14 years ago http/tests/xmlhttprequest/readystatechange.html fails on buildbot.
ap@webkit.org [Fri, 25 Jan 2008 14:42:37 +0000 (14:42 +0000)]
    http/tests/xmlhttprequest/readystatechange.html fails on buildbot.

        * platform/win/Skipped:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29791 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago Reviewed by Mark Rowe.
ap@webkit.org [Fri, 25 Jan 2008 11:45:18 +0000 (11:45 +0000)]
    Reviewed by Mark Rowe.

        <rdar://problem/5699933> http/tests/security/cross-frame-access-get.html is still failing

        * DumpRenderTree/win/DumpRenderTree.cpp: (main): Clear HTTP cache to ensure consistent test environment
        (matching Mac DRT).
        * DumpRenderTree/win/DumpRenderTree.vcproj: Link to CFNetwork.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29790 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago Build fix
sfalken@apple.com [Fri, 25 Jan 2008 01:13:38 +0000 (01:13 +0000)]
    Build fix

        * WebKit.vcproj/WebKit.sln:
        * WebKit.vcproj/WebKit.submit.sln:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29785 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago Build fix.
sfalken@apple.com [Fri, 25 Jan 2008 01:08:54 +0000 (01:08 +0000)]
    Build fix.

        * JavaScriptCore.vcproj/JavaScriptCore.sln:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29784 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago Build fix.
sfalken@apple.com [Fri, 25 Jan 2008 01:07:16 +0000 (01:07 +0000)]
    Build fix.

        * JavaScriptCore.vcproj/JavaScriptCoreSubmit.sln:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29783 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoWebKitTools:
weinig@apple.com [Fri, 25 Jan 2008 00:51:55 +0000 (00:51 +0000)]

        Reviewed by Ada Chan.

        Fix for <rdar://5131975> DumpRenderTree doesn't support undo/redo

        - Adds simplified UndoManager to windows DRT to allow testing the
          undo/redo code paths in WebCore and WebKit.  This is a temporary
          solution to an issue that should eventually be solved by having a
          shared UndoManager in WebKit.

        * DumpRenderTree/win/DumpRenderTree.cpp:
        * DumpRenderTree/win/UIDelegate.cpp:
        * DumpRenderTree/win/UIDelegate.h:


        Reviewed by Ada Chan.

        <rdar://5131975> DumpRenderTree doesn't support undo/redo

        * platform/win/Skipped: With undo/redo support enabled, we can now
        remove the undo tests from the Skipped list.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29782 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago Better fix for Bug 16996
aroben@apple.com [Thu, 24 Jan 2008 23:50:34 +0000 (23:50 +0000)]
    Better fix for Bug 16996

        Rubberstamped by Mitz.

        * loader/CachedFont.cpp:
        (WebCore::CachedFont::ensureCustomFontData): Don't call
        createFontCustomPlatformData if m_data is null.
        * platform/graphics/gtk/FontCustomPlatformData.cpp:
        (WebCore::createFontCustomPlatformData): Added an assertion.
        * platform/graphics/mac/FontCustomPlatformData.cpp:
        (WebCore::createFontCustomPlatformData): Ditto.
        * platform/graphics/qt/FontCustomPlatformData.cpp:
        (WebCore::createFontCustomPlatformData): Ditto.
        * platform/graphics/win/FontCustomPlatformData.cpp:
        (WebCore::createFontCustomPlatformData): Ditto.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29781 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago Fix Bug 16996: Crash in createFontCustomPlatformData when loading 0-byte font via...
aroben@apple.com [Thu, 24 Jan 2008 23:13:26 +0000 (23:13 +0000)]
 Fix Bug 16996: Crash in createFontCustomPlatformData when loading 0-byte font via @font-face


         Fix Bug 16996: Crash in createFontCustomPlatformData when loading
         0-byte font via @font-face


         Reviewed by Hyatt.

         Test: fast/loader/font-face-empty.html

         * platform/graphics/win/FontCustomPlatformData.cpp:
         (WebCore::createFontCustomPlatformData): Null-check the buffer


         Test for Bug 16996: Crash in createFontCustomPlatformData when loading
         0-byte font via @font-face


         Reviewed by Hyatt.

         * fast/loader/font-face-empty-expected.txt: Added.
         * fast/loader/font-face-empty.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29780 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoUpdate results for an editing test that changed from Acid3 fix.
hyatt@apple.com [Thu, 24 Jan 2008 22:02:11 +0000 (22:02 +0000)]
Update results for an editing test that changed from Acid3 fix.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29779 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago2008-01-24 Jon Honeycutt <jhoneycutt@apple.com>
jhoneycutt@apple.com [Thu, 24 Jan 2008 21:53:29 +0000 (21:53 +0000)]
2008-01-24  Jon Honeycutt  <jhoneycutt@apple.com>

        Reviewed by Anders.

        <rdar://problem/5588807> Crash in Flash when destroying plug-in (found
        using yahoo beta mail)

        Flash can dereference NULL in the call to NPP_URLNotify if a request
        made with NPN_PostURLNotify fails before NPP_NewStream is called.

        Work around this by creating a quirk, PluginQuirkFlashURLNotifyBug, and
        checking for this quirk before calling NPP_URLNotify for any request
        made with NPN_PostURLNotify. If the quirk is present, call NPP_NewStream
        and NPP_DestroyStream before calling NPP_URLNotify.

        * WebCore.vcproj/WebCore.vcproj:
        * plugins/PluginQuirkSet.h: Added. Moved quirks out of PluginViewWin so
        PluginViewWin and PluginStream could share it. Created a class,
        PluginQuirkSet, to store plug-in quirks
        * plugins/PluginStream.cpp:
        (WebCore::PluginStream::PluginStream): Copy the PluginQuirkSet for this
        (WebCore::PluginStream::destroyStream): Check for the FlashURLNotifyBug
        * plugins/PluginStream.h:
        * plugins/win/PluginViewWin.cpp:
        (WebCore::PluginViewWin::performRequest): Pass the quirks when creating
        the PluginStream
        (WebCore::PluginViewWin::didReceiveResponse): Same
        (WebCore::PluginViewWin::wndProc): Use add / contains instead of
        bit ops
        (WebCore::PluginViewWin::userAgent): Same
        (WebCore::PluginViewWin::invalidateRect): Same
        (WebCore::PluginViewWin::~PluginViewWin): Same
        (WebCore::PluginViewWin::determineQuirks): Same
        (WebCore::PluginViewWin::setParameters): Same
        (WebCore::PluginViewWin::PluginViewWin): Same
        (WebCore::PluginViewWin::init): Same
        (WebCore::PluginViewWin::setCallingPlugin): Same
        * plugins/win/PluginViewWin.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29778 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoAdd test results for Acid3 buckets parsing issue.
hyatt@apple.com [Thu, 24 Jan 2008 21:52:07 +0000 (21:52 +0000)]
Add test results for Acid3 buckets parsing issue.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29777 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoAdd test of Acid3 buckets parsing issue.
hyatt@apple.com [Thu, 24 Jan 2008 21:51:37 +0000 (21:51 +0000)]
Add test of Acid3 buckets parsing issue.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29776 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago http://bugs.webkit.org/show_bug.cgi?id=16982
hyatt@apple.com [Thu, 24 Jan 2008 21:49:25 +0000 (21:49 +0000)]

        Make sure to make <head> the current block if it is created before a <body> already exists.

        Reviewed by Dan

        * html/HTMLParser.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29775 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago http://bugs.webkit.org/show_bug.cgi?id=16993
ap@webkit.org [Thu, 24 Jan 2008 19:40:05 +0000 (19:40 +0000)]
        <rdar://problem/5704331> REGRESSION: Loading page leads to many unexpected redirections

        Rolled out r29590, which was not a correct fix for <rdar://problem/5692566>.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29774 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoWebCore:
antti@apple.com [Thu, 24 Jan 2008 18:24:59 +0000 (18:24 +0000)]

        Reviewed by Adam.

        If the media is playing and the load stalls the playback wont restart by seeking backwards.

        We should be in CAN_PLAY state if the current time is less than the maximum loaded time. Rate
        tricks were for streaming case and are not necessary now since that is now disabled.

        Windows patch coming separately.

        Test: http/tests/media/video-play-stall-seek.html

        * platform/graphics/mac/MediaPlayerPrivateQTKit.mm:


        Reviewed by Adam.

        Test that playback can be resumed by seeking backwards after a stall.

        * http/tests/media/video-play-stall-seek-expected.txt: Added.
        * http/tests/media/video-play-stall-seek.html: Added.
        * platform/win/Skipped:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29773 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago Reviewed by Adam.
antti@apple.com [Thu, 24 Jan 2008 18:18:30 +0000 (18:18 +0000)]
    Reviewed by Adam.

        Fix <rdar://problem/5684815>
        After navigating back to the page that contains a <audio> that has been muted, the audio is still heard

        - get rid of the separate muted state in MediaPlayer, maintain the state in the cross platform code only
        - remove volumeChanged() callbacks from HTMLMediaElement and OS X MediaPlayer, they were not used for anything
        - rename updateMediaPlayer -> updatePlayState which tells more about what it actually does

        * html/HTMLMediaElement.cpp:
        * html/HTMLMediaElement.h:
        * platform/graphics/MediaPlayer.cpp:
        * platform/graphics/MediaPlayer.h:
        * platform/graphics/mac/MediaPlayerPrivateQTKit.h:
        * platform/graphics/mac/MediaPlayerPrivateQTKit.mm:
        * platform/graphics/win/MediaPlayerPrivateQuickTimeWin.h:
        * platform/graphics/win/QTMovieWin.cpp:
        * platform/graphics/win/QTMovieWin.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29772 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago Build fix
aroben@apple.com [Thu, 24 Jan 2008 17:38:33 +0000 (17:38 +0000)]
    Build fix

        * WebView.cpp: Added a comma.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29771 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago Allow directories containing ChangeLogs to be passed to resolve-ChangeLogs
aroben@apple.com [Thu, 24 Jan 2008 17:27:26 +0000 (17:27 +0000)]
     Allow directories containing ChangeLogs to be passed to resolve-ChangeLogs

         Reviewed by David Kilzer.

         * Scripts/resolve-ChangeLogs:
         (sub findChangeLog): Return a ChangeLog contained within the specified
         directory if one exists.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29770 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago Reviewed by Adam Roben.
ap@webkit.org [Thu, 24 Jan 2008 16:36:54 +0000 (16:36 +0000)]
    Reviewed by Adam Roben.

        <rdar://problem/5702734> REGRESSION(r28936/r15890): Shift+del does not work in win Safari

        * WebView.cpp: Fixed keyDownEntries to list this shortcut (and related ones, for consistency).

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29769 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoHolger Hans Peter Freyther <holger.freyther@trolltech.com>
hausmann@webkit.org [Thu, 24 Jan 2008 14:52:05 +0000 (14:52 +0000)]
Holger Hans Peter Freyther <holger.freyther@trolltech.com>

* Fix "QObject::startTimer: QTimer can only be used with threads started with QThread"
* The JSC GCController gets automatically destructed on application exit, it will stop
  its timer and we try to schedule the next timer. Now the GCController can get destructed
  after our QApplication is gone. This will trigger the above warning, we can avoid this
  by checking if our qApp is still around.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29768 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoMorten Johan Sørvig <msorvig@trolltech.com>
hausmann@webkit.org [Thu, 24 Jan 2008 14:27:58 +0000 (14:27 +0000)]
Morten Johan Sørvig <msorvig@trolltech.com>

Qt/Mac: Make sure the scrollbars does not overlap the grow box.

When showing only one scrollbar we need to move it so it doesn't overlap the
grow box. This is similar to the code in QAbstractScrollArea.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29767 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoDavid Boddie <dboddie@trolltech.com>
hausmann@webkit.org [Thu, 24 Jan 2008 14:24:48 +0000 (14:24 +0000)]
David Boddie <dboddie@trolltech.com>

Documentation tidying.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29766 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoJarek Kobus <jkobus@trolltech.com>
hausmann@webkit.org [Thu, 24 Jan 2008 14:23:50 +0000 (14:23 +0000)]
Jarek Kobus <jkobus@trolltech.com>

Text for translations were used wrongly as comments

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29765 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoMichael Goddard <michael.goddard@trolltech.com>
hausmann@webkit.org [Thu, 24 Jan 2008 14:18:34 +0000 (14:18 +0000)]
Michael Goddard  <michael.goddard@trolltech.com>


Fix for CSS crash in -webkit-border-image.
Reviewed by Darin.

While parsing -webkit-border-image, store
the border widths as naked pointers rather
than as OwnPtrs, since they point to the
middle of an array.

Test: fast/css/border-image-crash.html

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29764 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoHolger Hans Peter Freyther <holger.freyther@trolltech.com>
hausmann@webkit.org [Thu, 24 Jan 2008 14:13:36 +0000 (14:13 +0000)]
Holger Hans Peter Freyther <holger.freyther@trolltech.com>

* We have a KURL->QUrl conversion on KURL itself, make use of it.
* This conversion is supposed to be loss-free

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29763 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoHolger Hans Peter Freyther <holger.freyther@trolltech.com>
hausmann@webkit.org [Thu, 24 Jan 2008 14:08:12 +0000 (14:08 +0000)]
Holger Hans Peter Freyther <holger.freyther@trolltech.com>

* Attempt to fix showing of popup again after a popup has been shown but
  the user did not select any item (e.g. clicked outside of the popup)
* Keep track of the popup status using the virtuals showPopup and hidePopup
  and do not go into recursion when hiding a popup.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29762 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoHolger Hans Peter Freyther <holger.freyther@trolltech.com>
hausmann@webkit.org [Thu, 24 Jan 2008 14:01:48 +0000 (14:01 +0000)]
Holger Hans Peter Freyther <holger.freyther@trolltech.com>

* Assert the presence of the PopupMenuClient

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29761 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoMichael Goddard <michael.goddard@trolltech.com>
hausmann@webkit.org [Thu, 24 Jan 2008 13:56:49 +0000 (13:56 +0000)]
Michael Goddard <michael.goddard@trolltech.com>

Fix QDateTime to JS Date conversion.
Several conversion errors (some UTC related, some month
offset related) and the conversion distance for Date
to DateTime conversion weights were fixed (it should never
be better to convert a JS Number into a Date rather than
an int).

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29760 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoMichael Goddard <michael.goddard@trolltech.com>
hausmann@webkit.org [Thu, 24 Jan 2008 13:51:26 +0000 (13:51 +0000)]
Michael Goddard <michael.goddard@trolltech.com>

Add support for calling QObjects.
Add support for invokeDefaultMethod (via a call to
a specific slot), and also allow using it as a
constructor, like QtScript.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29759 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoMichael Goddard <michael.goddard@trolltech.com>
hausmann@webkit.org [Thu, 24 Jan 2008 13:44:46 +0000 (13:44 +0000)]
Michael Goddard <michael.goddard@trolltech.com>

Code style cleanups.
Add spaces before/after braces in inline function.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29758 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoMichael Goddard <michael.goddard@trolltech.com>
hausmann@webkit.org [Thu, 24 Jan 2008 13:41:14 +0000 (13:41 +0000)]
Michael Goddard <michael.goddard@trolltech.com>

Code style cleanups.
Remove spaces and unneeded declared parameter names.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29757 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoMichael Goddard <michael.goddard@trolltech.com>
hausmann@webkit.org [Thu, 24 Jan 2008 13:37:49 +0000 (13:37 +0000)]
Michael Goddard <michael.goddard@trolltech.com>

Clear stale RuntimeObjectImps.
Since other objects can have refs to the QtInstance,
we can't rely on the QtInstance being deleted when the
RuntimeObjectImp is invalidate or deleted.  This
could result in a stale JSObject being returned for
a valid Instance.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29756 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoVersioning.
mrowe@apple.com [Thu, 24 Jan 2008 07:44:50 +0000 (07:44 +0000)]

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29755 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago <rdar://problem/5702947> WebCore: CGContextDrawPDFDocument will be deprecated
ddkilzer@apple.com [Thu, 24 Jan 2008 00:47:00 +0000 (00:47 +0000)]
    <rdar://problem/5702947> WebCore: CGContextDrawPDFDocument will be deprecated

        Reviewed by Darin.

        No test cases added since there is no change in behavior.

        * platform/graphics/cg/PDFDocumentImage.cpp:
        (WebCore::PDFDocumentImage::draw): Replace CGContextDrawPDFDocument() with
        calls to CG methods that perform the equivalent work.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29753 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago Reviewed by Hyatt, Adele.
antti@apple.com [Wed, 23 Jan 2008 23:08:54 +0000 (23:08 +0000)]
    Reviewed by Hyatt, Adele.

        Timers for media controls keep running when the page is in the page cache.

        Stop timers in RenderMedia and get rid of the controls tree when the page goes to the cache.

        * html/HTMLMediaElement.cpp:
        * html/HTMLMediaElement.h:
        * rendering/RenderMedia.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29752 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago Reviewed by Adam.
antti@apple.com [Wed, 23 Jan 2008 22:26:01 +0000 (22:26 +0000)]
    Reviewed by Adam.

        Video that had already played to the end would start playing (from beginning) on back/forward navigation.

        Use setPausedInternal() instead of pause(). This avoid generating pause/play events and matches specification text.

        * html/HTMLMediaElement.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29751 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago Reviewed by Darin.
ap@webkit.org [Wed, 23 Jan 2008 22:20:55 +0000 (22:20 +0000)]
    Reviewed by Darin.

        <rdar://problem/4200075> Missing support for accented chars in mailto forms

        With some combinations of form parameters, this was resulting in regressed behavior.

        Tests: fast/forms/mailto/get-non-ascii-text-plain-latin-1.html

        * html/HTMLFormElement.cpp:
        (WebCore::HTMLFormElement::formData): Ignore accept-charset for mailto forms, which are always
        encoded as UTF-8.
        (WebCore::HTMLFormElement::isMailtoForm): Added.
        (WebCore::HTMLFormElement::dataEncoding): Added (factored out from formData).
        (WebCore::HTMLFormElement::submit): When round-tripping in text/plain case, use utf-8 encoding.
        * html/HTMLFormElement.h: Added a isMailtoForm() function. It may be a bit unfortunate that
        its result is calculated repeatedly, but this doesn't seem to be anywhere near hot code.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29750 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoWebCore:
sfalken@apple.com [Wed, 23 Jan 2008 22:04:16 +0000 (22:04 +0000)]

        Reviewed by Alice Liu.

        <rdar://problem/5658727> Undo/redo for pasted or typed text in yahoo mail becomes disabled after one round of being selected

        * rendering/RenderTextControl.cpp:
        (WebCore::RenderTextControl::updateFromElement): If value == text() and
        m_innerText is empty, then value and text() are "". In that case both DOM
        mutations are no-ops and there is no reason to clear the Undo stack.


        <rdar://problem/5698732> Copyright strings should include 2008

        Reviewed by Sam.

        * WebKit.vcproj/WebKit.rc:


        <rdar://problem/5698732> Copyright strings should include 2008

        Reviewed by Sam.

        * Drosera/win/Drosera.vcproj/Drosera.rc:
        * DumpRenderTree/win/TestNetscapePlugin/TestNetscapePlugin.rc:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29749 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoWebCore:
justin.garcia@apple.com [Wed, 23 Jan 2008 21:51:13 +0000 (21:51 +0000)]

        Reviewed by Alice Liu.

        <rdar://problem/5658727> Undo/redo for pasted or typed text in yahoo mail becomes disabled after one round of being selected

        * rendering/RenderTextControl.cpp:
        (WebCore::RenderTextControl::updateFromElement): If value == text() and
        m_innerText is empty, then value and text() are "". In that case both DOM
        mutations are no-ops and there is no reason to clear the Undo stack.


        Reviewed by Alice Liu.

        <rdar://problem/5658727> Undo/redo for pasted or typed text in yahoo mail becomes disabled after one round of being selected

        * editing/undo/5658727-expected.txt: Added.
        * editing/undo/5658727.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29748 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago Fix <rdar://5698502> Inspector window reverts to default position/size unexp...
aroben@apple.com [Wed, 23 Jan 2008 21:08:15 +0000 (21:08 +0000)]
     Fix <rdar://5698502> Inspector window reverts to default position/size unexpectedly

         Reviewed by Hyatt.

         * WebInspectorClient.cpp:
         (defaultWindowRect): Added.
             - Use defaultWindowRect() to set the initial size/position of the
               Inspector window.
             - Pass the Inspector window's client rect to
               IWebView::initWithFrame so that the WebView will be initially
               sized to fill the window.
         (WebInspectorClient::showWindow): Don't move/resize the window back to
         the defaults every time this method is called.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29747 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago Fix <rdar://5700620> REGRESSION (r28190): "Open frame in new tab" instead of "Open...
aroben@apple.com [Wed, 23 Jan 2008 21:07:47 +0000 (21:07 +0000)]
 Fix <rdar://5700620> REGRESSION (r28190): "Open frame in new tab" instead of "Open link in new tab" @ digg.com


         Fix <rdar://5700620> REGRESSION (r28190): "Open frame in new tab" instead of "Open link in new tab" @ digg.com

         Test: fast/events/contextmenu-scrolled-page-with-frame.html

         Reviewed by Dan.

         * WebView.cpp:
         (WebView::handleContextMenuEvent): Translate the mouse coordinates
         into document coordinates before performing the hit-test. This is what
         EventHandler does internally.


         Add a test for <rdar://5700620> REGRESSION (r28190): "Open frame in
         new tab" instead of "Open link in new tab" @ digg.com

         Reviewed by Dan.

         * fast/events/contextmenu-scrolled-page-with-frame-expected.txt: Added.
         * fast/events/contextmenu-scrolled-page-with-frame.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29746 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoWebCore:
sfalken@apple.com [Wed, 23 Jan 2008 19:43:14 +0000 (19:43 +0000)]

        <rdar://problem/5699509> Allow file upload dialog to be localized.

        Reviewed by Darin.

        * platform/LocalizedStrings.h:
        * platform/win/FileChooserWin.cpp:


        Update localization exceptions.

        Reviewed by Darin.

        * English.lproj/Localizable.strings:
        * StringsNotToBeLocalized.txt:


        <rdar://problem/5699509> Allow file upload dialog to be localized.

        Reviewed by Darin.

        * English.lproj/Localizable.strings: Updated.
        * WebCoreLocalizedStrings.cpp:
        (WebCore::uploadFileText): Added.
        (WebCore::allFilesText): Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29745 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago Reviewed by Darin.
dsmith@webkit.org [Wed, 23 Jan 2008 17:49:23 +0000 (17:49 +0000)]
    Reviewed by Darin.

        Added a modified version of the MooTools SlickSpeed Selector Test/Benchmark to show off querySelectorAll. MIT licensed.

        * perf/slickspeed: Added.
        * perf/slickspeed/config.ini: Added.
        * perf/slickspeed/footer.html: Added.
        * perf/slickspeed/frameworks: Added.
        * perf/slickspeed/frameworks/DomQuery.js: Added.
        * perf/slickspeed/frameworks/dummy.js: Added.
        * perf/slickspeed/frameworks/jquery-1.2.1.js: Added.
        * perf/slickspeed/frameworks/prototype.js: Added.
        * perf/slickspeed/header.html: Added.
        * perf/slickspeed/index.php: Added.
        * perf/slickspeed/instructions.txt: Added.
        * perf/slickspeed/logo.png: Added.
        * perf/slickspeed/selectors.list: Added.
        * perf/slickspeed/style.css: Added.
        * perf/slickspeed/system: Added.
        * perf/slickspeed/system/index.php: Added.
        * perf/slickspeed/system/slickspeed.js: Added.
        * perf/slickspeed/system/template.php: Added.
        * perf/slickspeed/template.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29744 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago Fix behavior of type-to-select in <select> menus
aroben@apple.com [Wed, 23 Jan 2008 17:38:17 +0000 (17:38 +0000)]
     Fix behavior of type-to-select in <select> menus

         In r27825 we started posting both a WM_KEYDOWN and WM_CHAR message
         whenever a key was pressed in the dropdown menu. However, since posted
         messages go through TranslateMessage, the WM_KEYDOWN was generating
         another WM_CHAR message, so two WM_CHAR messages in a row would reach
         the WebView. This caused the searching to happen twice for each key,
         so typing "dog" would search for "ddoogg".

         Reviewed by Darin.

         * platform/win/PopupMenuWin.cpp:
         (WebCore::PopupWndProc): Repost the WM_KEYDOWN message to the WebView.
         This is will generate a WM_CHAR message which will trigger the
         type-to-select behavior. This is very similar to our behavior prior to
         r27825, except that we're now calling PostMessage instead of

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29743 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago - updated results following <http://trac.webkit.org/projects/webkit/changeset...
mitz@apple.com [Wed, 23 Jan 2008 16:18:24 +0000 (16:18 +0000)]
    - updated results following <trac.webkit.org/projects/webkit/changeset/29721>

        * platform/mac/fast/frames/viewsource-attribute-expected.checksum:
        * platform/mac/fast/frames/viewsource-attribute-expected.png:
        * platform/mac/fast/frames/viewsource-attribute-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29742 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago2008-01-23 Alp Toker <alp@atoker.com>
alp@webkit.org [Wed, 23 Jan 2008 15:55:41 +0000 (15:55 +0000)]
2008-01-23  Alp Toker  <alp@atoker.com>

        Rubber-stamped by Mark Rowe.

        Remove whitespace after -I in automake include lists.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29741 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoHolger Hans Peter Freyther <holger.freyther@trolltech.com>
hausmann@webkit.org [Wed, 23 Jan 2008 11:16:45 +0000 (11:16 +0000)]
Holger Hans Peter Freyther <holger.freyther@trolltech.com>

* Fix leaking of sub frames (WebCore::Frame). We keep one reference too many.
  This was found while working on the page-cache, other ports are not affected.
* Rubber stamped by Simon

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29740 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoHolger Hans Peter Freyther <holger.freyther@trolltech.com>
hausmann@webkit.org [Wed, 23 Jan 2008 11:14:06 +0000 (11:14 +0000)]
Holger Hans Peter Freyther <holger.freyther@trolltech.com>

* Load images incrementally. This will show warnings on the console
  and we want to fix them for Qt4.4. If that happens we have to remove
  the comment from the enum inside the ImageDecoderQt.cpp

* Rubber stamped by Simon

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29739 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoRohan McGovern <rohan.mcgovern@trolltech.com>
hausmann@webkit.org [Wed, 23 Jan 2008 11:11:27 +0000 (11:11 +0000)]
Rohan McGovern <rohan.mcgovern@trolltech.com>

Fix Qtopia compilation with QT_NO_TOOLTIP

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29738 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoGeir Vattekar <gvatteka@trolltech.com>
hausmann@webkit.org [Wed, 23 Jan 2008 11:11:13 +0000 (11:11 +0000)]
Geir Vattekar <gvatteka@trolltech.com>

Doc: Replaced \code with snippets in the docs

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29737 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoLars Knoll <lars@trolltech.com>
hausmann@webkit.org [Wed, 23 Jan 2008 11:10:57 +0000 (11:10 +0000)]
Lars Knoll <lars@trolltech.com>

Fix rendering of the Scrollbar as well as mouse handling for some styles.

The QStyle expects that that painter is set up to clip to the scrollbar bounds
and some of the styles seem to paint somewhat outside of these bounds. Clipping to
the scrollbar bounds before drawing removes some artifacts.

Also set m_opt.rect.topLeft to (0/0) in the mouse handlers. Fixes some issues with
the oxygen style.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29736 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoMichael Brasser <michael.brasser@trolltech.com>
hausmann@webkit.org [Wed, 23 Jan 2008 09:55:46 +0000 (09:55 +0000)]
Michael Brasser <michael.brasser@trolltech.com>

Rendering fixes for Qtopia.

Make the focus ring work with line breaks (until we have QPainterPath::simplify) and
call QStyle with better arguments when painting themed elements.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29735 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoHolger Hans Peter Freyther <holger.freyther@trolltech.com>
hausmann@webkit.org [Wed, 23 Jan 2008 09:55:28 +0000 (09:55 +0000)]
Holger Hans Peter Freyther <holger.freyther@trolltech.com>

* FormData::flatten does not include to be uploaded files. Create our own QIODevice
  for the QNetworkAccessManager and feed the complete content of the FormData* to
  the (HTTP) backend.
* Try to optimize the ::readData implementation to fill the buffer as best as possible

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29734 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoHolger Hans Peter Freyther <holger.freyther@trolltech.com>
hausmann@webkit.org [Wed, 23 Jan 2008 09:55:08 +0000 (09:55 +0000)]
Holger Hans Peter Freyther <holger.freyther@trolltech.com>

* Coding-Style fixes

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29733 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoMichael Brasser <michael.brasser@trolltech.com>
hausmann@webkit.org [Wed, 23 Jan 2008 09:48:22 +0000 (09:48 +0000)]
Michael Brasser <michael.brasser@trolltech.com>

use PopupClient's font for popup menu. Otherwise, in Qtopia, the popup's font is much bigger (QApplication::font()?) and the text doesn't fit.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29732 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoBenjamin Meyer <bmeyer@trolltech.com>
hausmann@webkit.org [Wed, 23 Jan 2008 09:44:09 +0000 (09:44 +0000)]
Benjamin Meyer <bmeyer@trolltech.com>

Tweak key presses even to match commonly expected behavior
- space key == page down
- page down moved down not a page, but slightly less then a page so you don't loose your spot when reading.
- Use font height rather then a hard coded "10" for left, right, up, down
- Ctrl-Up moves to the top of the page
- Ctrl-Down move to the bottom of the page
- Backspace == GoBack
- Shift-Backspace == GoForward

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29731 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoImplemented FrameLoaderClient::startDownload() and FrameLoaderClient::download().
hausmann@webkit.org [Wed, 23 Jan 2008 09:43:52 +0000 (09:43 +0000)]
Implemented FrameLoaderClient::startDownload() and FrameLoaderClient::download().

Added two signals to QWebPage to handle downloading of links and handling of
unsupported content.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29730 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoMichael Goddard <michael.goddard@trolltech.com>
hausmann@webkit.org [Wed, 23 Jan 2008 09:39:04 +0000 (09:39 +0000)]
Michael Goddard <michael.goddard@trolltech.com>

Reworked the JavaScriptCore Qt bindings:

* Add initial support for string and variant arrays, as well
  as sub QObjects in the JS bindings.

* Don't expose fields marked as not scriptable by moc.

* Add support for dynamic properties and accessing named
  QObject children of an object (like QtScript and older
  IE DOM style JS).
* Add support for custom toString methods.

* Fine tune some bindings to be closer to QtScript.
  Make void functions return undefined, and empty/
  null QStrings return a zero length string.

* Create framework for allowing more direct method calls.
  Since RuntimeMethod doesn't allow us to add additional
  methods/properties to a function, add these classes.
  Start prototyping object.signal.connect(...).

* Add signal support to the Qt bindings.
  Allow connecting to signals (object.signal.connect(slot)),
  disconnecting, and emitting signals.  Currently chooses
  the first signal that matches the name, so this will need

* Add property names, and resolve signals closer to use.
  Enumerating properties now returns some of the Qt properties
  and signals.  Slots and methods aren't quite present.  Also,
  resolve signal connections etc. closer to the time of use, so
  we can do more dynamic resolution based on argument type etc.
  Still picks the first one with the same name, at the moment.

* Make signature comparison code consistent.
  Use the same code for checking meta signatures in
  the method and fallback getters, and avoid a
  QByteArray construction when we can.

* Fix minor memory leak, and handle pointers better.
  Delete the private object in the dtors, and use RefPtrs
  for holding Instances etc.

* Handle method lookup better.
  Allow invocation time method lookup based on the arguments,
  which is closer to QtScript behaviour.  Also, cache the
  method lists and delete them in the QtClass dtor (stops
  a memory leak).

* Improve JS to Qt data type conversions.
  Add some support for Date & RegExp JS objects,
  and provide some metrics on the quality of the

* A couple of fixes for autotest failures.
  Better support for converting lists, read/write only
  QMetaProperty support, modified slot search order...)

* Update JS DRT controller for Qt JS binding changes.
  There were two functions that needed some changes
  so that the layout tests would work, so this makes
  a few tests pass again.

* Bump the timeout for layout tests up to 11s.
  At least some tests have an internal timeout of
  10 seconds, so make the waitUntilDone approach
  wait at least 11s.  fast/dom/open-and-close-by-DOM.html
  is one of these - now the failure message is more

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29729 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago Reviewed by Adele Peterson.
mitz@apple.com [Wed, 23 Jan 2008 04:01:09 +0000 (04:01 +0000)]
    Reviewed by Adele Peterson.

        - fix http://bugs.webkit.org/show_bug.cgi?id=16905
          <rdar://problem/5692407> REGRESSION (3.0.4-TOT): "menu" pseudocolor is badly chosen

        The regression test for this is fast/css/css2-system-color.html which
        is currently disabled.

        * rendering/RenderThemeMac.mm:
        (WebCore::menuBackgroundColor): Added. Uses HIThemeDrawMenuBackground
        to draw the menu item background into a bitmap graphics context and
        returns the color of the pixel at (0, 0).
        (WebCore::RenderThemeMac::systemColor): Changed to call
        menuBackgroundColor for the CSS2 menu color.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29728 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago Reviewed by Oliver Hunt.
darin@apple.com [Wed, 23 Jan 2008 00:29:49 +0000 (00:29 +0000)]
    Reviewed by Oliver Hunt.

        - turn full repainting back on for <canvas> until we get incremental
          repaint working properly

        Ollie and I came up with a plan for testing this, but for now we should
        just fall back on the "repaint everything" behavior we have always had.

        * html/HTMLCanvasElement.cpp:
        (WebCore::HTMLCanvasElement::willDraw): Just call repaint.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29727 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoJavaScriptCore:
andersca@apple.com [Tue, 22 Jan 2008 23:16:37 +0000 (23:16 +0000)]

        Reviewed by Darin and Adam.

        div element on microsoft site has wrong left offset.

        Return true even if NPN_GetProperty returns null or undefined. This matches Firefox
        (and is what the Silverlight plug-in expects).

        * bindings/NP_jsobject.cpp:


        Reviewed by Darin and Adam.

        div element on microsoft site has wrong left offset.

        Add new method for testing the return vale of NPN_GetProperty.

        * DumpRenderTree/TestNetscapePlugIn.subproj/PluginObject.cpp:


        Reviewed by Darin and Adam.

        div element on microsoft site has wrong left offset.

        Add test.

        * plugins/netscape-get-property-return-value-expected.txt: Added.
        * plugins/netscape-get-property-return-value.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29726 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago2008-01-22 Alp Toker <alp@atoker.com>
alp@webkit.org [Tue, 22 Jan 2008 21:23:44 +0000 (21:23 +0000)]
2008-01-22  Alp Toker  <alp@atoker.com>

        GTK+/qmake build fix.

        * WebCore.pro:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29725 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago Reviewed by Adam Roben.
ap@webkit.org [Tue, 22 Jan 2008 20:46:16 +0000 (20:46 +0000)]
    Reviewed by Adam Roben.

        <rdar://problem/5670257> editing/selection/4895428-3.html makes editing/selection/5131716-2.html fail

        * DumpRenderTree/win/EventSender.cpp: (makeEventSender): Reset static variables for a new test.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29724 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago2008-01-22 Christian Dywan <christian@imendio.com>
alp@webkit.org [Tue, 22 Jan 2008 20:45:18 +0000 (20:45 +0000)]
2008-01-22  Christian Dywan  <christian@imendio.com>

        Reviewed by Alp Toker.

        [GTK] API: WebKitWebSettings is not usable

        Implement WebKitWebSettings.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29723 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago Fix <rdar://5693599> REGRESSION (r29369): scroll bars are drawing when they shouldn...
aroben@apple.com [Tue, 22 Jan 2008 19:33:26 +0000 (19:33 +0000)]
 Fix <rdar://5693599> REGRESSION (r29369): scroll bars are drawing when they shouldn't be on an ad on cnn.com


         Add HTMLFrameOwnerElement::scrollingMode

         This new virtual method is overridden by HTMLFrameElementBase to
         provide the scrolling mode set on the frame or iframe element. Object
         elements always have scrolling=auto, so there's no need to override
         this method for HTMLPlugInElement.

         Reviewed by Hyatt.

         * html/HTMLFrameElementBase.h: Declare scrollingMode as virtual.
         * html/HTMLFrameOwnerElement.h:
         (WebCore::HTMLFrameOwnerElement::scrollingMode): Added.


         Fix <rdar://5693599> REGRESSION (r29369): scroll bars are drawing when they shouldn't be on an ad on cnn.com

         Now that we have multiple FrameViews per Frame on Windows, we need to
         make sure each FrameView gets its scrolling mode set correctly. On Mac
         this works because WebFrameView holds the scrolling mode and a single
         WebFrameView is shared by many FrameViews. On Windows, we have no such
         shared object, so we just ask the Frame's owner element for the
         scrolling mode.

         Tests: fast/frames/frame-scrolling-attribute.html

         Reviewed by Hyatt.

         * WebFrame.cpp:
         (WebFrame::transitionToCommittedForNewPage): Set the scrolling mode
         on the FrameView based on the scrolling mode of the Frame's owner


         Tests for <rdar://5693599> REGRESSION (r29369): scroll bars are drawing when they shouldn't be on an ad on cnn.com

         Reviewed by Hyatt.

         * fast/frames/frame-scrolling-attribute.html: Added.
         * fast/frames/iframe-scrolling-attribute.html: Added.
         * platform/mac/fast/frames/frame-scrolling-attribute-expected.checksum: Added.
         * platform/mac/fast/frames/frame-scrolling-attribute-expected.png: Added.
         * platform/mac/fast/frames/frame-scrolling-attribute-expected.txt: Added.
         * platform/mac/fast/frames/iframe-scrolling-attribute-expected.checksum: Added.
         * platform/mac/fast/frames/iframe-scrolling-attribute-expected.png: Added.
         * platform/mac/fast/frames/iframe-scrolling-attribute-expected.txt: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29722 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago Fix for <rdar://problem/5698481> REGRESSION: Web Inspector source view gutter...
hyatt@apple.com [Tue, 22 Jan 2008 17:36:26 +0000 (17:36 +0000)]
    Fix for <rdar://problem/5698481> REGRESSION: Web Inspector source view gutter double divider and no padding

        Now that cellpadding cascades properly on <td>s, we need to make the view source padding rules in the user
        agent sheet have !important in order to override author settings.

        Reviewed by Tim

        * css/view-source.css:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29721 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoReviewed by Alp.
zimmermann@webkit.org [Tue, 22 Jan 2008 16:09:01 +0000 (16:09 +0000)]
Reviewed by Alp.
Allow to use SVGPaintServerSolid without RenderObject/RenderStyle. Used later to draw SVG Fonts in HTML.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29720 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoReviewed by Alp.
zimmermann@webkit.org [Tue, 22 Jan 2008 14:57:47 +0000 (14:57 +0000)]
Reviewed by Alp.

Initialize missing variables, m_lineGap / m_lineSpacing.
Only noticeable when trying to render HTML text using SVG Fonts (which is not possible with trunk, but with my local version).

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29719 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago2008-01-22 Alp Toker <alp@atoker.com>
alp@webkit.org [Tue, 22 Jan 2008 13:53:37 +0000 (13:53 +0000)]
2008-01-22  Alp Toker  <alp@atoker.com>

        Reviewed by Nikolas.

        Fix Cairo SVG fonts following changes in r29700.

        Set the font description's details as needed in FontPlatformData and
        eliminate GraphicsContext::setPlatformFont() since it was redundant
        and assumed all fonts were Cairo fonts.

        * platform/graphics/GraphicsContext.cpp:
        * platform/graphics/cairo/GraphicsContextCairo.cpp:
        * platform/graphics/gtk/FontPlatformDataGtk.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29718 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoReviewed by Eric.
zimmermann@webkit.org [Tue, 22 Jan 2008 11:10:45 +0000 (11:10 +0000)]
Reviewed by Eric.

- fix http://bugs.webkit.org/show_bug.cgi?id=16967
<rdar://problem/5699344> Reproducible crash when navigating back to a page using SVG fonts

Test: svg/custom/font-platformDestroy-crash.svg

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29717 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoLars Knoll <lars@trolltech.com>
hausmann@webkit.org [Tue, 22 Jan 2008 10:03:54 +0000 (10:03 +0000)]
Lars Knoll <lars@trolltech.com>

setup a family for cursive and fantasy fonts as well.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29716 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoLars Knoll <lars@trolltech.com>
hausmann@webkit.org [Tue, 22 Jan 2008 10:03:40 +0000 (10:03 +0000)]
Lars Knoll <lars@trolltech.com>

cleanup the CookieJarQt implementation and implement cookiesEnabled.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29715 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoSimon Hausmann <hausmann@webkit.org>
hausmann@webkit.org [Tue, 22 Jan 2008 10:03:26 +0000 (10:03 +0000)]
Simon Hausmann <hausmann@webkit.org>

Don't compile the ICO plugin when building against Qt >= 4.4

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29714 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoBenjamin Meyer <bmeyer@trolltech.com>
hausmann@webkit.org [Tue, 22 Jan 2008 10:03:14 +0000 (10:03 +0000)]
Benjamin Meyer <bmeyer@trolltech.com>

Implement ChromeClientQt::setToolTip
Implement ChromeClientQt::mouseDidMoveOverElement

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29713 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoLars Knoll <lars@trolltech.com>
hausmann@webkit.org [Tue, 22 Jan 2008 08:38:52 +0000 (08:38 +0000)]
Lars Knoll <lars@trolltech.com>

Ported the font implementation to use the new QFont features of Qt 4.4.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29712 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago Reviewed by Nikolas & Oliver.
eric@webkit.org [Tue, 22 Jan 2008 06:48:57 +0000 (06:48 +0000)]
    Reviewed by Nikolas & Oliver.

        Tested by svg/custom/svg-features.html

        * dom/DOMImplementation.cpp:
        (WebCore::isSVG10Feature): use ENABLE(SVG_*) conditionals
        (WebCore::isSVG11Feature): use ENABLE(SVG_*) conditionals

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29711 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoJavaScriptCore:
ggaren@apple.com [Tue, 22 Jan 2008 06:18:10 +0000 (06:18 +0000)]

        Reviewed by Maciej Stachowiak.

        Fixed http://bugs.webkit.org/show_bug.cgi?id=16909
        REGRESSION: Amazon.com crash (ActivationImp)

        (and a bunch of other crashes)

        Plus, a .7% SunSpider speedup to boot.

        Replaced the buggy currentExec and savedExec mechanisms with an
        explicit ExecState stack.

        * kjs/collector.cpp:
        (KJS::Collector::collect): Explicitly mark the ExecState stack.

        (KJS::Collector::reportOutOfMemoryToAllExecStates): Slight change in
        behavior: We no longer throw an exception in any global ExecStates,
        since global ExecStates are more like pseudo-ExecStates, and aren't
        used for script execution. (It's unclear what would happen if you left
        an exception waiting around in a global ExecState, but it probably
        wouldn't be good.)


        Reviewed by Maciej Stachowiak.

        Adapted WebCore to the fix for http://bugs.webkit.org/show_bug.cgi?id=16909
        REGRESSION: Amazon.com crash (ActivationImp)

        * bindings/js/kjs_proxy.cpp:
        (WebCore::KJSProxy::~KJSProxy): No convenient way to make this assertion
        anymore. (It wasn't firing for anyone, anyway, so it's no big loss.)

        * bindings/objc/WebScriptObject.mm:
        (+[WebScriptObject throwException:]): Use the ExecState stack, instead
        of currentExec.
        (-[WebScriptObject setException:]): ditto. Also, a slight change in
        behavior: If no ExecStates are active, we no longer throw an exception
        in the global ExecState. The JavaScriptCore ChangeLog explains why.
        This also matches the behavior of +throwException.


        Layout test for http://bugs.webkit.org/show_bug.cgi?id=16909
        REGRESSION: Amazon.com crash (ActivationImp)

        * fast/js/exec-state-marking-expected.txt: Added.
        * fast/js/exec-state-marking.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29710 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoNot reviewed. Try to fix Qt build, after the rmdir() fixes.
zimmermann@webkit.org [Tue, 22 Jan 2008 00:52:38 +0000 (00:52 +0000)]
Not reviewed. Try to fix Qt build, after the rmdir() fixes.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29709 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoNot reviewed. Another Wx build fix.
zimmermann@webkit.org [Tue, 22 Jan 2008 00:49:41 +0000 (00:49 +0000)]
Not reviewed. Another Wx build fix.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29708 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago - try to fix Qt build
darin@apple.com [Tue, 22 Jan 2008 00:30:41 +0000 (00:30 +0000)]
    - try to fix Qt build

        * platform/qt/FileSystemQt.cpp:
        (WebCore::deleteEmptyDirectory): QDir::root().rmdir() instead of QDir::rmdir().

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29707 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoNot reviewed. Yet another win/gtk build fix.
zimmermann@webkit.org [Tue, 22 Jan 2008 00:05:28 +0000 (00:05 +0000)]
Not reviewed. Yet another win/gtk build fix.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29706 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoNot reviewed. Another build fix for Gtk/Linux & Wx/Mac.
zimmermann@webkit.org [Mon, 21 Jan 2008 23:43:32 +0000 (23:43 +0000)]
Not reviewed. Another build fix for Gtk/Linux & Wx/Mac.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29705 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoNot reviewed. Build fix for Qt/Gtk & Wx.
zimmermann@webkit.org [Mon, 21 Jan 2008 23:26:06 +0000 (23:26 +0000)]
Not reviewed. Build fix for Qt/Gtk & Wx.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29704 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoNot reviewed. Build fix for ports that don't build SVG: wx.
zimmermann@webkit.org [Mon, 21 Jan 2008 23:16:59 +0000 (23:16 +0000)]
Not reviewed. Build fix for ports that don't build SVG: wx.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29703 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoNot reviewed. (First) Build fix for Qt/Gtk.
zimmermann@webkit.org [Mon, 21 Jan 2008 23:12:25 +0000 (23:12 +0000)]
Not reviewed. (First) Build fix for Qt/Gtk.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29702 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago2008-01-21 Jan Michael Alonzo <jmalonzo@unpluggable.com>
alp@webkit.org [Mon, 21 Jan 2008 23:06:23 +0000 (23:06 +0000)]
2008-01-21  Jan Michael Alonzo  <jmalonzo@unpluggable.com>

        Reviewed by Alp Toker.

        Get errors when cross-compile webkit-gtk

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29701 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoReviewed by Eric. Older parts reviewed by Dan.
zimmermann@webkit.org [Mon, 21 Jan 2008 22:56:53 +0000 (22:56 +0000)]
Reviewed by Eric. Older parts reviewed by Dan.

Fixes: http://bugs.webkit.org/show_bug.cgi?id=16880 (SVGCSSFontFace should die, instead integrate within the FontCache.)
Fixes: http://bugs.webkit.org/show_bug.cgi?id=16784 (In-document fonts work only as the first child to font-face-src)
Proper integration of SVG Fonts within the existing custom font concept, removing the need for a custom SVGCSSFontFace.
Integrate within the CSSFontFaceSource logic for caching support, as well as the framework for external SVG Fonts.

Proper handling of missing glyphs, using <missing-glyph> element. If that is not specified in a font, fallback to non-SVG font rendering.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29700 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago2008-01-21 Alp Toker <alp@atoker.com>
alp@webkit.org [Mon, 21 Jan 2008 22:43:52 +0000 (22:43 +0000)]
2008-01-21  Alp Toker  <alp@atoker.com>

        GTK+ build fix for breakage introduced in r29698.

        * WebCoreSupport/ChromeClientGtk.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29699 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoWebCore:
darin@apple.com [Mon, 21 Jan 2008 22:23:28 +0000 (22:23 +0000)]

        Reviewed by John Sullivan.

        - fix <rdar://problem/5644324> Delegate-less WebKit clients should have no databases

        - remove the default database quota setting

        - simplify the quota-related client calls by merging the one for a new database with
          the one for an existing database, adding a feature where you can get details about
          a database being created using the same functions that work on the other databases,
          and removing the parameters and return values, instead having clients use a function
          to set the quota

        - fix unsafe multi-thread access to the database tracker's quota map
        - fix bug in deleteAllDatabases where it would iterate a HashMap while modifying it

        The tracker database is now only created when we set the quota for a database origin.
        Thus asking for info about databases won't cause anything to be written to disk.

        * WebCore.base.exp: Updated.

        * bindings/js/GCController.cpp: Added an #if to get rid of an unused function warning.

        * page/Chrome.cpp: Removed database-related functions. There's no problem having the
        code deal directly with the client.
        * page/Chrome.h: Ditto. Also made Chrome inherit from Noncopyable.

        * page/ChromeClient.h: Replaced the two quota-related functions with a single one.
        The details about the state of databases in the origin are now available by asking
        for database details. There's also no need to pass the security origin, since
        it's easy to get that from the frame's document.

        * page/Settings.cpp: Removed the default quota setting.
        * page/Settings.h: Ditto.

        * platform/FileSystem.h: Added deleteEmptyDirectory.
        * platform/gtk/FileSystemGtk.cpp:
        (WebCore::deleteEmptyDirectory): Added.
        * platform/posix/FileSystemPOSIX.cpp:
        (WebCore::deleteEmptyDirectory): Added.
        * platform/qt/FileSystemQt.cpp:
        (WebCore::deleteEmptyDirectory): Added.
        * platform/win/FileSystemWin.cpp:
        (WebCore::deleteEmptyDirectory): Added.
        * platform/wx/FileSystemWx.cpp:
        (WebCore::deleteEmptyDirectory): Added placeholder.

        * storage/DatabaseDetails.h: Removed the isValid() function since its name is
        confusing -- we removed our other isValid() functions. For the few callers that
        need this, it's fine to just check name().isEmpty(). Made the member functions
        all const.

        * storage/DatabaseTracker.cpp:
        (WebCore::DatabaseTracker::DatabaseTracker): Removed m_defaultQuota initialization.
        Added m_proposedDatabase and m_thread.
        (WebCore::DatabaseTracker::setDatabaseDirectoryPath): Got rid of code that would
        open the tracker database here. This might slightly speed up launch time, since
        we won't do the I/O until we have to, and before we were actually creating an SQL
        database in this code that's run when we go to the first webpage.
        (WebCore::DatabaseTracker::trackerDatabasePath): Added.
        (WebCore::DatabaseTracker::openTrackerDatabase): Added a boolean parameter telling
        this function whether to create the database. Made this function safe to call if
        the database is already open. Also made the function fail quietly if the path has
        not been set yet.
        (WebCore::DatabaseTracker::canEstablishDatabase): Got rid of the call to the
        establishEntryForOrigin function. The origin is now put in the tracker database
        when the quota is set to a non-zero value. When judging if there's enough space
        for the new database, require at least one byte even if estimatedSize is 0, and
        check for overflow. Also added code here to populate the origins map, which
        guarantees it will be ready when the database uses it on another thread later.
        Also changed this to call the new ChromeClient function.
        (WebCore::DatabaseTracker::hasEntryForDatabase): Added code to open the
        tracker database, since that's no longer done by setDatabaseDirectoryPath.
        (WebCore::DatabaseTracker::originPath): Added.
        (WebCore::DatabaseTracker::fullPathForDatabase): Added code so that this will
        return a null string for the proposed database if called from within the
        ChromeClient function. Also switched from empty string to null string for the
        error cases.
        (WebCore::DatabaseTracker::populateOrigins): Added code to open the tracker
        (WebCore::DatabaseTracker::databaseNamesForOrigin): Ditto.
        (WebCore::DatabaseTracker::detailsForNameAndOrigin): Ditto. Also added code that
        will return the details of the proposed database if called from within the
        ChromeClient function. This is how the client can learn of the display name and
        the estimated size of the new database.
        (WebCore::DatabaseTracker::setDatabaseDetails): Added code to open the tracker
        (WebCore::DatabaseTracker::quotaForOrigin): Made this code OK to call on a non-main
        thread by using m_quotaMapGuard to guard access to the map. Other code runs on the
        main thread only, and only functions that write to the map use the lock.
        (WebCore::DatabaseTracker::setQuota): Changed this function so it can insert the
        initial quota as well as updating an existing quota. Added locking since this
        function modifies the quota map. Added code to open the tracker database. Added
        an early exit if the quota is already correct, which is guarantees that if you
        set a quota to 0 it won't trigger creation of a tracker database.
        (WebCore::DatabaseTracker::addDatabase): Added code to open the tracker database.
        (WebCore::DatabaseTracker::deleteAllDatabases): Made a copy of the quota map before
        iterating it to find all the origins. This fixes a problem with the old code where
        it would modify the map while iterating it, which gives assertions in debug builds
        and unpredictable results.
        (WebCore::DatabaseTracker::deleteOrigin): Replaced deleteDatabasesWithOrigin with
        this function. Added code to open the tracker database if needed. Added code to
        delete the origin from the tracker database, and to close the tracker database
        and delete files and directories as needed if we are deleting the final origin.
        (WebCore::DatabaseTracker::deleteDatabase): Added code to open the tracker database
        if needed.

        * storage/DatabaseTracker.h: Renamed databasePath to databaseDirectoryPath for
        clarity, including the data member, and the getter and setter functions.
        Replaced deleteDatabasesWithOrigin with deleteOrigin. Removed the functions
        dealing with default origin quota. There is no default any more; origins start
        with no quota and the client must set a quota. Added trackerDatabasePath and
        originPath helper functions. Added a boolean parameter to openTrackerDatabase
        to tell it whether to create the database or not. Removed the
        establishEntryForOrigin function. Renamed m_originQuotaMap to just m_quotaMap,
        and added m_quotaMapGuard. Added a QuotaMap typedef. Added m_proposedDatabase,
        which holds the origin and details for the current proposed database during
        the client callback function that must decide whether to grant quota. Added a
        m_thread data member for debugging use to assert if functions that can only
        be called on a single thread are misused.

        * storage/SQLTransaction.cpp:
        (WebCore::SQLTransaction::deliverQuotaIncreaseCallback): Changed to call the
        new exceededDatabaseQuota function instead of the old one.

        * svg/graphics/SVGImageEmptyClients.h: Updated for the change to ChromeClient.


        Reviewed by John Sullivan.

        - updated for changes to database functions

        * WebCoreSupport/ChromeClientGtk.cpp:
        * WebCoreSupport/ChromeClientGtk.h:
        * WebView/webkitprivate.cpp:


        Reviewed by John Sullivan.

        - fix <rdar://problem/5644324> Delegate-less WebKit clients should have no databases
        - add a missing export of WebDatabaseExpectedSizeKey
        - implement deleteOrigin: and remove deleteDatabasesWithOrigin:

        * Storage/WebDatabaseManager.mm:
        (-[WebDatabaseManager detailsForDatabase:withOrigin:]): Updated to check for a null
        name instead of calling isValid().
        (-[WebDatabaseManager deleteOrigin:]): Implemented.
        (WebKitInitializeDatabasesIfNecessary): Updated for name change.

        * Storage/WebDatabaseManagerPrivate.h: Removed deleteDatabasesWithOrigin:.

        * WebCoreSupport/WebChromeClient.h: Updated for changes to ChromeClient.
        * WebCoreSupport/WebChromeClient.mm:
        (WebChromeClient::exceededDatabaseQuota): Replaced the two different client functions
        we had before with a single one.

        * WebKit.exp: Added missing export for WebDatabaseExpectedSizeKey.

        * WebView/WebPreferenceKeysPrivate.h: Removed WebKitDefaultDatabaseQuotaKey.
        * WebView/WebPreferences.m:
        (+[WebPreferences initialize]): Removed the default for WebKitDefaultDatabaseQuotaKey.
        * WebView/WebPreferencesPrivate.h: Removed defaultDatabaseQuota and

        * WebView/WebUIDelegatePrivate.h: Replaced the two different database quota delegate
        methods we had before with a single one.

        * WebView/WebView.mm:
        (-[WebView _preferencesChangedNotification:]): Removed the code to set the
        default database origin quota in WebCore::Settings based on WebPreferences.

        * WebView/WebViewInternal.h: Removed delegate method dispatch functions for unusual
        types of parameters that the database UI delegate methods had before.


        Reviewed by John Sullivan.

        - updated for changes to ChromeClient database functions

        * WebCoreSupport/ChromeClientQt.cpp:
        * WebCoreSupport/ChromeClientQt.h:


        Reviewed by John Sullivan.

        - fix <rdar://problem/5644324> Delegate-less WebKit clients should have no databases
        - fix incorrect value for WebDatabaseDisplayNameKey in header

        * Interfaces/IWebDatabaseManager.idl: Renamed detailsForDatabaseWithOrigin to
        detailsForDatabase. Replaced deleteDatabasesWithOrigin with deleteOrigin.
        Renamed deleteDatabaseWithOrigin to deleteDatabase. Fixed incorrect value for
        * Interfaces/IWebPreferencesPrivate.idl: Removed defaultDatabaseQuota and
        * Interfaces/IWebUIDelegatePrivate.idl: Replaced the two database quota functions
        with a single one.

        * WebChromeClient.cpp:
        (WebChromeClient::exceededDatabaseQuota): Replaced the two database quota functions
        with a single one.
        * WebChromeClient.h: Ditto.

        * WebDatabaseManager.cpp:
        (WebDatabaseManager::detailsForDatabase): Changed to check for a name of null to
        detect a nonexistent database rather than using isValid.
        (WebDatabaseManager::deleteOrigin): Updated for name change (actually a semantic
        change too, but both have the same name).
        (WebDatabaseManager::deleteDatabase): Updated for name change.
        (WebKitSetWebDatabasesPathIfNecessary): Ditto.

        * WebDatabaseManager.h: Updated for name changes.

        * WebPreferenceKeysPrivate.h: Removed WebKitDefaultDatabaseQuotaKey.
        * WebPreferences.cpp:
        (WebPreferences::initializeDefaultSettings): Removed the default for
        * WebPreferences.h: Removed defaultDatabaseQuota and setDefaultDatabaseQuota.
        * WebView.cpp:
        (WebView::notifyPreferencesChanged): Removed the code to set the
        default database origin quota in WebCore::Settings based on IWebPreferencesPrivate.


        Reviewed by John Sullivan.

        - updated for changes to ChromeClient database functions

        * WebKitSupport/ChromeClientWx.cpp:
        * WebKitSupport/ChromeClientWx.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29698 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoRubber stamped by Oliver.
zimmermann@webkit.org [Mon, 21 Jan 2008 21:52:09 +0000 (21:52 +0000)]
Rubber stamped by Oliver.
Add missing layout test results.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29697 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoWebCore:
hyatt@apple.com [Mon, 21 Jan 2008 21:19:30 +0000 (21:19 +0000)]

        Fix for http://bugs.webkit.org/show_bug.cgi?id=16935

        Fix a bug in layers where the positioned ancestor was being computed incorrectly (the root should not
        automatically be included).

        Reviewed by olliej

        Added fast/layers/positioned-inside-root-with-margins.html

        * rendering/RenderLayer.cpp:



        Reviewed by olliej

        * fast/layers/positioned-inside-root-with-margins.html: Added.
        * platform/mac/fast/layers/positioned-inside-root-with-margins-expected.checksum: Added.
        * platform/mac/fast/layers/positioned-inside-root-with-margins-expected.png: Added.
        * platform/mac/fast/layers/positioned-inside-root-with-margins-expected.txt: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29696 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago Adding media/{audio,video}-controls-rendering.html back to the skipped list...
adele@apple.com [Mon, 21 Jan 2008 19:00:08 +0000 (19:00 +0000)]
    Adding media/{audio,video}-controls-rendering.html back to the skipped list, since they're still failing on some machines.

        * platform/win/Skipped:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29695 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years ago Reviewed by Mitz and Adam.
darin@apple.com [Mon, 21 Jan 2008 16:59:24 +0000 (16:59 +0000)]
    Reviewed by Mitz and Adam.

        * Scripts/svn-create-patch: Sort ChangeLog files first. Also slightly improved the
        sorting speed by doing all sort criteria in a single pass instead of three sorts.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29694 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoBenjamin Meyer <bmeyer@trolltech.com>
hausmann@webkit.org [Mon, 21 Jan 2008 11:06:20 +0000 (11:06 +0000)]
Benjamin Meyer <bmeyer@trolltech.com>

Change QWebHistoryInterface::addHistoryEntry() from const to non-const

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29693 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoAdam Treat <treat@kde.org>
hausmann@webkit.org [Mon, 21 Jan 2008 11:06:06 +0000 (11:06 +0000)]
Adam Treat <treat@kde.org>

* Don't leak the d-pointer in QWebSettings.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29692 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoSimon Hausmann <hausmann@webkit.org>
hausmann@webkit.org [Mon, 21 Jan 2008 10:47:15 +0000 (10:47 +0000)]
Simon Hausmann <hausmann@webkit.org>

Update the WebKit version number from WebKit/win/WebKit.vcproj/VERSION.

This has the fortunate side-effect that gmail sends us sensible HTML/JS again :)

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29691 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoFix KURL to QUrl conversion.
hausmann@webkit.org [Mon, 21 Jan 2008 10:41:43 +0000 (10:41 +0000)]
Fix KURL to QUrl conversion.

Stick to the encoded version of the URL and in addition to KURL we encode the
characters mentioned in section 2.4.3 of RFC 2396 as QUrl requires these, too.

This fixes fast/css/import-rule-regression-11590.html,
fast/css/import-style-update.html, svg/hixie/processing-model/003.xml and

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29690 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoFix focus chain handling and cycling through focusable objects (links) using tab...
hausmann@webkit.org [Mon, 21 Jan 2008 10:40:53 +0000 (10:40 +0000)]
Fix focus chain handling and cycling through focusable objects (links) using tab/backtab.

* Fix GraphicsContext::drawFocusRing to also draw single focus rects.
* Implemented QWebPage::focusNextPrevChild by sending fake tab/shift-tab events
  and make the return value depend on whether we successfully determined a focusable
  node or not.
* Changed QWebView::focusNextPrevChild() to call the base QWidget implementation correctly
  if we could not handle the focus chain ourselves.
* Changed the focus policy of QWebView to correctly use WheelFocus instead of ClickFocus.
* Made ChromeClientQt::canTakeFocus() and takeFocus() dummy method since they are only
  used to control the situation of stepping out of the focus chain inside the page.
* Made inclusion of links in the focus chain configurable through QWebSettings::LinksIncludedInFocusChain.
  The layout tests expect this to be disabled but for the user it seems sensible to have it
  on by default, hence the default in qwebsettings.cpp

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29689 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoFix access key support and fast/forms/legend-access-key.html
hausmann@webkit.org [Mon, 21 Jan 2008 10:40:16 +0000 (10:40 +0000)]
Fix access key support and fast/forms/legend-access-key.html

SVN revision 26664 changed the default access key for the non-mac build to Alt
while this test relies on Ctrl.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29688 268f45cc-cd09-0410-ab3c-d52691b4dbfc

14 years agoFix error reporting when parsing X(HT)ML fragments.
hausmann@webkit.org [Mon, 21 Jan 2008 10:39:57 +0000 (10:39 +0000)]
Fix error reporting when parsing X(HT)ML fragments.

We use the regular XMLTokenizer write()/.../end() sequence to parse those
fragments, but we should not report any parsing errors inline inside the
content. Instead we should just return the error and the DOM layer will take
care of throwing an exception.

This fixes fast/innerHTML/innerHTML-changing-document-properties.xhtml

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29687 268f45cc-cd09-0410-ab3c-d52691b4dbfc