WebKit-https.git
13 years ago2007-10-18 Eric Seidel <eric@webkit.org>
eseidel [Thu, 18 Oct 2007 07:30:13 +0000 (07:30 +0000)]
2007-10-18  Eric Seidel  <eric@webkit.org>

        Reviewed by Maciej.

        Fix crashers in SVGViewSpec::parseViewSpec
        http://bugs.webkit.org/show_bug.cgi?id=15504

        Test is blocked by bug 15503, landed as:
        * svg/dom/viewspec-parser.html-disabled

        * bindings/scripts/CodeGeneratorObjC.pm: support classes where all parents are interfaces
        * ksvg2/svg/SVGViewSpec.cpp:
        (WebCore::SVGViewSpec::parseViewSpec):
        * ksvg2/svg/SVGViewSpec.idl: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26740 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Rubber stamped by Oliver.
mjs [Thu, 18 Oct 2007 07:27:04 +0000 (07:27 +0000)]
    Rubber stamped by Oliver.

        * tests/string-tagcloud.js: Remove stray print()

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26739 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Oliver.
mjs [Thu, 18 Oct 2007 07:17:41 +0000 (07:17 +0000)]
    Reviewed by Oliver.

        Add some string test cases.

        * tests/string-base64.js: Added. Does base64 encoding and decoding.
        * tests/string-tagcloud.js: Added. Parses a JSON string and
        generates tagcloud HTML markup based on the data.
        * tests/string-unpack-code.js: Added. Unpack compressed versions of
        four of the most popular JavaScript libraries.
        * tests/LIST: Added the new tests.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26738 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Make it possible to have Windows-specific tests and results
aroben [Thu, 18 Oct 2007 03:56:34 +0000 (03:56 +0000)]
     Make it possible to have Windows-specific tests and results

         When searching for tests to run and the Skipped file, we will only
         look in platform/win and the cross-platform directory. When looking
         for expected results, we will look in platform/win, then
         platform/mac-leopard, then platform/mac, then finally the
         cross-platform directory.

         Reviewed by Sam.

         * Scripts/run-webkit-tests:
         (sub expectedDirectoryForTest): Search in mac-leopard and mac before
         searching in the cross-platform directory.
         (sub buildPlatformHierarchy): Removed some unneeded calls to
         dirname/basename.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26734 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Add back the call to register WebKit that we had before WebKitInitializer...
aroben [Thu, 18 Oct 2007 03:53:00 +0000 (03:53 +0000)]
     Add back the call to register WebKit that we had before WebKitInitializer existed

         This is needed to ensure that the right WebKit is used when
         instantiating COM objects.

         Reviewed by Oliver.

         * DumpRenderTree/win/debug_internal.vsprops: Define
         DEBUG_WEBKIT_HAS_SUFFIX.
         * DumpRenderTree/win/DumpRenderTree.cpp:
         (initialize):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26733 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoWebCore:
darin [Thu, 18 Oct 2007 03:28:22 +0000 (03:28 +0000)]
WebCore:

        Reviewed by Darin, Adam, and Maciej.

        http://bugs.webkit.org/show_bug.cgi?id=12988
        First element (in document order) is not returned when other duplicate ID-ed elements were created first

        Reset the element id cache when an id is added and there is a duplicate for that id.

        * dom/Document.cpp:
        (WebCore::Document::addElementById):

LayoutTests:

        Reviewed by Darin, Adam, and Maciej.

        Testcase for:
        http://bugs.webkit.org/show_bug.cgi?id=12988
        First element (in document order) is not returned when other duplicate ID-ed elements were created first

        * fast/dom/duplicate-ids-document-order.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26732 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Import File::Basename in webkitdirs since we use it
aroben [Thu, 18 Oct 2007 03:17:36 +0000 (03:17 +0000)]
    Import File::Basename in webkitdirs since we use it

        Reviewed by NOBODY.

        * Scripts/webkitdirs.pm:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26730 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago - Updated the tiger leaks list to make the internal bots green.
kmccullo [Thu, 18 Oct 2007 01:07:54 +0000 (01:07 +0000)]
    - Updated the tiger leaks list to make the internal bots green.

        * Scripts/run-webkit-tests:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26729 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Build fix for VC++ Express
aroben [Thu, 18 Oct 2007 00:15:29 +0000 (00:15 +0000)]
    Build fix for VC++ Express

        Reviewed by NOBODY.

        * FindSafari/FindSafari.vcproj: Explicitly link against advapi32.lib
        and ole32.lib.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26727 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Fix an ASSERT on launch on Windows
aroben [Thu, 18 Oct 2007 00:10:18 +0000 (00:10 +0000)]
    Fix an ASSERT on launch on Windows

        Fix dictated to me by Brady.

        * loader/icon/IconDatabase.cpp:
        (WebCore::IconDatabase::syncThreadMainLoop): Make sure we hold
        m_syncLock before going into the loop.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26726 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago2007-10-17 Mark Rowe <mrowe@apple.com>
bdash [Wed, 17 Oct 2007 23:51:34 +0000 (23:51 +0000)]
2007-10-17  Mark Rowe  <mrowe@apple.com>

        Mac build fix.

        * Configurations/WebCore.xcconfig:  Make it possible to include CoreGraphics header files.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26725 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Fix for clean builds needed after r26683
aroben [Wed, 17 Oct 2007 23:39:15 +0000 (23:39 +0000)]
    Fix for clean builds needed after r26683

        Reviewed by NOBODY.

        * WebView.cpp: Remove #include of non-existant file.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26724 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Windows build fix
aroben [Wed, 17 Oct 2007 23:10:09 +0000 (23:10 +0000)]
    Windows build fix

        Reviewed by NOBODY.

        * platform/graphics/AffineTransform.h: Don't #include
        ApplicationServices.h.
        * platform/graphics/cg/GraphicsContextCG.cpp: Add an #include that's
        needed now that we're not #including ApplicationServices.h.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26723 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago2007-10-17 Mark Rowe <mrowe@apple.com>
bdash [Wed, 17 Oct 2007 22:57:12 +0000 (22:57 +0000)]
2007-10-17 Mark Rowe <mrowe@apple.com>

        Gtk build fix.  Move non-pthread stubs from Threading.h to ThreadingNone.cpp to prevent
        multiple-definition link errors.

        * WebCore.pro:
        * platform/Threading.h:
        * platform/ThreadingNone.cpp: Added.
        (WebCore::createThread):
        (WebCore::waitForThreadCompletion):
        (WebCore::detachThread):
        (WebCore::Mutex::Mutex):
        (WebCore::Mutex::~Mutex):
        (WebCore::Mutex::lock):
        (WebCore::Mutex::tryLock):
        (WebCore::Mutex::unlock):
        (WebCore::ThreadCondition::ThreadCondition):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26722 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Rubber-stamped by Adele.
aliceli1 [Wed, 17 Oct 2007 22:37:23 +0000 (22:37 +0000)]
    Rubber-stamped by Adele.

        Correct uses of initKeyboardEvent for tests that test accesskey.
        Accesskey modifier is ctrl for mac and alt for other platforms.

        * fast/events/access-key-self-destruct.html:
        * fast/forms/access-key.html:
        * fast/forms/legend-access-key.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26721 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoJavaScriptCore:
darin [Wed, 17 Oct 2007 21:25:59 +0000 (21:25 +0000)]
JavaScriptCore:

        Reviewed by Mark Rowe.

        - fix http://bugs.webkit.org/show_bug.cgi?id=15543
          <rdar://problem/5545639> REGRESSION (r26697):
          GoogleDocs: Can't create new documents or open existing ones

        Test: fast/js/regexp-non-character.html

        * pcre/pcre_compile.c: (check_escape): Take out the checks for valid characters
        in the \u sequences -- not needed and actively harmful.

LayoutTests:

        Reviewed by Mark Rowe.

        - test for http://bugs.webkit.org/show_bug.cgi?id=15543
          <rdar://problem/5545639> REGRESSION (r26697):
          GoogleDocs: Can't create new documents or open existing ones

        * fast/js/regexp-non-character-expected.txt: Added.
        * fast/js/regexp-non-character.html: Added.
        * fast/js/resources/regexp-non-character.js: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26720 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoRemove stray change marker.
darin [Wed, 17 Oct 2007 21:15:16 +0000 (21:15 +0000)]
Remove stray change marker.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26718 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoRemove stray merge marker.
darin [Wed, 17 Oct 2007 21:13:59 +0000 (21:13 +0000)]
Remove stray merge marker.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26717 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoJavaScriptCore:
andersca [Wed, 17 Oct 2007 21:05:38 +0000 (21:05 +0000)]
JavaScriptCore:

        Reviewed by Oliver.

        * wtf/Platform.h:
        #define USE_PTHREADS on Mac.

WebCore:

        Reviewed by Oliver.

        * WebCore.xcodeproj/project.pbxproj:
        * WebCore.vcproj/WebCore.vcproj:

        * config.h:
        #define USE_PTHREADS on Windows.

        * loader/icon/IconDatabase.cpp:
        (WebCore::IconDatabase::getOrCreateIconRecord):
        (WebCore::IconDatabase::getOrCreatePageURLRecord):
        Update for mutex changes.

        * platform/Threading.h:
        (WebCore::):
        (WebCore::ThreadSafeShared::ThreadSafeShared):
        (WebCore::ThreadSafeShared::ref):
        (WebCore::ThreadSafeShared::deref):
        (WebCore::ThreadSafeShared::hasOneRef):
        (WebCore::ThreadSafeShared::refCount):
        (WebCore::ThreadSafeShared::isThreadSafe):
        Add a new Shared base class that uses a mutex to manage its refcount.

        (WebCore::createThread):
        (WebCore::waitForThreadCompletion):
        (WebCore::detachThread):
        New functions for thread creation.

        (WebCore::Mutex::Mutex):
        (WebCore::Mutex::~Mutex):
        (WebCore::Mutex::lock):
        (WebCore::Mutex::tryLock):
        (WebCore::Mutex::unlock):
        Move the definitions to ThreadingPthreads.cpp

        (WebCore::ThreadCondition::ThreadCondition):
        (WebCore::ThreadCondition::~ThreadCondition):
        (WebCore::ThreadCondition::wait):
        (WebCore::ThreadCondition::signal):
        (WebCore::ThreadCondition::broadcast):
        Move the definitions to ThreadingPthreads.cpp

        * platform/pthreads: Added.
        * platform/pthreads/ThreadingPthreads.cpp: Added.
        Add pthread specific implementation of the thread creation functions, Mutex and ThreadCondition.

win:

        Update for locking primitive changes.

        * WebIconDatabase.cpp:
        (WebIconDatabase::scheduleNotificationDelivery):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26716 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Darin Adler.
ggaren [Wed, 17 Oct 2007 21:01:27 +0000 (21:01 +0000)]
    Reviewed by Darin Adler.

        Merged DeclaredFunctionImp into FunctionImp (the base class) because
        the distinction between the two was unused.

        Removed codeType() from FunctionImp because FunctionImp and its
        subclasses all returned FunctionCode, so it was unused, practically
        speaking.

        Removed a different codeType() from GlobalFuncImp because it was unused.
        (Perhaps it was vestigial from a time when GlobalFuncImp used to
        inherit from FunctionImp.)

        * bindings/runtime_method.cpp:
        * bindings/runtime_method.h:
        * kjs/function.cpp:
        (KJS::):
        (KJS::FunctionImp::FunctionImp):
        (KJS::FunctionImp::callAsFunction):
        (KJS::FunctionImp::construct):
        (KJS::FunctionImp::execute):
        (KJS::FunctionImp::processVarDecls):
        * kjs/function.h:
        (KJS::FunctionImp::implementsConstruct):
        (KJS::FunctionImp::scope):
        * kjs/function_object.cpp:
        (FunctionProtoFunc::callAsFunction):
        (FunctionObjectImp::construct):
        * kjs/nodes.cpp:
        (FuncDeclNode::processFuncDecl):
        (FuncExprNode::evaluate):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26715 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Add support for creating a layer when a transform is in effect. Add support...
hyatt [Wed, 17 Oct 2007 20:28:09 +0000 (20:28 +0000)]
    Add support for creating a layer when a transform is in effect.  Add support to RenderStyle for
        handing back a computed AffineTransform.

        Reviewed by Mitz Pettel

        * css/CSSStyleSelector.cpp:
        (WebCore::CSSStyleSelector::adjustRenderStyle):
        * rendering/RenderBox.cpp:
        (WebCore::RenderBox::setStyle):
        * rendering/RenderObject.cpp:
        (WebCore::RenderObject::RenderObject):
        (WebCore::RenderObject::requiresLayer):
        (WebCore::RenderObject::setStyle):
        * rendering/RenderObject.h:
        (WebCore::RenderObject::hasTransform):
        (WebCore::RenderObject::setHasTransform):
        * rendering/RenderStyle.cpp:
        (WebCore::RenderStyle::applyTransform):
        * rendering/RenderStyle.h:
        (WebCore::RenderStyle::hasTransform):
        * rendering/RenderTableCell.cpp:
        (WebCore::RenderTableCell::requiresLayer):
        * rendering/RenderTableRow.h:
        (WebCore::RenderTableRow::requiresLayer):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26710 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Mark Rowe.
darin [Wed, 17 Oct 2007 20:24:10 +0000 (20:24 +0000)]
    Reviewed by Mark Rowe.

        - fix <rdar://problem/5183775> Uninitialized memory in -[WebDynamicScrollBarsView updateScrollers]

        * WebView/WebDynamicScrollBarsView.m: (-[WebDynamicScrollBarsView updateScrollers]):
        Change code path so it doesn't dispatch a method that returns an NSSize passing a nil object.
        It's safe to do that for functions that return integers or pointers, but not structures.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26709 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Add a new nightly launcher script for Windows
aroben [Wed, 17 Oct 2007 20:12:42 +0000 (20:12 +0000)]
     Add a new nightly launcher script for Windows

         This new script is for versions of Safari that don't delay-load
         WebKit. I had to make FindSafari be able to print out the nightly
         launcher script on stdout because Windows shell scripts can't capture
         the output from a command into an environment variable.

         Reviewed by Steve.

         * FindSafari/FindSafari.cpp:
         (_tmain): Added /printSafariLauncher option.
         * Scripts/run-webkit-nightly.cmd: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26708 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoStop delay-loading WebKit in DRT and teskjs
aroben [Wed, 17 Oct 2007 20:12:19 +0000 (20:12 +0000)]
Stop delay-loading WebKit in DRT and teskjs

JavaScriptCore:

        Make testkjs not delay-load WebKit

        Soon, delay-loading WebKit will be impossible (because we will be
        using __declspec(thread) for thread-local storage). This change
        prepares testkjs for the future.

        Reviewed by Sam.

        * JavaScriptCore.vcproj/JavaScriptCore.sln: Removed WebKitInitializer,
        added FindSafari.
        * JavaScriptCore.vcproj/testkjs/testkjs.vcproj: Don't link against
        WebKitInitializer, don't delay-load WebKit.
        * kjs/testkjs.cpp: Don't use WebKitInitializer.

WebKitTools:

        Replace WebKitInitializer with FindSafari

        We now rely on setting the PATH environment variable to tell Windows
        where to find WebKit and its dependencies (similar to
        DYLD_FRAMEWORK_PATH on Mac).

        This change also make DumpRenderTree no longer delay-load WebKit.

        Reviewed by Sam.

        * DumpRenderTree/DumpRenderTree.sln: Removed WebKitInitializer, added
        FindSafari.
        * DumpRenderTree/win/DumpRenderTree.cpp: Don't use WebKitInitializer.
        * DumpRenderTree/win/DumpRenderTree.vcproj: Don't link against
        WebKitInitializer, don't delay-load WebKit.
        * Scripts/run-javascriptcore-tests: Call setPathForRunningWebKitApp.
        * Scripts/run-webkit-tests: Ditto.
        * Scripts/webkitdirs.pm:
        (sub setPathForRunningWebKitApp): Added.
        * WebKitInitializer/WebKitInitializer.cpp: Removed.
        * WebKitInitializer/WebKitInitializer.h: Removed.
        * WebKitInitializer/WebKitInitializer.vcproj: Removed.
        * WebKitInitializer/debug.vsprops: Removed.
        * WebKitInitializer/debug_internal.vsprops: Removed.
        * WebKitInitializer/release.vsprops: Removed.

WebKit/win:

        Remove WebKitInitializer

        Reviewed by Sam.

        * WebKit.vcproj/WebKit.sln:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26707 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Fix Bug 15532: run-safari fails if Safari is installed in a non-default location
aroben [Wed, 17 Oct 2007 20:11:24 +0000 (20:11 +0000)]
 Fix Bug 15532: run-safari fails if Safari is installed in a non-default location

 WebKitTools:

         Fix Bug 15532: run-safari fails if Safari is installed in a non-default location

         http://bugs.webkit.org/show_bug.cgi?id=15532

         FindSafari simply prints the location of an installed Safari.exe on
         stdout.

         Reviewed by Darin.

         * FindSafari/FindSafari.cpp: Copied from WebKitTools/WebKitInitializer/WebKitInitializer.cpp.
         (getStringValue):
         (getWebViewCLSID):
         (getInstalledWebKitDirectory):
         (_tmain):
         * FindSafari/FindSafari.vcproj: Added.
         * Scripts/webkitdirs.pm:
         (sub installedSafariPath): Added. Calls FindSafari on Windows.

 WebKit/win:

         Add FindSafari

         Reviewed by Darin.

         * WebKit.vcproj/WebKit.sln:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26706 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Add support for mapping of the transform CSS property into RenderStyles....
hyatt [Wed, 17 Oct 2007 19:47:00 +0000 (19:47 +0000)]
    Add support for mapping of the transform CSS property into RenderStyles.  Everything is now ready for the front
        end to use.

        Reviewed by Mitz Pettel

        * WebCore.xcodeproj/project.pbxproj:
        * css/CSSStyleSelector.cpp:
        (WebCore::CSSStyleSelector::applyProperty):
        * css/CSSTransformValue.h:
        (WebCore::CSSTransformValue::type):
        (WebCore::CSSTransformValue::values):
        * platform/graphics/AffineTransform.cpp:
        (WebCore::AffineTransform::skew):
        * platform/graphics/AffineTransform.h:
        * rendering/RenderStyle.cpp:
        (WebCore::StyleTransformData::StyleTransformData):
        (WebCore::StyleTransformData::operator==):
        (WebCore::StyleTransformData::transformDataEquivalent):
        * rendering/RenderStyle.h:
        (WebCore::TransformOperation::~TransformOperation):
        (WebCore::TransformOperation::operator!=):
        (WebCore::TransformOperation::isScaleOperation):
        (WebCore::TransformOperation::isRotateOperation):
        (WebCore::TransformOperation::isSkewOperation):
        (WebCore::TransformOperation::isTranslateOperation):
        (WebCore::TransformOperation::isMatrixOperation):
        (WebCore::ScaleTransformOperation::ScaleTransformOperation):
        (WebCore::ScaleTransformOperation::isScaleOperation):
        (WebCore::ScaleTransformOperation::operator==):
        (WebCore::ScaleTransformOperation::apply):
        (WebCore::RotateTransformOperation::RotateTransformOperation):
        (WebCore::RotateTransformOperation::isRotateOperation):
        (WebCore::RotateTransformOperation::operator==):
        (WebCore::RotateTransformOperation::apply):
        (WebCore::SkewTransformOperation::SkewTransformOperation):
        (WebCore::SkewTransformOperation::isSkewOperation):
        (WebCore::SkewTransformOperation::operator==):
        (WebCore::SkewTransformOperation::apply):
        (WebCore::TranslateTransformOperation::TranslateTransformOperation):
        (WebCore::TranslateTransformOperation::isTranslateOperation):
        (WebCore::TranslateTransformOperation::operator==):
        (WebCore::TranslateTransformOperation::apply):
        (WebCore::MatrixTransformOperation::MatrixTransformOperation):
        (WebCore::MatrixTransformOperation::isMatrixOperation):
        (WebCore::MatrixTransformOperation::operator==):
        (WebCore::MatrixTransformOperation::apply):
        (WebCore::StyleTransformData::operator!=):
        (WebCore::RenderStyle::transform):
        (WebCore::RenderStyle::setTransform):
        (WebCore::RenderStyle::initialTransform):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26705 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoWindows build fix part 2.
aroben [Wed, 17 Oct 2007 19:30:07 +0000 (19:30 +0000)]
Windows build fix part 2.

JavaScriptCore:

        Windows build fix part 2.

        Fix was by Darin, reviewed by Anders and Adam.

        * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: Add
        FastMallocPCRE.cpp to the project, and let Visual Studio have its way
        with the post-build step.
        * pcre/pcre.h: Don't DLL export the entry points just because this
        is Win32 -- this is an internal copy of PCRE and should be private.
        * pcre/pcre_compile.c: Fix an uninitialized variable warning --
        there's no real problem but it's better to quiet the compiler by
        tweaking the code slightly than turn off the warning entirely.

WebCore:

        * page/Frame.cpp: I had to touch this file when fixing the Windows
        build. Checking it in in case it helps others, too.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26704 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago2007-10-17 Eric Seidel <eric@webkit.org>
eseidel [Wed, 17 Oct 2007 18:33:25 +0000 (18:33 +0000)]
2007-10-17  Eric Seidel  <eric@webkit.org>

        Reviewed by Mark Rowe.

        Remove a couple more uses of svg_dynamic_cast.

        No functional changes, no tests.

        * ksvg2/svg/SVGPatternElement.cpp:
        (WebCore::SVGPatternElement::buildPattern):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26703 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago2007-10-17 Mark Rowe <mrowe@apple.com>
bdash [Wed, 17 Oct 2007 18:05:16 +0000 (18:05 +0000)]
2007-10-17  Mark Rowe  <mrowe@apple.com>

        Reviewed by Eric and Geoff.

        <rdar://problem/5453743> Repro ASSERT in CachedResource::setEncodedSize() loading image in background tab
        http://bugs.webkit.org/show_bug.cgi?id=15191

        * loader/DocLoader.cpp:
        (WebCore::DocLoader::setAutoLoadImages): Don't start a load if the CachedImage is already loading.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26702 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Windows build fix.
aroben [Wed, 17 Oct 2007 17:54:08 +0000 (17:54 +0000)]
    Windows build fix.

        Reviewed by Anders.

        * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj: Disable
        some mismatched signed/unsigned comparison warnings.
        * pcre/pcre_exec.c:
        (match): #if-out some labels that don't seem to exist.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26701 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoWebCore:
justing [Wed, 17 Oct 2007 17:51:27 +0000 (17:51 +0000)]
WebCore:

        Reviewed by Harrison.

        <rdar://problem/5481523>
        GoogleDocs: Safari hangs when indenting a particular table twice

        Fixed by fixing problems with the selection preservation done by IndentOutdentCommand.
        It is now more difficult to create selections that cause hangs.  Those are covered by:
        <rdar://problem/5543472>

        * editing/IndentOutdentCommand.cpp:
        (WebCore::indexForVisiblePosition): Compute indices from the first VisiblePosition
        in the document, instead of the first Position.
        Use rangeCompliantEquivalents when creating the Range that we pass to rangeLength.
        Tell TextIterator::rangeLength that we're doing selection preservation, so that it
        will emit characters between all VisiblePositions.
        (WebCore::IndentOutdentCommand::indentRegion): Fixed a bug where the range and location
        passed to rangeFromLocationAndLength were reversed.
        Tell rangeFromLocationAndLength that we're doing doing selection preservation, as
        above.
        * editing/TextIterator.cpp:
        (WebCore::CharacterIterator::CharacterIterator):
        (WebCore::TextIterator::rangeLength): Rename emitSpacesForReplacedElements
        to forSelectionPreservation, to match the name of the boolean inside TextIterator and
        to match its meaning after r25522.
        (WebCore::TextIterator::rangeFromLocationAndLength): Ditto.

LayoutTests:

        Reviewed by Harrison.

        <rdar://problem/5481523>
        GoogleDocs: Safari hangs when indenting a particular table twice

        * editing/execCommand/5481523.html: Added.
        * platform/mac/editing/execCommand/5481523-expected.checksum: Added.
        * platform/mac/editing/execCommand/5481523-expected.png: Added.
        * platform/mac/editing/execCommand/5481523-expected.txt: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26700 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago2007-10-17 Mark Rowe <mrowe@apple.com>
bdash [Wed, 17 Oct 2007 07:45:27 +0000 (07:45 +0000)]
2007-10-17  Mark Rowe  <mrowe@apple.com>

        Gtk build fix.

        * JavaScriptCore.pri: Add FastMallocPCRE.cpp.
        * pcre/pcre_get. #if out two functions that depend on pcre_get_stringnumber, which
        is currently unavailable for UTF-16.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26699 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Anders.
mjs [Wed, 17 Oct 2007 07:12:51 +0000 (07:12 +0000)]
    Reviewed by Anders.

        * TODO: Updated with info on how to do the stats properly.

        * hosted/sunspider-driver.html: Reduce timeout from 500 to 200 to
        avoid overly slowing down the benchmark.

        Added some 3D-related test cases.

        * tests/3d-cube.js: Added.
        * tests/3d-morph.js: Added.
        * tests/3d-raytrace.js: Added.
        * tests/LIST:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26698 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Geoff.
darin [Wed, 17 Oct 2007 05:38:39 +0000 (05:38 +0000)]
    Reviewed by Geoff.

        - merged PCRE changes between 6.4 and 6.5

        * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
        * JavaScriptCore.xcodeproj/project.pbxproj:
        Removed pcre_config.c, pcre_globals.c, pcre_info.c, pcre_maketables.c,
        pcre_printint.src, pcre_refcount.c, pcre_study.c, pcre_try_flipped.c,
        pcre_ucp_findchar.c, pcre_version.c, and ucptable.c. Added pcre_ucp_searchfuncs.c.

        * pcre/AUTHORS:
        * pcre/LICENCE:
        * pcre/MERGING:
        * pcre/dftables.c:
        * pcre/pcre-config.h:
        * pcre/pcre.h:
        * pcre/pcre.pri:
        * pcre/pcre_compile.c:
        * pcre/pcre_exec.c:
        * pcre/pcre_fullinfo.c:
        * pcre/pcre_get.c:
        * pcre/pcre_internal.h:
        * pcre/pcre_maketables.c:
        * pcre/pcre_ord2utf8.c:
        * pcre/pcre_tables.c:
        * pcre/pcre_ucp_searchfuncs.c: Copied from pcre/pcre_ucp_findchar.c.
        * pcre/pcre_xclass.c:
        * pcre/ucp.h:
        * pcre/ucpinternal.h:
        * pcre/ucptable.c:
        Updated with new versions from the PCRE 6.5 release, merged with changes.

        * pcre/pcre_config.c: Removed.
        * pcre/pcre_globals.c: Removed.
        * pcre/pcre_info.c: Removed.
        * pcre/pcre_printint.src: Removed.
        * pcre/pcre_refcount.c: Removed.
        * pcre/pcre_study.c: Removed.
        * pcre/pcre_try_flipped.c: Removed.
        * pcre/pcre_ucp_findchar.c: Removed.
        * pcre/pcre_version.c: Removed.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26697 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Mitz.
darin [Wed, 17 Oct 2007 04:51:50 +0000 (04:51 +0000)]
    Reviewed by Mitz.

        - fix http://bugs.webkit.org/show_bug.cgi?id=15536
          need to cache missing glyph so we're not slow on pages that show missing glyphs

        - <rdar://problem/5404359> UI thread stall (>60sec) in MLANG running stress test
          (related to font data caching)

        * platform/Font.cpp: (WebCore::Font::glyphDataForCharacter): Check for a null value
        for fontData rather than for glyph to detect non-cached entries in the glyph data.

        * platform/gtk/GlyphPageTreeNodeGtk.cpp: (WebCore::GlyphPage::fill):
        * platform/mac/GlyphPageTreeNodeMac.cpp: (WebCore::GlyphPage::fill):
        * platform/win/GlyphPageTreeNodeWin.cpp: (WebCore::GlyphPage::fill):
        Set fontData to 0 for missing glyph entries. Also fixed the Windows and GTK versions
        to return the proper value for haveGlyphs.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26696 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago2007-10-16 Mark Rowe <mrowe@apple.com>
bdash [Wed, 17 Oct 2007 03:47:35 +0000 (03:47 +0000)]
2007-10-16  Mark Rowe  <mrowe@apple.com>

        Gtk build fix.  Replace use of assert with ASSERT_NOT_REACHED.

        * platform/gtk/FontCacheGtk.cpp:
        (WebCore::FontCache::platformInit):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26695 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoWebCore:
ddkilzer [Wed, 17 Oct 2007 02:02:53 +0000 (02:02 +0000)]
WebCore:

        Reviewed by Timothy.

        <rdar://problem/5544354> Wrong delegate method called in WebFrameLoaderClient::dispatchDidReceiveServerRedirectForProvisionalLoad()

        * WebCoreSupport/WebFrameLoaderClient.mm:
        (WebFrameLoaderClient::dispatchDidReceiveServerRedirectForProvisionalLoad): Fixed selector name.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26694 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Sam.
kmccullo [Wed, 17 Oct 2007 00:57:12 +0000 (00:57 +0000)]
    Reviewed by Sam.

        - Minor mac improvements based on observations I made while implementing the
        Windows code.

        * Drosera/mac/DebuggerApplication.mm: Attach does not need to create the server
        then hand it to the Client who will retain it, it should just be the client who
        creates and owns the server.
        (-[DebuggerApplication attach:]):
        * Drosera/mac/DebuggerClient.h: Can now be created with just a server name.
        * Drosera/mac/DebuggerClient.mm: Creates the server.
        (-[DebuggerClient initWithServerName:]):
        * Drosera/mac/ServerConnection.mm: Does not need to include 2 header files, but
        can instead forward declare what it needs.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26693 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago removing fixed tests
aliceli1 [Wed, 17 Oct 2007 00:36:35 +0000 (00:36 +0000)]
    removing fixed tests

        * platform/win/Skipped:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26692 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Adam.
darin [Tue, 16 Oct 2007 23:52:44 +0000 (23:52 +0000)]
    Reviewed by Adam.

        - fix http://bugs.webkit.org/show_bug.cgi?id=15534
          WebScriptObject.h has some formatting and editorial mistakes

        * bindings/objc/WebScriptObject.h: Fix wording and formatting.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26691 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Darin Adler.
ggaren [Tue, 16 Oct 2007 23:35:01 +0000 (23:35 +0000)]
    Reviewed by Darin Adler.

        Removed KJS_VERBOSE because it was getting in the way of readability,
        and the messages didn't seem very helpful.

        * kjs/function.cpp:
        (KJS::FunctionImp::callAsFunction):
        (KJS::FunctionImp::passInParameters):
        * kjs/lookup.h:
        (KJS::lookupPut):
        * kjs/object.cpp:
        (KJS::JSObject::put):
        * kjs/value.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26690 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Darin Adler.
ggaren [Tue, 16 Oct 2007 23:29:55 +0000 (23:29 +0000)]
    Reviewed by Darin Adler.

        Removed the Parameter class because it was a redundant wrapper around
        Identifier.

        * kjs/function.cpp:
        (KJS::FunctionImp::passInParameters):
        (KJS::FunctionImp::getParameterName):
        * kjs/nodes.cpp:
        (FunctionBodyNode::addParam):
        * kjs/nodes.h:
        (KJS::FunctionBodyNode::):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26689 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Darin Adler.
ggaren [Tue, 16 Oct 2007 23:25:33 +0000 (23:25 +0000)]
    Reviewed by Darin Adler.

        Global replace of assert with ASSERT.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26688 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Rubber-stamped by Adele.
aliceli1 [Tue, 16 Oct 2007 22:34:06 +0000 (22:34 +0000)]
    Rubber-stamped by Adele.

        fixing incorrect usage of JS String search method.

        * editing/deleting/delete-by-word-001.html:
        * editing/deleting/delete-by-word-002.html:
        * fast/forms/input-text-option-delete.html:
        * fast/forms/listbox-onchange.html:
        * fast/forms/listbox-selection.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26687 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago - try to fix the Qt build
darin [Tue, 16 Oct 2007 22:02:41 +0000 (22:02 +0000)]
    - try to fix the Qt build

        * config.h: Don't use DisallowCType.h, since it's incompatible with some C++ headers
        that are used in some Qt-specific source files.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26686 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Geoff Garen.
darin [Tue, 16 Oct 2007 21:53:20 +0000 (21:53 +0000)]
    Reviewed by Geoff Garen.

        * Scripts/run-webkit-tests: Added --add-platform-exceptions; useful when you
        want to turn failing tests into platform-specific test results. Also did a
        number of tweaks, including fixing a potential bug where expectedDirectoryForTest
        would take the type of results into account only some of the time.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26685 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Rename WebKit_debug.dll to WebKit.dll for the Debug configuration
aroben [Tue, 16 Oct 2007 21:43:14 +0000 (21:43 +0000)]
 Rename WebKit_debug.dll to WebKit.dll for the Debug configuration

 JavaScriptCore:

         Updated testkjs for the rename of WebKit_debug.dll to WebKit.dll for the Debug configuration

         Reviewed by Kevin McCullough.

         * JavaScriptCore.vcproj/debug.vsprops: Added WebKitDLLConfigSuffix.
         * JavaScriptCore.vcproj/debug_internal.vsprops: Ditto.
         * JavaScriptCore.vcproj/release.vsprops: Ditto.
         * JavaScriptCore.vcproj/testkjs/testkjs.vcproj: Use
         WebKitDLLConfigSuffix when referring to WebKit.dll, and fixed a typo
         in the name of icuuc36[_debug].dll.

 WebKitTools:

         Updated for rename of WebKit_debug.dll to WebKit.dll for the Debug configuration

         Reviewed by Kevin McCullough.

         * DumpRenderTree/win/DumpRenderTree.vcproj: Use WebKitDLLConfigSuffix
         when referring to WebKit.dll.
         * DumpRenderTree/win/debug.vsprops: Added WebKitDLLConfigSuffix.
         * DumpRenderTree/win/debug_internal.vsprops: Ditto.
         * DumpRenderTree/win/release.vsprops: Ditto.
         * Scripts/run-safari: Don't pass /debug to Safari anymore because we
         never have a _debug suffix on WebKit.dll.
         * WebKitInitializer/WebKitInitializer.cpp:
         (initializeWebKit): Only use the _debug suffix for WebKit.dll when
         specified.
         * WebKitInitializer/debug_internal.vsprops: Added
         DEBUG_WEBKIT_HAS_SUFFIX preprocessor definition.

 WebKit/win:

         Rename WebKit_debug.dll to WebKit.dll for the Debug configuration

         This is needed so that we can stop having Safari delay-load WebKit,
         which in turn is needed so that Maciej can land a JavaScript speedup
         which breaks delay-loading.

         Reviewed by Kevin McCullough.

         * WebKit.vcproj/WebKit.vcproj: Use WebKitDLLConfigSuffix for the name
         of our DLL and module definition file.
         * WebKit.vcproj/debug.vsprops: Added WebKitDLLConfigSuffix.
         * WebKit.vcproj/debug_internal.vsprops: Ditto.
         * WebKit.vcproj/release.vsprops: Ditto.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26684 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoWebCore:
darin [Tue, 16 Oct 2007 21:40:11 +0000 (21:40 +0000)]
WebCore:

        Reviewed by Adele.

        - fix http://bugs.webkit.org/show_bug.cgi?id=15525
          transpose (control-T) should do the last two characters on a line if at end of line
        - remove unused CommandByName class
        - move transpose command implementation here from Mac OS X WebKit

        * editing/Editor.h:
        * editing/Editor.cpp:
        (WebCore::execTranspose): Added.
        (WebCore::Editor::transpose): Added. This has the transpose implementation from
        WebCoreFrameBridge and WebHTMLView, translated into WebCore-style C++ and with
        a special case for the end of a paragraph.

        * page/Frame.h: Removed unused command() function member.
        * page/FramePrivate.h: Removed unused m_command data member.
        * page/Frame.cpp: Ditto.

        * page/mac/WebCoreFrameBridge.h: Removed rangeOfCharactersAroundCaret method.
        * page/mac/WebCoreFrameBridge.mm: Ditto.

        * WebCore.pro: Removed CommandByName.h/cpp.
        * WebCore.vcproj/WebCore.vcproj: Ditto.
        * WebCore.xcodeproj/project.pbxproj: Ditto.

        * editing/CommandByName.cpp: Removed.
        * editing/CommandByName.h: Removed.

WebKit:

        Reviewed by Adele.

        - moved transpose command implementation into WebCore

        * WebView/WebHTMLView.mm: Removed transpose: and replaced it with standard WebCore forwarding.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26683 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoJavaScriptCore:
ggaren [Tue, 16 Oct 2007 21:35:50 +0000 (21:35 +0000)]
JavaScriptCore:

        Reviewed by Maciej Stachowiak.

        Re-structured variable and function declaration code.

        Command-line JS iBench shows no regression.

        Here are the changes:

        1. Function declarations are now processed at the same time as var
        declarations -- namely, immediately upon entry to an execution context.
        This does not match Firefox, which waits to process a function
        declaration until the declaration's containing block executes, but it
        does match IE and the ECMA spec. (10.1.3 states that var and function
        declarations should be processed at the same time -- namely, "On
        entering an execution context." 12.2 states that "A Block does not
        define a new execution scope.")

        2. Declaration processing proceeds iteratively now, rather than
        recursively, storing the nodes is finds in stacks. This will later
        facilitate an optimization to hold on to the gathered declaration nodes,
        rather than re-fetching them in every function call.
        [ http://bugs.webkit.org/show_bug.cgi?id=14868 ]

        Modified these tests because they expected the incorrect Mozilla
        behavior described above:

        * tests/mozilla/ecma_3/Function/scope-001.js:
        * tests/mozilla/js1_5/Scope/regress-184107.js:

LayoutTests:

        Reviewed by Maciej Stachowiak.

        Layout tests for http://bugs.webkit.org/show_bug.cgi?id=15478
        Declare vars and functions iteratively upon entering an execution context

        * fast/js/function-declarations-expected.txt: Added.
        * fast/js/function-declarations.html: Added.
        * fast/js/var-declarations-expected.txt: Added.
        * fast/js/var-declarations.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26682 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Adele.
aliceli1 [Tue, 16 Oct 2007 21:18:52 +0000 (21:18 +0000)]
    Reviewed by Adele.

        Fix <rdar://5134108> "fast/forms/select-accesskey.html and focus-selection-input.html fail on Windows"

        Adding code to choose the right accesskey modifier based on a useragent check for existance of mac platform.
        * fast/forms/focus-selection-input.html:
        * fast/forms/select-accesskey.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26681 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago - try to fix the GTK build
darin [Tue, 16 Oct 2007 20:53:21 +0000 (20:53 +0000)]
    - try to fix the GTK build

        * kjs/ustring.cpp: Include ASCIICType.h, not ASCIICtype.h.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26680 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago - try to fix the Windows build
darin [Tue, 16 Oct 2007 20:44:28 +0000 (20:44 +0000)]
    - try to fix the Windows build

        * kjs/date_object.cpp: (KJS::parseDate): A couple instances of isspace were
        in here. Not sure why it wasn't failing elsewhere. Changed to isASCIISpace.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26679 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago - try to fix the GTK build
darin [Tue, 16 Oct 2007 20:34:01 +0000 (20:34 +0000)]
    - try to fix the GTK build

        * kjs/ustring.cpp: Include ASCIICType.h.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26678 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoWebKitTools:
darin [Tue, 16 Oct 2007 20:28:58 +0000 (20:28 +0000)]
WebKitTools:

        Reviewed by Adele.

        * DumpRenderTree/win/EventSender.cpp: (keyDownCallback):
        Remove now-unneeded control/alt key hack; the tests have been updated.

LayoutTests:

        Reviewed by Adele.

        - http://bugs.webkit.org/show_bug.cgi?id=15533
          some tests have hard-coded Mac-specific modifier keys

        - <rdar://problem/5102974> Layout tests fail on Windows because
          they have hard-coded Mac-specific modifier keys
        - <rdar://problem/5075449> fast/forms/listbox-onchange.html
          is failing on Windows due to modifier key differences
        - <rdar://problem/5134134> test 12 in fast/forms/listbox-selection.html
          fails on Windows

        * editing/deleting/delete-by-word-001.html: Use control key instead of alt key to
        delete a word on non-Mac platforms. Also use character 8 instead of 127.
        * editing/deleting/delete-by-word-002.html: Ditto.
        * fast/forms/input-text-option-delete.html: Ditto.
        * fast/forms/listbox-onchange.html: Use control key instead of meta key to toggle
        a single list item on non-Mac platforms.
        * fast/forms/listbox-selection.html: Ditto.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26677 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoJavaScriptCore:
darin [Tue, 16 Oct 2007 20:13:24 +0000 (20:13 +0000)]
JavaScriptCore:

        Reviewed by Maciej and Geoff (and looked over by Eric).

        - http://bugs.webkit.org/show_bug.cgi?id=15519
          eliminate use of <ctype.h> for processing ASCII

        * wtf/ASCIICType.h: Added.
        * wtf/DisallowCType.h: Added.

        * kjs/config.h: Include DisallowCType.h.

        * kjs/date_object.cpp:
        (KJS::skipSpacesAndComments):
        (KJS::findMonth):
        (KJS::parseDate):
        * kjs/function.cpp:
        (KJS::decode):
        * kjs/ustring.cpp:
        (KJS::UString::toDouble):
        Use ASCIICType.h functions instead of ctype.h ones.

WebCore:

        Reviewed by Maciej and Geoff (and looked over by Eric).

        - http://bugs.webkit.org/show_bug.cgi?id=15519
          eliminate use of <ctype.h> for processing ASCII

        * ForwardingHeaders/wtf/ASCIICType.h: Added.
        * ForwardingHeaders/wtf/DisallowCType.h: Added.

        * WebCorePrefix.h: Get rid of inclusion of <ctype.h>.
        * config.h: Include DisallowCType.h.

        * css/CSSParser.cpp:
        (WebCore::ParseString::lower):
        * css/CSSPrimitiveValue.cpp:
        (WebCore::isCSSTokenizerIdentifier):
        * css/CSSStyleDeclaration.cpp:
        (WebCore::propertyID):
        * html/HTMLSelectElement.cpp:
        (WebCore::stripLeadingWhiteSpace):
        * html/HTMLTokenizer.cpp:
        (WebCore::tagMatch):
        * loader/FTPDirectoryParser.cpp:
        (WebCore::parseOneFTPLine):
        * loader/TextResourceDecoder.cpp:
        (WebCore::TextResourceDecoder::checkForHeadCharset):
        * platform/DeprecatedCString.cpp:
        (WebCore::DeprecatedCString::lower):
        (WebCore::DeprecatedCString::upper):
        (WebCore::DeprecatedCString::find):
        (WebCore::DeprecatedCString::contains):
        * platform/DeprecatedString.cpp:
        (WebCore::equalCaseInsensitive):
        (WebCore::isCharacterAllowedInBase):
        (WebCore::DeprecatedString::find):
        (WebCore::DeprecatedString::contains):
        (WebCore::toIntegralType):
        * platform/DeprecatedString.h:
        (WebCore::DeprecatedChar::isSpace):
        (WebCore::DeprecatedChar::lower):
        (WebCore::DeprecatedChar::upper):
        * platform/KURL.cpp:
        (WebCore::KURL::parse):
        * platform/StringImpl.cpp:
        (WebCore::isSpace):
        (WebCore::StringImpl::containsOnlyWhitespace):
        (WebCore::StringImpl::isLower):
        (WebCore::StringImpl::lower):
        (WebCore::StringImpl::find):
        (WebCore::StringImpl::reverseFind):
        (WebCore::equalIgnoringCase):
        * platform/TextEncodingRegistry.cpp:
        (WebCore::TextEncodingNameHash::equal):
        (WebCore::TextEncodingNameHash::hash):
        (WebCore::atomicCanonicalTextEncodingName):
        * platform/mac/KeyEventMac.mm:
        (WebCore::keyIdentifierForKeyEvent):
        * platform/win/KeyEventWin.cpp:
        (WebCore::keyIdentifierForWindowsKeyCode):
        * platform/win/PopupMenuWin.cpp:
        (WebCore::isASCIIPrintable):
        (WebCore::PopupWndProc):
        * plugins/win/PluginViewWin.cpp:
        (WebCore::capitalizeRFC822HeaderFieldName):
        * rendering/RenderText.cpp:
        (WebCore::RenderText::widthFromCache):
        Use ASCIICType.h functions instead of ctype.h ones.

WebKit:

        Reviewed by Maciej and Geoff (and looked over by Eric).

        - http://bugs.webkit.org/show_bug.cgi?id=15519
          eliminate use of <ctype.h> for processing ASCII

        * ForwardingHeaders/wtf/ASCIICType.h: Added.
        * ForwardingHeaders/wtf/DisallowCType.h: Added.

        * WebKitPrefix.h: Include DisallowCType.h.

        * Misc/WebNSURLExtras.mm: (-[NSURL _web_URLWithLowercasedScheme]): Use toASCIILower.
        * WebView/WebHTMLView.mm:
        (-[WebHTMLView callWebCoreCommand:]): Use toASCIIUpper.
        (-[WebTextCompleteController filterKeyDown:]): Add a list of specific character codes,
        instead of using ispunct.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26676 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoFix build bustage caused by merge with changes to CSSStyleSelector.cpp
hyatt [Tue, 16 Oct 2007 20:05:50 +0000 (20:05 +0000)]
Fix build bustage caused by merge with changes to CSSStyleSelector.cpp

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26675 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoWebCore:
mitz [Tue, 16 Oct 2007 20:04:42 +0000 (20:04 +0000)]
WebCore:

        Reviewed by Dave Hyatt.

        - fix http://bugs.webkit.org/show_bug.cgi?id=15296
          REGRESSION (r19716-r19719): width not specified for <marquee> tag is causing the alignment issue.

        Test: fast/block/float/marquee-shrink-to-avoid-floats.html

        * rendering/RenderObject.cpp:
        (WebCore::RenderObject::shrinkToAvoidFloats): Changed to return 'true' for <marquee>s with auto width.

LayoutTests:

        Reviewed by Dave Hyatt.

        - test for http://bugs.webkit.org/show_bug.cgi?id=15296
          REGRESSION (r19716-r19719): width not specified for <marquee> tag is causing the alignment issue.

        * fast/block/float/marquee-shrink-to-avoid-floats.html: Added.
        * platform/mac/fast/block: Added.
        * platform/mac/fast/block/float: Added.
        * platform/mac/fast/block/float/marquee-shrink-to-avoid-floats-expected.checksum: Added.
        * platform/mac/fast/block/float/marquee-shrink-to-avoid-floats-expected.png: Added.
        * platform/mac/fast/block/float/marquee-shrink-to-avoid-floats-expected.txt: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26674 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoWebCore:
bdakin [Tue, 16 Oct 2007 19:51:33 +0000 (19:51 +0000)]
WebCore:

        Reviewed by Darin.

        Fix for <rdar://problem/5517118> REGRESSION: 9A570 - Safari renders
        Amazon seller page way too small

        The font size at Amazon was way too small because the calls to
        getComputedStyle and getPropertyValue did not cause the div to
        attach because there was a pending stylesheet. The fix is to call
        updateLayoutIgnorePendingStylesheets() instead of just
        updateLayout().

        * css/CSSComputedStyleDeclaration.cpp:
        (WebCore::CSSComputedStyleDeclaration::getPropertyCSSValue):

LayoutTests:

        Reviewed by Darin.

        Fix for <rdar://problem/5517118> REGRESSION: 9A570 - Safari renders
        Amazon seller page way too small

        * fast/css/pendingStylesheetFontSize.html: Added.
        * platform/mac/fast/css/pendingStylesheetFontSize-expected.checksum: Added.
        * platform/mac/fast/css/pendingStylesheetFontSize-expected.png: Added.
        * platform/mac/fast/css/pendingStylesheetFontSize-expected.txt: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26673 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago - Updated leaks list because our internal bot upgraded it's version of leopar...
kmccullo [Tue, 16 Oct 2007 19:48:28 +0000 (19:48 +0000)]
    - Updated leaks list because our internal bot upgraded it's version of leopard and
        one of the leaks was fixed between that revision and the bot's previous version of
        leopard.

        * Scripts/run-webkit-tests:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26671 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Tim.
kmccullo [Tue, 16 Oct 2007 19:45:42 +0000 (19:45 +0000)]
    Reviewed by Tim.

        - I encapsulated out the server connecton functionality because it is a specific part
        of Drosera that will be platform dependant until the WebScriptDebugServer can be moved
        into WebCore and C++.  But if it is encapsulated out it can be easily replaced on
        Windows.  So the majority of this patch is moving preexisting functionality.

        - Also I removed the prefix header and changed to use config.h because it is more
        standard on how we use config.h in windows and it didn't make much sense to keep
        the prefix header and the config.h

        * Drosera/DebuggerDocument.cpp:
        (DebuggerDocument::DebuggerDocument):
        * Drosera/DebuggerDocument.h:
        * Drosera/config.h:
        * Drosera/mac/DebuggerApplication.mm:
        (-[DebuggerApplication attach:]):
        * Drosera/mac/DebuggerClient.h:
        * Drosera/mac/DebuggerClient.mm:
        (-[DebuggerClient initWithServerConnection:]):
        (-[DebuggerClient dealloc]):
        (-[DebuggerClient windowDidLoad]):
        (-[DebuggerClient windowWillClose:]):
        (-[DebuggerClient webView:didFinishLoadForFrame:]):
        (-[DebuggerClient webView:didReceiveTitle:forFrame:]):
        * Drosera/mac/DebuggerDocumentPlatform.mm:
        (DebuggerDocument::platformPause):
        (DebuggerDocument::platformResume):
        (DebuggerDocument::platformStepInto):
        (DebuggerDocument::platformEvaluateScript):
        (DebuggerDocument::getPlatformCurrentFunctionStack):
        (DebuggerDocument::getPlatformLocalScopeVariableNamesForCallFrame):
        (DebuggerDocument::platformValueForScopeVariableNamed):
        (DebuggerDocument::platformLog):
        * Drosera/mac/Drosera.pch: Removed.
        * Drosera/mac/Drosera.xcodeproj/project.pbxproj:
        * Drosera/mac/ServerConnection.h: Added.
        * Drosera/mac/ServerConnection.mm: Added.
        (-[ServerConnection initWithServerName:]):
        (-[ServerConnection dealloc]):
        (-[ServerConnection setGlobalContext:]):
        (-[ServerConnection pause]):
        (-[ServerConnection resume]):
        (-[ServerConnection stepInto]):
        (-[ServerConnection switchToServerNamed:]):
        (-[ServerConnection applicationTerminating:]):
        (-[ServerConnection serverConnectionDidDie:]):
        (-[ServerConnection webView:didLoadMainResourceForDataSource:]):
        (-[ServerConnection webView:didParseSource:baseLineNumber:fromURL:sourceId:forWebFrame:]):
        (-[ServerConnection webView:failedToParseSource:baseLineNumber:fromURL:withError:forWebFrame:]):
        (-[ServerConnection webView:didEnterCallFrame:sourceId:line:forWebFrame:]):
        (-[ServerConnection webView:willExecuteStatement:sourceId:line:forWebFrame:]):
        (-[ServerConnection webView:willLeaveCallFrame:sourceId:line:forWebFrame:]):
        (-[ServerConnection webView:exceptionWasRaised:sourceId:line:forWebFrame:]):
        (-[ServerConnection currentFrame]):
        (-[ServerConnection webScriptAttributeKeysForScriptObject:]):
        (-[ServerConnection knownServers]):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26668 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago2007-10-15 Jon Honeycutt <jhoneycutt@apple.com>
honeycutt [Tue, 16 Oct 2007 19:36:56 +0000 (19:36 +0000)]
2007-10-15  Jon Honeycutt  <jhoneycutt@apple.com>

        Reviewed by Ollie.

        <rdar://5530789>: REGRESSION(303-310A19): Crash opening .wma files with
        MediaPlayer for the first time

        Failure to setup the stream in PluginView::didReceiveResponse will lead,
        in a full-page plugin, to the main document load being cancelled. This
        is the case with at least two versions of Windows Media Player, which
        cancels the stream and brings up its own "Welcome to Windows Media
        Player" dialog.

        As part of the main document load cancellation, m_pluginView is set to
        null, and the crash came from dereferencing this pointer. This patch
        adds a null check.

        * WebFrame.cpp:
        (WebFrame::finishedLoading): Fix some typos
        (WebFrame::committedLoad): Added a null check

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26667 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Get transform-origin(x/y) mapped into the front end RenderStyle. Adds a...
hyatt [Tue, 16 Oct 2007 19:06:45 +0000 (19:06 +0000)]
    Get transform-origin(x/y) mapped into the front end RenderStyle.  Adds a new StyleTransformData struct that
        holds transform-origin (and that will also eventually hold the parsed transform operations as well).

        Reviewed by Beth

        * css/CSSStyleSelector.cpp:
        (WebCore::CSSStyleSelector::applyProperty):
        * rendering/RenderStyle.cpp:
        (WebCore::StyleTransformData::StyleTransformData):
        (WebCore::StyleTransformData::operator==):
        (WebCore::StyleRareNonInheritedData::StyleRareNonInheritedData):
        (WebCore::StyleRareNonInheritedData::operator==):
        (WebCore::RenderStyle::RenderStyle):
        (WebCore::RenderStyle::diff):
        * rendering/RenderStyle.h:
        (WebCore::StyleTransformData::operator!=):
        (WebCore::RenderStyle::transformOriginX):
        (WebCore::RenderStyle::transformOriginY):
        (WebCore::RenderStyle::setTransformOriginX):
        (WebCore::RenderStyle::setTransformOriginY):
        (WebCore::RenderStyle::initialTransformOriginX):
        (WebCore::RenderStyle::initialTransformOriginY):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26666 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoWebCore:
adele [Tue, 16 Oct 2007 18:30:55 +0000 (18:30 +0000)]
WebCore:

        Reviewed by Alice.

        Fix for <rdar://problem/5538793> REGRESSION: Failing layout test: fast/forms/select-change-listbox-to-popup.html (due to landing feature branch)

        * html/HTMLSelectElement.cpp: (WebCore::HTMLSelectElement::parseMappedAttribute):  When switching between types of select elements, call setRecalcListItems
          since recalcListItems updates the default selection in different ways for the different controls.

LayoutTests:

        Reviewed by Alice.

        Updating with correct results for:
        <rdar://problem/5538793> REGRESSION: Failing layout test: fast/forms/select-change-listbox-to-popup.html (due to landing feature branch)

        * platform/mac/fast/forms/select-change-listbox-to-popup-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26665 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago rubber-stamped by Darin.
aliceli1 [Tue, 16 Oct 2007 18:22:47 +0000 (18:22 +0000)]
    rubber-stamped by Darin.

        flipping my previous fix around so that mac is special-cased instead of win.

        * dom/Document.cpp:
        (WebCore::Document::defaultEventHandler):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26664 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Clean up transform parsing. Fix a crash, ditch the extra parse context class...
hyatt [Tue, 16 Oct 2007 18:21:09 +0000 (18:21 +0000)]
    Clean up transform parsing.  Fix a crash, ditch the extra parse context class, and use RefPtr/PassRefPtr more
        to simplify the code.

        Reviewed by aroben

        * css/CSSParser.cpp:
        (WebCore::CSSParser::parseValue):
        (WebCore::CSSParser::parseTransform):
        * css/CSSParser.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26663 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Adele.
aliceli1 [Tue, 16 Oct 2007 17:57:00 +0000 (17:57 +0000)]
    Reviewed by Adele.

        fixed <rdar://5085596> Accesskeys don't work

        * dom/Document.cpp:
        (WebCore::Document::defaultEventHandler):
        accesskey modifier is platform-depedent.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26662 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Adam Roben
sullivan [Tue, 16 Oct 2007 16:55:13 +0000 (16:55 +0000)]
    Reviewed by Adam Roben

        Cleaned up localizable strings

        * English.lproj/Localizable.strings:
        updated
        * StringsNotToBeLocalized.txt:
        updated
        * WebKit.xcodeproj/project.pbxproj:
        StringsNotToBeLocalized.txt recently moved but project file wasn't updated to match; now it is

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26661 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoRemoved conflict markers.
ddkilzer [Tue, 16 Oct 2007 11:19:54 +0000 (11:19 +0000)]
Removed conflict markers.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26656 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Maciej.
ap [Tue, 16 Oct 2007 07:04:40 +0000 (07:04 +0000)]
    Reviewed by Maciej.

        http://bugs.webkit.org/show_bug.cgi?id=15506
        editing/pasteboard/paste-into-anchor-text.html accesses an external resource

        Removed the external resource reference and made the test dump as text.

        * editing/pasteboard/paste-into-anchor-text-expected.txt: Copied from platform/mac/editing/pasteboard/paste-into-anchor-text-expected.txt.
        * editing/pasteboard/paste-into-anchor-text.html:
        * platform/mac/editing/pasteboard/paste-into-anchor-text-expected.checksum: Removed.
        * platform/mac/editing/pasteboard/paste-into-anchor-text-expected.png: Removed.
        * platform/mac/editing/pasteboard/paste-into-anchor-text-expected.txt: Removed.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26654 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoWebKitTools:
ddkilzer [Tue, 16 Oct 2007 05:53:03 +0000 (05:53 +0000)]
WebKitTools:

        Reviewed by Eric.

        - fix http://bugs.webkit.org/show_bug.cgi?id=15002
          Script to automatically search nightly builds for regressions (bisect-builds)

        * Scripts/bisect-builds: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26653 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago2007-10-15 Mark Rowe <mrowe@apple.com>
bdash [Tue, 16 Oct 2007 05:17:06 +0000 (05:17 +0000)]
2007-10-15  Mark Rowe  <mrowe@apple.com>

        Reviewed by John.

        <rdar://problem/5494040> Reproducible assertion failure in WebCore::IconDatabase::iconForPageURL

        A race condition between icon database import and Safari asking for an icon for the empty URL
        was resulting in a PageURLRecord being created for a URL that can never be retained.  By enforcing
        that PageURLRecord's cannot be created for the empty URL we can prevent the assertion from failing.

        * loader/icon/IconDatabase.cpp:
        (WebCore::IconDatabase::iconForPageURL):
        (WebCore::IconDatabase::getOrCreatePageURLRecord): Bail out early if the URL is empty.
        (WebCore::IconDatabase::performURLImport): Don't create the PageURLRecord if the URL is empty.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26652 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago2007-10-15 Jon Honeycutt <jhoneycutt@apple.com>
honeycutt [Tue, 16 Oct 2007 04:51:40 +0000 (04:51 +0000)]
2007-10-15  Jon Honeycutt  <jhoneycutt@apple.com>

        Reviewed by Ollie.

        Use OwnPtr for m_deliveryData, and fix a bug where we were memmove()ing
        over m_deliveryData instead of its data buffer

        * plugins/win/PluginStreamWin.cpp:
        (WebCore::PluginStreamWin::~PluginStreamWin): Don't delete
        m_deliveryData
        (WebCore::PluginStreamWin::deliverData): Fix memmove() call
        (WebCore::PluginStreamWin::didReceiveData): Assign using OwnPtr::set()
        * plugins/win/PluginStreamWin.h: Use OwnPtr

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26651 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Oliver.
sfalken [Tue, 16 Oct 2007 03:39:05 +0000 (03:39 +0000)]
    Reviewed by Oliver.

        Add missing null check.

        * platform/network/cf/AuthenticationCF.cpp:
        (WebCore::core):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26646 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Darin Adler.
kdecker [Tue, 16 Oct 2007 03:30:38 +0000 (03:30 +0000)]
    Reviewed by Darin Adler.

        Fixed: <rdar://problem/5520541> REGRESSION: Broken image when forwarding certain email on Tiger

        * WebCoreSupport/WebFrameBridge.mm: The problem was that we were loading Mail's WebKit plug-in
        too soon, which borked some necessary housekeeping on behalf of Mail. The fix is to add a quirk
        that treats Tiger Mail's WebKit plug-in like a Netscape plug-in, thus ensuring the plug-in will
        load during first layout and not attach time. For this plug-in, loading at first layout is
        expected and is consistent with Safari 2 behavior.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26645 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago2007-10-15 Mark Rowe <mrowe@apple.com>
bdash [Tue, 16 Oct 2007 02:34:04 +0000 (02:34 +0000)]
2007-10-15  Mark Rowe  <mrowe@apple.com>

        Fix the no-SVG build again.  Add new transform CSS properties to the switch statements.

        * css/CSSComputedStyleDeclaration.cpp:
        (WebCore::CSSComputedStyleDeclaration::getPropertyCSSValue):
        * css/CSSStyleSelector.cpp:
        (WebCore::CSSStyleSelector::applyProperty):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26644 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoJavaScriptCore:
mjs [Tue, 16 Oct 2007 01:30:22 +0000 (01:30 +0000)]
JavaScriptCore:

        Reviewed by Darin.

        - fixes for "New JavaScript benchmark"
        http://bugs.webkit.org/show_bug.cgi?id=15515

        * kjs/testkjs.cpp:
        (TestFunctionImp::callAsFunction): Implement "load" for compatibility
        with SpiderMonkey.
        (TestFunctionImp::): ditto
        (doIt): ditto
        (kjsmain): Drop useless --> from output.

SunSpider:

        Reviewed by Darin.

        - New JavaScript benchmark
        http://bugs.webkit.org/show_bug.cgi?id=15515

        A start on a new JavaScript benchmark, with standalone and
        browser-hosted drivers.

        The test content is still incomplete and the drivers could use
        some more features but there is enough here to make a good start.

        The drivers compute a 95% confidence interval on the mean for the
        whole test, each category, and each individual test to make it
        easier to tell whether differences are statistically
        significant. The confidence interval can be narrowed by running
        the test more times.

        * TODO: Added. Remaining things that need to be done.
        * sunspider: Added. Perl script that acts as the standalone test driver.
        * resources/sunspider-standalone-driver.js: Added. JavaScript part
        of standalone test driver.
        * resources/sunspider-analyze-results.js: Added. JavaScript statistical
        analysis code.
        * resources/TEMPLATE.html: Added. Template for browser-hosted tests.
        * make-hosted: Added. Script to generate browser-hosted tests.
        * hosted/sunspider.html: Added. Start page for browser-hosted test.
        * hosted/sunspider-driver.html: Added. Driver for browser-hosted tests.
        * hosted/sunspider-results.html: Added. Results page for browser-hosted
        tests.
        * hosted/sunspider-record-result.js: Added. Helper file for browser-hosted
        tests.
        * tests/LIST: Added. List of tests to use.
        * tests/bitops-3bit-bits-in-byte.js: Added. Some initial test content.
        * tests/bitops-bits-in-byte.js: Added. Ditto.
        * tests/bitops-bitwise-and.js: Added. Ditto.
        * tests/math-cordic.js: Added. Ditto.

WebKitTools:

        Reviewed by Darin.

        - New JavaScript benchmark
        http://bugs.webkit.org/show_bug.cgi?id=15515

        * Scripts/run-sunspider: Added. Wrapper to run sunspider on the
        current development or release build of JavaScriptCore.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26639 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Mitz.
darin [Mon, 15 Oct 2007 23:54:32 +0000 (23:54 +0000)]
    Reviewed by Mitz.

        - http://bugs.webkit.org/show_bug.cgi?id=15518
          Remove the unused attemptFontSubstitution feature and do some
          cleanup of glyphDataForCharacter.

        * platform/Font.h:
        * platform/Font.cpp:
        (WebCore::WidthIterator::advance): Eliminated uneeded attemptFontSubstitution
        and cluster parameters to the glyphDataForCharacter function.
        (WebCore::Font::glyphDataForCharacter): Removed the cluster parameter and
        attemptFontSubstitution parameter. Removed the isUpper check before calling
        toUpper in the small caps code path. Split up the inner loop into separate
        copies for non-small-caps and small caps. Broke out the system fallback
        page handling so it's not inside the loop. Changed system fallback so it uses
        the character and breaks it up into UTF-16 as needed, instead of using a
        passed-in "character cluster".

        * platform/TextStyle.h:
        (WebCore::TextStyle::TextStyle): Removed attemptFontSubstitution.
        (WebCore::TextStyle::applyWordRounding): Ditto.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26638 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Fixed layout test for reals. Changed \r\n to \n, fixing mass confusion.
ggaren [Mon, 15 Oct 2007 23:52:50 +0000 (23:52 +0000)]
    Fixed layout test for reals. Changed \r\n to \n, fixing mass confusion.

        * fast/js/do-while-expression-value-expected.txt:
        * fast/js/do-while-expression-value.html:
        * fast/js/nested-function-scope.html:
        * fast/js/while-expression-value-expected.txt:
        * fast/js/while-expression-value.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26637 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago2007-10-15 Jon Honeycutt <jhoneycutt@apple.com>
honeycutt [Mon, 15 Oct 2007 23:18:46 +0000 (23:18 +0000)]
2007-10-15  Jon Honeycutt  <jhoneycutt@apple.com>

        Reviewed by Ollie.

        Remove an operator precedence warning on Windows

        * editing/markup.cpp:
        (WebCore::escapeContentText): Use (a) | (b)
        (WebCore::appendEscapedContent): Use (a) | (b)

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26636 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoWebCore:
kmccullo [Mon, 15 Oct 2007 22:59:48 +0000 (22:59 +0000)]
WebCore:

        Reviewed by Darin.

        - <rdar://problem/5238818> window.resizeTo doesn't restrict the resized window to the size of the screen
        - Now we take the doc into account when resizing.
        - Also I found an issue where we would resize to outside the window, because the resize would be smaller than the window size but the location of the window would make the resize go off-screen.  Now we move the window back into the screen.

        * bindings/js/kjs_window.cpp:
        (KJS::adjustWindowRect):
        (KJS::WindowFunc::callAsFunction):

LayoutTests:

        Reviewed by Darin.

        - <rdar://problem/5238818> window.resizeTo doesn't restrict the resized window to the size of the screen
        - Now we take the dock into account when resizing.
        - Also I found an issue where we would resize to outside the window, because the resize would be smaller than the window size but the location of the window would make the resize go off-screen.  Now we move the window back into the screen.

        * fast/dom/Window/window-resize-expected.txt:
        * fast/dom/Window/window-resize.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26635 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago2007-10-15 Mark Rowe <mrowe@apple.com>
bdash [Mon, 15 Oct 2007 22:34:27 +0000 (22:34 +0000)]
2007-10-15  Mark Rowe  <mrowe@apple.com>

        Reviewed by Sam.

        Fix 'run-webkit-tests --qt' complaining about the --qt argument.

        * Scripts/webkitdirs.pm:
        - Change checkArgv to remove the options from @ARGV to prevent
        them from interfering with further option processing.
        - Fix logic error in determineIsQt that would prevent it from bailing out early.
        - Change isOSX to mean OS X and not Qt or Gtk.  Most of our uses of isOSX assume that this is the meaning
        already, so this change fixes several broken areas of the scripts including 'run-webkit-tests --qt' on a Mac
        incorrectly using the Mac results, and 'run-webkit-tests --gtk' attempting to build the Mac DumpRenderTree.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26633 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago2007-10-15 Jon Honeycutt <jhoneycutt@apple.com>
honeycutt [Mon, 15 Oct 2007 22:27:31 +0000 (22:27 +0000)]
2007-10-15  Jon Honeycutt  <jhoneycutt@apple.com>

        Reviewed by Anders.

        <rdar://5510700> Repro crash loading embedded Windows Media Player
        content

        Anders noticed that the older Windows Media Player plugin (npdsplay.dll)
        will crash if it receives its resource requests out of order, whereas we
        were immediately fulfilling untargeted, non-JavaScript requests and
        scheduling targeted or JavaScript resource requests. Implement his
        suggested fix, which is to schedule PluginRequests for all resource
        requests

        * plugins/win/PluginViewWin.cpp:
        (WebCore::getString):
        (WebCore::PluginViewWin::performRequest): If this non-JavaScript request
        has no target, create a stream for it
        (WebCore::PluginViewWin::load): Schedule PluginRequests for all resource
        requests

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26632 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago2007-10-15 Mark Rowe <mrowe@apple.com>
bdash [Mon, 15 Oct 2007 22:12:26 +0000 (22:12 +0000)]
2007-10-15  Mark Rowe  <mrowe@apple.com>

        Gtk and Qt build fix.  Add new .cpp file to project.

        * WebCore.pro:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26631 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago * platform/mac/FontDataMac.mm:
sullivan [Mon, 15 Oct 2007 21:59:48 +0000 (21:59 +0000)]
    * platform/mac/FontDataMac.mm:
        D'oh! Added missing #import to fix build

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26630 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Geoff Garen
sullivan [Mon, 15 Oct 2007 21:42:57 +0000 (21:42 +0000)]
    Reviewed by Geoff Garen

        Replaced NS_DURING/NS_HANDLER with @try/@catch throughout WebKit

        I made the following changes:
        - replaced NS_DURING with @try, and added opening brace if there wasn't one
        - replaced NS_HANDLER with @catch (NSException *localException), and added braces if there weren't any
        - removed NS_ENDHANDLER, and added a closing brace if there wasn't one
        - in a couple of places, fixed indentation therein

        * Misc/WebIconDatabase.mm:
        (objectFromPathForKey):
        * WebView/WebHTMLView.mm:
        (-[WebHTMLView drawSingleRect:]):
        (-[WebHTMLView beginDocument]):
        (-[WebHTMLView deleteToMark:]):
        * WebView/WebView.mm:
        (-[WebView initWithCoder:]):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26629 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Geoff
sullivan [Mon, 15 Oct 2007 21:41:36 +0000 (21:41 +0000)]
    Reviewed by Geoff

        * platform/mac/FontDataMac.mm:
        (WebCore::FontData::smallCapsFontData):
        replace NS_DURING/NS_HANDLER with WebCore-style BEGIN/END_BLOCK_OBJC_EXCEPTIONS

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26628 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Changed line ending style of results file to fix layout test that was
ggaren [Mon, 15 Oct 2007 21:25:07 +0000 (21:25 +0000)]
    Changed line ending style of results file to fix layout test that was
        failing locally on my machine.

        * fast/js/nested-function-scope-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26627 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago The CSS WG has been discussing the concept of rotation. In the latest draft...
hyatt [Mon, 15 Oct 2007 21:15:48 +0000 (21:15 +0000)]
    The CSS WG has been discussing the concept of rotation.  In the latest draft specification, two properties
        have been introduced: rotation and rotation-point.  I believe these properties should be generalized to
        allow for arbitrary transformations (e.g., translation, skew, rotate, scale).

        This patch adds support for the back-end parsing of two new CSS properties: transform and transform-origin.
        They are directly analogous to the properties specified by the CSS WG, except that they are designed to allow
        for arbitrary affine transforms rather than just rotation.

        Reviewed by Beth Dakin

        * WebCore.xcodeproj/project.pbxproj:
        * WebCore.vcproj/WebCore.vcproj:
        * css/CSSComputedStyleDeclaration.cpp:
        (WebCore::CSSComputedStyleDeclaration::getPropertyCSSValue):
        * css/CSSParser.cpp:
        (WebCore::CSSParser::parseValue):
        (WebCore::TransformParseContext:::m_list):
        (WebCore::TransformParseContext::list):
        (WebCore::TransformParseContext::failed):
        (WebCore::TransformParseContext::addValue):
        (WebCore::TransformOperationInfo::TransformOperationInfo):
        (WebCore::TransformOperationInfo::type):
        (WebCore::TransformOperationInfo::argCount):
        (WebCore::TransformOperationInfo::unit):
        (WebCore::TransformOperationInfo::unknown):
        (WebCore::TransformOperationInfo::hasCorrectArgCount):
        (WebCore::CSSParser::parseTransform):
        (WebCore::CSSParser::parseTransformOrigin):
        * css/CSSParser.h:
        * css/CSSPropertyNames.in:
        * css/CSSTransformValue.cpp: Added.
        (WebCore::CSSTransformValue::CSSTransformValue):
        (WebCore::CSSTransformValue::~CSSTransformValue):
        (WebCore::CSSTransformValue::addValue):
        (WebCore::CSSTransformValue::cssText):
        * css/CSSTransformValue.h: Added.
        (WebCore::CSSTransformValue::):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26626 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Removed unnecessary #include.
ggaren [Mon, 15 Oct 2007 20:58:28 +0000 (20:58 +0000)]
    Removed unnecessary #include.

        * API/JSObjectRef.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26625 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Double-reverse build fix. My tree was out of date.
ggaren [Mon, 15 Oct 2007 20:57:12 +0000 (20:57 +0000)]
    Double-reverse build fix. My tree was out of date.

        * kjs/nodes.cpp:
        (NumberNode::evaluate):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26624 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago2007-10-15 Mark Rowe <mrowe@apple.com>
bdash [Mon, 15 Oct 2007 20:53:35 +0000 (20:53 +0000)]
2007-10-15  Mark Rowe  <mrowe@apple.com>

        Reviewed by Oliver.

        Fix the no-SVG build.

        * DerivedSources.make: Use the correct file as input to generate CSSPropertyNames.h.
        * rendering/RenderObject.cpp:
        (WebCore::objectIsRelayoutBoundary):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26623 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Build fix.
ggaren [Mon, 15 Oct 2007 20:47:23 +0000 (20:47 +0000)]
    Build fix.

        * kjs/nodes.cpp:
        (NumberNode::evaluate):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26622 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoJavaScriptCore:
ggaren [Mon, 15 Oct 2007 20:44:09 +0000 (20:44 +0000)]
JavaScriptCore:

        Reviewed by Darin Adler.

        Removed surprising self-named "hack" that made nested functions
        available as named properties of their containing functions, and placed
        containing function objects in the scope chains of nested functions.

        There were a few reasons to remove this "hack:"

        1. It contradicted FF, IE, and the ECMA spec.

        2. It incurred a performance penalty, since merely parsing a function
        required parsing its body for nested functions (and so on).

        3. SVN history contains no explanation for why it was added. It was just
        legacy code in a large merge a long, long time ago.

        [ Patch broken off from http://bugs.webkit.org/show_bug.cgi?id=14868 ]

        * kjs/nodes.cpp:
        (FuncDeclNode::processFuncDecl):

LayoutTests:

        Reviewed by Darin Adler.

        Removed surprising self-named "hack" that made nested functions
        available as named properties of their containing functions, and placed
        containing function objects in the scope chains of nested functions.

        [ Patch broken off from http://bugs.webkit.org/show_bug.cgi?id=14868 ]

        Changed this test to reflect correct behavior:

        * fast/js/kde/function-expected.txt:
        * fast/js/kde/resources/function.js: (Also removed tab characters.)

        Added this test to flesh out the behavior more:

        * fast/js/nested-function-scope-expected.txt: Added.
        * fast/js/nested-function-scope.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26621 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Darin Adler.
ggaren [Mon, 15 Oct 2007 20:41:39 +0000 (20:41 +0000)]
    Reviewed by Darin Adler.

        Removed the concept of AnonymousCode. It was unused, and it doesn't
        exist in the ECMA spec.

        [ Patch broken off from http://bugs.webkit.org/show_bug.cgi?id=14868 ]

        * kjs/Context.cpp:
        (KJS::Context::Context):
        * kjs/function.h:
        (KJS::):
        * kjs/nodes.cpp:
        (ReturnNode::execute):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26620 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoJavaScriptCore:
ggaren [Mon, 15 Oct 2007 20:40:50 +0000 (20:40 +0000)]
JavaScriptCore:

        Reviewed by Darin Adler.

        Made function parameters DontDelete. This matches FF and the vague
        description in ECMA 10.1.3. It's also required in order to make
        symbol table based lookup of function parameters valid. (If the
        parameters aren't DontDelete, you can't guarantee that you'll find
        them later in the symbol table.)

        [ Patch broken off from http://bugs.webkit.org/show_bug.cgi?id=14868 ]

        * kjs/function.cpp:
        (KJS::FunctionImp::passInParameters):

LayoutTests:

        Committed revision 26618.

        Made function parameters DontDelete.

        * fast/js/delete-function-parameter-expected.txt: Added.
        * fast/js/delete-function-parameter.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26619 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Maciej Stachowiak.
ggaren [Mon, 15 Oct 2007 20:39:10 +0000 (20:39 +0000)]
    Reviewed by Maciej Stachowiak.

        Some Vector optimizations. These are especially important when using
        Vector as a stack for implementing recursive algorithms iteratively.

        [ Broken off from http://bugs.webkit.org/show_bug.cgi?id=14868 ]

        1. Added shrink(), which is a version of resize() that you can call
        to save a branch / improve code generation and inlining when you know
        that the vector is not getting bigger.

        2. Changed subclassing relationship in VectorBuffer to remove a call to
        fastFree() in the destructor for the inlineCapacity != 0 template
        specialization. This brings inline Vectors one step closer to true
        stack-allocated arrays.

        Also changed abort() to CRASH(), since the latter works better.

        * wtf/Vector.h:
        (WTF::VectorBufferBase::allocateBuffer):
        (WTF::VectorBufferBase::deallocateBuffer):
        (WTF::VectorBufferBase::VectorBufferBase):
        (WTF::VectorBufferBase::~VectorBufferBase):
        (WTF::):
        (WTF::VectorBuffer::VectorBuffer):
        (WTF::VectorBuffer::~VectorBuffer):
        (WTF::VectorBuffer::deallocateBuffer):
        (WTF::VectorBuffer::releaseBuffer):
        (WTF::Vector::clear):
        (WTF::Vector::removeLast):
        (WTF::::operator):
        (WTF::::fill):
        (WTF::::shrink):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26618 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoJavaScriptCore:
ggaren [Mon, 15 Oct 2007 20:36:57 +0000 (20:36 +0000)]
JavaScriptCore:

        Reviewed by Maciej Stachowiak.

        Fixed http://bugs.webkit.org/show_bug.cgi?id=15490
        Iteration statements sometimes incorrectly evaluate to the empty value
        (KDE r670547).

        [ Broken off from http://bugs.webkit.org/show_bug.cgi?id=14868 ]

        This patch is a merge of KDE r670547, with substantial modification
        for performance.

        It fixes do-while statements to evaluate to a value. (They used
        to evaluate to the empty value in all cases.)

        It also fixes SourceElementsNode to maintain the value of abnormal
        completions like "break" and "continue."

        It also re-works the main execution loop in SourceElementsNode so that
        it (1) makes a little more sense and (2) avoids unnecessary work. This
        is a .28% speedup on command-line JS iBench.

        * kjs/nodes.cpp:
        (DoWhileNode::execute):
        (SourceElementsNode::execute):

LayoutTests:

        Reviewed by Maciej Stachowiak.

        Layout tests for http://bugs.webkit.org/show_bug.cgi?id=15490
        Iteration statements sometimes incorrectly evaluate to the empty value
        (KDE r670547)

        * fast/js/do-while-expression-value-expected.txt: Added.
        * fast/js/do-while-expression-value.html: Added.
        * fast/js/while-expression-value-expected.txt: Added.
        * fast/js/while-expression-value.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26617 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoWebCore:
darin [Mon, 15 Oct 2007 19:02:41 +0000 (19:02 +0000)]
WebCore:

        Reviewed by Sam.

        - fix http://bugs.webkit.org/show_bug.cgi?id=15520
          alert(undefined) should say "undefined" in the alert

        Test: fast/dom/Window/alert-undefined.html

        * page/DOMWindow.idl: Remove ConvertUndefinedOrNullToNullString from alert, confirm, and prompt.
        It turns out that the default behavior is what the other browsers do (browsers tested: IE 7,
        Firefox 2).

LayoutTests:

        Reviewed by Sam.

        - fix http://bugs.webkit.org/show_bug.cgi?id=15520
          alert(undefined) should say "undefined" in the alert

        * fast/dom/Window/alert-undefined.html: Added tests of null and both confirm and prompt.
        * fast/dom/Window/alert-undefined-expected.txt: Updated.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26616 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years ago Reviewed by Sam.
darin [Mon, 15 Oct 2007 18:47:56 +0000 (18:47 +0000)]
    Reviewed by Sam.

        - added logging of window.prompt and window.confirm

        * DumpRenderTree/mac/UIDelegate.mm:
        (-[UIDelegate webView:runJavaScriptAlertPanelWithMessage:initiatedByFrame:]):
        Implement the recommended delegate rather than the deprecated one.
        (-[UIDelegate webView:runJavaScriptConfirmPanelWithMessage:initiatedByFrame:]):
        Added.
        (-[UIDelegate webView:runJavaScriptTextInputPanelWithPrompt:defaultText:initiatedByFrame:]):
        Added.

        * DumpRenderTree/win/UIDelegate.h:
        * DumpRenderTree/win/UIDelegate.cpp:
        (UIDelegate::runJavaScriptAlertPanelWithMessage): Got rid of unnecessary check for null string.
        (UIDelegate::runJavaScriptConfirmPanelWithMessage): Added.
        (UIDelegate::runJavaScriptTextInputPanelWithPrompt): Added.
        (UIDelegate::webViewAddMessageToConsole): Got rid of unnecessary check for null string.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26615 268f45cc-cd09-0410-ab3c-d52691b4dbfc

13 years agoWebCore:
adele [Mon, 15 Oct 2007 18:39:05 +0000 (18:39 +0000)]
WebCore:

        Reviewed by Darin.

        Fix for http://bugs.webkit.org/show_bug.cgi?id=15252
        <rdar://problem/5498184> REGRESSION: <select multiple> doesn't scroll to top when old options are removed and new ones are added, leaving listbox empty-looking

        * rendering/RenderListBox.cpp: (WebCore::RenderListBox::calcHeight): If the scrollbar is disabled, make sure the scroll offset gets reset to 0.  In general,
          we don't want to unnecessarily adjust the scroll offset, but in this case, there won't be an obvious way for the user to adjust the scroller position once it's disabled.

LayoutTests:

        Reviewed by Darin.

        Test for http://bugs.webkit.org/show_bug.cgi?id=15252
        <rdar://problem/5498184> REGRESSION: <select multiple> doesn't scroll to top when old options are removed and new ones are added, leaving listbox empty-looking

        * fast/forms/listbox-scroll-after-options-removed-expected.txt: Added.
        * fast/forms/listbox-scroll-after-options-removed.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@26614 268f45cc-cd09-0410-ab3c-d52691b4dbfc