KJS::BitwiseNotNode::):
-        (KJS::LogicalNotNode::):
-        (KJS::MultNode::):
-        (KJS::AddNode::):
-        (KJS::ShiftNode::):
-        (KJS::RelationalNode::):
-        (KJS::EqualNode::):
-        (KJS::BitOperNode::):
-        (KJS::BinaryLogicalNode::):
-        (KJS::ConditionalNode::):
-        (KJS::AssignResolveNode::):
-        (KJS::AssignBracketNode::):
-        (KJS::AssignDotNode::):
-        (KJS::AssignErrorNode::):
-        (KJS::CommaNode::):
-        (KJS::AssignExprNode::):
-        (KJS::VarDeclListNode::):
-        (KJS::VarStatementNode::):
-        (KJS::EmptyStatementNode::):
-        (KJS::ExprStatementNode::):
-        (KJS::IfNode::):
-        (KJS::DoWhileNode::):
-        (KJS::WhileNode::):
-        (KJS::ForNode::):
-        (KJS::ContinueNode::):
-        (KJS::BreakNode::):
-        (KJS::ReturnNode::):
-        (KJS::WithNode::):
-        (KJS::LabelNode::):
-        (KJS::ThrowNode::):
-        (KJS::TryNode::):
-        (KJS::ParameterNode::):
-        (KJS::Parameter::):
-        (KJS::FunctionBodyNode::):
-        (KJS::FuncExprNode::):
-        (KJS::FuncDeclNode::):
-        (KJS::SourceElementsNode::):
-        (KJS::CaseClauseNode::):
-        (KJS::ClauseListNode::):
-        (KJS::SwitchNode::):
-
-2007-04-24  Oliver Hunt  <oliver@apple.com>
-
-        GTK Build fix, ::findEntry->KJS::findEntry
-
-        * kjs/lookup.cpp:
-        (KJS::Lookup::findEntry):
-        (KJS::Lookup::find):
-
-2007-04-23  Maciej Stachowiak  <mjs@apple.com>
-
-        Reviewed by Geoff.
-        
-        - compile most of JavaScriptCore as one file for 4% JS iBench speed improvement 
-
-        * JavaScriptCore.xcodeproj/project.pbxproj: Add AllInOneFile.cpp, and remove files it includes
-        from the build.
-        * kjs/AllInOneFile.cpp: Added.
-        * kjs/dtoa.cpp: Renamed CONST to CONST_ to avoid conflict.
-        (Bigint::):
-        (Bigint::nrv_alloc):
-        * kjs/lookup.cpp: Use "namspace KJS { ... }" instead of "using namespace KJS;" 
-
-2007-04-23  Maciej Stachowiak  <mjs@apple.com>
-
-        Build fix, not reviewed.
-
-        * kjs/collector.h: Fix struct/class mismatch.
-
-2007-04-23  Maciej Stachowiak  <mjs@apple.com>
-
-        Reviewed by Darin.
-
-        - raise ALLOCATIONS_PER_COLLECTION to 4000, for 3.7% iBench speed improvement
-         
-        Now that the cell size is smaller and the block size is bigger, we can fit 4000 objects in
-        the two spare cells the collector is willing to keep around, so collect a bit less often.
-        
-        * kjs/collector.cpp:
-
-2007-04-23  Maciej Stachowiak  <mjs@apple.com>
-
-        Reviewed by Darin and Geoff.
-        
-        - move mark and collectOnMainThreadOnly bits into separate bitmaps
-        
-        This saves 4 bytes per cell, allowing shrink of cell size to 32,
-        which leads to a .8% speed improvement on iBench.
-        
-        This is only feasible because of all the previous changes on the branch.
-
-        * kjs/collector.cpp:
-        (KJS::allocateBlock): Adjust for some renames of constants. 
-        (KJS::Collector::markStackObjectsConservatively): Now that cells are 32 bytes (64 
-        bytes on 64-bit) the cell alignment check can be made much more strict, and also
-        obsoletes the need for a % sizeof(CollectorCell) check. Also, we can mask off the low
-        bits of the pointer to have a potential block pointer to look for.
-        (KJS::Collector::collectOnMainThreadOnly): Use bitmap.
-        (KJS::Collector::markMainThreadOnlyObjects): Use bitmap.
-        (KJS::Collector::collect): When sweeping, use bitmaps directly to find mark bits.
-        * kjs/collector.h:
-        (KJS::): Move needed constants and type declarations here.
-        (KJS::CollectorBitmap::get): Bit twiddling to get a bitmap value.
-        (KJS::CollectorBitmap::set): Bit twiddling to set a bitmap bit to true.
-        (KJS::CollectorBitmap::clear): Bit twiddling to set a bitmap bit to false.
-        (KJS::CollectorBitmap::clearAll): Clear whole bitmap at one go.
-        (KJS::Collector::cellBlock): New operation, compute the block pointer for
-        a cell by masking off low bits.
-        (KJS::Collector::cellOffset): New operation, compute the cell offset for a
-        cell by masking off high bits and dividing (actually a shift).
-        (KJS::Collector::isCellMarked): Check mark bit in bitmap
-        (KJS::Collector::markCell): Set mark bit in bitmap.
-        * kjs/value.h:
-        (KJS::JSCell::JSCell): No more bits.
-        (KJS::JSCell::marked): Let collector handle it.
-        (KJS::JSCell::mark): Let collector handle it.
-
-2007-04-23  Anders Carlsson  <andersca@apple.com>
-
-        Build fix.
-        
-        * kjs/regexp_object.h:
-        RegExpObjectImpPrivate is a struct, not a class.
-        
-2007-04-23  Maciej Stachowiak  <mjs@apple.com>
-
-        Reviewed by Darin.
-        
-        - shrink FunctionImp / DeclaredFunctionImp by 4 bytes, by moving parameter list to function body
-        
-        I reconciled this with a similar change in KDE kjs by Maks Orlovich <maksim@kde.org>.
-
-        * kjs/function.cpp:
-        (KJS::FunctionImp::callAsFunction):
-        (KJS::FunctionImp::passInParameters):
-        (KJS::FunctionImp::lengthGetter):
-        (KJS::FunctionImp::getParameterName):
-        * kjs/function.h:
-        * kjs/function_object.cpp:
-        (FunctionProtoFunc::callAsFunction):
-        (FunctionObjectImp::construct):
-        * kjs/nodes.cpp:
-        (FunctionBodyNode::addParam):
-        (FunctionBodyNode::paramString):
-        (FuncDeclNode::addParams):
-        (FuncDeclNode::processFuncDecl):
-        (FuncExprNode::addParams):
-        (FuncExprNode::evaluate):
-        * kjs/nodes.h:
-        (KJS::Parameter::Parameter):
-        (KJS::FunctionBodyNode::numParams):
-        (KJS::FunctionBodyNode::paramName):
-        (KJS::FunctionBodyNode::parameters):
-        (KJS::FuncExprNode::FuncExprNode):
-        (KJS::FuncDeclNode::FuncDeclNode):
-        * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: Disable 64-bit warnings because
-        they handle size_t badly.
-
-2007-04-23  Maciej Stachowiak  <mjs@apple.com>
-        Reviewed by Darin.
-
-        - shrink RegexpObjectImp by 4 bytes
-        
-        Somewhat inexplicably, this seems to be a .33% speedup on JS iBench.
-        
-        * kjs/regexp_object.cpp:
-        (KJS::RegExpObjectImpPrivate::RegExpObjectImpPrivate):
-        (RegExpObjectImp::RegExpObjectImp):
-        (RegExpObjectImp::performMatch):
-        (RegExpObjectImp::arrayOfMatches):
-        (RegExpObjectImp::getBackref):
-        (RegExpObjectImp::getLastMatch):
-        (RegExpObjectImp::getLastParen):
-        (RegExpObjectImp::getLeftContext):
-        (RegExpObjectImp::getRightContext):
-        (RegExpObjectImp::getValueProperty):
-        (RegExpObjectImp::putValueProperty):
-        * kjs/regexp_object.h:
-
-2007-04-23  Maciej Stachowiak  <mjs@apple.com>
-
-        Reviewed by Oliver.
-        
-        - change to 1-bit bitfields instead of 8-bit, this turns out to lead to a .51% speedup on JS iBench
-        
-        The 1-bit bitfields are actually faster than just plain bools, at least on Intel (go figure).
-
-        * kjs/property_map.h:
-
-2007-04-23  Maciej Stachowiak  <mjs@apple.com>
-
-        Reviewed by Darin.
-       
-        - shrink ArrayInstance objects by 4 bytes
-        http://bugs.webkit.org/show_bug.cgi?id=13386
-        
-        I did this by storing the capacity before the beginning of the storage array. It turns out
-        it is rarely needed and is by definition 0 when the storage array is null.
-        * kjs/array_instance.h:
-        (KJS::ArrayInstance::capacity): Get it from the secret stash
-        * kjs/array_object.cpp:
-        (allocateStorage): New function to encapsulate allocating the storage with extra space ahead
-        for the capacity.
-        (reallocateStorage): ditto for realloc
-        (ArrayInstance::ArrayInstance):
-        (ArrayInstance::~ArrayInstance):
-        (ArrayInstance::resizeStorage):
-
-2007-04-23  Darin Adler  <darin@apple.com>
-
-        Reviewed by Maciej.
-
-        - fix <rdar://problem/4840688> REGRESSION (r10588, r10621): JavaScript won't parse
-          modifications of non-references (breaks 300themovie.warnerbros.com, fedex.com)
-
-        Despite the ECMAScript specification's claim that you can treat these as syntax
-        errors, doing so creates some website incompatibilities. So this patch turns them back
-        into evaluation errors instead.
-
-        Test: fast/js/modify-non-references.html
-
-        * kjs/grammar.y: Change makeAssignNode, makePrefixNode, and makePostfixNode so that they
-        never fail to parse. Update rules that use them. Fix a little bit of indenting. Use
-        new PostfixErrorNode, PrefixErrorNode, and AssignErrorNode classes.
-
-        * kjs/nodes.h: Added an overload of throwError that takes a char* argument.
-        Replaced setExceptionDetailsIfNeeded and debugExceptionIfNeeded with handleException,
-        which does both. Added PostfixErrorNode, PrefixErrorNode, and AssignErrorNode classes.
-
-        * kjs/nodes.cpp: Changed exception macros to use handleException; simpler and smaller
-        code size than the two functions that we used before.
-        (Node::throwError): Added the overload mentioned above.
-        (Node::handleException): Added. Contains the code from both setExceptionDetailsIfNeeded
-        and debugExceptionIfNeeded.
-        (PostfixErrorNode::evaluate): Added. Throws an exception.
-        (PrefixErrorNode::evaluate): Ditto.
-        (AssignErrorNode::evaluate): Ditto.
-        (ThrowNode::execute): Call handleException instead of debugExceptionIfNeeded; this
-        effectively adds a call to setExceptionDetailsIfNeeded, which may help with getting
-        the correct file and line number for these exceptions.
-
-        * kjs/nodes2string.cpp:
-        (PostfixErrorNode::streamTo): Added.
-        (PrefixErrorNode::streamTo): Added.
-        (AssignErrorNode::streamTo): Added.
-
-2007-04-23  Maciej Stachowiak  <mjs@apple.com>
-
-        Reviewed by Darin.
-        
-        - fix test failures / crashes on PPC
-
-        * kjs/property_map.h: Make the bool fields explicitly 8-bit bitfields, since bool is a full
-        word there otherwise :-(
-
-2007-04-23  Maciej Stachowiak  <mjs@apple.com>
-
-        Reviewed by Darin.
-        
-        - fix more test case failures
-
-        * bindings/runtime_array.cpp:
-        (RuntimeArray::RuntimeArray): inherit from JSObject instead of ArrayInstance; it turns
-        out that this class only needs the prototype and classInfo from ArrayInstance, not the
-        actual class itself, and it was too big otherwise.
-        (RuntimeArray::getOwnPropertySlot):
-        * bindings/runtime_array.h:
-
-2007-04-23  Maciej Stachowiak  <mjs@apple.com>
-
-        Reviewed by Darin.
-        
-        - fix some test failures
-
-        * bindings/runtime_method.cpp:
-        (RuntimeMethod::RuntimeMethod): inherit from InternalFunctionImp instead of FunctionImpl,
-        otherwise this is too big
-        (RuntimeMethod::getOwnPropertySlot):
-        * bindings/runtime_method.h:
-
-2007-04-22  Maciej Stachowiak  <mjs@apple.com>
-
-        Reviewed by Darin.
-        
-        - discard the arguments List for an ActivationImp when the corresponding Context is destroyed (1.7% speedup)
-        http://bugs.webkit.org/show_bug.cgi?id=13385
-
-        Based an idea by Christopher E. Hyde <C.Hyde@parableuk.force9.co.uk>. His patch to do 
-        this also had many other List changes and I found this much simpler subset of the changes
-        was actually a hair faster.
-        
-        This optimization is valid because the arguments list is only kept around to
-        lazily make the arguments object. If it's not made by the time the function
-        exits, it never will be, since any function that captures the continuation will
-        have its own local arguments variable in scope.
-        
-        Besides the 1.7% speed improvement, it shrinks List by 4 bytes
-        (which in turn shrinks ActivationImp by 4 bytes).
-        
-        * kjs/Context.cpp:
-        (KJS::Context::~Context): Clear the activation's arguments list.
-        * kjs/function.cpp:
-        (KJS::ActivationImp::ActivationImp): Adjusted for list changes.
-        (KJS::ActivationImp::mark): No need to mark, lists are always protected (this doesn't
-        cause a ref-cycle for reasons stated above).
-        (KJS::ActivationImp::createArgumentsObject): Clear arguments list.
-        * kjs/function.h:
-        * kjs/list.cpp:
-        (KJS::List::List): No more needsMarking boolean
-        (KJS::List::operator=): ditto
-        * kjs/list.h:
-        (KJS::List::List): ditto
-        (KJS::List::reset): ditto
-        (KJS::List::deref): ditto
-
-2007-04-22  Maciej Stachowiak  <mjs@apple.com>
-
-        Reviewed by Darin.
-        
-        - shrink PropertyMap by 8 bytes and therefore shrink CELL_SIZE to 40 (for 32-bit; 
-        similar shrinkage for 64-bit)
-        http://bugs.webkit.org/show_bug.cgi?id=13384
-
-        Inspired by similar changes by Christopher E. Hyde <C.Hyde@parableuk.force9.co.uk>
-        done in the kjs-tweaks branch of KDE's kjs. However, this version is somewhat 
-        cleaner style-wise and avoids some of the negative speed impact (at least on gcc/x86) 
-        of his version.
-        
-        This is nearly a wash performance-wise, maybe a slight slowdown, but worth doing
-        to eventually reach cell size 32.
-        
-        * kjs/collector.cpp:
-        (KJS::):
-        * kjs/property_map.cpp:
-        (KJS::PropertyMap::~PropertyMap):
-        (KJS::PropertyMap::clear):
-        (KJS::PropertyMap::get):
-        (KJS::PropertyMap::getLocation):
-        (KJS::PropertyMap::put):
-        (KJS::PropertyMap::insert):
-        (KJS::PropertyMap::expand):
-        (KJS::PropertyMap::rehash):
-        (KJS::PropertyMap::remove):
-        (KJS::PropertyMap::mark):
-        (KJS::PropertyMap::containsGettersOrSetters):
-        (KJS::PropertyMap::getEnumerablePropertyNames):
-        (KJS::PropertyMap::getSparseArrayPropertyNames):
-        (KJS::PropertyMap::save):
-        (KJS::PropertyMap::checkConsistency):
-        * kjs/property_map.h:
-        (KJS::PropertyMap::hasGetterSetterProperties):
-        (KJS::PropertyMap::setHasGetterSetterProperties):
-        (KJS::PropertyMap::):
-        (KJS::PropertyMap::PropertyMap):
-
-2007-04-22  Maciej Stachowiak  <mjs@apple.com>
-
-        Reviewed by Darin.
-        
-        - change blocks to 64k in size, and use various platform-specific calls to allocate at 64k-aligned addresses
-        http://bugs.webkit.org/show_bug.cgi?id=13383
-        
-        * kjs/collector.cpp:
-        (KJS::allocateBlock): New function to allocate 64k of 64k-aligned memory
-        (KJS::freeBlock): Corresponding free
-        (KJS::Collector::allocate):
-        (KJS::Collector::collect):
-
-2007-04-22  Maciej Stachowiak  <mjs@apple.com>
-
-        Reviewed by Darin and Geoff.
-        
-        - remove the concept of oversize objects, now that there aren't any (for now
-        only enforced with an assert).
-        http://bugs.webkit.org/show_bug.cgi?id=13382
-
-        This change is a .66% speedup on JS iBench for 32-bit platforms, probably much more
-        for 64-bit since it finally gives a reasonable cell size, but I did not test that.
-        
-        * kjs/collector.cpp:
-        (KJS::): Use different cell size for 32-bit and 64-bit, now that there is no
-        oversize allocation.
-        (KJS::Collector::allocate): Remove oversize allocator.
-        (KJS::Collector::markStackObjectsConservatively): Don't check oversize objects.
-        (KJS::Collector::markMainThreadOnlyObjects): Ditto.
-        (KJS::Collector::collect): Ditto.
-
-2007-04-21  Mitz Pettel  <mitz@webkit.org>
-
-        Reviewed by Adam.
-
-        - fix http://bugs.webkit.org/show_bug.cgi?id=13428
-          REGRESSION (r20973-r20976): Failing ecma/Array/15.4.4.5-3.js
-
-        - fix http://bugs.webkit.org/show_bug.cgi?id=13429
-          REGRESSION (r20973-r20976): Crashing in fast/dom/plugin-attributes-enumeration.html
-
-        * kjs/array_object.cpp:
-        (ArrayInstance::sort): Free the old storage, not the new one.
-
-2007-04-20  Maciej Stachowiak  <mjs@apple.com>
-
-        Not reviewed, build fix.
-
-        - fix build problem with last change - -O3 complains more about uninitialized variables
-        
-        * pcre/pcre_compile.c:
-        (compile_branch):
-        (pcre_compile2):
-
-2007-04-20  Maciej Stachowiak  <mjs@apple.com>
-
-        Reviewed by Darin.
-        
-        - <rdar://problem/5149915> use mergesort when possible, since it leads to fewer compares (2% JS iBench speedup)
-
-        * kjs/array_object.cpp:
-        (ArrayInstance::sort): Use mergesort(3) on platforms that have it, since it tends
-        to do fewer compares than qsort; but avoid it very on large arrays since it uses extra
-        memory. Also added comments identifying possibly even better sorting algorithms
-        for sort by string value and sort by compare function.
-        * kjs/config.h:
-
-2007-04-20  Maciej Stachowiak  <mjs@apple.com>
-
-        Reviewed by Darin.
-        
-        - bump optimization flags up to -O3 for 1% JS iBench speed improvement 
-
-        * Configurations/Base.xcconfig:
-
-2007-04-20  Mark Rowe  <mrowe@apple.com>
-
-        Reviewed by Maciej.
-
-        Fix bogus optimisation in the generic pthread code path.
-
-        * kjs/collector.cpp:
-        (KJS::currentThreadStackBase):
-
-2007-04-20  Mark Rowe  <mrowe@apple.com>
-
-        Reviewed by Anders.
-
-        Improve FreeBSD compatibility, as suggested by Alexander Botero-Lowry.
-
-        * kjs/collector.cpp:
-        (KJS::currentThreadStackBase): FreeBSD requires that pthread_attr_t's are
-        initialized via pthread_attr_init before being used in any context.
-
-2007-04-19  Mark Rowe  <mrowe@apple.com>
-
-        Reviewed by Darin.
-
-        Fix http://bugs.webkit.org/show_bug.cgi?id=13401
-        Bug 13401: Reproducible crash calling myArray.sort(compareFn) from within
-        a sort comparison function
-
-        * kjs/array_object.cpp:
-        (ArrayInstance::sort): Save/restore the static variables around calls to qsort
-        to ensure nested calls to ArrayInstance::sort behave correctly.
-
-2007-04-12  Deneb Meketa  <dmeketa@adobe.com>
-
-        Reviewed by Darin Adler.
-
-        http://bugs.webkit.org/show_bug.cgi?id=13029
-        rdar://problem/4994849
-        Bug 13029: Permit NPAPI plug-ins to see HTTP response headers.
-        This doesn't actually change JavaScriptCore, but that's where npapi.h is.
-
-        * bindings/npapi.h:
-        Add headers member to NPStream struct.  Also increase NP_VERSION_MINOR to 18.
-        Increasing to >= 17 allows plug-ins to safely detect whether to look for
-        NPStream::headers.  Increasing from 17 to 18 reflects presence of NPObject
-        enumeration, which was added in a prior patch, and which has been agreed to
-        constitute version 18 by the plugin-futures list.  Also add other missing
-        bits of npapi.h to catch up from 14 to 18.  This includes features that are
-        not implemented in WebKit, but those are safely stubbed.
-
-2007-04-10  Geoffrey Garen  <ggaren@apple.com>
-
-        Reviewed by Mark Rowe.
-        
-        Fixed last check-in to print in release builds, too.
-
-        * kjs/collector.cpp:
-        (KJS::getPlatformThreadRegisters):
-
-2007-04-10  Geoffrey Garen  <ggaren@apple.com>
-
-        Reviewed by John Sullivan, Darin Adler.
-        
-        Fixed <rdar://problem/5121899> JavaScript garbage collection leads to 
-        later crash under Rosetta (should abort or leak instead?)
-        
-        Log an error message and crash if the kernel reports failure during GC.
-        We decided to do this instead of just leaking because we don't want people
-        to get the mistaken impression that running in Rosetta is a supported
-        configurtion.
-        
-        The CRASH macro will also hook into CrashReporter, which will tell us if 
-        many (any?) users run into this issue.
-
-        * kjs/collector.cpp:
-        (KJS::getPlatformThreadRegisters):
-
-2007-04-06  Krzysztof Kowalczyk  <kkowalczyk@gmail.com>
-
-        Reviewed by darin.
-
-        Coverity fix. Coverity says:
-        "Event var_deref_model: Variable "sourceRanges" tracked as NULL was passed to a
-        function that dereferences it"
-
-        * kjs/string_object.cpp:
-        (KJS::replace):
-
-2007-04-06  Geoffrey Garen  <ggaren@apple.com>
-
-        Rubber stamped by Adele Peterson.
-
-        * kjs/ExecState.h: Removed obsolete forward/friend declaration of 
-        RuntimeMethodImp.
-
-2007-04-05  Krzysztof Kowalczyk  <kkowalczyk@gmail.com>
-
-        Reviewed by darin.
-
-        Coverity fix. Coverity says:
-        "Event check_after_deref: Pointer "dateString" dereferenced before NULL check"
-
-        * kjs/date_object.cpp:
-        (KJS::parseDate):
-
-2007-04-05  Krzysztof Kowalczyk  <kkowalczyk@gmail.com>
-
-        Reviewed by darin.
-
-        Coverity fix. Coverity says:
-        "Event check_after_deref: Pointer "re" dereferenced before NULL check"
-
-        * pcre/pcre_study.c:
-        (pcre_study):
-
-2007-04-05  Krzysztof Kowalczyk  <kkowalczyk@gmail.com>
-
-        Reviewed by darin.
-
-        Coverity fixes. Coverity says:
-        "Event leaked_storage: Returned without freeing storage "buffer""
-        and:
-        "Event leaked_storage: Returned without freeing storage "script""
-
-        * kjs/testkjs.cpp:
-        (doIt):
-        (createStringWithContentsOfFile):
-
-2007-04-05  Krzysztof Kowalczyk  <kkowalczyk@gmail.com>
-
-        Reviewed by darin.
-
-        Coverity fix: in single-threaded case currentThreadIsMainThread is always true
-        so the code in if (!currentThreadIsMainThread) cannot possibly be reached
-        and Coverity complains about dead code.
-
-        * kjs/collector.cpp:
-        (KJS::Collector::collect):
-
-=== Safari-5522.6 ===
-
-2007-04-03  Kevin McCullough  <kmccullough@apple.com>
-
-        Reviewed by Adam.
-
-        - Testing a post-commit hook.
-
-        * JavaScriptCore.vcproj/testkjs/testkjs.vcproj:
-
-2007-04-03  Anders Carlsson  <andersca@apple.com>
-
-        Reviewed by Adam.
-
-        <rdar://problem/5107534>
-        http://bugs.webkit.org/show_bug.cgi?id=13265
-        REGRESSION: Crash in KJS::Bindings::convertValueToNPVariant
-        
-        * bindings/NP_jsobject.cpp:
-        (_NPN_InvokeDefault):
-        Return false if the object isn't a function. Set the return value to undefined by default
-        (to match Firefox).
-        
-2007-03-30  Anders Carlsson <andersca@apple.com>
-
-        Build fix.
-        
-        * bindings/NP_jsobject.cpp:
-        (_NPN_Enumerate):
-
-2007-03-30  Anders Carlsson  <andersca@apple.com>
-
-        Reviewed by Geoff.
-
-        Implement _NPN_Enumerate support.
-        
-        * JavaScriptCore.exp:
-        * bindings/NP_jsobject.cpp:
-        (_NPN_Enumerate):
-        * bindings/c/c_instance.cpp:
-        (KJS::Bindings::CInstance::getPropertyNames):
-        * bindings/c/c_instance.h:
-        * bindings/npapi.h:
-        * bindings/npruntime.h:
-        * bindings/npruntime_impl.h:
-        * bindings/runtime.h:
-        (KJS::Bindings::Instance::getPropertyNames):
-        * bindings/runtime_object.cpp:
-        (RuntimeObjectImp::getPropertyNames):
-        * bindings/runtime_object.h:
-        (KJS::RuntimeObjectImp::getInternalInstance):
-
-2007-03-28  Jeff Walden  <jwalden+code@mit.edu>
-
-        Reviewed by Darin.
-
-        http://bugs.webkit.org/show_bug.cgi?id=12963
-        Fix some inconsistencies in the Mozilla JS Array extras implementations
-        with respect to the Mozilla implementation:
-
-          - holes in arrays should be skipped, not treated as undefined,
-            by all such methods
-          - an element with value undefined is not a hole
-          - Array.prototype.forEach should return undefined
-
-        * kjs/array_object.cpp:
-        (ArrayInstance::getOwnPropertySlot):
-        (ArrayProtoFunc::callAsFunction):
-
-2007-03-27  Anders Carlsson  <acarlsson@apple.com>
-
-        Reviewed by Geoff.
-
-        * bindings/NP_jsobject.cpp:
-        (_NPN_InvokeDefault):
-        Call JSObject:call for native JavaScript objects.
-
-2007-03-26  David Carson  <dacarson@gmail.com>
-
-        Reviewed by Darin, landed by Anders.
-
-        Fix for: REGRESSION (r19559): Java applet crash
-        http://bugs.webkit.org/show_bug.cgi?id=13142
-        <rdar://problem/5080340>
-
-        The previous fix http://bugs.webkit.org/show_bug.cgi?id=12636 
-        introduced new JNIType to enum in jni_utility.h This is a 
-        problem on the Mac as it seems that the JNIType enum is also
-        used in the JVM, it is used to specify the return type in
-        jni_objc.mm
-        Corrected the fix by moving type to the end, and changing
-        jni_objc.mm to convert the new type to an old compatible
-        type.
-
-        * bindings/jni/jni_objc.mm:
-        (KJS::Bindings::dispatchJNICall):
-        * bindings/jni/jni_utility.h:
-
-2007-03-26  Christopher Brichford  <chrisb@adobe.com>
-
-        Reviewed/landed by Adam.
-
-        Bug 13198: Move build settings from project file to xcconfig file for apollo
-        port JSCore
-        http://bugs.webkit.org/show_bug.cgi?id=13198
-
-        - Moving build settings from xcode project file to xcconfig files.
-
-        * JavaScriptCore.apolloproj/mac/JavaScriptCore/JavaScriptCore.Debug.xcconfig:
-        * JavaScriptCore.apolloproj/mac/JavaScriptCore/JavaScriptCore.Release.xcconfig:
-        * JavaScriptCore.apolloproj/mac/JavaScriptCore/JavaScriptCore.xcconfig:
-        * JavaScriptCore.apolloproj/mac/JavaScriptCore/JavaScriptCore.xcodeproj/project.pbxproj:
-
-2007-03-26  Brady Eidson  <beidson@apple.com>
-
-        Rubberstamped by Anders and Maciej aand Geoff (oh my!)
-        
-        Since CFTypeRef is really void*, a RetainPtr couldn't be used. 
-        RefType was "void", which doesn't actually exist as a type.
-        Since RefType only existed for operator*(), and since that operator
-        doesn't make any sense for RetainPtr, I removed them!
-
-        * kjs/nodes.cpp: Touch this to force a rebuild and (hopefully) help the
-          compiler with dependencies
-        * wtf/RetainPtr.h: Nuke RefType and operator*()
-
-2007-03-26  Geoffrey Garen  <ggaren@apple.com>
-
-        Touched a file to (hopefully) help the compiler with RetainPtr dependencies.
-
-        * kjs/nodes.cpp:
-        (Node::deref):
-
-2007-03-24  Brady Eidson  <beidson@apple.com>
-
-        Reviewed by Adam
-
-        Whoops, RetainPtr should be in the WTF namespace
-
-        * wtf/RetainPtr.h:
-
-2007-03-24  Brady Eidson  <beidson@apple.com>
-
-        Reviewed by Adam
-        
-        <rdar://problem/5086210> - Move RetainPtr to WTF
-
-        * wtf/RetainPtr.h: Added
-        * JavaScriptCore.xcodeproj/project.pbxproj: Add it to the project file
-        * JavaScriptCore.vcproj/WTF/WTF.vcproj: Ditto
-
-
-2007-03-23  Christopher Brichford  <chrisb@adobe.com>
-
-        Reviewed/landed by Adam.
-
-        Bug 13175: Make apollo mac project files for JavaScriptCore actually
-        build something
-        http://bugs.webkit.org/show_bug.cgi?id=13175
-
-        - Changing apollo mac project files for JavaScriptCore such that they actually build
-        JavaScriptCore source code.
-
-        * JavaScriptCore.apolloproj/ForwardingSources/grammar.cpp: Added.
-        * JavaScriptCore.apolloproj/mac/JavaScriptCore/JavaScriptCore.xcconfig:
-        * JavaScriptCore.apolloproj/mac/JavaScriptCore/JavaScriptCore.xcodeproj/project.pbxproj:
-
-2007-03-24  Mark Rowe  <mrowe@apple.com>
-
-        Rubber-stamped by Darin.
-
-        * Configurations/JavaScriptCore.xcconfig: Remove unnecessary INFOPLIST_PREPROCESS.
-
-2007-03-22  Christopher Brichford  <chrisb@adobe.com>
-
-        Reviewed/landed by Adam.
-
-        Bug 13164: Initial version of mac JavaScriptCore project files for
-        apollo port 
-        http://bugs.webkit.org/show_bug.cgi?id=13164
-
-        - Adding mac project files for apollo port of JavaScriptCore. Currently project
-        just builds dftables.
-
-        * JavaScriptCore.apolloproj/mac/JavaScriptCore/JavaScriptCore.Debug.xcconfig: Added.
-        * JavaScriptCore.apolloproj/mac/JavaScriptCore/JavaScriptCore.Release.xcconfig: Added.
-        * JavaScriptCore.apolloproj/mac/JavaScriptCore/JavaScriptCore.xcconfig: Added.
-        * JavaScriptCore.apolloproj/mac/JavaScriptCore/JavaScriptCore.xcodeproj/project.pbxproj: Added.
-
-2007-03-21  Timothy Hatcher  <timothy@apple.com>
-
-        Reviewed by Darin.
-
-        <rdar://problem/5076599> JavaScriptCore has a weak export (vtable for KJS::JSCell)
-
-        * JavaScriptCore.exp: Remove __ZTVN3KJS6JSCellE.
-
-2007-03-21  Adele Peterson  <adele@apple.com>
-
-        Reviewed by Geoff.
-
-        * API/JSStringRef.cpp: (JSStringIsEqual): Added JSLock.
-
-2007-03-21  Zack Rusin  <zrusin@trolltech.com>
-
-        Fix the compile when USE(MULTIPLE_THREADS) isn't
-        defined
-
-        * kjs/JSLock.cpp:
-        (KJS::JSLock::currentThreadIsHoldingLock):
-
-2007-03-20  Maciej Stachowiak  <mjs@apple.com>
-
-        Reviewed by Geoff and Adam.
-        
-        - make USE(MULTIPLE_THREADS) support more portable
-        http://bugs.webkit.org/show_bug.cgi?id=13069
-        
-        - fixed a threadsafety bug discovered by testing this
-        
-        - enhanced threadsafety assertions in collector
-
-        * API/JSCallbackObject.cpp:
-        (KJS::JSCallbackObject::~JSCallbackObject): This destructor can't
-        DropAllLocks around the finalize callback, because it gets called
-        from garbage collection and we can't let other threads collect!
-
-        * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
-        * kjs/JSLock.cpp:
-        (KJS::JSLock::currentThreadIsHoldingLock): Added new function
-        to allow stronger assertions than just that the lock is held
-        by some thread (you can now assert that the current thread is
-        holding it, given the new JSLock design).
-        * kjs/JSLock.h:
-        * kjs/collector.cpp: Refactored for portability plus added some
-        stronger assertions.
-        (KJS::Collector::allocate):
-        (KJS::currentThreadStackBase):
-        (KJS::Collector::registerAsMainThread):
-        (KJS::onMainThread):
-        (KJS::PlatformThread::PlatformThread):
-        (KJS::getCurrentPlatformThread):
-        (KJS::Collector::Thread::Thread):
-        (KJS::destroyRegisteredThread):
-        (KJS::Collector::registerThread):
-        (KJS::Collector::markCurrentThreadConservatively):
-        (KJS::suspendThread):
-        (KJS::resumeThread):
-        (KJS::getPlatformThreadRegisters):
-        (KJS::otherThreadStackPointer):
-        (KJS::otherThreadStackBase):
-        (KJS::Collector::markOtherThreadConservatively):
-        (KJS::Collector::markStackObjectsConservatively):
-        (KJS::Collector::protect):
-        (KJS::Collector::unprotect):
-        (KJS::Collector::collectOnMainThreadOnly):
-        (KJS::Collector::markMainThreadOnlyObjects):
-        (KJS::Collector::collect):
-        * kjs/collector.h:
-        * wtf/FastMalloc.cpp:
-        (WTF::fastMallocSetIsMultiThreaded):
-        * wtf/FastMallocInternal.h:
-        * wtf/Platform.h:
-
-2007-03-19  Darin Adler  <darin@apple.com>
-
-        * kjs/value.h: Roll ~JSValue change out. It was causing problems. I'll do it right later.
-
-2007-03-19  Geoffrey Garen  <ggaren@apple.com>
-
-        Reviewed by John Sullivan.
-
-        Fixed <rdar://problem/5073380> REGRESSION: Crash occurs at WTF::fastFree() 
-        when reloading liveconnect page (applet)
-        
-        Best to use free when you use malloc, especially when malloc and delete
-        use completely different libraries.
-
-        * bindings/jni/jni_runtime.cpp:
-        (JavaMethod::~JavaMethod):
-
-2007-03-19  Andrew Wellington  <proton@wiretapped.net>
-
-        Reviewed by Maciej.
-
-        Really set Xcode editor to use 4 space indentation (http://webkit.org/coding/coding-style.html)
-
-        * JavaScriptCore.xcodeproj/project.pbxproj:
-
-2007-03-19  Darin Adler  <darin@apple.com>
-
-        Reviewed by Geoff.
-
-        - Changed list size threshold to 5 based on testing.
-
-        I was testing the i-Bench JavaScript with the list statistics
-        dumping on, and discovered that there were many 5-element lists.
-        The fast case for lists was for 4 elements and fewer. By changing
-        the threshold to 5 elements we get a measurable speedup. I believe
-        this will help real web pages too, not just the benchmark.
-
-        * kjs/list.cpp: Change constant from 4 to 5.
-
-2007-03-19  Darin Adler  <darin@apple.com>
-
-        * kjs/value.h: Oops, fix build.
-
-2007-03-19  Darin Adler  <darin@apple.com>
-
-        Reviewed by Geoff.
-
-        - remove ~JSValue; tiny low-risk performance boost
-
-        * kjs/value.h: Remove unneeded empty virtual destructor from JSValue.
-        The only class derived from JSValue is JSCell and it already has a
-        virtual destructor. Declaring an empty constructor in JSValue had one
-        good effect: it marked the destructor private, making it a compile
-        time error to try to destroy a JSValue; but that's not a likely
-        mistake for someone to make. It had two bad effects: (1) it caused gcc,
-        at least, to generate code to fix up the virtual table pointer to
-        point to the JSValue version of the virtual table inside the destructor
-        of all classes derived from JSValue directly or indirectly; (2) it
-        caused JSValue to be a polymorphic class so required a virtual table for
-        it. It's cleaner to not have either of those.
-
-2007-03-18  Maciej Stachowiak  <mjs@apple.com>
-
-        Reviewed by Mark.
-        
-        - avoid static construction (and global variable access) in a smarter, more portable way,
-        to later enable MUTLI_THREAD mode to work on other platforms and compilers.
-        
-        * kjs/CommonIdentifiers.cpp: Added. New class to hold all the shared identifiers.
-        (KJS::CommonIdentifiers::CommonIdentifiers):
-        (KJS::CommonIdentifiers::shared):
-        * kjs/CommonIdentifiers.h: Added.
-
-        * kjs/ExecState.h:
-        (KJS::ExecState::propertyNames): Hand the CommonIdentifiers instance here for easy access.
-        (KJS::ExecState::ExecState):
-
-        * API/JSObjectRef.cpp:
-        (JSObjectMakeConstructor):
-        * CMakeLists.txt:
-        * JavaScriptCore.exp:
-        * JavaScriptCore.pri:
-        * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
-        * JavaScriptCore.xcodeproj/project.pbxproj:
-        * JavaScriptCoreSources.bkl:
-        * bindings/runtime_array.cpp:
-        (RuntimeArray::getOwnPropertySlot):
-        (RuntimeArray::put):
-        * bindings/runtime_method.cpp:
-        (RuntimeMethod::getOwnPropertySlot):
-        * kjs/array_object.cpp:
-        (ArrayInstance::getOwnPropertySlot):
-        (ArrayInstance::put):
-        (ArrayInstance::deleteProperty):
-        (ArrayProtoFunc::ArrayProtoFunc):
-        (ArrayProtoFunc::callAsFunction):
-        (ArrayObjectImp::ArrayObjectImp):
-        * kjs/bool_object.cpp:
-        (BooleanPrototype::BooleanPrototype):
-        (BooleanProtoFunc::BooleanProtoFunc):
-        (BooleanProtoFunc::callAsFunction):
-        (BooleanObjectImp::BooleanObjectImp):
-        * kjs/completion.h:
-        (KJS::Completion::Completion):
-        * kjs/date_object.cpp:
-        (KJS::DateProtoFunc::DateProtoFunc):
-        (KJS::DateObjectImp::DateObjectImp):
-        (KJS::DateObjectFuncImp::DateObjectFuncImp):
-        * kjs/error_object.cpp:
-        (ErrorPrototype::ErrorPrototype):
-        (ErrorProtoFunc::ErrorProtoFunc):
-        (ErrorProtoFunc::callAsFunction):
-        (ErrorObjectImp::ErrorObjectImp):
-        (ErrorObjectImp::construct):
-        (NativeErrorPrototype::NativeErrorPrototype):
-        (NativeErrorImp::NativeErrorImp):
-        (NativeErrorImp::construct):
-        (NativeErrorImp::callAsFunction):
-        * kjs/function.cpp:
-        (KJS::FunctionImp::getOwnPropertySlot):
-        (KJS::FunctionImp::put):
-        (KJS::FunctionImp::deleteProperty):
-        (KJS::FunctionImp::getParameterName):
-        (KJS::DeclaredFunctionImp::construct):
-        (KJS::IndexToNameMap::unMap):
-        (KJS::Arguments::Arguments):
-        (KJS::ActivationImp::getOwnPropertySlot):
-        (KJS::ActivationImp::deleteProperty):
-        (KJS::GlobalFuncImp::GlobalFuncImp):
-        * kjs/function_object.cpp:
-        (FunctionPrototype::FunctionPrototype):
-        (FunctionProtoFunc::FunctionProtoFunc):
-        (FunctionProtoFunc::callAsFunction):
-        (FunctionObjectImp::FunctionObjectImp):
-        (FunctionObjectImp::construct):
-        * kjs/grammar.y:
-        * kjs/identifier.cpp:
-        * kjs/identifier.h:
-        * kjs/interpreter.cpp:
-        (KJS::Interpreter::init):
-        (KJS::Interpreter::initGlobalObject):
-        * kjs/interpreter.h:
-        * kjs/lookup.h:
-        * kjs/math_object.cpp:
-        (MathFuncImp::MathFuncImp):
-        * kjs/nodes.cpp:
-        (ArrayNode::evaluate):
-        (FuncDeclNode::processFuncDecl):
-        (FuncExprNode::evaluate):
-        * kjs/number_object.cpp:
-        (NumberPrototype::NumberPrototype):
-        (NumberProtoFunc::NumberProtoFunc):
-        (NumberObjectImp::NumberObjectImp):
-        * kjs/object.cpp:
-        (KJS::JSObject::put):
-        (KJS::JSObject::defaultValue):
-        (KJS::JSObject::hasInstance):
-        * kjs/object.h:
-        (KJS::JSObject::getOwnPropertySlot):
-        * kjs/object_object.cpp:
-        (ObjectPrototype::ObjectPrototype):
-        (ObjectProtoFunc::ObjectProtoFunc):
-        (ObjectObjectImp::ObjectObjectImp):
-        * kjs/regexp_object.cpp:
-        (RegExpPrototype::RegExpPrototype):
-        (RegExpProtoFunc::RegExpProtoFunc):
-        (RegExpObjectImp::RegExpObjectImp):
-        * kjs/string_object.cpp:
-        (KJS::StringInstance::getOwnPropertySlot):
-        (KJS::StringInstance::put):
-        (KJS::StringInstance::deleteProperty):
-        (KJS::StringPrototype::StringPrototype):
-        (KJS::StringProtoFunc::StringProtoFunc):
-        (KJS::StringProtoFunc::callAsFunction):
-        (KJS::StringObjectImp::StringObjectImp):
-        (KJS::StringObjectFuncImp::StringObjectFuncImp):
-        * kjs/testkjs.cpp:
-        (TestFunctionImp::TestFunctionImp):
-
-2007-03-18  Andrew Wellington  <proton@wiretapped.net>
-
-        Reviewed by Mark Rowe
-        
-        Set Xcode editor to use 4 space indentation (http://webkit.org/coding/coding-style.html) 
-
-        * JavaScriptCore.xcodeproj/project.pbxproj:
-
-2007-03-19  Mark Rowe  <mrowe@apple.com>
-
-        Rubber-stamped by Brady.
-
-        Update references to bugzilla.opendarwin.org with bugs.webkit.org.
-
-        * bindings/c/c_utility.cpp:
-        (KJS::Bindings::convertUTF8ToUTF16):
-        * kjs/function.cpp:
-        (KJS::FunctionImp::callAsFunction):
-        * kjs/grammar.y:
-        * kjs/keywords.table:
-        * kjs/lexer.cpp:
-        (KJS::Lexer::shift):
-
-2007-03-18  Geoffrey Garen  <ggaren@apple.com>
-
-        Reviewed by Oliver Hunt.
-        
-        Exposed some extra toUInt32 functionality, as part of the fix for
-        REGRESSION: Incomplete document.all implementation breaks abtelectronics.com 
-        (Style Change Through JavaScript Blanks Content)
-
-        * JavaScriptCore.exp:
-        * kjs/identifier.h:
-        (KJS::Identifier::toUInt32):
-
-2007-03-18  Geoffrey Garen  <ggaren@apple.com>
-
-        Removed duplicate export name.
-        
-        * JavaScriptCore.exp:
-
-2007-03-15  Geoffrey Garen  <ggaren@apple.com>
-
-        Reviewed by Maciej Stachowiak.
-        
-        Fixed <rdar://problem/5064964> Repro ASSERT failure in JS Bindings when 
-        closing window @ lowtrades.bptrade.com
-        
-        Unfortunately, the bindings depend on UString and Identifier as string 
-        representations. So, they need to acquire the JSLock when doing something
-        that will ref/deref their strings.
-
-        Layout tests, the original site, and Java, Flash, and Quicktime on the 
-        web work. No leaks reported. No automated test for this because testing 
-        the Java bindings, like math, is hard.
-        
-        * bindings/runtime.h: Made Noncopyable, just to be sure.
-        
-        * bindings/c/c_class.cpp: 
-        (KJS::Bindings::CClass::~CClass): Acquire the JSLock and explicitly clear the keys
-        in our hashtable, since they're UString::Reps, and ref/deref aren't thread-safe.
-        (KJS::Bindings::CClass::methodsNamed): Also acquire the JSLock when adding
-        keys to the table, since the table ref's them.
-        (KJS::Bindings::CClass::fieldNamed): ditto.
-
-        * bindings/c/c_utility.cpp: Removed dead function.
-        (KJS::Bindings::convertValueToNPVariant): Acquire the JSLock because doing
-        it recursively is pretty cheap, and it's just too confusing to tell whether
-        all our callers do it for us.
-        (KJS::Bindings::convertNPVariantToValue): ditto
-        * bindings/c/c_utility.h:
-
-        * bindings/jni/jni_class.cpp: Same deal as c_class.cpp.
-        (JavaClass::JavaClass):
-        (JavaClass::~JavaClass):
-
-        * bindings/jni/jni_instance.cpp: Same deal as c_utility.cpp.
-        (JavaInstance::stringValue):
-        * bindings/jni/jni_jsobject.cpp:
-        (JavaJSObject::convertValueToJObject):
-
-        * bindings/jni/jni_runtime.cpp:
-        (JavaMethod::~JavaMethod): Moved from header, for clarity.
-        (appendClassName): Made this static, so the set of callers is known, and
-        we can assert that we hold the JSLock. Also changed it to take a UString
-        reference, which makes the calling code simpler.
-        (JavaMethod::signature): Store the ASCII value we care about instead of
-        a UString, since UString is so much more hassle. Hold the JSLock while
-        building up the temporary UString.
-
-        * bindings/jni/jni_runtime.h: Nixed dead code in JavaMethod.
-        (KJS::Bindings::JavaString::JavaString): Hold a UString::Rep instead of
-        a UString, so we can acquire the JSLock and explicitly release it.
-        (KJS::Bindings::JavaString::_commonInit):
-        (KJS::Bindings::JavaString::~JavaString):
-        (KJS::Bindings::JavaString::UTF8String):
-        (KJS::Bindings::JavaString::uchars):
-        (KJS::Bindings::JavaString::length):
-        (KJS::Bindings::JavaString::ustring):
-
-        * bindings/jni/jni_utility.cpp:
-        (KJS::Bindings::convertArrayInstanceToJavaArray): Made this static, so 
-        the set of callers is known, and we can assert that we hold the JSLock. 
-        (KJS::Bindings::convertValueToJValue): Acquire the JSLock because doing
-        it recursively is pretty cheap, and it's just too confusing to tell whether
-        all our callers do it for us.
-
-        * bindings/objc/objc_runtime.h: Nixed some dead code.
-        * bindings/objc/objc_utility.mm:
-        (KJS::Bindings::convertNSStringToString): Same drill as above.
-
-2007-03-18  Alexey Proskuryakov  <ap@webkit.org>
-
-        Reviewed by Geoff.
-
-        http://bugs.webkit.org/show_bug.cgi?id=13105
-        REGRESSION: an exception raised when calculating base value of a dot expression is not returned
-
-        Test: fast/js/dot-node-base-exception.html
-
-        * kjs/nodes.cpp:
-        (FunctionCallDotNode::evaluate): Added the necessary KJS_CHECKEXCEPTIONVALUE.
-
-2007-03-18  Steve Falkenburg  <sfalken@apple.com>
-
-        Build fix.
-
-        * JavaScriptCore.vcproj/testkjs/testkjs.vcproj:
-
-2007-03-17  Timothy Hatcher  <timothy@apple.com>
-
-        Reviewed by Mark Rowe.
-
-        Made Version.xcconfig smarter when building for different configurations.
-        Now uses the 522+ OpenSource version for Debug and Release, while using the
-        full 522.4 version for Production builds. The system prefix is also computed
-        based on the current system, so 4522.4 on Tiger and 5522.4 on Leopard.
-
-        * Configurations/JavaScriptCore.xcconfig:
-        * Configurations/Version.xcconfig:
-
-2007-03-15  Maciej Stachowiak  <mjs@apple.com>
-
-        Not reviewed.
-        
-        - build fix
-
-        * wtf/TCSystemAlloc.cpp:
-
-2007-03-15  Maciej Stachowiak  <mjs@apple.com>
-
-        Reviewed by Geoff and Steve.
-        
-        - fix some portability issues with TCMalloc.
-
-        * JavaScriptCore.vcproj/WTF/WTF.vcproj:
-        * kjs/config.h:
-        * wtf/FastMalloc.cpp:
-        (WTF::SizeClass):
-        (WTF::InitSizeClasses):
-        (WTF::TCMalloc_PageHeap::Split):
-        (WTF::TCMalloc_PageHeap::RegisterSizeClass):
-        (WTF::TCMalloc_Central_FreeList::length):
-        (WTF::TCMalloc_ThreadCache::InitTSD):
-        (WTF::TCMalloc_ThreadCache::CreateCacheIfNecessary):
-        * wtf/TCSpinLock.h:
-        * wtf/TCSystemAlloc.cpp:
-        (TryVirtualAlloc):
-        (TCMalloc_SystemAlloc):
-
-2007-03-15  Timothy Hatcher  <timothy@apple.com>
-
-        Reviewed by John.
-
-        * Factored out most of our common build settings into .xcconfig files. Anything that was common in
-          each build configuration was factored out into the shared .xcconfig file.
-        * Adds a Version.xcconfig file to define the current framework version, to be used in other places.
-        * Use the new $(BUNDLE_VERSION) (defined in Version.xcconfig) in the preprocessed Info.plist.
-        * Use the versions defined in Version.xcconfig to set $(DYLIB_CURRENT_VERSION).
-
-        * Configurations/Base.xcconfig: Added.
-        * Configurations/DebugRelease.xcconfig: Added.
-        * Configurations/JavaScriptCore.xcconfig: Added.
-        * Configurations/Version.xcconfig: Added.
-        * Info.plist:
-        * JavaScriptCore.xcodeproj/project.pbxproj:
-
-2007-03-16  Shrikant Gangoda  <shrikant.gangoda@celunite.com>
-
-        Gdk build fix.
-
-        * kjs/DateMath.cpp:  gettimeofday comes from <sys/time.h> on Linux.
-
-2007-03-14  Kevin McCullough  <kmccullough@apple.com>
-
-        Reviewed by .
-
-        - Fixed one more build breakage
-
-        * kjs/date_object.cpp:
-        (KJS::formatLocaleDate):
-
-2007-03-14  Kevin McCullough  <kmccullough@apple.com>
-
-        Reviewed by .
-
-        - Fixed a build breakage.
-
-        * kjs/DateMath.cpp:
-        * kjs/date_object.cpp:
-        (KJS::formatLocaleDate):
-        (KJS::DateObjectImp::construct):
-
-2007-03-14  Kevin McCullough  <kmccullough@apple.com>
-
-        Reviewed by Geoff.
-
-        - rdar://problem/5045720
-        - DST changes in US affect JavaScript date calculations (12975)
-        This fix was to ensure we properly test for the new changes to DST in the US.
-        Also this fixes when we apply DST, now we correctly map most past years to current
-        DST rules.  We still have a small issue with years before 1900 or after 2100.
-        rdar://problem/5055038
-
-        * kjs/DateMath.cpp: Fix DST to match spec better.
-        (KJS::getCurrentUTCTime):
-        (KJS::mimimumYearForDST):
-        (KJS::maximumYearForDST):
-        (KJS::equivalentYearForDST):
-        (KJS::getDSTOffset):
-        * kjs/DateMath.h: Consolodated common funtionality.
-        * kjs/date_object.cpp: Consolodated common functionality.
-        (KJS::formatLocaleDate):
-        (KJS::DateObjectImp::construct):
-        * tests/mozilla/ecma/jsref.js: Added functions for finding the correct days when DST starts and ends.
-        * tests/mozilla/ecma/shell.js: Added back in the old DST functions for ease of merging with mozilla if needed.
-        * tests/mozilla/ecma_2/jsref.js: Added functions for finding the correct days when DST starts and ends.
-        * tests/mozilla/ecma_3/Date/shell.js: Added functions for finding the correct days when DST starts and ends.
-        * tests/mozilla/expected.html: Updated to show all date tests passing.
-
-=== Safari-5522.4 ===
-
-2007-03-13  Kevin McCullough  <kmccullough@apple.com>
-
-        Reviewed by .
-
-        - Adding expected failures until the are truly fixed. 
-        - rdar://problem/5060302
-
-        * tests/mozilla/expected.html:
-
-2007-03-12  Kevin McCullough  <kmccullough@apple.com>
-
-        Reviewed by .
-
-        - Actually update tests for new DST rules.
-
-        * tests/mozilla/ecma/Date/15.9.3.1-1.js:
-        * tests/mozilla/ecma/Date/15.9.3.1-2.js:
-        * tests/mozilla/ecma/Date/15.9.3.1-3.js:
-        * tests/mozilla/ecma/Date/15.9.3.1-4.js:
-        * tests/mozilla/ecma/Date/15.9.3.1-5.js:
-        * tests/mozilla/ecma/Date/15.9.3.2-1.js:
-        * tests/mozilla/ecma/Date/15.9.3.2-2.js:
-        * tests/mozilla/ecma/Date/15.9.3.2-3.js:
-        * tests/mozilla/ecma/Date/15.9.3.2-4.js:
-        * tests/mozilla/ecma/Date/15.9.3.2-5.js:
-        * tests/mozilla/ecma/Date/15.9.3.8-1.js:
-        * tests/mozilla/ecma/Date/15.9.3.8-2.js:
-        * tests/mozilla/ecma/Date/15.9.3.8-3.js:
-        * tests/mozilla/ecma/Date/15.9.3.8-4.js:
-        * tests/mozilla/ecma/Date/15.9.3.8-5.js:
-        * tests/mozilla/ecma/Date/15.9.5.10-1.js:
-        * tests/mozilla/ecma/Date/15.9.5.10-10.js:
-        * tests/mozilla/ecma/Date/15.9.5.10-11.js:
-        * tests/mozilla/ecma/Date/15.9.5.10-12.js:
-        * tests/mozilla/ecma/Date/15.9.5.10-13.js:
-        * tests/mozilla/ecma/Date/15.9.5.10-2.js:
-        * tests/mozilla/ecma/Date/15.9.5.10-3.js:
-        * tests/mozilla/ecma/Date/15.9.5.10-4.js:
-        * tests/mozilla/ecma/Date/15.9.5.10-5.js:
-        * tests/mozilla/ecma/Date/15.9.5.10-6.js:
-        * tests/mozilla/ecma/Date/15.9.5.10-7.js:
-        * tests/mozilla/ecma/Date/15.9.5.10-8.js:
-        * tests/mozilla/ecma/Date/15.9.5.10-9.js:
-        * tests/mozilla/ecma/jsref.js:
-        * tests/mozilla/ecma_2/jsref.js:
-        * tests/mozilla/ecma_3/Date/shell.js:
-
-2007-03-12  Kevin McCullough  <kmccullough@apple.com>
-
-        Reviewed by .
-
-        - Update tests for new DST rules.
-
-        * tests/mozilla/ecma/shell.js:
-
-2007-03-11  Geoffrey Garen  <ggaren@apple.com>
-
-        Reviewed by Oliver Hunt.
-        
-        Fixed <rdar://problem/4681051> Installer crashes in KJS::Collector::
-        markOtherThreadConservatively(KJS::Collector::Thread*) trying to install 
-        iLife 06 using Rosetta on an Intel Machine
-        
-        The problem was that our thread-specific data destructor would modify the
-        list of active JavaScript threads without holding the JSLock, corrupting
-        the list. Corruption was especially likely if one JavaScript thread exited 
-        while another was starting up.
-
-        * JavaScriptCore.exp:
-        * kjs/JSLock.cpp: Don't conflate locking the JSLock with registering a
-        thread, since the thread-specific data destructor needs to lock
-        without registering a thread. Instead, treat thread registration as a
-        part of the convenience of the JSLock object, and whittle down JSLock::lock()
-        to just the bits that actually do the locking.
-        (KJS::JSLock::lock):
-        (KJS::JSLock::registerThread):
-        * kjs/JSLock.h: Updated comments to mention the new behavior above, and
-        other recent changes.
-        (KJS::JSLock::JSLock):
-        * kjs/collector.cpp:
-        (KJS::destroyRegisteredThread): Lock here.
-        (KJS::Collector::registerThread): To match, assert that we're locked here.
-
-2007-03-10  Geoffrey Garen  <ggaren@apple.com>
-
-        Reviewed by Darin Adler.
-
-        Fixed <rdar://problem/4587763> PAC file: lock inversion between QT and 
-        JSCore causes a hang @ www.panoramas.dk
-        
-        With a PAC file, run-webkit-tests --threaded passes, the reported site
-        works, and all the Quicktime/JavaScript and Flash/JavaScript examples
-        I found through Google work, too.
-        
-        Any time JavaScript causes arbitrary non-JavaScript code to execute, it 
-        risks deadlock, because that code may block, trying to acquire a lock 
-        owned by a thread that is waiting to execute JavaScript. In this case,
-        the thread was a networking thread that was waiting to interpret a PAC file.
-        
-        Because non-JavaScript code may execute in response to, well, anything,
-        a perfect solution to this problem is impossible. I've implemented an
-        optimistic solution, instead: JavaScript will drop its lock whenever it
-        makes a direct call to non-JavaScript code through a bridging/plug-in API,
-        but will blissfully ignore the indirect ways it may cause non-JavaScript 
-        code to run (resizing a window, for example). 
-        
-        Unfortunately, this solution introduces significant locking overhead in 
-        the bridging APIs. I don't see a way around that.
-
-        This patch includes some distinct bug fixes I saw along the way:
-        
-        * bindings/objc/objc_instance.mm: Fixed a bug where a nested begin() call
-        would leak its autorelease pool, because it would NULL out _pool without
-        draining it.
-
-        * bindings/runtime_object.cpp:
-        (RuntimeObjectImp::methodGetter): Don't copy an Identifier to ASCII only
-        to turn around and make an Identifier from the ASCII. In an earlier 
-        version of this patch, the copy caused an assertion failure. Now it's 
-        just unnecessary work.
-        (RuntimeObjectImp::getOwnPropertySlot): ditto
-
-        * bindings/objc/objc_instance.h: Removed overrides of setVAlueOfField and
-        getValueOfField, because they did exactly what the base class versions did.
-        Removed overrides of Noncopyable declarations for the same reason.
-
-        * bindings/runtime.h: Inherit from Noncopyable instead of rolling our own.
-        * bindings/c/c_instance.h: ditto
-
-        And the actual patch:
-        
-        * API/JSCallbackConstructor.cpp: Drop all locks when calling out to C.
-        (KJS::JSCallbackConstructor::construct):
-        * API/JSCallbackFunction.cpp: ditto
-        (KJS::JSCallbackFunction::callAsFunction):
-        * API/JSCallbackObject.cpp: ditto
-        (KJS::JSCallbackObject::init):
-        (KJS::JSCallbackObject::~JSCallbackObject):
-        (KJS::JSCallbackObject::getOwnPropertySlot):
-        (KJS::JSCallbackObject::put):
-        (KJS::JSCallbackObject::deleteProperty):
-        (KJS::JSCallbackObject::construct):
-        (KJS::JSCallbackObject::hasInstance):
-        (KJS::JSCallbackObject::callAsFunction):
-        (KJS::JSCallbackObject::getPropertyNames):
-        (KJS::JSCallbackObject::toNumber):
-        (KJS::JSCallbackObject::toString):
-        (KJS::JSCallbackObject::staticValueGetter):
-        (KJS::JSCallbackObject::callbackGetter):
-        
-        * bindings/c/c_instance.cpp: Drop all locks when calling out to C.
-        (KJS::Bindings::CInstance::invokeMethod):
-        (KJS::Bindings::CInstance::invokeDefaultMethod):
-        * bindings/c/c_runtime.cpp: Drop all locks when calling out to C.
-        (KJS::Bindings::CField::valueFromInstance):
-        (KJS::Bindings::CField::setValueToInstance):
-        * bindings/jni/jni_objc.mm:
-        (KJS::Bindings::dispatchJNICall): Drop all locks when calling out to Java.
-
-        * bindings/objc/objc_instance.mm: The changes here are to accomodate the
-        fact that C++ unwinding of DropAllLocks goes crazy when you put it inside
-        a @try block. I moved all JavaScript stuff outside of the @try blocks, and 
-        then prefixed the whole blocks with DropAllLocks objects. This required some
-        supporting changes in other functions, which now acquire the JSLock for
-        themselves, intead of relying on their callers to do so.
-        (ObjcInstance::end):
-        (ObjcInstance::invokeMethod):
-        (ObjcInstance::invokeDefaultMethod):
-        (ObjcInstance::setValueOfUndefinedField):
-        (ObjcInstance::getValueOfUndefinedField):
-        * bindings/objc/objc_runtime.mm: Same as above, except I didn't want to
-        change throwError to acquire the JSLock for itself.
-        (ObjcField::valueFromInstance):
-        (ObjcField::setValueToInstance):
-        * bindings/objc/objc_utility.mm: Supporting changes mentioned above.
-        (KJS::Bindings::convertValueToObjcValue):
-        (KJS::Bindings::convertObjcValueToValue):
-
-        * kjs/JSLock.cpp: 
-        (1) Fixed DropAllLocks to behave as advertised, and drop the JSLock only 
-        if the current thread actually acquired it in the first place. This is 
-        important because WebKit needs to ensure that the JSLock has been 
-        dropped before it makes a plug-in call, even though it doesn't know if 
-        the current thread actually acquired the JSLock. (We don't want WebKit
-        to accidentally drop a lock belonging to *another thread*.)
-        (2) Used the new per-thread code written for (1) to make recursive calls
-        to JSLock very cheap. JSLock now knows to call pthread_mutext_lock/ 
-        pthread_mutext_unlock only at nesting level 0.
-        (KJS::createDidLockJSMutex):
-        (KJS::JSLock::lock):
-        (KJS::JSLock::unlock):
-        (KJS::DropAllLocks::DropAllLocks):
-        (KJS::DropAllLocks::~DropAllLocks):
-        (KJS::JSLock::lockCount):
-        * kjs/JSLock.h: Don't duplicate Noncopyable.
-        (KJS::JSLock::~JSLock):
-
-        * wtf/Assertions.h: Blind attempt at helping the Windows build.
-
-2007-03-08  MorganL  <morganl.webkit@yahoo.com>
-
-        Reviewed by Darin.
-
-        http://bugs.webkit.org/show_bug.cgi?id=13018
-        Bug 13018: allow embedders to override the definition of CRASH.
-
-        * wtf/Assertions.h: make it possible to override CRASH.
-
-2007-03-07  Anrong Hu  <huanr@yahoo.com>
-
-        Reviewed by Maciej.
-
-        Fix http://bugs.webkit.org/show_bug.cgi?id=12535
-        Bug 12535: Stack-optimizing compilers can trick GC into freeing in-use objects
-
-        * kjs/internal.cpp:
-        (KJS::StringImp::toObject): Copy val onto the stack so it is not subject to garbage collection.
-
-2007-03-07  Geoffrey Garen  <ggaren@apple.com>
-
-        Build fix for non-multiple-thread folks.
-        
-        Use a shared global in the non-multiple-thread case.
-
-        * wtf/FastMalloc.cpp:
-        (WTF::isForbidden):
-        (WTF::fastMallocForbid):
-        (WTF::fastMallocAllow):
-
-2007-03-07  Geoffrey Garen  <ggaren@apple.com>
-
-        Reviewed by Darin Adler.
-        
-        Fixed ASSERT failure I just introduced.
-        
-        Made the fastMalloc isForbidden flag per thread. (Oops!) We expect that
-        other threads will malloc while we're marking -- we just want to prevent
-        our own marking from malloc'ing.
-
-        * wtf/FastMalloc.cpp:
-        (WTF::initializeIsForbiddenKey):
-        (WTF::isForbidden):
-        (WTF::fastMallocForbid):
-        (WTF::fastMallocAllow):
-        (WTF::fastMalloc):
-        (WTF::fastCalloc):
-        (WTF::fastFree):
-        (WTF::fastRealloc):
-        (WTF::do_malloc):
-
-2007-03-07  Shrikant Gangoda  <shrikant.gangoda@celunite.com>
-
-        Reviewed by Maciej.
-
-        http://bugs.webkit.org/show_bug.cgi?id=12997
-
-        Wrap pthread-specific assertion in #if USE(MULTIPLE_THREADS).
-
-        * kjs/collector.cpp:
-        (KJS::Collector::markMainThreadOnlyObjects):
-
-2007-03-06  Geoffrey Garen  <ggaren@apple.com>
-
-        Reviewed by Maciej Stachowiak.
-        
-        Fixed <rdar://problem/4576242> | http://bugs.webkit.org/show_bug.cgi?id=12586
-        PAC file: malloc deadlock sometimes causes a hang @ www.apple.com/pro/profiles/ (12586)
-        
-        This is a modified version of r14752 on the branch.
-        
-        These changes just add debugging functionality. They ASSERT that we don't 
-        malloc during the mark phase of a garbage collection, which can cause a
-        deadlock.
-
-        * kjs/collector.cpp:
-        (KJS::Collector::collect):
-        * wtf/FastMalloc.cpp:
-        (WTF::fastMallocForbid):
-        (WTF::fastMallocAllow):
-        (WTF::fastMalloc):
-        (WTF::fastCalloc):
-        (WTF::fastFree):
-        (WTF::fastRealloc):
-        (WTF::do_malloc):
-        * wtf/FastMalloc.h:
-
-2007-03-06  Geoffrey Garen  <ggaren@apple.com>
-
-        Reviewed by Maciej Stachowiak.
-        
-        Fixed all known crashers exposed by run-webkit-tests --threaded. This covers:
-
-        <rdar://problem/4565394> | http://bugs.webkit.org/show_bug.cgi?id=12585 
-            PAC file: after closing a window that contains macworld.com, new window 
-            crashes (KJS::PropertyMap::mark()) (12585)
-        <rdar://problem/4571215> | http://bugs.webkit.org/show_bug.cgi?id=9211
-            PAC file: Crash occurs when clicking on the navigation tabs at http://www.businessweek.com/ (9211)
-        <rdar://problem/4557926> 
-            PAC file: Crash occurs when attempting to view image in slideshow mode 
-            at http://d.smugmug.com/gallery/581716 ( KJS::IfNode::execute (KJS::
-            ExecState*) + 312) if you use a PAC file
-
-        (1) Added some missing JSLocks, along with related ASSERTs.
-        
-        (2) Fully implemented support for objects that can only be garbage collected
-        on the main thread. So far, only WebCore uses this. We can add it to API
-        later if we learn that it's needed. 
-        
-        The implementation uses a "main thread only" flag inside each object. When 
-        collecting on a secondary thread, the Collector does an extra pass through 
-        the heap to mark all flagged objects before sweeping. This solution makes
-        the common case -- flag lots of objects, but never collect on a secondary 
-        thread -- very fast, even though the uncommon case of garbage collecting
-        on a secondary thread isn't as fast as it could be. I left some notes 
-        about how to speed it up, if we ever care.
-        
-        For posterity, here are some things I learned about GC while investigating:
-        
-        * Each collect must either mark or delete every heap object. "Zombie" 
-        objects, which are neither marked nor deleted, raise these issues:
-
-            * On the next pass, the conservative marking algorithm might mark a 
-            zombie, causing it to mark freed objects.
-
-            * The client might try to use a zombie, which would seem live because 
-            its finalizer had not yet run.
-
-        * A collect on the main thread is free to delete any object. Presumably, 
-        objects allocated on secondary threads have thread-safe finalizers.
-
-        * A collect on a secondary thread must not delete thread-unsafe objects.
-
-        * The mark function must be thread-safe.
-        
-        Line by line comments:
-
-        * API/JSObjectRef.h: Added comment specifying that the finalize callback 
-        may run on any thread.
-
-        * JavaScriptCore.exp: Nothing to see here.
-
-        * bindings/npruntime.cpp:
-        (_NPN_GetStringIdentifier): Added JSLock.
-
-        * bindings/objc/objc_instance.h:
-        * bindings/objc/objc_instance.mm:
-        (ObjcInstance::~ObjcInstance): Use an autorelease pool. The other callers 
-        to CFRelease needed one, too, but they were dead code, so I removed them 
-        instead. (This fixes a leak seen while running run-webkit-tests --threaded,
-        although I don't think it's specifically a threading issue.) 
-        
-        * kjs/collector.cpp:
-        (KJS::Collector::collectOnMainThreadOnly): New function. Tells the collector
-        to collect a value only if it's collecting on the main thread.
-        (KJS::Collector::markMainThreadOnlyObjects): New function. Scans the heap
-        for "main thread only" objects and marks them.
-
-        * kjs/date_object.cpp: 
-        (KJS::DateObjectImp::DateObjectImp): To make the new ASSERTs happy, allocate 
-        our globals on the heap, avoiding a seemingly unsafe destructor call at 
-        program exit time.
-        * kjs/function_object.cpp:
-        (FunctionPrototype::FunctionPrototype): ditto
-
-        * kjs/interpreter.cpp:
-        (KJS::Interpreter::mark): Removed boolean parameter, which was an incomplete
-        and arguably hackish way to implement markMainThreadOnlyObjects() inside WebCore.
-        * kjs/interpreter.h:
-
-        * kjs/identifier.cpp:
-        (KJS::identifierTable): Added some ASSERTs to check for thread safety 
-        problems.
-
-        * kjs/list.cpp: Added some ASSERTs to check for thread safety problems.
-        (KJS::allocateListImp):
-        (KJS::List::release):
-        (KJS::List::append):
-        (KJS::List::empty): Make the new ASSERTs happy.
-
-        * kjs/object.h:
-        (KJS::JSObject::JSObject): "m_destructorIsThreadSafe" => "m_collectOnMainThreadOnly".
-        I removed the constructor parameter because m_collectOnMainThreadOnly,
-        like m_marked, is a Collector bit, so only the Collector should set or get it.
-
-        * kjs/object_object.cpp:
-        (ObjectPrototype::ObjectPrototype): Make the ASSERTs happy.
-        * kjs/regexp_object.cpp:
-        (RegExpPrototype::RegExpPrototype): ditto
-
-        * kjs/ustring.cpp: Added some ASSERTs to check for thread safety problems.
-        (KJS::UCharReference::ref): 
-        (KJS::UString::Rep::createCopying):
-        (KJS::UString::Rep::create):
-        (KJS::UString::Rep::destroy):
-        (KJS::UString::null): Make the new ASSERTs happy.
-        * kjs/ustring.h:
-        (KJS::UString::Rep::ref): Added some ASSERTs to check for thread safety problems.
-        (KJS::UString::Rep::deref):
-
-        * kjs/value.h:
-        (KJS::JSCell::JSCell):
-
-2007-03-06  Geoffrey Garen  <ggaren@apple.com>
-
-        Reviewed by Maciej Stachowiak.
-        
-        2% speedup on super accurate JS iBench.
-
-        (KJS::Collector::collect): Removed anti-optimization to call
-        pthread_is_threaded_np() before calling pthread_main_np(). Almost all 
-        apps have more than one thread, so the extra call is actually worse.
-        Interestingly, even the single-threaded testkjs shows a speed gain
-        from removing the pthread_is_threaded_np() short-circuit. Not sure why.
-
-2007-03-04  Don Gibson  <dgibson77@gmail.com>
-
-        Reviewed by Nikolas Zimmermann.
-
-        - fix http://bugs.webkit.org/show_bug.cgi?id=12950
-          Assertions.cpp should not #define macros that are already defined
-
-        * wtf/Assertions.cpp: Don't #define WINVER and _WIN32_WINNT if they
-        are already defined.
-
-2007-03-02  Steve Falkenburg  <sfalken@apple.com>
-
-        Reviewed by Anders.
-        
-        Add unsigned int hash traits (matches existing unsigned long version)
-
-        * wtf/HashTraits.h:
-        (WTF::):
-
-2007-03-02  Adam Roben  <aroben@apple.com>
-
-        Reviewed by Kevin M.
-
-        Try to fix the Qt build.
-
-        * kjs/DateMath.cpp:
-        (KJS::msToGregorianDateTime): Removed unnecessary "struct" keyword.
-        * kjs/DateMath.h: Moved forward declarations to the top of the file
-        before they are used.
-        * kjs/date_object.cpp:
-        (KJS::formatLocaleDate): Changed to take a const GregorianDateTime&
-        since GregorianDateTime is Noncopyable.
-
-2007-03-02  Darin Adler  <darin@apple.com>
-
-        Reviewed by Kevin McCullough.
-
-        - fix http://bugs.webkit.org/show_bug.cgi?id=12867
-          REGRESSION: BenchJS test 7 (dates) is 220% slower than in Safari 2.0.4
-
-        * kjs/DateMath.h: Marked GregorianDateTime as noncopyable, since it has a non-trivial
-        destructor and not the correspoding copy constructor or assignment operator.
-        Changed the GregorianDateTime constructor to use member initialization syntax.
-        Fixed the destructor to use the array delete operator, since timeZone is an array.
-
-        * kjs/DateMath.cpp:
-        (KJS::daysInYear): Changed to call isLeapYear so the rule is not repeated twice.
-        (KJS::getUTCOffset): Added caching on PLATFORM(DARWIN), since we can rely on the
-        notify_check function and "com.apple.system.timezone" to let us know when the
-        offset has changed.
-
-2007-02-27  Geoffrey Garen  <ggaren@apple.com>
-
-        Reviewed by Darin Adler.
-        
-        Follow-up to fixing http://bugs.webkit.org/show_bug.cgi?id=12659 | <rdar://problem/4954306>
-        JS objects not collected after closing window @ ebay.com/maps.google.com
-        
-        Changed Interpreter cache of global constructors and prototypes from
-        ProtectedPtrs to bare, marked pointers. ProtectedPtrs are inefficient,
-        and they increase the risk of reference cycles. Also, Darin said something
-        about ProtectedPtrs giving him warts.
-        
-        Also changed data members to precise types from generic JSObject*'s.
-        
-        Layout tests and JS tests pass.
-
-        * kjs/SavedBuiltins.h:
-        * kjs/interpreter.cpp:
-        (KJS::Interpreter::init):
-        (KJS::Interpreter::~Interpreter):
-        (KJS::Interpreter::initGlobalObject): Moved Identifier::init() call to
-        constructor, for clarity.
-        (KJS::Interpreter::mark):
-        * kjs/interpreter.h:
-
-2007-02-27  Geoffrey Garen  <ggaren@apple.com>
-
-        Reviewed by Maciej Stachowiak.
-        
-        Fixed http://bugs.webkit.org/show_bug.cgi?id=12659 | <rdar://problem/4954306>
-        JS objects not collected after closing window @ ebay.com/maps.google.com
-
-        Don't GC in the Interpreter destructor. For that to work, the Interpreter
-        would have to NULL out all of its ProtectedPtrs before calling collect(). But 
-        we've decided that we don't want things to work that way, anyway. We want the
-        client to be in charge of manual GC so that it can optimize cases when
-        it will be destroying many interpreters at once
-        (e.g., http://bugs.webkit.org/show_bug.cgi?id=12900).
-        
-        Also removed Interpreter::collect() because it was redundant with 
-        Collector::collect().
-
-        * JavaScriptCore.exp:
-        * kjs/interpreter.cpp:
-        (KJS::Interpreter::~Interpreter):
-        * kjs/testkjs.cpp:
-        (TestFunctionImp::callAsFunction):
-
-2007-02-26  Krzysztof Kowalczyk  <kkowalczyk@gmail.com>
-
-        Reviewed by Adam Roben.
-
-        Rename *_SUPPORT defines to ENABLE_*.
-
-        * jscore.bkl:
-
-2007-02-26  Maciej Stachowiak  <mjs@apple.com>
-
-        Reviewed by Lars.
-        
-        - <rdar://problem/5021698> Disable experimental SVG features (12883)
-
-        * wtf/Platform.h: Add ENABLE() macro similar to HAVE() and USE(), to
-        allow nicer handling of optional WebKit features.
-
-2007-02-22  George Staikos  <staikos@kde.org>
-
-        Reviewed by Lars.
-
-        Add return values
-
-        * wtf/unicode/qt4/UnicodeQt4.h:
-        (WTF::Unicode::toLower):
-        (WTF::Unicode::toUpper):
-
-2007-02-22  Oscar Cwajbaum  <public@oscarc.net>
-
-        Reviewed by Maciej.
-
-        Fix ARM-specific alignment problem in FastMalloc
-        http://bugs.webkit.org/show_bug.cgi?id=12841
-
-        * wtf/FastMalloc.cpp:
-        Modify how pageheap_memory is declared to ensure proper alignment
-        on architectures such as ARM
-
-2007-02-20  Zack Rusin  <zrusin@trolltech.com>
-
-        Reviewed by Lars
-
-        Make sure that non-void methods always return something.
-
-        * wtf/unicode/qt4/UnicodeQt4.h:
-        (WTF::Unicode::toLower):
-        (WTF::Unicode::toUpper):
-        (WTF::Unicode::foldCase):
-
-2007-02-18  Kevin Ollivier  <kevino@theolliviers.com>
-
-        Reviewed by Adam Roben.
-
-        Fix cases where MSVC-specific code was identified as Win32 platform
-        code. (as it should be compiled for e.g. wx port when using MSVC too)
-        
-        * wtf/Assertions.h: 
-        * wtf/MathExtras.h:
-        * wtf/StringExtras.h:
-        changed PLATFORM(WIN) sections to COMPILER(MSVC) as necessary
-
-2007-02-17  Krzysztof Kowalczyk  <kkowalczyk@gmail.com>
-
-        Reviewed by Adam Roben.
-
-        Fix crashes on ARM due to different struct packing. Based on a patch
-        by Mike Emmel.
-        * kjs/ustring.cpp: compile-time assert to make sure sizeof(UChar) == 2
-        * kjs/ustring.h: pack UChar struct to ensure that sizeof(UChar) == 2
-        * wtf/Assertions.h: add COMPILE_ASSERT macro for compile-time assertions
-
-2007-02-16  George Staikos  <staikos@kde.org>
-
-        Reviewed by Maciej.
-
-        Fix uninitialized variable
-
-        * bindings/testbindings.cpp:
-        (myAllocate):
-
-2007-02-16  Anders Carlsson  <acarlsson@apple.com>
-
-        Reviewed by Mitz.
-
-        http://bugs.webkit.org/show_bug.cgi?id=12788
-        REGRESSION: Going back one page in history has a noticeable delay
-        
-        Um...if all elements in two vectors are equal, then I guess we could say that
-        the two vectors are equal too.
-        
-        * wtf/Vector.h:
-        (WTF::):
-
-2007-02-14  Anders Carlsson  <acarlsson@apple.com>
-
-        Reviewed by Darin.
-
-        Add new canCompareWithMemcmp vector trait and use it to determine whether
-        operator== can use memcmp.
-        
-        * wtf/Vector.h:
-        (WTF::):
-        (WTF::VectorTypeOperations::compare):
-        (WTF::operator==):
-        * wtf/VectorTraits.h:
-        (WTF::):
-
-2007-02-13  Brady Eidson  <beidson@apple.com>
-
-        Reviewed by Darin
-        
-        Tweaked vector a bit
-
-        * wtf/Vector.h:
-        (WTF::operator==):
-
-2007-02-13  Dex Deacon  <occupant4@gmail.com>
-
-        Reviewed by Darin.
-
-        - fix for http://bugs.webkit.org/show_bug.cgi?id=12750
-          Vector operator== was not defined correctly. It returned void,
-          did not accept const Vectors, and used an int instead of size_t. 
-
-        * wtf/Vector.h: fixed comparison operators
-        (WTF::operator==):
-        (WTF::operator!=):
-
-2007-02-10  David Carson  <dacarson@gmail.com>
-
-        Reviewed by Maciej.
-
-        - fix for http://bugs.webkit.org/show_bug.cgi?id=12636
-        Corrected the generation of method signatures when the parameter
-        is an Array. 
-        Added support for converting a Javascript array to a Java array.
-
-        * bindings/jni/jni_utility.h: added new type for array, array_type
-        * bindings/jni/jni_runtime.cpp: add support for new array type
-        (JavaField::valueFromInstance):
-        (JavaField::setValueToInstance):
-        (JavaMethod::JavaMethod):
-        (JavaMethod::signature):
-        * bindings/jni/jni_utility.cpp: add support for new array type
-        (KJS::Bindings::callJNIMethod):
-        (KJS::Bindings::callJNIStaticMethod):
-        (KJS::Bindings::callJNIMethodIDA):
-        (KJS::Bindings::JNITypeFromClassName):
-        (KJS::Bindings::signatureFromPrimitiveType):
-        (KJS::Bindings::JNITypeFromPrimitiveType):
-        (KJS::Bindings::getJNIField):
-        (KJS::Bindings::convertArrayInstanceToJavaArray): new method
-        converts the Javascript array to the requested Java array.
-        (KJS::Bindings::convertValueToJValue):
-
-2007-02-08  Anders Carlsson  <acarlsson@apple.com>
-
-        Reviewed by Geoff.
-
-        <rdar://problem/4930614>
-        Safari complains about "Slow Script" if GMail is left open and machine is busy
-        
-        <rdar://problem/4649516>
-        Turn off slow script dialog or crank up time that makes it come up
-        
-        <rdar://problem/4963589>
-        Slow script warning is displayed after closing of PROMPT or PRINT dialog
-        
-        Re-do the way script timeouts are handled. No longer use a unix timer that sends signals. Instead, add a 
-        tick count and increment it in loop bodies. If the tick count reaches a threshold, do a timeout check. If the total time executing
-        is higher than the timeout value, (possibly) interrupt the script. The timeout checker also adjusts the threshold dynamically
-        to prevent doing the timeout check too often.
-         
-        * JavaScriptCore.exp:
-        Remove pause and resume calls.
-        
-        * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
-        Add winmm.lib.
-
-        * kjs/interpreter.cpp:
-        (KJS::Interpreter::init):
-        (KJS::Interpreter::~Interpreter):
-        (KJS::Interpreter::startTimeoutCheck):
-        (KJS::Interpreter::stopTimeoutCheck):
-        (KJS::Interpreter::resetTimeoutCheck):
-        (KJS::getCurrentTime):
-        (KJS::Interpreter::checkTimeout):
-        * kjs/interpreter.h:
-        (KJS::Interpreter::timedOut):
-        * kjs/nodes.cpp:
-        (DoWhileNode::execute):
-        (WhileNode::execute):
-        (ForNode::execute):
-
-2007-02-07  Darin Adler  <darin@apple.com>
-
-        * JavaScriptCore.vcproj/JavaScriptCore.sln: Reenable testkjs.
-
-2007-02-07  Darin Adler  <darin@apple.com>
-
-        Reviewed by Geoff.
-
-        - another build fix; this time for sure
-
-        * pcre/pcre_exec.c: (match):
-        The compiler caught an incorrect use of the othercase variable across
-        a call to RMATCH in character repeat processing. Local variables can
-        change in the crazy NO_RECURSE mode that we use, so we instead need
-        the value in othercase to be in one of the special stack frame variables.
-        Added a new stack frame variable for this purpose named repeat_othercase.
-        Also noted a similar error in the non-UTF-16 side of the #ifdef, but
-        didn't try to fix that one. Also removed a SUPPORT_UCP #ifdef from the
-        PCRE_UTF16 side; that code doesn't work without the Unicde properties
-        table, and we don't try to use it that way.
-
-2007-02-06  Steve Falkenburg  <sfalken@apple.com>
-
-        Disable testkjs in sln until we figure out mysterious compiler warning.
-
-        * JavaScriptCore.vcproj/JavaScriptCore.sln:
-
-2007-02-06  Steve Falkenburg  <sfalken@apple.com>
-
-        Build fix by ggaren
-
-        * pcre/pcre_exec.c:
-        (match):
-
-2007-02-06  Darin Adler  <darin@apple.com>
-
-        Reviewed by Geoff.
-
-        - fix <rdar://problem/4979089> PCRE should avoid setjmp/longjmp even when compiler
-          is not GCC
-
-        Added a new code path that's slower and way uglier but doesn't rely on GCC's
-        computed gotos.
-
-        * pcre/pcre_exec.c: Added a numeric parameter to the RMATCH function. It must be
-        different at every RMATCH call site. Changed the non-GCC NO_RECURSE version of
-        the macro to use a label incorporating the number. Changed the RRETURN macro to
-        use a goto instead of longjmp.
-        (match): Added a different number at each callsite, using a perl script for the
-        first-time task. Going forward it should be easy to maintain by hand. Added a
-        switch statement at the bottom of the function. We'll get compile time errors
-        if we have anything in the switch statement that's never used in an RMATCH,
-        but errors in the other direction are silent except at runtime.
-
-2007-02-06  Darin Adler  <darin@apple.com>
-
-        Reviewed by John.
-
-        - fix <rdar://problem/4687840> 9A241: JavaScript RegExp 25-30x slower than on 10.4.7
-
-        I used Shark to figure out what to do. The test case is now 15% faster than with
-        stock Safari. Some other regular expression cases might still be a few % slower
-        than before, but the >10x slowdown is now completely gone.
-
-        1) Fix slowness caused by setjmp/longjmp by using computed goto instead.
-
-        Use GCC extensions - locally declared labels, labels as values, and computed goto -
-        instead of using setjmp/longjmp to implemement non-recursive version of the regular
-        expression system. We could probably make this even faster if we reduced the use
-        of malloc a bit too.
-
-        2) Fix slowness caused by allocating heapframe objects by allocating the first
-           16 of them from the stack.
-
-        3) Speed up use of malloc and free in PCRE by making it use fastMalloc and fastFree.
-
-        4) Speed up the test case by adding a special case to a UString function.
-
-        5) Made a small improvement to the innermost hottest loop of match by hoisting
-           the conversion from int to pcre_uchar out of the loop.
-
-        * JavaScriptCore.xcodeproj/project.pbxproj: Compile FastMallocPCRE.cpp, and don't
-        compile pcre_globals.c.
-
-        * wtf/FastMallocPCRE.cpp: Added. A copy of pcre_globals.c that uses FastMalloc.h.
-        This is better than code that sets the PCRE allocation globals because by doing it
-        this way there's guaranteed to be no problem with order of initialization.
-
-        * kjs/ustring.cpp: (KJS::UString::spliceSubstringsWithSeparators): Add a fast
-        special case when this is called for only one subrange and no seaprators. This
-        was happening a lot in the test case and it seems quite reasonable to optimize this.
-
-        * pcre/pcre_exec.c: Create a copy of the RMATCH and RRETURN macros that use goto
-        instead of setjmp/longjmp. Change code that calls pcre_stack_malloc to first use
-        storage on the stack inside the match function.
-        (match): Move initialization of utf8 up a couple lines to avoid "possibly used
-        uninitialized" warning. Use a local variable so we compare with pcre_uchar instead
-        of with int inside the inner "find a character" loop.
-
-2007-02-03  George Staikos  <staikos@kde.org>
-
-        Reviewed by Alexey.
-
-        -1 is not a valid point.  We can't handle anything > 0xffff anyway.
-        Fixes crash on cases like eval("x");
-
-        * wtf/unicode/qt4/UnicodeQt4.h:
-        (WTF::Unicode::category):
-
-2007-02-02  Darin Adler  <darin@apple.com>
-
-        Reviewed by Anders.
-
-        - fix copying and assigning a ListHashSet
-
-        No test because the code path with bugs I am fixing is not used yet.
-
-        * wtf/ListHashSet.h: Tweaked ListHashSetNodeAllocator a little bit for clarity.
-        Changed m_allocator to be an OwnPtr instead of doing an explicit delete.
-        Fixed bug in copy constructor where we'd have an uninitialized m_allocator.
-        Fixed bug in assignment operator where it would swap only the hash table, and
-        not the head, tail, and allocator pointers.
-
-2007-02-02  Geoffrey Garen  <ggaren@apple.com>
-
-        Reviewed by Maciej Stachowiak.
-        
-        Use WTFLog instead of fprintf for logging KJS::Node leaks.
-
-        * kjs/nodes.cpp:
-        (NodeCounter::~NodeCounter): Changed count to unsigned, updated
-        to match style guidelines.
-
-2007-02-02  Maciej Stachowiak  <mjs@apple.com>
-
-        - not reviewed, build fix
-
-        * wtf/ListHashSet.h:
-        (WTF::ListHashSetNodeAllocator::ListHashSetNodeAllocator): ummm, use union correctly
-
-2007-02-01  Maciej Stachowiak  <mjs@apple.com>
-
-        Reviewed by Darin.
-        
-        - use a custom allocator for ListHashSet, to fix ~1% perf regression using it for form control
-
-        * wtf/ListHashSet.h:
-        (WTF::ListHashSetNodeAllocator::ListHashSetNodeAllocator):
-        (WTF::ListHashSetNodeAllocator::allocate):
-        (WTF::ListHashSetNodeAllocator::deallocate):
-        (WTF::ListHashSetNode::operator new):
-        (WTF::ListHashSetNode::operator delete):
-        (WTF::ListHashSetNode::destroy):
-        (WTF::ListHashSetTranslator::translate):
-        (WTF::::ListHashSet):
-        (WTF::::~ListHashSet):
-        (WTF::::add):
-        (WTF::::unlinkAndDelete):
-        (WTF::::deleteAllNodes):
-
-2007-01-31  Maciej Stachowiak  <mjs@apple.com>
-
-        Reviewed by Adam.
-        
-        - fix sporadic crash
-
-        * wtf/ListHashSet.h:
-        (WTF::::remove): remove before deleting
-
-2007-01-31  Maciej Stachowiak  <mjs@apple.com>
-
-        Reviewed by Mark with help from Lars.
-        
-        - added new ListHashSet class, which combines a hashtable and a linked list to provide a set
-        that keeps elements in inserted order
-        
-        This is to assist in fixing the following:
-        <rdar://problem/4751164> REGRESSION: Safari places text on incorrect button when returning to a page via back [10541]
-        http://bugs.webkit.org/show_bug.cgi?id=10541
-
-        * JavaScriptCore.vcproj/WTF/WTF.vcproj:
-        * JavaScriptCore.xcodeproj/project.pbxproj:
-        * wtf/HashTable.h:
-        (WTF::HashTable::find):
-        (WTF::HashTable::contains):
-        (WTF::::find):
-        (WTF::::contains):
-        * wtf/ListHashSet.h: Added.
-        (WTF::ListHashSetNode::ListHashSetNode):
-        (WTF::ListHashSetNodeHashFunctions::hash):
-        (WTF::ListHashSetNodeHashFunctions::equal):
-        (WTF::ListHashSetIterator::ListHashSetIterator):
-        (WTF::ListHashSetIterator::get):
-        (WTF::ListHashSetIterator::operator*):
-        (WTF::ListHashSetIterator::operator->):
-        (WTF::ListHashSetIterator::operator++):
-        (WTF::ListHashSetIterator::operator--):
-        (WTF::ListHashSetIterator::operator==):
-        (WTF::ListHashSetIterator::operator!=):
-        (WTF::ListHashSetIterator::operator const_iterator):
-        (WTF::ListHashSetIterator::node):
-        (WTF::ListHashSetConstIterator::ListHashSetConstIterator):
-        (WTF::ListHashSetConstIterator::get):
-        (WTF::ListHashSetConstIterator::operator*):
-        (WTF::ListHashSetConstIterator::operator->):
-        (WTF::ListHashSetConstIterator::operator++):
-        (WTF::ListHashSetConstIterator::operator--):
-        (WTF::ListHashSetConstIterator::operator==):
-        (WTF::ListHashSetConstIterator::operator!=):
-        (WTF::ListHashSetConstIterator::node):
-        (WTF::ListHashSetTranslator::hash):
-        (WTF::ListHashSetTranslator::equal):
-        (WTF::ListHashSetTranslator::translate):
-        (WTF::::ListHashSet):
-        (WTF::::operator):
-        (WTF::::~ListHashSet):
-        (WTF::::size):
-        (WTF::::capacity):
-        (WTF::::isEmpty):
-        (WTF::::begin):
-        (WTF::::end):
-        (WTF::::find):
-        (WTF::::contains):
-        (WTF::::add):
-        (WTF::::remove):
-        (WTF::::clear):
-        (WTF::::unlinkAndDelete):
-        (WTF::::appendNode):
-        (WTF::::deleteAllNodes):
-        (WTF::::makeIterator):
-        (WTF::::makeConstIterator):
-        (WTF::deleteAllValues):
-
-2007-01-30  Darin Adler  <darin@apple.com>
-
-        * kjs/DateMath.cpp: Fix license header to reflect LGPL as the first license
-        mentioned. We still mention the option of using under MPL or GPL since some
-        of this code came from the Mozilla project with those license terms.
-
-2007-01-30  Simon Hausmann  <hausmann@kde.org>
-
-        Reviewed by Zack.
-
-        Turned JavaScriptCore from a separate library into an includable
-        project, to combine it all into libWebKitQt.
-
-        * JavaScriptCore.pri: Added.
-        * JavaScriptCore.pro: Removed.
-        * kjs/testkjs.pro:
-
-2007-01-29  Geoffrey Garen  <ggaren@apple.com>
-
-        Reviewed by Maciej Stachowiak.
-
-        Fixed <rdar://problem/4485644> REGRESSION: JavaScriptCore has init routines
-        
-        The TCMalloc module now initializes, if needed, inside GetCache() and 
-        fastMallocSetIsMultiThreaded(). We leverage the same synchronization 
-        technique used for enabling / disabling the single-threaded optimization 
-        to synchronize initialization of the library without requiring a lock 
-        for every malloc.
-        
-        1,251 runs of tcmalloc_unittest, 2 runs of a custom, massively multi-threaded 
-        tcmalloc_unittest, and my custom version of the PLT show no regressions.
-        Super-accurate JS iBench reports a .24% regression, which is right at the
-        limit of its error range, so I'm declaring victory.
-
-        * wtf/FastMalloc.cpp:
-        (WTF::fastMallocSetIsMultiThreaded): Initialize, if needed. (InitModule()
-        checks the "if needed" part.)
-        (WTF::TCMalloc_ThreadCache::GetCache): Restored original TCMalloc code
-        inside #ifdef, for posterity. Added new initialization logic.
-        (WTF::TCMalloc_ThreadCache::InitModule): Call InitTSD(), since we don't
-        have a static initializer to call it for us, now. This means that fastMalloc
-        is not usable as a general libc allocator, but it never was, and if it were
-        the general libc allocator, we wouldn't be here in the first place, so whatever.
-        (WTF::TCMalloc_ThreadCache::InitTSD): Don't try to take the pageheap_lock, 
-        since InitModule already has it.
-
-2007-01-29  Kevin McCullough  <KMcCullough@apple.com>
-
-        Reviewed by Geoff and Oliver.
-
-        - rdar://problem/4955561
-        - missusing JavaScript shouldn't crash webkit.  Now it doesn't, in this case.
-
-        * bindings/objc/objc_runtime.mm:
-        (ObjcFallbackObjectImp::callAsFunction):
-        * bindings/runtime_method.cpp:
-        (RuntimeMethod::callAsFunction):
-        * bindings/runtime_object.cpp:
-        (RuntimeObjectImp::callAsFunction):
-
-2007-01-28  Geoffrey Garen  <ggaren@apple.com>
-
-        Reviewed by Maciej Stachowiak.
-        
-        First step in fixing <rdar://problem/4485644> REGRESSION: JavaScriptCore 
-        has init routines
-        
-        Don't rely on a static initializer to store the main thread's ID (which
-        we would use to detect allocations on secondary threads). Instead, require 
-        the caller to notify fastMalloc if it might allocate on a secondary thread.
-        
-        Also fixed what seemed like a race condition in do_malloc.
-        
-        tcmalloc_unittest and my custom versions of JS iBench and PLT show no
-        regressions.
-
-        * wtf/FastMalloc.cpp:
-        (WTF::fastMallocSetIsMultiThreaded): 
-        (1) Renamed from "fastMallocRegisterThread", which was a misleading name because 
-        not all threads need to register with fastMalloc -- only secondary threads 
-        need to, and only for the purpose of disabling its single-threaded optimization. 
-
-        (2) Use the pageheap_lock instead of a custom one, since we need to synchronize
-        with the read of isMultiThreaded inside CreateCacheIfNecessary. This is a new
-        requirement, now that we can't guarantee that the first call to CreateCacheIfNecessary
-        will occur on the main thread at init time, before any other threads have been created.
-
-        (WTF::TCMalloc_ThreadCache::CreateCacheIfNecessary):
-        (WTF::do_malloc): Reverted WTF change only to call GetCache() if size <= kMaxSize.
-        The WTF code would read phinited without holding the pageheap_lock, which
-        seemed like a race condition. Regardless, calling GetCache reduces the number 
-        of code paths to module initialization, which will help in writing the 
-        final fix for this bug.
-
-2007-01-28  David Kilzer  <ddkilzer@webkit.org>
-
-        Reviewed by Darin.
-
-        - fix http://bugs.webkit.org/show_bug.cgi?id=9815
-          JavaScript TypeError loading Dean Edwards' JS compressor/obfuscator
-
-        Creating a function using 'new Function()' was not setting its prototype with the
-        same flags as 'function() { }'.
-
-        Test: fast/js/function-prototype.html
-
-        * kjs/function_object.cpp:
-        (FunctionObjectImp::construct): Change flags from DontEnum|DontDelete|ReadOnly to
-        Internal|DontDelete to match FuncDeclNode::processFuncDecl() and
-        FuncExprNode::evaluate() in kjs/nodes.cpp.
-
-2007-01-27  Geoffrey Garen  <ggaren@apple.com>
-
-        Reviewed by Beth Dakin.
-        
-        Added some missing JSLocks, which might fix <rdar://problem/4889707>.
-
-        We need to lock whenever we might allocate memory because our FastMalloc
-        implementation requires clients to register their threads, which we do
-        through JSLock. 
-        
-        We also need to lock whenever modifying ref-counts because they're not 
-        thread-safe.
-
-        * API/JSObjectRef.cpp:
-        (JSClassCreate): Allocates memory
-        (JSClassRetain): Modifies a ref-count
-        (JSClassRelease): Modifies a ref-count
-        (JSPropertyNameArrayRetain): Modifies a ref-count
-        (JSPropertyNameArrayRelease): Modifies a ref-count
-        * API/JSStringRef.cpp:
-        (JSStringRetain): Modifies a ref-count
-        * API/JSValueRef.cpp:
-        (JSValueIsInstanceOfConstructor): Might allocate memory if an exception
-        is thrown.
-
-2007-01-27  Lars Knoll <lars@trolltech.com>
-
-        Fix the Qt build.
-
-        * bindings/qt/qt_instance.h:
-
-2007-01-25  Geoffrey Garen  <ggaren@apple.com>
-
-        Reviewed by Maciej Stachowiak.
-        
-        Fixed <rdar://problem/4608404> WebScriptObject's _rootObject lack 
-        of ownership policy causes crashes (e.g., in Dashcode)
-        
-        The old model for RootObject ownership was either to (1) leak them or (2) assign
-        them to a single owner -- the WebCore::Frame -- which would destroy them 
-        when it believed that all of its plug-ins had unloaded.
-        
-        This model was broken because of (1) and also because plug-ins are not the only 
-        RootObject clients. All Bindings clients are RootObjects clients, including 
-        applications, which outlive any particular WebCore::Frame.
-        
-        The new model for RootObject ownership is to reference-count them, with a
-        throw-back to the old model: The WebCore::Frame tracks the RootObjects
-        it creates, and invalidates them when it believes that all of its plug-ins 
-        have unloaded.
-        
-        We maintain this throw-back to avoid plug-in leaks, particularly from Java.
-        Java is completely broken when it comes to releasing JavaScript objects. 
-        Comments in our code allege that Java does not always call finalize when 
-        collecting objects. Moreoever, my own testing reveals that, when Java does 
-        notify JavaScript of a finalize, the data it provides is totally bogus.
-        
-        This setup is far from ideal, but I don't think we can do better without
-        completely rewriting the bindings code, and possibly part of the Java
-        plug-in / VM.
-        
-        Layout tests pass. No additional leaks reported. WebCore/manual-tests/*liveconnect*
-        and a few LiveConnect demos on the web also run without a hitch.
-        
-        const RootObject* => RootObject*, since we need to ref/deref
-        
-        * bindings/NP_jsobject.cpp:
-        (jsDeallocate): deref our RootObjects. Also unprotect or JSObject, instead
-        of just relying on the RootObject to do it for us when it's invalidated.
-        (_isSafeScript): Check RootObject validity.
-        (_NPN_CreateScriptObject): ditto
-        (_NPN_Invoke): ditto
-        (_NPN_Evaluate): ditto
-        (_NPN_GetProperty): ditto
-        (_NPN_SetProperty): ditto
-        (_NPN_RemoveProperty): ditto
-        (_NPN_HasProperty): ditto
-        (_NPN_HasMethod): ditto
-        (_NPN_SetException): ditto
-
-        * bindings/runtime_root.cpp: 
-        Revived bit-rotted LIAR LIAR LIAR comment.
-        
-        LOOK: Added support for invalidating RootObjects without deleting them, 
-        which is the main goal of this patch. 
-
-        Moved protect counting into the RootObject class, to emphasize that 
-        the RootObject protects the JSObject, and unprotects it upon being invalidated.
-            addNativeReference => RootObject::gcProtect
-            removeNativeReference => RootObject::gcUnprotect
-            ProtectCountSet::contains => RootObject::gcIsProtected
-            
-        I know we'll all be sad to see the word "native" go.
-        
-        * bindings/runtime_root.h: Added ref-counting support to RootObject, with
-        all the standard accoutrements.
-
-        * bindings/c/c_utility.cpp:
-        (KJS::Bindings::convertValueToNPVariant): If we can't find a valid RootObject,
-        return void instead of just leaking.
-
-        * bindings/jni/jni_instance.cpp:
-        (JavaInstance::JavaInstance): Don't take a RootObject in our constructor;
-        be like other Instances and require the caller to call setRootObject. This
-        reduces the number of ownership code paths.
-        (JavaInstance::invokeMethod): Check RootObject for validity.
-        * bindings/jni/jni_instance.h: Removed private no-arg constructor. Having
-        an arg constructor accomplishes the same thing.
-
-        * bindings/jni/jni_jsobject.cpp:
-        (JavaJSObject::invoke): No need to call findProtectCountSet, because finalize()
-        checks for RootObject validity.
-        (JavaJSObject::JavaJSObject): check RootObject for validity
-        (JavaJSObject::call): ditto
-        (JavaJSObject::eval): ditto
-        (JavaJSObject::getMember): ditto
-        (JavaJSObject::setMember): ditto
-        (JavaJSObject::removeMember): ditto
-        (JavaJSObject::getSlot): ditto
-        (JavaJSObject::setSlot): ditto
-        (JavaJSObject::toString): ditto
-        (JavaJSObject::finalize): ditto
-        (JavaJSObject::createNative): No need to tell the RootObject to protect 
-        the global object, since the RootObject already owns the interpreter.
-
-        * bindings/jni/jni_runtime.cpp:
-        (JavaArray::JavaArray): Removed copy construcutor becaue it was unused.
-        Dead code is dangerous code.
-
-        * bindings/objc/objc_runtime.mm: Added WebUndefined protocol. Previous use
-        of WebScriptObject was bogus, because WebUndefined is not a subclass of
-        WebScriptObject.
-        (convertValueToObjcObject): If we can't find a valid RootObject,
-        return nil instead of just leaking.
-
-        * bindings/objc/objc_utility.mm:
-        (KJS::Bindings::convertValueToObjcValue): If we can't find a valid RootObject,
-        return nil instead of just leaking.
-
-2007-01-27  Andrew Wellington  <proton@wiretapped.net>
-
-        Reviewed by Maciej.
-        
-        Fix for Repeated string concatenation results in OOM crash
-        http://bugs.webkit.org/show_bug.cgi?id=11131
-
-        * kjs/operations.cpp:
-        (KJS::add): Throw exception if string addition result is null
-        * kjs/ustring.cpp:
-        (KJS::UString::UString): Don't call memcpy when malloc failed
-
-2007-01-25  Jan Kraemer  <camel@gmx.de>
-
-        Reviewed by Maciej
-
-        Fix for http://bugs.webkit.org/show_bug.cgi?id=12382
-
-        Fix crash on architectures with 32 bit ints and
-        64 bit longs (For example Linux on AMD64)
-
-        * kjs/dtoa.cpp: #define Long int as suggested in comment
-
-2007-01-24  Geoffrey Garen  <ggaren@apple.com>
-
-        Fixed up #include order for style. No review necessary.
-
-        * API/JSStringRef.cpp:
-
-2007-01-24  Geoffrey Garen  <ggaren@apple.com>
-
-        Reviewed by Maciej Stachowiak.
-
-        * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
-        Copy JSStringRefCF, in case anybody wants to use it. (I just added
-        it recently.)
-
-2007-01-24  Maciej Stachowiak  <mjs@apple.com>
-
-        Not reviewed, trivial property change.
-        
-        * JavaScriptCore.vcproj/JavaScriptCore.sln: remove svn:mime-type
-        property which made this binary.
-
-2007-01-25  Mark Rowe  <mrowe@apple.com>
-
-        Reviewed by Darin.
-
-        * Info.plist: Update copyright string.
-
-2007-01-24  Darin Adler  <darin@apple.com>
-
-        Reviewed by Mark Rowe.
-
-        * JavaScriptCore.xcodeproj/project.pbxproj: Changed to /usr/sbin/sysctl
-        so we don't rely on people's paths.
-
-2007-01-23  Alice Liu  <alice.liu@apple.com>
-
-        release build fix
-
-        * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
-        Copy APICasts.h
-
-2007-01-23  Geoffrey Garen  <ggaren@apple.com>
-
-        build fix
-
-        * API/JSStringRef.h:
-        * JavaScriptCore.xcodeproj/project.pbxproj:
-
-2007-01-24  Mark Rowe  <mrowe@apple.com>
-
-        Build fix for DumpRenderTree.
-
-        * JavaScriptCore.xcodeproj/project.pbxproj: Make JSStringRefCF.h public so it's copied into built framework.
-
-2007-01-23  Anders Carlsson  <acarlsson@apple.com>
-
-        Reviewed by Darin.
-
-        * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
-        Copy APICasts.h
-
-2007-01-23  Geoffrey Garen  <ggaren@apple.com>
-
-        Reviewed by Maciej Stachowiak.
-        
-        Fixed <rdar://problem/4885131> Move CFString function declarations from 
-        JSStringRef.h to JSStringRefCF.h
-        
-        Also removed remaining API FIXMEs and changed them into Radars.
-
-        * API/JSClassRef.cpp:
-        (OpaqueJSClass::OpaqueJSClass): Added Radar numbers for UTF8 conversion.
-
-        * API/JSContextRef.cpp:
-        (JSGlobalContextCreate): Replaced FIXME for NULL JSContextRef with Radar number.
-
-        * API/JSObjectRef.h: Removed FIXME, which is unprofessional in a public header.
-
-        * API/JSStringRef.cpp: Moved CF related implementations to JSStringRefCF.cpp.
-        (JSStringCreateWithUTF8CString): Replaced FIXME with Radar number.
-        * API/JSStringRef.h: Moved CF related declarations to JSStringRefCF.h. Added
-        #include of JSStringRefCF.h as a stopgap until clients start #including
-        it as needed by themselves.
-
-        * API/JSStringRefCF.cpp: Added.
-        (JSStringCreateWithCFString):
-        (JSStringCopyCFString): Replaced JSChar cast with UniChar cast, which is 
-        more appropriate for a CF call.
-        * API/JSStringRefCF.h: Added.
-        * JavaScriptCore.xcodeproj/project.pbxproj:
-
-2007-01-18  Sanjay Madhav  <sanjay12@gmail.com>
-
-        Reviewed by Darin.
-
-        Add JavaScriptCore define to help with tracing of when objects are marked.
-
-        * kjs/object.cpp:
-        (KJS::JSObject::mark):
-
-2007-01-18  Simon Hausmann  <hausmann@kde.org>
-
-        Reviewed by Zack.
-
-        * JavaScriptCore.pro: Remove generated files on make clean.
-        * pcre/pcre.pri:
-
-2007-01-16  Alexey Proskuryakov  <ap@webkit.org>
-
-        Reviewed by Maciej.
-
-        http://bugs.webkit.org/show_bug.cgi?id=12268
-        Give object prototypes their own names
-
-        * kjs/lookup.h: Append "Prototype" to ClassName in KJS_IMPLEMENT_PROTOTYPE.
-
-2007-01-16  Geoffrey Garen  <ggaren@apple.com>
-
-        Reviewed by Darin Adler.
-        
-        Added re-entrency checking to GC allocation and collection. It is an error
-        to allocate or collect from within a collection. We've had at least one 
-        case of each bug in the past.
-        
-        Added a comment to the API header, explaining that API clients must not
-        make this mistake, either.
-        
-        Layout tests and JS tests pass.
-
-        * API/JSObjectRef.h:
-        * kjs/collector.cpp:
-        (KJS::GCLock::GCLock):
-        (KJS::GCLock::~GCLock):
-        (KJS::Collector::allocate):
-        (KJS::Collector::collect):
-
-2007-01-14  Mark Rowe  <mrowe@apple.com>
-
-        Reviewed by Mitz.
-
-        Minor fixes to JavaScript pretty-printing.
-
-        * JavaScriptCore.exp:
-        * kjs/Parser.cpp:
-        (KJS::Parser::prettyPrint): Return line number and error message if parsing fails.
-        * kjs/Parser.h:
-        * kjs/nodes2string.cpp:
-        (ElementNode::streamTo): Include comma delimiters in array literals.
-        (PropertyNameNode::streamTo): Quote property names in object literals to handle the case when the property name is not a valid identifier.
-        * kjs/testkjs.cpp:
-        (doIt): Print any errors encountered while pretty-printing.
-
-2007-01-12  Anders Carlsson  <acarlsson@apple.com>
-
-        Reviewed by Darin.
-
-        * wtf/HashTraits.h:
-        Add hash traits for unsigned long and unsigned long long.
-
-2007-01-12  Geoffrey Garen  <ggaren@apple.com>
-
-        RS by Brady Eidson.
-        
-        Rolling back in r18786 with leaks fixed, and these renames slightly reworked:
-
-        Because they can return 0:
-        rootObjectForImp => findRootObject (overloaded for JSObject* and Interpreter*)
-        rootObjectForInterpreter => findRootObject (ditto)
-        findReferenceSet => findProtectCountSet
-
-2007-01-11  Geoffrey Garen  <ggaren@apple.com>
-
-        RS by Brady Eidson.
-        
-        Rolling out r18786 because it caused leaks.
-
-2007-01-11  Geoffrey Garen  <ggaren@apple.com>
-
-        Reviewed by Anders Carlsson.
-
-        Even more cleanup in preparation for fixing <rdar://problem/4608404> 
-        WebScriptObject's _executionContext lack of ownership policy causes 
-        crashes (e.g., in Dashcode)
-        
-        Layout tests pass.
-        
-        Renames:
-            ReferencesSet | ProtectCounts => ProtectCountSet (because it's a typename for a set of GC protect counts)
-            ReferencesByRootMap => RootObjectMap (because RootObjectToProtectCountSetMap would have been confusing)
-            pv => protectedValues
-            rootObjectForImp => getRootObject (overloaded for JSObject* and Interpreter*)
-            rootObjectForInterpreter => getRootObject (ditto)
-            findReferenceSet => getProtectCountSet
-            imp => jsObject
-        
-        (KJS::Bindings::getRootObjectMap): Changed to take advantage of built-in
-        facility for initializing static variables.
-
-        (KJS::Bindings::getProtectCountSet):
-        (KJS::Bindings::destroyProtectCountSet): Added. Helps encapsulate the fact
-        that getting a ProtectCountSet entails adding a RootObject to a hash table,
-        and destroying one entails the reverse.
-
-        (KJS::Bindings::getRootObject): Removed spurious NULL check.
-        
-        (KJS::Bindings::findReferenceSet): Renamed. Changed to use getRootObject()
-        instead of iterating on its own.
-
-        (KJS::Bindings::addNativeReference): Changed to use an early return instead
-        of indenting the whole function.
-        (KJS::Bindings::removeNativeReference): Ditto.
-
-2007-01-11  Geoffrey Garen  <ggaren@apple.com>
-
-        Reviewed by Anders Carlsson.
-
-        Even more cleanup in preparation for fixing <rdar://problem/4608404> 
-        WebScriptObject's _executionContext lack of ownership policy causes 
-        crashes (e.g., in Dashcode)
-        
-        Layout tests pass.
-        
-        Renames:
-            findRootObjectForNativeHandleFunction => createRootObject
-            FindRootObjectForNativeHandleFunctionPtr => CreateRootObjectFunction
-            
-        Also removed unnecessary use of "Bindings::" prefix.
-
-        * JavaScriptCore.exp:
-        * bindings/jni/jni_jsobject.cpp:
-        (JavaJSObject::createNative):
-        (JavaJSObject::convertValueToJObject):
-        (JavaJSObject::convertJObjectToValue):
-        * bindings/runtime_root.cpp:
-        (KJS::Bindings::RootObject::setCreateRootObject):
-        * bindings/runtime_root.h:
-        (KJS::Bindings::RootObject::createRootObject):
-
-2007-01-11  George Staikos  <staikos@kde.org>
-
-        Reviewed by Maciej
-
-        Appears to be Mac specific right now.
-
-        * kjs/config.h:
-
-2007-01-10  Lars Knoll <lars@trolltech.com>
-
-        Reviewed by Zack
-
-        Use the new functionality in Qt 4.3, to make
-        the methods closer compliant with the Unicode
-        spec.
-
-        Keep the old code so that it still compiles against
-        Qt 4.2.
-
-        * wtf/unicode/qt4/UnicodeQt4.h:
-        (WTF::Unicode::toLower):
-        (WTF::Unicode::toUpper):
-        (WTF::Unicode::toTitleCase):
-        (WTF::Unicode::foldCase):
-        (WTF::Unicode::isFormatChar):
-        (WTF::Unicode::isPrintableChar):
-        (WTF::Unicode::isSeparatorSpace):
-        (WTF::Unicode::isPunct):
-        (WTF::Unicode::isDigit):
-        (WTF::Unicode::isLower):
-        (WTF::Unicode::isUpper):
-        (WTF::Unicode::digitValue):
-        (WTF::Unicode::mirroredChar):
-        (WTF::Unicode::combiningClass):
-        (WTF::Unicode::decompositionType):
-        (WTF::Unicode::umemcasecmp):
-        (WTF::Unicode::direction):
-        (WTF::Unicode::category):
-
-2007-01-09  Darin Adler  <darin@apple.com>
-
-        - update 2007 Apple copyright for the new company name
-
-        * kjs/DateMath.cpp:
-
-2007-01-09  Darin Adler  <darin@apple.com>
-
-        - fix build
-
-        * kjs/string_object.cpp: (KJS::StringProtoFunc::callAsFunction):
-        Actually compile it this time.
-
-2007-01-09  Darin Adler  <darin@apple.com>
-
-        - fix build
-
-        * kjs/string_object.cpp: (KJS::StringProtoFunc::callAsFunction):
-        Change types.
-
-2007-01-09  Darin Adler  <darin@apple.com>
-
-        - fix build on platforms where Unicode::UChar is != uint16_t
-
-        * kjs/string_object.cpp: (KJS::StringProtoFunc::callAsFunction):
-        Change types.
-
-2007-01-09  Mitz Pettel  <mitz@webkit.org>
-
-        Reviewed by Darin.
-
-        - changes for http://bugs.webkit.org/show_bug.cgi?id=11078
-          Forms Don't Submit (ASP Pages)
-
-        * JavaScriptCore.exp:
-        * kjs/value.cpp:
-        (KJS::JSValue::toInt32): Folded toInt32Inline into this method, which was its
-        only caller.
-        (KJS::JSValue::toUInt32): Added a variant that reports if the conversion has
-        succeeded.
-        * kjs/value.h:
-
-2007-01-09  Darin Adler  <darin@apple.com>
-
-        Reviewed by Maciej.
-
-        - fix http://bugs.webkit.org/show_bug.cgi?id=12174
-          improve Unicode use (less WTF::Unicode:: prefix, centralized character names)
-
-        * wtf/unicode/icu/UnicodeIcu.h: Change parameter and return types
-        to UChar32 and UChar. Removed unneeded type casts and added some
-        const to functions that lacked it. Removed WTF::Unicode::memcmp.
-        (WTF::Unicode::umemcasecmp): Renamed from strcasecmp since this
-        doesn't work on 0-terminated strings as the str functions do.
-        * wtf/unicode/qt4/UnicodeQt4.h: Ditto.
-
-        - got rid of namespace prefixes from most uses of WTF::Unicode
-
-        * kjs/function.cpp:
-        (KJS::isStrWhiteSpace):
-        (KJS::escapeStringForPrettyPrinting):
-        * kjs/lexer.cpp:
-        (KJS::Lexer::isWhiteSpace):
-        (KJS::Lexer::isIdentStart):
-        (KJS::Lexer::isIdentPart):
-        * kjs/string_object.cpp:
-        (KJS::StringProtoFunc::callAsFunction):
-
-2007-01-07  David Kilzer  <ddkilzer@webkit.org>
-
-        Reviewed by Darin.
-
-        - fix http://bugs.webkit.org/show_bug.cgi?id=11917
-          setlocale() can return null
-
-        * kjs/date_object.cpp:
-        (KJS::DateProtoFunc::callAsFunction): Removed dead code.
-
-2007-01-07  David Carson  <dacarson@gmail.com>
-
-        Reviewed by Darin.
-
-        - fix http://bugs.webkit.org/show_bug.cgi?id=12100
-          JNI bindings should be available to non-Mac platforms that have JNI
-
-        Change JNI so that it is not wrapped in the PLATFORM(MAC) ifdef, enabling
-        other platforms who have JNI to use it.
-
-        * bindings/jni/jni_instance.h:  
-          Removed unnecessary include of <CoreFoundation/CoreFoundation.h>
-        * bindings/jni/jni_utility.cpp:
-        (KJS::Bindings::setJavaVM):
-        * bindings/jni/jni_utility.h:
-          Added new method for clients to set the JavaVM
-        * bindings/runtime.cpp:
-        (KJS::Bindings::Instance::createBindingForLanguageInstance):
-          Changed code to utilize new #if HAVE(JNI)
-        * kjs/config.h:
-          Added new #define for JNI, ie HAVE_JNI
-
-2007-01-07  David Carson  <dacarson@gmail.com>
-
-        Reviewed by Darin.
-
-        Fix http://bugs.webkit.org/show_bug.cgi?id=11431
-        ARM platform has some byte alignment issues
-
-        Fix for NaN being 4 bytes and it must start on a byte boundary
-        for ARM architectures.
-
-        * kjs/fpconst.cpp:
-        (KJS::):
-
-2007-01-04  David Kilzer  <ddkilzer@webkit.org>
-
-        Reviewed by Kevin McCullough.
-
-        - fix http://bugs.webkit.org/show_bug.cgi?id=12070
-          REGRESSION: KJS::getUTCOffset() caches UTC offset but ignores time zone changes
-
-        * kjs/DateMath.cpp:
-        (KJS::getUTCOffset): Don't cache UTC offset.
-
-2007-01-02  Darin Adler  <darin@apple.com>
-
-        - minor tweak (hope this doesn't re-break Windows)
-
-        * pcre/pcre_compile.c: Removed use of const pcre_uchar const * -- Mitz probably
-        meant const pcre_uchar *const, but I think we can do without the explicit const here.
-
-        * pcre/pcre_internal.h: Re-enabled warning C4114.
-
-2007-01-02  David Kilzer  <ddkilzer@webkit.org>
-
-        Reviewed by NOBODY (Windows build fix).
-
-        The MSVC compiler requires variables to be declared at the top of the enclosing block in C source.
-
-        Disable this warning to prevent MSVC from complaining about the 'const pcre_uchar const *' type:
-        warning C4114: same type qualifier used more than once
-
-        * pcre/pcre_compile.c:
-        (pcre_compile2): Moved variable declarations to top of their respective enclosing blocks.
-        * pcre/pcre_internal.h: Added pragma to disable compiler warning.
-
-2007-01-01  Mitz Pettel  <mitz@webkit.org>
-
-        Reviewed by Darin.
-
-        - fix http://bugs.webkit.org/show_bug.cgi?id=11849
-          REGRESSION (r18182): Google Calendar is broken (a regular expression containing a null character is not parsed correctly)
-
-        Modified pcre_compile() (and the functions that it calls) to work with patterns
-        containing null characters.
-
-        Covered by JavaScriptCore tests ecma_3/RegExp/octal-002.js and ecma_3/RegExp/regress-85721.js
-
-        * kjs/regexp.cpp:
-        (KJS::RegExp::RegExp): Changed to not null-terminate the pattern string and instead
-        pass its length to pcre_compile.
-        * pcre/pcre.h:
-        * pcre/pcre_compile.c:
-        (check_escape):
-        (get_ucp):
-        (is_counted_repeat):
-        (check_posix_syntax):
-        (compile_branch):
-        (compile_regex):
-        (pcre_compile): Added a parameter specifying the length of the pattern, which
-        is no longer required to be null-terminated and may contain null characters. 
-        (pcre_compile2):
-        * pcre/pcre_internal.h:
-        * tests/mozilla/expected.html: Updated for the two tests that this patch
-        fixes. Also updated failing results for ecma_3/RegExp/regress-100199.js
-        which were not updated after bug 6257 was fixed.
-
-2007-01-01  David Kilzer  <ddkilzer@webkit.org>
-
-        Reviewed by Darin.
-
-        - fix http://bugs.webkit.org/show_bug.cgi?id=12057
-          REGRESSION: JavaScript Date Is One Day In The Future in GMT time zone
-
-        Because Mac OS X returns geographically and historically accurate time zone information,
-        converting Jan 02, 1970 12:00:00 AM to local time then subtracting 24 hours did not work
-        in GMT (London - England) since it was in BST (+0100) all year in 1970[1].  Instead, the
-        UTC offset is calculated by converting Jan 01, 2000 12:00:00 AM to local time then
-        subtracting that from the same date in UTC.
-
-        [1] http://en.wikipedia.org/wiki/British_Summer_Time
-
-        * kjs/DateMath.cpp:
-        (KJS::getUTCOffset): Updated UTC offset calculation.
-        (KJS::getDSTOffset): Improved comment.
-
-2006-12-31  David Kilzer  <ddkilzer@webkit.org>
-
-        Reviewed by Geoff.
-
-        Update embedded pcre library from version 6.2 to 6.4.  Changes from pcre 6.2 to 6.3
-        did not include any files in JavaScriptCore/pcre.
-
-        All changes include renaming EXPORT to PCRE_EXPORT, renaming of ucp_findchar() to
-        _pcre_ucp_findchar(), or comment changes.  Additional changes noted below.
-
-        * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: Updated source file list.
-        * JavaScriptCore.xcodeproj/project.pbxproj: Renamed pcre_printint.c to pcre_printint.src
-        and changed it from a source file to a header file.
-        * JavaScriptCoreSources.bkl: Updated source file list.
-        * pcre/CMakeLists.txt: Updated source file list.
-        * pcre/pcre-config.h:
-        * pcre/pcre.h: Updated version.
-        * pcre/pcre.pri: Updated source file list.
-        * pcre/pcre_compile.c: Include pcre_printint.src #if DEBUG.
-        (pcre_compile2):
-        * pcre/pcre_config.c:
-        * pcre/pcre_exec.c:
-        (match):
-        * pcre/pcre_fullinfo.c:
-        * pcre/pcre_info.c:
-        * pcre/pcre_internal.h: Added header guard.  Removed export of _pcre_printint().
-        * pcre/pcre_ord2utf8.c:
-        * pcre/pcre_printint.c: Renamed to pcre_printint.src.
-        * pcre/pcre_printint.src: Added.  Renamed _pcre_printint() to pcre_printint().
-        * pcre/pcre_refcount.c:
-        * pcre/pcre_study.c:
-        * pcre/pcre_tables.c:
-        * pcre/pcre_try_flipped.c:
-        * pcre/pcre_ucp_findchar.c: Added contents of ucp_findchar.c.
-        * pcre/pcre_version.c:
-        * pcre/pcre_xclass.c:
-        (_pcre_xclass):
-        * pcre/ucp.h: Removed export of ucp_findchar().
-        * pcre/ucp_findchar.c: Removed.  Contents moved to pcre_ucp_findchar.c.
-
-2006-12-29  David Kilzer  <ddkilzer@webkit.org>
-
-        Reviewed by Geoff.
-
-        Update embedded pcre library from version 6.1 to 6.2.  From the pcre ChangeLog:
-
-        3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
-           operating environments where this matters.
-
-        5. Named capturing subpatterns were not being correctly counted when a pattern
-           was compiled. This caused two problems: (a) If there were more than 100
-           such subpatterns, the calculation of the memory needed for the whole
-           compiled pattern went wrong, leading to an overflow error. (b) Numerical
-           back references of the form \12, where the number was greater than 9, were
-           not recognized as back references, even though there were sufficient
-           previous subpatterns.
-
-        * pcre/dftables.c: Item 3.
-        (main):
-        * pcre/pcre.h: Updated version.
-        * pcre/pcre_compile.c: Item 5.
-        (read_repeat_counts):
-        (pcre_compile2):
-
-2006-12-29  Geoffrey Garen  <ggaren@apple.com>
-
-        Reviewed by Brian Dash... err... Mark Rowe.
-
-        More cleanup in preparation for fixing <rdar://problem/4608404> 
-        WebScriptObject's _executionContext lack of ownership policy causes 
-        crashes (e.g., in Dashcode)
-        
-        The key change here is to RootObject::RootObject().
-        
-        * JavaScriptCore.exp:
-
-        * bindings/c/c_utility.cpp:
-        (KJS::Bindings::convertValueToNPVariant): Changed to use new constructor.
-
-        * bindings/jni/jni_jsobject.cpp:
-        (JavaJSObject::createNative): Changed to use new constructor. Replaced
-        large 'if' followed by default condition with "if !" and explicit default
-        condition.
-
-        * bindings/objc/objc_runtime.mm:
-        (convertValueToObjcObject): Changed to use new constructor.
-
-        * bindings/runtime_root.cpp:
-        (KJS::Bindings::RootObject::destroy): "removeAllNativeReferences" => "destroy"
-        because this function actually destroys the RootObject.
-
-        * bindings/runtime_root.h: Changed Interpreter* to RefPtr<Interpreter>
-        to prevent a RootObject from holding a stale Interperter*.
-        
-        (KJS::Bindings::RootObject::RootObject): Changed constructor to take an 
-        Interpreter*, since it's pointless to create a RootObject without one.
-        Removed setRootObjectImp() and rootObjectImp() because they were just
-        a confusing way of setting and getting the Interpreter's global object.
-
-        (KJS::Bindings::RootObject::nativeHandle): "_nativeHandle" => "m_nativeHandle"
-        (KJS::Bindings::RootObject::interpreter): "_interpreter" => "m_interpreter"
-
-2006-12-28  George Staikos  <staikos@kde.org>
-
-        Reviewed by Olliej.
-
-        * bindings/qt/qt_instance.cpp: build
-        (KJS::Bindings::QtInstance::QtInstance):
-
-2006-12-28  Geoffrey Garen  <ggaren@apple.com>
-
-        Reviewed by Oliver Hunt.
-        
-        More cleanup. Layout tests pass.
-        
-        Use a helper function to initialize and access WebUndefined and WebScriptObject.
-
-        * bindings/objc/objc_runtime.h:
-        * bindings/objc/objc_runtime.mm:
-        (KJS::Bindings::webScriptObjectClass):
-        (KJS::Bindings::webUndefinedClass):
-        (convertValueToObjcObject):
-        * bindings/objc/objc_utility.mm:
-        (KJS::Bindings::convertValueToObjcValue):
-        (KJS::Bindings::convertObjcValueToValue):
-
-2006-12-28  Geoffrey Garen  <ggaren@apple.com>
-
-        Reviewed by Brady Eidson.
-        
-        Some cleanup in preparation for fixing <rdar://problem/4608404> 
-        WebScriptObject's _executionContext lack of ownership policy causes 
-        crashes (e.g., in Dashcode)
-        
-        I'm just trying to make heads or tails of this baffling code.
-        
-        Renamed "root" | "execContext" | "executionContext" => "rootObject", because
-        that's the object's (admittedly vague) type name.
-        
-        * bindings/runtime.cpp: Removed createLanguageInstanceForValue
-        because I'll give you a dollar if you can explain to me what it actually did.
-        
-        * bindings/runtime_root.cpp: Put everything in the KJS::Bindings namespace,
-        removing the KJS::Bindings prefix from individual functions and datatypes.
-        This matches the header and eliminates a lot of syntax cruft.
-        
-        * bindings/c/c_utility.cpp:
-        (KJS::Bindings::convertValueToNPVariant): Replaced use of createLanguageInstanceForValue
-        with call to _NPN_CreateScriptObject because that's what createLanguageInstanceForValue
-        actually did (but don't ask me for that dollar now; that's cheating.)
-
-        * bindings/objc/objc_utility.h:
-        * bindings/objc/objc_utility.mm:
-        (KJS::Bindings::convertValueToObjcValue): Removed. Its only purpose was 
-        to call a single function for WebKit, which WebKit can do on its own.
-
-        * kjs/interpreter.h: Removed rtti() because it was unused, and this class
-        is scheduled for demolition anyway.
-        
-        * kjs/interpreter.cpp: Removed createLanguageInstanceForValue because it had
-        nothing to do with the Interpreter, and nothing makes Chuck Norris more mad
-        than a function whose sole purpose is to call another function of the same
-        name. (Really, I asked him.)
-
-2006-12-26  Geoffrey Garen  <ggaren@apple.com>
-
-        Reviewed by Eric Seidel.
-
-        Some cleanup in preparation for fixing <rdar://problem/4740328> Safari 
-        crash on quit in _NPN_ReleaseObject from KJS::Bindings::CInstance::~CInstance
-
-        * bindings/c/c_instance.cpp:
-        * bindings/c/c_instance.h: Removed unused copy constructor and assignment
-        operator. They made tracking data flow more difficult. Unused code is also 
-        dangerous because it can succumb to bit rot with the stealth of a Ninja.
-        
-        Replaced #include with forward declaration to reduce header dependency.
-        
-        * bindings/npruntime.cpp: Sorted #includes.
-        (_NPN_GetStringIdentifier): Replaced assert with ASSERT.
-        (_NPN_GetStringIdentifiers): ditto
-        (_NPN_ReleaseVariantValue): ditto
-        (_NPN_CreateObject): ditto
-        (_NPN_RetainObject): ditto
-        (_NPN_ReleaseObject): ditto
-        (_NPN_DeallocateObject): ditto
-
-2006-12-20  Anders Carlsson  <acarlsson@apple.com>
-
-        * kjs/string_object.cpp:
-        (localeCompare):
-        Another speculative Win32 fix.
-
-2006-12-20  Anders Carlsson  <acarlsson@apple.com>
-
-        * kjs/string_object.cpp:
-        (localeCompare):
-        Speculative Win32 fix.
-
-2006-12-20  Anders Carlsson  <acarlsson@apple.com>
-
-        Reviewed by Darin.
-
-        <rdar://problem/4235733>
-        <http://bugs.webkit.org/?show_bug.cgi?id=10193>
-        support String.localeCompare.
-        
-        Implement localeCompare.
-        
-        * JavaScriptCore.xcodeproj/project.pbxproj:
-        * kjs/string_object.cpp:
-        (localeCompare):
-        (StringProtoFunc::callAsFunction):
-        * kjs/string_object.h:
-        (KJS::StringProtoFunc::):
-
-2006-12-20  Timothy Hatcher  <timothy@apple.com>
-
-        Reviewed by Mark Rowe.
-
-        * JavaScriptCore.xcodeproj/project.pbxproj: use GCC 4.0 for all the other test targets
-
-2006-12-20  Timothy Hatcher  <timothy@apple.com>
-
-        Reviewed by Mark Rowe.
-
-        <rdar://problem/4871613> JavaScriptCore-421.31's dftables target needs to override default compiler and use gcc-4.0
-
-        * JavaScriptCore.xcodeproj/project.pbxproj:
-
-2006-12-20  Lars Knoll <lars@trolltech.com>
-
-        Reviewed by David Hyatt
-
-        Added support to bind QObject's to 
-        JavaScript.
-
-        * JavaScriptCore.pro:
-        * bindings/qt/qt_class.cpp: Added.
-        (KJS::Bindings::QtClass::QtClass):
-        (KJS::Bindings::QtClass::~QtClass):
-        (KJS::Bindings::QtClass::classForObject):
-        (KJS::Bindings::QtClass::name):
-        (KJS::Bindings::QtClass::methodsNamed):
-        (KJS::Bindings::QtClass::fieldNamed):
-        * bindings/qt/qt_class.h: Added.
-        (KJS::Bindings::QtClass::constructorAt):
-        (KJS::Bindings::QtClass::numConstructors):
-        * bindings/qt/qt_instance.cpp: Added.
-        (KJS::Bindings::QtInstance::QtInstance):
-        (KJS::Bindings::QtInstance::~QtInstance):
-        (KJS::Bindings::QtInstance::operator=):
-        (KJS::Bindings::QtInstance::getClass):
-        (KJS::Bindings::QtInstance::begin):
-        (KJS::Bindings::QtInstance::end):
-        (KJS::Bindings::QtInstance::implementsCall):
-        (KJS::Bindings::QtInstance::invokeMethod):
-        (KJS::Bindings::QtInstance::invokeDefaultMethod):
-        (KJS::Bindings::QtInstance::defaultValue):
-        (KJS::Bindings::QtInstance::stringValue):
-        (KJS::Bindings::QtInstance::numberValue):
-        (KJS::Bindings::QtInstance::booleanValue):
-        (KJS::Bindings::QtInstance::valueOf):
-        * bindings/qt/qt_instance.h: Added.
-        (KJS::Bindings::QtInstance::getObject):
-        * bindings/qt/qt_runtime.cpp: Added.
-        (KJS::Bindings::convertValueToQVariant):
-        (KJS::Bindings::convertQVariantToValue):
-        (KJS::Bindings::QtField::name):
-        (KJS::Bindings::QtField::valueFromInstance):
-        (KJS::Bindings::QtField::setValueToInstance):
-        * bindings/qt/qt_runtime.h: Added.
-        (KJS::Bindings::QtField::QtField):
-        (KJS::Bindings::QtField::type):
-        (KJS::Bindings::QtMethod::QtMethod):
-        (KJS::Bindings::QtMethod::name):
-        (KJS::Bindings::QtMethod::numParameters):
-        * bindings/runtime.cpp:
-        (KJS::Bindings::Instance::createBindingForLanguageInstance):
-        * bindings/runtime.h:
-        (KJS::Bindings::Instance::):
-        * bindings/testbindings.pro: Added.
-        * bindings/testqtbindings.cpp: Added.
-        (MyObject::MyObject):
-        (MyObject::setTestString):
-        (MyObject::setTestInt):
-        (MyObject::testString):
-        (MyObject::testInt):
-        (MyObject::foo):
-        (Global::className):
-        (main):
-
-2006-12-19  Anders Carlsson  <acarlsson@apple.com>
-
-        Reviewed by Geoff.
-
-        Add -p option to testkjs which pretty prints the files instead of executing them.
-
-        * JavaScriptCore.exp:
-        * JavaScriptCore.xcodeproj/project.pbxproj:
-        * kjs/Parser.cpp:
-        (KJS::Parser::prettyPrint):
-        * kjs/Parser.h:
-        * kjs/testkjs.cpp:
-        (doIt):
-
-2006-12-19  Brady Eidson  <beidson@apple.com>
-
-        Rubberstamped by Lou
-        
-        Removed unneccessary "else"
-
-        * wtf/Assertions.cpp:
-
-2006-12-19  Timothy Hatcher  <timothy@apple.com>
-
-        Reviewed by Darin.
-
-        <rdar://problem/4891774> Local WebCore/WebBrowser builds fail in 9A328 due to warning about ObjC-2.0 language features
-
-        * JavaScriptCore.xcodeproj/project.pbxproj:
-
-2006-12-17  Simon Hausmann  <hausmann@kde.org>
-
-        Reviewed by Zack.
-
-        * kjs/testkjs.pro: Oops, make it also build on machines other than
-        mine :)
-
-2006-12-17  Simon Hausmann  <hausmann@kde.org>
-
-        Reviewed by Rob Buis.
-
-        * kjs/testkjs.pro: Added .pro file to build testkjs.
-
-2006-12-16  Alexey Proskuryakov  <ap@webkit.org>
-
-        Reviewed by Rob.
-
-        A deleted object was accessed to prepare RegExp construction error messages.
-
-        * kjs/regexp_object.cpp:
-        (RegExpObjectImp::construct): Wrap the RegExp into an OwnPtr.
-
-2006-12-16  Mitz Pettel  <mitz@webkit.org>
-
-        Reviewed by Alexey.
-
-        - fix http://bugs.webkit.org/show_bug.cgi?id=11814
-          REGRESSION(r18098): Find does not work with capital letters
-
-        Test: editing/execCommand/findString-3.html
-
-        * wtf/unicode/icu/UnicodeIcu.h:
-        (WTF::Unicode::foldCase): Changed to not return an error if the result fits
-        in the buffer without a null terminator.
-
-2006-12-13  Maciej Stachowiak  <mjs@apple.com>
-
-        Reviewed by Anders.
-        
-        - added equality and inequality operations for HashMap and Vector, useful for comparing more complex types
-
-        * wtf/HashMap.h:
-        (WTF::operator==):
-        (WTF::operator!=):
-        * wtf/Vector.h:
-        (WTF::operator==):
-        (WTF::operator!=):
-
-2006-12-12  Alexey Proskuryakov  <ap@webkit.org>
-
-        Reviewed by Geoff. Based on a patch by Maks Orlovich.
-
-        http://bugs.webkit.org/show_bug.cgi?id=6257
-        Throw errors on invalid expressions (KJS merge)
-
-        * kjs/regexp.cpp:
-        (KJS::RegExp::RegExp):
-        (KJS::RegExp::~RegExp):
-        (KJS::RegExp::match):
-        * kjs/regexp.h:
-        (KJS::RegExp::flags):
-        (KJS::RegExp::isValid):
-        (KJS::RegExp::errorMessage):
-        (KJS::RegExp::subPatterns):
-        Remember and report RegExp construction failures. Renamed data members not to start with underscores.
-
-        * kjs/regexp_object.cpp:
-        (RegExpObjectImp::construct): Raise an exception if RegExp construction fails.
-        (RegExpObjectImp::callAsFunction): Removed an obsolete comment.
-
-        * tests/mozilla/ecma_3/RegExp/regress-119909.js: Reduced the number of nested parentheses to 
-        a value supported by PCRE.
-
-2006-12-11  Alexey Proskuryakov  <ap@webkit.org>
-
-        Reviewed by Darin.
-
-        http://bugs.webkit.org/show_bug.cgi?id=9673
-        Add support for window.atob() and window.btoa()
-
-        * JavaScriptCore.exp: Export UString::is8Bit().
-        * JavaScriptCore.xcodeproj/project.pbxproj: Added StringExtras.h as 
-        a private header.
-
-2006-12-11  Darin Adler  <darin@apple.com>
-
-        Reviewed by Brady.
-
-        * JavaScriptCore.xcodeproj/project.pbxproj: Let Xcode update this
-        (I think Hyatt is using an old Xcode).
-
-2006-12-11  David Hyatt  <hyatt@apple.com>
-
-        Fix the failing layout test.  Just remove Unicode::isSpace and
-        revert StringImpl to do the same thing it was doing before.
-
-        Reviewed by darin
-
-        * wtf/unicode/icu/UnicodeIcu.h:
-        * wtf/unicode/qt4/UnicodeQt4.h:
-
-2006-12-09  George Staikos  <staikos@kde.org>
-
-        Reviewed by Zack.
-
-        Fix bison again on qmake build.
-
-        * JavaScriptCore.pro:
-
-2006-12-09  Lars Knoll <lars@trolltech.com>
-
-        Reviewed by Zack
-
-        Make it possible to build WebKit with qmake.
-
-        * JavaScriptCore.pro: Added.
-        * kjs/kjs.pro: Removed.
-        * pcre/pcre.pri: Added.
-
-2006-12-09  Zack Rusin  <zack@kde.org>
-
-        Fixing the compilation with platform kde after the icu changes.
-
-        * CMakeLists.txt:
-
-2006-12-09  Adam Roben  <aroben@apple.com>
-
-        Reviewed by Darin.
-
-        Some updates in reaction to r18098.
-
-        * wtf/unicode/icu/UnicodeIcu.h: Use !! to convert UBool to bool in all
-        cases.
-        (WTF::Unicode::toLower):
-        (WTF::Unicode::toUpper):
-        (WTF::Unicode::isDigit):
-        (WTF::Unicode::isSpace):
-        (WTF::Unicode::isPunct):
-        (WTF::Unicode::isLower):
-        (WTF::Unicode::isUpper):
-        * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
-        * JavaScriptCore.vcproj/WTF/WTF.vcproj:
-
-2006-12-09  George Staikos  <staikos@kde.org>
-
-        Patch by Lars Knoll, comment out ICU dependency on Qt platform (unused code).
-
-        Reviewed by Darin.
-
-        * bindings/c/c_utility.cpp:
-        (KJS::Bindings::convertUTF8ToUTF16):
-
-2006-12-08  David Hyatt  <hyatt@apple.com>
-
-        Land the new ICU abstraction layer.  Patch by Lars.
-
-        Reviewed by me
-
-        * JavaScriptCore.xcodeproj/project.pbxproj:
-        * wtf/Platform.h:
-        * wtf/unicode/UnicodeCategory.h: Removed.
-        * wtf/unicode/UnicodeDecomposition.h: Removed.
-        * wtf/unicode/UnicodeDirection.h: Removed.
-        * wtf/unicode/icu/UnicodeIcu.h:
-        (WTF::Unicode::):
-        (WTF::Unicode::foldCase):
-        (WTF::Unicode::toLower):
-        (WTF::Unicode::toUpper):
-        (WTF::Unicode::toTitleCase):
-        (WTF::Unicode::isDigit):
-        (WTF::Unicode::isSpace):
-        (WTF::Unicode::isPunct):
-        (WTF::Unicode::mirroredChar):
-        (WTF::Unicode::category):
-        (WTF::Unicode::direction):
-        (WTF::Unicode::isLower):
-        (WTF::Unicode::isUpper):
-        (WTF::Unicode::digitValue):
-        (WTF::Unicode::combiningClass):
-        (WTF::Unicode::decompositionType):
-        (WTF::Unicode::strcasecmp):
-        (WTF::Unicode::memset):
-        * wtf/unicode/qt4/UnicodeQt4.cpp: Removed.
-        * wtf/unicode/qt4/UnicodeQt4.h:
-        (WTF::Unicode::):
-        (WTF::Unicode::toLower):
-        (WTF::Unicode::toUpper):
-        (WTF::Unicode::toTitleCase):
-        (WTF::Unicode::foldCase):
-        (WTF::Unicode::isPrintableChar):
-        (WTF::Unicode::isLower):
-        (WTF::Unicode::isUpper):
-        (WTF::Unicode::digitValue):
-        (WTF::Unicode::combiningClass):
-        (WTF::Unicode::decompositionType):
-        (WTF::Unicode::strcasecmp):
-        (WTF::Unicode::memset):
-        (WTF::Unicode::direction):
-        (WTF::Unicode::category):
-
-=== Safari-521.32 ===
-
-2006-12-08  Adam Roben  <aroben@apple.com>
-
-        Reviewed by Anders.
-
-        This is a mo' better fix for ensuring we don't use macro definitions
-        of min/max.
-
-        * kjs/config.h:
-        * wtf/Vector.h:
-
-2006-12-07  Kevin Fyure  <digdog@macports.org>
-
-        Reviewed by Darin.
-
-        http://bugs.webkit.org/show_bug.cgi?id=11545
-        Disable the testcases do not follow the ECMA-262v3 specification.
-        
-        * tests/mozilla/expected.html: Update Results.
-        * tests/mozilla/js1_2/String/concat.js:
-        4 tests disabled. The result of concat Array object is not followinig
-        ECMA 15.5.4.6
-        * tests/mozilla/js1_2/function/Number.js:
-        1 test disabled. The result of Array object to Number object conversion 
-        is not following ECMA 9.3. And the test was duplicated in 
-        ecma/TypeConversion/9.3-1.js
-        * tests/mozilla/js1_2/function/String.js:
-        2 tests disabled. The result of Object/Array object to String object 
-        conversion is not following ECMA 15.5.1.1 and ECMA 9.8
-
-2006-11-30  Steve Falkenburg  <sfalken@apple.com>
-
-        Reviewed by Oliver.
-        
-        Move WTF from JavaScriptCore project into a new WTF project.
-
-        * JavaScriptCore.vcproj/JavaScriptCore.sln: Add WTF.vcproj to sln
-        * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: Remove WTF source files
-        * JavaScriptCore.vcproj/WTF/WTF.vcproj: Added.
-        * JavaScriptCore.vcproj/testkjs/testkjs.vcproj: Add dependency on WTF.lib
-
-2006-11-30  Geoffrey Garen  <ggaren@apple.com>
-
-        Reviewed by Beth Dakin.
-        
-        Fixed up garbage collection at window close time.
-
-        * kjs/interpreter.cpp:
-        (KJS::Interpreter::~Interpreter): Garbage collect here, since
-        destroying the interpreter frees the global object and 
-        therefore creates a lot of garbage.
-
-2006-11-20  W. Andy Carrel  <wac@google.com>
-
-        Reviewed by Maciej.
-
-        http://bugs.webkit.org/show_bug.cgi?id=11501
-        REGRESSION: \u no longer escapes metacharacters in RegExps
-        http://bugs.webkit.org/show_bug.cgi?id=11502
-        Serializing RegExps doesn't preserve Unicode escapes
-
-        * kjs/lexer.cpp:
-        (Lexer::Lexer):
-        (Lexer::setCode):
-        (Lexer::shift):
-        (Lexer::scanRegExp):
-        Push \u parsing back down into the RegExp object rather than in the
-        parser. This backs out r17354 in favor of a new fix that better 
-        matches the behavior of other browsers.
-
-        * kjs/lexer.h:
-        * kjs/regexp.cpp:
-        (KJS::RegExp::RegExp):
-        (KJS::sanitizePattern):
-        (KJS::isHexDigit):
-        (KJS::convertHex):
-        (KJS::convertUnicode):
-        * kjs/regexp.h:
-        Translate \u escaped unicode characters for the benefit of pcre.
-
-        * kjs/ustring.cpp:
-        (KJS::UString::append):
-        Fix failure to increment length on the first UChar appended to a 
-        UString that was copy-on-write.
-
-        * tests/mozilla/ecma_2/RegExp/properties-001.js:
-        Adjust tests back to the uniform standards.
-
-2006-11-20  Samuel Weinig  <sam@webkit.org>
-
-        Reviewed by Maciej.
-
-        Fix for http://bugs.webkit.org/show_bug.cgi?id=11647
-        Fix Win32 build
-
-        * kjs/config.h: define NOMINMAX instead of min/max
-        as themselves.
-        * wtf/Vector.h: put back hack to ensure that min/max
-        are not defined as macros.
-
-2006-11-19  Simon Hausmann  <hausmann@kde.org>
-
-        Reviewed by Zack.
-
-        http://bugs.webkit.org/show_bug.cgi?id=11649
-        Fix CMake Qt-only build without KDE CMake files
-
-        * CMakeLists.txt:
-        * pcre/CMakeLists.txt:
-
-2006-11-17  Anders Carlsson  <acarlsson@apple.com>
-
-        Reviewed by Adam.
-
-        Make sure that we always use std::min and std::max instead of macros.
-        
-        * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
-        * kjs/config.h:
-        * wtf/Vector.h:
-
-=== Safari-521.31 ===
-
-2006-11-12  Geoffrey Garen  <ggaren@apple.com>
-
-        Reviewed by Beth Dakin.
-        
-        Added project-wide setting to disable Microsoft's made-up deprecation 
-        warnings related to std:: functions. (Doesn't have any affect yet,
-        since we currently disable all deprecation warnings.)
-
-        * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
-
-2006-11-12  Mark Rowe  <bdash@webkit.org>
-
-        Reviewed by Mitz.
-
-        Clean up of JavaScriptCore bakefiles.
-
-        * JavaScriptCoreSources.bkl:
-        * jscore.bkl:
-
-2006-11-11  Alexey Proskuryakov  <ap@webkit.org>
-
-        Reviewed by Maciej.
-
-        http://bugs.webkit.org/show_bug.cgi?id=11508
-        Undisable some warnings for JSImmediate.h
-
-        Fix suggested by Don Gibson.
-
-        * kjs/JSImmediate.h:
-        Re-enable all MSVC warnings, move the remaining runtime checks
-        to compile-time.
-
-2006-11-10  Zalan Bujtas  <zalan.bujtas@nokia.com>
-
-        Reviewed by Maciej.
-
-        Added s60/symbian platform defines.
-        http://bugs.webkit.org/show_bug.cgi?id=11540
-
-        * wtf/Platform.h:
-
-=== Safari-521.30 ===
-
-2006-11-08  Ada Chan  <adachan@apple.com>
-
-        Reviewed by darin.
-        
-        Added a method to delete all the keys in a HashMap.
-
-        * wtf/HashMap.h:
-        (WTF::deleteAllPairFirsts):
-        (WTF::deleteAllKeys):
-
-2006-11-07  Anders Carlsson  <acarlsson@apple.com>
-
-        Reviewed by Geoff.
-
-        * API/JSClassRef.cpp:
-        (OpaqueJSClass::OpaqueJSClass):
-        Initialize cachedPrototype to 0.
-
-2006-11-06  Krzysztof Kowalczyk  <kkowalczyk@gmail.com>
-
-        Reviewed by Maciej.
-
-        Remove warning about garbage after #else. #else clause applies for all
-        non-mac platforms, not only win.
-
-        * kjs/date_object.cpp:
-
-2006-11-06  Mark Rowe  <bdash@webkit.org>
-
-        Reviewed by the wonderful Mitz Pettel.
-
-        http://bugs.webkit.org/show_bug.cgi?id=11524
-        Bug 11524: REGRESSION(r9842): Array.prototype.join should use ToString operator rather than calling toString on each element
-
-        * kjs/array_object.cpp:
-        (ArrayProtoFunc::callAsFunction): Use ToString operator on each element rather than calling their toString method.
-
-2006-11-03  Steve Falkenburg  <sfalken@apple.com>
-
-        Fix build
-
-        * kjs/JSImmediate.h:
-
-2006-11-03  Alexey Proskuryakov  <ap@nypop.com>
-
-        Reviewed by Darin.
-
-        http://bugs.webkit.org/show_bug.cgi?id=11504
-        Fix warnings on non 32 bit platforms
-
-        * kjs/JSImmediate.h:
-        (KJS::JSImmediate::NanAsBits):
-        (KJS::JSImmediate::oneAsBits):
-        Rewrite in a way that moves runtime checks to compile-time.
-
-        (KJS::):
-        (KJS::JSImmediate::fromDouble):
-        (KJS::JSImmediate::toDouble):
-
-2006-11-02  George Staikos <staikos@kde.org>
-
-        Reviewed by Maciej.
-
-        * collector.cpp:
-        Remove a deprecated pthreads call.
-
-2006-11-02  Anders Carlsson  <acarlsson@apple.com>
-
-        Reviewed by Maciej, landed by Anders.
-
-        * CMakeLists.txt:
-        Make KDE support optional.
-
-2006-11-01  Kevin McCullough  <KMcCullough@apple.com>
-
-        Reviewed by Brady.
-
-        - Fixes many JavaScriptCore tests in other timezones.  The root problem is that on mac localtime() returns historically accurate information for DST, but the JavaScript spec explicitly states to not take into account historical information but rather to interpolate from valid years.
-
-        * kjs/DateMath.cpp:
-        (KJS::equivalentYearForDST):
-        (KJS::getDSTOffsetSimple):
-        (KJS::getDSTOffset):
-
-2006-10-31  Geoffrey Garen  <ggaren@apple.com>
-
-        Reviewed by Beth.
-        
-        Fixed http://bugs.webkit.org/show_bug.cgi?id=11477
-        REGRESSION: GMail crashes in KJS::FunctionImp::callerGetter
-
-        * kjs/function.cpp:
-        (KJS::FunctionImp::argumentsGetter): Removed unnecessary braces.
-        (KJS::FunctionImp::callerGetter): More logical NULL checking.
-
-2006-10-31  Oliver Hunt  <oliver@apple.com>
-
-        Reviewed by Geoff.
-
-        Adding definition for PLATFORM(CI)
-
-        * wtf/Platform.h:
-
-2006-10-31  Vladimir Olexa  <vladimir.olexa@gmail.com>
-
-        Reviewed by Geoff.
-
-        http://bugs.webkit.org/show_bug.cgi?id=4166
-        Function object does not support caller property
-
-        Test: fast/js/caller-property.html
-
-        * kjs/function.cpp: 
-        (KJS::FunctionImp::callerGetter): added
-        (KJS::FunctionImp::getOwnPropertySlot): added if statement to handle callerGetter()
-        * kjs/function.h: added callerGetter() declaration
-        * kjs/identifier.h: added caller property macro
-        * tests/mozilla/expected.html: 
-
-2006-10-30  Kevin McCullough  <KMcCullough@apple.com>
-
-        Reviewed by Adam.
-
-        - Fix some timezone issues and JavaScriptCore date tests.  Addresses bugzilla 4930.
-
-        * kjs/DateMath.h:
-        (KJS::GregorianDateTime::GregorianDateTime): Here's the fix, to add parenthesis for order of precedence.
-        * kjs/date_object.cpp:
-        (KJS::DateProtoFunc::callAsFunction): 
-        (KJS::DateObjectImp::construct): memset not needed as GregorianDateTime initializes itself.
-
-2006-10-30  Darin Adler  <darin@apple.com>
-
-        Reviewed by John Sullivan.
-
-        * kjs/SavedBuiltins.h: Added needed include.
-        * wtf/OwnPtr.h: (WTF::OwnPtr::set): Fixed mistake in assertion.
-
-2006-10-28  Darin Adler  <darin@apple.com>
-
-        Reviewed by Maciej.
-
-        - renamed PassRefPtr::release to releaseRef to make it clearer that
-          it's the counterpart of adoptRef, and to make it harder to confuse
-          it with the safer-to-use RefPtr::release
-
-        * kjs/identifier.cpp:
-        (KJS::CStringTranslator::translate):
-        (KJS::UCharBufferTranslator::translate):
-        * kjs/ustring.cpp:
-        (KJS::UString::Rep::create):
-        * wtf/PassRefPtr.h:
-        (WTF::PassRefPtr::PassRefPtr):
-        (WTF::PassRefPtr::~PassRefPtr):
-        (WTF::PassRefPtr::get):
-        (WTF::PassRefPtr::releaseRef):
-        (WTF::PassRefPtr::operator->):
-        (WTF::PassRefPtr::operator=):
-        (WTF::adoptRef):
-        (WTF::static_pointer_cast):
-        (WTF::const_pointer_cast):
-        * wtf/RefPtr.h:
-        (WTF::RefPtr::RefPtr):
-        (WTF::RefPtr::operator=):
-
-2006-10-28  Darin Adler  <darin@apple.com>
-
-        Reviewed by Steve.
-
-        * kjs/grammar.y: Add definitions of YYMALLOC and YYFREE to fix
-        a warning some people see (not sure why others don't see it).
-
-        * JavaScriptCore.vcproj/JavaScriptCore/grammarWrapper.cpp: Touch
-        this file to force it to re-build grammar.cpp.
-
-2006-10-28  Darin Adler  <darin@apple.com>
-
-        Reviewed by Geoff.
-
-        - made changes so the code compiles with the highest warning level
-          under MSVC (disabling some warnings, making some code fixes)
-
-        * API/JSCallbackConstructor.cpp:
-        (KJS::JSCallbackConstructor::construct):
-        * API/JSCallbackFunction.cpp:
-        (KJS::JSCallbackFunction::callAsFunction):
-        * API/JSCallbackObject.cpp:
-        (KJS::JSCallbackObject::init):
-        (KJS::JSCallbackObject::construct):
-        (KJS::JSCallbackObject::callAsFunction):
-        * API/JSObjectRef.cpp:
-        (JSPropertyNameArrayGetNameAtIndex):
-        * API/JSStringRef.cpp:
-        (JSStringCreateWithCharacters):
-        * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
-        * bindings/c/c_utility.cpp:
-        (KJS::Bindings::convertUTF8ToUTF16):
-        (KJS::Bindings::coerceValueToNPVariantStringType):
-        (KJS::Bindings::convertValueToNPVariant):
-        * kjs/DateMath.h:
-        (KJS::GregorianDateTime::GregorianDateTime):
-        * kjs/ExecState.h:
-        (KJS::ExecState::hadException):
-        * kjs/JSImmediate.h:
-        (KJS::JSImmediate::fromDouble):
-        (KJS::JSImmediate::toDouble):
-        (KJS::JSImmediate::NanAsBits):
-        (KJS::JSImmediate::oneAsBits):
-        * kjs/Parser.h:
-        * kjs/PropertyNameArray.h:
-        (KJS::PropertyNameArray::size):
-        * kjs/array_object.cpp:
-        (ArrayObjectImp::callAsFunction):
-        * kjs/bool_object.cpp:
-        (BooleanObjectImp::callAsFunction):
-        * kjs/collector.cpp:
-        (KJS::Collector::allocate):
-        (KJS::Collector::markCurrentThreadConservatively):
-        (KJS::Collector::collect):
-        * kjs/completion.h:
-        (KJS::Completion::isValueCompletion):
-        * kjs/date_object.cpp:
-        (KJS::findMonth):
-        * kjs/debugger.cpp:
-        (Debugger::sourceParsed):
-        (Debugger::sourceUnused):
-        (Debugger::exception):
-        (Debugger::atStatement):
-        (Debugger::callEvent):
-        (Debugger::returnEvent):
-        * kjs/dtoa.cpp:
-        * kjs/error_object.cpp:
-        (ErrorObjectImp::callAsFunction):
-        (NativeErrorImp::callAsFunction):
-        * kjs/function.cpp:
-        (KJS::FunctionImp::processVarDecls):
-        (KJS::GlobalFuncImp::callAsFunction):
-        * kjs/function_object.cpp:
-        (FunctionPrototype::callAsFunction):
-        * kjs/grammar.y:
-        * kjs/identifier.cpp:
-        (KJS::CStringTranslator::translate):
-        (KJS::Identifier::add):
-        * kjs/internal.h:
-        * kjs/lexer.cpp:
-        (Lexer::lex):
-        (Lexer::isIdentStart):
-        (Lexer::isIdentPart):
-        (isDecimalDigit):
-        (Lexer::isHexDigit):
-        (Lexer::isOctalDigit):
-        (Lexer::matchPunctuator):
-        (Lexer::singleEscape):
-        (Lexer::convertOctal):
-        (Lexer::convertHex):
-        (Lexer::convertUnicode):
-        (Lexer::record8):
-        * kjs/lexer.h:
-        * kjs/math_object.cpp:
-        (MathFuncImp::callAsFunction):
-        * kjs/number_object.cpp:
-        (integer_part_noexp):
-        (intPow10):
-        (NumberProtoFunc::callAsFunction):
-        (NumberObjectImp::callAsFunction):
-        * kjs/object.cpp:
-        (KJS::JSObject::deleteProperty):
-        (KJS::JSObject::callAsFunction):
-        (KJS::JSObject::toBoolean):
-        (KJS::JSObject::toObject):
-        * kjs/object.h:
-        (KJS::JSObject::getPropertySlot):
-        * kjs/property_map.cpp:
-        (KJS::isValid):
-        (KJS::PropertyMap::put):
-        (KJS::PropertyMap::insert):
-        (KJS::PropertyMap::containsGettersOrSetters):
-        * kjs/property_map.h:
-        (KJS::PropertyMap::hasGetterSetterProperties):
-        * kjs/property_slot.h:
-        * kjs/string_object.cpp:
-        (StringInstance::getPropertyNames):
-        (StringObjectImp::callAsFunction):
-        (StringObjectFuncImp::callAsFunction):
-        * kjs/ustring.cpp:
-        (KJS::UString::Rep::computeHash):
-        (KJS::UString::UString):
-        (KJS::UString::from):
-        (KJS::UString::append):
-        (KJS::UString::ascii):
-        (KJS::UString::operator=):
-        (KJS::UString::find):
-        (KJS::UString::rfind):
-        * kjs/ustring.h:
-        (KJS::UChar::high):
-        (KJS::UChar::low):
-        (KJS::UCharReference::low):
-        (KJS::UCharReference::high):
-        * kjs/value.cpp:
-        (KJS::JSValue::toUInt16):
-        * kjs/value.h:
-        * pcre/pcre_compile.c:
-        (get_othercase_range):
-        * pcre/pcre_exec.c:
-        (match):
-        * pcre/pcre_internal.h:
-        * wtf/HashFunctions.h:
-        (WTF::intHash):
-        (WTF::PtrHash::hash):
-        * wtf/MathExtras.h:
-        (isnan):
-        (lround):
-        (lroundf):
-        * wtf/StringExtras.h:
-        (strncasecmp):
-        * wtf/unicode/icu/UnicodeIcu.h:
-        (WTF::Unicode::isPrintableChar):
-
-2006-10-26  W. Andy Carrel  <wac@google.com>
-
-        Reviewed by Maciej.
-
-        - Fix http://bugs.webkit.org/show_bug.cgi?id=7445 /
-          <rdar://problem/4614195> (and 7253 / <rdar://4694011>) by changing
-          inline regexps so that they can have \u escaped Unicode sequences and
-          still work properly.
-
-        * kjs/lexer.cpp:
-        (Lexer::Lexer): 
-        (Lexer::setCode):
-        (Lexer::shift): Looking ahead one additional character for the benefit
-        of scanRegExp
-        (Lexer::scanRegExp): Change code to support unicode escapes in inline 
-        regexps.
-        * kjs/lexer.h: Extra lookahead added.
-
-=== Safari-521.29 ===
-
-2006-10-26  Nikolas Zimmermann  <zimmermann@kde.org>
-
-        Reviewed by Darin.
-
-        Fix build with older gcc 3.3.4.
-
-        * kjs/DateMath.cpp: Remove inline prefix.
-        (KJS::equivalentYearForDST):
-
-2006-10-26  Darin Adler  <darin@apple.com>
-
-        Reviewed by John.
-
-        - fix iteration of properties of string objects (found because of a warning
-          emitted by the MSVC compiler)
-
-        * kjs/string_object.cpp: (StringInstance::getPropertyNames): Change code that
-        wants to format a number as a string to use UString::from. Before it was using
-        the UString constructor that makes a string from a character!
-
-        * kjs/ustring.h:
-        * kjs/ustring.cpp: Remove the dangerous and not all that helpful UString(char)
-        constructor.
-
-        * kjs/grammar.y: Change code to not depend on the UString(char) constructor.
-        This is potentially more efficient anyway because we could overload the + operator
-        some day to handle char* directly instead of creating a UString.
-
-        * kjs/nodes2string.cpp: (SourceStream::operator<<): Change code to not depend on
-        the UString(char) constructor.
-
-2006-10-25  Kevin McCullough  <KMcCullough@apple.com>
-
-        Reviewed by Steve (rubber stamp).
-
-        - Link against your local build of JavaScriptCore.lib first, this fixes some errors on release builds of testkjs.
-
-        * JavaScriptCore.vcproj/testkjs/testkjs.vcproj:
-
-2006-10-25  Geoffrey Garen  <ggaren@apple.com>
-
-        Reviewed by Lou.
-        
-        Removed duplicate symbol declaration.
-
-        * JavaScriptCore.xcodeproj/project.pbxproj:
-        * kjs/grammar.y:
-
-2006-10-24  Steve Falkenburg  <sfalken@apple.com>
-
-        Build config change
-
-        * JavaScriptCore.vcproj/testkjs/testkjs.vcproj:
-
-2006-10-24  Kevin McCullough  <KMcCullough@apple.com>
-
-        Reviewed by Brady.
-
-        - Fixes a date formatting issue on win.  Specifically strftime cannot handle some ranges of time so we shift time call strftime and then manipulate the returned string, if needed.
-
-        * kjs/date_object.cpp:
-        (KJS::):
-        (KJS::formatLocaleDate):
-        (KJS::DateProtoFunc::callAsFunction):
-
-2006-10-23  Kevin McCullough  <KMcCullough@apple.com>
-
-        Reviewed by 
-
-       - Build fix
-
-        * JavaScriptCore.xcodeproj/project.pbxproj:
-        * kjs/grammar.y:
-
-2006-10-23  Kevin McCullough  <KMcCullough@apple.com>
-
-        Reviewed by Maciej.
-
-        - Makes the toTM function an operator.  Was going to piggy back on a patch but the patch needs more work.
-
-        * kjs/DateMath.cpp:
-        (KJS::equivalentYearForDST):
-        * kjs/DateMath.h:
-        (KJS::GregorianDateTime::operator tm):
-        * kjs/date_object.cpp:
-        (KJS::formatTime):
-        (KJS::DateProtoFunc::callAsFunction):
-
-2006-10-23  Kevin McCullough  <KMcCullough@apple.com>
-
-        Reviewed by Maciej.
-
-       - Fixes two regressions on win.  Both are stack overflows. For one the number of recursions is capped at 100, and for the other, nested parenthesis pairs are not evaluated (since they would evaluate to whatever is in them anyway).
-
-        * kjs/grammar.y:
-        * kjs/object.cpp:
-
-2006-10-21  Steve Falkenburg  <sfalken@apple.com>
-
-        Reviewed by Adam.
-        
-        Add minimal compatibility with MSVCRT leak checker
-
-        * wtf/FastMalloc.h:
-
-2006-10-23  Kevin McCullough  <KMcCullough@apple.com>
-
-        Reviewed by Geof.
-
-        - Sets the lowercase range correctly in the test and consolidates a variable to make the test more readable.
-
-       * tests/mozilla/ecma/String/15.5.4.11-2.js:
-
-2006-10-21  Darin Adler  <darin@apple.com>
-
-        Reviewed by Anders.
-
-        - http://bugs.webkit.org/show_bug.cgi?id=11377
-          swap(Vector, Vector) should be O(1) instead of O(n)
-
-        * wtf/Vector.h:
-        (WTF::VectorBuffer::swap): Added.
-        (WTF::Vector::swap): Added.
-        (WTF::swap): Added overload that takes two Vector objects.
-
-2006-10-21  Darin Adler  <darin@apple.com>
-
-        Reviewed by Adam.
-
-        - http://bugs.webkit.org/show_bug.cgi?id=11376
-          build scripts should invoke make with "-j" option for multiple processors
-
-        * JavaScriptCore.xcodeproj/project.pbxproj: Pass -j `sysctl -n hw.ncpu` to make.
-
-2006-10-19  Kevin McCullough  <KMcCullough@apple.com>
-
-        Reviewed by Geof.
-
-        Changed test to make us pass Georgian case changing for Unicode 4.0 and 5.0.  This incorporates changes from the 1.4 revision of the same mozilla test.
-        On Tiger we are still using Unicode 4.0 but on win and Leopard we are using Unicode 5.0, so this test currently allows for either answer.
-
-        * tests/mozilla/ecma/String/15.5.4.11-2.js:
-
-2006-10-18  Maciej Stachowiak  <mjs@apple.com>
-
-        Reviewed by Geoff.
-        
-        - remove vestiges of KXMLCore name (former name of WTF).
-
-        * wtf/Assertions.h:
-        * wtf/FastMalloc.h:
-        (operator new):
-        (operator delete):
-        (operator new[]):
-        (operator delete[]):
-        * wtf/FastMallocInternal.h:
-        * wtf/Forward.h:
-        * wtf/GetPtr.h:
-        * wtf/HashCountedSet.h:
-        * wtf/HashFunctions.h:
-        * wtf/HashMap.h:
-        * wtf/HashSet.h:
-        * wtf/HashTable.h:
-        * wtf/HashTraits.h:
-        * wtf/ListRefPtr.h:
-        * wtf/MathExtras.h:
-        * wtf/Noncopyable.h:
-        * wtf/OwnArrayPtr.h:
-        * wtf/OwnPtr.h:
-        * wtf/PassRefPtr.h:
-        * wtf/Platform.h:
-        * wtf/RefPtr.h:
-        * wtf/StringExtras.h:
-        (snprintf):
-        * wtf/UnusedParam.h:
-        * wtf/Vector.h:
-        * wtf/VectorTraits.h:
-
-2006-10-17  Steve Falkenburg  <sfalken@apple.com>
-
-        Reviewed by Maciej.
-        
-        Adjust include paths
-
-        * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
-
-2006-10-17  Kevin McCullough  <KMcCullough@apple.com>
-
-        Reviewed by Darin.
-
-        Fixed a date issue where the UTC offset was not set in win.
-
-        * kjs/DateMath.cpp:
-        (KJS::getDSTOffsetSimple):
-        (KJS::getDSTOffset):
-        (KJS::msToGregorianDateTime):
-        * kjs/DateMath.h:
-        (KJS::):
-        (KJS::GregorianDateTime::GregorianDateTime):
-
-2006-10-17  Kevin McCullough  <KMcCullough@apple.com>
-
-        Reviewed by Brady.
-
-        Fixes a JavaScriptCore math issue on win.
-
-        * kjs/math_object.cpp:
-        (MathFuncImp::callAsFunction):
-        * wtf/MathExtras.h:
-        (wtf_atan2):
-
-2006-10-16  Kevin McCullough  <kmccullough@apple.com>
-
-        Reviewed by Geof.
-
-        Removed unecessary global specifiers.
-
-        * kjs/math_object.cpp:
-        (MathFuncImp::callAsFunction):
-
-2006-10-16  Kevin McCullough  <KMcCullough@apple.com>
-
-        Reviewed by John.
-
-        Fixes a compile order issue for testkjs on win.
-
-        * JavaScriptCore.vcproj/testkjs/testkjs.vcproj:
-
-2006-10-15  Krzysztof Kowalczyk  <kkowalczyk@gmail.com>
-
-        Reviewed by Anders.
-
-        Remove junk (as gcc calls it) after #else clause.
-
-        * wtf/FastMalloc.cpp:
-        (WTF::do_free):
-
-2006-10-14  Krzysztof Kowalczyk  <kkowalczyk@gmail.com>
-
-        Reviewed by Maciej.
-
-        Define KXMLCORE_USE_CURL for platforms that wish to use CURL as
-        networking, and set it for GDK build
-
-        * wtf/Platform.h:
-
-2006-10-13  Marvin Decker  <marv.decker@gmail.com>
-
-        Reviewed by Kevin McCullough.
-
-        Fixes http://bugs.webkit.org/show_bug.cgi?id=11283
-        Fixes Qt/Linux and Windows build
-
-        * kjs/DateMath.cpp:
-        * kjs/DateMath.h:
-        * kjs/date_object.cpp:
-        (KJS::DateProtoFunc::callAsFunction):
-
-2006-10-13  Kevin McCullough  <KMcCullough@apple.com>
-
-        Reviewed by Adam, Geoff, Darin.
-
-       Fixed displaying the UTC offset and time zone string, as well as renamed the GregorianDateTime structure and clean up. 
-
-        * ChangeLog:
-        * kjs/DateMath.cpp:
-        (KJS::getUTCOffset):
-        (KJS::getDSTOffsetSimple):
-        (KJS::gregorianDateTimeToMS):
-        (KJS::msToGregorianDateTime):
-        * kjs/DateMath.h:
-        (KJS::GregorianDateTime::GregorianDateTime):
-        (KJS::GregorianDateTime::~GregorianDateTime):
-        (KJS::GregorianDateTime::toTM):
-        * kjs/date_object.cpp:
-        (KJS::gmtoffset):
-        (KJS::formatDate):
-        (KJS::formatDateUTCVariant):
-        (KJS::formatTime):
-        (KJS::fillStructuresUsingTimeArgs):
-        (KJS::fillStructuresUsingDateArgs):
-        (KJS::DateInstance::getTime):
-        (KJS::DateInstance::getUTCTime):
-        (KJS::DateProtoFunc::callAsFunction):
-        (KJS::DateObjectImp::construct):
-        (KJS::DateObjectImp::callAsFunction):
-        (KJS::DateObjectFuncImp::callAsFunction):
-        (KJS::parseDate):
-        * kjs/date_object.h:
-
-2006-10-13  Kevin McCullough  <KMcCullough@apple.com>
-
-        Reviewed by Adam.
-
-        Gets JavaScripCore tests running on windows.
-
-        * Scripts/run-javascriptcore-tests:
-        * Scripts/webkitdirs.pm:
-
-2006-10-12  Geoffrey Garen  <ggaren@apple.com>
-
-        Reviewed by Maciej.
-        
-        Removed JSObjectMakeWithPrototype, clarified some comments. We really
-        don't want people to manage their own prototypes, so we don't want an
-        extra function in the API devoted to just that. People can still manage
-        their own prototypes if they really want by using JSObjectSetPrototype.
-
-        * API/JSClassRef.cpp:
-        (OpaqueJSClass::createNoAutomaticPrototype):
-        (OpaqueJSClass::create):
-        * API/JSClassRef.h:
-        * API/JSObjectRef.cpp:
-        (JSClassCreate):
-        (JSObjectMake):
-        * API/JSObjectRef.h:
-        * API/testapi.c:
-        (main):
-        * JavaScriptCore.exp:
-
-2006-10-12  Kevin McCullough  <KMcCullough@apple.com>
-
-        Reviewed by Adam.
-
-        Build breakage fix
-
-        * kjs/DateMath.cpp:
-        (KJS::msToTM):
-        * kjs/date_object.cpp:
-        (KJS::gmtoffset):
-
-2006-10-11  Kevin McCullough  <KMcCullough@apple.com>
-
-        Reviewed by Geoff.
-
-        Added our own tm struct to have a consistent set of fields, which lets us display the DST offset and timezone strings correctly.  Also there is some code cleanup.
-
-        * kjs/DateMath.cpp:
-        (KJS::timeToMS):
-        (KJS::getUTCOffset):
-        (KJS::getDSTOffsetSimple):
-        (KJS::dateToMS):
-        (KJS::msToTM):
-        (KJS::tmToKJStm):
-        (KJS::KJStmToTm):
-        * kjs/DateMath.h:
-        * kjs/date_object.cpp: 
-        (KJS::gmtoffset): 
-        (KJS::formatTime): 
-        (KJS::DateProtoFunc::callAsFunction): 
-        (KJS::DateObjectImp::construct):
-        (KJS::DateObjectImp::callAsFunction):
-        (KJS::DateObjectFuncImp::callAsFunction):
-        (KJS::parseDate):
-        * kjs/date_object.h:
-
-2006-10-09  Krzysztof Kowalczyk  <kkowalczyk@gmail.com>
-
-        Reviewed by Geoff.
-
-        Improve gdk build compiler flags (show warning, no rtti and exceptions).
-
-        * jscore.bkl:
-
-2006-10-06  Kevin McCullough  <KMcCullough@apple.com>
-
-        Reviewed by Brady.
-
-        DST and TimeZones were wrong in some cases, specifically on some of the dates where DST changes.
-
-        * kjs/DateMath.cpp:
-        (KJS::equivalentYearForDST):
-        (KJS::getUTCOffset):
-        (KJS::getDSTOffsetSimple):
-        (KJS::getDSTOffset):
-        (KJS::dateToMseconds):
-        (KJS::msToTM):
-        * kjs/DateMath.h:
-        * kjs/date_object.cpp:
-        (KJS::gmtoffset):
-
-2006-10-05  Darin Adler  <darin@apple.com>
-
-        Reviewed by Kevin McCullough.
-
-        * wtf/Assertions.cpp: Fix build when _DEBUG is not defined.
-
-2006-10-04  Kevin McCullough  <KMcCullough@apple.com>
-
-        Reviewed by Adam.
-
-        - Removed an unnecessary assert that was stopping many pages.  tm_gmtoff was not set for UTC time in mozilla but is always set for us.
-
-        * kjs/DateMath.cpp:
-        (KJS::getUTCOffset):
-        (KJS::msToTM):
-        * kjs/date_object.cpp:
-        (KJS::gmtoffset):
-        (KJS::formatTime):
-
-2006-10-04  Geoffrey Garen  <ggaren@apple.com>
-
-        Patch by Darin and me, reviewed by Maciej.
-
-        Fixed <rdar://problem/4518397> REGRESSION(?): Oft-seen but unrepro crash 
-              in JavaScript garbage collection (KJS::Collector::collect())
-               <rdar://problem/4752492> Crash in KJS::collect
-              
-        The issue here was allocating one garbage-collected object in the midst 
-        of allocating a second garbage-collected object. In such a case, the
-        zeroIfFree word lies.
-
-        * kjs/collector.cpp:
-        (KJS::Collector::allocate):
-        (KJS::Collector::collect):
-
-2006-10-04  Kevin McCullough  <KMcCullough@apple.com>
-
-        Reviewed by Adam.
-
-        - Layout test fix
-
-        * kjs/DateMath.cpp:
-        (KJS::dateToDayInYear): accept and correctly handle negative months
-
-2006-10-05  Kevin McCullough  <KMcCullough@apple.com>
-
-       build fix
-
-        * kjs/DateMath.cpp:
-        (KJS::dateToDayInYear):
-
-2006-10-05  Mark Rowe  <bdash@webkit.org>
-
-        Reviewed by maculloch.
-
-        Gdk build fix.
-
-        * JavaScriptCoreSources.bkl: Add DateMath.cpp to file list.
-
-2006-10-05  Kevin McCullough  <KMcCullough@apple.com>
-
-        Reviewed by aroben
-
-        - build fix
-
-        * JavaScriptCore.xcodeproj/project.pbxproj:
-
-2006-10-04  Nikolas Zimmermann  <zimmermann@kde.org>
-
-        Reviewed by Mitz.
-
-        Fix Qt/Linux build by adding DateMath.cpp to compilation.
-
-        * CMakeLists.txt: Also replace tabs with spaces.
-
-2006-10-04  Kevin McCullough  <KMcCullough@apple.com>
-
-        Reviewed by DethBakin.
-
-        - Apparently the build bot uses an older version of XCode which warns about conversions and the newest version does not.  I hope this fixes the build but I cann't be sure on my system.
-
-        * kjs/DateMath.cpp:
-        (KJS::msToYear):
-        (KJS::dayInYear):
-        (KJS::dateToDayInYear):
-
-2006-10-05  Darin Adler  <darin@apple.com>
-
-        Reviewed by Adam.
-
-        * wtf/Assertions.cpp: Changed assertion formatting to omit the "======"
-        lines so you can see more assertions in less space. Also improved format
-        of file/line information so it works with more development environments.
-
-2006-10-04  Kevin McCullough  <KMcCullough@apple.com>
-
-        Reviewed by Tim H.
-
-        - The build machine is more sensitive about automatic conversions.  These fixes exp
-licitly cast or change the input and return types of functions to avoid conversions.
-
-        * JavaScriptCore.xcodeproj/project.pbxproj:
-        * kjs/DateMath.cpp:
-        (KJS::):
-        (KJS::msToDays):
-        (KJS::msToYear):
-        (KJS::dayInYear):
-        (KJS::monthToDayInYear):
-        (KJS::dateToDayInYear):
-        (KJS::getDSTOffsetSimple):
-        (KJS::getDSTOffset):
-        (KJS::dateToMseconds):
-        (KJS::msToTM):
-
-2006-10-04  Kevin McCullough  <KMcCullough@apple.com>
-
-        Reviewed by GGaren
-
-        - This is a big makeover for our Date implemenetation.  This solves many platform specific issues, specifically dates before 1970, and simplifies some ugly code.  The purpose of this was to get us to pass many of the JavaScriptCore tests on windows.
-
-        * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
-        * JavaScriptCore.xcodeproj/project.pbxproj:
-        * kjs/DateMath.cpp: Added.
-        (KJS::):
-        (KJS::daysInYear):
-        (KJS::daysFrom1970ToYear):
-        (KJS::msFrom1970ToYear):
-        (KJS::msToDays):
-        (KJS::msToYear):
-        (KJS::isLeapYear):
-        (KJS::isInLeapYear):
-        (KJS::dayInYear):
-        (KJS::msToMilliseconds):
-        (KJS::msToWeekDay):
-        (KJS::msToSeconds):
-        (KJS::msToMinutes):
-        (KJS::msToHours):
-        (KJS::msToMonth):
-        (KJS::msToDayInMonth):
-        (KJS::monthToDayInYear):
-        (KJS::timeToMseconds):
-        (KJS::dateToDayInYear):
-        (KJS::equivalentYearForDST):
-        (KJS::getUTCOffset):
-        (KJS::getDSTOffsetSimple):
-        (KJS::getDSTOffset):
-        (KJS::localTimeToUTC):
-        (KJS::UTCToLocalTime):
-        (KJS::dateToMseconds):
-        (KJS::msToTM):
-        (KJS::isDST):
-        * kjs/DateMath.h: Added.
-        (KJS::):
-        * kjs/date_object.cpp:
-        (KJS::gmtoffset):
-        (KJS::formatTime):
-        (KJS::DateInstance::getTime):
-        (KJS::DateInstance::getUTCTime):
-        (KJS::DateProtoFunc::callAsFunction):
-        (KJS::DateObjectImp::construct):
-        (KJS::DateObjectFuncImp::callAsFunction):
-        (KJS::parseDate):
-        * kjs/testkjs.cpp:
-        * os-win32/stdint.h:
-
-2006-10-02  Nikolas Zimmermann  <zimmermann@kde.org>
-
-        Reviewed/landed by Adam.
-
-        Build testkjs on Qt/Linux.
-
-        * CMakeLists.txt:
-
-2006-10-02  Nikolas Zimmermann  <zimmermann@kde.org>
-
-        Reviewed by eseidel.  Landed by eseidel.
-
-        Fix win32 build, which has no inttypes.h
-
-        * wtf/Assertions.h:
-
-2006-10-02  Nikolas Zimmermann <zimmermann@kde.org>
-
-        Reviewed by eseidel & mjs.  Landed by eseidel.
-
-        Fix Qt/Linux build with older gcc 3.3.4.
-        http://bugs.webkit.org/show_bug.cgi?id=11116
-        * kjs/lookup.h: Move cacheGlobalObject into KJS namespace.
-        (KJS::cacheGlobalObject): Also remove GCC_ROOT_NS_HACK.
-        * wtf/Assertions.h: Include inttypes.h for uintptr_t.
-
-2006-09-28  Steve Falkenburg  <sfalken@apple.com>
-
-        Reviewed by Maciej.
-        
-        Use $(ConfigSuffix) set via vsprops files to add _debug
-        to end of debug filenames.
-
-        * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
-        * JavaScriptCore.vcproj/debug.vsprops: Added.
-        * JavaScriptCore.vcproj/dftables/dftables.vcproj:
-        * JavaScriptCore.vcproj/release.vsprops: Added.
-        * JavaScriptCore.vcproj/testkjs/testkjs.vcproj:
-
-2006-09-28  Darin Adler  <darin@apple.com>
-
-        Reviewed by Alice.
-
-        - support for change that should fix <rdar://problem/4733044>
-          REGRESSION: XML iBench shows 10% perf. regression (copying
-          strings while decoding)
-
-        * wtf/Vector.h: Changed VectorBuffer so that the general case
-        contains an instance of the 0 case, since deriving from it
-        was violating the Liskov Substitution Principle.
-        (WTF::VectorBuffer::releaseBuffer): Added. Releases the buffer so it can
-        be adopted by another data structure that uses the FastMalloc.h allocator.
-        Returns 0 if the internal buffer was being used.
-        (WTF::Vector::releaseBuffer): Added. Releases the buffer as above or creates
-        a new one in the case where the internal buffer was being used.
-
-2006-09-28  Maciej Stachowiak  <mjs@apple.com>
-
-        Reviewed by Geoff.
-        
-        - change garbage collection to happen at increments proportional to number of live objects, not always
-        every 1000 allocations
-
-        * kjs/collector.cpp:
-        (KJS::Collector::allocate):
-
-2006-09-28  Maciej Stachowiak  <mjs@apple.com>
-
-        Reviewed by Mitz.
-
-        - fixed REGRESSION (r16606): javascriptCore Crash on website load
-        
-        Plus style fixes.
-        
-            - fixed some possible off-by-one bugs
-            - use indexing, not iterators, for Vectors
-            - store Vector by pointer instead of by value to avoid blowing out FunctionImp size
-        
-        * kjs/function.cpp:
-        (KJS::FunctionImp::addParameter):
-        (KJS::FunctionImp::parameterString):
-        (KJS::FunctionImp::processParameters):
-        (KJS::FunctionImp::lengthGetter):
-        (KJS::FunctionImp::getParameterName):
-        * kjs/function.h:
-
-2006-09-27  Steve Falkenburg  <sfalken@apple.com>
-
-        Reviewed by Maciej.
-        
-        More build tweaks
-
-        * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
-        * JavaScriptCore.vcproj/JavaScriptCore/dstroot-to-sdk.cmd: Removed.
-
-2006-09-27  John Sullivan  <sullivan@apple.com>
-
-        * kjs/function.cpp:
-        (KJS::FunctionImp::getParameterName):
-        removed assertion that displeased gcc 4.0.1 (build 5420):
-        ASSERT(static_cast<size_t>(index) == index);
-
-2006-09-27  Kevin McCullough  <KMcCullough@apple.com>
-
-        Reviewed by GGaren.
-
-        Cleanup of previous fix which was to address Radar: 4752492
-
-        * kjs/function.cpp:
-        (KJS::FunctionImp::addParameter):
-        (KJS::FunctionImp::parameterString):
-        (KJS::FunctionImp::processParameters):
-        (KJS::FunctionImp::lengthGetter):
-        (KJS::FunctionImp::getParameterName):
-        * kjs/function.h:
-
-2006-09-27  Kevin McCullough  <KMcCullough@apple.com>
-
-        Reviewed by Adele.
-
-        Fixes a GC stack overflow crash.
-        The change is to move from a linked list implementation of Parameters to a Vector.
-        The problem with the linked list is that each one creates it's own stack frame when being destroyed and in extreme cases this caused the stack to overflow. 
-
-        * kjs/function.cpp:
-        (KJS::Parameter::Parameter):
-        (KJS::FunctionImp::addParameter):
-        (KJS::FunctionImp::parameterString):
-        (KJS::FunctionImp::processParameters):
-        (KJS::FunctionImp::lengthGetter):
-        (KJS::FunctionImp::getParameterName):
-        * kjs/function.h:
-
-2006-09-27  Steve Falkenburg  <sfalken@apple.com>
-
-        Fix last path fix.
-
-        * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
-
-2006-09-27  Steve Falkenburg  <sfalken@apple.com>
-
-        Set path before build.
-
-        * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
-
-2006-09-27  Sean Gies  <seangies@apple.com>
-
-        Reviewed by Adam Roben.
-
-        * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: Debug config should link to debug runtime.
-        * JavaScriptCore.vcproj/dftables/dftables.vcproj: Debug config should link to debug runtime.
-
-2006-09-27  Don Melton  <gramps@apple.com>
-
-        Reviewed by Adam Roben.
-
-        Changed line ending from DOS to UNIX format so it doesn't die running
-        on my machine. ;)
-
-        * JavaScriptCore.vcproj/JavaScriptCore/build-generated-files.sh:
-
-2006-09-23  Alexey Proskuryakov  <ap@nypop.com>
-
-        Reviewed by Maciej.
-
-        http://bugs.webkit.org/show_bug.cgi?id=10183
-        REGRESSION: obfuscated JS decoding breaks because of soft hyphen removal
-        (Fanfiction.net author pages not listing stories)
-
-        Rolled out the fix for bug 4139.
-
-        * kjs/lexer.cpp:
-        (Lexer::setCode):
-        (Lexer::shift):
-        * tests/mozilla/ecma/Array/15.4.5.1-1.js:
-        * tests/mozilla/expected.html:
-
-2006-09-22  Steve Falkenburg  <sfalken@apple.com>
-
-        Build fix.
-
-        * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
-
-2006-09-22  Darin Adler  <darin@apple.com>
-
-        Reviewed by Alice.
-
-        * wtf/Vector.h: Add an append that takes a pointer and length.
-        Generalize the existing Vector append to work on vectors with
-        any value for inlineCapacity. Change the append algorithm so
-        it doesn't check capacity each time through the loop.
-
-2006-09-22  Steve Falkenburg  <sfalken@apple.com>
-
-        Fix release build.
-
-        * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
-
-2006-09-21  Geoffrey Garen  <ggaren@apple.com>
-
-        Reviewed by Maciej.
-
-        * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
-        Updated to include the right path.
-        * wtf/FastMalloc.h: #include Platform.h, since we use Platform macros.
-
-=== Safari-521.27 ===
-
-2006-09-20  Anders Carlsson  <acarlsson@apple.com>
-
-        Reviewed by Dave Hyatt.
-
-        * wtf/MathExtras.h:
-        Get rid of lrint.
-
-2006-09-20  Sean Gies  <seangies@apple.com>
-
-        Reviewed by Steve Falkenburg.
-
-        * wtf/Assertions.cpp: Debug messages should go into debugger console.
-
-2006-09-20  David Hyatt  <hyatt@apple.com>
-
-        Add an implementation of lrint for Win32.
-
-        Reviewed by anders
-
-        * wtf/MathExtras.h:
-        (lrint):
-
-2006-09-15  Krzysztof Kowalczyk  <kkowalczyk@gmail.com>
-
-        Reviewed by Adam.
-
-        http://bugs.webkit.org/show_bug.cgi?id=10864
-        Bug 10864: Linux\GDK build fixes
-
-        * JavaScriptCoreSources.bkl:
-        * jscore.bkl:
-
-2006-09-15  Adam Roben  <aroben@apple.com>
-
-        Windows build fix.
-
-        * JavaScriptCore.vcproj/JavaScriptCore/build-generated-files.sh:
-
-2006-09-15  Anders Carlsson  <acarlsson@apple.com>
-
-        * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
-        Fix the release build.
-
-2006-09-15  Anders Carlsson  <acarlsson@apple.com>
-
-        Reviewed by Steve.
-
-        Add JavaScriptCore API to the build.
-        * API/JSBase.cpp:
-        * API/JSCallbackConstructor.cpp:
-        * API/JSCallbackFunction.cpp:
-        * API/JSCallbackObject.cpp:
-        * API/JSClassRef.cpp:
-        * API/JSContextRef.cpp:
-        * API/JSObjectRef.cpp:
-        * API/JSStringRef.cpp:
-        * API/JSValueRef.cpp:
-        * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
-        * os-win32/stdbool.h: Added.
-
-2006-09-12  Steve Falkenburg  <sfalken@apple.com>
-
-        Reviewed by Ada.
-        
-        Build tweaks (doing JavaScriptCore now since it doesn't have
-        dependencies).
-
-        * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
-        * JavaScriptCore.vcproj/JavaScriptCore/build-generated-files.sh:
-        * JavaScriptCore.vcproj/JavaScriptCore/dstroot-to-sdk.cmd: Added.
-        * JavaScriptCore.vcproj/dftables/dftables.vcproj:
-        * JavaScriptCore.vcproj/testkjs/testkjs.vcproj:
-
-2006-09-11  Brady Eidson  <beidson@apple.com>
-
-        Build fix - I think Tim's last checkin wasn't tested on Tiger, possibly.  I simply
-        commented out the undefined constants until he can have a chance to make the right call
-
-        * bindings/objc/objc_utility.mm:
-        (KJS::Bindings::objcValueTypeForType): Commented out undefined symbols
-
-2006-09-11  Timothy Hatcher  <timothy@apple.com>
-
-        Reviewed by Tim O. and Darin.
-
-        Add support for more method signatures affecting ObjC methods called from JavaScript:
-        - Added unsigned types and long long.
-        - Allow methods that use const, oneway, bycopy and byref type modifiers.
-
-        * bindings/objc/objc_instance.mm:
-        (ObjcInstance::invokeMethod):
-        * bindings/objc/objc_utility.h:
-        (KJS::Bindings::):
-        * bindings/objc/objc_utility.mm:
-        (KJS::Bindings::convertValueToObjcValue):
-        (KJS::Bindings::convertObjcValueToValue):
-        (KJS::Bindings::objcValueTypeForType):
-
-2006-09-05  Timothy Hatcher  <timothy@apple.com>
-
-        Reviewed by Tim O.
-
-        <rdar://problem/4715840> SEL is not char*
-
-        * bindings/objc/objc_class.mm:
-        (KJS::Bindings::ObjcClass::methodsNamed): use sel_getName instead of a char* cast.
-        * bindings/objc/objc_runtime.mm:
-        (ObjcFallbackObjectImp::callAsFunction): ditto
-
-2006-09-03  Alexey Proskuryakov  <ap@nypop.com>
-
-        Reviewed by Tim H.
-
-        http://bugs.webkit.org/show_bug.cgi?id=10693
-        Convert JavaScript arrays to AppleScript lists
-
-        * JavaScriptCore.exp: Export ArrayInstance::info and ArrayInstance::getItem().
-        * kjs/array_instance.h:
-        * kjs/array_object.cpp:
-        (ArrayInstance::getItem): Added a method to access array items from C++.
-
-2006-09-02  Krzysztof Kowalczyk  <kkowalczyk@gmail.com>
-
-        Reviewed by Tim H.
-
-        Bug 10454: Unix bakefile fixes
-        http://bugs.webkit.org/show_bug.cgi?id=10454
-
-        * JavaScriptCoreSources.bkl:
-
-2006-09-01  Nikolas Zimmermann  <zimmermann@kde.org>
-
-        Reviewed by hyatt.  Landed by eseidel.
-
-        Fix build on Linux.
-
-        * pcre/CMakeLists.txt: Add wtf/ include.
-
-2006-09-01  Nikolas Zimmermann  <zimmermann@kde.org>
-
-        Reviewed and landed by ap.
-
-        Fix build on Linux (C89 without gcc extensions enabled).
-
-        * pcre/pcre_internal.h: Use C style comments.
-        * wtf/Assertions.h: Use C style comments.
-        * wtf/Platform.h: Use C style comments.
-
-2006-09-01  Steve Falkenburg  <sfalken@apple.com>
-
-        Fix build.
-
-        * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
-        * JavaScriptCore.vcproj/dftables/dftables.vcproj:
-
-2006-08-31  Anders Carlsson  <acarlsson@apple.com>
-
-        Reviewed by Darin.
-        
-        Add new portability functions to MathExtras.h and add StringExtras.h which is for
-        string portability functions.
-        
-        * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
-        * bindings/c/c_instance.cpp:
-        * kjs/date_object.cpp:
-        * wtf/MathExtras.h:
-        (copysign):
-        (isfinite):
-        * wtf/StringExtras.h: Added.
-        (snprintf):
-        (strncasecmp):
-
-2006-08-31  Anders Carlsson  <acarlsson@apple.com>
-
-        Reviewed by Tim H.
-        
-        Fix Windows build.
-        
-        * JavaScriptCore.vcproj/dftables/dftables.vcproj:
-        * pcre/pcre_internal.h:
-
-2006-08-31  Timothy Hatcher  <timothy@apple.com>
-
-        Reviewed by Geoff.
-
-        Band-aid fix for PCRE to compile for ppc64 and x86_64 now that
-        we use -Wshorten-64-to-32. Adds an INT_CAST macro that ASSERTs
-        the value <= INT_MAX.
-
-        I filed <rdar://problem/4712064> to track the need to verify
-        PCRE's 64-bit compliance.
-
-        * pcre/pcre_compile.c:
-        (complete_callout):
-        (compile_branch):
-        (compile_regex):
-        (pcre_compile2):
-        * pcre/pcre_exec.c:
-        (match):
-        (pcre_exec):
-        * pcre/pcre_get.c:
-        (pcre_get_substring_list):
-        * pcre/pcre_internal.h:
-        * pcre/pcre_tables.c:
-        * pcre/pcre_try_flipped.c:
-        (_pcre_try_flipped):
-
-2006-08-30  Darin Adler  <darin@apple.com>
-
-        Reviewed by Tim Hatcher.
-
-        - add WTF::getPtr, a function template that makes it possible to write
-          generic code that gets a raw pointer out of any of our pointer types
-
-        * JavaScriptCore.xcodeproj/project.pbxproj:
-        * wtf/GetPtr.h: Added.
-        * wtf/ListRefPtr.h: (WTF::getPtr): Added.
-        * wtf/OwnArrayPtr.h: (WTF::getPtr): Added.
-        * wtf/OwnPtr.h: (WTF::getPtr): Added.
-        * wtf/PassRefPtr.h: (WTF::getPtr): Added.
-        * wtf/RefPtr.h: (WTF::getPtr): Added.
-
-2006-08-29  waylonis  <waylonis@google.com>
-
-        Reviewed, tweaked by ggaren.
-
-        - Added storage and accessor functions for ExecState as a fix for
-          http://bugs.webkit.org/show_bug.cgi?id=10114
-        
-        * kjs/ExecState.cpp:
-        (KJS::ExecState::ExecState):
-        * kjs/ExecState.h:
-        * kjs/context.h:
-        (KJS::Context::setExecState):
-        (KJS::Context::execState):
-
-2006-08-30  Nikolas Zimmermann  <zimmermann@kde.org>
-
-        Reviewed by Tim H.
-
-        Commit KDE related tweaks, to be able to
-        differentiate between a Qt-only or a KDE build.
-
-        * CMakeLists.txt: Install wtf-unity library.
-        * wtf/Platform.h: Add define for the KDE platform.
-
-2006-08-28  Darin Adler  <darin@apple.com>
-
-        Reviewed by Geoff.
-
-        * kjs/list.h: Use explicit in constructor (as appropriate).
-
-2006-08-24  Nikolas Zimmermann  <zimmermann@kde.org>
-
-        Reviewed, tweaked and landed by ap
-
-        http://bugs.webkit.org/show_bug.cgi?id=10467
-        WebKit should have Qt platform support (Part II)
-
-        * CMakeLists.txt: Adjust to Anders' build fixes.
-        * wtf/Platform.h: Fix define for the Qt platform (we don't use/need Cairo.)
-
-2006-08-23  David Hyatt  <hyatt@apple.com>
-
-        Fix Platform.h to include #defines for graphics features.
-
-        Reviewed by darin
-
-        * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
-        * wtf/Platform.h:
-
-2006-08-23  Anders Carlsson  <acarlsson@apple.com>
-
-        Reviewed by Darin.
-
-        Make the bindings compile without CoreFoundation.
-        
-        * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
-        * bindings/c/c_instance.cpp:
-        * bindings/c/c_utility.cpp:
-        (KJS::Bindings::convertUTF8ToUTF16):
-        * bindings/npapi.h:
-        * bindings/runtime.cpp:
-        (KJS::Bindings::Instance::createBindingForLanguageInstance):
-        (KJS::Bindings::Instance::createLanguageInstanceForValue):
-        * bindings/runtime_root.cpp:
-        * bindings/runtime_root.h:
-        * kjs/interpreter.cpp:
-        (KJS::Interpreter::createLanguageInstanceForValue):
-        * kjs/interpreter.h:
-
-2006-08-22  Anders Carlsson  <acarlsson@apple.com>
-
-        Reviewed by Darin.
-
-        Move the npruntime code over to using HashMap and the runtime_root code over to using 
-        HashMap and HashCountedSet.
-        
-        * bindings/NP_jsobject.cpp:
-        * bindings/c/c_utility.cpp:
-        (KJS::Bindings::identifierFromNPIdentifier):
-        * bindings/c/c_utility.h:
-        * bindings/jni/jni_jsobject.cpp:
-        (JavaJSObject::invoke):
-        * bindings/npruntime.cpp:
-        (getStringIdentifierMap):
-        (getIntIdentifierMap):
-        (_NPN_GetStringIdentifier):
-        (_NPN_GetIntIdentifier):
-        * bindings/runtime_root.cpp:
-        (getReferencesByRootMap):
-        (getReferencesSet):
-        (KJS::Bindings::findReferenceSet):
-        (KJS::Bindings::rootForImp):
-        (KJS::Bindings::rootForInterpreter):
-        (KJS::Bindings::addNativeReference):
-        (KJS::Bindings::removeNativeReference):
-        (RootObject::removeAllNativeReferences):
-        * bindings/runtime_root.h:
-
-2006-08-22  Anders Carlsson  <acarlsson@apple.com>
-
-        Reviewed by Geoff.
-        
-        Switch over the NPAPI and Java bindings to using HashMaps instead of dictionaries.
-        
-        * JavaScriptCore.xcodeproj/project.pbxproj:
-        * bindings/c/c_class.cpp:
-        (KJS::Bindings::CClass::CClass):
-        (KJS::Bindings::CClass::~CClass):
-        (KJS::Bindings::CClass::classForIsA):
-        (KJS::Bindings::CClass::methodsNamed):
-        (KJS::Bindings::CClass::fieldNamed):
-        * bindings/c/c_class.h:
-        * bindings/jni/jni_class.cpp:
-        (JavaClass::JavaClass):
-        (JavaClass::~JavaClass):
-        (JavaClass::methodsNamed):
-        (JavaClass::fieldNamed):
-        * bindings/jni/jni_class.h:
-        * bindings/objc/objc_class.h:
-        * bindings/objc/objc_class.mm:
-        (KJS::Bindings::deleteMethod):
-        (KJS::Bindings::deleteField):
-        (KJS::Bindings::):
-        (KJS::Bindings::ObjcClass::methodsNamed):
-        (KJS::Bindings::ObjcClass::fieldNamed):
-        * bindings/runtime.cpp:
-        * bindings/runtime.h:
-        * bindings/runtime_object.cpp:
-        (RuntimeObjectImp::fieldGetter):
-        (RuntimeObjectImp::getOwnPropertySlot):
-        (RuntimeObjectImp::put):
-        (RuntimeObjectImp::canPut):
-
-2006-08-21  Vladimir Olexa  <vladimir.olexa@gmail.com>
-
-        Reviewed by Darin.
-
-        http://bugs.webkit.org/show_bug.cgi?id=6252
-        JavaScript 1.6 Array.lastIndexOf
-
-        Test: fast/js/array-lastIndexOf.html
-
-        * kjs/array_object.cpp:
-        (ArrayProtoFunc::callAsFunction): Added a LastIndexOf case.
-        * kjs/array_object.h:
-        (KJS::ArrayProtoFunc::): Added LastIndexOf to enum.
-        * tests/mozilla/expected.html: Two more tests now pass.
-
-2006-08-20  Nikolas Zimmermann  <zimmermann@kde.org>
-
-        Reviewed by Maciej. Landed by rwlbuis.
-
-        Fixes parts of: http://bugs.webkit.org/show_bug.cgi?id=10463
-        WebKit should have Qt platform support
-
-        Removing obsolete QConstString/QString constructors in kjs code.
-
-        * kjs/identifier.h:
-        * kjs/ustring.h:
-
-2006-08-17  Nikolas Zimmermann  <zimmermann@kde.org>
-
-        Reviewed by Maciej. Landed by rwlbuis.
-
-        Fixes: http://bugs.webkit.org/show_bug.cgi?id=10463
-        WTF Changes needed for Qt platform code.
-
-        * wtf/Platform.h:
-        * wtf/unicode/UnicodeDecomposition.h: Added.
-        (WTF::Unicode::):
-        * wtf/unicode/UnicodeDirection.h: Added.
-        (WTF::Unicode::):
-        * wtf/unicode/qt4/UnicodeQt4.cpp: Added.
-        (WTF::Unicode::direction):
-        (WTF::Unicode::category):
-        (WTF::Unicode::decomposition):
-        * wtf/unicode/qt4/UnicodeQt4.h:
-        (WTF::Unicode::toLower):
-        (WTF::Unicode::toUpper):
-        (WTF::Unicode::isPrintableChar):
-        (WTF::Unicode::isSpace):
-        (WTF::Unicode::isPunct):
-        (WTF::Unicode::isDigit):
-        (WTF::Unicode::mirroredChar):
-        (WTF::Unicode::compare):
-
-2006-08-17  Nikolas Zimmermann  <zimmermann@kde.org>
-
-        Reviewed by Eric. Landed by rwlbuis.
-
-        Fixes: http://bugs.webkit.org/show_bug.cgi?id=10464
-        Offer a cmake build system for Qt platform.
-
-        * CMakeLists.txt: Added.
-        * pcre/CMakeLists.txt: Added.
-
-2006-08-17  Anders Carlsson  <acarlsson@apple.com>
-
-        Reviewed by Maciej.
-
-        * bindings/npapi.h:
-        Fix ifdef.
-
-2006-08-15  Steve Falkenburg  <sfalken@apple.com>
-
-        Reviewed by mjs.
-        
-        Build fix.
-
-        * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
-        * wtf/Assertions.h:
-
-2006-08-15  Mark Rowe  <opendarwin.org@bdash.net.nz>
-
-        Reviewed by Tim H.
-
-        Build fix:  DWARF and -gfull are incompatible with symbol separation.
-
-        * JavaScriptCore.xcodeproj/project.pbxproj:
-
-2006-08-15  Mark Rowe  <opendarwin.org@bdash.net.nz>
-
-        Reviewed by Tim H.
-
-        http://bugs.webkit.org/show_bug.cgi?id=10394
-        Bug 10394: WebKit Release and Production configurations should enable dead code stripping
-
-        * JavaScriptCore.xcodeproj/project.pbxproj:
-
-2006-08-15  Mark Rowe  <opendarwin.org@bdash.net.nz>
-
-        Reviewed by Tim H.
-
-        http://bugs.webkit.org/show_bug.cgi?id=10384
-        Bug 10384: Switch to DWARF for Release configuration
-
-        * JavaScriptCore.xcodeproj/project.pbxproj:
-
-2006-08-13  Maks Orlovich  <maksim@kde.org>
-
-        Reviewed (and tweaked a little) by Maciej.
-        
-        - shrank the size of JSObject by 8 bytes and made the corresponding reduction to the cell size, resulting
-        in a 1.2% speed improvement on JS iBench (and probably overall memory savings).
-
-        This was done by removing _scope and _internalValue data members
-        from JSObject and moving them only to the subclasses that actually
-        make use of them.
-        
-        * kjs/object.cpp: 
-        (KJS::JSObject::mark): No need to mark scope or internal value here.
-        * kjs/object.h:
-        (KJS::JSObject::JSObject): Don't initialize them.
-        * kjs/JSWrapperObject.cpp: Added. New base class for object types that
-        wrap primitive values (Number, String, Boolean, Date).
-        (KJS::JSWrapperObject::mark): 
-        * kjs/JSWrapperObject.h: Added.
-        (KJS::JSWrapperObject::JSWrapperObject):
-        (KJS::JSWrapperObject::internalValue):
-        (KJS::JSWrapperObject::setInternalValue):
-        * kjs/array_object.cpp:
-        (ArrayPrototype::ArrayPrototype): Don't set useless internal value.
-        * kjs/bool_object.cpp:
-        (BooleanInstance::BooleanInstance): Inherit from JSWrapperObject.
-        (BooleanProtoFunc::callAsFunction): Fixed to account for fact that not all
-        JSObjects have an internal value.
-        (BooleanObjectImp::construct): ditto.
-        * kjs/bool_object.h:
-        * kjs/collector.cpp: Lowered cell size to 48.
-        (KJS::Collector::allocate): meaningless whitespace change
-        * kjs/date_object.cpp:
-        (KJS::DateInstance::DateInstance): Inherit from JSWrapperObject.
-        (KJS::DateProtoFunc::callAsFunction): adjusted for move of internalValue
-        (KJS::DateObjectImp::construct): ditto
-        * kjs/date_object.h:
-        * kjs/error_object.cpp:
-        (ErrorPrototype::ErrorPrototype): don't set internal value
-        * kjs/function.cpp: move _scope and related handling here
-        (KJS::FunctionImp::mark): mark scope
-        * kjs/function.h:
-        (KJS::FunctionImp::scope): moved here from JSObject
-        (KJS::FunctionImp::setScope): ditto
-        * kjs/number_object.cpp:
-        (NumberInstance::NumberInstance): inherit from JSWrapperObject
-        (NumberProtoFunc::callAsFunction): adjusted
-        (NumberObjectImp::construct): adjusted
-        * kjs/number_object.h: shring RegExp-related objects a little
-        * kjs/regexp_object.cpp:
-        (RegExpPrototype::RegExpPrototype): Adjust for size tweaks
-        (RegExpObjectImp::RegExpObjectImp): ditto
-        * kjs/regexp_object.h:
-        * kjs/string_object.cpp:
-        (StringInstance::StringInstance): inherit from JSWrapperObject
-        (StringProtoFunc::callAsFunction): adjusted
-        * kjs/string_object.h:
-        * JavaScriptCore.exp: Exported new methods as needed.
-        * JavaScriptCore.xcodeproj/project.pbxproj: Added new files to build.
-
-2006-08-04  Brady Eidson  <beidson@apple.com>
-
-        Reviewed by Geoff's rubber stamp
-
-        Fix a build break on Intel hardware causes by adapting stricter
-        compiler warnings (-Wshorten-64-to-32)
-
-        * API/testapi.c:
-        (assertEqualsAsNumber): manually cast some doubles to floats
-        (main): ditto
-
-2006-08-04  Sam Weinig  <sam.weinig@gmail.com>
-
-        Reviewed by Darin.
-
-        - patch for http://bugs.webkit.org/show_bug.cgi?id=10192
-          Make WebCore (and friends) compile with -Wshorten-64-to-32
-
-          * Adds -Wshorten-64-to-32 flag to Xcode project.
-          * Adds explicit casts where OK.
-
-        * API/JSNodeList.c:
-        (JSNodeList_item):
-        (JSNodeList_getProperty):
-        * JavaScriptCore.xcodeproj/project.pbxproj:
-
-2006-08-04  Adam Roben  <aroben@apple.com>
-
-        Reviewed by Anders.
-
-        * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: Convert
-        spaces to tabs
-
-2006-08-03  Sam Weinig  <sam.weinig@gmail.com>
-
-        Reviewed by Darin.
-
-        - patch for http://bugs.webkit.org/show_bug.cgi?id=10176
-          Make WebCore compile with -Wundef
-
-          * Adds -Wundef flag to Xcode project
-          * Converts #ifs to #ifdef and #ifndefs where needed.
-          * Added #define YYMAXDEPTH 10000 in kjs/grammar.y
-            to fix a warning from within Bison.
-
-        * JavaScriptCore.xcodeproj/project.pbxproj:
-        * bindings/jni/jni_jsobject.cpp:
-        (JavaJSObject::getSlot):
-        (JavaJSObject::setSlot):
-        * bindings/npapi.h:
-        * bindings/objc/objc_class.mm:
-        (KJS::Bindings::ObjcClass::methodsNamed):
-        (KJS::Bindings::ObjcClass::fieldNamed):
-        * bindings/objc/objc_instance.mm:
-        (ObjcInstance::invokeMethod):
-        * bindings/objc/objc_runtime.mm:
-        (ObjcMethod::getMethodSignature):
-        (ObjcField::name):
-        (ObjcField::type):
-        * kjs/grammar.y:
-        * kjs/identifier.h:
-
-2006-08-03  Anders Carlsson  <acarlsson@apple.com>
-
-        Reviewed by John Sullivan.
-
-       * wtf/HashSet.h:
-        (WTF::::operator):
-        Return *this in operator=
-
-2006-08-03  Adam Roben  <aroben@apple.com>
-
-        Reviewed by Anders.
-
-        - Fixed Windows build
-
-        * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
-        * wtf/MathExtras.h: Implement inline versions of these functions
-        (nextafter):
-        (nextafterf):
-
-2006-08-02  Adam Roben  <aroben@apple.com>
-
-        Reviewed by Darin.
-
-        - Fixed build
-
-        * kjs/date_object.cpp:
-        (KJS::formatTime):
-
-2006-07-29  Darin Adler  <darin@apple.com>
-
-        - Removed tabs from these source files that still had them.
-          We don't use them; that way source files look fine in editors
-          that have tabs set to 8 spaces or to 4 spaces.
-        - Removed allow-tabs Subversion property from the files too.
-
-        * bindings/NP_jsobject.cpp:
-        * bindings/c/c_utility.cpp:
-        * bindings/jni/jni_runtime.cpp:
-        * bindings/jni/jni_utility.cpp:
-        * bindings/objc/objc_utility.mm:
-        * bindings/runtime.cpp:
-        * bindings/runtime_method.cpp:
-        * bindings/testbindings.cpp:
-        * bindings/testbindings.mm:
-        * kjs/date_object.cpp:
-        * kjs/function.cpp:
-        * kjs/list.cpp:
-        * kjs/nodes.cpp:
-        * kjs/nodes.h:
-        * kjs/string_object.cpp:
-        * kjs/ustring.cpp:
-
-2006-07-29  Darin Adler  <darin@apple.com>
-
-        * tests/mozilla/expected.html: Update test results now that regress-185165.js
-        is succeeding. I suspect Anders fix for bug 4620655 is the reason.
-
-2006-07-29  Sam Weinig  <sam.weinig@gmail.com>
-
-        Reviewed by Darin.
-
-        - patch for http://bugs.webkit.org/show_bug.cgi?id=10080
-          Adopt pedantic changes from the Unity project to improve 
-          cross-compiler compatibility
-
-          Changes include:
-          * Removing trailing semicolon from namespace braces.
-          * Removing trailing comma from last enum declaration.
-          * Updating to match style guidelines.
-          * Adding missing newline to the end of the file.
-          * Turning on gcc warning for missing newline at the end of a source file
-            (GCC_WARN_ABOUT_MISSING_NEWLINE in Xcode, -Wnewline in gcc).
-          * Alphabetical sorting of Xcode source list files. 
-          * Replace use of non-portable variable-size array with Vector.
-          * Use C-style comments instead of C++ comments in files that might
-            be included by either C or C++ files.
-
-        * API/JSCallbackConstructor.cpp:
-        (KJS::JSCallbackConstructor::construct):
-        * API/JSCallbackFunction.cpp:
-        (KJS::JSCallbackFunction::callAsFunction):
-        * API/JSCallbackObject.cpp:
-        (KJS::JSCallbackObject::construct):
-        (KJS::JSCallbackObject::callAsFunction):
-        * JavaScriptCore.xcodeproj/project.pbxproj:
-        * JavaScriptCorePrefix.h:
-        * bindings/jni/jni_class.cpp:
-        (JavaClass::fieldNamed):
-        * bindings/jni/jni_class.h:
-        * bindings/jni/jni_instance.cpp:
-        (JavaInstance::JavaInstance):
-        (JavaInstance::valueOf):
-        * bindings/jni/jni_objc.mm:
-        (KJS::Bindings::dispatchJNICall):
-        * bindings/jni/jni_runtime.cpp:
-        (JavaParameter::JavaParameter):
-        (JavaArray::JavaArray):
-        * bindings/jni/jni_runtime.h:
-        * bindings/jni/jni_utility.h:
-        * bindings/objc/objc_instance.h:
-        * bindings/runtime_array.h:
-        * kjs/collector.h:
-        * kjs/config.h:
-        * kjs/ustring.cpp:
-        * wtf/Platform.h:
-
-2006-07-29  Mike Emmel  <mike.emmel@gmail.com>
-
-        Reviewed by Darin.
-
-        - fixes for Linux build
-
-        * JavaScriptCoreSources.bkl: Added new files to build, kjs/PropertyNameArray.cpp
-        and kjs/testkjs.cpp, and removed old files.
-
-2006-07-24  Dan Waylonis  <waylonis@google.com>
-
-        Reviewed and tweaked a bit by Darin.
-
-        - fix http://bugs.webkit.org/show_bug.cgi?id=9902
-          jsNull and NSNull not properly converted between JS and ObjC
-
-        * bindings/objc/objc_utility.mm: (KJS::Bindings::convertObjcValueToValue):
-        Added case for converting NSNull to jsNull.
-
-2006-07-24  Rob Buis  <buis@kde.org>
-
-        Reviewed by Darin.
-
-        http://bugs.webkit.org/show_bug.cgi?id=4258
-        Date().toString() only includes GMT offset, not timezone string
-
-        Use the info in tm_zone to append timezone abbreviation
-        to Date().toString().
-
-        * kjs/date_object.cpp:
-        (KJS::formatTime):
-
-2006-07-24  Rob Buis  <buis@kde.org>
-
-        Reviewed by Darin.
-
-        http://bugs.webkit.org/show_bug.cgi?id=5257
-        setYear() does not match FireFox/IE behavior
-
-        Make sure the right values end up in tm_year.
-
-        * kjs/date_object.cpp:
-        (KJS::formatTime):
-
-2006-07-23  Mark Rowe  <opendarwin.org@bdash.net.nz>
-
-        Reviewed by Maciej.
-
-        Bug 9686: [Drosera] Need the ability to break into Drosera on Javascript exceptions
-        http://bugs.webkit.org/show_bug.cgi?id=9686
-
-        JavaScriptCore portion of the fix.
-
-        * JavaScriptCore.exp: Update symbol for change in argument type.
-        * kjs/debugger.cpp:
-        (Debugger::detach): Clear map of recent exceptions.
-        (Debugger::hasHandledException): Track the most recent exception
-        thrown by an interpreter.
-        (Debugger::exception): Change exception argument to a JSValue.
-        * kjs/debugger.h:
-        * kjs/nodes.cpp:
-        (Node::debugExceptionIfNeeded): Notify the debugger of an exception
-        if it hasn't seen it before.
-        (ThrowNode::execute): Notify the debugger that an exception is being thrown.
-        * kjs/nodes.h:
-
-    2006-07-23  Geoffrey Garen  <ggaren@apple.com>
-
-        Patch by Eric Albert, reviewed by Darin and me.
-        
-        - Fixed <rdar://problem/4645931> JavaScriptCore stack-scanning code 
-        crashes (Collector::markStackObjectsConservatively)
-        
-        * bindings/jni/jni_jsobject.cpp: On 64bit systems, jint is a long, not an
-        int.
-        (JavaJSObject::getSlot):
-        (JavaJSObject::setSlot):
-        * kjs/collector.cpp:
-        (KJS::Collector::markCurrentThreadConservatively): Use a pointer instead of
-        an int as 'dummy,' because on LP64 systems, an int is not pointer-aligned,
-        and we want to scan the stack for pointers.
-        * JavaScriptCore.xcodeproj/project.pbxproj: After a tense cease-fire, the 
-        XCode war has started up again!
-
-=== Safari-521.20 ===
-
-2006-07-21  Geoffrey Garen  <ggaren@apple.com>
-
-        Reviewed by Darin.
-
-        <rdar://problem/4507265> REGRESSION: overlays don't work on HousingMaps.com (Google Maps-based site)
-
-        - Added support for strings that masquerade as undefined. Currently used
-        by WebCore to implement undetectable style.filter.
-        
-        The name is a little long, but it's only used in one line of code, so I
-        thought clarity should win over brevity.
-
-        * JavaScriptCore.exp:
-        * JavaScriptCore.xcodeproj/project.pbxproj:
-        * kjs/object.h:
-        * kjs/string_object.h:
-        (KJS::StringInstanceThatMasqueradesAsUndefined::StringInstanceThatMasqueradesAsUndefined):
-        (KJS::StringInstanceThatMasqueradesAsUndefined::masqueradeAsUndefined):
-        (KJS::StringInstanceThatMasqueradesAsUndefined::toBoolean):
-
-=== Safari-521.19 ===
-
-2006-07-20  Steve Falkenburg  <sfalken@apple.com>
-
-        Fix the build
-
-        * kjs/function.cpp:
-        (KJS::escapeStringForPrettyPrinting):
-
-2006-07-19  Anders Carlsson  <acarlsson@apple.com>
-
-        Reviewed by Darin.
-
-        <rdar://problem/4620655> REGRESSION(10.4.7-10.5): preview button for a blogger.com post doesn't work
-        
-        * kjs/nodes2string.cpp:
-        (StringNode::streamTo):
-        Return the escaped string.
-        
-        (RegExpNode::streamTo):
-        Use the correct syntax.
-        
-        * kjs/function.cpp:
-        (KJS::escapeStringForPrettyPrinting):
-        * kjs/function.h:
-        Add escape function which escapes a string for pretty-printing so it can be parsed again.
-        
-        * wtf/unicode/icu/UnicodeIcu.h:
-        (WTF::Unicode::isPrintableChar):
-        New function.
-
-2006-07-18  Maciej Stachowiak  <mjs@apple.com>
-
-        Reviewed by Adele Peterson.
-        
-        <rdar://problem/4589530> REGRESSION: null character in JS string causes parse error (works in Tiger and in other browsers)
-
-        * kjs/lexer.cpp:
-        (Lexer::shift):
-        (Lexer::lex):
-        (Lexer::record16):
-        (Lexer::scanRegExp):
-        * kjs/lexer.h:
-
-2006-07-18  Tim Omernick  <timo@apple.com>
-
-        Reviewed by Tim Hatcher.
-
-        Removed a misleading comment; we recently added support for the NPNVPluginElementNPObject
-        variable.
-
-        * bindings/npapi.h:
-
-=== Safari-521.18 ===
-
-2006-07-18  Timothy Hatcher  <timothy@apple.com>
-
-        Made the following headers public:
-
-        * JavaScriptCore.h
-        * JSBase.h
-        * JSContextRef.h
-        * JSObjectRef.h
-        * JSStringRef.h
-        * JSValueRef.h
-
-        * JavaScriptCore.xcodeproj/project.pbxproj:
-
-2006-07-17  Geoffrey Garen  <ggaren@apple.com>
-
-        Reviewed by Maciej.
-        
-        - Added automatic prototype creation for classes.
-        
-        A class stores a weak reference to a prototype, which is cleared when
-        the prototype is garbage collected, to avoid a reference cycle.
-        
-        We now have an attributes field in JSClassDefinition, that currently is
-        used only to override automatic prototype creation when you want to manage your
-        own prototypes, but can be extended in the future for other nefarious purposes.
-        
-        Similarly, we have JSObjectMake and JSObjectMakeWithPrototype, the latter
-        allowing you to manage your own prototypes.
-        
-        JSObjectMakeConstructor is more interesting now, able to make a constructor
-        on your behalf if you just give it a class.
-        
-        - Removed bogus old code from minidom.js.
-        
-        - Tweaked the headerdocs.
-        
-        - Added more GC testing, which caught some leaks, and tested more funny 
-        edge cases in lookup, which caught a lookup bug. Removed some testing 
-        we used to do with MyObject because it was redundant with the new, cool 
-        stuff.
-        
-        While fixing the lookup bug I retracted this change:
-        
-            "If a static setProperty callback returns 'false', to indicate that the
-            property was not set, we no longer forward the set request up the class
-            chain, because that's almost certainly not what the programmer expected."
-
-        Returning false when setting a static property is a little silly, but you can see
-        it being useful when shadowing a base class's static properties, and, regardless
-        of usefullness, this is the defined behavior of the setProperty callback.
-        
-        - Plus a little ASCII art, for the kids.
-
-2006-07-17  Timothy Hatcher  <timothy@apple.com>
-
-        Reviewed by Maciej.
-
-        <rdar://problem/4634874> WebScriptObject and WebUndefined are no longer defined by WebKit
-
-        Moves WebScriptObject and WebUndefined up to WebCore.
-        This change does create an upwards-dependancy on WebScriptObject existing
-        in the loaded process, but this code path in JavaScriptCore does not get used
-        unless it is through WebKit/WebCore. Moving all of the binding code out of
-        JavaScriptCore might make sense in the future.
-
-        * JavaScriptCore.exp:
-        * JavaScriptCore.xcodeproj/project.pbxproj:
-        * bindings/objc/WebScriptObject.h: Replaced.
-        * bindings/objc/WebScriptObject.mm: Removed.
-        * bindings/objc/WebScriptObjectPrivate.h: Removed.
-        * bindings/objc/objc_class.h:
-        * bindings/objc/objc_instance.h:
-        * bindings/objc/objc_instance.mm:
-        (ObjcInstance::~ObjcInstance):
-        * bindings/objc/objc_runtime.h:
-        * bindings/objc/objc_runtime.mm:
-        (convertValueToObjcObject):
-        * bindings/objc/objc_utility.mm:
-        (KJS::Bindings::convertValueToObjcValue):
-        (KJS::Bindings::convertObjcValueToValue):
-        (KJS::Bindings::createObjcInstanceForValue):
-
-2006-07-17  Darin Adler  <darin@apple.com>
-
-        * API/JSBase.h: Fix comment formatting where things used to be lined up but
-        are now ragged. Got rid of spaces that attempted to line things up.
-        * API/JSObjectRef.h: Ditto. Also add missing periods for a couple of comments.
-
-2006-07-17  Geoffrey Garen  <ggaren@apple.com>
-
-        Reviewed by Maciej.
-        
-        - Removed the exception parameter from the initialize callback and, by extension,
-        JSObjectMake. We have never had a need for exceptions when iniitializing,
-        so the parameter seemed likely to "get in the way." 
-        
-        Also, an exception in JavaScript is thrown in response to input -- 
-        "invalid URL", "index not a number", etc., so it's the job of the 
-        constructor function, not the initialize method, to throw.
-        
-        If initialize *really* wants to throw, it can communicate the throw to
-        the constructor through the constructed object's private data (e.g., set
-        it to NULL, signaling to the consntructor that initialization failed.)
-        
-        - Added JSObjectMakeWithData, which enables a constructor to set private
-        data on an object *before* it has been initialized. That way, the initialize
-        methods can properly operate on the data. 
-
-        * API/JSNode.c: Moved ref into the initialize method, for better encapsulation,
-        now that it's possible.
-        * API/JSNodeList.c: ditto
-        * API/minidom.c:
-        (main): Do more aggressive garbage collection to test ref/deref and 
-        initialize/finalize.
-        * API/minidom.js: store childNodes in a temporary so it doesn't get re-created
-        like a thousand times. This makes debugging ref/deref easier
-
-2006-07-17  Geoffrey Garen  <ggaren@apple.com>
-
-        Reviewed by Maciej.
-        
-        - Changed the initialize callback to run from least derived class (parent
-        class) to most derived class. This enables C++ style initialization,
-        and derived class overriding of member data.
-        
-        - Added excpetion propopgation to JSObjectMake, to support initialize
-        exceptions, and generally round out our policy of making function
-        signatures as long as possible.
-
-        * API/JSCallbackObject.h: Use ExecState instead of ContextRef, cuz we're
-        in C++ land now.
-
-2006-07-17  Geoffrey Garen  <ggaren@apple.com>
-
-        Reviewed by Maciej.
-        
-        - Changed JSObjectMakeConstructor to JSObjectMakeConstructorWithCallback,
-        to match JSObjectMakeFunctionWithCallback.
-        
-        - Added prototype parameter, so the generated constructor
-        automatically works with hasInstance / instanceof
-        
-        - Moved hasInstance implementation from InternalFunctionImp to JSObject
-        so that subclasses can inherit it without inheriting function-related baggage.
-        More refactoring here would be good, but this seems like a good short-term
-        solution.
-
-        (KJS::JSCallbackFunction::implementsHasInstance): override and return false,
-        because callback functions aren't constructors.
-
-2006-07-17  Maciej Stachowiak  <mjs@apple.com>
-
-        Reviewed by Geoff.
-        
-        - add a JSContextRef parameter to all JSValueRef, JSObjectRef, and JSContextRef operations;
-        except JSObject{Get,Set}PrivateData which can be assumed to be simple pure accessors.
-        
-        Also renamed the parameter "context" to "ctx" because it makes the code read better with this pervasive
-        but usually uninteresting parameter.
-
-        * API/JSBase.cpp:
-        (JSEvaluateScript):
-        (JSCheckScriptSyntax):
-        (JSGarbageCollect):
-        * API/JSBase.h:
-        * API/JSCallbackObject.cpp:
-        (KJS::JSCallbackObject::JSCallbackObject):
-        (KJS::JSCallbackObject::init):
-        (KJS::JSCallbackObject::getOwnPropertySlot):
-        (KJS::JSCallbackObject::put):
-        (KJS::JSCallbackObject::deleteProperty):
-        (KJS::JSCallbackObject::toNumber):
-        (KJS::JSCallbackObject::toString):
-        * API/JSContextRef.cpp:
-        (JSGlobalContextCreate):
-        (JSGlobalContextRetain):
-        (JSGlobalContextRelease):
-        (JSContextGetGlobalObject):
-        * API/JSContextRef.h:
-        * API/JSNode.c:
-        (JSNodePrototype_appendChild):
-        (JSNodePrototype_removeChild):
-        (JSNodePrototype_replaceChild):
-        (JSNode_getNodeType):
-        (JSNode_getFirstChild):
-        (JSNode_prototype):
-        * API/JSNodeList.c:
-        (JSNodeListPrototype_item):
-        (JSNodeList_length):
-        (JSNodeList_getProperty):
-        (JSNodeList_prototype):
-        * API/JSObjectRef.cpp:
-        (JSObjectMake):
-        (JSObjectMakeFunctionWithCallback):
-        (JSObjectMakeConstructor):
-        (JSObjectMakeFunction):
-        (JSObjectGetPrototype):
-        (JSObjectSetPrototype):
-        (JSObjectHasProperty):
-        (JSObjectGetProperty):
-        (JSObjectSetProperty):
-        (JSObjectGetPropertyAtIndex):
-        (JSObjectSetPropertyAtIndex):
-        (JSObjectDeleteProperty):
-        (JSObjectIsFunction):
-        (JSObjectCallAsFunction):
-        (JSObjectIsConstructor):
-        (JSObjectCallAsConstructor):
-        (JSObjectCopyPropertyNames):
-        * API/JSObjectRef.h:
-        * API/JSStringRef.cpp:
-        * API/JSValueRef.cpp:
-        (JSValueGetType):
-        (JSValueIsUndefined):
-        (JSValueIsNull):
-        (JSValueIsBoolean):
-        (JSValueIsNumber):
-        (JSValueIsString):
-        (JSValueIsObject):
-        (JSValueIsObjectOfClass):
-        (JSValueIsEqual):
-        (JSValueIsStrictEqual):
-        (JSValueIsInstanceOfConstructor):
-        (JSValueMakeUndefined):
-        (JSValueMakeNull):
-        (JSValueMakeBoolean):
-        (JSValueMakeNumber):
-        (JSValueMakeString):
-        (JSValueToBoolean):
-        (JSValueToNumber):
-        (JSValueToStringCopy):
-        (JSValueToObject):
-        (JSValueProtect):
-        (JSValueUnprotect):
-        * API/JSValueRef.h:
-        * API/minidom.c:
-        (print):
-        * API/testapi.c:
-        (MyObject_getProperty):
-        (MyObject_deleteProperty):
-        (MyObject_callAsFunction):
-        (MyObject_callAsConstructor):
-        (MyObject_convertToType):
-        (print_callAsFunction):
-        (main):
-
-2006-07-16  Geoffrey Garen  <ggaren@apple.com>
-
-        Approved by Maciej, RS by Beth.
-        
-        JSObjectMakeFunction -> JSObjectMakeFunctionWithCallback
-        JSObjectMakeFunctionWithBody -> JSObjectMakeFunction
-        
-        because the latter is more common, and more fundamental, than the former.
-
-        * API/APICast.h:
-        (toJS):
-        * API/JSBase.h:
-        * API/JSCallbackObject.cpp:
-        (KJS::JSCallbackObject::getOwnPropertySlot):
-        (KJS::JSCallbackObject::put):
-        (KJS::JSCallbackObject::deleteProperty):
-        (KJS::JSCallbackObject::getPropertyNames):
-        (KJS::JSCallbackObject::staticValueGetter):
-        (KJS::JSCallbackObject::staticFunctionGetter):
-        * API/JSClassRef.cpp:
-        (OpaqueJSClass::OpaqueJSClass):
-        (OpaqueJSClass::~OpaqueJSClass):
-        * API/JSClassRef.h:
-        * API/JSObjectRef.cpp:
-        (JSClassCreate):
-        (JSObjectMakeFunctionWithCallback):
-        (JSObjectMakeFunction):
-        (OpaqueJSPropertyNameArray::OpaqueJSPropertyNameArray):
-        (JSObjectCopyPropertyNames):
-        * API/JSObjectRef.h:
-        * API/minidom.c:
-        (main):
-        * API/testapi.c:
-        (main):
-        * ChangeLog:
-        * JavaScriptCore.exp:
-
-2006-07-16  Geoffrey Garen  <ggaren@apple.com>
-
-        Laughed at by Beth.
-        
-        Replace __JS with OpaqueJS because the former, while used by CF, is
-        a prefix that's triply-reserved by the compiler. (_* is reserved in global 
-        names, _[A-Z] is reserved in all names, and __ is reserved in all names 
-        in C++.)
-        
-        Opaque is an alternative used by other Mac OS X framewokrs.
-
-        * API/APICast.h:
-        (toJS):
-        * API/JSBase.h:
-        * API/JSCallbackObject.cpp:
-        (KJS::JSCallbackObject::getOwnPropertySlot):
-        (KJS::JSCallbackObject::put):
-        (KJS::JSCallbackObject::deleteProperty):
-        (KJS::JSCallbackObject::getPropertyNames):
-        (KJS::JSCallbackObject::staticValueGetter):
-        (KJS::JSCallbackObject::staticFunctionGetter):
-        * API/JSClassRef.cpp:
-        (OpaqueJSClass::OpaqueJSClass):
-        (OpaqueJSClass::~OpaqueJSClass):
-        * API/JSClassRef.h:
-        * API/JSObjectRef.cpp:
-        (JSClassCreate):
-        (OpaqueJSPropertyNameArray::OpaqueJSPropertyNameArray):
-        (JSObjectCopyPropertyNames):
-
-2006-07-16  Darin Adler  <darin@apple.com>
-
-        - try to fix Windows build
-
-        * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
-        Added some recently added files, removed some recently removed.
-
-2006-07-16  Geoffrey Garen  <ggaren@apple.com>
-
-        Reviewed by Maciej.
-        
-        - Change getProperty* to return undefined, rather than NULL, for missing
-        properties, since that's what the spec says. Also added exception out
-        parameters to the *Index functions, because they can call through to the
-        regular functions, which can throw for custom objects.
-
-        * API/JSObjectRef.cpp:
-        (JSObjectGetProperty):
-        (JSObjectGetPropertyAtIndex):
-        (JSObjectSetPropertyAtIndex):
-        * API/JSObjectRef.h:
-        * API/testapi.c:
-        (main):
-
-2006-07-16  Geoffrey Garen  <ggaren@apple.com>
-
-        Reviewed by Maciej.
-        
-        - Properly document and handle NULL callbacks for static properties. We
-        throw an exception in any case other than a ReadOnly property with a NULL
-        setProperty callback, because a NULL callback almost certainly indicates 
-        a programming error. Also throw an exception if hasProperty returns true
-        for a property that getProperty can't get.
-        
-        - If a static setProperty callback returns 'false', to indicate that the
-        property was not set, we no longer forward the set request up the class
-        chain, because that's almost certainly not what the programmer expected.
-
-        * API/JSCallbackObject.cpp:
-        (KJS::JSCallbackObject::getOwnPropertySlot):
-        (KJS::JSCallbackObject::put):
-        (KJS::JSCallbackObject::staticValueGetter):
-        (KJS::JSCallbackObject::staticFunctionGetter):
-        (KJS::JSCallbackObject::callbackGetter):
-        * API/JSObjectRef.h:
-        * API/minidom.js:
-        * API/testapi.c:
-        (MyObject_hasProperty):
-        * API/testapi.js:
-
-2006-07-16  Geoffrey Garen  <ggaren@apple.com>
-
-        Reviewed by Maciej.
-        
-        - Added names to functions.
-        
-        - Removed GetPrivate/SetPrivate from callbackFunctions and callbackConstructors.
-        The private data idiom is that a JS object stores its native implementation
-        as private data. For functions and constructors, the native implementation is nothing
-        more than the callback they already store, so supporting private data, too,
-        confuses the idiom. If you *really* want, you can still create a custom 
-        function with private data.
-
-        * API/JSCallbackConstructor.cpp:
-        * API/JSCallbackConstructor.h:
-        * API/JSCallbackFunction.cpp:
-        (KJS::JSCallbackFunction::JSCallbackFunction):
-        * API/JSCallbackFunction.h:
-        * API/JSCallbackObject.cpp:
-        (KJS::JSCallbackObject::staticFunctionGetter):
-        * API/JSObjectRef.cpp:
-        (JSObjectMakeFunction):
-        (JSObjectMakeFunctionWithBody):
-        (JSObjectGetPrivate):
-        (JSObjectSetPrivate):
-        * API/JSObjectRef.h:
-        * API/minidom.c:
-        (main):
-        * API/testapi.c:
-        (main):
-
-2006-07-15  Maciej Stachowiak  <mjs@apple.com>
-
-        Reviewed by Darin.
-        
-        - switch property lists to be vector+set of Identifiers instead of list of References
-        
-        This has the following benefits:
-        
-        - no duplicates in property lists
-        - simplifies API calls
-        - probably more efficient, since linked list is gone
-        - entirely removed Reference, ReferenceList and ProtectedReference types from the API
-
-        * kjs/PropertyNameArray.cpp: Added.
-        (KJS::PropertyNameArray::add): Check set, if not already there, add to
-        vector.
-        * kjs/PropertyNameArray.h: Added.
-        (KJS::PropertyNameArray::PropertyNameArray): Newly added type, combines
-        a set and a vector to make a unique but ordered list of identifiers.
-        (KJS::PropertyNameArray::begin): ditto
-        (KJS::PropertyNameArray::end): ditto
-        (KJS::PropertyNameArray::size): ditto
-        (KJS::PropertyNameArray::operator[]): ditto
-        * kjs/array_instance.h:
-        * kjs/array_object.cpp:
-        (ArrayInstance::getPropertyNames): renamed from getPropertyList, updated
-        for PropertyNameArray
-        (ArrayInstance::setLength): updated for PropertyNameArray
-        (ArrayInstance::pushUndefinedObjectsToEnd): ditto
-        * kjs/nodes.cpp:
-        (ForInNode::execute): updated for PropertyNameArray
-        * kjs/nodes.h:
-        * kjs/object.cpp:
-        (KJS::JSObject::getPropertyNames): renamed from getPropertyList, updated
-        for PropertyNameArray
-        * kjs/object.h:
-        * kjs/property_map.cpp:
-        (KJS::PropertyMap::getEnumerablePropertyNames): updated for PropertyNameArray
-        (KJS::PropertyMap::getSparseArrayPropertyNames): ditto
-        * kjs/property_map.h:
-        * kjs/protected_reference.h: Removed.
-        * kjs/reference.cpp: Removed.
-        * kjs/reference.h: Removed.
-        * kjs/reference_list.cpp: Removed.
-        * kjs/reference_list.h: Removed.
-        * kjs/scope_chain.cpp:
-        (KJS::ScopeChain::print): Use PropertyNamesArray instead of ReferenceList.
-        * kjs/string_object.cpp:
-        (StringInstance::getPropertyNames): Updated for new approach.
-        * kjs/string_object.h:
-        * kjs/ustring.h:
-        * API/APICast.h:
-        (toJS): Added overload for PropertyNameAccumulatorRef / PropertyNameArray*
-        (toRef): ditto
-        * API/JSBase.h:
-        * API/JSCallbackObject.cpp:
-        (KJS::JSCallbackObject::getPropertyNames): Fixed for new API.
-        * API/JSCallbackObject.h:
-        * API/JSObjectRef.cpp:
-        (__JSPropertyNameArray::__JSPropertyNameArray): Type used for a publicly vended
-        JSPropertyNameArrayRef.
-        (JSObjectCopyPropertyNames): New API call - renamed / refactored from 
-        JSObjectCreatePropertyList
-        (JSPropertyNameArrayRetain): new retain call for JSPropertyNameArray.
-        (JSPropertyNameArrayRelease): new release call for - " -.
-        (JSPropertyNameArrayGetCount): Instead of having to use a stateful enumerator you
-        can now get the count and items in any order.
-        (JSPropertyNameArrayGetNameAtIndex): See above.
-        (JSPropertyNameAccumulatorAddName): What you add properties to is now an opaque 
-        accumulator object.
-        * API/JSObjectRef.h: Prototyped new functions, removed old ones
-        * JavaScriptCore.exp: Updated exported symbols.
-        * JavaScriptCore.xcodeproj/project.pbxproj: Added new files, removed old.
-        * API/testapi.c:
-        (MyObject_getPropertyNames): Renamed / fixed callback to fit new paradigm.
-        (main): Updated for new API.
-
-2006-07-15  Darin Adler  <darin@apple.com>
-
-        - oops, missed a few more arrays that had to be const
-
-        * API/JSNode.c:
-        (JSNodePrototype_appendChild): Added const.
-        (JSNodePrototype_removeChild): Ditto.
-        (JSNodePrototype_replaceChild): Ditto.
-        (JSNode_construct): Ditto.
-        * API/JSNodeList.c:
-        (JSNodeListPrototype_item): Ditto.
-        * API/JSObjectRef.cpp:
-        (JSObjectMakeFunctionWithBody): Ditto.
-        (JSObjectCallAsFunction): Ditto.
-        (JSObjectCallAsConstructor): Ditto.
-        * API/minidom.c:
-        (print): Ditto.
-        * API/testapi.c:
-        (MyObject_callAsFunction): Ditto.
-        (MyObject_callAsConstructor): Ditto.
-        (print_callAsFunction): Ditto.
-        (myConstructor_callAsConstructor): Ditto.
-
-2006-07-15  Darin Adler  <darin@apple.com>
-
-        Reviewed by Maciej.
-
-        * API/JSNode.h: Made an array parameter const.
-        * API/JSObjectRef.h: Made array parameters const. Fixed a comment.
-
-2006-07-15  Geoffrey Garen  <ggaren@apple.com>
-
-        Reviewed by Maciej.
-        
-        - JSObjectMakeFunctionWithBody includes a function name and named parameters now.
-
-        * API/JSObjectRef.cpp:
-        (JSObjectMakeFunctionWithBody):
-        * API/JSObjectRef.h:
-        * API/testapi.c:
-        (assertEqualsAsUTF8String): More informative failure reporting.
-        (main): Test more function cases.
-
-2006-07-15  Geoffrey Garen  <ggaren@apple.com>
-
-        Reviewed by Maciej.
-
-        - Moved the arguments passed to JSClassCreate into a single structure,
-        called JSClassDefinition. This will enable easier structure 
-        migration/versioning in the future, if necessary.
-        
-        - Added support for class names.
-        
-        - kJSClassDefinitionNull replaces kJSObjectCallbacksNone.
-        
-        - JSClass is becoming a fairly complex struct, so I migrated all of its
-        implementation other than reference counting to the sruct.
-        
-        - Also moved JSClass* functions in the API to JSObjectRef.cpp, since they're
-        declared in JSObjectRef.h
-        
-        - Also added some more informative explanation to the class structure doc.
-        
-2006-07-15  Darin Adler  <darin@apple.com>
-
-        Reviewed by Geoff.
-
-        - fix http://bugs.webkit.org/show_bug.cgi?id=8395
-          <rdar://problem/4613467>
-          REGRESSION: RegEx seems broken for hex escaped non breaking space
-
-        Test: fast/js/regexp-extended-characters-more.html
-
-        * pcre/pcre_exec.c:
-        (match): Got rid of utf16Length local variable to guarantee there's no
-        extra stack usage in recursive calls. Fixed two places in the PCRE_UTF16
-        code that were using the length variable, which is the UTF-8 length of
-        a character in the pattern, to move in the UTF-16 subject string. Instead
-        they hardcode lengths of 1 and 2 since the code already handles BMP
-        characters and surrogate pairs separately. Also fixed some DPRINTF so
-        I could compile with DEBUG on.
-        (pcre_exec): Changed a place that was checking for multibyte characters
-        in the subject string to use ISMIDCHAR. Instead it was using hardcoded
-        logic that was right for UTF-8 but wrong for UTF-16.
-
-        * pcre/pcre_compile.c: (pcre_compile2): Fixed a DPRINTF so I could compile
-        with DEBUG on.
-
-2006-07-14  Geoffrey Garen  <ggaren@apple.com>
-
-        RS by Maciej.
-        
-        Global replace in the API of argc/argv with argumentCount/arguments.
-
-2006-07-14  Geoffrey Garen  <ggaren@apple.com>
-
-        Reviewed by Maciej.
-        
-        - Finalized exception handling in the API.
-        
-        setProperty can throw because it throws for built-in arrays. getProperty
-        and deleteProperty can throw because setProperty can throw and we want
-        to be consistent, and also because they seem like "actions." callAsFunction,
-        callAsConstructor, and hasInstance can throw, because they caan throw for
-        all built-ins.
-        
-        toBoolean can't throw because it's defined that way in the spec.
-        
-        - Documented that toBoolean and toObject can't be overridden by custom 
-        objects because they're defined that way in the spec.
-
-=== Safari-521.17 ===
-
-2006-07-14  Geoffrey Garen  <ggaren@apple.com>
-
-        Reviewed by Maciej.
-        
-        - Implemented ref-counting of JSContexts by splitting into two datatypes:
-        JSGlobalContext, which you can create/retain/release, and JSContext, which
-        you can't.
-        
-        Internally, you retain a JSGlobalContext/ExecState by retaining its
-        interpreter, which, in the case of a global ExecState, owns it.
-        
-        - Also made ~Interpreter() protected to catch places where Interpreter
-        is manually deleted. (Can't make it private because some crazy fool
-        decided it would be a good idea to subclass Interpreter in other frameworks.
-        I pity da fool.)
-
-        * API/APICast.h: 
-        (toJS): Added cast for new JSGlobalContext
-        * API/JSStringRef.h: Changed vague "you must" language to more specific
-        (but, ultimately, equally vague) "behavior is undefined if you don't" 
-        language.
-        (KJS::Interpreter::Interpreter): Factored more common initialization into
-        init()
-        * kjs/interpreter.h:
-        (KJS::Interpreter::ref): new
-        (KJS::Interpreter::deref): new
-        (KJS::Interpreter::refCount): new
-        * kjs/testkjs.cpp: 
-        (doIt): Ref-count the interpreter.
-
-2006-07-14  Maciej Stachowiak  <mjs@apple.com>
-
-        Reviewed by Geoff.
-        
-        - removed bool return value from JSObjectSetProperty, since it is inefficient and
-        also doesn't work quite right
-        - added JSObjectGetPropertyAtIndex and JSObjectSetPropertyAtIndex
-
-        * API/JSObjectRef.cpp:
-        (JSObjectSetProperty): Removed return value and canPut stuff.
-        (JSObjectGetPropertyAtIndex): Added.
-        (JSObjectSetPropertyAtIndex): Added.
-        * API/JSObjectRef.h: Prototyped and documented new functions.
-
-2006-07-14  Geoffrey Garen  <ggaren@apple.com>
-
-        Reviewed by Beth.
-        
-        Moved JSCheckScriptSyntax, JSEvaluateScript, and JSGarbageCollect into
-        JSBase.h/.cpp. They don't belong in the value-specific or context-specific 
-        files because they're not part of the value or context implementations.
-
-        * API/JSBase.h:
-        * API/JSContextRef.cpp:
-        (JSContextGetGlobalObject):
-        * API/JSContextRef.h:
-        * API/JSValueRef.cpp:
-        (JSValueUnprotect):
-        * API/JSValueRef.h:
-        * JavaScriptCore.xcodeproj/project.pbxproj:
-
-2006-07-13  Timothy Hatcher  <timothy@apple.com>
-
-        Reviewed by Maciej.
-
-        Moved JavaScriptCore to be a public framework.
-
-        * JavaScriptCore.xcodeproj/project.pbxproj:
-
-2006-07-13  Mark Rowe  <opendarwin.org@bdash.net.nz>
-
-        Reviewed by Geoffrey.
-
-        http://bugs.webkit.org/show_bug.cgi?id=9742
-        Bug 9742: REGRESSION: WebKit hangs when loading <http://www.vtbook.com>
-
-        * kjs/value.h:
-        (KJS::JSValue::getUInt32): Only types tagged as numeric can be converted to UInt32.
-
-2006-07-13  Geoffrey Garen  <ggaren@apple.com>
-
-        Pleasing to Maciej.
-        
-        - Renamed JSEvaluate -> JSEvaluateScript, JSCheckSyntax -> JSCheckScriptSyntax
-        - Added exception out parameters to JSValueTo* and JSValueIsEqual because
-        they can throw
-        - Removed JSObjectGetDescription because it's useless and vague, and 
-        JSValueToString/JSValueIsObjectOfClass do a better job, anyway
-        - Clarified comments about "IsFunction/Constructor" to indicate that they
-        are true of all functions/constructors, not just those created by JSObjectMake*
-
-2006-07-12  Geoffrey Garen  <ggaren@apple.com>
-
-        RS by Beth.
-        
-        Finished previously approved JSInternalString -> JSString conversion 
-        by renaming the files.
-
-        * API/JSCallbackObject.cpp:
-        * API/JSInternalStringRef.cpp: Removed.
-        * API/JSInternalStringRef.h: Removed.
-        * API/JSStringRef.cpp: Added.
-        * API/JSStringRef.h: Added.
-        * API/JavaScriptCore.h:
-        * JavaScriptCore.xcodeproj/project.pbxproj:
-
-2006-07-12  Geoffrey Garen  <ggaren@apple.com>
-
-        Reviewed by Maciej.
-        
-        - Removed context and exception parameters from JSObjectGetPropertyEnumerator,
-        removing the spurious use of ExecState inside JavaScriptCore that made
-        us think this was necessary in the first place.
-
-        (StringInstance::getPropertyList): Use getString instead of toString because
-        we know we're dealing with a string -- we put it there in the first place.
-        While we're at it, store the string's size instead of retrieving it each time 
-        through the loop, to avoid the unnecessary killing of puppies.
-        * kjs/string_object.h:
-
-2006-07-12  Maciej Stachowiak  <mjs@apple.com>
-
-        Reviewed by Geoff.
-        
-        - add handling of hasInstance callback for API objects
-
-        * API/JSCallbackObject.cpp:
-        (KJS::JSCallbackObject::implementsHasInstance): Check if callback is present.
-        (KJS::JSCallbackObject::hasInstance): Invoke appropriate callback.
-        * API/JSCallbackObject.h:
-        * API/JSClassRef.cpp:
-        * API/JSObjectRef.h:
-        * API/testapi.c:
-        (MyObject_hasInstance): Test case; should match what construct would do.
-        * API/testapi.js:
-
-2006-07-11  Geoffrey Garen  <ggaren@apple.com>
-
-        Reviewed by Maciej.
-        
-        - Implemented a vast number of renames and comment clarifications 
-        suggested during API review.
-        
-        JSInternalString -> JSString
-        JS*Make -> JSValueMake*, JSObjectMake*
-        JSTypeCode -> JSType
-        JSValueIsInstanceOf -> JSValueIsInstanceOfConstructor (reads strangely well in client code)
-        JSGC*Protect -> JSValue*Protect
-        JS*Callback -> JSObject*Callback
-        JSGetPropertyListCallback -> JSObjectAddPropertiesToListCallback
-        JSPropertyEnumeratorGetNext -> JSPropertyEnumeratorGetNextName
-        JSString* -> 
-            JSStringCreateWithUTF8CString, JSStringGetUTF8CString,
-            JSStringGetMaximumUTF8CStringSize JSStringIsEqualToUTF8CString, 
-            JSStringCreateWithCFString, JSStringCopyCFString, JSStringCreateWithCharacters.
-        
-        - Changed functions taking a JSValue out arg and returning a bool indicating
-        whether it was set to simply return a JSValue or NULL.
-        
-        - Removed JSStringGetCharacters because it's more documentation than code,
-        and it's just a glorified memcpy built on existing API functionality.
-        
-        - Moved standard library includes into the headers that actually require them.
-        
-        - Standardized use of the phrase "Create Rule."
-        
-        - Removed JSLock from make functions that don't allocate.
-        
-        - Added exception handling to JSValueToBoolean, since we now allow
-        callback objects to throw exceptions upon converting to boolean.
-        
-        - Renamed JSGCCollect to JSGarbageCollect.
-
-2006-07-10  Geoffrey Garen  <ggaren@apple.com>
-
-        Reviewed by Darin.
-        
-        - Changed public header includes to the <JavaScriptCore/ style.
-        - Changed instances of 'buffer' to 'string' since we decided on
-        JSInternalString instead of JSStringBuffer.
-
-        * API/JSContextRef.h:
-        * API/JSInternalStringRef.cpp:
-        (JSStringMake):
-        (JSInternalStringRetain):
-        (JSInternalStringRelease):
-        (JSValueCopyStringValue):
-        (JSInternalStringGetLength):
-        (JSInternalStringGetCharactersPtr):
-        (JSInternalStringGetCharacters):
-        (JSInternalStringGetMaxLengthUTF8):
-        (JSInternalStringGetCharactersUTF8):
-        (CFStringCreateWithJSInternalString):
-        * API/JSInternalStringRef.h:
-        * API/JSNode.c:
-        (JSNodePrototype_appendChild):
-        (JSNode_getNodeType):
-        * API/JSObjectRef.cpp:
-        (JSObjectCallAsConstructor):
-        * API/JSValueRef.h:
-        * API/JavaScriptCore.h:
-        * API/minidom.c:
-        (main):
-        (print):
-        * API/testapi.c:
-        (MyObject_getPropertyList):
-        (myConstructor_callAsConstructor):
-        (main): I noticed that we were prematurely releasing some string buffers,
-        so I moved their release calls to the end of main(). I got rid of 'Buf' in *Buf
-        (sometimes changing to 'IString', when necessary to differentiate a variable) 
-        to match the buffer->string change.
-
-=== Safari-521.16 ===
-
-2006-07-10  Darin Adler  <darin@apple.com>
-
-        * kjs/value.cpp: (KJS::JSValue::toInt32Inline): Added inline keyword one more place.
-        Just in case.
-
-2006-07-10  Darin Adler  <darin@apple.com>
-
-        - fix the release build
-
-        * kjs/value.h:
-        * kjs/value.cpp:
-        (KJS::JSValue::toInt32Inline): Move the code here to an inline.
-        (KJS::JSValue::toInt32): Call the inline from both overloaded toInt32 functions.
-
-2006-07-10  David Kilzer  <ddkilzer@kilzer.net>
-
-        Reviewed by Darin.
-
-        - fix http://bugs.webkit.org/show_bug.cgi?id=9179
-          Implement select.options.add() method
-
-        * JavaScriptCore.exp: Added overloaded KJS::JSValue::toInt32() method.
-        * JavaScriptCore.xcodeproj/project.pbxproj: Altered attributes metadata for
-        kjs/value.h to make it available as a forwarded header.
-        * kjs/lookup.h:
-        (KJS::lookupPut): Extracted a lookupPut() method from the existing lookupPut() method.
-        The new method returns a boolean value if no entry is found in the lookup table.
-        * kjs/value.cpp:
-        (KJS::JSValue::toInt32): Overloaded toInt32() method with boolean "Ok" argument.
-        * kjs/value.h: Ditto.
-
-2006-07-10  Geoffrey Garen  <ggaren@apple.com>
-
-        No review necessary. Removed bogus file I accidentally checked in before.
-
-        * API/JSInternalSringRef.h: Removed.
-
-2006-07-10  Geoffrey Garen  <ggaren@apple.com>
-
-        Reviewed by Darin.
-        
-        Added exception out parameter to API object callbacks, removed semi-bogus
-        JSContext(.*)Exception functions.
-        
-        To make these calls syntactically simple, I added an exceptionSlot()
-        method to the ExecState class, which provides a JSValue** slot in which to 
-        store a JSValue* exception.
-
-        * API/APICast.h:
-        (toRef):
-        * API/JSCallbackConstructor.cpp:
-        (KJS::JSCallbackConstructor::construct):
-        * API/JSCallbackFunction.cpp:
-        (KJS::JSCallbackFunction::callAsFunction):
-        * API/JSCallbackObject.cpp:
-        (KJS::JSCallbackObject::init):
-        (KJS::JSCallbackObject::getOwnPropertySlot):
-        (KJS::JSCallbackObject::put):
-        (KJS::JSCallbackObject::deleteProperty):
-        (KJS::JSCallbackObject::construct):
-        (KJS::JSCallbackObject::callAsFunction):
-        (KJS::JSCallbackObject::getPropertyList):
-        (KJS::JSCallbackObject::toBoolean):
-        (KJS::JSCallbackObject::toNumber):
-        (KJS::JSCallbackObject::toString):
-        (KJS::JSCallbackObject::staticValueGetter):
-        (KJS::JSCallbackObject::callbackGetter):
-        * API/JSContextRef.cpp:
-        (JSCheckSyntax):
-        * API/JSContextRef.h:
-        * API/JSNode.c:
-        (JSNodePrototype_appendChild):
-        (JSNodePrototype_removeChild):
-        (JSNodePrototype_replaceChild):
-        (JSNode_getNodeType):
-        (JSNode_getChildNodes):
-        (JSNode_getFirstChild):
-        (JSNode_construct):
-        * API/JSNode.h:
-        * API/JSNodeList.c:
-        (JSNodeListPrototype_item):
-        (JSNodeList_length):
-        (JSNodeList_getProperty):
-        * API/JSObjectRef.h:
-        * API/minidom.c:
-        (print):
-        * API/testapi.c:
-        (MyObject_initialize):
-        (MyObject_hasProperty):
-        (MyObject_getProperty):
-        (MyObject_setProperty):
-        (MyObject_deleteProperty):
-        (MyObject_getPropertyList):
-        (MyObject_callAsFunction):
-        (MyObject_callAsConstructor):
-        (MyObject_convertToType):
-        (print_callAsFunction):
-        (myConstructor_callAsConstructor):
-        (main):
-        * JavaScriptCore.exp:
-        * kjs/ExecState.h:
-        (KJS::ExecState::exceptionHandle):
-
-2006-07-10  Geoffrey Garen  <ggaren@apple.com>
-
-        Reviewed by Darin.
-
-        Improved type safety by implementing opaque JSValue/JSObject typing through 
-        abuse of 'const', not void*. Also fixed an alarming number of bugs
-        exposed by this new type safety.
-        
-        I made one design change in JavaScriptCore, which is that the JSObject
-        constructor should take a JSValue* as its prototype argument, not a JSObject*,
-        since we allow the prototype to be any JSValue*, including jsNull(), for
-        example.
-        
-        * API/APICast.h:
-        (toJS):
-        * API/JSBase.h:
-        * API/JSCallbackConstructor.cpp:
-        (KJS::JSCallbackConstructor::construct):
-        * API/JSCallbackFunction.cpp:
-        (KJS::JSCallbackFunction::callAsFunction):
-        * API/JSCallbackObject.cpp:
-        (KJS::JSCallbackObject::JSCallbackObject):
-        (KJS::JSCallbackObject::getOwnPropertySlot):
-        (KJS::JSCallbackObject::put):
-        (KJS::JSCallbackObject::construct):
-        (KJS::JSCallbackObject::callAsFunction):
-        (KJS::JSCallbackObject::staticFunctionGetter):
-        * API/JSCallbackObject.h:
-        * API/JSContextRef.cpp:
-        (JSEvaluate):
-        * API/JSNode.c:
-        (JSNodePrototype_appendChild):
-        (JSNodePrototype_removeChild):
-        (JSNodePrototype_replaceChild):
-        * API/JSObjectRef.cpp:
-        (JSObjectMake):
-        (JSFunctionMakeWithBody):
-        (JSObjectGetProperty):
-        (JSObjectCallAsFunction):
-        (JSObjectCallAsConstructor):
-        * API/JSObjectRef.h:
-        * API/testapi.c:
-        (main):
-        * ChangeLog:
-        * kjs/object.h:
-        (KJS::JSObject::JSObject):
-
-2006-07-10  Geoffrey Garen  <ggaren@apple.com>
-
-        Approved by Maciej, Darin.
-        
-        Renamed JSStringBufferRef to JSInternalStringRef. "Internal string" means the 
-        JavaScript engine's internal string representation, which is the most 
-        low-level and efficient representation to use when interfacing with JavaScript.
-
-        * API/APICast.h:
-        (toJS):
-        (toRef):
-        * API/JSBase.h:
-        * API/JSCallbackObject.cpp:
-        (KJS::JSCallbackObject::getOwnPropertySlot):
-        (KJS::JSCallbackObject::put):
-        (KJS::JSCallbackObject::deleteProperty):
-        (KJS::JSCallbackObject::staticValueGetter):
-        (KJS::JSCallbackObject::callbackGetter):
-        * API/JSContextRef.cpp:
-        (JSEvaluate):
-        (JSCheckSyntax):
-        * API/JSContextRef.h:
-        * API/JSInternalStringRef.cpp: Added.
-        (JSStringMake):
-        (JSInternalStringCreate):
-        (JSInternalStringCreateUTF8):
-        (JSInternalStringRetain):
-        (JSInternalStringRelease):
-        (JSValueCopyStringValue):
-        (JSInternalStringGetLength):
-        (JSInternalStringGetCharactersPtr):
-        (JSInternalStringGetCharacters):
-        (JSInternalStringGetMaxLengthUTF8):
-        (JSInternalStringGetCharactersUTF8):
-        (JSInternalStringIsEqual):
-        (JSInternalStringIsEqualUTF8):
-        (JSInternalStringCreateCF):
-        (CFStringCreateWithJSInternalString):
-        * API/JSInternalStringRef.h: Added.
-        * API/JSNode.c:
-        (JSNodePrototype_appendChild):
-        (JSNode_getNodeType):
-        (JSNode_getChildNodes):
-        (JSNode_getFirstChild):
-        * API/JSNodeList.c:
-        (JSNodeList_length):
-        (JSNodeList_getProperty):
-        * API/JSObjectRef.cpp:
-        (JSFunctionMakeWithBody):
-        (JSObjectGetDescription):
-        (JSObjectHasProperty):
-        (JSObjectGetProperty):
-        (JSObjectSetProperty):
-        (JSObjectDeleteProperty):
-        (JSPropertyEnumeratorGetNext):
-        (JSPropertyListAdd):
-        * API/JSObjectRef.h:
-        * API/JSStringBufferRef.cpp: Removed.
-        * API/JSStringBufferRef.h: Removed.
-        * API/JSValueRef.h:
-        * API/JavaScriptCore.h:
-        * API/minidom.c:
-        (main):
-        (print):
-        * API/testapi.c:
-        (assertEqualsAsUTF8String):
-        (assertEqualsAsCharactersPtr):
-        (assertEqualsAsCharacters):
-        (MyObject_hasProperty):
-        (MyObject_getProperty):
-        (MyObject_setProperty):
-        (MyObject_deleteProperty):
-        (MyObject_getPropertyList):
-        (print_callAsFunction):
-        (myConstructor_callAsConstructor):
-        (main):
-        * JavaScriptCore.exp:
-        * JavaScriptCore.xcodeproj/project.pbxproj:
-
-2006-07-08  Tim Omernick  <timo@apple.com>
-
-        Reviewed by Maciej.
-
-        Added an OpenGL drawing model to the Netscape Plug-in API.
-
-        * bindings/npapi.h:
-
-2006-07-08  Timothy Hatcher  <timothy@apple.com>
-
-        Reviewed by Maciej.
-
-        Moved KJS_GetCreatedJavaVMs to jni_utility.cpp.
-        Switched KJS_GetCreatedJavaVMs over to use dlopen and dlsym
-        now that NSAddImage, NSLookupSymbolInImage and NSAddressOfSymbol
-        are deprecated in Leopard.
-
-        * JavaScriptCore.exp:
-        * JavaScriptCore.xcodeproj/project.pbxproj:
-        * bindings/jni/jni_utility.cpp:
-        (KJS::Bindings::KJS_GetCreatedJavaVMs):
-        * bindings/softlinking.c: Removed.
-        * bindings/softlinking.h: Removed.
-
-2006-07-08  Geoffrey Garen  <ggaren@apple.com>
-
-        Reviewed by Anders.
-        
-        - Make JSObjectGetProperty return a JSValue or NULL, like JSEvaluate does.
-
-        * API/JSObjectRef.cpp:
-        (JSObjectGetProperty):
-        * API/JSObjectRef.h:
-        * API/testapi.c:
-        (main):
-
-2006-07-08  Geoffrey Garen  <ggaren@apple.com>
-
-        Style change -- no review necessary.
-        
-        Use 0 instead of NULL in API .cpp files, to match our style guidelines.
-
-        * API/JSContextRef.cpp:
-        (JSEvaluate):
-        * API/JSObjectRef.cpp:
-        (JSFunctionMakeWithBody):
-        (JSObjectCallAsFunction):
-        (JSObjectCallAsConstructor):
-        * API/JSValueRef.cpp:
-        (JSValueToObject):
-
-2006-07-08  Geoffrey Garen  <ggaren@apple.com>
-
-        Reviewed by TimO.
-
-        - Added ability to pass NULL for thisObject when calling JSObjectCallAsFunction,
-        to match JSEvaluate.
-        
-        * API/JSObjectRef.cpp:
-        (JSObjectCallAsFunction):
-        * API/JSObjectRef.h:
-        * API/testapi.c:
-        (main):
-
-=== Safari-521.15 ===
-
-2006-07-07  Geoffrey Garen  <ggaren@apple.com>
-
-        Reviewed by Maciej.
-        
-        - Standardized which functions take a JSContext as an argument. The rule is:
-        if you might execute JavaScript, you take a JSContext, otherwise you don't.
-        
-        The FIXME in JSObjectRef.h requires refactoring some parts of Interpreter,
-        but not API changes, so I'm putting it off until later.
-
-        * API/JSCallbackObject.cpp:
-        (KJS::JSCallbackObject::JSCallbackObject):
-        (KJS::JSCallbackObject::init):
-        * API/JSCallbackObject.h:
-        * API/JSContextRef.cpp:
-        (JSContextCreate):
-        * API/JSContextRef.h:
-        * API/JSObjectRef.cpp:
-        (JSObjectMake):
-        (JSPropertyEnumeratorGetNext):
-        * API/JSObjectRef.h:
-        * API/testapi.c:
-        (MyObject_initialize):
-        (main):
-        * JavaScriptCore.exp:
-        * kjs/array_object.cpp:
-        (ArrayInstance::setLength):
-        (ArrayInstance::pushUndefinedObjectsToEnd):
-        * kjs/nodes.cpp:
-        (ForInNode::execute):
-        * kjs/reference.cpp:
-        (KJS::Reference::getPropertyName):
-        (KJS::Reference::getValue):
-        * kjs/reference.h:
-        * kjs/scope_chain.cpp:
-        (KJS::ScopeChain::print):
-
-2006-07-06  Geoffrey Garen  <ggaren@apple.com>
-
-        Reviewed by Maciej.
-        
-        More API action.
-        
-        - Headerdoc finished
-
-        Semantic Changes:
-        - Added a JSContextRef argument to many functions, because you need a 
-        JSContextRef for doing virtually anything. I expect to add this argument
-        to even more functions in a future patch.
-        
-        - Removed the globalObjectPrototype argument to JSContextCreate because
-        you can't create an object until you have a context, so it's impossible
-        to pass a prototype object to JSContextCreate. That's OK because (1) there's
-        no reason to give the global object a prototype and (2) if you really want
-        to, you can just use a separate call to JSObjectSetPrototype.
-        
-        - Removed the JSClassRef argument to JSClassCreate because it was unnecessary,
-        and you need to be able to make the global object's class before you've
-        created a JSContext.
-        
-        - Added an optional exception parameter to JSFunctionMakeWithBody because anything
-        less would be uncivilized.
-        
-        - Made the return value parameter to JSObjectGetProperty optional to match
-        all other return value parameters in the API.
-        
-        - Made JSObjectSetPrivate/JSObjectGetPrivate work on JSCallbackFunctions
-        and JSCallbackConstructors. You could use an abstract base class or strategic
-        placement of m_privateData in the class structure to implement this, but 
-        the former seemed like overkill, and the latter seemed too dangerous.
-        
-        - Fixed a bug where JSPropertyEnumeratorGetNext would skip the first property.
-
-        Cosmetic Changes:
-        - Reversed the logic of the JSChar #ifdef to avoid confusing headerdoc
-        
-        - Removed function names from @function declarations because headeroc
-        can parse them automatically, and I wanted to rule out manual mismatch.
-
-        - Changed Error::create to take a const UString& instead of a UString*
-        because it was looking at me funny.
-        
-        - Renamed JSStringBufferCreateWithCFString to JSStringBufferCreateCF
-        because the latter is more concise and it matches JSStringBufferCreateUTF8.
-        
-        * API/JSCallbackObject.cpp:
-        (KJS::JSCallbackObject::getOwnPropertySlot):
-        (KJS::JSCallbackObject::put):
-        (KJS::JSCallbackObject::deleteProperty):
-        (KJS::JSCallbackObject::getPropertyList):
-        (KJS::JSCallbackObject::toBoolean):
-        (KJS::JSCallbackObject::toNumber):
-        (KJS::JSCallbackObject::toString):
-        * API/JSClassRef.cpp:
-        (JSClassCreate):
-        * API/JSContextRef.cpp:
-        (JSContextCreate):
-        (JSContextSetException):
-        * API/JSContextRef.h:
-        * API/JSNode.c:
-        (JSNodePrototype_class):
-        (JSNode_class):
-        * API/JSNodeList.c:
-        (JSNodeListPrototype_class):
-        (JSNodeList_class):
-        * API/JSObjectRef.cpp:
-        (JSObjectGetProperty):
-        (JSObjectGetPrivate):
-        (JSObjectSetPrivate):
-        (JSObjectCallAsFunction):
-        (JSObjectCallAsConstructor):
-        (JSPropertyEnumeratorGetNext):
-        * API/JSObjectRef.h:
-        * API/JSStringBufferRef.cpp:
-        (JSStringBufferCreateCF):
-        * API/JSStringBufferRef.h:
-        * API/JSValueRef.cpp:
-        (JSValueIsInstanceOf):
-        * API/JSValueRef.h:
-        * API/minidom.c:
-        (main):
-        * API/minidom.js:
-        * API/testapi.c:
-        (MyObject_hasProperty):
-        (MyObject_setProperty):
-        (MyObject_deleteProperty):
-        (MyObject_getPropertyList):
-        (MyObject_convertToType):
-        (MyObject_class):
-        (main):
-        * JavaScriptCore.exp:
-
-2006-07-07  Geoffrey Garen  <ggaren@apple.com>
-
-        Reviewed by John.
-        
-        - Fixed a few crashes resulting from NULL parameters to JSClassCreate.
-
-        * API/JSClassRef.cpp:
-        (JSClassCreate):
-        (JSClassRelease):
-        * API/testapi.c: Added test for NULL parameters.
-        (main):
-
-2006-07-07  Geoffrey Garen  <ggaren@apple.com>
-
-        Reviewed by John, mocked by Darin.
-        
-        - Changed JSEvaluate to take a JSObjectRef instead of a JSValueRef as
-        "this," since "this" must be an object.
-
-        * API/JSContextRef.cpp:
-        (JSEvaluate):
-        * API/JSContextRef.h:
-
-2006-07-07  Geoffrey Garen  <ggaren@apple.com>
-
-        Reviewed by John.
-        
-        - More headerdoc
-
-        * API/JSBase.h:
-        * JavaScriptCore.xcodeproj/project.pbxproj:
-
-2006-07-05  Geoffrey Garen  <ggaren@apple.com>
-
-        RS by Beth.
-        
-        Renamed JSCharBufferRef, which was universally unpopular, to JSStringBufferRef,
-        which, hopefully, will be less unpopular.
-
-        * API/APICast.h:
-        (toJS):
-        (toRef):
-        * API/JSBase.h:
-        * API/JSCallbackObject.cpp:
-        (KJS::JSCallbackObject::getOwnPropertySlot):
- &