Add the .cpp files to the TestNetscapePlugIn target.

        * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28459 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Geoff.

        - Wait until local variable data is fully constructed before notifying the debugger of entering
        or leaving a call frame.

        * kjs/function.cpp:
        (KJS::FunctionImp::callAsFunction):
        * kjs/nodes.cpp:
        (KJS::FunctionBodyNode::execute):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28458 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Geoff.

        Rename the TestNetscapePlugIn .c files to be .cpp.

        * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
        * DumpRenderTree/TestNetscapePlugIn.subproj/PluginObject.c: Removed.
        * DumpRenderTree/TestNetscapePlugIn.subproj/PluginObject.cpp: Copied from DumpRenderTree/TestNetscapePlugIn.subproj/PluginObject.c.
        * DumpRenderTree/TestNetscapePlugIn.subproj/TestObject.c: Removed.
        * DumpRenderTree/TestNetscapePlugIn.subproj/TestObject.cpp: Copied from DumpRenderTree/TestNetscapePlugIn.subproj/TestObject.c.
        * DumpRenderTree/TestNetscapePlugIn.subproj/main.c: Removed.
        * DumpRenderTree/TestNetscapePlugIn.subproj/main.cpp: Copied from DumpRenderTree/TestNetscapePlugIn.subproj/main.c.
        * DumpRenderTree/win/TestNetscapePlugin/TestNetscapePlugin.vcproj:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28457 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Fix 64-bit debug build.

Reviewed by Kevin Decker.

* platform/text/mac/TextCodecMac.cpp:
(WebCore::TextCodecMac::decode): OSStatus is a long for 32-bit but int on 64-bit,
so we need to cast it to be able to use a single format specifier on both.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28456 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Build fix for GCC 4.2. Cast via a union to avoid strict-aliasing issues.

Reviewed by Oliver.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28455 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    * platform/win/Skipped: Add another test failing due to fonts.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28454 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Fix case of keypresses from the Windows implementation of eventSender.keyDown

        This fixes several regression tests.

        Reviewed by Alice.

        * DumpRenderTree/win/EventSender.cpp:
        (keyDownCallback): Virtual keycodes for ASCII characters are always
        uppercase, so we need to check the case of the original character
        passed in to eventSender.keyDown.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28453 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Corrected a Radar number in my previous ChangeLog entry.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28452 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Hyatt.

        <rdar://problem/5602071> Safari displays each pixel of 300 dpi US Patent TIFF Images, making them extremely large

        * loader/FrameLoader.cpp:
        (WebCore::FrameLoader::shouldUsePlugin): Allow plug-ins (with the exception of QuickTime) to
        takeover TIFF once again.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28451 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Add another Inspector screenshot for an upcoming blog post.

        * blog-files/InspectorDatabaseBrowser.png: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28450 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Fix testkjs in 64-bit.

When built for 64-bit the TCMalloc spin lock uses pthread mutexes rather than a custom spin lock
implemented in assembly.  If we fail to initialize the pthread mutex, attempts to lock or unlock
it will fail and trigger a call to abort.

Reviewed by Darin.

* wtf/FastMalloc.cpp: Initialize the spin lock so that we can later lock and unlock it.
* wtf/TCSpinLock.h: Add an Init method to the optimised spin lock.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28449 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-12-05  Luca Bruno  <lethalman88@gmail.com>

        Reviewed by Alp Toker.

        http://bugs.webkit.org/show_bug.cgi?id=16290
        [GTK] Pasting clipboard in rich text.

        * platform/gtk/PasteboardGtk.cpp:
        (Pasteboard::documentFragment): Implement rich text paste support.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28448 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Hyatt.

        Reverting fix for <rdar://problem/5499918> REGRESSION: insertion point in input field with custom border cuts holes in focus ring interior edges

        We need to just fix the underlying focus ring repaint problem.

        * rendering/RenderTextControl.cpp: (WebCore::RenderTextControl::setStyle):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28447 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

        Reviewed by Darin.

        - <rdar://5621435>
        - Security Fix. Instead of having it off by default, WebKit now must
        explicitly turn off local-resource restriction when needed for backwards
        compatibility reasons.

        * loader/FrameLoader.cpp:

WebKit/mac:

        Reviewed by Darin.

        - <rdar://5621435>
        - Security Fix. Instead of having it off by default, WebKit now must
        explicitly turn off local-resource restriction when needed for backwards
        coimpatibility reasons.

        * WebView/WebView.mm:
        (-[WebView _commonInitializationWithFrameName:groupName:]):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28446 268f45cc-cd09-0410-ab3c-d52691b4dbfc

     Expose WebView::setProhibitsMainFrameScrolling through IWebViewPrivate

         Reviewed by Steve.

         * Interfaces/IWebViewPrivate.idl: Added declaration.
         * WebInspectorClient.cpp:
         (WebInspectorClient::createPage): Updated for change to
         setProhibitsMainFrameScrolling signature.
         * WebView.cpp:
         (WebView::setProhibitsMainFrameScrolling): Updated and robustified.
         * WebView.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28445 268f45cc-cd09-0410-ab3c-d52691b4dbfc

     Set the menu item identifier for all menu items

         This allows WebKit clients to distinguish between submenus in the
         context menu.

         Reviewed by Darin.

         No test possible.

         * platform/win/ContextMenuItemWin.cpp:
         (WebCore::ContextMenuItem::ContextMenuItem):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28444 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Learn from Tim's mistakes

        * DumpRenderTree/win/DumpRenderTree.cpp:
        (runTest): Reset the authorAndUserStylesEnabled preference for each
        test.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28443 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

        Reviewed by Darin Adler.

        - WebCore part of fixing <rdar://problem/5111082> Flash popup video ad doesn't close when clicked at http://www.firestonecompleteautocare.com/

        * manual-tests/nested-plug-ins.html: Added.
        * manual-tests/resources/nested-plug-ins-inner-frame.html: Added.
        * manual-tests/resources/nested-plug-ins-outer-frame.html: Added.
        * platform/ScrollView.h:
        * platform/Widget.h:
        (WebCore::Widget::attachToWindow): Added. Tells the widget that it is connected
        via visible ScrollViews to the root ScrollView of the WebView.
        (WebCore::Widget::detachFromWindow): Added. Tells the widget that it is no longer
        connected via visible ScrollViews to the root ScrollView of the WebView.
        * platform/win/ScrollViewWin.cpp:
        (WebCore::ScrollView::ScrollViewPrivate::ScrollViewPrivate): Added
        m_visible and m_attachedToWindow members.
        (WebCore::ScrollView::attachToWindow): Added. Attaches all children if this
        view is visible.
        (WebCore::ScrollView::detachFromWindow): Added. Detaches all children if this
        view is visible (otherwise they would be detached already).
        (WebCore::ScrollView::show): Added. Attaches all children if this view
        is attached.
        (WebCore::ScrollView::hide): Added. Detaches all children if this view
        is attached (otherwise they would be detached already).
        (WebCore::ScrollView::isAttachedToWindow): Added.
        * platform/win/WidgetWin.cpp:
        (WebCore::Widget::setParent): Added calls to attachToWindow() and detachFromWindow()
        depending on whether the parent is set and attached or not.
        * plugins/win/PluginViewWin.cpp:
        (WebCore::PluginViewWin::setFocus):
        (WebCore::PluginViewWin::show): Changed to show the plug-in only if
        this view is attached.
        (WebCore::PluginViewWin::hide): Changed to hide the plug-in only if this
        view is attached (otherwise it would be hidden already).
        (WebCore::PluginViewWin::attachToWindow): Added. Shows the plug-in if this view
        is visible.
        (WebCore::PluginViewWin::detachFromWindow): Added. Hides the plug-in if this view
        is visible (otherwise it would be hidden already).
        (WebCore::PluginViewWin::PluginViewWin):
        * plugins/win/PluginViewWin.h: Added m_attachedToWindow member.

WebKit/win:

        Reviewed by Darin Adler.

        - WebKit/win part of fixing <rdar://problem/5111082> Flash popup video ad doesn't close when clicked at http://www.firestonecompleteautocare.com/

        * WebView.cpp:
        (WebView::initWithFrame): Attach the main frame's view.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28442 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Kevin Deckers rubberstamp

        Disclose and export the Databases Directory defaults key

        * Storage/WebDatabaseManager.mm:
        * Storage/WebDatabaseManagerPrivate.h:
        * WebKit.exp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28441 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-12-05  Alp Toker  <alp@atoker.com>

        Reviewed by Mark Rowe.

        Initialize GTK+ and WebKit so the tests can run.

        Reorganize the headers a little.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28440 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-12-05  Michael Natterer  <mitch@imendio.com>

        Reviewed by Alp Toker.

        * WebView/webkitwebview.cpp: split key and button event handlers
        into separate press and release functions.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28439 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-12-05  Luca Bruno  <lethalman88@gmail.com>

        Reviewed by Alp Toker.

        Handle events for Home and End keys.

        * WebView/webkitwebview.cpp:
        (webkit_web_view_key_event):
        * WebCoreSupport/EditorClientGtk.cpp:
        (EditorClient::handleKeypress):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28438 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-12-05  Michael Natterer  <mitch@imendio.com>

        Reviewed by Alp Toker.

        * WebView/webkitwebview.cpp
        * WebCoreSupport/ChromeClientGtk.cpp
        * WebCoreSupport/FrameLoaderClientGtk.cpp: canonicalize signal names.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28437 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Adam Roben.

        - adding a test for http://bugs.webkit.org/show_bug.cgi?id=16097

        * fast/dom/ImageDocument-image-deletion-expected.txt: Added.
        * fast/dom/ImageDocument-image-deletion.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28436 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Fix gtk build

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28435 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Merge TCMalloc r38

Reviewed by Mark Rowe and Geoff Garen.

It also result in a performance progression between 0.5% and
0.9% depending on the test, however most if not all of this
gain will be consumed by the overhead involved in the later
change to release memory to the system.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28434 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-12-04  Alp Toker  <alp@atoker.com>

        Potential GTK+ build fix for the glib-genmarshal issue.

        Split the glib-genmarshal qmake generator in two; one for the source
        file, another for the header.

        * WebCore.pro:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28433 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Remove IWebScriptScope include.

        * Drosera/win/DebuggerDocumentPlatform.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28432 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-12-04  Alp Toker  <alp@atoker.com>

        GTK+ build fix.

        * platform/PlugInInfoStore.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28431 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Anders.

        <rdar://problem/5629125> PluginInfoStore needs the ability to return the name of a plug-in for a given MIME type

        * page/mac/WebCoreViewFactory.h: Added pluginNameForMIMEType: method.
        * platform/PlugInInfoStore.h:
        * platform/gtk/TemporaryLinkStubs.cpp: Added stub version of pluginNameForMIMEType.
        (PlugInInfoStore::pluginNameForMIMEType):
        * platform/mac/PlugInInfoStoreMac.mm:
        (WebCore::PlugInInfoStore::pluginNameForMIMEType): Added.
        * platform/qt/PlugInInfoStoreQt.cpp:
        (WebCore::PlugInInfoStore::pluginNameForMIMEType): Added stub version of pluginNameForMIMEType.
        * plugins/win/PlugInInfoStoreWin.cpp:
        (WebCore::PlugInInfoStore::pluginNameForMIMEType): Added stub version of pluginNameForMIMEType.

        * WebCoreSupport/WebViewFactory.mm:
        (-[WebViewFactory pluginNameForMIMEType:]): Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28426 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

        Reviewed by Darin.

        Fix for <rdar://problem/5499918> REGRESSION: insertion point in input field with custom border cuts holes in focus ring interior edges

        * rendering/RenderTextControl.cpp: (WebCore::RenderTextControl::setStyle):
        Adjust outline-offset so it doesn't extend beyond the border to interfere with typing

LayoutTests:

        Reviewed by Darin.

        Test for <rdar://problem/5499918> REGRESSION: insertion point in input field with custom border cuts holes in focus ring interior edges

        * fast/forms/textfield-outline.html: Added.
        * platform/mac/fast/forms/textfield-outline-expected.checksum: Added.
        * platform/mac/fast/forms/textfield-outline-expected.png: Added.
        * platform/mac/fast/forms/textfield-outline-expected.txt: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28423 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Provide API for setting the default storage quota per database origin

        Reviewed by Jon.

        * Interfaces/IWebPreferencesPrivate.idl: added SPI to get/set default database quota
        * WebPreferenceKeysPrivate.h:
        * WebPreferences.cpp:
        (WebPreferences::initializeDefaultSettings): default database quota is 5MB
        (WebPreferences::longlongValueForKey):
        (WebPreferences::setLongLongValue):
        (WebPreferences::defaultDatabaseQuota):
        (WebPreferences::setDefaultDatabaseQuota):
        * WebPreferences.h:
        * WebView.cpp:
        (WebView::notifyPreferencesChanged): propagate default database quota change to WebCore settings.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28422 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Jon.

        Always call NPP_NewStream, NPP_DestroyStream and NPP_URLNotify for JavaScript
        result streams, even if the JavaScript expression threw an exception or returned undefined.
        This matches what Firefox does.

        * plugins/win/PluginStreamWin.cpp:
        (WebCore::PluginStreamWin::sendJavaScriptStream):
        Handle the result string being null, and set the correct NPReason if it is.

        * plugins/win/PluginViewWin.cpp:
        (WebCore::PluginViewWin::performRequest):
        Always send the JavaScript stream, regardless of the result could be converted to a string or not.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28421 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Mark Rowe

        Tweaked the way we typedef and cast these objc_msgSend calls

        * WebView/WebView.mm:
        (CallDelegateReturningUnsignedLongLong):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28420 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Rubber stamped by Mark Rowe.

        Define CF as platform for mac and revert r28409

        * DumpRenderTree/DumpRenderTree.h:
        * DumpRenderTree/mac/DumpRenderTreeMac.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28419 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Adam Roben.

        Remove http/tests/incremental/frame-focus-before-load.html from the skipped list
        as it is no longer crashing. <rdar://problem/5507278>

        * platform/win/Skipped:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28418 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

        Reviewed by Hyatt.

        Fix for <rdar://problem/5622464> controls should show up for audio element

        * css/html4.css: Added styles for audio controls.  This will need to be refined, since we may want a different look for audio controls.
        * html/HTMLMediaElement.cpp:
        (WebCore::HTMLMediaElement::attributeChanged): When the controls attribute changes detach and reattach if necessary.
        (WebCore::HTMLMediaElement::rendererIsNeeded): Added.
        (WebCore::HTMLMediaElement::createRenderer): Added.  Uses RenderMedia.
        * html/HTMLMediaElement.h:
        * rendering/RenderMedia.cpp: (WebCore::RenderMedia::RenderMedia): Added constructor that doesn't take intrinsic size.
        * rendering/RenderMedia.h:

LayoutTests:

        Reviewed by Hyatt.

        Test for <rdar://problem/5622464> controls should show up for audio element

        * media/audio-controls-rendering.html: Added.
        * platform/mac/media/audio-controls-rendering-expected.checksum: Added.
        * platform/mac/media/audio-controls-rendering-expected.png: Added.
        * platform/mac/media/audio-controls-rendering-expected.txt: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28417 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Brady Eidson (with help from Mark Rowe)

        Fixed return values for unsigned-long-long delegate methods

        * WebView/WebView.mm:
        (CallDelegateReturningUnsignedLongLong):
        redid the change that Brady did at home over the weekend but forgot to check in

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28416 268f45cc-cd09-0410-ab3c-d52691b4dbfc

JavaScriptCore:

        Reviewed by Sam.

        Make isSafeScript const.

        * kjs/JSGlobalObject.h:
        (KJS::JSGlobalObject::isSafeScript):

WebCore:

        Reviewed by Sam.

        Add isSafeScript(const JSGlobalObject*) so we won't get the JSGlobalObject
        implementation, which always returns true(!).

        * bindings/js/kjs_window.cpp:
        (KJS::Window::isSafeScript):
        Make the static isSafeScript method take two JSGlobalObjects.

        * bindings/js/kjs_window.h:
        (KJS::Window::isSafeScript):
        Implement isSafeScript and have it call the static method.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28415 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Define CF as platform for windows.

        Reviewed by Adam Roben.

        * DumpRenderTree/win/DumpRenderTreeWin.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28414 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebKit/win:

        Reviewed by Adam and Darin.

        - Fixed a crasher where the top of the stack was not updated correctly.
        And set the eol style on IWebScriptCallFrame.idl file.

        * Interfaces/IWebScriptCallFrame.idl:
        * WebScriptDebugger.cpp:
        (WebScriptDebugger::atStatement):
        (WebScriptDebugger::leaveFrame):

WebKitTools:

        Reviewed by Adam and Darin.

        - Removed a needless BSTR cleanup.

        * Drosera/win/DebuggerDocumentPlatform.cpp:
        (DebuggerDocument::getPlatformCurrentFunctionStack):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28413 268f45cc-cd09-0410-ab3c-d52691b4dbfc

 Hook up the authorAndUserStylesEnabled preference on Windows

 WebKit/win:

         Hook up the authorAndUserStylesEnabled preference

         This is the Windows counterpart of r28071.

         Reviewed by Maciej.

         * Interfaces/IWebPreferencesPrivate.idl: Added new method
         declarations.
         * WebPreferenceKeysPrivate.h: Added a new preference key.
         * WebPreferences.cpp:
         (WebPreferences::initializeDefaultSettings): Set the default for the
         new preference.
         (WebPreferences::setAuthorAndUserStylesEnabled): Added.
         (WebPreferences::authorAndUserStylesEnabled): Added.
         * WebPreferences.h:
         * WebView.cpp:
         (WebView::notifyPreferencesChanged): Communicate the new preference
         down to WebCore::Settings.

 WebKitTools:

         Fix <rdar://5614497> setAuthorAndUserStylesEnabled is not implemented in DRT

         Reviewed by Maciej.

         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
         (LayoutTestController::setAuthorAndUserStylesEnabled): Implemented.

 LayoutTests:

         Remove a now-passing test from the Windows Skipped file

         Reviewed by Maciej.

         * platform/win/Skipped:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28412 268f45cc-cd09-0410-ab3c-d52691b4dbfc

 Make the implementation of Frame::setNeedsReapplyStyles cross-platform

 WebCore:

         Make the implementation of Frame::setNeedsReapplyStyles cross-platform

         Reviewed by Hyatt.

         No change in functionality.

         * page/Frame.cpp:
         (WebCore::Frame::setNeedsReapplyStyles): Added. Invalidates the
         FrameView so that FrameView::layout will be called, which ends up
         calling reapplyStyles.
         (WebCore::Frame::needsReapplyStyles): Added.
         (WebCore::Frame::reapplyStyles): Renamed from reparseConfiguration.
         (WebCore::FramePrivate::FramePrivate):
         * page/Frame.h:
         * page/FramePrivate.h: Added new boolean member.
         * page/FrameView.cpp:
         (WebCore::FrameView::layout): Call Frame::reapplyStyles if needed.
         (WebCore::FrameView::needsLayout): Say that we need layout if the
         Frame needs styles reapplied.
         * page/mac/WebCoreFrameBridge.h: Removed setNeedsReapplyStyles.
         * page/mac/WebCoreFrameBridge.mm:
         (-[WebCoreFrameBridge reapplyStylesForDeviceType:]): Updated for
         rename.

         * bridge/win/FrameWin.cpp: Removed Frame::setNeedsReapplyStyles.
         * page/gtk/FrameGtk.cpp: Ditto.
         * page/mac/FrameMac.mm: Ditto.
         * page/qt/FrameQt.cpp: Ditto.
         * platform/wx/TemporaryLinkStubs.cpp: Ditto.

 WebKit/mac:

         Remove -[WebFrameBridge setNeedsReapplyStyles]

         This functionality is now WebCore's responsibility.

         Reviewed by Hyatt.

         * WebCoreSupport/WebFrameBridge.mm:
         * WebKit.order:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28411 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-12-04  Alp Toker  <alp@atoker.com>

        Prospective Win DRT build fix.

        * DumpRenderTree/DumpRenderTree.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28409 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Adele.

        Pass intrinsic size to the base class.

        * rendering/RenderMedia.cpp:
        (WebCore::RenderMedia::RenderMedia):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28408 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-12-04  Jon Honeycutt  <jhoneycutt@apple.com>

        Reviewed by Ollie.

        Don't crash if our JavaPlugin DLL can't be found

        * WebFrame.cpp:
        (WebFrame::createJavaAppletWidget): Added null check

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28405 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Darin

        Added deleteOrigin: SPI, which isn't fully implemented

        * Storage/WebDatabaseManagerPrivate.h:
        * Storage/WebDatabaseManager.mm:
        (-[WebDatabaseManager deleteOrigin:]):
        just calls deleteDatabasesWithOrigin: for now, but needs to delete origin itself too

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28404 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Darin.

        Fix for <rdar://problem/5578209> REGRESSION: Crash in WebCore::DeprecatedValueListImpl::Private::copyList when going back on zoomerang.com/survey

        * loader/FrameLoader.cpp: (WebCore::FrameLoader::scheduleHistoryNavigation):
          When the new url is exactly the same as the old url, don't treat it like a fragment navigation, since its going to cause a reload.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28403 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Mark Rowe.

        Remove a check for early versions of Leopard CFNetwork now that
        Leopard has shipped.

        * Misc/WebKitVersionChecks.h: Remove WEBKIT_FIRST_CFNETWORK_VERSION_WITH_LARGE_DISK_CACHE_FIX.
        * WebView/WebView.mm:
        (+[WebView _setCacheModel:]): Remove the early Leopard CFNetwork check.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28401 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-12-04  Kevin Decker  <kdecker@apple.com>

        Reviewed by Darin.

        Revised fix for: <rdar://problem/5586978> REGRESSION (Safari 2-3): WebKit sometimes doesn't invoke Flash's NPP_SetWindow function and causes a hang

        This fix is exactly the same as chageset 28359 with the exception of an added early return in
        updateAndSetWindow to cover the additional case of when a plug-in isn't drawable.

        The CG-based Flash player would sometimes hang because (for CoreGraphics-based plug-ins) our code would
        only call into the NPP_SetWindow() function when we tell the plug-in to draw. This created havoc with
        Flash because Flash expects the browser to call NPP_SetWindow() and provide a valid graphics context
        regardless of whether or not it actually needs to draw.

        * Plugins/WebBaseNetscapePluginView.mm:
        (-[WebBaseNetscapePluginView saveAndSetNewPortStateForUpdate:]): Removed an incorrect comment and toned
        down an ASSERT that was too strict.
        (-[WebBaseNetscapePluginView updateAndSetWindow]): Removed an early return for CoreGraphics-based plug-ins
        which would sometimes altogether prevent updating the PortState and calling into a plug-ins NPP_SetWindow()
        function. Also tweaked a comment and added an early return if the plug-in can't draw.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28400 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

        Reviewed by Kevin Decker.

        - added an assertion for a situation that leads to a Safari assertion:
          a commit that is not followed by a load completion

        - removed some unneeded FrameLoader stuff -- maybe some day we'll get this
          class cut down to size

        * loader/FrameLoader.cpp:
        (WebCore::FrameLoader::FrameLoader): Added boolean for assertion.
        (WebCore::FrameLoader::checkLoadCompleteForThisFrame): Added code to assert
        that we already sent a commit before sending a complete.
        (WebCore::FrameLoader::continueFragmentScrollAfterNavigationPolicy): Call
        client directly instead of indirecting through a function.
        (WebCore::FrameLoader::loadItem): Ditto.
        (WebCore::FrameLoader::goToItem): Ditto.
        (WebCore::FrameLoader::addHistoryForCurrentLocation): Get the private browsing
        state from Settings instead of FrameLoaderClient. Also call client directly
        instead of indirecting through a function.
        (WebCore::FrameLoader::updateHistoryForReload): Call client directly instead
        of indirecting through a function.
        (WebCore::FrameLoader::dispatchDidCommitLoad): Added code to assert
        that we did not yet send a commit and set the flag that will be used to check
        that we don't do this twice in a row without an intervening completion call.

        * loader/FrameLoader.h: Added a boolean for the assertion. Removed six
        now-unneeded functions.

        * loader/FrameLoaderClient.h: Removed now-unneeded privateBrowsingEnabled function.

        * loader/ResourceLoader.cpp: (WebCore::ResourceLoader::willCacheResponse): Get
        the private browsing state from Settings instead of FrameLoaderClient.

        * svg/graphics/SVGImageEmptyClients.h: Removed obsolete privateBrowsingEnabled.

WebKit/gtk:

        Reviewed by Kevin Decker.

        * WebCoreSupport/FrameLoaderClientGtk.cpp: Removed obsolete privateBrowsingEnabled.
        * WebCoreSupport/FrameLoaderClientGtk.h: Ditto.

WebKit/mac:

        Reviewed by Kevin Decker.

        * WebCoreSupport/WebFrameLoaderClient.h: Removed obsolete privateBrowsingEnabled.
        * WebCoreSupport/WebFrameLoaderClient.mm: Ditto.
        * WebKit.order: Ditto.

WebKit/qt:

        Reviewed by Kevin Decker.

        * WebCoreSupport/FrameLoaderClientQt.cpp: Removed obsolete privateBrowsingEnabled.
        * WebCoreSupport/FrameLoaderClientQt.h: Ditto.

WebKit/win:

        Reviewed by Kevin Decker.

        * WebFrame.cpp: Removed obsolete privateBrowsingEnabled.
        * WebFrame.h: Ditto.

WebKit/wx:

        Reviewed by Kevin Decker.

        * WebKitSupport/FrameLoaderClientWx.cpp: Removed obsolete privateBrowsingEnabled.
        * WebKitSupport/FrameLoaderClientWx.h: Ditto.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28399 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Update StringsNotToBeLocalized.txt.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28398 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

        Reviewed by Adele.

        Fix <rdar://problem/5624802>
        Controller displays outside of <video> element when width and height aren't specified

        - Explicitly adjust size of the controls tree root.
        - Switch RenderMedia base class back to RenderReplaced. It is easier to manage the shadow tree
          by hand. This also allows better code sharing with other replaced element classes.
        - Move duplicated layout and painting code from subclasses up to RenderReplaced.

        Test: media/video-controls-rendering.html

        * rendering/RenderHTMLCanvas.cpp:
        (WebCore::RenderHTMLCanvas::paintReplaced):
        * rendering/RenderHTMLCanvas.h:
            Share code.
        * rendering/RenderImage.cpp:
        (WebCore::RenderImage::paintReplaced):
        (WebCore::RenderImage::minimumReplacedHeight):
        * rendering/RenderImage.h:
            Share code.
        * rendering/RenderMedia.cpp:
        (WebCore::RenderMediaControlShadowRoot::RenderMediaControlShadowRoot):
        (WebCore::RenderMediaControlShadowRoot::setParent):
            Add a subclass just to get through encapsulation to use setParent().
        (WebCore::MediaControlShadowRootElement::MediaControlShadowRootElement):
            Move initialization to the constructor.
        (WebCore::RenderMedia::RenderMedia):
        (WebCore::RenderMedia::~RenderMedia):
        (WebCore::RenderMedia::layout):
            Resize and layout the controller root by hand.
        (WebCore::RenderMedia::firstChild):
        (WebCore::RenderMedia::lastChild):
        (WebCore::RenderMedia::removeChild):
            Support one child renderer for controls.
        (WebCore::RenderMedia::createControlsShadowRoot):
        * rendering/RenderMedia.h:
        (WebCore::RenderMedia::isMedia):
            Switch base class to RenderReplaced, delete the now unnecessary stuff.
        * rendering/RenderReplaced.cpp:
        (WebCore::RenderReplaced::layout):
        (WebCore::RenderReplaced::paint):
        * rendering/RenderReplaced.h:
        (WebCore::RenderReplaced::minimumReplacedHeight):
        (WebCore::RenderReplaced::paintReplaced):
            Share code.
        * rendering/RenderVideo.cpp:
        (WebCore::RenderVideo::videoSizeChanged):
            Simplify, just request relayout.
        (WebCore::RenderVideo::paintReplaced):
        (WebCore::RenderVideo::layout):
        * rendering/RenderVideo.h:
            Share code.

LayoutTests:

        Reviewed by Adele.

        Test for <rdar://problem/5624802>
        Controller displays outside of <video> element when width and height aren't specified

        * media/remove-from-document.html: Improve coverage by doing relayout before finishing.
        * media/video-controls-rendering.html: Added.
        * platform/mac/media/video-controls-rendering-expected.checksum: Added.
        * platform/mac/media/video-controls-rendering-expected.png: Added.
        * platform/mac/media/video-controls-rendering-expected.txt: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28397 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-12-04  Michael Natterer  <mitch@imendio.com>

        Reviewed by Alp Toker.

        * WebCore.pro: fix build of webkit-marshal by building the
        functions into a separate .cpp file instead of including them in
        the header. This way we can use the marshallers from more than one
        file without using bad hacks.

        Add the generated sources directory to DEPENDPATH.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28396 268f45cc-cd09-0410-ab3c-d52691b4dbfc

JavaScriptCore:

        Reviewed by Geoff.

        - fix first part of http://bugs.webkit.org/show_bug.cgi?id=16220
          <rdar://problem/5625221> Crash opening www.news.com (CNet)

        Test: fast/js/regexp-overflow.html

        * pcre/pcre_compile.cpp:
        (calculateCompiledPatternLengthAndFlags): Add room for the additional BRA/KET that
        was generated in the compile code but not taken into account here.

LayoutTests:

        Reviewed by Geoff.

        - test for first part of http://bugs.webkit.org/show_bug.cgi?id=16220
          <rdar://problem/5625221> Crash opening www.news.com (CNet)

        * fast/js/regexp-overflow-expected.txt: Updated.
        * fast/js/resources/regexp-overflow.js: Added test cases.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28395 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Build fix.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28392 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Windows build fix

        * WebScriptCallFrame.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28391 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-12-04  Alp Toker  <alp@atoker.com>

        Reviewed by Mark Rowe.

        Link to Planet WebKit.

        * nav.inc:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28390 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-12-04  Pierre-Luc Beaudoin  <pierre-luc.beaudoin@collabora.co.uk>

        Reviewed by Alp Toker.

        http://bugs.webkit.org/show_bug.cgi?id=16145
        [gtk] Implement media support in GTK backend

        This implements the media tags of HTML5 on the GTK+ port based on the
        Mac port. Media tests pass although some tests required a small modifications:
        As per the HTML5 spec, the loading of the clip should start only after
        all script are done.  But in the case of the tests, the loading starts
        before some callbacks get registered.  This has been discussed with
        Antti Koivisto, and the tests should be updated.

        There is only one known issue: one time in 10 tries, loading a buffered
        clip will never end buffering.  The buffering will stall at 97% for no
        apparent reasons.  Reloading the page usually works around the problem.
        On the GStreamer side, some codecs don't return correct values, therefore
        they make the tests fail.  For instance H.264 will make the video-end
        test fail.  This should be fixed in GStreamer 0.10.15.

        This version displays video in a pop up window.  A place holder is drawn
        on the page where the video should appear.

        By default, it is turned off in WebCore.pro until GStreamer/X
        detection issues are sorted out.

        * WebCore.pro:
        Disable video for now
        * platform/Logging.cpp:
        (WebCore::):
        * platform/Logging.h:
        Added a logging level for Media stuff
        * platform/graphics/Movie.cpp:
        * platform/graphics/gtk/MoviePrivateGStreamer.cpp: Added.
        (WebCore::moviePrivateErrorCallback):
        (WebCore::moviePrivateEOSCallback):
        (WebCore::moviePrivateStateCallback):
        (WebCore::moviePrivateBufferingCallback):
        (WebCore::moviePrivateWindowIDCallback):
        (WebCore::MoviePrivate::MoviePrivate):
        (WebCore::MoviePrivate::~MoviePrivate):
        (WebCore::MoviePrivate::load):
        (WebCore::MoviePrivate::play):
        (WebCore::MoviePrivate::pause):
        (WebCore::MoviePrivate::duration):
        (WebCore::MoviePrivate::currentTime):
        (WebCore::MoviePrivate::seek):
        (WebCore::MoviePrivate::setEndTime):
        (WebCore::MoviePrivate::addCuePoint):
        (WebCore::MoviePrivate::removeCuePoint):
        (WebCore::MoviePrivate::clearCuePoints):
        (WebCore::MoviePrivate::startCuePointTimerIfNeeded):
        (WebCore::MoviePrivate::cancelSeek):
        (WebCore::MoviePrivate::cuePointTimerFired):
        (WebCore::MoviePrivate::paused):
        (WebCore::MoviePrivate::seeking):
        (WebCore::MoviePrivate::naturalSize):
        (WebCore::MoviePrivate::hasVideo):
        (WebCore::MoviePrivate::setVolume):
        (WebCore::MoviePrivate::setMuted):
        (WebCore::MoviePrivate::setRate):
        (WebCore::MoviePrivate::dataRate):
        (WebCore::MoviePrivate::networkState):
        (WebCore::MoviePrivate::readyState):
        (WebCore::MoviePrivate::maxTimeBuffered):
        (WebCore::MoviePrivate::maxTimeSeekable):
        (WebCore::MoviePrivate::maxTimeLoaded):
        (WebCore::MoviePrivate::bytesLoaded):
        (WebCore::MoviePrivate::totalBytesKnown):
        (WebCore::MoviePrivate::totalBytes):
        (WebCore::MoviePrivate::cancelLoad):
        (WebCore::MoviePrivate::updateStates):
        (WebCore::MoviePrivate::loadStateChanged):
        (WebCore::MoviePrivate::rateChanged):
        (WebCore::MoviePrivate::sizeChanged):
        (WebCore::MoviePrivate::timeChanged):
        (WebCore::MoviePrivate::volumeChanged):
        (WebCore::MoviePrivate::didEnd):
        (WebCore::MoviePrivate::loadingFailed):
        (WebCore::MoviePrivate::setRect):
        (WebCore::MoviePrivate::setVisible):
        (WebCore::MoviePrivate::paint):
        (WebCore::MoviePrivate::getSupportedTypes):
        (WebCore::MoviePrivate::createGSTPlayBin):
        * platform/graphics/gtk/MoviePrivateGStreamer.h: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28389 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-12-04  Luca Bruno  <lethalman88@gmail.com>

        Reviewed by Alp Toker.

        Fix a regression in key press propagation in r28386.

        Fix indentation (was off by two spaces).

        * WebView/webkitwebview.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28388 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Add the missing ChangeLog entries for r28349.

WebCore/ChangeLog:
2007-12-04  Holger Hans Peter Freyther <holger.freyther@trolltech.com>

        Reviewed by Simon.

        * Implement the InspectorClient for the Qt port
        * It does not support highlighting of nodes yet
        * Use QRC to open the internal page. The important thing is the
        '/' in the URL to make WebCore::Document::completeURL behave the
        way we want.
        * To make the InspectorClient work we will have to mark qrc as secure.

        * WebCore.pro:
        * loader/FrameLoader.cpp:
        (WebCore::localSchemes):
        * page/inspector/WebKit.qrc: Added.

WebKit/qt:
2007-12-04  Holger Hans Peter Freyther <holger.freyther@trolltech.com>

        Reviewed by Simon.

        * Implement the InspectorClient for the Qt port
        * It does not support highlighting of nodes yet
        * Use QRC to open the internal page. The important thing is the
        '/' in the URL to make WebCore::Document::completeURL behave the
        way we want.
        * To make the InspectorClient work we will have to mark qrc as secure.

        * Api/qwebpage.cpp:
        (QWebPagePrivate::QWebPagePrivate):
        * Api/qwebpage.h:
        * WebCoreSupport/InspectorClientQt.cpp:
        (WebCore::InspectorClientWebPage::hideEvent):
        (WebCore::InspectorClientQt::InspectorClientQt):
        (WebCore::InspectorClientQt::inspectorDestroyed):
        (WebCore::InspectorClientQt::createPage):
        (WebCore::InspectorClientQt::localizedStringsURL):
        (WebCore::InspectorClientQt::showWindow):
        (WebCore::InspectorClientQt::closeWindow):
        (WebCore::InspectorClientQt::attachWindow):
        (WebCore::InspectorClientQt::detachWindow):
        * WebCoreSupport/InspectorClientQt.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28387 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-12-04  Luca Bruno  <lethalman88@gmail.com>

        Reviewed by Alp Toker.

        http://bugs.webkit.org/show_bug.cgi?id=15911
        [GTK] Use GtkBindingSet to make key bindings user-configurable

        This patch doesn't cover the full range of bindings, only the ones
        that seem obviously correct and have clear public API.

        * WebCoreSupport/EditorClientGtk.cpp:
        (WebKit::EditorClient::handleKeypress): do not handle clipboard operations and select-all
        * WebView/webkitwebview.cpp: add cut, copy, paste and select-all signals and allow binding sets (issue #15911 and #16144)
        * WebView/webkitwebview.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28386 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-12-04  Alp Toker  <alp@atoker.com>

        Fix a clobbered copyright header.

        * DumpRenderTree/gtk/DumpRenderTree.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28385 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-12-04  Xan Lopez  <xan@gnome.org>

        Reviewed by Alp Toker.

        http://bugs.webkit.org/show_bug.cgi?id=15561
        GTK port needs DumpRenderTree implementation

        Start work on the GTK+ DRT.

        This does not work yet, and there are a few lingering style issues
        (nothing major) but this patch has been stuck in the bug tracker for
        too long already.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28384 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Roll out r28359 as it causes serious problems with tabbed browsing in Safari when viewing pages containing plugins (<rdar://problem/5627529>).

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28383 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-12-04  Alp Toker  <alp@atoker.com>

        Rubber stamped by Mark Rowe.

        Remove two unused variables to silence warnings.

        * html/CanvasRenderingContext2D.cpp:
        (WebCore::CanvasRenderingContext2D::fill):
        (WebCore::CanvasRenderingContext2D::stroke):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28382 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Rubberstamped by Tim.

        Make the bullets lickable.

        * wwwroot/planetwebkit.css:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28381 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Adam.

        Make it sexy with awesome icon and tweaks from Tim.

        * templates/index.html.tmpl:
        * wwwroot/ie.css: Added.
        * wwwroot/images/ie-nav-blue.png: Added.
        * wwwroot/planetwebkit.css: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28380 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Added bug reference.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28379 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Move JavaScriptCore thread testing code to pthread specific directory
        and remove the dependance on CoreFoundation by using WTF::HashSet.

        Reviewed by Geoff and Oliver.

        * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
        * DumpRenderTree/ForwardingHeaders/wtf/HashMap.h: Added.
        * DumpRenderTree/ForwardingHeaders/wtf/HashSet.h: Added.
        * DumpRenderTree/ForwardingHeaders/wtf/Vector.h: Added.
        * DumpRenderTree/JavaScriptThreading.h: Copied from DumpRenderTree/mac/JavaScriptThreading.h.
        * DumpRenderTree/mac/JavaScriptThreading.cpp: Removed.
        * DumpRenderTree/mac/JavaScriptThreading.h: Removed.
        * DumpRenderTree/pthreads: Added.
        * DumpRenderTree/pthreads/JavaScriptThreadingPthreads.cpp: Copied from DumpRenderTree/mac/JavaScriptThreading.cpp.
        (javaScriptThreads):
        (runJavaScriptThread):
        (startJavaScriptThreads):
        (stopJavaScriptThreads):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28377 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Versioning.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28375 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-12-03  Andrew Bonventre  <andybons@google.com>

        Reviewed by Darin Adler.

        * Drosera/debugger.js: Fixed javascript error where document property
        should have been used instead of contentDocument in switchFunction that
        was breaking the select symbol dropdown menu.  This is because
        window.frames will return a Window object and not a frame object like
        the author was originally expecting.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28374 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

        Reviewed by Darin Adler.

        - fix <rdar://problem/5134044> fast/frames/frame-src-attribute.html fails on Windows

        * platform/win/ScrollViewWin.cpp:
        (WebCore::ScrollView::visibleWidth): Do not return negative values.
        (WebCore::ScrollView::visibleHeight): Ditto.

LayoutTests:

        Reviewed by Darin Adler.

        - remove passing test

        * platform/win/Skipped:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28372 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

        Reviewed by Dave Hyatt.

        - fix <rdar://problem/5346452> Resize event doesn't fire on body element inside a frame

        Test: fast/events/resize-subframe.html

        * page/FrameView.cpp:
        (WebCore::FrameViewPrivate::FrameViewPrivate): Added a timer used when
        deferring tasks that need to be done after layout.
        (WebCore::FrameViewPrivate::reset):
        (WebCore::FrameView::~FrameView):
        (WebCore::FrameView::layout): Moved the updating of widget positions,
        loading plug-ins and sending events queued up during layout into
        performPostLayoutTasks(). performPostLayoutTasks() is called after
        layout unless the layout was triggered by a previous layout's post-
        layout tasks. In the latter case, performPostLayoutTasks() is scheduled
        to run later.
        (WebCore::FrameView::performPostLayoutTasks): Performs work that needs
        to be done after layout but which can result in arbitrary code
        execution and therefore may re-invalidate the layout. This includes
        updating widget positions, loading plug-ins, and dispatching layout-
        related DOM events (scroll, overflow and resize).
        (WebCore::FrameView::postLayoutTimerFired):
        (WebCore::FrameView::dispatchScheduledEvents):
        * page/FrameView.h:
        * page/mac/WebCoreFrameBridge.h:
        * page/mac/WebCoreFrameBridge.mm: Removed -sendResizeEvent since this
        is handled by FrameView now.

WebKit/gtk:

        Reviewed by Dave Hyatt.

        - fix <rdar://problem/5346452> Resize event doesn't fire on body element inside a frame

        * WebView/webkitwebview.cpp: Remove the call to sendResizeEvent() since
        FrameView sends it now.

WebKit/mac:

        Reviewed by Dave Hyatt.

        - fix <rdar://problem/5346452> Resize event doesn't fire on body element inside a frame

        * WebView/WebHTMLView.mm:
        (-[WebHTMLView layoutToMinimumPageWidth:maximumPageWidth:adjustingViewSize:]): Removed the code
        that checked if the view had resized and sent the resize event, since
        FrameView sends resize events now.
        * WebView/WebHTMLViewInternal.h:

WebKit/win:

        Reviewed by Dave Hyatt.

        - fix <rdar://problem/5346452> Resize event doesn't fire on body element inside a frame

        * WebView.cpp:
        (WebViewWndProc): Removed call to sendResizeEvent() since FrameView
        sends them now.

LayoutTests:

        Reviewed by Dave Hyatt.

        - test for <rdar://problem/5346452> Resize event doesn't fire on body element inside a frame

        * fast/events/resize-subframe-expected.txt: Added.
        * fast/events/resize-subframe.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28371 268f45cc-cd09-0410-ab3c-d52691b4dbfc

JavaScriptCore:

        Reviewed by Geoff.

        - fix http://bugs.webkit.org/show_bug.cgi?id=15618
          <rdar://problem/5619353> REGRESSION: Stack overflow/crash in KJS::equal (15618)

        Test: fast/js/recursion-limit-equal.html

        * kjs/operations.cpp: (KJS::equal): Check the exception from toPrimitive.

LayoutTests:

        Reviewed by Geoff.

        - test for http://bugs.webkit.org/show_bug.cgi?id=15618
          <rdar://problem/5619353> REGRESSION: Stack overflow/crash in KJS::equal (15618)

        * fast/js/recursion-limit-equal-expected.txt: Added.
        * fast/js/recursion-limit-equal.html: Added.
        * fast/js/resources/recursion-limit-equal.js: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28370 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebKit/win:

        Reviewed by Adam.

        - <rdar://5618942> Drosera: Console window does not process everything
        correctly.
        - <rdar://5619005> Drosera: could be sped up by moving the
        WebScriptScope stuff into the WebScriptCallFrame.
        - Now the console can correctly process objects and does not receive
        notifications from JavaScriptCore about the JavaScript in Drosera's
        own process.

        * Interfaces/IWebScriptCallFrame.idl: Moved the variable lookup
        functions from WebScriptScope to here.
        * Interfaces/WebKit.idl: Removed WebScriptScope.
        * WebKit.vcproj/Interfaces.vcproj: Ditto.
        * WebKit.vcproj/WebKit.vcproj: Ditto.
        * WebKit.vcproj/WebKitGUID.vcproj: Ditto.
        * WebScriptCallFrame.cpp: Moved WebScriptScope functionality into here.
        (WebScriptCallFrame::stringByEvaluatingJavaScriptFromString): Now
        returns a string value for things that are not strings.
        (WebScriptCallFrame::variableNames): Moved from WebScriptScope.
        (WebScriptCallFrame::valueForVariable): Moved from WebScriptScope.
        (WebScriptCallFrame::jsValueToString): Helper functions that converts
        any JSValue to a String.
        * WebScriptCallFrame.h: Moved WebScriptScope functionality into here.
        * WebScriptDebugger.cpp: Added nested guards so that Drosera does not
        receive notifications about JavaScript that Drosera is running just the
        JavaScript WebKit is running.
        (WebScriptDebugger::WebScriptDebugger):
        (WebScriptDebugger::sourceParsed):
        (WebScriptDebugger::callEvent):
        (WebScriptDebugger::atStatement):
        (WebScriptDebugger::returnEvent):
        (WebScriptDebugger::exception):
        * WebScriptDebugger.h: Ditto.
        * WebScriptScope.cpp: Removed.
        * WebScriptScope.h: Removed.

WebKitTools:

        Reviewed by Adam.

        - <rdar://5618942> Drosera: Console window does not process everything
        correctly.
        - <rdar://5619005> Drosera: could be sped up by moving the
        WebScriptScope stuff into the WebScriptCallFrame.
        - Now the console can correctly process objects and does not receive
        notifications from JavaScriptCore about the JavaScript in Drosera's
        own process.

        * Drosera/win/DebuggerDocumentPlatform.cpp:
        (DebuggerDocument::getPlatformLocalScopeVariableNamesForCallFrame):
        (DebuggerDocument::platformValueForScopeVariableNamed):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28369 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    - fix a copy-and-paste-o

        * bindings/npruntime.cpp:
        (_NPN_GetIntIdentifier):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28368 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Check to see if we are building a debug root

        * Scripts/check-for-global-initializers:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28367 268f45cc-cd09-0410-ab3c-d52691b4dbfc

JavaScriptCore:

        Reviewed by Darin Adler.

        - fix an ASSERT when getIntIdentifier is called with 0 or -1

        * bindings/npruntime.cpp:
        (_NPN_GetIntIdentifier): We cannot use the hashmap for 0 and -1 since
        they are the empty value and the deleted value. Instead, keep the
        identifiers for those two integers in a static array.

WebKitTools:

        Reviewed by Darin Adler.

        - added a testGetIntIdentifier() method to TestNetscapePlugIn

        * DumpRenderTree/TestNetscapePlugIn.subproj/PluginObject.c:
        (pluginInvoke):

LayoutTests:

        Reviewed by Darin Adler.

        - added a test of the behavior of getIntIdentifier with the integers
          0 and -1

        * plugins/getintidentifier-special-values-expected.txt: Added.
        * plugins/getintidentifier-special-values.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28366 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Darin.

        Change WebViewGetResourceLoadDelegateImplementations and WebViewGetFrameLoadDelegateImplementations
        to return a pointer to the implementation struct instead of a copy of the struct. This changes
        all of the callers to dereference the pointer to access the struct fields.

        * Plugins/WebNullPluginView.mm:
        (-[WebNullPluginView reportFailure]):
        * WebCoreSupport/WebFrameBridge.mm:
        * WebCoreSupport/WebFrameLoaderClient.mm:
        (WebFrameLoaderClient::dispatchDidLoadResourceFromMemoryCache):
        (WebFrameLoaderClient::assignIdentifierToInitialRequest):
        (WebFrameLoaderClient::dispatchWillSendRequest):
        (WebFrameLoaderClient::dispatchDidReceiveAuthenticationChallenge):
        (WebFrameLoaderClient::dispatchDidCancelAuthenticationChallenge):
        (WebFrameLoaderClient::dispatchDidReceiveResponse):
        (WebFrameLoaderClient::willCacheResponse):
        (WebFrameLoaderClient::dispatchDidReceiveContentLength):
        (WebFrameLoaderClient::dispatchDidFinishLoading):
        (WebFrameLoaderClient::dispatchDidFailLoading):
        (WebFrameLoaderClient::dispatchDidHandleOnloadEvents):
        (WebFrameLoaderClient::dispatchDidReceiveServerRedirectForProvisionalLoad):
        (WebFrameLoaderClient::dispatchDidCancelClientRedirect):
        (WebFrameLoaderClient::dispatchWillPerformClientRedirect):
        (WebFrameLoaderClient::dispatchDidChangeLocationWithinPage):
        (WebFrameLoaderClient::dispatchWillClose):
        (WebFrameLoaderClient::dispatchDidStartProvisionalLoad):
        (WebFrameLoaderClient::dispatchDidReceiveTitle):
        (WebFrameLoaderClient::dispatchDidCommitLoad):
        (WebFrameLoaderClient::dispatchDidFailProvisionalLoad):
        (WebFrameLoaderClient::dispatchDidFailLoad):
        (WebFrameLoaderClient::dispatchDidFinishDocumentLoad):
        (WebFrameLoaderClient::dispatchDidFinishLoad):
        (WebFrameLoaderClient::dispatchDidFirstLayout):
        * WebView/WebView.mm:
        (WebViewGetResourceLoadDelegateImplementations):
        (WebViewGetFrameLoadDelegateImplementations):
        (-[WebView _dispatchDidReceiveIconFromWebFrame:]):
        * WebView/WebViewInternal.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28365 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Fix the build, move JSDOMWindow.h to the Copy Generated Headers phase.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28364 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Darin.

        http://bugs.webkit.org/show_bug.cgi?id=15767
        Text decorations do not respect "visibility" property

        Only paint decorations if visibility is "visible".

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28363 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Brady Eidson.

        <rdar://problem/5539913> 188 crashes in WebViewGetFrameLoadDelegateImplementations
        <rdar://problem/5586095> CrashTracer: [USER] 5000+ crashes in Safari and Dashboard in dispatchDidFailLoading
        <rdar://problem/5607081> CrashTracer: [USER] 2150 crashes in Safari at com.apple.WebKit:
        WebViewGetResourceLoadDelegateImplementations + 28

        * WebView/WebView.mm:
        (-[WebView _cacheResourceLoadDelegateImplementations]): If the delegate is nil, bzero the
        implementation cache. This just prevents us from calling getMethod() multiple times just to zero.
        (-[WebView _cacheFrameLoadDelegateImplementations]): Ditto.
        (WebViewGetResourceLoadDelegateImplementations): Return a zeroed implementations struct
        if the WebView is nil. This fixes the crashes.
        (WebViewGetFrameLoadDelegateImplementations): Ditto.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28362 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Removing passing tests

        * platform/win/Skipped:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28361 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Fix build.

        * WebCore.vcproj/WebCore.vcproj:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28360 268f45cc-cd09-0410-ab3c-d52691b4dbfc

i        Reviewed by Darin.

        Fixed: <rdar://problem/5586978> REGRESSION (Safari 2-3): WebKit sometimes doesn't invoke Flash's NPP_SetWindow function and causes a hang

        The CG-based Flash player would sometimes hang because (for CoreGraphics-based plug-ins) our code would only call into the NPP_SetWindow()
        function when we tell the plug-in to draw. This created havoc with Flash because Flash expects the browser to call NPP_SetWindow()
        and provide a valid graphics context regardless of whether or not it actually needs to draw.

        * Plugins/WebBaseNetscapePluginView.mm:
        (-[WebBaseNetscapePluginView saveAndSetNewPortStateForUpdate:]): Removed an incorrect comment and toned down an ASSERT that was too strict.
        (-[WebBaseNetscapePluginView updateAndSetWindow]): Removed an early return for CoreGraphics-based plug-ins which would sometimes
        altogether prevent updating the PortState and calling into a plug-ins NPP_SetWindow() function.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28359 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Reviewed by David.

Fixes: http://bugs.webkit.org/show_bug.cgi?id=15645
Don't allow selectors to interfere with the SVG shadow tree (<use> internal non-exposed tree).

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28357 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Rubber stamped by Mark.
Commit missing layout test results from my last commit (fix for bug 15403)

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28356 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-12-03  Holger Hans Peter Freyther  <holger.freyther@trolltech.com>

        Rubber stamped by Mark.

        Always include config.h at the beginning of the file.

        * Api/qwebsettings.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28355 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

        Reviewed by Adam Roben.

        - fix <rdar://problem/5538651> REGRESSSION: domfuzz: null deref in WebCore::Document::canReplaceChild

        Test: fast/dom/Document/replaceChild-null-oldChild.html

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

LayoutTests:

        Reviewed by Adam Roben.

        - test for <rdar://problem/5538651> REGRESSSION: domfuzz: null deref in WebCore::Document::canReplaceChild

        * fast/dom/Document/replaceChild-null-oldChild-expected.txt: Added.
        * fast/dom/Document/replaceChild-null-oldChild.html: Added.
        * fast/dom/Document/resources/replaceChild-null-oldChild.js: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28354 268f45cc-cd09-0410-ab3c-d52691b4dbfc

* Avoid crashes by making sure everything is layouted before
  we start painting. This avoids a crash in Widget::invalidateRect
  because QPainter::begin would fail
* The QWebFrame::layout() methods and calls are left untouched because
  this would be an API decision.

Signed-off-by: Simon
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28353 268f45cc-cd09-0410-ab3c-d52691b4dbfc

* Implement GraphicsContext::fillRoundedRect similiar to the
  CoreGraphics implementation only addPath is not called this
  seems to be only needed for CoreGraphics.

Signed-off-by: Simon
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28352 268f45cc-cd09-0410-ab3c-d52691b4dbfc

* SubClass QWebPage to handle hide and close events.
* Forward these events to the InspectorController
* The other options would have been using an eventFilter
  on the webpage and making InspectorClient a QObject or
  creating a QObject Observer.
* Provide a simple QWebPage* createWindow implementation. This is needed
  to make FrameLoaderClientQt::dispatchCreatePage stop crashing in the case
  of the inspector client.

Signed-off-by: Simon
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28351 268f45cc-cd09-0410-ab3c-d52691b4dbfc

* Coding Style fixes

Signed-off-by: Simon
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28350 268f45cc-cd09-0410-ab3c-d52691b4dbfc

* Implement the InspectorClient for the Qt port
* It does not support highlighting of nodes yet
* Use QRC to open the internal page. The important thing is the
 '/' in the URL to make WebCore::Document::completeURL behave the
  way we want.
* To make the InspectorClient work we will have to mark qrc as secure.

Signed-off-by: Simon
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28349 268f45cc-cd09-0410-ab3c-d52691b4dbfc

* Allow adding the Inspect ContextMenuItem to our Menu
* Open the Inspector when Inspect was activated. This requires
  that we keep the innerNonSharedNode from the HitTest around. This
  forces us to include <wtf/RefPtr.h> in the private header. It is
  the first non Qt header but should be okay.

Signed-off-by: Simon
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28348 268f45cc-cd09-0410-ab3c-d52691b4dbfc

* Add the Developer Extras to the WebAttribute and propagate it
  to WebCore::Settings. This will enable the Inspect Element menu item
* CodingStyle fixes in QWebSettings
* Enable the Developer Extras in the QtLauncher

Signed-off-by: Simon
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28347 268f45cc-cd09-0410-ab3c-d52691b4dbfc

JavaScriptCore:

        Reviewed by Mitz.

        - fix http://bugs.webkit.org/show_bug.cgi?id=15848
          <rdar://problem/5619330> REGRESSION: Assertion failure viewing comments page on digg.com

        Test: fast/js/sparse-array.html

        * kjs/array_instance.cpp:
        (KJS::ArrayInstance::inlineGetOwnPropertySlot): Check sparse array cutoff before looking
        in hash map. Can't avoid the branch because we can't look for 0 in the hash.
        (KJS::ArrayInstance::deleteProperty): Ditto.

LayoutTests:

        Reviewed by Mitz.

        - test for http://bugs.webkit.org/show_bug.cgi?id=15848
          <rdar://problem/5619330> REGRESSION: Assertion failure viewing comments page on digg.com

        * fast/js/resources/sparse-array.js: Added.
        * fast/js/sparse-array-expected.txt: Added.
        * fast/js/sparse-array.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28346 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2007-12-03  Alp Toker  <alp@atoker.com>

        Reviewed by Mark Rowe.

        Don't delete the decoder if it's already been created. The one we have
        is fine.

        Cairo image cleanups. Nothing substantial.

        Notify ImageObservers where appropriate.

        * platform/graphics/cairo/ImageCairo.cpp:
        (WebCore::BitmapImage::draw):
        (WebCore::Image::drawPattern):
        * platform/graphics/cairo/ImageSourceCairo.cpp:
        (WebCore::ImageSource::setData):
        * platform/graphics/gtk/ImageGtk.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28345 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Fixed crash when rendering form elements with Qt 4.4

QPainter::device() is not guaranteed to return a QWidget, so do the safe "cast" with the
help of QPaintDevice::devType().
Also fall back to the QApplication::style() if we don't have a widget.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@28344 268f45cc-cd09-0410-ab3c-d52691b4dbfc